Anda di halaman 1dari 437

DENSIFICATION OF BIOMASS

A DISSERTATION
SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL
OF THE UNIVERSITY OF MINNESOTA
BY

NALLADURAI KALIYAN

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS


FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

PROF. R. VANCE MOREY, Adviser

MARCH 2008
UMI Number: 3302306

INFORMATION TO USERS

The quality of this reproduction is dependent upon the quality of the copy
submitted. Broken or indistinct print, colored or poor quality illustrations and
photographs, print bleed-through, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.

UMI
UMI Microform 3302306
Copyright 2008 by ProQuest LLC.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.

ProQuest LLC
789 E. Eisenhower Parkway
PO Box 1346
Ann Arbor, Ml 48106-1346
Nalladurai Kaliyan 2008
ACKNOWLEDGMENTS

I am greatly indebted to my adviser, Professor R. Vance Morey, for giving me


this great opportunity to pursue my Ph.D. in the Bioproducts and Biosystems
Engineering (BBE) Department, University of Minnesota. I thank Dr. Morey for his
excellent advice, financial as well as personal support, friendliness, and patience. Dr.
Morey provided very good learning and working environments, and was involved with
each and every aspect of this thesis work, especially during the pilot-scale briquetting
and pelleting studies. I learned from Dr. Morey how to think in new ways and perfect
my work.

My deepest gratitude to Professor William F. Wilcke (BBE) for serving on the


preliminary and final oral exam committees. Dr. Wilcke was my M.S. adviser. Many
thanks to Dr. Wilcke for providing constructive comments and suggestions on this
thesis. Also, I extend my thanks to Dr. Wilcke for his friendliness. I am very much
thankful to Professor R. Roger Ruan (BBE) for serving on the preliminary and final
oral exam committees. I greatly appreciate Dr. Ruan's suggestions on the thesis, in
particular for studying the micro-structural analyses. My sincere thanks to Professor
William T. Y. Tze (BBE) for serving on the final oral exam committee, for his
suggestions on this thesis, and for providing assistance with interpreting DSC data on
the glass transition temperatures. I would like to thank Professor Christopher W.
Macosko (Department of Chemical Engineering and Materials Science) for serving on
the preliminary oral exam committee. Also, many thanks to Dr. Macosko for
providing insightful comments and suggestions on the thesis. Dr. Macosko's Rheology
course greatly helped me in the development of constitutive modeling for this thesis. I
would like to thank Dr. Shri Ramaswamy (BBE) for serving on the preliminary
written exam committee and for his suggestions on this thesis.

I wish to acknowledge the Initiative for Renewable Energy and the


Environment (IREE), University of Minnesota for providing financial assistance.
Also, I would like to acknowledge West Central Research and Outreach Center
i
(WCROC), University of Minnesota, Morris, MN for supplying biomass samples;
Bepex International LLC, Minneapolis, MN for proving access to their facility and
assistance with the roll press briquetting study; Agricultural Utilization Research
Institute (AURI), Waseca, MN for providing access to their facility and assistance
with the hammer milling of biomass feedstocks and pelleting study; Uniscope Inc.,
Johnstown, CO for supplying the chemical feed binders at no cost; and Chippewa
Valley Ethanol Company LLC, Benson, MN for providing the corn DDGS samples.

Many thanks to Dr. Greg Cuomo (WCROC) and Mr. Mike Reese (WCROC)
for their assistance and hospitality whenever my adviser and I went to WCROC for
collecting biomass samples. I would like to thank Mr. Michael White (Bepex), Mr.
Ron Halland (Bepex), and Mr. Mark Koester (Bepex) for helping with the roll press
briquetting study. A special appreciation to Mr. White for sharing his experience and
knowledge on the roll press briquetting. I would like to thank Mr. Alan Doering
(AURI) for his assistance on the purchase of biomass for the pilot-scale studies, and
the pelleting study at AURI. Also, I extend my thanks to Mr. Doering for sharing his
experience and knowledge on pelleting. My special thanks to Matthew De Kam (BBE)
for his assistance with the pilot-scale briquetting and pelleting studies. Also, many
thanks to Matt for all the help and friendship during my Ph.D. program.

My thanks are due to Dr. Robert Seavey (BBE) and Laura Moya (BBE) for
their assistance with the INSTRON universal testing machine. I wish to thank Mr.
Verlyn Johnson (BBE) for his help with the purchases and instrumentation required
for the lab-scale densification studies. Thanks a lot to Mr. Geoff Harms (St. Paul
Apparatus Shop) for his assistance for making the lab-scale densification apparatuses.
Thanks to Mr. Gilbert Ahlstrand (Imaging Center) for his assistance with SEM
imaging. I would like to thank the University of Minnesota for providing wonderful
library facilities and services, especially the Interlibrary Loan service.

I would like to thank BBE professors Dr. Kevin Janni, Dr. Mrinal
Bhattacharya, Dr. Jonathan Chaplin, Dr. Bruce Wilson, Dr. John Nieber, Dr. Philip
ii
Goodrich, and Dr. Charles Clanton for their help, mentorships, and friendliness.
Thanks very much to BBE scientists and staff, particularly Dr. Blanca Martinez, Ms.
Deborah Hansen, Ms. Kathryn Younie, Ms. Kathleen Evans, Ms. Dorit Hafner, Ms.
Barbara Oliver, Ms. Michele Schermann, Dr. Aida Mendez, Mr. Bradley Hansen, and
Mr. Brian Hetchler for all of the help provided by them during my graduate studies in
the BBE department.

I would like to thank Dr. Sudhagar Mani (University of Georgia) for his
suggestions while designing the lab-scale densification apparatus and for providing
valuable comments on our densification results published in ASABE meeting papers.
Many thanks to Dr. S. Sokhansanj (University of British Columbia, Canada), Dr. L.G.
Tabil (University of Saskatchewan, Canada), Dr. O.O. Fasina (Auburn University,
Alabama), and Dr. Klein Ueleji (Purdue University) for their comments and
appreciations on our densification studies when we presented our research at the
ASABE Annual International Meetings. Also, many thanks to my undergraduate
professors Dr. K. Alagusundaram, Dr. V. Kumar, Dr. S. Santhana Bosu, and Dr. S.
Ganapathy for their blessings, friendships, and moral support from India.

I would like to thank my BBE friends Brian Ashman, Dr. Fei Yu, Shaobo
Deng, Dr. Udai Singh, Omar Ibne Abdul Aziz, Kevin Hennessy, Reid Pulley, and Dr.
Caleb Arica for their help and warm support. Also, I would like to express my thanks
to my friends K. Navaneetha Carman (Virginia), Dr. Kathiravan Krishnamurthy
(Alabama), P. Gayathri (Manitoba, Canada), Dr. Jacob John (Twin Cities, MN), and
Dr. Ranjit Bhagyam (Twin Cities, MN) for their moral support.

I would like to thank my mother, Sellam Kaliyan, and my brother, Nagarajan


Kaliyan, for their love, support, and encouragement from India.

Finally, I would like to thank God for being with me and for helping me
successfully complete my Ph.D.

iii
DEDICATION

My hard work involved in this dissertation is dedicated to my grandfather,


Rengan Umaiyan, and my father, Kaliyan Rengan, who live in my memories.

IV
ABSTRACT

Corn stover and switchgrass are potential feedstocks for bio-energy and bio-
products industries. These biomass materials are often collected from the field as bales
with bulk densities of 100 to 200 kg m" . Because of low bulk density, corn stover and
switchgrass are difficult to handle, transport, store, and use in their natural forms.
Densification of corn stover and switchgrass can greatly reduce the problems and costs
of handling, transportation, and storage. Therefore, the overall objective of this
research was to study the densification mechanisms of corn stover and switchgrass
through laboratory-scale and pilot-scale densification experiments, and constitutive
modeling.

Using a laboratory scale piston-cylinder densification apparatus, the effects of


compression pressure, moisture content, particle size, preheating temperature, and
commercial binders on the densification behaviors of corn stover and switchgrass were
studied. Corn stover and switchgrass briquettes (19 to 20-mm diameter) with unit
densities of 417 to 1161 kg m"3 were produced. The briquettes produced had
maximum durability (tumbling can method) of about 97% for corn stover, and about
80% for switchgrass.

Using a pilot-scale roll press briquetting machine, briquettes (almond shape


with 28.7 to 31.3-mm in length) with bulk densities of 351 to 527 kg m" were
produced. Using a pilot-scale ring-die pelleting machine, pellets (cylindrical shape
with 9.6 to 9.8-mm diameter) with bulk densities of 534 to 616 kg m" were produced.
The energy consumption for briquetting and pelleting ranged from 1 to 4% of the
energy in the biomass material. The roll press briquetting appears to be a promising
low-cost, low-energy, high-capacity approach for commercial production of biomass
briquettes.

v
Commercially, high quality briquettes and pellets could be produced from corn
stover and switchgrass without adding chemical binders (i.e., additives) by activating
(softening) the natural binding components such as water soluble carbohydrates,
lignin, protein, starch, and fat in the corn stover and switchgrass through moisture [15-
20% (w.b.) for corn stover, and 8-15% (w.b.) for switchgrass] and temperature [no
preheating for corn stover (25C), and preheating to the glass transition temperature of
switchgrass (>75C)].

An elasto-visco-plastic solid (constitutive) model adequately characterized the


compression/densification behaviors of corn stover and switchgrass grinds.

VI
TABLE OF CONTENTS

Page
ACKNOWLEDGMENTS i
DEDICATION iv
ABSTRACT v
TABLE OF CONTENTS vii
LIST OF TABLES xv
LIST OF FIGURES xx
NOMENCLATURE xxv

Chapter 1 INTRODUCTION AND OBJECTIVES 1


1.1 Problem Statement 1
1.2 Objectives 4
1.3 Thesis Outline 5
1.4 References 6

Chapter 2 REVIEW OF LITERATURE 7


2.1 Biomass Feedstocks 7
2.1-1 Biomass Harvesting and Collection 8
2.1.1-1 Corn stover 8
2.1.1-2 Switchgrass 10
2.1-2 Biomass Storage 11
2.1-3 Biomass Transportation 13
2.1-4 Biomass Drying 16
2.1.4-1 Field drying 16
2.1.4-2 Artificial drying 17
2.1-5 Size Reduction 19
2.1-6 Densification 22
2.1.6-1 Densified products 23

vii
2.1.6-2 Densification machines 25
2.1.6.2-1 Screw extruder 25
2.1.6.2-2 Piston press 26
2.1.6.2-3 Pelleting machine 26
2.1.6.2-4 Forage cubers 28
2.1.6.2-5 Roll press briquetting machine 29
2.1.6.2-6 Roll compaction / dry granulation 30
2.1.6.2.6-1 Scale-up Procedure for Roll
Press 31
2.1.6.2-7 Mobile densification machines 32
2.2 Laboratory Scale Densification 34
2.3 Process Variables Affecting Densification 35
2.3-1 Effect of Pressure 35
2.3-2 Effect of Moisture Content 36
2.3-3 Effect of Particle Size 37
2.3-4 Effect of Temperature 37
2.3-5 Effect of Binders 38
2.4 Specific Energy Consumption for Densification 39
2.5 Densification Cost Economics 40
2.6 Bonding Mechanisms 43
2.6-1 Wood 43
2.6-2 Animal Feed Pellets 44
2.6-3 Biomass Materials 45
2.6.3-1 Biomass compositions 47
2.7 Commercial Binders and Lubricants 48
2.8 Strength of Agglomerates 50
2.9 Conclusions 51
2.10 References 52

viii
Chapter 3 STRENGTH AND DURABILITY OF DENSIFIED
PRODUCTS 66
3.1 Abstract 66
3.2 Introduction 66
3.3 Objectives 69
3.4 Binding Mechanisms 70
3.4-1 Solid Bridges 70
3.4-2 Attraction Forces between Solid Particles 70
3.4-3 Mechanical Interlocking Bonds 71
3.4-4 Adhesion and Cohesion Forces 71
3.4-5 Interfacial Forces and Capillary Pressure 71
3.5 Measurement of Strength and Durability 72
3.5-1 Compressive Resistance 72
3.5-2 Durability 73
3.5-3 Impact Resistance 76
3.5-4 Water Resistance 77
3.6 Factors Affecting Strength and Durability 78
3.6-1 Effect of Feed Constituents 78
3.6.1-1 Starch 80
3.6.1-2 Protein 80
3.6.1-3 Fiber 81
3.6.1-4 Fat/oil 82
3.6.1-5 Lignin and extractives 83
3.6.1-6 Other feed constituents 83
3.6-2 Effect of Feed Moisture Content 83
3.6-3 Effect of Feed Particle Size 86
3.6-4 Effect of Steam Conditioning / Preheating 87
3.6-5 Effect of Binders / Additives 91
3.6-6 Effect of Mixing of Feed / Biomass Materials 93
3.6-7 Densification Equipment Variables 95

IX
3.6.7-1 Effect of pressure 95
3.6.7-2 Pellet mill variables 97
3.6.7.2-1 Die dimension 97
3.6.7.2-2 Die speed 98
3.6.7.2-3 Gap between roller and die 99
3.6.7.2-4 Specific energy input to the pellet
mill 99
3.6.7.2-5 Double pelleting 100
3.6.7.2-6 Expander 100
3.6.7.2-7 Ripener 101
3.6.7-3 Roll press variables 102
3.6-8 Effect of Post-Production Conditions 103
3.6.8-1 Time of measurement 103
3.6.8-2 Cooling / drying 104
3.6.8-3 Storage conditions 105
3.7 Acceptance Levels for Strength and Durability 107
3.8 Discussion 109
3.9 Conclusions 112
3.10 References 113

Chapter 4 DENSIFICATION CHARACTERISTICS OF CORN


STOVER AND SWITCHGRASS - LABORATORY
SCALE STUDY 137
4.1 Abstract 137
4.2 Introduction 137
4.3 Objectives 138
4.4 Materials and Methods 139
4.4-1 Biomass Samples 139
4.4-2 Briquette Compression and Ejection Apparatuses 140
4.4-3 Briquetting Procedure 140

x
4.4-4 Preheating of Biomass Grinds 141
4.4-5 Differential Scanning Calorimeter Experiments 143
4.4-6 Briquetting Experiments 144
4.4-7 Briquette Properties 144
4.4-8 Statistical Analysis 145
4.5 Results and Discussion 147
4.5-1 Biomass Grinds 147
4.5-2 Glass transition Temperature 147
4.5-3 Densification Process Variables 149
4.5.3-1 Effect of pressure 150
4.5.3-2 Effect of moisture content 152
4.5.3-3 Effect of particle size 153
4.5.3-4 Effect of preheating temperature 154
4.6 Summary and Conclusions 159
4.7 References 160
4.8 Appendix 185

Chapter 5 STRATEGIES TO IMPROVE DURABILITY OF


SWITCHGRASS BRIQUETTES - LABORATORY SCALE
STUDY 217
5.1 Abstract 217
5.2 Introduction 217
5.3 Objectives 219
5.4 Materials and Methods 220
5.4-1 Biomass and Binders 220
5.4-2 Briquetting Procedure 221
5.4-3 Briquetting Experiments 223
5.4-4 Briquette Properties 225
5.5 Results and Discussion 226
5.5-1 Effect of Fine Grinding 226

xi
5.5-2 Effect of Biomass Based Binders 226
5.5-3 Effect of Chemical Binders 228
5.5-4 Strategy to Improve Durability of Switchgrass 229
Briquettes
5.5-5 Implications of Briquette Ejection Curves 231
5.6 Conclusions 234
5.7 References 235

Chapter 6 ROLL PRESS BRIQUETTING AND PELLETING -


PILOT SCALE STUDY 251
6.1 Abstract 251
6.2 Introduction 251
6.3 Objectives 254
6.4 Materials and Methods 254
6.4-1 Biomass Feedstock 254
6.4-2 Roll Press Briquetting 255
6.4-3 Pelleting 257
6.4-4 Machine and Biomass Feedstock Variables Studied 259
6.4-5 Properties of Briquettes and Pellets 261
6.5 Results and Discussion 263
6.5-1 Characteristics of Biomass Feedstock 263
6.5-2 Roll Press Briquetting 265
6.5.2-1 Effect of roll press briquetting machine 265
variables
6.5.2-2 Effect of particle size 267
6.5.2-3 Effect of moisture content 268
6.5.2-4 Effect of steam conditioning temperature 269
6.5-3 Pelleting 271
6.5.3-1 Effect of pelleting machine variables 271
6.5.3-2 Effect of biomass feedstock variables 273

xii
6.5-4 Comparison of Roll Press Briquetting with Pelleting 275
6.5-5 Binding Mechanisms of Corn Stover and Switchgrass 277
6.6 Summary and Conclusions 278
6.7 References 281
6.8 Appendix 307

Chapter 7 CONSTITUTIVE MODELS FOR DENSIFICATION OF


BIOMASS 312
7.1 Abstract 312
7.2 Introduction 312
7.2-1 Densification Mechanism 313
7.2-2 Constitutive Models 314
7.2.2-1 Empirical models 315
7.2.2-2 Rational models 317
7.2-3 Rheological Models 318
7.3 Objectives 322
7.4 Materials and Methods 322
7.4-1 Constitutive Model Development 322
7.4-2 Constitutive Model Parameter Estimation 325
7.4-3 Biomass Compression Experimental Procedure 326
7.4-4 Biomass Compression Conditions 328
7.5 Results and Discussion 329
7.6 Conclusions 334
7.7 References 334
7.8 Appendix 363

Chapter 8 SUMMARY AND OVERALL CONCLUSIONS 365

Chapter 9 RECOMMENDATIONS FOR FUTURE WORK 370

xiii
Appendix A EFFECT OF ENHANCED DEAERATION DURING
DENSIFICATION OF CORN STOVER AND
SWITCHGRASS - LABORATORY SCALE STUDY 373
A.l Abstract 373
A.2 Introduction 373
A.3 Objectives 376
A.4 Materials and Methods 377
A.4-1 Biomass Samples 377
A.4-2 Briquetting Procedure 377
A.4-3 Deaeration Without Steel Filter 379
A.4-4 Deaeration With Steel Filter 380
A.4-5 Briquetting Experiments 381
A.4-6 Briquette Properties 382
A.5 Results and Discussion 382
A.5-1 Corn Stover 382
A.5-2 Switchgrass 383
A.6 Conclusions 385
A.7 References 385

Bibliography 392

xiv
LIST OF TABLES

Table Caption Page


3.1 Pellet durability index (%) of various types of commercial feed
pellets determined by four durability measurement methods
(Winowiski, 1998). 124
3.2 Effect of fat added to the feed before pelleting on the pellet
durability. 124
3.3 Effect of lignin and extractives content on the pellet durability
(Bradfield and Levi, 1984). 125
3.4 Recommended feed particle size distribution to produce high
quality pellets (Payne, 2006). 125
3.5 Effect of feed particle size on the pellet durability. 126
3.6 Effect of steam conditioning on the quality of pellets made from
corn-soybean meal based poultry layer diet (Skoch et al., 1981)
w
. 126
3.7 Effect of binders (additives) on the pellet quality (Tabil et al.,
1997). 127
3.8 Effect of mixing of different feed materials on the
pellet/briquette durability. 128
3.9 Quality of water hyacinth briquettes made from different
densification equipment (Koser et al., 1982). 129
3.10 Effect of pressure on the physical quality of densified products. 130
3.11 Effect of pellet mill die-size on starch gelatinization (measured
after pelleting) and pellet durability for a poultry layer ration
(Heffner and Pfost, 1973). 130
3.12 Effect of expander conditioning on the pellet quality (van der
Poeletal, 1997). 131
3.13 Minimum feed pellet quality recommended for feed mills
(Major, 1984)[a]. 131

xv
3.14 Generalized optimum densification conditions [a]. 132-135
3.15 Tested strategies to improve pellet durability (Winowiski, 136
2006).
4.1 Properties of corn stover and switchgrass grinds. 173
4.2 Glass transition temperatures of corn stover, switchgrass, and
lignin. 173
4.3a Unit density and durability of corn stover briquettes. 174-175
4.3b Unit density and durability of switchgrass briquettes. 176-177
4.4a Stability and moisture content of corn stover briquettes
measured after one week of storage at room temperature. 178
4.4b Stability and moisture content of switchgrass briquettes
measured after one week of storage at room temperature. 179
4.5a Specific energy required for briquetting corn stover. 180
4.5b Specific energy required for briquetting switchgrass. 181
4.6a Properties of corn stover and switchgrass briquettes (pressure -
150 MPa). 182-183
4.6b Specific energy required for briquetting corn stover and
switchgrass (pressure = 150 MPa). 184
4.A1 ANOVA using univariate general linear model (GLM) on the
unit density of corn stover briquettes. 185-186
4.A2 ANOVA using univariate GLM on the durability of corn stover
briquettes. 187-188
4.A3 ANOVA using univariate GLM on the unit density of
switchgrass briquettes. 189-190
4. A4 ANOVA using univariate GLM on the durability of switchgrass
briquettes. 191-192
4.A5 One-way ANOVA followed by Tukey's procedure (T-test) to
investigate the significant effect of treatment factors within a
group of tests. 193-197
4.A6 Independent samples T-test on the unit density or durability of

xvi
corn stover briquettes (moisture conditioning versus
preheating). 197-203
4.A7 Independent samples T-test on the unit density or durability of
switchgrass briquettes (moisture conditioning versus
preheating). 203-208
4.A8 Independent samples T-test to determine the effect of particle
size on corn stover briquette density or durability when
moisture conditioning to 15 to 20% (w.b.). 209-210
4.A9 Statistical analyses (one-way ANOVA followed by T-test) to
determine the effect of moisture content on switchgrass
briquette density or durability at the preheating temperature of
100C. 211-213
4.A10 Statistical analyses (one-way ANOVA followed by T-test) to
determine the effect of particle size on switchgrass briquette
density or durability at the preheating temperature of 100C. 214-216
5.1 Properties of switchgrass and corn stover grinds. 241
5.2 Compositions of corn stover, switchgrass, and corn Distillers
Dried Grains with Solubles (DDGS). 242
5.3 Cost of the binders used for the study. 243-244
5.4 Properties of switchgrass briquettes (maximum compression
pressure = 150 MPa; preheating temperature = 100C). 245-246
5.5 Stability and moisture content of switchgrass briquettes
measured after one week of storage at room temperature
(nominal grind moisture content = 10% (w.b.); maximum
compression pressure =150 MPa; preheating temperature =
100C). 247-248
5.6 Specific energy required for briquetting switchgrass (nominal
grind moisture content = 10% (w.b.); maximum compression
pressure = 150 MPa; preheating temperature = 100C). 249-250
6.1 Properties of corn stover and switchgrass grinds used for roll

xvn
press briquetting and pelleting experiments. 294
6.2 Compositions of corn stover and switchgrass grinds. 295
6.3 Specifications of the roll press briquetting machine. 296
6.4 Specifications of the pelleting machine. 297
6.5 Properties of corn stover briquettes measured immediately after
making the briquettes (mean standard deviation, if given). 298
6.6 Properties of corn stover briquettes measured after one week of
storage at room temperature (mean standard deviation, if
given). 299
6.7 Properties of switchgrass briquettes measured immediately after
making the briquettes (mean standard deviation, if given). 300
6.8 Properties of switchgrass briquettes measured after one week of
storage at room temperature (mean standard deviation, if
given). 301
6.9 Specific energy consumption for the roll press briquetting
process (mean standard deviation, if given). 302
6.10 Properties of corn stover and switchgrass pellets measured
immediately after making the pellets (mean standard
deviation, if given). 304
6.11 Properties of corn stover and switchgrass pellets measured after
one week of storage at room temperature (mean standard
deviation, if given). 305
6.12 Specific energy consumption for the pelleting process (mean
standard deviation, if given). 306
6.A1 SEM images of corn stover and switchgrass grinds (before roll
press briquetting or pelleting). 307
6.A2 SEM images of corn stover briquettes made using roll press
briquetting machine. 308
6. A3 SEM images of switchgrass briquettes made using roll press
briquetting machine. 309

xviii
6.A4 SEM images of corn stover pellets made using ring-die
pelleting machine (no preheating/steam conditioning was
involved). 310
6.A5 SEM images of switchgrass pellets made using ring-die
pelleting machine (no preheating/steam conditioning was
involved). 311
7.1 Properties of corn stover and switchgrass briquettes. 347
7.2 Constitutive model parameters for uniaxial compression of corn
stover and switchgrass. 349
7.3 Effects of densification conditions on constitutive model
parameters for corn stover. [The model parameters (mean of
three replications) are plotted as a function of maximum stress
involved for estimating the model parameters at various stress
ranges.] 357
7.4 Effects of densification conditions on constitutive model
parameters for switchgrass. [The model parameters (mean of
three replications) are plotted as a function of maximum stress
involved for estimating the model parameters at various stress
ranges.] 360
7.A1 Example calculation showing estimation of model parameters
for the measured data in the stress range of 100 to 150 MPa. 363
A.l Properties of corn stover and switchgrass briquettes. 389
A.2 Moisture content and stability of corn stover and switchgrass
briquettes measured after one week of storage at room
temperature. 390
A.3 Specific energy required for briquetting corn stover and
switchgrass. 391

xix
LIST OF FIGURES

Figure Caption Page


1.1 Energy consumption in the U.S. in 2004 (EIA, 2005). 1
2.1 Densification of biomass for efficient and economic use of
biomass feedstocks in bio-based industries. 63
2.2 A typical biomass densification process flow diagram. 63
2.3 Schematics of high pressure agglomeration machines (Pietsch,
1991). (a) Reciprocating piston and die press, (b) rotary
tableting machine, (c) ram extrusion or plunger press, (d) roller
compacting machine, and (e) roller briquetting machine. 64
2.4 Schematics of pelleting machines (Pietsch, 1991). (a) screw
extruder, (b) pelleting machine with flat-die and muller-type
press rollers, (c) pelleting machine with one solid and one
hollow roll, (d) pelleting machine with two hollow rolls, (e)
pelleting machine with ring-die and internal press roll, and (f)
gear-type pelleter. 65
4.1a Schematic of briquette compression apparatus (figure not to
scale). 164
4.lb Uniaxial, piston-cylinder densification (compression) apparatus.
Pressure is applied by the Instron universal testing machine. 164
4.2a Schematic of briquette ejection apparatus in operation (figure
not to scale). 165
4.2b Briquette ejection apparatus located below the piston-cylinder to
eject the briquette. Pressure is applied by the Instron universal
testing machine. 165
4.3a Modifications to the densification cylinder during preheating of
biomass grind. 166
4.3b Preheating of biomass grind using heating tapes (covered with
insulation). 166

xx
4.3c Preheating of biomass grind using heating tapes (insulation is
removed to show the heating tape). 167
4.3d (A): A specially designed cap (consisting of a pressure relief
valve, an O-ring seal, and a K-type thermocouple) used to cover
the top of the cylinder (die) during preheating; (B): Inclusion of
an O-ring seal in the bottom steel plate of the compression
apparatus during preheating and compression. 167
4.4 Typical briquette compression (A), and ejection (B) curves. 168
4.5 DSC thermograms for corn stover, switchgrass, and lignin. 168
4.6 Durability tester (tumbler or tumbling can) used for measuring
the durability of briquettes based on ASAE Standards S269.4
(2003c). 169
4.7a Corn stover grind obtained using a hammer mill screen size of
4.6 mm (geometric mean particle diameter = 0.80 mm). 170
4.7b Switchgrass grind obtained using a hammer mill screen size of
4.6 mm (geometric mean particle diameter = 0.64 mm). 170
4.8 Corn stover briquettes made at pressures of 30, 60, 100, 130 and
150 MPa (left to right) (geometric mean particle diameter = 0.80
mm; grind moisture content = 10% w.b.; preheating temperature
= 25C). 171
4.9 Corn stover briquettes made at 100 MPa pressure for five
replications (geometric mean particle diameter = 0.66 mm;
moisture content = 10% w.b.; preheating temperature = 25C;
durability = 58.3%). 171
4.10 Switchgrass briquettes made at 100 MPa pressure for five
replications (geometric mean particle diameter = 0.56 mm;
moisture content = 10% w.b.; preheating temperature = 25C;
durability = 0.0%). 171
4.11 Corn stover briquettes after durability test (pressure =150 MPa;
geometric mean particle diameter = 0.66 mm; moisture content

xxi
= 10% w.b.; preheating temperature - 75C; durability =
96.8%). 172
4.12 Switchgrass briquettes after durability test (pressure =150 MPa;
geometric mean particle diameter = 0.56 mm; moisture content
= 10% w.b.; preheating temperature = 100C; durability =
67.3%). 172
5.1 Compression curve obtained while forming switchgrass
briquettes. Compression pressure of 150 MPa was reached
earlier at the particle size of 0.26 mm than at the particle size of
0.56 mm. This is because the bulk density of the switchgrass
grind for the particle size of 0.26 mm was higher than for the
particle size of 0.56 mm (table 5.1). 238
5.2 Briquette ejection curves (A, B, and C) for switchgrass
briquettes made using binding agents at different inclusion rates
(% wt), and at a compression pressure of 150 MPa and 100C. 239-240
5.3 Effect of temperature on the briquette ejection pressure for
Super-bind binder mixed with switchgrass. 241
6.1 Schematics of conventional ring-die pelleting (A), and roll press
briquetting (B) machines. 285
6.2 Roll press briquetting machine (A: front view of the roll press
where one half of the conical cover for the screw feeder is
removed, and the arrows indicate the direction of operation. B:
side view of one of the rolls showing the briquette pockets.). 286
6.3 Flow diagram of roll press briquetting of corn stover and
switchgrass. 287
6.4 Corn stover and switchgrass briquettes produced. 288
6.5 Corn stover and switchgrass pellets produced. 288
6.6 Durability tester (tumbler or tumbling can) used for measuring
the durability of briquettes and pellets based on AS ABE
Standards S269.4 (2003c). 289

xxn
6.7 Corn stover briquettes (whole briquettes plus broken pieces of
briquettes) retained on a 3.2 mm (1/8 in.) screen after a
durability test. 290
6.8 Switchgrass briquettes (whole briquettes plus broken pieces of
briquettes) retained on a 3.2 mm (1/8 in.) screen after a
durability test. 291
6.9 Crushing strength testers for briquettes (A), and pellets (B). 291
6.10 SEM micrographs of corn stover and switchgrass grinds,
briquettes, and pellets (magnification at X600). 292
6.11 UV auto-fluorescence images of corn stover and switchgrass
grinds, briquettes, and pellets (magnification at XI45). Auto-
fluorescence color interpretation (Rost, 1995): green or yellow-
green for protein compounds; brilliant blue or bluish-white for
lignin, cutin, suberin, or phenolic acids such as ferulic acid; and
whitish fluorescence for cutin (cuticle). Pure carbohydrates
(cellulose, hemicellulose, starch, and lipid/fat molecules) do not
fluoresce (Rost, 1995). 293
7.1 Typical pressure-time curve for compression process (Pietsch,
2002). 340
7.2 Mechanical elements used to describe stress-strain behavior of
materials. Spring element represents an elastic solid, and the
dashpot element represents a viscous fluid, (a = stress; E =
modulus of elasticity; and r\ = coefficient of viscosity.) 341
7.3 Mechanical analogy of biomass grind for the development of
constitutive model for the compression process. This
mechanical analogy system represents an elasto-visco-plastic
solid. The dashpot viscous element (r\) represents the time
dependent loss, the strain hardening spring element (E, R, and
n) represents the elastic and plastic deformation, and the
Coulomb friction element (<7f) represents the time independent

xxiii
loss during uniaxial compression of biomass grind. 342
7.4 Identification of inertial deformation and elasto-visco-plastic
deformation stages from the compression curve of corn stover
or switchgrass grind. 343-344
7.5 A comparison of constitutive model versus measured data. In
addition to the total compression curve for 0 to 150 MPa, the
compression curves are shown for different stress range
segments. The model curve represents the Eq. 5 with the natural
< H ^
strain substituted as In , natural strain rate
KH-(Vxt);
f
V
substituted as and the model parameters (E, R, n,
yH-(Vxt)

r\, and oi) estimated for one replication data on the measured
stress, natural strain, and natural strain rate. [H = measured
initial height of the grind at time t = 0 (i.e., at the stress of 0
MPa) inside the die (m); V - the set speed of the INSTRON
crosshead (m s"1); t = elapsed time since the start of compression
345-346
(s)]
A. 1 Schematic of uniaxial densification apparatus showing
deaeration (i.e., air removal) paths during compression of
biomass grind. 388
A.2 Pictures of base plates (114.29-mm diameter x 11.74-mm thick)
with and without steel filter (21.01-mm diameter x 1.50-mm
thick) at their centers. (A) Solid base plate (no steel filter) for no
deaeration through the bottom of the die; (B) Base plate with
steel filter (top side) for deaeration through the bottom of the
die; (C) Base plate with steel filter (bottom side) showing four
3.25-mm diameter holes to allow air to escape. 388

xxiv
NOMENCLATURE

A = constant in Heckel (1961) equation, MPa"


B = constant in Heckel (1961) equation
D = roll diameter, cm
DM = dry matter
d.b. = dry basis
E = elastic modulus, MPa
H = initial height of the grind at time t = 0 inside the die, m m
h - roll gap, cm
L = roll width, cm
m = number of briquettes per revolution
M S R = mean square residuals
n = strain hardening exponent
nron = roll speed, min"1
P = applied pressure, MPa
R = strength coefficient or plastic modulus, MPa
2
R = coefficient of determination
SSR = sum of squared residuals
t = time, s
Vb = briquette volume, cm 3
w.b. = wet basis

s = natural strain
Sj = measured natural strain corresponding to the end of the inertial deformation
sp = measured natural strain at the first data points (i.e., knots) for a stress range
a = stress (i.e., applied pressure), MPa
a; = measured stress corresponding to the end of the inertial deformation, MPa
af = frictional loss factor, MPa
dp = measured stress at the first data points (i.e., knots) for a stress range, MPa
r| = viscous coefficient or coefficient of viscosity, MPa-s
AH = total displacement of the crosshead of the Instron recorded until time t, m m
Pb = briquette density, g cm" 3
pc = compact density, g cm"3
pr = relative density of the compacted material

xxv
Chapter 1
INTRODUCTION AND OBJECTIVES

1.1 Problem Statement

Biomass is any organic matter that is available on a renewable or recurring


basis including but not limited to agricultural crops and residues, forest residues, wood
and wood residues, plants (including aquatic plants), animal wastes, and municipal
residues. Biomass is generally produced by green plants in a sustainable manner from
water and carbon dioxide (CO2) by photosynthesis. The energy stored in the biomass
can be recovered by several conversion processes such as fermentation, combustion,
and gasification. In 2004, biomass contributed nearly 2.9 quadrillion Btu to the U.S.
energy supply, nearly 3% of the total U.S. energy consumption of about 100.3
quadrillion Btu (fig. 1.1).

Total = 100.278 Quadrillion Btu Total = 6.117 Quadrillion Btu

Natural Gas
-Solar 1 %
23%

-Biomass 47%
Renewable
Energy 3eothermal 6%
6%
Nuclear -Hydroelectric 45%
Energy
Power -Wind 2%
Petroleum 8%
40%

Figure 1.1. Energy consumption in the U.S. in 2004 (EIA, 2005).

Concerns about global climate change and air quality have increased interest in
biomass and other energy sources that are potentially C02-neutral and less polluting
(Cook and Beyea, 2000). Among the alternative renewable energy sources that are
seriously considered for large scale implementation (biomass, wind, solar and

1
geothermal), biomass is unique and attractive because (DOE, 2005): (i) biomass is the
only current renewable source of liquid transportation fuels and thus, use of biomass
can reduce oil imports - one of the nation's most pressing energy and security needs;
(ii) biomass has great potential to provide heat, steam, and power to industry, and to
heat residential and commercial building space; (iii) biomass can provide feedstock to
produce a wide variety of chemicals and materials or bioproducts; and (iv) increased
use of biomass can create jobs and enhance income in the rural sector.

Although using biomass feedstocks to produce liquid transportation fuels (e.g.,


ethanol), chemicals, heat and power has received greater attention in the U.S.,
utilization of biomass feedstocks will be challenging because of the inherent
characteristics of the biomass such as high moisture content and low bulk density. For
example, corn (Zea mays L.) stover may have moisture content in the range of 30 to
75% (w.b) after harvest (Shinners et al, 2003). The bulk density of corn stover bales
may range from 95 to 150 kg m" in the moisture content range of 23 to 40% (w.b)
(Shinners et al., 2003). Therefore, there is a need to upgrade the quality of the biomass
feedstocks through pre-processing such as drying, size reduction, and densification for
efficient utilization of biomass in the bioenergy and bioproducts industries.

About 80% of the agricultural residues in the U.S. are corn stover (Walsh et
al., 2005). Corn stover is the aboveground portion of the corn plant, including stalks,
leaves, cobs and husks, but excluding the corn kernels. Annually about 230 million
dry tons of corn stover can be collected in the U.S. (Walsh et al., 2005), which can be
used for producing ethanol, chemicals, biomaterials, process heat, and electricity.
Switchgrass (Panicum virgatum L.), a native perennial warm-season grass, has been
identified as a dedicated energy crop for producing bio-fuels and renewable energy
(ORNL, 2005; McLaughlin and Kszos, 2005). Use of biomass such as corn stover and
switchgrass in place of fossil fuels could reduce the nation's dependence on foreign
oil, reduce greenhouse gas emissions, increase farm income, and create job
opportunities in rural areas.

2
Corn stover and switchgrass can be collected only during a limited harvest
season and, therefore, they should be stored to feed the processing facilities year-
round. Corn stover and switchgrass are often stored in baled forms, which involve a
lot of handling, storage and transportation costs because of low bulk density. One of
the solutions to reducing storage and transportation costs is densification of the
biomass materials. Densification (pelleting, cubing, compaction, or briquetting) of
particulate matter is achieved by forcing the particles together by applying a
mechanical force to create inter-particle bonding, which makes well-defined shapes
and sizes such as pellets, cubes, and briquettes. Because the densified biomass will be
in more consistent form, and easier to handle and feed into the processing equipment,
densification of biomass materials can greatly reduce the problems and costs of
handling, transportation, and storage. Moreover, increasing feedstock density is
essential to successful adoption of biomass for a range of applications.

3
1.2 Objectives

The overall objective of this dissertation research is to study the densification


mechanisms of corn stover and switchgrass through experiments and modeling.
Understanding the densification mechanisms is important to produce high quality
pellets, briquettes, or compacted/granulated biomass with less inputs, energy, and cost.

The specific objectives of this research are to:

1. Study the effect of process variables (pressure, moisture content, particle size,
and preheating temperature) on the densification characteristics of corn stover
and switchgrass using a laboratory densification apparatus.
2. Study the effect of adding commercial binders on the densification
characteristics of switchgrass using a laboratory densification apparatus.
3. Study the densification behaviors of corn stover and switchgrass in pilot-scale,
commercial densification equipment (roll press briquetting and pelleting
machines) for the best densification conditions derived from the laboratory
studies.
4. Investigate the binding mechanisms of corn stover and switchgrass by micro-
structural analyses.
5. Develop constitutive models for densification of corn stover and switchgrass to
understand their densification mechanisms.

The densification process involves two other important unit operations, drying
and size reduction (i.e., grinding) of biomass, before pressing of biomass in
densification machines at predetermined moisture, particle size, and other
densification conditions. For this thesis work, only the densification process was
studied in detail. However, the literature was reviewed for the possible methods of
drying and size reduction of corn stover and switchgrass (Chapter 2).
4
1.3 Thesis Outline

The thesis is arranged such that each chapter stands alone to make it easier to
submit various parts for publication.

Chapter 2 provides the background and related literature for the densification
of biomass materials. The information from Chapter 2 assisted in identifying
commercial densification machines that can be readily adapted to densify the biomass
materials in the U.S. Chapter 3 consists of a concentrated literature review on the
factors influencing the strength and durability of densified biomass products. This
review helped understand and select various densification conditions to be evaluated
for corn stover and switchgrass in order to produce good quality densified products.

Chapter 4 details the laboratory experiments conducted to study the


densification characteristics of corn stover and switchgrass and to obtain process
conditions to produce high density, durable densified products (Objective 1). Chapter
5 presents the laboratory experiments conducted to study the effect of adding various
commercial binders on the durability of switchgrass briquettes (Objective 2). To check
the applicability of the results from the laboratory densification studies, densification
behaviors of corn stover and switchgrass were studied in pilot-scale roll press
briquetting and pelleting machines as detailed in Chapter 6 (Objective 3). Chapter 6
also includes the binding mechanisms identified due to the natural binders in the corn
stover and switchgrass through the micro-structural analyses (Objective 4).

Chapter 7 documents the densification mechanisms of corn stover and


switchgrass revealed by the constitutive models under various densification conditions
studied in the laboratory (Objective 5). Chapter 8 summarizes the results and
conclusions obtained from the laboratory-scale and pilot-scale densification studies,
and from the constitutive modeling of the densification/compression process. Chapter
9 lists the potential future research areas that may help for the energy- and cost-

5
effective production of high quality densified products from corn stover and
switchgrass. Appendix A demonstrates that the current design of the densification
apparatus along with the densification procedure used for the laboratory scale
densification experiments is sufficient to remove (i.e., deaerate) almost all of the inter-
particle air from the biomass grinds during compression/densification.

1.4 References

Cook, J., and J. Beyea. 2000. Bioenergy in the United States: progress and
possibilities. Biomass and Bioenergy 18(6): 441-455.
DOE. 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The
Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.
Department of Energy (DOE), Office of Scientific and Technical Information.
Available at: http://www.osti.gov/bridge. Accessed 18 June 2005.
EI A. 2005. Renewable Energy Trends 2004. Washington, DC: Energy Information
Administration. Available at: http://www.eia.doe.gov/. Accessed 20 October
2005.
McLaughlin, S.B., and L.A. Kszos. 2005. Development of switchgrass (Panicum
virgatum) as a bioenergy feedstock in the United States. Biomass and
Bioenergy 28(6): 515-535.
ORNL. 2005. Biofuels from Switchgrass: Greener Energy Pastures. Oak Ridge, TN:
Oak Ridge National Laboratory (ORNL). Available at:
http://bioenergy.ornl.gov/papers/misc/switgrs.html. Accessed 18 June 2005.
Shinners, K.J., B.N. Binversie, and P. Savoie. 2003. Harvest and storage of wet and
dry corn stover as a biomass feedstock. ASAE Paper No. 036088. St. Joseph,
MI: ASABE.
Walsh, M.E., R.L. Perlack, A. Turhollow, D.T. Ugarte, D.A. Becker, R.L. Graham,
S.E. Slinsky, and D.E. Ray. 2005. Biomass Feedstock Availability in the
United States: 1999 State Level Analysis. Oak Ridge, TN: Oak Ridge National
Laboratory. Available at: http://bioenergy.ornl.gov/resourcedata/. Accessed 18
June 2005.

6
Chapter 2
REVIEW OF LITERATURE

2.1 Biomass Feedstocks

Herbaceous crops are plants including annuals (e.g., corn), which die at the end
of a growing season and must be replanted from seed, and perennials (e.g.,
switchgrass), which die back each year in temperate climates but reestablish
themselves each spring from rootstock (Brown, 2003). The residues (i.e., biomass)
from the herbaceous crops are potential feedstocks for renewable energy applications.
The challenge is transporting a large amount of biomass feedstock from the field to a
conversion facility economically.
Sokhansanj et al. (2003) presented a handling system for biomass collection
and processing it into feedstock for biorefining. The system was divided into three
activity centers: field, countryside (preprocessing center), and refinery. Each of these
activity centers play parts in converting biomass from its raw form to a finished
feedstock for delivery to biorefmery. Activities in the field include those for
harvesting and packaging, hauling and temporary storage in the field or moving to an
intermediate storage. Using existing equipment, the field operations involved are
similar to forage harvest and collection, mainly cutting, swathing, chopping or baling
and hauling. Preprocessing activities are those processes that reduce variability in
biomass. These operations may include size reduction, drying, densification (cubing or
pelleting or briquetting), and storing larger quantities of feedstock. Biorefmery
activities include operations such as receiving, sorting, storing, washing, and grinding.
The price of biomass feedstock is made up of two distinct components
(Sokhansanj et al., 2003): (i) "farm gate price" that includes production costs and
collection costs plus net return to the crop producer, and (ii) "delivered price" that
includes farm gate costs plus handling (loading and unloading), preprocessing,
storage, and transportation costs and some net return to the feedstock supplier.
Opportunities to reduce the farm gate price when using existing equipment come

7
from: (1) eliminating some of the field operations (e.g., integrating cutting and raking
operations; eliminating baling and instead using field chopped biomass; and whole-
crop harvesting and fractionation), and (2) improving field performance of the existing
equipment (e.g., cutting wider swath; increasing tractor speed; and making slightly
denser bales). Biomass delivered price could be reduced by densification (e.g.,
granulating) or comminuting biomass so that it can be safely and easily handled by the
existing handling equipment.
Figure 2.1 shows the role of the densification process in the biomass supply
chain from the field to the conversion facility. Figure 2.2 shows a typical process flow
diagram for the densification process. The major operations or pre-processing steps
involved in the preparation of biomass feedstocks to transport them from the field to
the conversion facility are discussed below.

2.1-1 Biomass Harvesting and Collection

2.1.1-1 Corn stover

Residue from the corn crop is called corn stover, which consists of all the
above ground, non-grain fractions of the plant including the stalk, leaf, cob, and husk.
On a dry mass basis, corn stover comprises of about 50% stalk, 20% leaf, 20% cob,
and 10% husk (Sokhansanj et al., 2002). Corn stover is the most abundant agricultural
residue in the U.S.; it represents 80% of the total agricultural residue, and is
concentrated in the Midwestern region of the U.S. (USDA, 2002). About 217 million
dry t/year of corn stover is available nationwide. The corn stover yield may range from
6 to 12 t DM/ha/year depending on geographic area, crop maturity, and crop variety
(Edens et al., 2002; Shinners et al., 2003; Savoie and Descoteaux, 2004). For the
sustainability of soil for organic matter replenishment and protection against water or
wind erosion, current harvesting methods offer the possibility of collecting 50 to 60%
of stover after grain combining. Therefore, the harvestable stover ranges from 3 to 6 t
DM/ha/year. According to Kadam and McMillan (2003), in the U.S., about 80-100

8
million dry t /year (estimated value) of corn stover can be sustainably collected and a
majority of which would be available for bio-ethanol production.
Corn stover is currently an underutilized biomass. Presently less than 10% of
corn stover is collected for bedding and animal feed by local farmers and for domestic
use (Kadam and McMillan, 2003; Sokhansanj et al., 2003). Potential uses of corn
stover are: after hammer milling directly fed in a boiler furnace, ethanol production,
particle board production, pulp and paper, chemicals such as furfural from corncobs,
and corn stover mulch to prevent soil erosion and for slope stability (Kadam and
McMillan, 2003). Morey et al. (2006) proposed the use of corn stover for heat and
electricity generation at ethanol plants and found that there was a significant annual
energy cost savings for the 150 million L ethanol per year plant.
Corn stover is harvested during mid September-November in the U.S. Midwest
(Sokhansanj et al., 2002). Conventionally, a corn stover windrow behind the combine
is collected using a round baler. A typical bulk density of round bales of 1.8 m
diameter x 1.5 m width is 150 dry kg m" at 25% (w.b.) moisture content (Shinners et
al., 2006a). The total energy input required for shredding, baling, and stacking of corn
stover bales may be about 314 MJ/t at a cost of $38/t (Sokhansanj et al., 2006a and
2006b).
Corn stover is most often harvested as a dry product and packaged in large
round or large square bales, typically involving the following steps after grain
harvesting (i.e., combining): shredding, field drying, raking into a windrow, baling,
gathering bales, transporting to storage, unloading, and storing. Problems with this
system include difficult field drying due to short day length and low ambient air
temperatures, short harvesting window between grain harvest and first snow cover,
frequent weather delays, soil contamination of stover during shredding and raking, and
low harvesting efficiency. Too many field operations can result in high costs per unit
harvested mass.
Single-pass harvesting systems which produce grain and stover harvest in
separate crop streams are currently being studied to eliminate the number of field
operations and costs. Single-pass harvesting of corn stover would result in 51 to 53%

9
(w.b.) moisture content, which could be stored in bag silos at densities of 120 to 193
kg DM m"3 (Hoskinson et al., 2007; Shinners et al., 2006a and 2007).

2.1.1-2 Switchgrass

Perennial grasses have several advantages over annual crops such as lower
establishment costs, reduced soil erosion, increased water quality, and enhanced
wildlife habitat. Switchgrass is a warm-season perennial grass, which reduces
greenhouse gas by increasing carbon storage in the soil. Switchgrass has a yield of 3 to
301 DM/ha/year depending on soil fertility, location, cultivars, and number of cuts per
year (Kumar and Sokhansanj et al., 2007). Switchgrass has been evaluated as a biofuel
crop in the Midwest, the Southern and Northern Great Plains of the U.S., south eastern
Canada, and Europe (Adler et al., 2006). Switchgrass has also been identified as a
fiber source in Canada for pelleting in the alfalfa pellet industry to run the facility year
round (Samson et al., 2000).
Switchgrass may be harvested in fall (August-November) or spring (April-
May) using standard silage (hay) equipment (Cundiff and Marsh, 1996; Shinners et al.,
2006b). Switchgrass is cut at about 100 mm stubble height. The switchgrass was cut
and conditioned with a disk cutterbar windrower equipped with urethane conditioning
rolls (Shinners et al., 2006b). For swathed crop, a single rotor rotary rake was used to
narrow the swath into a windrow of appropriate width for the baler to pick-up. Baling
was done using a round baler. The maximum density of round baled (1.8 m diameter x
1.5 m wide) switchgrass was 139 kg m"3 at 13% moisture content (Cundiff and Marsh,
1996). Spring harvested switchgrass is better than fall harvested switchgrass because
spring harvested switchgrass has higher energy content, lower ash content, lower
nitrogen content, and lower potassium content (Samson et al., 2000).
Since hay harvesting does not involve separation of plant parts (e.g., grain),
equipment much simpler than combined harvesting and threshing machinery can be
employed. Haymaking involves separate steps of cutting the crop (mowing),
conditioning (rollers or flails that slightly crush the plant stems allowing more rapid

10
release of moisture) the cut crop to improve field drying (with possible tedding,
turning the crop to improve field drying), arranging the crop into long rows in the field
(windrowing), and either baling the crop or loading the loose hay into wagons (Brown,
2003). However, conventional hay harvesting equipment may not adequately handle
the switchgrass due to the higher yield per year of perennial grasses than forage crops
(Shinners et al., 2006b). A great deal of work related to harvesting, drying, packaging,
and storing of perennial grasses in a system approach is required (Shinners et al.,
2006b).

2.1-2 Biomass Storage

Biomass materials are harvested only once per year or just a few times per
year. Processing to energy, fuels, chemicals or bio-materials is year-round activity as
manufacturing facilities are too expensive to remain idle most of the year. Most
biomass materials undergo substantial degradation in the course of a few months or
even days if exposed to the elements such as rain after harvest. Therefore, proper
storage to preserve plant materials for periods of a year or even longer is critical to the
successful development of a bio-based products industry (Brown, 2003).
Traditional methods of long-term storage of crop residues involve storing the
dried bales (indoor or outdoor), ensiling a crop, or immediately processing the crop to
obtain a more stable intermediate product such as cubes (Brown, 2003). The choice
depends on the nature of the crop and the kind of processing it will ultimately
undergo.
Ensiling was developed for humid climates, where field drying is impractical,
but it is finding increasing application in drier climates. The biomass harvested at 40
to 50% (w.b.) moisture content and stored under anaerobic conditions to promote
partial oxidation of sugars to organic acids, which act as preservative and suspend
further microbial degradation of the crop. Chemical additives such as organic acids or
pulverized lime stone may be added to control the pH. Storage systems include
horizontal (e.g., bunker silos) and vertical (e.g., metal or concrete conventional silos)

11
silos. Losses during ensiled storage are inevitable. If properly preserved, biomass
losses can be kept to 5 to 10%, depending on the sugar content of the biomass.
Improper ensiling can result in losses as high as 30% (Brown, 2003).
High moisture corn stover (40 to 70% w.b.) has been used for animal feed by
ensiling in silage bags or silos. Ensiling could provide an opportunity for intermediate
storage on farm and pre-treating the corn stover destined for conversion to ethanol or
lactic acid (Muck and Shinners, 2006). If the field wilting does not allow natural
drying of corn stover to less than 30% (w.b.) moisture because of frequent rain or cool
weather, wet storage is likely to be the preferred method of conservation (Shinners et
al., 2003). Storing corn stover bales indoors has less dry matter loss (3.3%) compared
to outdoor storage (18.1%) (Shinners and Binversie, 2004).
Assuming that corn stover bale collection and transport are fully mechanized
using automated loaders and bale transporters, Sokhansanj and Turhollow (2002)
calculated the costs for all operations up to and including stacking in (on-farm) storage
as $22/dry t for the round baling system, and $24/dry t for the rectangular baling
system. The difference in cost of the two systems was due to the additional raking
operation and higher capital cost of equipment for rectangular baling and transport,
and smaller load for the rectangular bale transport.
Shinners et al. (2006b) found that switchgrass had a moisture content of 46 to
66% (w.b.) at cutting. When switchgrass was placed in a wide-swath by tedding, the
moisture content would be reduced to a baling moisture content of < 20% (w.b.) in a
day. Round bale (twine or net wrap) density averaged 203 kg m"3 at 20% (w.b.)
moisture content. Dry bales stored outdoors for 9-11 months averaged 3.4, 7.7, 8.3,
and 14.9% dry matter loss for round bales wrapped with plastic film, net wrap, plastic
twine, and sisal twine, respectively. Bales stored indoors averaged 3.0% DM loss.
Preservation by ensiling in a tube of plastic film produced average DM loss of 1.1%
(Shinners et al., 2006b). Switchgrass bales exposed to rainfall during outdoor storage
experienced much greater weathering (including leaching, ultraviolent degradation and
erosion) (Wiselogel et al., 1996). Due to weathering and biochemical reactions
produced by microbial life during outdoor storage of bales, loss of dry matter,

12
fermentable carbohydrates, and heating values would occur, which has a negative
economic impact.
Cundiff and Marsh (1996) reported about 13% storage losses for round bales
of switchgrass. The square bales (1.2 m high x 1.2 m wide x 2.4 m long) must be
stored in covered storage (3% storage/handling loss), while net-wrapped round bales
(1.8 m diameter x l . 5 m wide) can be stored on crushed rock (5% storage/handling
loss). Total harvest plus (on-farm) storage cost, per dry Mg of switchgrass-delivered
basis, were $20 for the round bale and $27 for the square bale (Cundiff and Marsh,
1996). Because of the requirement for covered storage, the square bale will have more
difficulty competing economically with the round bale. However, as yield increases
above 9 dry t/ha and storage losses for round bales stored outside increase above 5%,
the difference in cost would decrease.
Handling baled crops is very labor intensive. Square bales consumed about
20% less time for breaking and chopping compared to round bales of switchgrass
(Jannasch et al., 2001).

2.1-3 Biomass Transportation

Sokhansanj et al. (2006) estimated that for combined heat and power
applications requiring 53 MW of total energy for 150 million L/year ethanol plant,
corn stover of 500 t/day was required. De Kam et al. (2007) estimated that for
combined heat and power applications in a dry-grind ethanol plant producing 190
million liters of ethanol per year, the amount of corn stover at 13% (w.b.) moisture
content required ranged from 416 to 575 t/day. Mukunda et al. (2006) calculated that
for production of 100 million gallons per year cellulosic ethanol (biorefinery) from
corn stover yielding 72 gallons per dry ton required 3968 dry tons of corn stover/day
of operation.
Therefore, there is a need to find ways to cost effectively transport the large
amount of biomass (corn stover or switchgrass) required for the operation of bio-based
industries. The biomass can be transported to the bio-based industries as bales, ground

13
particles, or densified forms (pellets/briquettes). Bales (round or square) can be
transported using flatbed trucks. Square bales can be transported more efficiently with
trucks (more dry mass per truck) due to their form than round bales. Square bales are
easy to handle and stack in enclosed spaces. However, round balers have a lower
purchase cost than square balers. The other disadvantage of square bales is that it is
very difficult to get the moisture out of the bales when rained on (Cundiff et al., 2004).
The ground or densified (pellets/briquettes) biomass can be transported using truck
boxes. Transportation by ships is used to export pelleted biomass.
Sokhansanj et al. (2006) showed that for a day of operation, 27 truck loads of
square bales with 18 t/load, 44 truck loads of chopped biomass with 11 t/load, and 12
truck loads of pellets with 40 t/load need to be transported to the 150 million L/year
ethanol plant for combined heat and power applications. They found that transporting
to the plant (from about a distance of 70 km), unloading and storing on-site would be
cheaper and easier with pelleting than bales or chopped biomass. The cost analysis for
corn stover by Sokhansanj et al. (2006) showed that collection, pre-processing and
transport to the ethanol plant required $48, $63, and $67/t of baled, chopped (ground),
and pelleted corn stover, respectively. Transport cost of chopped biomass was the
highest due to low bulk density. Additional cost ($25/t) was required for baled
biomass to make it suitable for combustion at the plant. A least on-site fuel preparation
cost ($14/t) was estimated for pelleted biomass due to low on-site storage cost
compared to those of bales and chopped biomass. Dry chops were the most expensive
form of biomass delivered to the combined heat and power plant followed by pellets
and square bales.
Sokhansanj et al. (2006) recommended compaction of chopped biomass or
granulation (forming clumps of compacted particles not as large as pellets or
briquettes) biomass particles to increase the bulk density to reduce the cost of biomass
delivered to the ethanol plant while keeping the pre-processing cost less than the
pelleting cost. They further concluded that granulated biomass would require less fuel
preparation time on-site (such as hammer milling and pneumatic conveying) compared

14
to bales, chopped biomass, and pellets. Also, more efficient collection and
transportation systems need to be developed to reduce the biomass delivery cost.
Mukunda et al. (2006) calculated that for production of 100 million gallons per
year cellulosic ethanol (biorefinery) from corn stover yielding 72 gallons per dry ton
required 53 truck loads (19.5 dry tons/load) of square bales with 6 unloading stations
at the plant to meet the daily feedstock demand (i.e., 3968 dry tons/day or 227 truck
loads/day) while maintaining a steady inventory of 10 days of demand at any given
point of the year. According to Mukunda et al. (2006), for delivering square baled corn
stover from farm-storage sites to ethanol plants for producing lignocellulosic ethanol,
unloading of bales from the trucks or semi-trailers at the plant appeared to be a
bottleneck in terms of additional requirements on the unloading facility and
equipment, time spent by the vehicles at the plant, and possible additional traffic. They
concluded that logistics in an alternate pathway with an intermediate pre-processing
and storage (e.g., pelleting/briquetting) facility needs to be explored to reduce the
above mentioned transportation problems.
Once the bio-energy industry is established in the U.S., large quantities of
biomass will be transported on public roadways and stored in a distributed network of
strategic locations. There are issues on maximum size and weight of the load in
storage or in transit. Also, biomass in its natural form (not dried and densified) is
susceptible to spontaneous heating and incidental fires. Therefore, biomass materials
may require a higher insurance premium (Sokhansanj et al., 2003). A detailed study on
cost economics for handling different forms of biomass feedstock at an end user (e.g.,
ethanol plant) either for combined heat and power or bio-ethanol production is
required.

15
2.1-4 Biomass Drying

The critical moisture content for safe storage of most agricultural products is
less than 15% (w.b.). Under cooler conditions, biomass can be stored at higher
moisture contents of up to 20% (w.b.) for a short time. Drying removes moisture from
a crop to attain a moisture level low enough to stop the growth of microorganisms.
Freshly cut herbaceous biomass may have a moisture content of 60 to 85% (w.b.). The
target moisture content at forage harvest would range between 20 and 30% (w.b.)
(Brown, 2003).
Biomass can be dried in the field or with a help of a mechanical dryer. To dry a
ton of biomass containing 50% moisture down to 10% moisture would require about
1.5 GJ of energy, which is about 18% of the energy content of the fresh biomass
(Brown, 2003). Thus, field drying is employed whenever possible.

2.1.4-1 Field drying

The moisture content of the alfalfa brought to the pelleting industry may range
from 65 to 80%> (w.b.) for freshly chopped crop, and from 15 to 40% (w.b.) for pre-
wilted crop (Patil et al., 1993). To save money required for drying, generally the
alfalfa is cut and allowed to dry in the field (wilting) (Patil et al., 1993). Field drying
rate of alfalfa was found to be affected by solar insolation, dry bulb temperature, vapor
pressure deficit, soil moisture content, and swath density (Rotz and Chen, 1985).
Mechanical conditioning (e.g., shredding and then mat forming in the field, or roller
crushing) and chemical conditioning (e.g., potassium carbonate treatment) could
enhance the field drying of hay and grasses (Shinners et al., 1989; Savoie et al, 1990;
Patil etal., 1993).
During the harvest time of corn, the moisture content of grain, cob, husk, leaf,
and stalk averaged 29%, 51%), 56%>, 28%, and 72%> (w.b.), respectively (Morey and
Thimsen, 1980). The moisture content is affected by relative maturity and variety,
among other factors. With field drying, depending on the geographic location, the

16
baled stover moisture content would range from 15 to 30% (w.b.) (Edens et al, 2002;
Shinners et al., 2003). Possible field drying time available (about one to three weeks),
and rain or snow after the cut crop was laid on the ground would affect the moisture
content of the baled crop (Shinners et al., 2003). Based on Shinners et al. (2006b),
switchgrass could be dried in the field to a baling moisture content of about 20%
(w.b.) or less.

2.1.4-2 Artificial drying

In climates where field drying is not practical, to avoid dry matter losses in the
field due to damp cool weather and rain, barn drying of hay is employed (Brown,
2003). After the forage is cut, it is allowed to field-dry down to a moisture content of
40 to 45% (w.b.). The material is then collected with a self-loading wagon and stored
in a drying barn, where either ambient air or air that has been heated 10 to 40C above
ambient is circulated through the stacked or palleted hay bales to reduce the moisture
to around 12 to 15% (w.b.) (Carson and Kreider, 1988). Bales are dried in groups of
60 to 84 as a batch drying system (Carson and Kreider, 1988; Plue and Bilanski,
1990).
Forced-air drying of baled hay has been practiced for over half a century
(Weaver et al., 1947). It is widely used in North America to remove excess water in
hay bales that otherwise spoil during storage. For forced-air drying systems, heat from
propane gas, solar energy or ambient air have been used (Baker and Shove, 1978;
Parker et al., 1992). The conventional approach with forced-airy drying is to stack hay
bales up to 4 m high and to ventilate the bales with ambient air. Forced-air drying with
ambient air can reduce the moisture content from 30% to 15% after four weeks.
Problems with the heated air drying systems include over drying (<12 to 15% moisture
content), non-homogeneity of moisture in the bales, and low energy efficiency.
However, heated air drying is feasible for baled hay within 2 to 3 days at good
operating conditions.

17
Savoie and Descoteaux (2004) conducted experiments on a forced-air drying
system for corn stover bales. Two layers of 18 corn stover bales (0.81 m * 0.89 m x
1.52 m) were dried from 56% to 18% (w.b.) moisture content in a batch type, heated-
air dryer (61C for 90 h). The energy cost for drying from 55 to 12% (w.b.) moisture
content ranged between $56 and $98/t DM depending on the type of dryer and the
source of energy (combustion of natural gas or electrical heat pump). Reducing the
energy cost below $10/t DM would require field drying stover below 20% (w.b.)
moisture. The energy efficiency of the dryer ranged from 13 to 30%. They also found
that there were several technical hurdles that need to be addressed before considering
artificial drying of corn stover bales on a larger scale. One is heterogeneity of moisture
content at harvest. The difference in moisture contents between bales (i.e., the
difference between maximum and minimum moisture) was initially on the order of 15
to 20 percentage points. This difference was seen to continue or even increase after
bale drying.
In densification plants, direct contact co-current type rotary drum dryers are
often used for drying forages such as timothy grass and alfalfa (particle size of 20 to
80 mm) to dry them from 30% to 10% (w.b.) (Mani et al., 2005a). This type of dryer
uses hot flue gas at about 250 to 600C. Drum dryers produce uniform product quality
due to the long residence time and relatively good mixing compared with other types
of dryers. Energy consumption of rotary dryers ranges from 3 to 8 GJ/t depending
upon the operating conditions and initial moisture contents (Mani et al., 2005a).
Generally, the drum dryers are fueled by natural gas or wood waste.
Super heated steam dryers, simple fluidized bed dryers, flash dryers, spouted
bed dryers, and belt dryers are also used in Europe for drying biomass materials (Stahl
et al., 2004; Thek and Obernberger, 2004). For finely ground particles, flash drying is
an effective method, but requires careful design. Most of the dryers in the U.S. run on
natural gas. A variety of biomass heaters are commercially available to be integrated
with biomass dryers (Mani et al., 2005a).

18
2.1-5 Size Reduction

Size reduction is crucial to the densification process. Size reduction is also an


important pre-processing operation for biomass energy conversion. The optimal
particle size of the feedstock depends on the conversion process (e.g., liquefaction,
combustion, or gasification). Particle size reduction increases the total surface area,
pore size of the material, and the number of contact points for inter-particle bonding in
the compaction process (Drzymala, 1993). The smaller the substrate particles, the
higher the reaction rate because more surface area is exposed to the enzymes and
microorganisms that promote the process. Size reduction of lignocellulosic biomass is
important to eliminate mass and heat transfer limitations during the hydrolysis
reactions (Schell and Harwood, 1994).
Coarse grinding using a tub grinder is done to make the biomass feedstock
suitable for a dryer. Tub grinders are used to chop (25 to 100 mm in length)
agricultural and forest residues (Arthur et al., 1982; Jannasch et al., 2001). Also, a
hydraulic piston pressing the hay against a grid of knives may be used to cut the round
or square bales into short stem pieces (Sokhansanj and Turhollow, 2004). This method
is preferred to the traditional tub grinding because the leaves are preserved better and
the length of cut is more precise and uniform.
In North America, hammer mills and roller mills are used for fine grinding of
biological materials such as grains and forage crops (Lobo, 2002). Roller mills are
limited to non-fibrous product (i.e., not for use on hay or fodder). Schell and Harwood
(1994) found that the energy requirement for milling of lignocellulosic biomass with
hammer milling was less than for disk milling. In addition, hammer mills have wide
applicability in biomass size reduction because of their simple design, ruggedness,
versatility to grind a variety of materials (e.g., breaking clumped distillers dried grain
with solubles (DDGS) or regrinding the pellets), and low maintenance.
Hammer mills consist of fixed or swinging hammers mounted on a rotor
assembly which is on a rotating shaft, a screen, and a fan. The screen (perforated
metal) is mounted below, above and/or around the hammers through which the

19
reduced product must pass. The product being ground remains in the grinder until it is
small enough to pass through the holes in the screen. The hammers do not touch the
screen. As a general rule, if the hammers are farther away from the screen, it will be
beneficial for chopping hay or fodder, and they will use less horse power and create
less fine material. On the other hand, if a fine product is desired, the hammers should
be very close to the screen which will increase horsepower requirements (Naylor and
Smith, 1981).
In the hammer mill, the feed material is initially struck (impacted) by rotating
hammers and then thrown (impacted) against grid (breaker) plates. Normally, the size
reduction due to impact occurs above the center line of the rotor, while the size
reduction due to attrition (rubbing action between the hammers and screen) occurs
below the center line of the rotor. Also, size reduction occurs by cutting action by the
edge of the hammers (Beven, 1977).
The rate of feed to the hammer mills is always controlled. After the product
passes through the screen it is normally picked up by a pneumatic systems supplied by
a fan and conveyed to a cyclone collector or filter and then into a bin. The negative
pressure collection system is popular for two reasons: (1) the negative pressure
increases the capacity of the mill, and (2) automatically adds the dust control system.
Alternative methods are gravity or mechanical conveyor systems.
Energy consumption for grinding biomass in hammer mills is affected by
hammer mill variables [grinding (feeding) rate, hammer tip speed, screen size, and
clearance between the hammer and the screen], and biomass variables [species,
physical and chemical composition, variety, maturity, parts of the plant (stem vs. leaf),
time after harvesting, and storage history] (Balk, 1964; Sah et al., 1977; Mani et al.,
2002a; Yu et al, 2003).
Lab units normally require drives in the range of 0.4 to 3.7 kW (0.5 to 5 HP)
(Beven, 1977). Industrial units require drives which fall in the range of 15.0 to 1865
kW (20 to 2500 HP) (Beven, 1977). Capacities of hammer mills can be as low as 1/8
ton/h to 2 ton/h for the lab type units or range from 5 ton/h to 2500 ton/h for industrial
units (Beven, 1977).

20
Mani et al. (2002b) studied the grinding characteristics of wheat straw, barley
straw, corn stover, and switchgrass with a hammer mill (1.5 kW) using three different
screen sizes (0.8, 1.6, and 3.2 mm). Switchgrass required the most grinding energy and
corn stover used the least energy. Large hammer mill screen size resulted in reduced
energy requirements for all of the biomass tested. Because of low fiber content and
presence of sponge vascular tissues in the stem, it was expected that corn stover would
consume less energy. For grinding corn stover with 3.2 mm hammer mill screen size
with a chop size of 12.5 mm at 6 to 12% (w.b.) moisture contents, the specific energy
consumption ranged from 25 to 40 MJ/t. For grinding switchgrass with 3.2 mm
hammer mill screen size with a chop size of 7.5 mm at 8 to 12% (w.b.) moisture
contents, the specific energy consumption ranged from 86 to 101 MJ/t. Samson et al.
(2000) reported a specific energy consumption of about 162 MJ/t of switchgrass for a
hammer mill with a screen size of 5.6 mm. Jannasch et al. (2001) reported a specific
energy of 202 MJ/t of switchgrass for hammer mill (112 kW) screen sizes of 5.6 and
2.8 mm.

21
2.1-6 Densification

Densification of agricultural residues can be achieved by baling, cubing,


pelleting, or briquetting. Bulk density of loose plant-based biomass ranges from 50 to
130 kg m"3 depending on the plant species and particle size and its distribution
(Sokhansanj and Turhollow, 2004). As a solid fuel, biomass is somewhat similar to
coal. For example, bituminous coal has roughly 30200 kJ kg"1, whereas agricultural
residues average about 18600 kJ kg"1 on a dry basis (Olsson et al., 2003). When
biomass is densified, its bulk volume can be much closer to that of coal (1.1 to 1.5
m3/Mg). Agricultural residues such as loose straw can have a bulk volume in the range
of 24.7 to 49.4 m3/Mg, although chopping (12.3 to 49.5 m3/Mg) or baling (4.9 to 9.0
m3/Mg) this type of biomass material significantly increases its density (Olsson et al.,
2003). The tendency toward low weight per unit volume translates into higher
transportation costs. Feeding the low bulk density materials into combustion units
requires special consideration for handling and processing these materials.
The bulk density of straw compressed to briquettes or pellets is 5 to 10 fold
compared to that of baled straw (Wilen et al., 1987). High-density, fabricated biomass
shapes simplify the logistics of handling and storage, improve biomass stability,
facilitate the feeding of solid biomass fuels to furnaces and reactors, and offer high
energy density, clear burning solid fuels that in some cases can approach the heating
value of coal. A significant part of the biomass supply chain involves transportation.
Standardized fuel characteristics allow for an easy adjustment of the overall
transportation chain and a fully automatic operation of the end use technology, as in
the case for fossil fuel energy (Holley, 1983; Kaltschmitt and Weber, 2006).
Solid biofuels are most widely used for combustion in small stoves to heating
plants, and are readily used for electricity generation based on the conventional steam
cycle. In the future, solid biofuels might be used for provision of liquid fuels based on
a gasification process for the provision of a gas used for the synthesis of liquid fuels.
In Europe, solid biofuels are the most important bioenergy carrier so far (Kaltschmitt
and Weber, 2006). The use of pellets in co-firing with coal enhanced its utilization

22
dramatically in Europe. Some European countries are planning to import wood pellets
from Africa, America or neighboring European countries due to the shortage of wood
on a long term-basis (Hoque et al., 2006).

Some of the added advantages of the densified products are listed below:

1. No or low content of dust, therefore, increased safety during handling


2. Freely flowing
3. Improved storage and handling characteristics
4. Improved metering/dosing capabilities
5. No segregation of co-agglomerated components
6. Defined shape and size
7. Increased bulk density
8. Higher energy density
9. Improved product appeal (value addition)
10. Increased sales value (marketability)
11. Dry and can be stored without degrading

2.1.6-1 Densified products

The original use of biomass pellets in the U.S. was as a fodder. Alfalfa, other
grasses, and some straws were pelletized, and sold as livestock feed. The most
common densified forage products are pellets, cubes, dense bales, bagged chopped
hay, and wafers (Sokhansanj and Wood, 1991).
Pellets are usually in the form of hardened forage cylinders 3/16 inch to 3A inch
in diameter with a length larger than the diameter (1/2 to 1 inches). The unit density of
pellets is on the order of 961 to 1121 kg m"3 (60 to 70 lb ft"3). Bulk density of pellets
may be as high as 700 kg m"3 (Sokhansanj and Turhollow, 2004).

23
Cubes are larger size pellets, usually in the form of a square cross section made
of chopped forage. Cube sizes are from V2 to 1.5 inches in width. The length of a cube
is usually equal or longer (1 to 4 inches) than the dimensions of the cross section. The
density of individual cubes may range from 641 to 801 kg m"3 (40 to 50 lb ft"3). Cubes
are less dense than pellets with bulk density ranging from 450 to 550 kg m"3
(Sokhansanj and Turhollow, 2004). Cubing is not recommended as a densification
method since the large (32 mm x 32 mm x 50 mm) cube with low surface-to-volume
ratio results in a very slow gasification (Miles, 1984). The large cubes are virtually
impossible to handle in automatic feeding equipment.
Dense bales are unchopped forage compressed into square bales. The bales are
usually 14 x 14 x 16 to 18 inches, and weigh about 60 lb each. The compressed bale
density (25 to 30 lb ft"3) is less than the cube density. Bagged-chopped-hay is chopped
hay compressed into 50 lb bags. The bags have dimensions of 8 x 16 x 32 inches. The
particle size in the chopped hay is larger than the ones in cubes. Wafers are pressed
hay made from long or chopped hay, the diameters may be as much as 6 inches but
unlike pellets or cubes, wafers are thin, up to 2 inches thick. The commercial use of
the term wafer is not common (Sokhansanj and Wood, 1991).
Solid fuels from wood or crop residues are usually made as pellets (6 to 20 mm
in diameter) and briquettes (50 to 100 mm in diameter). For pelletization, ring or flat
die pelleting presses are used. For manufacturing wood pellets, ring-die pellet mills
are commercially used. Most pellet mills in North America produce pellets with 6.4
mm diameter. For briquetting, impact press, piston or screw press, roll press, and die
presses are commercially available. Briquettes are usually larger in diameter and
length than pellets (Wilen et al., 1987). Larger briquettes (e.g., 3 3A inches in diameter
and 11 inches in length) produced by extruder type of machines are called logs (Jones
and Jones, 1980).
The major outlets for bio-pellets are the domestic market (heating), and the
energy market (boiler fuels, furnaces, and heat and power). Most of the U.S. pellets
are bagged and marketed for domestic pellet stoves. In Canada, pellets produced from

24
sawdust and wood shavings are exported to European counties such as Sweden and
Denmark (Mani et al., 2006).

2.1.6-2 Densification machines

Generally, two techniques are most widely used for size enlargement of
particulate materials: tumble agglomeration and pressure agglomeration (Pietsch,
1991). In tumble agglomeration, agglomerates are formed during suitable movement
of the particulate materials containing binder in equipment such as balling discs,
balling cones and balling drums. Whereas in pressure agglomeration, high forces are
applied to a mass of particulate materials within a confined volume to increase the
density. Pressure agglomeration is accomplished in piston, roller, and extrusion
presses as well as in pelleting machines (figs. 2.3 and 2.4).
The most popular commercially used pressure agglomeration equipment for
densification of biomass materials are screw extruders, ram (piston-type) extrusion
presses, ring-die pelletizers, flat-die pelletizers, cubers, and roller (or roll) presses
(Pietsch, 1991). The selection of the type of agglomeration machine is determined by
several factors such as investment cost, operational costs (i.e., energy, spare parts, and
maintenance), quality of the densified products, flexibility in feedstocks, and safety
aspects (Svenningson and Hosier, 1987).

2.1,6.2-1 Screw extruder

In an extruder, the raw material is compressed by a screw through a die to


form compacted particles, often of cylindrical shape. The diameter of extrusion
products may range from 20 to 100-mm. Extrusion technologies are commonly used
for biomass densification in Asian countries such as India, Japan, and Thailand
(Bhattacharya et al., 1985; Grover and Mishra, 1996a). The extrusion press has been
found to have several problems in briquetting biomass materials such as danger of fire
in the material which can get heated to a temperature close to the ignition point, and

25
the material gets partly pyrolyzed on the surface which can produce hazardous gases.
A typical energy demand for a screw press is 360 MJ/t (Svenningson and Hosier,
1987).

2.1.6.2-2 Piston press

In piston presses, the raw material is compressed by a piston through a die to


form briquettes. In piston presses, high acceleration forces at the return point of the
piston limit the number of cycles per unit time. At high speeds, feeding the die cavity
is also difficult, especially for fine powders. Therefore, the capacity of such machines
is limited (Pietsch, 1991). Also, the piston press is not flexible with all feedstocks; for
example, it was impossible to make piston briquettes from coffee husk (Svenningson
and Hosier, 1987). A typical energy consumption by piston press is 300 MJ/t
(Svenningson and Hosier, 1987).

2.1.6.2-3 Pelleting machine

In pelletizing equipment, the material is pressed through open-ended


cylindrical holes (dies) made in the periphery of a ring. One to three small rotating
rolls push the feed material into the die holes from inside of the ring towards the
outside of the ring. The skin friction between the feed particles and the wall of the die
resists the free flow of feed and thus the particles are compressed against each other
inside the die to form pellets. One or two adjustable knives placed outside the ring cut
the pellets into desired lengths. The diameter of the pellets may range from 4.8 to 19-
mm. Pelletizing technology is commonly used in the U.S. for producing animal feed.
Pellet mills (annular or flat dies) have production rates of as high as 25 t/h (Pietsch,
1991).
Animal feed or wood pelleting involves five important stages: drying, milling
(size reduction), pressing, cooling, and screening (Khankari et al., 1989a; Mani et al.,
2006). The wood pelleting procedure is as follows. Green woodchips may contain 50

26
to 65% (w.b.) moisture content. So the material is usually dried to 7 to 8% (w.b.)
moisture content. For wood pelleting, moisture content of up to 20% (w.b.) is used.
All existing types of dryers (drum dryer, steam dryer (direct or indirect), and hot-air
dryer) have been used for drying woodchips. Dried product passes over a permanent
magnet to remove any ferrous metal before milling. A hammer mill is used for
milling. The grinding energy is converted into heat, which further removes moisture
from the material.
In the pelleting mill, the feed mash (i.e., mixed feed ingredients) flows into the
screw feeder, which delivers a constant amount of meal into the conditioning chamber
where steam, and liquids or external binders are added. The conditioning chamber has
an agitator to blend the additives. The conditioned mash then flows into the pelleting
chamber where the pellets are formed. Pellets leave the pellet mill at about 100C and
need to be cooled down to about 25C to harden and stabilize the pellets. A counter
current pellet cooler is used for pellet cooling. Cooled wood pellets are finally passed
over a vibratory pellet screen to separate fines from the pellets. The fines are returned
to the process, whereas the dust free wood pellets are ready for bulk storage or
packaging.
During pelletizing of wood, no additives are used, but steam conditioning may
be done. During the extrusion process, the lignin component in the cellulose material
migrates to the pellet surface. During cooling, the lignin forms a surface skin on the
pellet that protects it from shattering and from any rapid change in moisture content
before the pellet fuel is used. A certain amount of moisture is removed by the pellet
mill.
Samson et al. (2000) conducted preliminary pelleting of switchgrass using 1.5-
kW (2-HP) or 18.7-kW (25-HP) pellet mills. They found that the bulk density of
pellets was 615 kg m"3 (4.8 mm diameter) with pellet hardness of >30 on the Pfizer
tablet hardness tester. A preliminary study in a commercial scale pelleting plant (2 t/h)
produced switchgrass pellets with bulk density of 561 to 641 kg m"3 (35 to 40 lb ft"3)
with an energy consumption by the pelleting mill of 268 MJ/t (Jannasch et al., 2001).
Jannasch et al. (2001) did not produce durable pellets, and they concluded that further

27
study would be required to produce durable pellets using binding agents, steam
conditioning, or adjusting pellet machine variables such as die length-to-diameter
ratio. Using a 1.5-kW (2-HP) pellet mill, Colley et al. (2006) produced switchgrass
pellets (4.8 mm diameter) with bulk densities of 536 to 708 kg m"3 and durabilities of
78 to 97%. However, Colley et al. (2006) did not study the effects of biomass material
or machine variables, and process conditions on the pelleting characteristics of
switchgrass.

2.1.6.2-4 Forage cubers

Cubing of biomass such as alfalfa involves baling, chopping of bales in tub


grinders, drying of chopped biomass in drum dryers at 200 to 400C, mixing with
water and 2% (by wt.) binders such as bentonite or hydrated lime, and compressing in
a cuber, cooling/drying the cubes to 10-12% (w.b.) moisture content, storing the cubes
in flat storage, and bagging the cubes for transportation via rail cars or trucks
(Sokhansanj and Turhollow, 2004). The cuber die and press roller (wheel) are similar
to the ring-die and press roll in a pellet mill. In the case of pelleting machines, the die
ring contains several rows of die holes of circular cross section, whereas the cubers
usually have one row of holes in the die ring. Cubers have a larger diameter auger, die
ring, and a single press wheel. The working principle of the cuber is the same as that
of the pellet mill. The auger moves the chopped biomass uniformly towards the
opening in the die ring. The press roller compresses the biomass into the ring-die (i.e.,
holes with square cross section) to from cubes. A knife located outside the ring-die
cuts the cubes with desired lengths. The pressures in the cuber range from 24 to 34
MPa. Cubes exit the cuber warm, at a temperature of more than 60C. It is necessary
to add a binder to increase the durability of cubes.

28
2.1.6.2-5 Roll press briquetting machine

The roll press (or roller press) was originally used for making coal briquettes at
the end of 19th century (Dec, 2004). Nowadays, a large number of other materials are
compacted using the roll press. The basic concept of roll pressing is to squeeze
(compact) particulate material between two rollers rotating in opposite directions. The
product size may range from 10 to 40-mm.
The basic working components of the machine are two counter-rotating rollers.
The working surface of the rollers has briquetting pockets with specific shapes cut into
the rollers. Proper feeding to the rolls is important for ensuring successful roll press
operation. A simple gravity type feeder was usually used in earlier roll presses. A
force feed system is commonly used with almost all modern machines. The most
widely used are screw type feeders. These devices, in addition to controlling the mass
flow and pre-compaction of feed material, frequently have important secondary
effects. For example, they may crush the particles to achieve a more favorable size
consistency. With fine materials, deaeration can be achieved with proper screw feeder
design. Shear deformation and friction during the material flow in a screw can be
beneficial, such as developing better contact area between the individual particles,
which results in higher binding forces during roll compaction. Shear deformation and
friction can also be utilized for heating the particles to activate some particular binding
forces. A screw feeder will also be essential when an automatic control system is
planned for controlling the roll press operating parameters.
The advantages of roll pressing are conceptual simplicity, great versatility as to
the nature and particle size of feed material as well as final product, possibility of
producing a strong agglomerate without any additives, and ability to process large
quantities of material at low cost. Due to their high compaction pressures (up to 160
kN/cm of working roll width) and outputs (up to 60 t/h), and low specific energy
consumption (20 to 60 MJ/t), Drzymala (1993) concluded that roll presses can be used
for large-scale briquetting of ores, metal chips, ferruginous sludge, and coal compared
to other briquetting machines such as piston presses.

29
2.1.6.2-6 Roll compaction /dry granulation

Compaction/dry granulation is a controlled crushing of pre-compacted


powders densified by passing them between two counter-rotating rolls (Kleinebudde,
2004). The advantages of roll compaction are greater production capacity, more
control over operating parameters and dwell time, and minimal need for powder
lubricant. The aims of roll compaction/dry granulation are an improved handling of
the powders due to a larger particle size and a better flowability. Dust problems are
minimized or avoided. Roll compaction is also best for compacting corrosion- or
abrasion-resistance material. Roll compaction is widely used for compaction of drugs
and drug formulations, inorganic materials, and herbal materials.
Compaction in a roll press is a continuous process. Friction between the
material and roll surface brings the powder towards the narrow space between the rolls
(gap), where the powder is exposed to high stress leading to the formation of
compacted product. If the rolls are smooth, fluted or knurled, the material will be
compacted into dense ribbons (flakes, sheets, strips), whereas pocket rolls will form
briquettes. The space between the rolls, where different mechanisms occur, is
generally divided into three zones: (1) the feeding zone, where the stresses are small
and densification is solely due to rearrangements of particles, (2) the compaction zone,
where the pressing forces become effective and the particles deform plastically and/or
break, and (3) extrusion zone, where a sheet of compacted material exits the rolls.
A major advantage of dry granulation over wet granulation is the absence of
water or any organic solvents (Kleinebudde, 2004). The compaction technique is
easily scalable, which offers conceptual simplicity and low operational costs. The
main process variables are compaction pressure or specific compaction force (i.e.,
compaction force per cm of roll width), speed of feeding screw, and roll speed. The
most difficult problem with the roll press is feeding of the proper quantity of material
into the rolls (Komarek, 1963).
Roll compaction requires low power and maintenance, and its tolerance of
tramp materials make it a very promising replacement for pelleting and cubing (Miles,

30
1984). For example, dry woodchips [hammer milled with a screen size of 4.8 mm
(3/16 in.)], roll compacted, pulverized/screen producing dense flakes (<8 mesh or 2.4
mm) with bulk density of 481 kg m"3 (30 lb ft"3) at 10% moisture content required a
specific energy of 194 MJ/t. For the same final densified product conditions, pelleting
required a specific energy of 304 MJ/t (Miles, 1984).

2.1.6.2.6-1 Scale-up Procedure for Roll Press

From the data collected on a laboratory size (or pilot scale) roll press, it is
possible to estimate the performance of commercial size equipment (Dec and
Komarek, 1997). In order to do this effectively, the following scale up procedure may
be required (Dec and Komarek, 1997).
First, size of commercial equipment has to be determined based on desired
machine capacity. Simple formulas for calculating the briquette or compact production
rate are presented below.

Briquette production rate = Vb x m x nron x Pb x 60/106 (t/h)


Compact production rate = J t x D x L x h x nron x pc x 60/106 (t/h)

Where,
Vb = briquette volume, cm3
m = number of briquettes per revolution
riroii = roll speed, min"1
pb or p c = briquette or compact density, g cm"
D = roll diameter, cm
L = roll width, cm
h = roll gap, cm

The roll speed (nron) has to be determined by selecting one of the following
scaling up criteria: either the angular or tangential velocity of the rolls. Scaling up

31
with constant value of tangential velocity is safe, but it significantly limits the capacity
of a machine with large roll diameter. Scaling up with constant angular velocity of the
rolls can sometimes cause problems with lower commercial briquette or compact
quality compared with the test results.
The second important criterion would be the compaction work per unit weight
of the material. Keeping this variable constant, as machine size is increased,
determines the desired size roll drive for a commercial machine. The last important
criterion to consider is average compaction pressure. To satisfy this criterion, the
desired roll force for a commercial press must be calculated based on optimum roll
separating force (i.e., force per unit width of the roll) recorded during laboratory trials.
The roll separating force can be simply obtained by monitoring pressure in the
hydraulic system supporting the movable roll. With the briquetting process, the
calculated briquette pressure or briquette mean pressure is calculated as roll separating
force divided by the surface determined by the briquette circumferential length and the
roll width.

2.1.6.2-7Mobile densification machines

Wilen et al. (1987) reported a portable pelletization unit consisted of a straw


(barley, rye, wheat, or oat) chopper to which small bales could be fed, feed equipment
for chopped straw, and a flat-die pelleting press. After the pelletization, the pellets
were cooled on a belt cooler. The output was 0.5 to 1 t/h, depending on the raw
material. This equipment did not include a power source. Electric power requirement
was 120 kW. A larger system could pellet 4 to 6 t/h. They also reported that for a unit
producing 3 t/h of straw pellets, the specific consumption of electricity would be 288
to 360 MJ/t.
Biotruck 2000 is a unique, moving vehicle developed in Europe that
continuously performs all of the operations in the field from harvesting agricultural
biomass to pellet production (Hartman, 1996; Klass, 1998). The operating sequence
consists of the integration into one machine of continuous crop harvesting, size

32
reduction to about 0.6-mm pieces, heating the pieces to temperatures between 80 and
120C using the waste heat of the engine, and compressing the heated pieces in a
toothed-wheel pelleting press. No binder is used. The production rate is 3 to 8 t/h with
bulk density of pelleted cereal crops ranging from 300 to 500 kg m"3.
A manually operated closed-end piston press type briquetting machine
(pressure of <0.2 MPa) (Olorunnisola, 2007), and a tractor-powered single-screw
extrusion type densification machine have been developed (Munoz-Hernandez et al.,
2006). Also, Varadharaju and Gothandapani (1998) designed a screw press to pellet
(10-mm diameter) decomposed coir pith. The throughout of this machine was 100
kg/h. These small scale densification machines were developed to make animal feed or
fuel pellets/briquettes. These machines can be easily moved to the biomass source to
densify the biomass materials.
Mobile densification machines have several limitations. The capacity of a
mobile pelleting/briquetting machine is lower (2 to 8 t/h) than the capacity of a baler
(10 to 20 t/h). The maximum capacity of a field cuber would be 5440 t/year whereas
the same cuber size in a stationary mode would have a capacity of 10890 t/year
(Curley et al., 1973). Also, it is difficult to provide the required power. The operating
time is limited depending upon the daylight conditions, and adverse weather.
Experiences from the field cubing and pelleting of biomass lead to the conclusion that
stationary densification processes may be a better option than mobile densification
processes for processing large volumes of biomass.

33
2.2 Laboratory Scale Densification

Laboratory experiments are conducted to answer the fundamental question of


whether the given material can be effectively compacted (pressed). Before selecting or
designing any commercial densification machine, it is necessary to know the
mechanical behavior of the material and the processing conditions. The effect of
process variables (die geometry, speed of compression, temperature, and pressure), the
effect of material variables (moisture content, size and shape of particles, size
distribution of particles, biochemical and mechanical characteristics), and the effect of
additives (binders or lubricants) could be studied using laboratory-scale densification
apparatuses.
Usually, laboratory-scale closed-end dies (6 to 75 mm in diameter) are used to
study the densification behaviors of materials up to a forming pressure of 200 MPa or
higher (O'Doghearty, 1989). A closed-end die densification apparatus was used to
simulate compression type of densification machines such as roll press
briquetting/compaction, piston press, or wafering machine (O'Doghearty, 1989). In
only two studies were laboratory scale open-end dies used to study densification
behaviors to simulate the pelleting/extrusion process (Payne et al., 1973; Munoz-
Hernandez et al., 2006). In both systems, mechanical pressure was applied by a
hydraulic press or universal testing machine. The speed of compression was limited by
these compression machines. For example, Munoz-Hernandez et al. (2006) used a 6.4
mm diameter die at a pellet velocity of 27.2 mm/s (60 to 80 MPa) to compress an
alfalfa hay and corn stover mixture, compared to a field-scale extruder operating at
pellet velocities of 5 to 70 mm/s. Testing with prototypes on a real scale would result
in better end results; however, the cost of the prototypes can be a major limitation.
Adapa et al. (2005) used a single cubing plunger-die (square opening of 30 mm
x 30 mm) assembly to study the cubing characteristics of fractionated alfalfa chops.
The compaction pressure range used was 1.2 to 14 MPa.
O'Doghearty (1989) reported that the mode of compression (loading at
constant stain rate vs. impact loading), speed of compression, die diameter, and mass

34
of sample charged to the die would affect the density and quality of the compacts
prepared using laboratory scale densification apparatuses. However, the speed of
compression (e.g., 12.7 to 50.8 mm/min), die diameter, and mass of sample were
generally kept constant for a particular laboratory scale densification study.

2.3 Process Variables Affecting Densification

Both the characteristics of the feedstock such as moisture content, particle size,
and temperature, and the densification equipment variables such as maximum pressure
created by the equipment, speed of compression, size and shape of the die, and
temperature of the die have been found to affect the densification behaviors of the
powder materials. The effects of some of the important process variables such as
applied pressure, moisture content, particle size, temperature, and binders on the
densification process are reviewed below. A detailed review on the effect of various
densification variables on the quality (strength and durability) of the densified
products is provided in Chapter 3.

2.3-1 Effect of Pressure

The higher the densification pressure, the higher the density, strength, and
durability of densified products. Increasing pressure increases the specific energy
consumption for densification. Bruhn et al. (1959) found that higher pressure (up to 69
MPa) produced more dense and more durable pellets, and pressures lower than 28
MPa produced fragile pellets. Graham and Bilanski (1984) reported that increasing
pressure from 15 to 45 MPa, pellet density increased from 1137 to 1459 kg m"3 for
grass, and from 1267 to 1402 kg m"3 for alfalfa. Khankari et al. (1989b) observed that
increasing the pressure from 50 to 200 MPa greatly increased the density and relative
stability to impact forces, and reduced the free volume expansion of the pellets made
from rice hulls.

35
2.3-2 Effect of Moisture Content

Moisture content of the feedstock plays a major role in determining the


density, durability, and strength of the densified products. Pickard et al. (1961)
showed that 10% (w.b.) moisture content hay produced better quality compacted hay
than 20% (w.b.) moisture content hay. Chancellor (1962) reported that the higher the
moisture content of alfalfa hay, the lower the compacted density. Gustafson and
Kjelgaard (1963) studied the compaction of hay for a wide range of moisture contents
(28 to 44% w.b.) and found that the density of the product decreased as the moisture
content increased. Mohsenin and Zaske (1976) investigated the compaction of alfalfa
in the moisture content range of 8 to 25% (w.b.). They found that alfalfa pellets made
at 19% (w.b.) moisture content had the highest durability. Haussmann (1975) reported
that 15% (w.b.) moisture content is favorable for briquetting sawdust, sander dust,
wood shavings, and peanut hulls. Orth and Lowe (1977) reported that grass briquettes
produced through extrusion had the highest density in the moisture content range of 14
to 16% (w.b.). Wamukonya and Jenkins (1995) produced high quality briquettes from
agricultural residues and wood wastes in the moisture content range of 12 to 20%
(w.b.).
According to Grover and Mishra (1996a), water acts as a film type binder by
strengthening the bonding in briquettes. Also, water helps in promoting bonding by
van der Waals' forces by increasing the true area of contact of the particles. The right
amount of moisture develops self-bonding properties in lignocellulosic substances at
elevated temperatures and pressures prevalent in briquetting machines (Shen, 1987).
Mani et al. (2002b) reported that increasing the moisture content decreased the density
of pellets made from wheat straw, barley straw, corn stover, and switchgrass.

36
2.3-3 Effect of Particle Size

The general trend is that the smaller the particle size, the better the quality of
the densified products. Payne (1978) suggested that medium- or fine-ground feed
constituents would be desirable in pelleting because these sizes would provide greater
surface area for moisture addition during steam conditioning. Shen (1987) reported
that quality of the pellets made from forestry residues with particle size of 1.7 mm was
consistently better than those made with the particle size of 5 mm. Hill and Pulkinen
(1988) noted that an increase in hammer mill screen size used for grinding alfalfa from
2.8 to 6.4 mm reduced the durability of alfalfa pellets by more than 15%; however, the
power consumption decreased by about 10%. Lindley and Vossoughi (1989) found
that grinding the biomass material to smaller particle size resulted in briquettes with
higher density, better durability and lower water absorption rate. Grover and Mishra
(1996a) observed that presence of different size particles improved the compaction
dynamics and contributed to high strength of biomass briquettes. Mani et al. (2002b
and 2004) reported that corn stover pellets made from the grind obtained from a
hammer mill screen with 1.6 mm openings were 5 to 16% more dense than the pellets
made from the grind obtained from a 3.2-mm hammer mill screen.

2.3-4 Effect of Temperature

Apart from the frictional heat generated due to compression, heat is added to
the densification process either by means of preheating the feed material or by heating
the die. Heating biomass materials softens biomass constituents such as lignin and
protein and improves particle bonding during densification. Reece (1966) reported that
pre-heating hay to 60-70C produced more stable compacted particles than unheated
hay. Hall and Hall (1968) observed that the higher the temperature, the lower the force
needed to provide a given degree of compaction. Smith et al. (1977) produced wheat
straw briquettes with high relaxed densities of 1200 to 1300 kg m"3 at die temperatures
of 80 to 140C for a heating time of 40 min. Edwards and Tarn (1987) found that the

37
density of compacted western hemlock sawdust increased with increasing temperature
up to 190C. They also found that the effect of pressure was relatively small when
temperature was 120C or higher. Hill and Pulkinen (1988) estimated that increasing
the feed conditioning temperature from 60 to 104C increased pellet durability by 30
to 35% during alfalfa pelleting. Khankari et al. (1989b) found that increasing
temperature of rice hulls from 30 to 225C increased the pellet density by 81.7% at a
compaction pressure of 118 MPa. Aqa and Bhattacharya (1992) found that pre-heating
sawdust to 130C and heating the die to 300C reduced the specific energy
consumption by 41.2%. Grover and Mishra (1996b) found that the briquette density of
mustard stalk increased by 8.9 times when heated to 120C compared to the density of
briquettes made at room temperature under a compaction pressure of 25 MPa. They
mentioned that the resistance of the biomass material to compaction decreased when
the temperature was increased.

2.3-5 Effect of Binders

Binders help reduce the elastic spring-back, and increase the strength and
durability of densified products. Some of the constituents of biomass material such as
starch, protein, and lignin act as binders under the right compression conditions. Under
high pressure, the protein and pectin components might be squeezed out of the stem
and leaf walls, which would provide better bonding and stabilization of the compacted
material (Pickard et al., 1961; Reece, 1966; Rehkugler and Buchele, 1968). Mohsenin
and Zaske (1976) reported that binding of alfalfa pellets was caused by the liberation
of protoplasm during compression.
Binders may also be added as foreign materials. Butler and McColly (1959)
used three binding agents (bentonite, Ceredex No. 265, and molasses) and found no
improvement in the density of hay compacts. The most widely used binders for
pelleting of animal feeds are calcium lignosulfonate, colloids, bentonite, starches,
proteins, and calcium hydroxide (Pfost, 1964; Tabil, 1996). Pfost (1964), and Pfost
and Young (1973) reported that there was a significant increase in feed pellet

38
durability when they added about 2.6% (by weight) of colloids and calcium
lignosulfonate as binders. Hill and Pulkinen (1988) tested six binding agents
(bentonite, lignosulfonate, lignosite, neutralized liquid lignosite, liquid molasses, and
barley grain) for alfalfa pelleting and found no significant improvement in pellet
durability. In their review paper, Sokhansanj and Wood (1991) concluded that the use
of binding materials to increase the pellet density and durability did not appear to be
practical. Tabil et al. (1997) studied the performance of different binders during alfalfa
pelleting. They concluded that addition of binders to low quality alfalfa grind
significantly improved the pellet durability and hardness to levels comparable to that
of pellets from high quality chop. They also suggested that use of 0.5% of either
hydrated lime (calcium hydroxide) or pea starch as binders would be sufficient for the
production of the most durable pellets.

2.4 Specific Energy Consumption for Densification

O'Doghearty (1989) reviewed specific energy required for densification of


various biomass materials. Specific energy required for laboratory-scale closed-end
dies ranged from 5 to 40 MJ/t for compressing biomass materials such as barley straws
and cotton stalks (Abd-Elrahim et al., 1981; O'Doghearty, 1989). A fraction of the
energy required is used to overcome the friction in pushing the densified product out
of the die after compression. Bellinger and McColly (1961) reported that the pushing
energy for circular dies was up to 2 MJ/t for alfalfa, which was about 10-15% of the
total energy applied. For ejecting briquettes from the circular dies, the specific energy
ranged from 0.2 to 2 MJ/t for biomass materials such as wheat straw and lucerne
(O'Doghearty, 1989). This is about 2 to 15% of the forming energy.
A field hay cubing machine required a specific energy of 44 MJ/t
(O'Doghearty, 1989). Densification of straw in a reciprocating piston press required
specific energy of 277 MJ/t (O'Doghearty, 1989). A ring-die extrusion press required
specific energy of 120 to 220 MJ/t for grass and lucerne, and 90 to 180 MJ/t for straw
(O'Doghearty, 1989). Samson et al. (2000) reported that a pelleting process required

39
108 MJ/t, 162 MJ/t, and 300 MJ/t for pelleting alfalfa, switchgrass, and straw,
respectively. Komarek (1963) reported that the specific energy required for roll press
briquetting of materials that require low pressure (e.g., coal), medium pressure (e.g.,
organic solids), high pressure (e.g., lime), and very high pressure (e.g., metal powders)
was 8-16, 16-32, 32-64, and >64 MJ/t, respectively.
Reed et al. (1980) showed that the work of compression measured in
laboratory closed-end dies is lower by a factor of 2 to 10 than that consumed in
operating compression machines (e.g., roll press briquetting machine). This is to be
expected because the work measured in a laboratory scale densification apparatus does
not include motor and bearing losses associated with commercial equipment, and the
laboratory measurements are made under idealized conditions. Therefore, the
laboratory results probably represent a lower limit to the work required.
In a typical biomass pelleting operation, manufacturing one tonne of pellets
may require 300-3500 MJ for drying, 100-180 MJ for grinding, and 100-300 MJ for
densification (Mani et al., 2005b). The total energy requirement for a wood pelleting
process (hammer milling, pelleting, screening, and pellet cooling) was estimated at
1690 MJ/t (Mani et al., 2006). In general, 1-3% of the energy contained in the
feedstock is required for densification, 2-3% for size reduction, and 7-15% for
feedstock drying depending on the initial moisture content of the raw biomass.
Moreover, the energy efficiency of biomass densification is in the range of 88-93%
(Balatinecz, 1983).

2.5 Densification Cost Economics

In the U.S., the wholesale cost of wood waste pellets was in the range of $85 to
$ 140/t (Klass, 1998), and the cost of premium wood pellets (i.e., pellets produced
from hardwood or softwood sawdust containing no tree bark) was $ 176/t (Holt et al.,
2004).
Sokhansanj and Turhollow (2004) estimated that the cost for dry corn stover
bales including final grinding was $60/t and the cost for the corn stover cubes was

40
$80/t delivered to an end user located 64 km from the farm. Both costs include a
payment of $11/ dry tonne to the farmer, and the cubing cost includes drying cost and
profit. The cost of the cubes after cubing is less than the cost for handling bales. The
difference in the costs of the two systems, after an intermediate storage, is in the way
each material is handled. Bales require stacking and unstacking both at truck loading
and at the storage sites, whereas cubes are loaded and unloaded in bulk. Baled biomass
needs to be ground at the conversion plant before using while cubes likely do not need
this process. Therefore, opportunities exist to reduce the cost of cubing to levels equal
to baling operations (Sokhansanj and Turhollow, 2004).
Some of the ways the cost of cubing can be reduced are: (i) grinding the
biomass in field to a bulk density of about 128 dry kg m"3 would avoid the baling cost,
although the density of the ground biomass would not be increased appreciably
compared to that of bales; (ii) operating the cubing facility >300 days/year with
multiple feedstocks; and (iii) achieving higher density cubes with higher cuber
throughput.
Samson et al. (2000) reported that depending upon the raw material cost,
switchgrass pellets cost from $72/t to $102/t, without including drying cost. Thek and
Obernberger (2004) reported that the wood pellet production cost in Sweden and
Austria was $78/t and $113/t, respectively. The main cost difference between the two
countries was due to the larger plant capacity and the lower electricity price in
Sweden.
Mani et al. (2006) estimated that the cost of biomass pellets for a plant capacity
of 6 t/h was $51/t of pellets assuming a raw material cost of $10/t and drying biomass
from 40% to 10% (w.b.) moisture content using dry wood shavings as fuel in rotary
drum dryers. Pellet plants with a capacity of more than 10 t/h decreased the costs to
roughly $40/t of pellets. They found that the raw material cost was the largest,
followed by personnel cost, drying cost, and pelleting mill cost. At a raw material cost
of $50/t, the pellet production cost with 20% profit was $132/t (i.e., $8/GJ), which is
almost equal to the current natural gas price (Mani et al., 2006).

41
Jones and Jones (1980) reported that the densified product cost at larger
pelleting plants (18 ton/h) compared well with the cost of imported, unsubsidized oil
in Canada. The use (combustion / gasification) of densified wood fuel would
considerably reduce the capital costs of the conversion systems compared to the use of
wet chips. These cost reductions and the advantages of fewer handling problems,
better direct combustion and gasification characteristics, and the reduction in storage
space and transportation costs offset the high cost of the densified wood fuel (Jones
and Jones, 1980).
The major factors affecting pellet cost are feedstock cost, drying cost, plant
capacity (economies of scale), and plant operating time (Jones and Jones, 1980; Thek
and Obernberger, 2004; Mani et al., 2006). Use of emission control devices may
increase the pellet production cost. Cost of pelleting can be minimized if drying of
biomass, steam conditioning, and additives (binders) are avoided for production of
biomass pellets. Mani et al. (2006) evaluated the cost of pellets when drying the
biomass using several fuels such as wet sawdust, dry biomass, fuel pellets, natural gas,
and coal. They concluded that wet sawdust and coal resulted in lowest pellet
production cost. Steam conditioning cost may be reduced by densifying the material
after drying while its temperature is still reasonably high. Also, waste heat from a
heating plant or combined heat and power plant can be used for heating the feedstock
if the densification plant is integrated with the other heat production plants. By mixing
other biomass materials or waste papers with the feedstock, the requirement of
additives can be avoided.
Future work is required to estimate the cost of densification of agricultural
residues with various initial moisture contents involving collection, baling, and
transportation to an intermediate storage located more than 7.5 km from the farmstead.
The densification plant may be located such that the biomass comes from various
biomass collection sites year around to the plant as well as close to an end user such as
an ethanol plant.

42
2.6 Bonding Mechanisms

2.6-1 Wood

To produce sufficient bonding area, especially in the absence of a binder, the


plasticization of wood polymers above their glass transition temperature is necessary
(Back, 1987). Hydrogen bonding at lignin and cellulose surface areas is considered to
be responsible for the main type of bonding in the press-drying operation of wood.
Covalent bonds are thought to form between the wood polymer chains in the inter-
fiber bonding area of hardboard during hot pressing operations. The London-van der
Waals dispersion forces are considered to be of some importance. The covalent bonds
are the strongest, the hydrogen bonds intermediate, and the non-polar van der Waals's
forces weakest (Back, 1987).
The hydrogen bonds and other secondary bonds predominate in inter-fiber
bonding areas. The strength of these bonds falls off significantly with increasing
moisture present in the bonding area. Hemicellulose is the part of the raw material that
swells most easily. In amorphous areas, secondary bonds are broken by water layers.
Covalent forces are not broken, or can be broken to a minor extent by excessive
swelling forces. Covalent bonds are most evident as an increase in wet strength and
wet moduli (elastic and shear), and also swelling restrictions. Lignin surfaces are not
as active in hydrogen bonding as carbohydrate surfaces. However, lignin appears to be
the most reactive wood polymer for auto-crosslinking reactions. Fats are helpful for
auto-crosslinking between polymers, thereby increasing the strength and moisture
resistance of hardboards. A low pH would increase the rate of auto-crosslinking
(Back, 1987).
Oleophilic materials such as fats, resin, fatty acids, and waxes have a low
cohesive strength. Only non-polar London dispersion forces usually are active
cohesion forces in such materials. A low strength only is achievable wherever such
particles or surface layers are present in the fiber to fiber interface (Back, 1987).

43
Inter-polymer bond strength falls off with increasing temperature. At the glass
transition temperature or main softening temperature, this intermolecular bonding is
reduced very significantly while the chain mobility and the inner volume is increased
to such an extent that chain ends and larger chain segments are able to rotate around
their axes. Also, within the span of 50C above the glass transition temperature, the
viscosity of a polymer drops significantly to show pronounced flow characteristics.
Thus, diffusion of polymer chains and chain ends from one fiber into the proximity of
an adjacent fiber is greatly facilitated, promoting bonding area, especially under
applied pressure. On cooling, these bonds are consolidated (Back, 1987).
In the dry state (i.e., 0% moisture content), the glass transition temperature is
around 220C for amorphous cellulose, around 170C for hemicelluloses, and about
200C for native lignin, naturally varying somewhat with structure and molecular
weight. Water is a plasticizer, or softener, for all these polymers. Thus, it reduces the
glass transition temperature. But native lignin absorbs only a limited amount of water
(about 2%). Above this water content, the glass transition temperature of lignin
remains constant. Cellulose and hemicellulose absorb fairly large amounts of water
and their glass transition temperature drops to 0C or below in wet fibers. In a wet
process, the critical temperature is the glass transition temperature of moist lignin,
approximately 115C (Back, 1987).

2.6-2 Animal Feed Pellets

Thomas and van der Poel (1996) reviewed the binding mechanisms in animal
feed pellets. The general mechanisms for binding feed particles are divided into solid-
solid interactions between particles, capillary forces in a three-phase system of water,
air, and solid material, so-called "liquid necking", adhesive and cohesive forces
between particulates and binders, and interactions between particles due to folding and
plying. The liquid necking is a solid bridge between feed particles created by
solubilisation and subsequent crystallization (or glass transition) of feedstuff

44
components such as starch, sugars, and fats. They concluded that binding in pellets
most probably is due to liquid necking.
The quality of feed pellets (durability and hardness) is affected by the
ingredient composition of the feed diet. The natural binding agents thought to improve
pellet quality include starch, protein, and lignin. Oil or fat was found to reduce pellet
quality. Differences in the physico-chemical properties of feed due to processing
history, geographical and climatic origin, and cultivar would show some differences in
the quality of pellet produced due to possible alteration on the binding action (Thomas
etal., 1998).

2.6-3 Biomass Materials

The characteristics and properties of biomass such as species, variety, time


from harvest (storage effect such as deterioration), particle size, moisture content, and
temperature have a considerable effect on its compression behavior (O'Doghearty,
1989). The following three possible mechanisms may contribute to the formation of a
durable briquette in hay materials (Pickard et al., 1961): (1) complete crushing of the
plant stems, (2) adhesion of stems, and (3) interlacing of stem and leaf material. For
lucerne, the degree of adhesion increased with increasing moisture content. This effect
was attributed to the maceration of the stem and leaf materials, and extrusion of
protein and pectin from the plant cells (Pickard et al., 1961). Similarly, Rehkugler and
Buchele (1969) reported that the percentage of protein and the relative concentration
of pectin, together with the percentage of stem material, were important factors
affecting the adhesion of biomass particles. Pickard et al. (1961) also found that at
high moisture of 30% (w.b.), coherent straw briquettes could not be produced because
the cell structure remained largely intact at high moisture levels so that the stems were
not flattened. Huang and Yoerger (1961) macerated lucerne before briquetting to
break down the cell structure and release the cell contents so as to provide greater
adhesion and less resilience of the material. Thus, briquettes of higher density were
formed from macerated material than from unmacerated material.

45
Tabil (1996) reported that alfalfa particles compact in three stages: particle
rearrangement, elastic-plastic deformation, local melting of materials or melting of
asperities of the particles which depends on the applied pressure, the area and number
of contact points, and is enhanced by increased temperature. Melting of asperities of
the alfalfa particles would form solid bridges upon cooling, thus creating strong
pellets.
At high temperature, plasticizing of protein in alfalfa would help in the binding
of particles. Soluble fiber in alfalfa may contribute to the adhesion of particles during
pelleting. Tabil (1996) found that the mechanical (compaction) characteristics of
alfalfa were more important than the binding contribution from the chemical
components to produce durable pellets from alfalfa chops with different qualities.
Water is one of the most useful agents that is employed as a binder and
lubricant. Water is particularly suitable as an aid in briquetting mixtures containing
water-soluble constituents such as starches, sugars, soda ash, sodium phosphate,
potassium salts, and calcium chloride (Moore, 1965). Water acts as a film type binder
by strengthening and promoting bonding via van der Waal's forces by increasing the
contact area of the particles. A thin film of water around the particles would exhibit
bonds via capillary sorption between particles. With the help of heat, water induces a
wide range of physical and chemical changes such as thermal softening of the
biomass, denaturation of proteins, gelatinization of starch, and solubulization and
consecutive recrystallisation of sugars and salts. These physico-chemical changes
affect binding properties of the biomass particles. The optimum moisture content for
pelleting cellulosic material may range from 8 to 12% (w.b.).
If the base powder has appreciable solubility in the liquid binder (e.g., water),
the liquid binder wets and spreads in the interstices between primary particles, forming
liquid bridges that hold particles together by capillary and viscous forces. Due to the
subsequent drying, the liquid evaporates from the bridges to leave solid bridges
(necks) between particles. The solid bridges will be formed by recrystallization (or
precipitation) of the base powder. The solid bridges impart mechanical strength to the
compact (Bika et al., 2005).

46
2.6.3-1 Biomass compositions

Lignocellulose is the term used to describe the three-dimensional polymeric


composites formed by plants as structural material (cell wall). It consists of variable
amounts of cellulose, hemicellulose, and lignin. Lignin and hemicellulose form a
sheath that surrounds the cellulose portion of the biomass (Bodig and Jayne, 1982).
Hardwoods, softwoods, and herbaceous materials have distinct compositions from one
another.
The organic constituents of wood can be classified as either (1) cell wall
components or (2) extraneous substances, commonly called extractives (Bodig and
Jayne, 1982). The cell wall components are the structural members of the wood cell,
and largely govern the physical properties of wood. Although normally not present in
large amounts, extractives can be important modifiers of physical properties.
Extractives sometimes affect specific gravity and equilibrium moisture content, and
therefore indirectly alter a number of related physical properties. Extractives include
resins, fats and fatty acids, phenolics, phytosterols, and other compounds. The
extractives content depends on the plant species.
Cell wall components can be classified as either carbohydrates or phenolics.
All the carbohydrates are essentially linear polymers (polysaccharides) and account
for roughly 75% of the woody substance. Most of the phenolic material is lignin, a
three-dimensional, complex polymeric substance. Hardwoods contain more cellulose
(45% vs. 41%) and less lignin (22% vs. 28%) than do softwoods. Their total
hemicellulose contents are approximately the same (30%).
Cellulose is the primary component of the cell wall. It is the single most
abundant organic chemical in nature. It retains the same structure regardless of
whether its source is wood, grass, or other plants. Structurally cellulose is the simplest
of the cell wall components - a linear polymer of glucose units. Hemicellulose is
composed of two carbohydrates polymers, xylose and mannose. Hemicelluloses are
more closely associated with lignin, and are in an amorphous state. Lignin is a
phenolic polymer, differing from the carbohydrates in its water repellency. Lignin is

47
completely amorphous and under normal conditions begins to soften at temperatures
of 165-175C (Bodig and Jayne, 1982). Water depresses and broadens the range of
softening temperatures of lignin. Consequently molecular displacement by visGOus
flow under pressure becomes more prevalent under conditions of elevated temperature
and moisture content. Very fine mechanical milling of wood can liberate "milled
wood" lignin, which is thought to be very close to natural lignin in composition and
chemistry (Brown, 2003).
The ash content of biomass materials is an indicator of mineral constituents,
which are primarily calcium and magnesium carbonates, oxalates, and silica crystals
(Bodig and Jayne, 1982).
The effect of biomass compositions on the quality (strength and durability) of
densified products is discussed in Chapter 3.

2.7 Commercial Binders and Lubricants

To improve both the compaction process and final product quality of densified
products, binders or lubricants are used (Dec, 2004). Binders are additives that
increase the strength of the agglomerates, whereas lubricants decrease the coefficient
of friction between the compacted material and the densification machine surface
and/or between the particles during compaction.
Based on Pietsch (1991), binders can be divided into three basic groups: (1)
matrix type (e.g., paraffin, clay, dry starch, and dry sugars); (2) film type (e.g., starch,
bentonite, gums, lignosulfonates, and alginates); and (3) chemical binders (e.g.,
calcium hydroxide). The effectiveness of matrix binders depends on how well the
particles are embedded in a more or less continuous matrix. Film-type binders require
water to be effective. Water will dissolve the surface of the crystals or particles of
soluble material, and when evaporated, it causes recrystallization across the particle
boundaries. The effectiveness of the chemical binders depends on the chemical
reaction between the binder and the material being agglomerated.

48
Lubricants that are mixed with the material are classified as internal and are in
general added as solids (e.g., talc, magnesium stearate, and dry starch). If they are
used to lubricate the machine surface, they are called external and are applied mostly
in a liquid form (e.g., water, oils, and glycerine) (Pietsch, 1991).
Proper selection of binders or lubricants is very important and in many cases
critical to both technical and economic feasibility of briquetting or compacting
processes (Dec, 2004). The chemistry of the material to be compacted and the
considered binder or binder system must be studied to determine their compatibility.
Contamination of the product can be introduced by binders containing sulfur,
phosphorous, silica or other substances. Use of any additives can be costly and will
require additional equipment for proportioning, mixing, and handling. Considering the
cost and many limitations in the choice of a binder, there is tendency to use higher
compaction pressure or increase temperature of the feed material to eliminate the
necessity for binders (Dec, 2004).
In the animal feed manufacturing industry, binders are used to produce good
quality pellets to avoid crumbling upon handling (Tabil, 1996; Thomas et al., 1998).
Bentonite is a colloidal hydrated magnesium aluminum silicate clay used for making
animal feed. Sodium bentonite is preferred over calcium bentonite due to the high
swelling capacity of sodium bentonite in water. Predominantly used pellet binders are
lignin sulfonates (or lignosulfonates), which are sulfonate salts from sulfite pulping in
paper manufacturing. Starches and proteins are added as binders. With adequate heat
and moisture, starches and proteins are activated during pelleting to result in binding.
Calcium hydroxide (hydrated lime) is also used for animal feed pelleting. Molasses
(byproducts from cane sugar manufacturing) are also claimed to have binding
potentials for densifying animal feed, coal, and steel (Siegerink, 2007).
Carboxymethylcellulose (CMC) is used as a binder in animal feed production. CMC
gives a viscous solution/paste when mixed with water. Hydration occurs during
conditioning. Subsequent compaction brings particulates close together and allows for
the development of ionic attractions between the CMC and feed particulates (Thomas
etal., 1998).

49
2.8 Strength of Agglomerates

One of the most important characteristics of agglomerates is their strength. The


strength of agglomerates is frequently determined by crushing, drop, abrasion, impact,
bending, cutting, and shear strength tests. The strength values given by these tests are
empirical and cannot be predicted by theory because it is not known which stress
component causes the agglomerate to fail.
Three models have been proposed to model strength of the agglomerates based
on a few selected binding mechanisms (Rumpf, 1962; Pietsch, 1991). The three
models are: (1) the entire pore volume of the agglomerate is filled with a substance
(e.g., hardening binder) that transmit forces and causes strength. In this model, the
agglomerate strength is the lowest of the strength of the binder substance included, the
strength of the adhesion between binder and particles, and the strength of particles
forming the agglomerate; (2) the pore volume of the agglomerate is entirely filled with
liquid, where the tensile strength of the agglomerate is approximated by capillary
pressure; and (3) the agglomerate strength is caused by solid bridges, and that the
entire solid binder material is uniformly distributed at all contact and coordination
points of the primary particles forming the agglomerate. Theoretical models are also
available to approximate the adhesion between two individual particles due to a liquid
bridge, van der Waals' forces, electrostatic forces, or magnetic forces (Pietsch, 1991).

50
2.9 Conclusions

For bioenergy or biorefinery applications, a large amount of biomass (about 400 to


4000 t/day) is required for a day of operation of a conversion facility. Densified
biomass such as pellets/briquettes or compacted, granulated biomass appears to be
better than baled or chopped (ground) biomass to reduce the cost and problems
with the handling, transporting, and storing biomass feedstocks for these
applications.
Due to the current interest in renewable energy production from corn stover and
switchgrass, a few preliminary laboratory studies on the densification
characteristics of corn stover and switchgrass have been conducted elsewhere.
However, elaborate laboratory densification studies are required to determine the
densification conditions to produce high density and high quality densified
products from corn stover and switchgrass with minimum input energy.
Among the existing commercial densification machines, either pelleting or roll
press briquetting/compaction machines could be used for densification of large
volume of biomass feedstocks (about 25 to 60 t/h). Some preliminary pelleting
performance of switchgrass has been conducted; however, no pelleting study on
corn stover has been reported in the literature. Studies on roll press
briquetting/compaction of corn stover and switchgrass have not been conducted in
the past.
Studies on cost economics for densifying corn stover and switchgrass are required.
Studies on logistics and cost for year-round supply of densified biomass feedstocks
from multiple sources to a conversion facility are needed.

51
2.10 References

Abd-Elrahim, Y.M., A.S. Huzayyin, and I.S. Taha. 1981. Dimensional analysis and
wafering cotton stalks. Trans. ASAE 24(4): 829-832.
Adapa, P.K., L.G. Tabil, G.J. Schoenau, and S. Sokhansanj. 2005. Physical properties
of cubes from fractionated sun-cured and dehydrated alfalfa chops. ASAE
Paper No. 056161. St. Joseph, Mich.: ASABE.
Adler, P.R., M.A. Sanderson, A.A. Boateng, P.J. Weimer, and H.G. Jung. 2006.
Biomass yield and biofuel quality of switchgrass harvested in fall or spring.
Agron.J. 98(6): 1518-1525.
Aqa, S., and S.C. Bhattacharya. 1992. Densification of preheated sawdust for energy
conservation. Energy 17(6): 575-578.
Arthur, J.F., R.A. Kepner, J.B. Dobie, G.E. Miller, and P.S. Parsons. 1982. Tub
grinder performance with crop and forest residues. Trans. ASAE 25(6): 1488-
1494.
Back, E.L. 1987. The bonding mechanism in hardboard manufacture. Holzforschung
41(4): 247-258.
Baker, J.L., and G.C. Shove. 1978. Solar drying of commercially produced large
round hay bales. ASAE Paper No. 78-3066. St. Joseph, Mich.: ASABE.
Balatinecz, J.J. 1983. The potential role of densification in biomass utilization. In
Biomass Utilization, 181-190. W.A. Cote, ed. New York, NY: Plenum Press.
Balk, W.A. 1964. Energy requirements for dehydrating and pelleting coastal
bermudagrass. Trans. ASAE 7(3): 349-351, 355.
Bellinger, P.L., and H.H. McColly. 1961. Energy requirements for forming hay
pellets. Agri. Eng. 42(5): 244-247.
Beven, R.R. 1977. A Review of Industrial Size Reduction Equipment Used in the
Processing of Coal. Report prepared for the United States Energy Research and
Development Administration under contract EX-76-C-01-2475. Danville, PA:
Kennedy Van Saun Corporation.

52
Bhattacharya, S.C., R. Bhatia, M.N. Islam, and N. Shah. 1985. Densified biomass in
Thailand: potential, status and problems. Biomass 8(4): 255-266.
Bika, D., G.I. Tardos, S. Panmai, L. Farber, and J. Michaels. 2005. Strength and
morphology of solid bridges in dry granules of pharmaceutical powders.
Powder Technol. 150(2): 104-116.
Bodig, J., and B.A. Jayne. 1982. Mechanics of Wood and Wood Composites. New
York, NY: Van Nostrand Reinhold Company.
Brown, R.C. 2003. Biorenewable Resources, Engineering New Products from
Agriculture. Ames, IA: Iowa State Press.
Bruhn, H.D., A. Zimmerman, and R.P. Niedermeier. 1959. Developments in pelleting
forage crops. Agric. Eng. 40(2): 204 - 207.
Butler, J.L., and H.F. McColly. 1959. Factors affecting the pelleting of hay. Agric.
Eng. 40(8): 442-446.
Carson, J.M., and B.G. Kreider. 1988. Forced heated air drying of hay. ASAE Winter
Meeting Paper No. 88-6583. St. Joseph, Mich.: ASABE.
Chancellor, W.J. 1962. Formation of hay wafers with impact loads. Agric. Eng. 43(3):
136-138, 149.
Colley, Z., O.O. Fasina, D. Bransby, and Y.Y. Lee. 2006. Moisture effect on the
physical characteristics of switchgrass pellets. Trans. ASABE 49(6): 1845-
1851.
Cundiff, J.S., and L.S. Marsh. 1996. Harvest and storage costs for bales of switchgrass
in the southeastern United States. Bioresource Technol. 56(1): 95-101.
Cundiff, S.J., P.P. Ravula, and R.D. Grisso. 2004. Management system for biomass
delivery at a plant conversion. ASAE Paper No. 046169. St. Joseph, Mich.:
ASABE.
Curley, R.G., J.B. Dobie, and P.S. Parsons. 1973. Comparison of stationary and field
cubing of forage. Trans. ASAE 16(2): 361-364, 366.
De Kam, M.J., R.V. Morey, and D.G. Tiffany. 2007. Integrating biomass to produce
heat and power at ethanol plants. ASABE Paper No. 076232. St. Joseph,
Mich.: ASABE.

53
Dec, R.T. 2004. Processing of industrial wastes in the roller press for recovery, recycle
or safe disposal. REWAS'2004 Global symposium on recycling, waste
treatment and clean technology, Madrid, Spain, 26-29 September, 2004.
Dec, R.T., and R.K. Komarek. 1997. Testing and scaling up compaction process using
small laboratory roll press. The Institute for Briquetting and Agglomeration
25th Biennial Conference, Charleston, SC, November 3-5, 1997.
Drzymala, Z. 1993. Industrial Briquetting - Fundamental and Methods. Studies in
mechanical engineering. Vol. 13. Warszawa: PWN-Polish Scientific
Publishers.
Edens, W.C., L.O. Pordesimo, and S. Sokhansanj. 2002. Field drying characteristics
and mass relationships of corn stover fractions. ASAE Paper No. 026015. St.
Joseph, Mich.: ASABE.
Edwards, W.C., and P. Tarn. 1987. Equipment for production of a water resistant
densified fuel from forest biomass. In Proc. of the Sixth Canadian Bioenergy R
& D Seminar, 214-218. London: Elsevier Applied Science.
Graham, V.A., and W.K. Bilanski. 1984. Non-linear viscoelastic behavior during
forage wafering. Trans. ASAE 27(6): 1661-1665.
Grover, P.D., and S.K. Mishra. 1996a. Biomass Briquetting: Technology and
Practices. Regional Wood Energy Development Program in Asia, Field
Document No. 46. Bangkok, Thailand: Food and Agriculture Organization of
the United Nations.
Grover, P.D., and S.K. Mishra. 1996b. Proceedings of the International Workshop on
Biomass Briquetting. Regional Wood Energy Development Program in Asia,
RWEDP Report No. 23. Bangkok, Thailand: Food and Agriculture
Organization of the United Nations.
Gustafson, A.S., and W.L. Kjelgaard. 1963. Hay pellet geometry and stability. Agric.
Eng. 44(8): 442 - 445.
Hall, G.E., and C.W. Hall. 1968. Heated die wafer formation of alfalfa and betmuda
grass. Trans. ASAE \\{A): 578-581.

54
Hartman, H. 1996. The self-propelled briquetting machine for biofuels - features and
chances of the Haimer-Biotruck 2000. In Biomassfor Energy and the
Environment. Proceedings of the 9th European Bioenergy Conf, Copenhagen,
Denmar, 24-27 June 1996, 839-845. Amsterdam: Pergamon Press, Elsevier.
Haussmann, F. 1975. Briquetting wood waste by the Fred Haussmann method.
Institute for Briquetting and Agglomeration Proceedings, Vol. 14. pp. 75-90.
Hill, B and D.A. Pulkinen. 1988. A Study of the Factors Affecting Pellet Durability
and Pelleting Efficiency in the Production of Dehydrated Alfalfa Pellets. A
Special Report. Tisdale, SK, Canada: Saskatchewan Dehydrators Association.
Holley, C.A. 1983. The densification of biomass by roll briquetting. Proceedings of
the IBA 18th Biennial Conference, 95-102.
Holt, G.A., J.L. Simonton, M.G. Beruvides, and A. Canto. 2004. Utilization of cotton
gin by-products for the manufacturing of fuel pellets: an economic perspective.
AppliedEng. inAgric. 20(4): 423-430.
Hoque, M., S. Sokhansanj, T. Bi, S. Mani, L. Jafari, L. Lim, P. Zaini, S. Melin, T.
Sowlati, and M. Afzal. 2006. Economics of pellet production for export
market. CSBE/SCGAB Paper No. 06-103. The Canadian Society for
Bioengineering, CSBE/SCGAB 2006 Annual Conference, Edmonton, Alberta,
Canada.
Hoskinson, R.L., D.L. Karlen, S.J. Birrell, C.W. Radtke, and W.W. Wilhelm. 2007.
Engineering, nutrient removal, and feedstock conversion evaluations of four
corn stover harvest scenarios. Biomass and Bioenergy 31(2-3): 126-136.
Huang, B.K., and R.R. Yoerger. 1961. Maceration as a pretreatment in hay wafering.
Trans. ASAE 4(1): 69-71.
Jannasch, R., Y. Quan, and R. Samson. 2001. A Process and Energy Analysis of
Pelletizing Switchgrass. Final Report. Ste. Anne de Bellevue, QC: Resource
Efficient Agricultural Production (REAP-Canada, www.reap-canada.com).
Jones, D., and J. Jones. 1980. Wood chips vs. densified biomass: an economic
comparison. Symposium papers "Energy from Biomass and Wastes IV". Hotel

55
Royal Plaza, Lake Buena Vista, Florida, Jan 21-25, 1980. Sponsored by
Institute of Gas Technology, Chicago, pp. 223-249.
Kadam, K.L., and J.D. McMillan. 2003. Availability of corn stover as a sustainable
feedstock for bioethanol production. Bioresource Technol. 88(1): 17-25.
Kaltschmitt, M., and M. Weber. 2006. Markets for solid biofuels within the EU-15.
Biomass and Bioenergy 30(11): 897-907.
Khankari, K.K., R.V. Morey, J.E. Fruin, and D.W. Halbach. 1989a. Engineering cost
estimates for pelleting Minnesota soybean meal. Department of Agricultural
and Applied Economics Staff Paper No. P89-42. St. Paul, MN: University of
Minnesota.
Khankari, K.K., M. Shrivastava, and R. Weggel. 1989b. Densification characteristics
of rice hulls under hot uniaxial compression. Institute for Briquetting and
Agglomeration Proceedings, Vol. 21. pp. 143-147.
Klass, D.L. 1998. Biomass for Renewable Energy, Fuels, and Chemicals. San Diego,
CA: Academic Press, Elsevier.
Kleinebudde, P. 2004. Roll compaction/dry granulation: pharmaceutical applications.
Eur. J. Pharm. Biopharm. 58(2): 317-326.
Komarek, K.R. 1963. The roll type briquette machine. Proceedings of the IB A 8th
Biennial Briquetting Conference, 35-37.
Kumar, A., and S. Sokhansanj. 2007. Switchgrass {Panicum virgatum, L.) delivery to
a biorefinery using integrated biomass supply analysis and logistics (IBSAL)
model. Bioresource Technol. 98(5): 1033-1044.
Lindley, J. A., and M. Vossoughi. 1989. Physical properties of biomass briquets.
Trans. ASAE 32(2): 361-366.
Lobo, P. 2002. The right grinding solution for you: roll, horizontal or vertical. Feed
Management 53(3): 23-26.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2002a. Grinding performance and physical
properties of selected biomass. ASAE Paper No. 026175. St. Joseph, Mich.:
ASABE.

56
Mani, S., L.G. Tabil, and S. Sokhansanj. 2002b. Compaction behavior of some
biomass grinds. AIC Paper No. 02-305. Saskatoon, Saskatchewan: AIC 2002
Meeting, CSAE/SCGR Program.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004. Evaluation of compaction equations
applied to four biomass species. Can. Biosys. Eng. 46(3): 3.55-3.61.
Mani, S., S. Sokhansanj, and X. Bi. 2005a. Modeling of forage drying in single and
triple pass rotary drum dryers. ASAE Paper No. 056082. St. Joseph, Mich.:
ASABE.
Mani, S., S. Sokhansanj, X. Bi, and L.G. Tabil. 2005b. Modeling of biomass drying
and densification processes. ASAE Paper No. 056144. St. Joseph, Mich.:
ASABE.
Mani, S., S. Sokhansanj, X. Bi, and A. Turhollow. 2006. Economics of producing fuel
pellets from biomass. Applied Eng. inAgric. 22(3): 421-426.
Miles, T.R. 1984. Biomass preparation for thermochemical conversion. In
Thermochemical Processing of Biomass, 69-90. A.V. Bridgwater, ed. London:
Butterworths.
Mohsenin, N., and J. Zaske. 1976. Stress relaxation and energy requirements in
compaction of unconsolidated materials. J! Agric. Eng. Res. 21(2): 193-205.
Moore, J. 1965. Processing applications for roll-type briquetting-compacting
machines. In Proc.9' Biennial Briquetting Conf., 2-16. Denver, CO:
International Briquetting Association.
Morey, R.V., and D.R. Thimsen. 1980. Two-stage combustion to provide heat for
drying corn. ASAE Winter Meeting Paper No. 80-3507. St. Joseph, Mich.:
ASABE.
Morey, R.V., D.G. Tiffany, and D.L. Hatfield. 2006. Biomass for electricity and
process heat at ethanol plants. Applied Eng. inAgric. 22(5): 723-728.
Muck, R.E., and K.J. Shinners. 2006. Effect of inoculants on the ensiling of corn
stover. ASABE Paper No. 061013. St. Joseph, Mich.: ASABE.

57
Mukunda, A., K.E. Ileleji, and H. Wan. 2006. Simulation of corn stover logistics from
on-farm storage to an ethanol plant. ASABE Paper No. 066177. St. Joseph,
Mich.: ASABE.
Munoz-Hernandez, G., J. Dominguez-Dominguez, and O. Alvarado-Mancilla. 2006.
An easy laboratory method for optimizing the parameters for the mechanical
densification process: an evaluation with an extruder. Agricultural Engineering
International: the CIGR Ejournal, Manuscript PM 06 015. Vol. VIII. July,
2006.
Naylor, J.L., and R. Smith. 1981. Roller mills vs hammermills on corn. ASAE Paper
No. 81-3028. St. Joseph, Mich.: ASABE.
O'Doghearty, M.J. 1989. A review of the mechanical behavior of straw when
compressed to high densities. J. Agric. Eng. Res. 44: 241-265.
Olorunnisola, A. 2007. Production of fuel briquettes from waste paper and coconut
husk admixtures. Agricultural Engineering International: the CIGR Ejournal,
Manuscript EE 06 006. Vol. IX. February, 2007.
Olsson, M., J. Kjallstrand, and G. Petersson. 2003. Specific chimney emissions and
biofuel characteristics of softwood pellets for residential heating in Sweden.
Biomass and Bioenergy 24(1): 51-57.
Orth, H.W., and R. Lowe. 1977. Influence of temperature on wafering in a continuous
extrusion process. J. Agric. Eng. Res. 22(3): 283-289.
Parker, B.F., G.M. White, M.R. Lindley, R.S. Gates, M. Collins, S. Lowry, and T.C.
Bridges. 1992. Forced-air drying of baled alfalfa hay. Trans. ASAE 35(2): 607-
615.
Patil, R.T., S. Sokhansanj, E.A. Arinze, and G.J. Schoenau. 1993. Methods of
expediting drying rates of chopped alfalfa. Trans. ASAE 36(6): 1799-1803.
Payne, F.A., I.J. Ross, H.E. Hamilton, and J.D. Fox. 1973. Short time, high
temperature extrusion of chicken excreta. ASAE Paper No. 72-450. St. Joseph,
Mich.: ASAE.
Payne, J.D. 1978. Improving quality of pelleted feeds. Milling Feed and Fertilizer,
161(5): 34-41.

58
Pfost, H.B. 1964. The effect of lignin binders, die thickness and temperature on the
pelleting process. Feedstuffs 36(22): 20, 54.
Pfost, H.B., and L.R. Young. 1973. Effect of colloidal binder and other factors on
pelleting. Feedstuffs 45(49): 21-22.
Pickard, E.G., W.M. Roll, and J.H. Ramser. 1961. Fundamentals of hay wafering.
Trans. ASAE 4(1): 65-68.
Pietsch, W. 1991. Size Enlargement by Agglomeration. New York, NY: John Wiley &
Sons.
Plue, P.S., and W.K. Bilanski. 1990. On-farm drying of large round bales. Applied
Eng. inAgric. 6(4): 418-421.
Reece, F.N. 1966. Temperature, pressure and time relationships in forming dense hay
wafers. Trans. ASAE 9(6): 749-751.
Reed, T.B., G. Trezek, and L. Diaz. 1980. Biomass densification energy requirements.
In Thermo chemical Conversion of Solid Wastes and Biomass, 169-178. J.L.
Jones, S.B. Radding, S. Takaoka, A.G. Buekens, M.Hiraoka, and R. Overend,
eds. Washington, DC: ASC Symposium Series 130, American Chemical
Society.
Rehkugler, G.E, and W.F. Buchele. 1969. Bio-mechanics of forage wafering. Trans.
ASAE\2{\): 1-8.
Rotz, C.A., and Y. Chen. 1985. Alfalfa drying model for the field environment. Trans.
ASAE2%(5): 1686-1691.
Rumpf, H. 1962. The strength of granules and agglomerates. In Agglomeration, 379-
419. W.A. Knepper, ed. New York, NY: John Wiley and Sons.
Sah, P.C., B.P.N. Singh, U. Chandra, and B.N. Sachchan. 1977. Grinding of wheat
straw in a hammer mill. J. Agric. Eng. (India) XIV (3): 108-112.
Samson, R., P. Duxbury, M. Drisdelle, and C. Lapointe. 2000. Assessment of
Pelletized Biofuels. A Report. Ste. Anne de Bellevue, QC: Resource Efficient
Agricultural Production (REAP-Canada, www.reap-canada.com).
Savoie, P., and S. Descoteaux. 2004. Artificial drying of corn stover in mid-size bales.
ASAE Paper No. 048009. St. Joseph, Mich.: ASABE.
59
Savoie, P., S. Beauregard, and C. Lague. 1990. Rapid hay drying with super hay
conditioning. CSAE Paper No. 90-502. Saskatoon: CSAE.
Schell, D.J., and C. Harwood. 1994. Milling of lignocellulosic biomass: results of
pilot-scale testing. Applied Biochemistry and Biotechnology 45/46: 159-168.
Shen, K.C. 1987. Development of a waterproof densified solid fuel pellet from
forestry residues. In Froc. of the Sixth Canadian Bioenergy R&D Seminar,
209 - 213. London, UK: Elsevier Applied Science.
Shinners, K.J., and B.N. Binversie. 2004. Harvest and storage of wet corn stover
biomass. ASAE Paper No. 041159. St. Joseph, Mich.: ASABE.
Shinners, K.J., R.G. Koegel, and R.J. Straub. 1989. Rapid field drying of forages
utilizing shredded mat technology. In Proc. Eleventh International Congress
on Agricultural Engineering, Dublin, 4-8 September 1989, 2063-2070. V.A.
Dodd and P.M. Grace, eds.
Shinners, K.J., B.N. Binversie, and P. Savoie. 2003. Harvest and storage of wet and
dry corn stover as a biomass feedstock. ASAE Paper No. 036088. St. Joseph,
Mich.: ASABE.
Shinners, K.J., G.C. Boettcher, J.T. Munk, M.F. Digman, R.E. Muck, and P.J.
Weimer. 2006a. Single-pass, split-stream of corn grain and stover:
characteristic performance of three harvester configurations. ASABE Paper
No. 061015. St. Joseph, Mich.: ASABE.
Shinners, K.J., G.C. Boettcher, R.E. Muck, P.J. Wiemer, and M.D. Casler. 2006b.
Drying, harvesting and storage characteristics of perennial grasses as biomass
feedstocks. ASABE Paper No. 061012. St. Joseph, Mich.: ASABE.
Shinners, K.J., D.S. Hoffman, G.C. Boettcher, and J.T. Munk. 2007. Harvesting rate,
power requirements and fuel use for single-pass harvesting of corn stover.
ASABE Paper No. 071018. St. Joseph, Mich.: ASABE.
Siegerink, A. 2007. Molasses as a binder in agglomeration processes. Proceedings of
the Institute for Briquetting and Agglomeration 30th Biennial Conference.
Manitowish Waters, WI: IBA.

60
Smith, I.E., S.D. Probert, R.E. Stokes, and R.J. Hansford. 1977. The briquetting of
wheat straw. J. Agric. Eng. Res. 22(2): 105-111.
Sokhansanj, S., and A. Turhollow. 2002. Baseline cost for corn stover collection.
Applied Eng. in Agric. 18(5): 38-43.
Sokhansanj, S., and H.C. Wood. 1991. Engineering aspects of forage processing for
pellets, cubes, dense chops and bales. Advances in Feed Technology 5: 6-23.
Sokhansanj, S., and A.F. Turhollow. 2004. Biomass densification - cubing operations
and costs for corn stover. Applied Eng. in Agric. 20(4): 495-499.
Sokhansanj, S., A. Turhollow, J. Cushman, and J. Cundiff. 2002. Engineering aspects
of collecting corn stover for bioenergy. Biomass and Bioenergy 23(5): 347-
355.
Sokhansanj, S., J. Cushman, and L. Wright. 2003. Collection and delivery of feedstock
biomass for fuel and power production. Agricultural Engineering
International: the CIGR Journal of Scientific Research and Development.
Invited Overview Paper. Vol. V. February 2003.
Sokhansanj, S., A. Turhollow, S. Tagore, and S. Mani. 2006a. Integrating biomass
feedstock with an existing grain handling system for biofuels. ASABE Paper
No. 066189. St. Joseph, Mich.: ASABE.
Sokhansanj, S., A. Kumar, and A.F. Turhollow. 2006b. Development and
implementation of integrated biomass supply analysis and logistics model
(IBSAL). Biomass and Bioenergy 30(10): 838-847.
Stahl, M , K. Granstrom, J. Berghel, and R. Renstorm. 2004. Industrial processes for
biomass for drying and their effects on the quality properties of wood pellets.
Biomass and Bioenergy 27(6): 621-628.
Svenningson, P.J., and R. Hosier. 1987. Biomass briquettes in the Dominican
Republic. Part II: technical analyses. Biomass 13(4): 275-291.
Tabil, L.G. Jr. 1996. Pelleting and binding characteristics of alfalfa. Ph.D. diss.
Saskatoon, SK, Canada: University of Saskatchewan, Department of
Agricultural and Bioresource Engineering.

61
Tabil, L.G. Jr., S. Sokhansanj, and R.T. Tyler 1997. Performance of different binders
during alfalfa pelleting. Can. Agric. Eng. 39(1): 17-23.
Thek, G., and I. Obernberger. 2004. Wood pellet production costs under Austrian and
in comparison to Swedish framework conditions. Biomass and Bioenergy
27(6): 671-693.
Thomas, M., and A.F.B. van der Poel. 1996. Physical quality of pelleted animal feed.
1. Criteria for pellet quality. Animal Feed Set Tech. 61(1-4): 89-112.
Thomas, M., T. van Vliet, and A.F.B. van der Poel. 1998. Physical quality of pelleted
animal feed 3. Contribution of feedstuff components. Animal Feed Sci. Tech.
70(1-2): 59-78.
USDA. 2002. Crop production - annual summary (PCP-BB). Washington, DC:
National Agricultural Statistical Service, U.S. Department of Agriculture
(USDA).
Varadharaju, N., and L. Gothandapani. 1998. Design and development of equipment
for pelleting decomposed coir pith. Agricultural Mechanization in Asia, Africa,
and Latin America 29(2): 33-34, 38.
Wamukonya, L., and B. Jenkins. 1995. Durability and relaxation of sawdust and wheat
straw briquettes as possible fuels for Kenya. Biomass and Bioenergy 8(3): 175-
179.
Weaver, J.W. Jr., CD. Grinnells, and R.L. Lovvorn. 1947. Drying baled hay with
forced air. Agri. Eng. 28 (7): 301-304.
Wilen, C , P.Stahlberg, K. Siplila, and J. Ahokas. 1987. Pelletization and combustion
of straw. In Energy from Biomass and Wastes X, 469-484. D.L. Klass, ed.
London: Elsevier Applied Science Publishers.
Wiselogel, A.E., F.A. Agblevor, D.K. Johnson, S. Deutch, J.A. Fennell, and M.A.
Sanderson. 1996. Compositional changes during storage of large round
switchgrass bales. Bioresource Technol. 56(1): 103-109.
Yu, M., A.R. Womac, and L.O. Pordesimo. 2003. Review of biomass size reduction
technology. ASAE Paper No. 036077. St. Joseph, Mich.: ASABE.

62
Biomass (e.g., corn Densification of biomass at
stover and switchgrass) Transportation of a central densification
harvesting, collection, bales facility (e.g., pelleting /
and on-farm storage in briquetting / compaction)
bales

i'

Bio-energy and Bio- Transportation of Densified products


products Industries densified biomass
products (pellets / briquettes /
(reserve storage and use compacted, granulated
of densified biomass ^ biomass)
products)

Figure 2.1. Densification of biomass for efficient and economic use of biomass
feedstocks in bio-based industries.

Biomass Chopping Drying chops in Hammer milling


bales (from bales in tub rotary dryers (fine grinding)
farm) grinders

v
Storing / packing Recycling of fines Densification
- (pelleting /
of densified
products briquetting /
4 f compaction)
(ready for
transporting to
bio-based i '
industries) Screen!ngto Cooling / Pellets /
remove fines drying of briquettes /
densified granulated
prodiicts biomas s

Figure 2.2. A typical biomass densification process flow diagram.

63
ure 2.3. Schematics of high pressure agglomeration machines (Pietsch, 1991). (a)
Reciprocating piston and die press, (b) rotary tableting machine, (c) ram
extrusion or plunger press, (d) roller compacting machine, and (e) roller
briquetting machine.

64
ure 2.4. Schematics of pelleting machines (Pietsch, 1991). (a) screw extruder, (b)
pelleting machine with flat-die and muller-type press rollers, (c) pelleting
machine with one solid and one hollow roll, (d) pelleting machine with
two hollow rolls, (e) pelleting machine with ring-die and internal press
roll, and (f) gear-type pelleter.

65
Chapter 3
STRENGTH AND DURABILITY OF DENSIFIED PRODUCTS

3.1 Abstract

The effectiveness of a densification process to create strong and durable


bonding in densified products such as pellets, briquettes, and cubes can be determined
by testing the strength (i.e., compressive resistance, impact resistance, and water
resistance), and durability (i.e., abrasion resistance) of the densified products. These
tests can indicate the maximum force/stress that the densified products can withstand,
and the amount of fines produced during handling, transportation, and storage. In this
Chapter, the procedures used for measuring the strength and durability of the densified
products are discussed. The effects of constituents of the feed such as starch, protein,
fiber, fat, lignin, and extractives; feed moisture content; feed particle size and its
distribution; feed conditioning temperature/preheating of feed; added binders; and
densification equipment variables (forming pressure; and pellet mill and roll press
variables) on the strength and durability of the densified products are reviewed. This
Chapter will help select process parameters to produce strong and durable densified
products from new biomass feedstocks or animal feed formulations. Guidelines for
developing standards on criteria for the acceptance levels of strength and durability of
the densified products are presented.

3.2 Introduction

Currently, there is tremendous interest in using biomass materials in the U.S.


for producing liquid transportation fuels, combined heat and power, chemicals, and
bio-products. In addition to numerous advantages, use of biomass materials in place of
fossil fuels would result in low emissions of greenhouse gases and acid rain-inducing
gases. In order to make the biomass materials available for a variety of applications,
the challenges with the use of biomass materials in their original form must be

66
resolved. Because of high moisture content, irregular shape and sizes, and low bulk
density, biomass is very difficult to handle, transport, store, and utilize in its original
form. One solution to these problems is densification of biomass materials into pellets,
briquettes, or cubes. Densification increases the bulk density of biomass from an
initial bulk density (including baled density) of 40 to 200 kg m"3 to a final bulk density
of 600 to 800 kg m"3 (Holley, 1983; Mani et al., 2003; Obernberger and Thek, 2004;
McMullen et al., 2005). Thus, densification of biomass materials could reduce the
costs of transportation, handling, and storage. Because of uniform shape and sizes,
densified products can be easily handled using standard handling and storage
equipment, and can be easily adopted in direct-combustion or co-firing with coal,
gasification, pyrolysis, and in other biomass-based conversions. Pelleting of animal
feed has shown the following advantages: increased bulk density, improved ease of
handling, better flow properties than the (ground) feed, reduced ingredient
segregation, less feed wastage, improved animal performance (e.g., higher feed intake,
feed conversion, and daily weight gain), lower energy expenditure while eating,
reduced microbial load, and improved palatability (Behnke, 1994; Franke and Rey,
2006).

Generally, two techniques are used for size enlargement of particulate


materials: tumble agglomeration and pressure agglomeration (Pietsch, 2002). In
tumble agglomeration, agglomerates are formed during suitable movement of the
particulate materials containing binder in equipment such as balling discs, balling
cones, and balling drums. Whereas in pressure agglomeration, high forces are applied
to a mass of particulate materials within a confined volume to increase the density.
Conventional processes for pressure-assisted densification can be classified into three
types: extruding, pelleting, and briquetting (Li and Liu, 2000; Pietsch, 2002). In an
extruder, the raw material is compressed by a screw or a piston through a die to form
compacted particles, often of cylindrical shape. The diameter of extrusion products
may range from 20 to 100-mm. Extrusion technologies are commonly used for
biomass densification in Asian countries such as India, Japan, and Thailand

67
(Bhattacharya et al, 1985; Grover and Mishra, 1996; Varadharaju and Gothandapani,
1998).
In pelleting equipment, the feed material is pressed through open-ended
cylindrical holes (dies) made in the periphery of a ring. One to three small rotating
rolls push the feed material into the die holes from inside of the ring towards the
outside of the ring. The skin friction between the feed particles and the wall of the die
resists the free flow of feed and thus the particles are compressed against each other
inside the die to form pellets. One or two adjustable knives placed outside the ring cut
the pellets into desired lengths. The diameter of the pellets may range from 4.8 to 19.0
mm and the length of the pellets may range from 12.7 to 25.4 mm. Pelleting
technology is commonly used in the U.S. and Europe for producing animal feed.
Cubing equipment is similar to that of pelleting equipment. The cubers usually have
one row of holes in the die ring, and the holes have a square cross section (Sokhansanj
and Wood, 1991). Cube sizes are from 12.7 x 12.7 mm to 38.1 x 38.1 mm in cross
section, and from 25.4 to 101.6 mm in length.
In a roll press briquetter/compactor, material is densified by compression
between two counter-rotating rolls. Initial densification of the material may occur
through compressing the material with a tapered auger in the feed mechanism. The
primary purpose of this initial densification phase is to remove air from low bulk
density material. The final compaction occurs because of high pressures created as the
material flows between the two rollers. The roll surfaces have pockets to form
briquettes of the desired size and shape when the material passes between the rolls.
The briquettes are easily separated and handled after leaving the machine. The
densified products are mostly of pillow-shape with a size of 10 to 40 mm or larger.
The roll press is generally used for densifying coals, minerals, and metals in the U.S.
and Europe (Pietsch, 2002).
Before entering into densification equipment, feed is ground to a desired
particle size, subjected to pretreatments such as mixing with additives or fats, and
conditioned with steam or expander to increase the temperature and/or moisture level.
The temperature of the densified products after leaving the densification equipment is

68
generally higher than the conditioned feed due to frictional heat developed in the
pressing systems, and therefore, densified products are cooled using ambient air before
storing/packaging (Thomas and van der Poel, 1996).
For success of the densification process, the quality of the densified products
must meet the consumer requirements and market standards. One of the important
quality factors is the amount of fines present in the densified products. Therefore, to
avoid fines production, densified products must withstand the rigors of handling and
transportation. Forces that cause pellet (or any densified product) damage (i.e.,
fragmentation and abrasion of pellets) during handling, transportation, and storage
may be divided into three general classes: compression, impact, and shear.
Compression forces result in a crushing action; impact forces result in shattering both
on the surface of the pellet and along any natural cleavage planes of the pellet; and
shearing forces result in abrasion of pellet edges and surfaces (Young et al., 1963).
Therefore, testing the densified products to estimate the amount of damage or
fines that could be observed at the point of utilization in terms of strength and
durability would help optimize the feed material, pre-conditioning processes, and
densification equipment to produce high quality densified products.

3.3 Objectives

The overall goal of this Chapter is to help develop strategies to improve the
strength and durability of the densified products, and to provide conditions to produce
strong and durable densified products from biomass feedstocks or feed formulations
whose densification characteristics are unknown. The specific objective of this
Chapter is to review the literature to study the effects of feedstock, pre-conditioning
processes, and densification equipment variables on the strength and durability of the
densified products such as pellets, briquettes, and cubes.

69
3.4 Binding Mechanisms

The strength and durability of the densified products depend on the physical
forces that bond the particles together. Understanding the particle binding mechanisms
is important in order to determine which test should be used to measure the strength
and durability of the densified products. The binding forces that act between the
individual particles in densified products have been categorized into five major groups
(Rumpf, 1962; Pietsch, 2002). They are briefly discussed below.

3.4-1 Solid Bridges

Due to the application of high pressures and temperatures, solid bridges may
be developed by diffusion of molecules from one particle to another at the points of
contact. Solid bridges may also be formed between particles due to crystallization of
some ingredients, chemical reaction, hardening of binders, and solidification of melted
components. Solid bridges are mainly formed after cooling/drying of densified
products.

3.4-2 Attraction Forces between Solid Particles

Short-range forces such as molecular [valance forces (i.e., free chemical


bonds), hydrogen bridges, and van der Waals' forces], electrostatic, and magnetic
forces can cause solid particles to adhere to each other if the particles are brought
close enough together. Valance forces are effective only if the inter-particle distance is
about 10 A. van der Waals' forces are believed to make the most contribution to all
intermolecular attractive effects and are partly responsible for the adhesion between
particles less than 0.1 urn apart. Electrostatic forces help binding when there is excess
charge or an electrical double layer, which may be created during grinding or by inter-
particle friction. If magnetic forces exist in the powder system, this could contribute to
the particle binding. The effectiveness of short-range forces diminishes dramatically as
the size of the particles or inter-particle distance increases.
70
3.4-3 Mechanical Interlocking Bonds

During the compression process, fibers, flat-shaped particles, and bulky


particles can interlock or fold about each other resulting in interlocking bonds.
Mechanical interlocking bonds can resist the disruptive forces caused by elastic
recovery following compression.

3.4-4 Adhesion and Cohesion Forces

Highly viscous binders (e.g., molasses and tar) adhere to the surfaces of solid
particles to generate strong bonds that are very similar to those of solid bridges. Many
viscous binders harden after cooling and form solid bridges. Thin adsorption layers (at
least 3-nm thick) are immobile and can form strong bonds between adjacent particles
either by smoothing out surface roughness and increasing the inter-particle contact
area or by decreasing the inter-particle distance and allowing the intermolecular
attractive forces to participate in the bonding mechanism.

3.4-5 Interfacial Forces and Capillary Pressure

Presence of liquids such as free moisture between particles, especially in a wet


agglomeration process, causes cohesive forces between particles. Mutual attraction of
particles is brought about by the surface tension of the liquid bridges. Therefore, the
forces that bond the particles are derived from the interfacial tension at the liquid-gas
interface. The capillary state is reached when all the void space within the agglomerate
is completely filled with the liquid. However, the bonds created by interfacial forces
and capillary pressure disappear once the liquid evaporates, and possibly some other
binding mechanisms may take over.
The above five binding mechanisms have been observed and postulated for the
densification of pharmaceutical powders (Ghebre-Sellassie, 1989), animal feeds
(Behnke, 1994; Tabil, 1996), and biomass materials (Mohsenin and Zaske, 1976;
Lindley and Vossoughi, 1989; Tabil and Sokhansaj, 1996).

71
3.5 Measurement of Strength and Durability

The effectiveness of the inter-particle bonds created during the densification


process has been measured in terms of strength and durability. Procedures used for
measuring the compressive resistance, abrasive resistance, impact resistance, and
water resistance of the densified products are discussed below.

3.5-1 Compressive Resistance

Compressive resistance (or crushing resistance or hardness) is the maximum


crushing load a pellet/briquette can withstand before cracking or breaking.
Compressive resistance of the densified products is determined by a diametrical
compression test. A single pellet/briquette is placed between two flat, parallel platens
which have facial areas greater than the projected area of the pellet/briquette. An
increasing load is applied at a constant rate, until the test specimen fails by cracking or
breaking. The load at fracture is read off a recorded stress-strain curve, which is the
compressive strength and reported as force or stress. In the diametrical compression
test, actually the tensile strength of the densified products is determined. Tensile
strength is related to the adhesion forces between particles at all contact points in the
agglomerate (Tabil, 1996; Pietsch, 2002). The diametrical compression test is
predominantly used for testing tablets in the pharmaceutical industry (Tabil, 1996).
The compressive resistance test simulates the compressive stress due to weight
of the top pellets on the lower pellets during storage in bins or silos, crushing of pellets
in a screw conveyor, and crushing (i.e., chewing) of feed pellets between animal teeth.
The Kahl tester, Stokes tester, Schleuniger tester, tablet hardness testers, universal
testing machine (Instron), and Kramer shear strength tester have been used to measure
the compressive strength of the densified products (Behnke, 1994; Tabil, 1996;
Thomas and van der Poel, 1996; Franke and Rey, 2006). Li and Liu (2000) used
ASTM method C39-96, which was developed for concrete briquettes (ASTM, 1998a)
to measure the compressive resistance of biomass logs (48.5-mm diameter x 50-mm
length).
72
The compressive resistance test provides a quick measure of the quality of
pellets as soon as pellets are produced from the pellet mill and aids in adjusting the
pelleting process to improve pellet quality. However, this test will not indicate the dust
or fines production potential of the pellets during handling, transportation, and storage.
Fines content in feed pellets affects animal performance (Stark, 1994). Usually, the
abrasive resistance (i.e., durability) test is used to adjust the pelleting conditions and
process to improve pellet quality during production of pellets in feed mills
(Winowiski, 1998).
Several researchers reported that it was difficult to obtain repeatability of the
results from the compressive resistance test for the same quality of pellets (Tabil,
1996; Thomas and van der Poel, 1996; Li and Liu, 2000; Franke and Rey, 2006).
During the compression resistance test, Li and Liu (2000) observed that sawdust logs
were squeezed (shortened) by approximately 1/3 of the size before the final failure
occurred. Therefore, the compression resistance may not be a reliable measure of
quality of the densified biomass products.

3.5-2 Durability

The durability or abrasive resistance test simulates either mechanical or


pneumatic handling. These tests can help control the densification process, and thus,
pellet quality in the feed manufacturing industry. In the feed industry, high durability
means high quality pellets. Durability is the prevalent form of measurement and
expression of pellet quality in the feed pelleting industry. The tumbling can, Holmen
tester, and Ligno tester, which are described in the following paragraphs, have been
used for measuring durability of the densified products.
The tumbling can method (ASABE Standards, 2003) is used to estimate pellet
quality in terms of pellet durability index (PDI), or, simply percent durability. This test
simulates the mechanical handling of pellets and predicts the possible fines produced
due to mechanical handling. During tumbling, pellets abrade and produce fines due to
impact and shearing of pellets over each other and over the wall of the tumbling can.

73
After tumbling 500 g of pellets for 10 min at 50 rpm, the pellets are sieved using a
sieve size of about 0.8 times the pellet diameter. The PDI or durability is calculated as
the ratio of weight after tumbling over the weight before tumbling, multiplied by 100.
A detailed procedure can be found at ASABE Standards (2003). According to
Winowiski (1998), the tumbling can method is the most often used method in the feed
manufacturing industry in the U.S.
Some feed mill operators have modified the tumbling can technique by adding
steel nuts, bolts, or ball bearings to the tumbling can along with the pellets before
tumbling. The modified method has been used to test very hard pellets such as dairy
feed pellets (Winowiski, 1998). Stevens (1987) and Cavalcanti (2004) used five to six
12.7 mm (14 in.) hexagonal nuts to add more severity to the durability test in order to
more accurately determine the treatment effects during pelleting of corn- and wheat-
based swine diets. Several authors have used the tumbling can method with some
variations in the speed of tumbling, length of time of tumbling, sieve size, or the
amount of samples tumbled (Richards, 1990; Raghavan and Conkle, 1991).
The Holmen durability tester simulates pneumatic handling of pellets. The
Holmen pellet durability tester pneumatically circulates a sample of pellets through a
square conduit of pipe or tubing with right-angled bends, and the pellets are impacted
repeatedly on hard surfaces. When pellets strike the right angle corners of the tester,
they fracture. A sample is circulated in the Holmen tester for 30 to 120 s. The
remaining pellets are collected, sieved with sieve size about 80% of the pellet
diameter, weighed, and the PDI is calculated. The Holmen tester is more widely used
in Europe than in North America because it simulates the pneumatic conveyors
common in European feed mills. The Holmen tester requires a smaller sample (i.e.,
100 g) than the tumbling can method (i.e., 500 g). The Holmen tester is harsher than
the tumbling can method and, therefore, yields lower PDI (Thomas and van der Poel,
1996; Franke and Rey, 2006).
The Ligno tester was recently introduced in the feed industry (Winowiski,
1998). This device uses air to rapidly circulate 100 g of pellets around a perforated
chamber for 30 s. The chamber is an inverted square pyramid with perforated sides.

74
Forced air is the destructive force. The chamber is set at 60 millibar pressure. Higher
pressures can be set when large diameter pellets [greater than 4.8 mm (3/16-in.)] are
tested. Fines are removed continuously during the test and there is no need to screen
the pellet. This method is faster than either the tumbling can or the Holmen tester. The
Ligno tester is adapted in an Austrian standard for measuring durability of wood
pellets or briquettes (ONORM M 7135, 2000; cited by Obernberger and Thek, 2004).
Based on this standard, 100-g of pellets are exposed to an air stream of 70 millibar for
60 s, and the percentage of fines that pass through a 3.15-mm sieve is taken as the
measure of durability.
Based on the results of some commercial feed pellets, the following
conclusions were drawn (table 3.1): the Ligno tester is more destructive than the
tumbling can tester, and the Ligno tester nearly matches the destructive force of the
Holmen tester. Winowiski (1998) concluded that the Ligno tester showed more
sensitivity to factors (e.g., added binder or fat) that influence pellet durability than
other durability testing methods. McEllhiney (1988) reported that durability values
obtained from the Holmen tester for a cycle time of 30 s were approximately equal to
the durability values obtained from the tumbling can method with a cycle time of 10
min. Temmerman et al. (2006) observed no clear correlation between the durability
values obtained by the tumbling can method and the Ligno tester. They advised not to
compare the durability values from these two methods. In addition, they found that the
tumbling can method produced more repeatable and reproducible durability results
than the Ligno tester.
AS ABE Standards (2003) also have procedures for measuring durability of
briquettes, cubes and crumbles. To determine the abrasive resistance of biomass logs
(48.5-mm diameter x 50-mm length), Li and Liu (2000) used the ASTM method
D441-86 developed for coal (ASTM, 1998b), where the samples were tumbled at 60
rpm for 40 min. In the literature, several other instruments have been tried to measure
the abrasive resistance of the densified products: Stein breakage tester (Sokhansanj et
al., 1991), Ro-Tap sieve shaker (Raghavan and Conkle, 1991), and Dural tester

75
(Adapa et al., 2003). The Ro-Tap sieve shaker-based durability test may simulate mild
handling and vibration of pellets during transport (Raghavan and Conkle, 1991).

3.5-3 Impact Resistance

Impact (or drop or shattering) resistance test may simulate the forces
encountered during emptying of densified products from trucks onto ground, or from
chutes into bins. Pietsch (2002) suggested that drop tests can be used to determine safe
heights for pellet handling.
Sah et al. (1980), Khankari et al. (1989), Shrivastava et al. (1989), and Al-
Widyan and Al-Jalil (2001) used a drop resistance test to determine the durability of
biomass pellets and briquettes. Pellets were dropped from a 1.85-m height onto a
metal plate four times. The weight retained as the percentage of the initial weight was
taken as the pellet/briquette durability.
Lindley and Vossoughi (1989) measured the shatter resistance as the
percentage loss of weight from shattering. Each briquette (50-mm diameter x 18-mm
thick) was subjected to 10 repeated drops from one meter height onto a concrete
surface. The percent loss was then calculated. Raghavan and Conkle (1991) developed
a similar test for coal pellets.
Li and Liu (2000) used ASTM method D440-86, the drop shatter test
developed for coal (ASTM, 1998c), to estimate the impact resistance of biomass logs
(48.5-mm diameter x 50-mm length). The biomass logs were dropped twice from 1.83
m onto a concrete floor. Then, an impact resistance index (IRI) was calculated as
described by Richards (1990): IRI = (100 x N)/n, where N is the number of drops, and
n is the total number of pieces after N drops. The maximum value of IRI is 200. Small
pieces weighing less than 5% of the original weight of the logs were not included in
the IRI calculation.
We noted that several authors have used the term "durability" to report the
impact resistance. In this Chapter, the term "durability" is used to present the
durability values obtained from the tumbling can method, unless indicated otherwise.

76
3.5-4 Water Resistance

Short-term exposure to rain or high humidity conditions during transportation


and storage could adversely affect the quality of the densified products. Lindley and
Vossoughi (1989) measured the water resistance as the percentage of water absorbed
by a briquette (50-mm diameter x 18-mm thick) when immersed in water. Each
briquette was immersed in water at 27C for 30 s.
Fasina and Sokhansani (1996), Tabil (1996), and several other researchers
determined the effect of high relative humidity environment on the strength and
durability of the pellets. They stored the pellets in an environmental chamber for
several combinations of relative humidity (70 to 90%) and temperature (10 to 40C).
Compressive resistance and durability (mostly tumbling can method) of the pellets
were measured with time as the moisture content of the pellets increased to different
values.
The selection of the type of a physical quality test (i.e., compressive resistance,
abrasive resistance, or impact resistance) used to test a particular densified product
depends on the type of forces encountered between manufacturing and utilization. The
right test is the one that simulates the way the densified products are handled,
conveyed, and stored, and results in approximately the amount of fragmentation and
fines that are produced. For feed pellets, hardness and durability are usually
determined. Additional pellet quality properties are required depending on animal
species. For example, fish-feeds require sinking velocity, water absorption index, and
water solubility index; however, they are not discussed in this Chapter.

Richards (1990) observed direct relationships between the compressive


resistance, abrasive resistance, and impact resistance of coal pellets. Wood (1987) and
Tabil (1996) found that the logarithm of hardness and durability of feed pellets had a
linear relationship. However, Tabil (1996) observed that some of the alfalfa pellets
with high durability had low hardness values. This is because hardness and durability
are two distinct pellet quality measures. Also, compressive resistance, abrasive

77
resistance, and impact resistance test results cannot be compared to each other since it
is not known which stress component causes the densified products to fail (Pietsch,
2002).

3.6 Factors Affecting Strength and Durability

3.6-1 Effect of Feed Constituents

Reece (1966) investigated the effect of quality of alfalfa hay on wafer


durability. He attributed hay quality to differences in the amount of stems. Lower
wafer durability (49.5%) was obtained for hay containing 58% stems, 19% leaves, and
23% fines that passed through 2.4 mm screen (mostly broken leaves). Higher wafer
durability (above 90%) was obtained for hay containing 44% stems, 41% leaves, and
15% fines that passed through 2.4 mm screen (mostly broken leaves). Similarly,
Adapa et al. (2003) fractionated alfalfa hay into leaves and stems, and mixed different
percentages of leaves with stems and studied the effect on pellet durability. Leaf
contents in the range of 25 to 100% produced higher pellet durability (75 to 85%) than
that for 0% leaves (i.e., 100% stems), where the pellet durability was 43 to 73%.
Israelsen et al. (1981) studied the effect of replacing feed ingredients from a
base formulation on the pellet durability. They replaced either barley or cotton seed
meal with a number of other feed ingredients from the base formulation of a dairy diet
(28.7% alkali-treated straw pellets, 28.7% barley, 28.7% cotton seed meal, 4.0% fat,
and 10.0%) molasses). To replace barley, they used beet pulp, barley malt culms, citrus
pulp, coconut meal, alfalfa meal, grass, wheat bran, or palm kernel cake. To replace
cotton seed meal, sunflower seed meal, soybean meal, or rapeseed meal were used.
The base formulation had pellet durability of 97.2%. Significantly (P < 0.05) higher
durability was obtained for beet pulp (98.9% pellet durability) or barley malt culms
(98.3% pellet durability) than for the base formulation, and the rest of the barley
substitutes resulted in about 96.8 to 97.6% pellet durability. The cotton seed meal
substitutes resulted in lower durability values than that of the base formulation. The

78
pellet durability values were 4.9, 94.5, or 91.2% for formulation with sunflower seed
meal, soybean meal, or rapeseed meal, respectively.
Because of the inherent variability in the physico-chemical properties of the
raw materials, the effect of different feed ingredients on the strength and durability of
the densified products may be studied in terms of the effect of feed constituents such
as starch, protein, fiber, and fat. Studying the effect of feed ingredients in terms of
feed constituents would be useful because the results from one study can be easily
extended to another study where a new feed formulation or biomass material is to be
densified. Also, the binding behaviors of the feed constituents would be similar
regardless of the type of raw material. The commonly observed binding characteristics
of feed constituents are gelatinization of starch, denaturation of protein, and
solubilization and subsequent recrystallization of sugars and salts (Thomas et al.,
1998).

79
3.6.1-1 Starch

Starch acts as a binder (Wood, 1987; Thomas et al., 1998). Native starch has
less binding ability than gelatinized starch. In the presence of heat and moisture,
gelatinization of starch occurs. Mechanical shearing during the densification process
also improves starch gelatinization. Stevens (1987) and Cavalcanti (2004) used a
differential scanning calorimeter (DSC) apparatus to study corn starch gelatinization.
They noted that starch gelatinization occurred by two mechanisms: (i) hydration and
swelling of starch granules and ultimately, disruption of the crystalline structure due to
the combined effect of temperature and moisture; and (ii) disruption of starch granules
by shear friction as the feed mash expelled through the pellet die. The greater the
percentage of starch gelatinization, the higher the pellet durability (Heffner and Pfost,
1973). Wood (1987) studied the effects of adding raw starch or pre-gelatinized starch
on pellet hardness and durability. He found that pre-gelatinized starch resulted in
higher pellet hardness and durability values than those for raw starch.

3.6.1-2 Protein

Protein will plasticize under heat and act as a binder, which helps increase the
strength of the densified products (Winowiski, 1988; Briggs et al., 1999). During feed
pelleting, the combined effects of heat, moisture, and shear result in protein
denaturation, which induces the binding functionality of protein (Wood, 1987;
Thomas et al., 1998). Proteins derived from cereal grains with dough-forming
capabilities such as wheat (Stevens, 1987; Winowiski, 1988), rye, and barley (Moran,
1989), and soybean meal (Cavalcanti, 2004) would help improve the pellet durability.
Cavalcanti (2004) reported that protein derived from corn (i.e., corn gluten meal) had
a negative effect on pellet durability.
Wood (1987) observed higher pellet hardness and durability when adding raw
protein rather than denatured protein. Starch (raw or pregelatinized) had a minimal
effect on pellet durability when material was pelleted with raw protein. The average

80
Holmen pellet durability for rations containing 40% raw starch:60% raw protein was
85%, and, for rations containing 40% pregelatinized starch:60% raw protein, the
average Holmen pellet durability was 96%. However, Wood (1987) also observed that
the effect of pregelatinized starch was more prominent when material was pelleted
with denatured protein. A feed mixture containing native starch and denatured protein,
the worst case scenario, had little or no binding capacity and gave pellets which were
very weak (Holmen pellet durability of 20%>, and pellet hardness of 19.6 N.)
Hill and Pulkinen (1988) reported that protein content of alfalfa in the range of
14.5 to 20% did not influence the pellet durability. They found that the average pellet
durability for this protein content range was about 80%. In the absence of binders (i.e.,
additives), Tabil et al. (1997) noted that alfalfa chops containing 22% crude protein
produced pellets with durability of 65% and hardness of 507 N. Whereas, alfalfa chops
containing 17.6% crude protein produced pellets with durability of 54.6% and
hardness of 425 N. Briggs et al. (1999) found that increasing the protein content in a
poultry diet from 16.3 to 21% increased the pellet durability from 76 to 89%.

3.6.1-3 Fiber

Fiber can be classified as water-soluble and water-insoluble. Water-soluble


fibers increase the viscosity of feed and positively affect the structural integrity of
pellets. Water-insoluble fibers may entangle and fold between particles (Rumpf,
1962). Due to their resilience characteristics (e.g., stiffness and elasticity), fibers may
not form good bonds between particles or fibers (Thomas et al., 1998). Also, presence
of large fibers in the pellet matrix may result in weak spots for fragmentation. To
decrease the resilience characteristics of fibers, chemical agents such as NaOH, CaO,
or urea have been added to high fiber feeds (e.g., wheat straw) (Thomas et al., 1998).
Addition of these chemicals might degrade the cell wall structure and separate lignin
from cellulose, thus, reducing the resilience characteristics of the fibers, and ultimately
this would help increase pellet durability. High water content (19 to 23% w.b.) has
been found to reduce the resilience characteristics of fibrous biomass materials such as

81
hay, bark, and sawdust (Mohsenin and Zaske, 1976). Hill and Pulkinen (1988) found
that increasing crude fiber content from 18.5 to 26.5% increased the durability of
alfalfa pellets by about 5%. Angulo et al. (1995) reported that animal feeds containing
4.4, 7.0, and 14.0% of fiber content resulted in pellet durability of 93.8, 96.1, and
97.4%, respectively.

3.6.1-4 Fat / oil

Inclusion of fat/oil (animal or vegetable based) in feed results in lower pellet


durability (Richardson and Day, 1976; Stark, 1994; Angulo et al., 1996; Briggs et al.,
1999; Cavalcanti, 2004). This is because fat acts as a lubricant between the feed
particles, and between the feed and the pellet mill die-wall. Due to low friction in the
die, pressure in the die is decreased which would result in pellets with lower
durability. Due to the hydrophobic nature of the fat, fat added to the feed inhibits the
binding properties of the water-soluble components in the feed such as starch, protein,
and fiber (Thomas et al., 1998). In the absence of added fat, sometimes the (natural)
fat in the cell-wall may come out of the cell and act as binding component between
particles and make solid bridges, which may positively influence pellet durability
(Thomas et al., 1998). To obtain maximum pellet quality, fat addition before pelleting
should not exceed 1.5% (Vest, 1993). According to Briggs et al. (1999), the oil content
should not exceed 5.6% in rations formulated using high-oil corn and soybean meal.
Cavalcanti (2004) studied the binding functionality of starch, protein, and fat in 13
different model feeds derived from mixing different percentages of corn, soybean
meal, and soybean oil. He found that fat inclusion levels beyond 6.5% are deleterious
to pellet durability. Moreover, high levels of fat (> 6.5%) affected the binding
functionality of starch and protein, and the fat content was the dominant factor
determining the pellet durability. Table 3.2 gives several examples where fat had
negative impact on pellet quality. Generally, fat addition slightly increases pellet
production rate and reduces the specific energy requirement (Richardson and Day,
1976).

82
3.6.1-5 Lignin and extractives

Lignin acts as a binder in situ in the feed material. At elevated temperatures,


lignin softens and helps the binding process. Bradfield and Levi (1984) reported that
as lignin plus extractives content increased above a threshold level of 34% in wood
samples, pellet durability decreased (table 3.3). Bradfield and Levi (1984) postulated
that the auto-adhesive action of thermally softened, non-crystalline wood polymers
was like that of a mastic with little internal strength of its own. Initially, pellet
durability increased as the non-crystalline wood polymers acted as an adhesive
between crystalline zones. However, above a threshold "excessive mastic" between
crystallites reduced the strength and durability of the pellets.

3.6.1-6 Other feed constituents

Physical and chemical properties of calcium and phosphorus sources (e.g.,


dicalcium phosphate and deflourinated phosphate) and other minerals added in animal
feeds have been found to affect pellet durability (Turner, 1995; Behnke, 2006). The
degree of abrasiveness of these minerals could have affected the pelleting performance
of the feed.

3.6-2 Effect of Feed Moisture Content

Water acts as both a binding agent and a lubricant. Water helps develop van
der Waals' forces by increasing the area of contact between particles (Grover and
Mishra, 1996). Water aids briquetting when water-soluble compounds are present in
the feed such as starch, sugar, soda ash, sodium phosphate, potassium salt, and
calcium chloride (Lindley and Vossoughi, 1989). Several studies showed that strength
and durability of the densified products increased with increasing moisture content
until an optimum is reached.

83
Reece (1966) found that alfalfa hay with moisture contents of 10 to 23% (w.b.)
produced wafers with durability of 80 to 90%. He observed that it was difficult to
make wafers when the moisture content was 25% and above. At high moistures (>
25% w.b.), probably due to the incompressibility of water, moisture trapped within the
particles may prevent complete flattening and the release of natural binders from the
particles (Pickard et al., 1961). Smith et al. (1977) showed that increasing moisture
content from 10 to 15% (w.b.) increased the durability of wheat straw briquettes from
73 to 81%. Holley (1983) stated that biomass materials with 10 to 20% (w.b.) moisture
contents can be briquetted in roll presses. Hill and Pulkinen (1988) recommended an
optimum moisture content of 8 to 9% (w.b.) for producing alfalfa pellets in pellet
mills. Generally, moisture contents of 11.0 to 12.0% (w.b.) are used for wheat- and
corn-based feed pelleting (Stevens, 1987). Turner (1995) cautioned that pellet-mill
dies tend to choke when the moisture content of the conditioned feed mash is 16 to
18% (w.b.). During pelleting, high fiber feeds are not able to absorb moisture in the
conditioning chamber so water remains on the surface of the particles, causing excess
particle-to-particle lubrication. This condition causes the center of the pellet to extrude
faster than the exterior, thus forming "Christmas tree"-shaped pellets, which reduces
pellet durability (Winowiski, 2006). Li and Liu (2000) concluded that good quality
wood logs can be produced with initial moisture contents of 6 to 12% (w.b.); however,
the optimum moisture content is around 8% (w.b.). According to Obernberger and
Thek (2004), production of high quality pellets is possible only if the moisture content
of the feed is between 8.0 and 12.0% (w.b.), and water contents above or below this
range would lead to lower quality pellets.

Usually, steam conditioning adds about 1.0 to 6.0% of moisture content before
pelleting (Young et al., 1963; Pfost, 1964; Maier and Gardecki, 1992). Water (1.5 to
3.5%) added through steam conditioning produced 1 to 2% higher pellet durability
than for direct addition of tap water (Thomas et al., 1997). Lund (1984) postulated that
a minimum water to starch ratio of 0.3:1 is required for gelatinization, and a water to
starch ratio of 1.5:1 would be required for complete starch gelatinization. Fairchild
and Greer (1999) and Moritz et al. (2002) concluded that increasing water to starch

84
ratio prior to pelleting significantly improved starch gelatinization, and thus, pellet
durability.
Mohsenin and Zaske (1976) found that alfalfa wafers made at 19% (w.b.)
moisture content had durability of more than 90%. They also found that wood-based
wafers produced at a moisture content of 25% (w.b.) had zero percent durability. They
indicated that protoplasm liberated during compression contributed to the binding of
particles. Srivastava et al. (1981) reported that increasing moisture content of grass
hay (mixed with 20% alfalfa) from 3 to 11% (w.b.), the wafer (150-mm diameter)
durability rating (based on an impact resistance test) increased from 27 to 96%.
Further increase in moisture content from 11 to 25% (w.b.) decreased the wafer
durability rating from 96 to 20%. They concluded that the best binding effect could be
obtained at around the moisture content of 11% (w.b.) for this feedstock.
Koser et al. (1982) obtained maximum briquette quality when the moisture
content of water hyacinth was 8 to 12% (w.b.). Too dry (< 8% w.b. moisture content)
and too wet (>20% w.b. moisture content) water hyacinth could not be briquetted
because of very high re-expansion of briquettes. O'Dogherty and Wheeler (1984)
found that durability of wheat-straw wafers (50-mm diameter) was about 97% for
initial moisture contents of 10 to 20% (w.b.). Increasing the initial moisture content
from 20 to 35% (w.b.) decreased wafer durability from about 97 to 85%. Conversely,
Al-Widyan and Al-Jalil (2001) showed that durability (based on an impact resistance
test) of olive cake briquettes increased from about 10 to 100% with increasing initial
moisture contents from 20 to 35% (w.b.). For corn stover, Kaliyan and Morey (2005)
found that increasing moisture content from 10 to 15% (w.b.) increased briquette
durability from 62 to 84%.
Komarek (1991) reported that crushing strength of briquettes made from sub-
bituminous coal, lignite, or peat was reduced from about 1127.8 to 245.2 N when the
moisture content was increased from 5 to 30% (w.b.).

85
3.6-3 Effect of Feed Particle Size

Particle size is an important influencer of pellet durability. Generally, the finer


the grind, the higher the durability. Fine particles usually accept more moisture than
large particles, and therefore, undergo a higher degree of conditioning. Also, large
particles are fissure points that cause cracks and fractures in pellets (MacBain, 1966).
The recommended particle size for good pellet quality is 0.6 to 0.8 mm (Turner,
1995). Franke and Rey (2006) recommended a particle size of 0.5 to 0.7 mm to
produce durable pellets. They also mentioned that particle sizes of greater than 1.0-
mm will act as predetermined breaking points in the pellet. Although, fine particles
produce more durable pellets, fine grinding is undesirable because of increased cost of
production. A mixture of different particle sizes would give optimum pellet quality
because the mixture of particles will make inter-particle bonding with nearly no inter-
particle spaces (MacBain, 1966; Payne, 1978; Grover and Mishra, 1996). Table 3.4
presents the suggested feed particle size distribution to produce good quality pellets
(Payne, 2006).
From the survey of several U.S. feed mills, Vest (1993) concluded that
hammer mill screen sizes ranging from 2.8 to 4.8 mm were used in the U.S. feed mills.
Feed mills that used a hammer mill screen size of 3.2-mm or 4.0 mm-3.2 mm split
screen (two different screen sizes on the same hammer mill, 4.0 mm screen on the
upside and 3.2 mm screen on the downside) produced high quality pellets. Feed mills
that used a hammer mill screen size of 4.0 mm had lowest pellet quality.
Reece et al. (1985) reported that corn ground using 3.18 mm (geometric mean
diameter of particles = 679 um), 6.35 mm (geometric mean diameter of particles =
987 um), and 9.53 mm (geometric mean diameter of particles = 1289 um) hammer
mill screen sizes produced pellet durability of 91.0, 91.3, and 92.5%, respectively, for
a broiler diet. Singh and Kashyap (1985) observed that decreasing average particle
size of rice husk from 5.14 to 4.05 mm increased the briquette durability from 84.1 to
95% at a pressure of 31.2 MPa and 25% molasses as binder. Stevens (1987) observed
no differences in the pellet durability when pelleting a corn-based swine diet ground

86
using a hammer mill and a roller mill. Also, particle size did not have much impact on
the pellet durability of corn-based swine diet whereas particle size had significant
effect on wheat-based swine diet. Table 3.5 shows that smaller particle size resulted in
higher pellet durability for several feedstocks.

3.6-4 Effect of Steam Conditioning / Preheating

The temperature of the feed is usually increased to activate the inherent binders
or added binders (i.e., additives). Elevated temperatures also promote plastic
deformation of thermo-plastic particles. Plastic deformation of particles is important
for making permanent bonding. Feed temperature can be increased by either direct or
indirect heating. Direct heating includes friction, fluidized bed heating, and steam
conditioning. Indirect heating includes conduction-based heating systems such as
band-electrical resistance heaters, and hot-oil circulation heat exchangers. Preheating
temperature is limited to 300C to avoid decomposition of biomass materials (Grover
andMishra, 1996).
Steam conditioning helps to release and activate natural binders and natural
lubricants in the feed, and activates artificial binders which are mixed with the feed,
promotes starch gelatinization and protein denaturation. Thus, steam conditioning
helps to produce durable and hard pellets. Also, steam conditioning helps reduce germ
and bacterial counts (Franke and Rey, 2006). Usually, the steam conditioner has
paddles to mix the steam and added liquids with the feed. Steam conditioning takes
place around 20 to 255 s in the conditioner depending on throughput, rpm of the
paddle bar, and the degree of fill. Due to the condensation of the steam, a thick film of
water is created around the particles, which, together with the temperature increase,
facilitates binding between particles through capillary sorption between particles and
physico-chemical changes of the feed such as thermal softening (i.e., glass transition),
starch gelatinization, and protein denaturation (Thomas et al., 1997).
Steam quality (i.e., the percentage of steam in the vapor phase) is an important
steam property that can affect pellet durability. High quality steam has more energy to

87
raise the feed mash temperature than low quality steam. Steam quality of 70 to 80% is
important to produce high quality pellets (Gilpin et al., 2002).
Steam pressures of 6.9 to 689.4 kPag (10 to 100 psig) have been used in the
feed industry. Steam pressures of 103 kPag (15 psig) or less are considered to be low
pressures (Stevens, 1987). MacBain (1966) indicated that low pressure steam gave the
best quality of pellets for high starch formulations, and high pressure steam gave the
best quality of pellets for high protein formulations. Payne (1978) made
recommendations on the steam conditioning pressure and feed temperature for
different raw materials to produce good quality pellets: high starch and low fiber
rations [103.4 kPag (15 psig) and 80 to 85C], dairy rations [448.2 KPag (65 psig) and
60 to 65C], high protein rations [448.2 KPag (65 psig) and 80C], and heat sensitive
rations such as milk powder and sugar [103.4 kPag (15 psig) and < 55C]. Also,
Winowiski (1985) studied the effect of steam conditioning on the pellet durability of
high starch, high protein, and high fiber content feeds. Increase in temperature in the
range of 54 to 88C increased the durability of feeds containing high starch and high
protein contents, but decreased the pellet durability of high fiber feed. To improve the
binding, he recommended addition of low pressure steam (i.e., high heat and high
moisture) to high starch feed, and high pressure steam (i.e., high heat and limited
moisture) to high protein feed. High fiber feed did not accept moisture and thus, it
required limited steam addition to improve pellet durability. Thomas et al. (1997)
generalized that low steam pressure, relative to heat, is used where more water is
needed for conditioning (e.g., starch gelatinization). High steam pressure is used in
cases where relatively low amounts of water and higher temperatures are needed for
conditioning the feed (e.g., protein denaturation).

Stevens (1987) concluded that there was no effect (P < 0.05) of the steam
pressures, 138 and 552 kPag (20 and 80 psig), at a conditioning temperature of 65C
on the pellet mill production rate, efficiency, and pellet durability for corn- and wheat-
based swine diets. However, a corn-based diet resulted in lower pellet durability
(58%) than for the wheat-based diet (91%). Recently, Briggs et al. (1999) also found
no significant effect of steam pressure in the range of 138 to 552 kPag (20 to 80 psig)

88
on pellet durability for diets containing corn and soybean meal. Vest (1993) reported
that feed mills that used steam pressures of 0 to 62 kPag (0 to 9 psig) produced pellets
with lower durability than feed mills that used steam pressures of 138 to 269 kPag (20
to 39 psig).
Stevens (1987) determined the degree of starch gelatinization in corn pellets
using a differential scanning calorimeter (DSC). At steam conditioning temperatures
of 23C and 80C, the percentage of starch gelatinization both on the outer surface of
the pellets (about 2 mm outer layers) and whole corn pellets was 42 to 58%, and 21 to
28%, respectively. The possible reasons reported were that at the low temperature (i.e.,
23C), friction and mechanical shear damage of starch granules in the pellet-mill die
could have been higher than that at the high temperature (i.e., 80C), where the feed
mash might have been conditioned such that the feed mash slips through the die-hole
very easily and faster. The measured temperature of pellets was 69 and 84C when the
feed temperature was 23 and 80C, respectively. This suggests that the higher the
temperature difference between the feed and pellet, the higher the degree of starch
gelatinization.
Skoch et al. (1981) found that steam conditioning improved the pellet
durability of poultry layer diet (table 3.6). Their results were similar to those of
Stevens (1987). Higher starch gelatinization occurred for pelleting without steam
conditioning (i.e., 27C) than with steam conditioning (i.e., 65 or 80C). They
attributed this effect to the fact that pressing the feed mash through the die causes
more starch damage during dry pelleting than during steam conditioning. Steam
appeared to act as lubricant and decrease mechanical friction.
Reece (1966) used infra-red lamps to heat alfalfa hay to produce wafers. He
found that increasing temperature from 27 to 60C increased the wafer durability from
10 to 72%o. Additional heating above 60C did not appear to improve durability when
the moisture content of the hay was 13 to 16% (w.b.). Hill and Pulkinen (1988)
reported that increasing temperature of alfalfa from 60 to 93C by steam conditioning
increased the pellet durability by about 30 to 35 percentage points (i.e., the durability
values increased from about 40 to 75%). However, they observed slightly dark color

89
pellets at conditioning temperatures higher than 82C, and suggested avoiding
temperatures greater than 82C if green color pellets are desired. Tabil and Sokhansanj
(1996) found that steam conditioning to less than 90C (but >65C) resulted in alfalfa
pellet durability of 66.5 to 69.9% whereas steam conditioning to more than 90C (but
<100C) produced pellet durability of 82.6 to 84.8%.
Khankari et al. (1989) reported that to pellet rice husk with an initial moisture
content of 9.2% to get a target pellet durability (based on an impact resistance test) of
95%) required a preheating temperature of 175C at higher pressure (i.e., 186 MPa) or
a preheating temperature of 225C at lower pressure (i.e., 118 MPa). For cubing
biomass materials, temperatures less than 38C (100F) should be used (Sokhansanj
and Wood, 1991).
Iyengar (1959) noted an increase in the crushing strength of coal briquettes
from 1.7 to 24.0 MPa when increasing preheating temperature from 30 to 250C.
Komarek (1991) found that increasing the preheating temperature from 140 to 240C
increased the crushing strength of bituminous coal briquettes made using a roll press
from 44.1 to 892.4 N.
Increasing retention time of feed mash inside the steam conditioner from 5 to
15 s increased the pellet durability by five percentage points. By adjusting mixer
paddle configuration, slowing the rotational speed of the mixing paddle shaft, or using
longer conditioners, the residence time of feed mash inside the steam conditioner can
be increased (Briggs et al., 1999).

90
3.6-5 Effect of Binders /Additives

A binder (or additive) can be a liquid or solid that forms a bridge, film, matrix,
or causes a chemical reaction to make strong inter-particle bonding. Steam
conditioning or preheating is essential to provide heat and moisture to activate the
inherent or added binders. Selection of binders mainly depends on cost and
environmental friendliness of the binders (Tabil, 1996; Marrero, 1999). When strength
and durability values of pellets do not match with the quality standards or marketing
requirements, additives in the range of 0.5 to 5% (by weight) are added to the feed to
increase the pellet quality or to minimize the pellet quality variations (Tabil, 1996).
More than 50 organic and inorganic binders have been employed for densification
(Pietsch, 2002). Commonly used binders in the animal feed industry are lignosulfonate
(a byproduct from pulp and paper industries), bentonite (a clay mineral), modified
cellulose binders (e.g., sodium carboxymethylcellulose), molasses, starches, and
proteins (Payne, 1978; Tabil, 1996; Thomas et al., 1997). In some European countries,
addition of binders is prohibited (Obernberger and Thek, 2004). In Austria, biological
additives rich in starch content (e.g., maize and rye flour) of only 2% (by weight) are
allowed for wood pellet production (Obernberger and Thek, 2004).

Pfost (1964) studied the effect of adding 1 to 2% calcium lignosulfonate with


four types of turkey and poultry rations. Adding 1 to 2% calcium lignosulfonate
resulted in pellet durability of 90 to 97%. He concluded that adding binder was
essential to significantly improve the pellet durability of high fat rations. Pfost and
Young (1973) found that addition of bentonite (2.4% by weight) improved the
durability of feed pellets by about six percentage points. Salmon (1985) reported that
addition of sodium bentonite (2.5%) did not improve the pellet durability when the
added fat content was 0 or 9%, but pellet durability increased by about two to three
percentage points when the added fat content was 3 or 6% in a turkey ration. Payne
(1978) reported that hardness of dairy feed pellets (measured after 24 h) was 166 N
with lignosulfonate (2.5%), and 131.4 N without lignosulfonate (i.e., 27% increase in
strength due to addition of 2.5% lignosulfonate).

91
Angulo et al. (1995) used sepiolite, hydrated magnesium silicate clay, as
binder at a rate of 20 g/kg of feed to increase the durability of feed pellets (made from
the diet of pig grower, sow, or rabbit) with high levels of fat content. They found that
the average durability of pellets at an added fat level of 5 g/kg of feed was 95.8%
without sepiolite and 96.8% with sepiolite. At an added fat level of 40 g/kg of feed,
the average pellet durability was 86.5% without sepiolite and 93.6% with sepiolite.
The study by Dobie (1975) showed that grass hay could not be formed into
good quality cubes without the application of a good binding agent. Cubes produced
from blue panic grass, Timothy grass, range hay, Pangola grass, and Congo grass
without a binder had durability of 15 to 44%. By adding 5.0 to 10% (by weight) of a
liquid binder (containing ammonium lignin sulfonate, wood sugars, 2.4% nitrogen in
the form of ammonia, and 50% water), durability of cubes increased to 93 to 97%.
Addition of about 5% (by weight) of the same liquid binder to bagasse pith increased
the durability of the cubes from 88.8 to 99.3%.
Singh and Singh (1982) reported that adding 10 to 25% (by weight) of
molasses, sodium silicate, or a mixture of 50% molasses and 50% sodium silicate with
rice straw produced briquettes with 40 to 80%o durability at a particle size 0.15 mm
and pressure of 29.4 MPa. The higher the amount of binders added, the higher the
briquette durability. Binding of molasses might occur due to solid bridges created by
recrystallization of sugars or glass transition after cooling/drying of pellets (Thomas et
al., 1998).
Hill and Pulkinen (1988) concluded that addition (by weight) of any of the
following six binders did not improve alfalfa pellet durability over the control: 4%
bentonite, 1.5%) Perma-Pel (lignosulfonate), 1.5%) lignosite 458, 4% of neutralized
liquid lignosite, 4% of liquid molasses, and 40% of ground barley grain. Tabil et al.
(1997) classified the quality of alfalfa chops based on the durability of pellets obtained
from them without an added binder: low quality when durability was less than 70%;
medium quality when durability was between 70 and 80%; and high quality when
durability was more than 80%. Addition (by weight) of 0.2% of collagen protein,
1.25% of lignosulfonate, 5.0% of bentonite, 1.9% of hydrated lime, or 0.74% of pea

92
starch as a binder considerably improved the durability and hardness of low quality
alfalfa chops (table 3.7), and the durability and hardness values were similar to those
obtained for high quality alfalfa chops with the addition of similar levels of binders
(Tabil et al., 1997). Pellets from medium and high quality chops did not respond
favorably to the addition of binders. They observed that the durability and hardness of
pellets with hydrated lime was highest for all three chop qualities. This may probably
be due to the formation and subsequent hardening of calcium carbonate. Furthermore,
pea starch increased pellet durability without necessarily increasing hardness (table
3.7). They suggested that adding 0.5% of either hydrated lime or pea starch would be
sufficient to improve the durability and hardness of pellets made from low quality
chops.
To increase the compressive strength of coal briquettes, about 5 to 20% (by
weight) of pitches (petroleum byproducts) have been used (Clarke and Marsh, 1989).
The finer the coal, the more binder it requires (Komarek, 1991). Kim et al. (2002) used
pulp black liquor (a byproduct from the pulp production industry) to increase the
strength of bio-coal briquettes (20-30% corn stalk and 70-80%) coal), which had an
initial breaking strength of 0.5 MPa. Adding 5 to 20% of pulp black liquor increased
the breaking strength of briquettes from 5 to 17 MPa.

3.6-6 Effect of Mixing ofFeed /Biomass Materials

Mixing a feed ingredient or biomass material having high natural binding


capacity with the base feed to improve the strength and durability of the densified
products has been exploited in several studies. Blending 15 to 20% (by weight) of beet
pulp or ground barley increased the durability of rice straw cubes to 80% from an
initial durability of 30% with no added binding materials (Waelti and Dobie, 1973).
Also, rice straw mixed with 25% (by weight) of ground almond hulls resulted in cubes
with 75% durability (Waelti and Dobie, 1973). Mixing of 25, 50, 75, and 100% (by
weight) of rice husk (which had about 40% lignin) with water hyacinth (which had
about 8% lignin) resulted in briquettes with compressive strength of 9.7, 13.4, 25.3,

93
and 37.9 MPa (Bhattacharya et al., 1985). In the study by Winowiski (1988), adding
15 to 60% of wheat to a turkey breeder ration that had an initial pellet durability of
32% increased the pellet durability to 45 to 74%. Addition of 5 to 20% (by weight)
waste paper (Kraft paper, newspaper, or any waste paper) with wheat straw produced
briquettes with compressive strength of 15 to 35 MPa (Demirbas, 1999).
Bradfield and Levi (1984) found that pure hardwoods (red maple, southern red
oak, sweetgum, tupelo, white oak, and yellow poplar) did not produce pellets and they
blocked the pellet-mill die with or without steam addition. Whereas, mixing 15 to 35%
bark with the pure wood produced pellets with about 93 to 99% durability. When
pelleting coal mixed with wood, Chen et al. (1989) found that addition of bark was
required to make pellets without the addition of binders. Pellet durability increased as
the percentage of bark in the furnish (i.e., feed material) increased from 33.3 to 50%
(table 3.8). They mentioned that the higher percentage of extractives (waxes, resins,
and starches) and lignin in the bark could have helped increase the bonding and the
overall pellet strength. They also found that addition of oak furnish with the coal
produced less durable pellets than addition of aspen furnish. This is probably due to
differences in the chemical composition, and the amounts of lignin, hemicellulose, and
extractives of the wood species.
Mechanical strength of lignite coal has been found to increase with 10 to 30%
(by weight) addition of biomass materials such as paper mill waste, sawdust, and pine
cone (Yaman et al., 2001). For instance, at 30% addition of biomass, compressive
strength of briquettes was 32.7 MPa for paper mill waste, 33.2 MPa for sawdust, and
25.0 MPa for pine cone. With no added biomass materials, compressive strength of
lignite coal briquettes was 17.6 MPa. On the other hand, Yaman et al. (2001) found
that addition of olive refuse had a detrimental effect on the mechanical strength of
briquettes.
Table 3.8 presents examples of blending of different feed materials to help
increase the quality of the densified products. Although mixing various feed materials
to improve the strength and durability of the densified products may result in lower
cost of densification, much higher percentage of binding feed ingredient or biomass

94
material (i.e., >20% by weight) must be mixed with the base feed in order to achieve
an equal amount of pellet durability compared to that of chemical binders (i.e., <5%
by weight).

3.6-7Densification Equipment Variables

The pellet mill and roll press are commonly used for densification of various
powder materials in the U.S. Therefore, equipment variables of these two presses are
discussed in detail in the following sections.
Koser et al. (1982) investigated the densification of water hyacinth in a
laboratory press and in three continuous presses (table 3.9). The three presses resulted
in good quality briquettes; however, the pellet mill and ram-extrusion press resulted in
higher throughput than the roll press [corresponding data were not reported by Koser
et al. (1982)]. The compressive strength of pellets made in the pellet mill was lower
than that for the briquettes made in the ram-extrusion press. This was attributed to the
possible breaking points created in the soft pellets due to cutting of pellets with a knife
after leaving the die in the pellet mill. Furthermore, table 3.9 shows that density and
durability values obtained in the laboratory using a closed-end die were similar to
those obtained from the three commercial types of presses. No other study like that by
Koser et al. (1982) was found in the literature.

3.6.7-1 Effect of pressure

Application of pressure by the densification equipment to the feed particles


enables different binding mechanisms. Under high pressure, the natural binding
components such as starch, protein, lignin, and pectin in the feed or biomass materials
are squeezed out of the particles, which contribute to inter-particle bonding. In a pellet
mill, pressures on the order of 100 to 150 MPa (and more) are expected (Koser et al.,
1982; Thomas et al., 1997). In a roll press, pressures of 100 to 200 MPa (and above)
can be obtained (Dec, 1999; Dec, 2002; Dec et al., 2003). Normally, the effect of

95
pressure is studied in laboratory using a closed-end-die and piston assembly, where the
pressure is applied to the powder mass by a universal testing machine or hydraulic
press.
Srivastava et al. (1981) found that increasing pressure from 5 to 44 MPa
increased the wafer (150-mm diameter) durability rating (based on an impact
resistance test) of grass hay (mixed with 20% alfalfa) from 5 to 91%. Singh and
Kashyap (1985) found that increasing pressure from 7.8 to 31.2 MPa increased the
durability of rice husk briquettes from 80 to 95% where the briquettes were made from
rice husk with an average particle size of 4.05 mm and with 25% (by wt.) molasses
added. Shrivastava et al. (1989) reported that rice husk could not form briquettes when
the pressure was in the range of 17.2 to 51.5 MPa at room temperature (i.e., 30C).
When rice husk was preheated from 30 to 225C, the briquette durability (based on an
impact resistance test) was increased from 7 to 72% at a pressure of 68.6 MPa, and
from 40 to 99.8% at a pressure of 257.4 MPa. Sokhansanj et al. (1999) showed that
increasing pressure from 15 to 100 MPa increased the tensile strength of alfalfa pellets
from 0.5 to 2.0 MPa. They also observed that increasing the pressure above 100 MPa
and up to 165 MPa, tensile strength of the pellets started to decrease and leveled off at
1.5 to 2.0 MPa.
Chin and Siddiqui (2000) measured the shear strength of biomass briquettes.
They found that increasing the densification pressure from 1 to 10 MPa increased the
shear strength of briquettes from 27.5 to 95.7 N, from 1.2 to 4.6 N, from 1.3 to 6.7 N,
from 10 to 73.3 N, and from 10 to 36.2 N for sawdust, rice husk, peanut shell, coconut
fiber, and palm fiber, respectively. Li and Liu (2000) studied the densification
behaviors of oak sawdust, oak mulch, oak bark, oak chips, pine sawdust, cottonwood
sawdust, and cottonwood mulch in the pressure range of 34 to 138 MPa. They found
that increasing pressure increased the abrasive resistance, impact resistance, and
compressive resistance of these biomass logs. Table 3.10 shows that increasing
pressure increased the quality of pellets/briquettes made from wood and biomass
materials.

96
3.6.7-2 Pellet mill variables

3.6.7.2-1 Die dimension

Heffner and Pfost (1973) found that smaller die size produced higher
gelatinization of poultry layer feed and higher pellet durability (table 3.11). They also
reported that some gelatinization occurred in the mixer and conditioning chamber, but
most in the pellet die. Their data suggest that the larger the die length to diameter (1/d)
ratio, the higher the pellet durability (table 3.11). Increasing die-thickness (i.e., die
length) or decreasing die-diameter would increase the amount of shear applied to the
feed, which may positively affect the pellet strength and durability. However, too
much shear (i.e., excessively longer die or very small die-diameter) will block the
pellet mill (Heffner and Pfost, 1973). The coefficient of friction between the feed and
the roll or die surface would determine the magnitude of the shear force.
Changing from a thick die (4.8-mm diameter x 57.2-mm length) with no
binder to a thin die (4.8-mm diameter x 44.5-mm length) with 2% calcium
lignosulfonate binder caused no decrease in durability of turkey starter pellets;
however, energy requirements were decreased by 37% or pelleting capacity was
increased by 37% (Pfost, 1964).
Hill and Pulkinen (1988) found that increasing die 1/d ratio from 5 to 9
increased the durability of alfalfa pellets from 50 to 80%. Also, they observed that this
effect was similar to increasing the conditioning temperature from 60 to 87.8C which
resulted in the same magnitude of increase in pellet durability (i.e., from 50 to 80%).
Their study also showed that a die with 1/d of 10 (d = 6.4 mm) produced about ten
percentage points higher durability than a die with 1/d of 8 (d = 6.4 mm) at the
conditioning temperature range of 70 to 104C. The measured durability values at 1/d
ratios of 10 and 8 were 72 to 80%, and 58 to 70%, respectively.
Tabil and Sokhansanj (1996) reported that at die 1/d ratios of 4.1 (d = 7.8 mm),
and 7.31 (d = 6.1 mm), alfalfa pellet durability was 49.8, and 65.8%, respectively.
Furthermore, they found that smaller die (6.1-mm diameter) generally resulted in

97
better durability. However, smaller die (6.1 mm diameter) plugged when moisture
content was more than 10% (w.b.). Larger die (7.8-mm diameter) handled moisture
contents greater than 10% (w.b.), but durability of the pellets was low. Thomas and
van der Poel (1996) reported that a pellet mill die with 5-mm diameter x 25-mm
length, and 5-mm diameter x 35-mm length resulted in barley-pellet hardness of 125,
and 172 N, and durability of 91.8, and 98.1%, respectively, with steam conditioning at
70C. Franke and Rey (2006) suggested that using a large die length would be a
strategy to improve the durability of feed pellets with high fat content.

3.6.7.2-2 Die speed

Die plugging was observed at a die speed of 126 rpm for a corn-based swine
diet (Stevens, 1987). Die speed of 150 to 268 rpm did not affect pellet durability of
corn-based swine diet, and resulted in pellet durability of about 90% (Stevens, 1987).
Also, die speed of 126 to 268 rpm did not affect pellet durability of a wheat-based
swine diet, and resulted in pellet durability of about 98% (Stevens, 1987). High die-
speed (about 10 m/s) is suggested for small pellets (3 to 6-mm diameter), and low die-
speed (about 6 to 7 m/s) is suggested for large pellets and cubes (Thomas et al., 1997).
Heinemans (1991) recommended a low peripheral die speed of 4 to 5 m/s for low bulk
density materials, where a large volume of air must be expelled during pelleting.
Commercial pellet mills can operate at speeds of 60 to 500 rpm, but they
normally operate at a die speed of 375 rpm, which was found to be unsatisfactory for
pelleting alfalfa (Hill and Pulkinen, 1988). Hill and Pulkinen (1988) observed that
either low speed (60 to 125 rpm) or high speed (>344 rpm) did not produce good
quality pellets. They used an optimum speed of 187 rpm for pelleting alfalfa. Tabil
and Sokhansanj (1996) reported that die speeds of 501 and 565 rpm were unsuccessful
in pelleting alfalfa. They concluded that the optimum speed for pelleting alfalfa was
250 rpm.

98
3.6.7.2-3 Gap between roller and die

Robohm and Apelt (1989a) and Robohm (1992) studied the effect of the
distance (gap) between the roller and the die on the strength and durability of pellets in
a flat-die press and in a ring-die press. For both press types, increasing gap-size (about
2 to 2.5 mm) increased pellet hardness and durability. A further increase in gap-size
(about 4 to 5 mm) caused pellet hardness and durability to decrease. The initial
increase in pellet quality was due to a dense layer of material compressed through the
die as a result of increased shear and prolonged pre-compression. A further increase in
gap-size resulted in decreased stability of the feed mash on the edge of the roller and
die because of sideways leaking of the feed mash. Robohm and Apelt (1989a) reported
that pig-feed pellet durability was 96.5, 97.5, 97.7, 97.5, and 97.2 for the gap-size of 0,
1, 2, 3, and 4 mm, respectively, for the flat-die press. For the ring-die press,
throughput, pig-feed pellet durability, and pellet strength were 3.7 t/h, 97.7%, and 57
N, respectively, for zero gap-size; and 5.7 t/h, 97.9%, and 59 N, respectively, for a
gap-size of 5 mm (Robohm and Apelt, 1989a; Robohm, 1992).

3.6.7.2-4 Specific energy input to the pellet mill

Tabil and Sokhansanj (1996) related the specific energy consumed by the
pellet-mill motor to the alfalfa pellet durability. Their study showed that increasing
specific energy input from 95 to 119 MJ/t increased the pellet durability from 25 to
80%. Payne (2004) documented the pellet durability as a function of electrical energy
input to the pellet-mill motor from more than 100 commercial feed mills over a period
of 28 years (1975 to 2003). His report showed that increasing the energy input to the
pellet mill increased the pellet durability. To produce poultry, pig, ruminant, fin fish,
and shrimp pellets with an acceptable durability (i.e., >92% Holmen pellet durability),
specific energy supplied to the pellet press motor must be 36, 43 to 54, 72 to 90, 43,
and 43 MJ/tonne of feed, respectively (Payne, 2006).

99
3.6.7.2-5 Double pelleting

Double pelleting equipment has been used to increase the hardness and
durability of cattle feed consisting of high fiber content (with high resilient behavior),
and swine diet consisting of high fat levels (McEllhiney, 1987; Robohm and Apelt,
1989b). A double pelleting system consists of two presses, serially connected to each
other. The first press, equipped with a conventional barrel type conditioner and a
relatively thin die (e.g., 5-mm diameter x 25-mm length), is used to pre-compress the
feed. Actual pelleting is performed by the second press, equipped with a thicker die
(e.g., 5-mm diameter x 40-mm length). The drawback with double pelleting system is
that it requires more specific energy for pelleting. Robohm and Apelt (1989b) found
that specific energy required for double pelleting was about 29 to 47 MJ/t higher than
that of the single pelleting system.

3.6.7.2-6 Expander

Expander equipment has been used to pre-densify (i.e., de-aerate), shear, and
mix the feed mash, which would alter the structure of the feed particles such that the
binding properties of the particles are enhanced, thereby, improving pellet durability
(Thomas et al., 1997; van der Poel et al., 1997). An expander consists of a conveying
screw with mixing bolts mounted inside a barrel. The screw exerts shearing, mixing,
and transport action to the feed. This moves the feed to a moving cone at the outlet of
the expander, which creates an annular shaped gap. The position of the cone is
controlled by the power take-up of the expander drive. The expander is capable of
raising the temperature of the feed material to above 100C through mechanical shear,
without adding moisture, thus gelatinizing the starch better and improving the binding
characteristics of the feed and producing better quality pellets (Turner, 1995).
Expander conditioning is widely used in Europe, and recently it was introduced in the
U.S. (Behnke, 2006).

100
Robohm (1991) showed that using an expander at pressures of 20 and 30-bar
increased the durability of corn-based pellets to 85% and 89%, respectively, from an
initial control pellet durability of 66%). For raw materials having high pelleting ability
(e.g., wheat-based feed), use of an expander device may not be justified (Robohm,
1991). Peisker (1992) used the expander for conditioning broiler feed with 6 to 13%
fat and obtained pellet durability values of 98.2 to 99.6% for these high-fat feeds.
Moreover, use of an expander increased the total pelleting energy and cost, van der
Poel et al. (1997) concluded that pelleting pig-grower diet with expander conditioning
resulted in higher pellet hardness and durability than pelleting without expander
conditioning (table 3.12).
DeFrain et al. (2003) evaluated an expander as an alternative to steam
conditioning (66C) to pellet feed containing raw soybean hulls and corn steep liquor.
They found that although the expander increased the pellet production (by 250 kg/h)
and pellet durability (by 1 to 2%), the expander consumed about 4 times more
electrical energy (about 150 to 180 MJ/t) than that consumed by the pellet mill (about
40 to 50 MJ/t). Therefore, they concluded that the additional energy expenditure did
not justify the expander use as an alternative method of thermal processing for this
feed mixture.

3.6.7.2- 7 Ripener

A ripener, conditioning equipment with or without steam addition, is used to


increase the feed mash conditioning time. The dwell time of pre-conditioned feed
mash in the ripener is about 15 to 30 min. The ripener helps increase the pellet
durability of high-molasses feeds (7 to 9% of molasses), coarsely ground feeds, low
cereal grain feeds, and feeds with high liquid content (McEllhiney, 1987).

101
3.6.7-3 Roll press variables

Performance of a roll press is controlled by five major parameters (Pietsch,


1997): (1) screw feeder speed (typical screw speeds are 7.5 rpm to 300 rpm); (2) roller
speed (typical roll speeds are 1 to 16 rpm or 0.01 to 1.5 m/s); (3) gap between the
rollers; (4) pressing force expressed as force per unit roller width [typical force ranges
are 10 to 150 kN/cm (Pietsch, 2002)]; and (5) system rigidity (i.e., the response of an
hydraulic pressurizing system to variations in the screw-feeder delivery rate and the
bulk density of the feed; the rolls are pressurized by the hydraulic system).
Komarek (1991) found that increasing pressure applied to the rolls from 120 to
240 MPa increased the crushing strength of bituminous coal briquettes from 225.6 to
892.4 N. Similarly, Dec (1999) showed that increasing roll force from 50 to 60 kN/cm
(roll torque ranged from 4500 to 6000 N-m) increased the crushing strength of the
quick lime briquettes from 75 to 1300 N. Increasing the specific energy supplied to
roll press (i.e., energy input to the roll drive and to the screw feeder) from 54 to 144
MJ/t increased the crushing strength of bituminous coal briquettes from 313.8 to
1333.7 N (Komarek, 1991). Dec (2002) also reported that increasing energy input to
the roll press from 18 to 72 MJ/t increased the durability of municipal sludge
briquettes from 74 to 89%.
The roll press has been used in the pharmaceutical industry to make granules
and tablets (Kleinebudde, 2004). From the review by Kleinebudde (2004), the main
variables that influence the roll pressing process are speed of feeding screws (vertical
or horizontal), roll speed, and compaction pressure (i.e., compaction force per cm of
roll width). Results from a few studies are discussed below to show the effects of roll
press variables that would affect the physical quality of granules and tablets made
from pharmaceutical powders. Inghelbrecht and Remon (1998) investigated the effects
of hydraulic pressure applied to the rolls (2.5 to 17.5 MPa), vertical screw feeder
speed (100 to 1000 rpm), and roll speed (3 to 13 rpm) on the quality of the roller
compacted granules of four types of lactose powders. Good quality (i.e., high
durability based on a tumbling method for pharmaceutical granules) granules were

102
obtained at a pressure of 6.9 MPa, screw feeder speed of 1000 rpm, and roll speed of 7
rpm. They concluded that pressure was the most important parameter, followed by the
roll speed and by screw feeder speed, affecting the quality of lactose granules.
Rambali et al. (2001) studied the effects of compaction force (4 and 8 kN), gap
between the rolls (0.8 and 1.0 mm), and the type of rolls (smooth and ribbed) on the
crushing strength of Miconazole-buccal tablets. They concluded that lower
compaction force, smooth rolls, and larger gap produced stronger tablets. Conversely,
Von Eggelkraut-Gottanka et al. (2002) found that crushing strength of tablets made
from an herbal extract increased from 40 to 140 N when compression force applied to
the rolls increased from 5 to 20 kN/cm. The contradicting results for the effect of roll
force on tablet quality between the two studies [Rambali et al. (2001) and Von
Eggelkraut-Gottanka et al. (2002)] may be attributed to the differences in the
mechanical properties of the powders used in these two studies.

3.6-8 Effect of Post-Production Conditions

3.6.8-1 Time of measurement

Strength of pellets can be measured immediately after manufacturing, which is


called "green strength", or after some curing time (usually, one week), which is called
"cured strength". Usually, the time of measurement is reported along with the strength
and durability values. Mohsenin and Zaske (1976) observed that the time of tumbling
affected the durability of wood, alfalfa, and hay wafers. Higher durability values were
obtained when tumbling immediately after making the wafers than after some storage
period of 3 to 45 min. During the storage period, they observed that wafers dried and
expanded and thus, resulted in lower durability values. Payne (1978) found that dairy
feed pellets had hardness values of 78.5 and 131.4 N when measured immediately
after cooling (i.e., 0 h) and after 24 h, respectively. In this case, solid bridges and
related binding mechanisms could have developed over a period of 24 h and thus,
improved pellet strength.

103
3.6.8-2 Cooling / drying

Pellets leave the pellet press at temperatures of 60 to 95C and moisture


contents of 12 to 18% (w.b.) (Maier et al., 1992). The excess heat and moisture should
be removed from the pellet for safe storage. Pellets are cooled (mostly using forced
air) immediately after the die to within 5C of ambient temperature, and to within
0.5% of the original moisture content of the feed ahead of the conditioner (Turner,
1995). In general, the final moisture content of the pellets should be less than 13%
(w.b.) (Maier et al., 1992). The cooling time may range from about 4 to 15 min
(Stevens, 1987; Maier and Bakker-Arkema, 1992). Pellets that are not properly cooled
can have a reduced durability due to stresses in the pellet between the (cooled) outer
layer and the (still) warmer center, which induces cracks in the pellets. During the
cooling process, soluble components in the feed recrystallize and create bonds
between particles, and viscosity of some components would increase and thus help in
maintaining structural integrity of the pellets (Thomas et al., 1997).
In the feed industry, the primary types of pellet coolers used are the vertical
(cross flow), horizontal (cross flow), carousel (mixed flow), and bunker (counter flow)
type coolers (Maier et al, 1992; Thomas et al., 1997). The horizontal (or belt) single-
or double-deck cooler is the predominant type of cooler used in the feed industry
(Maier et al., 1992). Air properties, amount of airflow, feed material characteristics,
pellet size, pellet bed depth, and residence time in the cooler would affect the cooling
process and also the strength and durability of pellets (Maier and Bakker-Arkema,
1992; Thomas et al, 1997). Friedrich and Robohm (1968) obtained the highest pellet
durability when cooling feed pellets (5-mm diameter) at air velocities of 0.74 to 0.98
m/s (30 < Re < 40; laminar flow) in a three-deck horizontal cooler.

104
3.6.8-3 Storage conditions

Koser et al. (1982) found that after submersion in water at room temperature
for 30 min, the compressive strength of water hyacinth briquettes decreased from 22.6
to 8.8 MPa. Lindley and Vossoughi (1989) estimated that briquettes made from
sunflower stalk, wheat straw, and flax straw absorbed about 9.9, 32.3, and 38.1% of
water after briquettes were immersed for 30 s in water at room temperature. Li and Liu
(2000) observed that biomass logs made from oak sawdust, pine saw dust, and
cottonwood sawdust could not tolerate more than 5 min immersion in water at room
temperature. The biomass logs swelled rapidly when contacting water, and
disintegrated within a few minutes. These studies show that short-term exposure to
rain would be detrimental to the physical quality of the densified products.
Fasina and Sokhansanj (1996) reported that increase in moisture content by
more than 3 to 5% due to storage under high relative humidity (70 to 90%) had
detrimental effect on durability of alfalfa pellets. During storage, an increase in the
moisture content of alfalfa pellets from 7.5% (w.b.) to about 12.5% (w.b.) increased
the durability of pellets from 81 to 85%. However, an increase in the moisture content
of alfalfa pellets from 7.5% (w.b.) to about 19.0% (w.b.) reduced the durability of
pellets from 81 to 75%. Fasina and Sokhansanj (1996) postulated that a small
percentage of increase in moisture content (about 4%) could have helped strengthen
the bond between the individual particles in the pellet due to the binding forces of
water molecules. However, increasing moisture content by more than 4% increased
the volume of the pellet and free water in the particles, which could have reduced the
binding forces between the individual particles in the pellet, and thus, a decrease in the
durability of pellets. Also, Fasina and Sokhansanj (1996) found that increasing the
moisture content of alfalfa pellets from 7.5% (w.b.) to 23.0% (w.b.) increased the
pellet volume by 30%.
Tabil (1996) also studied the effect of high humidity (90%) storage on alfalfa
pellet durability and hardness (i.e., crushing strength). His results are similar to those
of Fasina and Sokhansanj (1996). Tabil (1996) found that increasing moisture content

105
of the pellets from 6.3% (w.b) to 10% (w.b.) did not affect the pellet durability.
Increasing the moisture content beyond 12 to 14% reduced the durability of pellets.
For example, increasing the pellet moisture content from 5.2% (w.b.) to 13.7% (w.b.)
reduced the durability from 82.6 to 48.6%, and increased the pellet volume from 0 to
17.6%. Tabil (1996) found significant reduction in the pellet hardness values when the
pellet moisture content was increased from 5.2 to 8% (w.b.) and more. For example,
increasing the pellet moisture content from 5.2% (w.b.) to 13.7% (w.b.) reduced the
pellet hardness from 666 to 249 N. Tabil (1996) concluded that pellet hardness was
more sensitive to moisture change during storage than pellet durability. Furthermore,
Tabil (1996) attributed the decrease in pellet durability and hardness to the excess
absorbed moisture which weakened the bonds between the particles and increased the
volumetric expansion of the pellets.
Khoshtaghaza et al. (1999) studied the effect of temperature (8 to 40C) and
relative humidity (60 to 90%) on the physical quality of alfalfa cubes (25 x 25 * 40
mm) after 66 or 90 d of storage period. Generally, durability decreased as temperature
and relative humidity increased. There was no effect on durability at low relative
humidity levels (59 to 66%). At the temperature of 31.3C and relative humidity of
81%, durability decreased from 90 to 50% after 66 d of storage. Similarly, after 90 d
of storage at the temperature of 39.1 C and relative humidity of 86%, durability
decreased from 88 to 12%.
In a high humidity storage study, McMullen et al. (2005) found that an
increase in the moisture content of the poultry litter pellets from 6.0 to 10.4% (w.b.)
increased the durability from 92 to 95%. Further increase in moisture content from
10.4 to 22% (w.b.) reduced the durability from 95 to 88%.

106
3.7 Acceptance Levels for Strength and Durability

Currently, there are no standardized criteria on the acceptance levels for


strength (i.e., compressive resistance, impact resistance, and water resistance) and
durability (i.e., abrasive resistance) of the densified products made from the biomass
materials in the U.S. Developing such standards is important for maintaining uniform
quality of the densified products in the market. To set criteria for the physical quality
of feed pellets, animal response to the pellet quality is important. For example, feed
intake by livestock is reduced if the pellet is very hard or has low durability (i.e., high
fines content) (Thomas et al., 1998). Table 3.13 presents the guideline
recommendations for minimum hardness and durability levels (based on feed pellet
diameter) developed by Holmen Chemicals (England) company (Major, 1984). Alfalfa
pellet durability (or pellet quality) was considered to be high, medium, and low when
the pellet durability was above 80%, between 70 and 80%, and below 70%,
respectively (Adapa et al., 2003).
To set criteria on the strength and durability for biomass
pellets/briquettes/cubes, the rigors of handling, transportation, and storage, and
weather conditions of the locations where the products are transported/exported
(Khoshtaghaza et al., 1999) must be considered. Dobie (1961) suggested that a small
amount of fines up to 5% (by weight) would be an acceptable level. More than 5% of
fines content may create problems: storage capacity may be reduced, and the flow
characteristics of the hay wafers may be worsened. Waelti and Dobie (1973) classified
the rice straw cubes as "good" when the durability was between 80 and 90%, and as
"very good" when durability was 90% and above.
For coal briquettes, Richards (1990) recommended an acceptance level for
compressive strength as 375 kPa, an acceptance level for the impact resistance index
(IRI) [previously described] as 50, and an acceptance level for the water resistance as
95%. Water resistance was defined as 100 minus percentage of water absorbed after
30 minutes of immersion in water at room temperature. Raghavan and Conkle (1991)
suggested acceptance levels on the compressive strength of coal pellets/briquettes as

107
0.0068, 0.0271, and 0.0068 N/mm3 (volume basis) for green, cured (i.e., after one
week), and wet pellets/briquettes, respectively. Their acceptance criterion for impact
resistance was that pellets/briquettes should remain intact after dropping from both 1
and 5 ft onto a concrete floor.
Richards (1990) set an acceptance criterion for durability of sub-bituminous
coal briquettes as 95%. He used a tumbling test that involved tumbling 60 briquettes,
each weighing about 20 g (i.e., total weight was 1200 g) for 100 revolutions at 50 rpm.
The charge was screened on a 3.0-mm screen and the percentage oversize was
measured as the durability. Raghavan and Conkle (1991) developed acceptance
criteria for raw coals, and coal pellets/briquettes. Their acceptance criterion for
minimum durability for raw coals was 80%, and for coal pellets/briquettes was > 28%.
To measure durability, about 500-g of coal or coal pellets/briquettes were tumbled at
50 rpm for 10 min. The percentage of original weight of coal or coal pellets/briquettes
retained on a 4.76 mm screen was calculated as the durability.
According to Austrian standards for wood pellets (ONORM M 7135, 2000),
fines that pass through a 3.15-mm sieve after the Ligno durability test should be less
than 2.3% (by weight) (Obernberger and Thek, 2004). This corresponds to a pellet
durability value of 97.7% assuming that the size of feed particles is less than 3.15 mm.
Also, Obernberger and Thek (2004) indicated that similar standards on pellet/briquette
durability have been developed by Sweden.
The U.S. Pellet Fuels Institute (PFI) (Arlington, VA:
http://www.pelletheat.org) has established standards for residential pellet fuel made
from biomass materials such as wood, paper, and agricultural residues. According to
PFI standards, there are two grades of pellet fuel - premium grade and standard grade.
For both pellet fuel grades, the fines that pass through a 3.2-mm screen should be no
more than 0.5% by weight to avoid dust while loading, and to avoid problems with
pellet flow during operation. It appears that the criterion set on the PFI standards for
the fines content is based on the sieving tests at the time of packing, and not based on
any physical tests performed to estimate the durability (or fines).

108
3.8 Discussion

In this Chapter, literature on the densification of animal feed, wood, biomass,


coal, and other powder materials such as pharmaceutical powders was reviewed. In the
past, strength and durability of densified products such as pellets, briquettes, and cubes
have been estimated based on compressive resistance, abrasive resistance, impact
resistance, and water resistance. Several procedures have been used for measuring
these mechanical strength characteristics of densified products. For animal feed
pellets, compressive resistance (i.e., hardness) and durability are usually determined.
In the feed industry, several types of hand-operated or semi-automatic instruments
have been used to measure pellet hardness. The tumbling can method descried in the
ASABE Standards (2003) is the method predominantly used for measuring durability
of feed pellets in the U.S. It should be noted that the durability test given in ASABE
Standards (2003) would indicate the ability of the densified products to retain
structural integrity, and it does not classify the fines (based on size) produced in the
durability test.
Feedstock variables (starch, protein, fiber, fat, lignin and extractives, moisture
content, and particle size and its distribution), pre-conditioning processes (steam
conditioning/preheating, and addition of binders), and densification equipment
variables (forming pressure, pellet mill and roll press variables) affected the strength
and durability of densified products. The factors that increase the strength and
durability would also increase the density and specific energy consumption. However,
high density or high specific energy input does not necessarily indicate high strength
or durability of the densified products. Factors affecting density and specific energy
can be found elsewhere (Chapter 2; Mani et al., 2003). Densification may result in
products with bulk density of 600 to 800 kg m" . This is about 2 to 10 fold increase in
the bulk density of the biomass. The U.S. PFI recommends a minimum bulk density of
641 kg m"3 (40 lb ft"3) for both premium and standard pellet fuel grades. According to
Swedish standard SS 187120, minimum requirement for bulk density of biomass
pellets is 600 kg m"3 (Obernberger and Thek, 2004). In general, the specific energy

109
required for pelleting (i.e., energy consumed by the pellet mill motor) may range from
14 to 144 MJ/t (Israelsen et al., 1981; Stevens, 1987; Tabil et al., 1997). In addition,
steam conditioning/preheating the feed may require considerable amount of energy.
For example, Skoch et al. (1981) estimated that steam conditioning to increase the
temperature from 27 to 80C consumed about 94 MJ/t. Bhattacharya et al. (1985) and
Smith et al. (1977) indicated that specific energy required for densification of biomass
including preheating energy (but excluding the energy involved for drying and size
reduction) was about 6 to 9% of the energy in the biomass materials.
According to Behnke (1994), Turner (1995), and Thomas et al. (1997),
contribution of different factors on the durability of feed pellets can be categorized as
follows: diet formulation = 40%; particle size = 20%; steam conditioning = 20%; die
specifications = 15%; and cooling/drying = 5%. When including an expander in the
conditioning process, the distribution becomes: diet formulation = 25%; particle size =
15%; conditioning (steam and expander) = 40%; die specifications = 15%; and
cooling/drying = 5%. The above classifications show that the effect of densification
equipment per se on the durability of densified products is small compared to the
cumulative effects of the variables related to the feedstock and pre-conditioning
processes. Thus, conditions that give good quality densified products in one type of
densification equipment (e.g., pellet mill) may be applied to another type of
densification equipment (e.g., roll press), at least as starting conditions.
Table 3.14 summarizes the optimum conditions for densification, and
strategies to improve the quality of the densified products. From table 3.14, starting
conditions for densification of biomass feedstock or animal feed in different
densification equipment can be obtained. Also, information presented in this Chapter
would help formulate/adjust feed ingredients that would lead to high quality densified
products. Table 3.15 shows an example of how to systematically manipulate the
densification conditions/variables in order to improve the pellet quality.
In reality, factors related to the feed material characteristics, pre-conditioning
processes, and densification equipment variables interact with one another. Therefore,
these variables should be optimized using statistical- or mathematical-based

110
optimization procedures to produce strong and durable densified products. While
optimizing these variables, in addition to strength and durability, specific energy
consumption, production rate, maintenance or production cost should also be
considered (Dec, 1999). Also, each feed material requires specific optimum
densification conditions.
Minimal handling and properly designed handling systems would also help
maintain the initial strength and durability of the densified products. For example,
pneumatic conveying of pellets at 10 m/s produced about 3.5% less fines than that at
30 m/s (Thomas and van der Poel, 1996). During pneumatic handling, pipe bends with
350 to 600-mm radius produced less fines than pipe bends with 150-mm radius
(Aarseth, 2004). Mina-Boac et al. (2006) studied the repeated mechanical handling of
feed pellets (6.4-mm diameter) in a bucket elevator-bin system. They found that after
eight transfers between two commercial-size bins using a bucket elevator, the initial
durability of the pellets (i.e., 93%) was not affected by the repeated handling process.
However, the repeated handling process resulted in about 33% more broken pellets (<
5.6-mm diameter) than that measured initially (i.e., 18%), and 0.07% (by weight of
pellets handled) dust after each transfer. This study shows that pellets with an initial
durability of > 93% can be safely handled in mechanical handling systems with less
fines production.

Proper and timely troubleshooting of the problems at different stages of the


densification process, and operating the densification equipment by skilled persons
would help reduce the chances of producing poor quality densified products (Maier
and Gardecki, 1992; Payne, 2006). Automation of the densification equipment
operation and related processes is in progress in this field.
In the U.S., there are no standardized criteria for acceptable levels of strength
(i.e., compressive resistance, impact resistance, and water resistance) and durability
(i.e., abrasive resistance) of the densified products made from biomass materials. From
a marketing point of view, there is a need to develop standards on the quality of
different types of densified products (i.e., pellets, briquettes, and cubes), especially to
set acceptable levels for strength and durability values. The U.S. PFI standards on fuel

111
pellet quality need to be updated to include a durability test-based criterion for fines
content. The selection of the type of durability test (i.e., tumbling can method vs.
Holmen tester or Ligno tester) depends on how the densified products would be
handled after manufacturing and until consumption (i.e., mechanical vs. pneumatic
handling).

3.9 Conclusions

Densification of biomass materials into pellets, briquettes, or cubes could


reduce costs and problems with handling, transportation, storage, and utilization of
low bulk density biomass materials. To produce good quality (i.e., high strength and
durability) densified products from biomass feedstocks whose densification
characteristics are unknown, directions or ways to make strong and durable densified
products are needed. We reviewed the literature to develop strategies to improve
strength and durability of the densified products or to provide conditions to densify
new biomass feedstocks or feed formulations that lead to strong and durable densified
products. The following conclusions could be drawn:

Factors related to the feedstock (starch, protein, fiber, fat, lignin and extractives,
moisture content, and particle size and its distribution), pre-conditioning processes
(steam conditioning/preheating, and addition of binders), and densification
equipment (forming pressure, and pellet mill and roll press variables) affect the
strength and durability of the densified products. Also, post-production conditions
such as cooling/drying and high humidity storage conditions would influence
strength and durability of the densified products.

Generalized optimum conditions related to the feed material, pre-conditioning


processes, and densification equipment were established for manufacturing high
quality densified products by pelleting, briquetting, or cubing.

112
There is a need to develop standards on the minimum acceptable levels for the
strength (i.e., compressive resistance, impact resistance, and water resistance), and
durability (i.e., abrasive resistance) of the densified products made from biomass
feedstocks and animal feeds in the U.S. Guidelines for developing such standards
are presented.

3.10 References

Aarseth, K.A. 2004. Attrition of feed pellets during pneumatic conveying: the
influence of velocity and bend radius. Biosys. Eng. 89(2): 197-213.
Adapa, P.K., G.J. Schoenau, L.G. Tabil, S. Sokhansanj, and B. Crerar. 2003. Pelleting
of fractionated alfalfa products. ASAE Paper No. 036069. St. Joseph, Mich.:
ASABE.

Al-Widyan, M.I., and H.F. Al-Jalil. 2001. Stress-density relationship and energy
requirement of compressed olive cake. Applied Eng. inAgric. 17(6): 749-753.
Angulo, E., J. Brufau, and E. Esteve-Garcia. 1995. Effect of sepiolite on pellet
durability in feeds differing in fat and fibre content. Animal Feed Set Tech.
53(3-4): 233-241.
Angulo, E., J. Brufau, and E. Esteve-Garcia. 1996. Effect of a sepiolite product on
pellet durability in pig diets differing particle size and in broiler starter and
finisher diets. Animal Feed Sci. Tech. 63(1-4): 25-34.

ASABE Standards. 2003. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: ASABE.
ASTM. 1998a. C39-96: Standard test method of compressive strength of cylindrical
concrete specimens. Annual book of ASTM Standards, vol. 04.02. West
Conshohocken, PA: American Society for Testing and Materials, pp. 17-21.

113
ASTM. 1998b. D441-86: Standard test method of tumbler test for coal. Annual book
of ASTM Standards, vol. 05.05. West Conshohocken, PA: American Society
for Testing and Materials, pp. 192-194.

ASTM. 1998c. D440-86: Standard test method of drop shatter test for coal. Annual
book of ASTM Standards, vol. 05.05. West Conshohocken, PA: American
Society for Testing and Materials, pp. 188-191.

Behnke, K.C. 1994. Factors affecting pellet quality. Proc. of Maryland Nutrition
Conference. Department of Poultry Science and Animal Science, College of
Agriculture, University of Maryland, College Park.
Behnke, K.C. 2006. The art (science) of pelleting. Feed Technology Technical Report
Series, American Soybean Association International Marketing Southeast Asia,
Singapore, pp. 5-9.
Bhattacharya, S.C., G.Y. Saunier, N. Shah, and N. Islam. 1985. Densification of
biomass residues in Asia. In Bioenergy 84. Vol III. Biomass Conversion, 559-
563. H. Egneus and A. Ellegard, eds. London: Elsevier Applied Science
Publishers.

Bradfield, J., and M.P. Levi. 1984. Effect of species and wood to bark ratio on
pelleting of southern woods. Forest Products J. 34(1): 61-63.

Briggs, J.L., D.E. Maier, B.A. Watkins, and K.C. Behnke. 1999. Effects of ingredients
and processing parameters on pellet quality. Poultry Sci. 78(10): 1464-1471.
Cavalcanti, W.B. 2004. The effect of ingredient composition on the physical quality of
pelleted feeds: a mixture experimental approach. Ph.D. diss. Manhattan, KS:
Kansas State University.
Chen, P.Y.S., J.G. Haygreen, and M.A. Graham. 1989. An evaluation of wood/coal
pellets made in a laboratory pelletizer. Forest Products J. 39(7/8): 53-58.
Chin, O.C., and K.M. Siddiqui. 2000. Characteristics of some biomass briquettes
prepared under modest die pressures. Biomass and Bioenergy 18(3): 223-228.
Clarke, D.E., and H. Marsh. 1989. Factors influencing properties of coal briquettes.
Fwe/68(8): 1031-1038.

114
Dec, R.T. 1999. Selection of proper roll press operating parameters. Proc. of the
Institute for Briquetting and Agglomeration (IBA), Vol. 26. pp. 29-36.
Dec, R.T. 2002. Optimization and control of roller press operating parameters. Powder
Handling & Process. 14(3): 222-225.

Dec, R.T., A. Zavaliangos, and J.C. Cunningham. 2003. Comparison of various


methods for analysis of powder compaction in roller press. Powder Tech.
130(1-3): 265-271.
DeFrain, J.M., J.E. Shirley, K.C. Behnke, E.C. Titgemeyer, and R.T. Ethington. 2003.
Development and evaluation of a pelleted feedstuff containing condensed corn
steep liquor and raw soybean hulls for dairy cattle diets. Animal Feed Sci.
Tech. 107(1-4): 75-86.
Demirbas, A. 1999. Physical properties of briquettes from waste paper and wheat
straw mixtures. Energy Conversion and Mgmt. 40(4): 437-445.
Dobie, J.B. 1961. Materials-handling systems for hay wafers. Agric. Eng. 42(12): 692-
697.
Dobie, J.B. 1975. Cubing tests with grass forages and similar roughages sources.
Trans. ASAE \S(5): 864-866.
Fairchild, F., and D. Greer. 1999. Pelleting with precise mixture moisture control.
Feed Int. 20(8): 32-36.
Fasina, O.O., and S. Sokhansanj. 1996. Storage and handling characteristics of alfalfa
pellets. Powder Handling & Processing 8(4): 361-365.
Franke, M., and A. Rey. 2006. Pelleting quality. World Grain May 2006: 78-79.
Friedrich, W., and K.F. Robohm. 1968. Die Abriebfestigkeit von Pellets und ihre
Abhangigheit vom Pressprozess insbesondere der Kuhlung. Kraftfutter 52: 59-
64.
Ghebre-Sellassie, I. 1989. Pharmaceutical Pelletization Technology. New York, NY:
Marcel Dekker.

115
Gilpin, A.S., TJ. Herrman, K.C. Behnke, and F.J. Fairchild. 2002. Feed moisture,
retention time, and steam as quality and energy utilization determinants in the
pelleting process. AppliedEng. inAgric. 18(3): 331-338.

Grover, P.D., and S.K. Mishra. 1996. Biomass Briquetting: Technology and Practices.
Regional Wood Energy Development Program in Asia, Field Document No.
46. Bangkok, Thailand: Food and Agriculture Organization of the United
Nations.

Heffner, L.E., and H.B. Pfost. 1973. Gelatinization during pelleting. Feedstuff45(23):
33.

Heinemans, H. 1991. The interaction of practical experience and the construction of


new pelleting and cooling machinery. Advances in Feed Tech. 6: 24-38.
Hill, B., and D.A. Pulkinen. 1988. A Study of the Factors Affecting Pellet Durability
and Pelleting Efficiency in the Production of Dehydrated Alfalfa Pellets.
Saskatchewan, Canada: Saskatchewan Dehydrators Association.

Holley, C.A. 1983. The densification of biomass by roll briquetting. Proc. of the
Institute for Briquetting and Agglomeration (IBA), Vol. 18. pp. 95-102.
Inghelbrecht, S., and J.P. Remon. 1998. The roller compaction of different types of
lactose. Int. J. Pharm. 166(2): 135-144.

Israelsen, M., J. Busk, and J. Jensen. 1981. Pelleting properties of dairy compounds
with molasses, alkali-treated straw and other byproducts. Feedstuff's 7: 26-28.
Iyengar, M. 1959. The problem of briquetting slack coal in India. Proc. of 6th Biennial
Briquetting Conf. of the International Briquetting Association, August 24-26,
Glacier Park, MT. pp. 20-29.
Kaliyan, N., and R.V. Morey. 2005. Densification of corn stover. ASAE Paper No.
056134. St. Joseph, Mich.: ASABE.
Khankari, K.K., M. Shrivastava, and R.V. Morey. 1989. Densification characteristics
of rice hulls. ASAE Paper No. 89-6093. St. Joseph, Mich.: ASABE.
Khoshtaghaza, M.H., S. Sokhansanj, and B.D. Gossen. 1999. Quality of alfalfa cubes
during shipping and storage. Applied Eng. inAgric. 15(6): 671-676.
116
Kim, H., G. Lu, T. Li, and M. Sadakata. 2002. Binding and desulfurization
characteristics of pulp black liquor in biocoalbriquettes. Environ. Sci. Tech.
36(7): 1607-1612.

Kleinebudde, P. 2004. Roll compaction/dry granulation: pharmaceutical applications.


Eur. J. Pharm. Biopharm. 58(2): 317-326.
Komarek, R.K. 1991. Binderless briquetting of peat, lignite, sub-bituminous and
bituminous coals in roll presses. Proc. of the Institute for Briquetting and
Agglomeration (IBA), Vol. 22. pp. 223-242.
Koser, H.J.K., G. Schmalstieg, and W. Siemers. 1982. Densification of water
hyacinth-basic data. Fuel 61(9): 791-798.

Li, Y., and H. Liu. 2000. High-pressure densification of wood residues to form an
upgraded fuel. Biomass andBioenergy 19(3): 177-186.
Lindley, J.A., and M. Vossoughi. 1989. Physical properties of biomass briquets.
Trans. ASAE 32(2): 361-366.
Lund, D. 1984. Influence of time, temperature, moisture, ingredients and processing
conditions on starch gelatinization. CRC Crit. Rev. Food Sci. Nut. 20: 249-273.
MacBain, R. 1966. Pelleting Animal Feed. Chicago, IL: American Feed
Manufacturing Association.
Maier, D.E., and F.W. Bakker-Arkema. 1992. The counterflow cooling of feed pellets.
J. Agric. Eng. Res. 53: 305-319.

Maier, D.E., and J. Gardecki. 1992. Feed mash conditioning field case studies. ASAE
Paper No. 92-1541. St. Joseph, Mich.: ASABE.
Maier, D.E., R.L. Kelley, and F.W. Bakker-Arkema. 1992. In-line, chilled air pellet
cooling. FeedMgmt. 43(1): 28-32.
Major, R. 1984. The pneumatic method. FeedMgmt. 35(6): 20-26.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2003. An overview of compaction of biomass
grinds. Powder Handling & Processing 15(3): 160-168.
Marrero, T.R. 1999. Theory and application of binders: an update. Proc. of the
Institute for Briquetting and Agglomeration (IBA), Vol. 26. pp. 103-109.
117
McEllhiney, R.R. 1987. What's new in pelleting. FeedMgmt. 38(2): 32-34.
McEllhiney, R.R. 1988. Mill management feedback. FeedMgmt. 39(6): 37-39.
McMullen, J., 0 . 0 . Fasina, C.W. Wood, and Y. Feng. 2005. Storage and handling
characteristics of pellets from poultry litter. Applied Eng. inAgric. 21(4): 645-
651.
Mina-Boac, J., R.G. Maghirang, and M.E. Casada. 2006. Durability and breakage of
feed pellets during repeated elevator handling. ASABE Paper No. 066044. St.
Joseph, Mich.: ASABE.

Mohsenin, N., and J. Zaske. 1976. Stress relaxation and energy requirements in
compaction of unconsolidated materials. J. Agric. Eng. Res. 21(2): 193-205.
Moran, E.T., Jr. 1989. Effect of pellet quality on the performance of meat birds. In
Recent Advances in Animal Nutrition, 87-108. W. Haresign and D.J.A. Cole,
eds. London, England: Butterworths.
Moritz, J.S., K.J. Wilson, K.R. Cramer, R.S. Beyer, L.J. McKinney, W.B. Cavalcanti,
and X. Mo. 2002. Effect of formulation density, moisture, and surfactant on
feed manufacturing, pellet quality, and broiler performance. J. Appl. Poult.
Res. 11: 155-163.
O'Dogherty, M.J., and J.A. Wheeler. 1984. Compression of straw to high densities in
closed cylindrical dies. J. Agric. Eng. Res. 29(1): 61-72.
Obernberger, I., and G. Thek. 2004. Physical characterisation and chemical
composition of densified biomass fuels with regard to their combustion
behavior. Biomass and Bioenergy 27(6): 653-669.
ONORM M 7135. 2000. Compressed wood or compressed bark in natural state -
pellets and briquettes, requirements and test specifications. Vienna, Austria:
Osterreichisches Normungsinstitut.
Payne, J.D. 1978. Improving quality of pellet feeds. Milling Feed and Fertilizer
162(May): 34-41.
Payne, J.D. 2004. Predicting pellet quality and production efficiency. World Grain
August 2004: 68-70.
118
Payne, J.D. 2006. Troubleshooting the pelleting process. Feed Technology Technical
Report Series, American Soybean Association International Marketing
Southeast Asia, Singapore, pp. 17-23.
Peisker, M. 1992. Improving feed quality by expansion. Int. Milling Flour Feed March
1992: 15-20.
Pfost, H.B. 1964. The effect of lignin binders, die thickness and temperature on the
pelleting process. Feedstuffs 36(22): 20, 54.

Pfost, H.B., and L.R. Young. 1973. Effect of colloidal binders and other factors on
pelleting. Feedstuffs 45(49): 21-22.
Pickard, G.E., W.M. Roll, and J.H. Ramser. 1961. Fundamentals of hay wafering.
Trans. ASAE 4(1): 65-68.
Pietsch, W. 1997. Granulate dry particulate solids by compaction and retain key
powder particle properties. Chemical Eng. Progress April 1997: 24-46.
Pietsch, W. 2002. Agglomeration Processes - Phenomena, Technologies, Equipment.
Weinheim: Wiley-VCH.
Raghavan, J.K., and H.N. Conkle. 1991. Physical characteristic measurements for
reconstituted coal pellets. Proc. of the Institute for Briquetting and
Agglomeration (IBA), Vol. 22. pp. 85-96.
Rambali, B., L. Baert, E. Jans, and D.L. Massart. 2001. Influence of the roll compactor
parameter settings and the compression pressure on the buccal bio-adhesive
tablet properties. Int. J. Pharm. 220(1-2): 129-140.
Reece, F.N. 1966. Temperature, pressure, and time relationships in forming dense hay
wafers. Trans. ASAE 9(6): 749-751.
Reece, F.N., B.D. Lott, and J.W. Deaton. 1985. The effects of hammer mill screen size
on ground corn particle size, pellet durability, and broiler performance. Poultry
Sci. 65: 1257-1261.

Richards, S.R. 1990. Physical testing of fuel briquettes. Fuel Process. Tech. 25(2): 89-
100.

119
Richardson, W., and E.J. Day. 1976. Effect of varying levels of added fat in broiler
diets on pellet quality. Feedstuff's 48(20): 24.

Robohm, K.F. 1991. Erfahrungen zum Einsatz des Druckkonditioneurs in Verbindung


mit Futtermittelpressen. Int. ZDS-Fachtagung SIS-11, Extrusions-Tagung '91,
23-25 September, Solingen, Germany, pp: 1-10.
Robohm, K.F. 1992. Adjustable roll gap: benefits to energy demand, throughput and
pellet durability. Feed Int. 13(3): 30-35.
Robohm, K.F., and J. Apelt. 1989a. Die automatische Spaltweitenverstellung. Die
Muhle+ Mischfuttertechnik 126: 271-275.
Robohm, K.F., and J. Apelt. 1989b. Verhoging van de flexibiliteit door gebruik van
het perssysteem met voorverdichting. De Molenaar 92: 615-625.
Rumpf, H. 1962. The strength of granules and agglomeration. In Agglomeraion, 379-
418. W.A. Knepper, ed. New York, NY: John Wiley.
Sah, P., B. Singh, and U. Agrawal. 1980. Compaction behavior of straw. J. Agric.
Eng.-India 18(1): 89-96.
Salmon, R.E. 1985. Effects of pelleting, added sodium bentonite and fat in a wheat-
based diet on performance and carcass characteristics of small turkeys. Animal
FeedSci. Tech. 12(3): 223-232.
Shrivastava, M., P. Shrivastava, and K.K. Khankari. 1989. Densification
characteristics of rice husk under cold and hot compression. In Agricultural
Engineering: Proceedings of the 11th International Congress on Agricultural
Engineering, Dublin, 4-8 September 1989, 2441-2443. V.A. Dodd and P. M.
Grace, eds. Rotterdam: A.A. Balkema Pub.

Singh, A., and Y. Singh. 1982. Briquetting of paddy straw. Agric. Mech. in Asia,
Africa and Latin America 13(4): 42-44.
Singh, D., and M.M. Kashyap. 1985. Mechanical and combustion characteristics of
paddy husk briquettes. Agric. Wastes \3(3): 189-196.
Skoch, E.R., K.C. Behnke, C.W. Deyoe, and S.F. Binder. 1981. The effect of steam-
conditioning rate on the pelleting process. Animal FeedSci. Tech. 6(1): 83-90.
120
Smith, I.E., S.D. Probert, R.E. Stokes, and R.J. Hansford. 1977. The briquetting of
wheat straw. J, Agric. Eng. Res. 22(2): 105-111.
Sokhansanj, S., and H.C. Wood. 1991. Engineering aspects of forage processing for
pellets, cubes, dense chops and bales. Advances in Feed Tech. 5: 6-23.
Sokhansanj, S., R.T. Patil, G. Ahmadnia, O.O. Fasina, and J. Irudayaraj. 1991.
Procedures for evaluating durability and density of forage cubes and pellets.
CASE Paper No. 91-402. Canadian Society of Agricultural Engineering-
Agricultural Institute of Canada Annual Conference, July 29-31, 1991 -
Fredericton, New Brunswick.

Sokhansanj, S., L. Tabil, Jr., and W. Yang. 1999. Characteristics of plant tissue to
form pellets. Powder Handling & Process. 11(2): 149-159.
Srivastava, A.C., W.K. Bilanski, and V.A. Graham. 1981. Feasibility of producing
large-size hay wafers. Can. Agric. Eng. 23(2): 109-112.
Stark, C.R. 1994. Pellet quality. I. Pellet quality and its effect on swine performance.
II. Functional characteristics of ingredients in the formation of quality pellets.
Ph.D. diss. Manhattan, KS: Kansas State University.
Stevens, C.A. 1987. Starch gelatinization and the influence of particle size, steam
pressure and die speed on the pelleting process. Ph.D. diss. Manhattan, KS:
Kansas State University.
Tabil, L., Jr., and S. Sokhansanj. 1996. Process conditions affecting the physical
quality of alfalfa pellets. Applied Eng. in Agric. 12(3): 345-350.
Tabil, L.G., Jr. 1996. Binding and pelleting characteristics of alfalfa. Ph.D. diss.
Saskatoon, Saskatchewan, CA: University of Saskatchewan, Department of
Agricultural and Bioresource Engineering.
Tabil, L.G., Jr., S. Sokhansanj, and R.T. Tyler. 1997. Performance of different binders
during alfalfa pelleting. Can. Agric. Eng. 39(1): 17-23.
Temmerman, M., F. Rabier, P.D. Jensen, H. Hartmann, and T. Bohm. 2006.
Comparative study of durability test methods for pellets and briquettes.
Biomass andBioenergy 30(11): 964-972.

121
Thomas, M., and A.F.B. van der Poel. 1996. Physical quality of pelleted animal feed.
1. Criteria for pellet quality. Animal FeedSci. Tech. 61(l-4):89-l 12.

Thomas, M., D.J. van Zuilichem, and A.F.B. van der Poel. 1997. Physical quality of
pelleted animal feed. 2. Contribution of processes and its conditions. Animal
Feed Sci. Tech. 64(2-4): 173-192.
Thomas, ML, T. van Vliet, and A.F.B. van der Poel. 1998. Physical quality of pelleted
animal feed: 3. Contribution of feedstuff components. Animal Feed Sci. Tech.
70(1-2): 59-78.
Turner, R. 1995. Bottomline in feed processing: achieving optimum pellet quality.
FeedMgmt. 46(12): 30-33.
van der Poel, A.F.B., H.M.P. Fransen, and M.W. Bosch. 1997. Effect of expander
conditioning and/or pelleting of a diet containing tapioca, pea and soybean
meal on the total tract digestibility in growing pigs. Animal Feed Sci. Tech.
66(1-4): 289-295.

Varadharaju, N., and L. Gothandapani. 1998. Design and development of equipment


for pelleting decomposed coir pith. Agric. Mech. in Asia, Africa and Latin
America 29(2): 33,34,38.
Vest, L. 1993. Southeastern survey: factors which influence pellet production and
quality. FeedMgmt. 44(5): 60-68.
Von Eggelkraut-Gottanka, S.G., S.A. Abed, W. Muller, and P.C. Schmidt. 2002.
Roller compaction and tabletting of St. John's wort plant dry extract using a
gap width and force controlled roller compactor. I. Granulation and tabletting
of eight different extract batches. Pharm. Dev. Tech. 7(4): 433-445.

Waelti, H., and J.B. Dobie .1973. Cubability of rice straw as affected by various
binders. Trans. ASAE 16(2): 380-383.
Wamukonya, L., and B. Jenkins. 1995. Durability and relaxation of sawdust and
wheat-straw briquettes as possible fuels for Kenya. Biomass and Bioenergy
8(3): 175-179.
Winowiski, T. 1985. Optimizing pelleting temperature. FeedMgmt. 36(7): 28-33.

122
Winowiski, T. 1988. Wheat and pellet quality. FeedMgmt. 39(9): 58-64.
Winowiski, T. 1998. Examining a new concept in measuring pellet quality: which test
is best? FeedMgmt. 49(1): 23-26.
Winowiski, T.S. 2006. Factors that affect pellet quality and trouble-shooting the
pelleting process. Technical Bulletin on Feed Technology, Vol. FT23-1995.
Singapore: American Soybean Association. Available at: www.asasea.com.
Accessed 8 August 2006.
Wood, J.F. 1987. The functional properties of feed raw materials and the effect on the
production and quality of feed pellets. Animal Feed Sci. Tech. 18(1):1-17.
Yaman, S., M. Sahan, H. Haykiri-Acma, K. Sesen, and S. Kticukbayrak. 2001. Fuel
briquettes from biomass-lignite blends. Fuel Process. Tech. 72(1): 1-8.
Young, L.R., H.B. Pfost, and A.M. Feyerherm. 1963. Mechanical durability of feed
pellets. Trans. ASAE 6(2): 145-147, 150.

123
Table 3.1. Pellet durability index (%) of various types of commercial feed pellets
determined by four durability measurement methods (Winowiski, 1998).
Type of feed Tumbling can Tumbling can + Holmen tester Ligno tester
pellet two %-in.
hexagonal nuts
Broiler, 2% fat 89.1 68.2 68.5 64.9
Turkey grower 94.7 82.0 84.6 87.2
Beef 96.1 91.6 89.0 94.0
Dairy 96.7 91.0 90.2 93.6
Pig Starter 95.5 83.8 80.5 82.4
Rabbit 98.4 96.5 96.5 97.7

Table 3.2. Effect of fat added to the feed before pelleting on the pellet durability.
Type of feed pellet Fat added Pellet durability (%) Reference
before pelleting
(%)
Broiler finisher diet 1.0 82.0 Richardson and Day
(1976)
2.0 78.0
3.0 70.8
4.0 68.4
5.33 49.2
Pig grower diet 2.3 93.8 Anguloetal. (1995)
5.7 81.9
Sow diet 2.4 96.1 Anguloetal. (1995)
5.9 85.7
Rabbit diet 2.6 97.4 Anguloetal. (1995)
5.8 91.8
Corn-soybean meal 2.9 89.0 Briggsetal. (1999)
based diet
7.5 60.0

124
Table 3.3 Effect of lignin and extractives content on the pellet durability (Bradfield
and Levi, 1984).
Wood 85% wood:: 15% bark 65% wood:: 35% bark
species Lignin Extractives Pellet Lignin Extractives Pellet
(%) (%) durability (%) (%) durability
(%) (%)
Red 20.65 3.83 98.8 20.78 5.87 98.8
maple
Southern 19.9 6.12 98.5 21.59 7.63 99.2
red oak
Sweetgum 22.32 5.1 99.0 24.38 6.77 99.4
Tupelo 27.68 6.55 98.4 29.17 8.18 97.8
White oak 20.74 7.88 99.0 22.51 8.51 99.4
Yellow 18.75 5.79 99.0 18.34 8.63 99.2
poplar
Loblolly 27.3 8.81 97.2 31.59 9.31 93.6
pine

Table 3.4. Recommended feed particle size distribution to produce high quality pellets
(Payne, 2006).
Sieve size (mm) Percentage of material retained on the sieve (%)
3.0 up to 1%
2.0 up to 5%
1.0 around 20%
0.5 around 30%
0.25 around 24%
<0.25 not less than 20%

125
Table 3.5. Effect of feed particle size on the pellet durability.
Feed type Hammer mill Geometric Pellet Reference
screen size (mm) mean diameter durability (%)
of particles
(mm)
Wheat-based 1.58 0.54 97.4 Stevens (1987)
swine diet
3.18 0.80 96.7
6.35 1.0 92.4
Pig grower diet 3.0 Not available 88.2 Angulo et al.
(1996)
6.0 Not available 82.4
Alfalfa 2.4 Not available 75.4 Tabil and
Sokhansanj (1996)
3.2 Not available 69.0
Alfalfa 2.78 Not available 47.0 Hill and Pulkinen
(1988)
4.76 Not available 41.0
6.35 Not available 37.0
Corn stover 3.0 0.66 75.2 Kaliyan and Morey
(2005)
4.6 0.80 61.6

Table 3.6. Effect of steam conditioning on the quality of pellets made from corn-
soybean meal based poultry layer diet (Skoch et al., 1981) w .
Temperature Temperature Temperature of Moisture of Moisture of Moisture Pellet
of feed before of feed after pellets after the feed before feed after of pellets durability
conditioning conditioning pellet-mill die conditioning conditioning after (%)|c|
(C) (C) ("C) (% w.b.) (% w.b.) cooling
(% w.b.)

27 27 68 10.4 10.8 9.4 79.1


27 65 75 10.7 13.1 10.8 93.5
27 80 87 11.3 13.8 11.2 96.5

[al x
Hammer mill screen size used for grinding the feed = 3.2 mm; Die dimension = 4.8-mm diameter
50.8-mm length; Pellet production rate = 655 kg/h.
!bl
Pellets were cooled in a vertical cooler for 15 min.
[c)
Durability of pellets was measured after cooling the pellets within 5C of ambient temperature.

126
Table 3.7. Effect of binders (additives) on the pellet quality (Tabil et al., 1997).
Feed type Binder Inclusion level Pellet Pellet
(% by weight) durability hardness
|a|
(%) (N)
Alfalfa (Trial 1- Control (without 0.0 65.1 507
steam conditioning binder)
to >92C)
Collagen protein 0.2 83.6 646
Hydrated lime 1.9 88.9 783
Lignosulfonate 1.25 85.8 654
Bentonite 5.0 88.8 701
Pea starch 0.74 86.4 573
Alfalfa (Trial 2- Control (without 0.0 54.6 425
steam conditioning binder)
to >92C)
Hydrated lime 0.5 69.7 471
Hydrated lime 1.0 70.2 502
Pea starch (rep. 1) 1.0 62.9 439
Pea starch (rep. 2) 1.0 62.7 467
0.5% hydrated lime - 54.0 472
and 0.5% pea starch

w
Pellet durability was measured using the Dural tester as described by Sokhansanj et al. (1991). The
diameter of the pellets was 6.1 mm. The Dural tester was operated at 1600 rpm for 30 s. After agitation
in the Dural tester, pellets were sieved using a 5.95-mm round hole sieve to remove the fines. The pellet
durability was estimated as the percentage of original pellet mass retained on the 5.95 mm sieve.

127
Table 3.8. Effect of mixing of different feed materials on the pellet/briquette
durability.
Feed Pellet/briquette durability (%)
Data from Chen et al. (1989) for aspen
furnish
33.3% bark: 44.4% wood: 22.2% coal 98.4
33.3% bark: 22.2% wood: 44.4% coal 96.6
33.3% bark: 0% wood: 66.7% coal 91.9
50.0% bark: 33.3% wood: 16.7% coal 98.3
50.0% bark: 16.7% wood: 33.3% coal 97.4
50.0% bark: 0% wood: 50.0% coal 95.2
Data from Wamukonya and Jenkins (1995)
Wheat straw 46.5
Sawdust 82.6
50% wheat straw: 50% sawdust 51.5
25% wheat straw: 75% sawdust 67.6

128
Table 3.9. Quality of water hyacinth briquettes or pellets made from different
densification equipment (Koser et al., 1982).
Densification Briquette shape, Briquette Briquette or Compressive
equipment size density (kg m~3) pellet durability strength of
(%) |f| briquettes or
pellets (MPa)
Laboratory Cylinder, 23-mm 1200 >99.0 Not measured
press w diameter
Pellet mill [bJ Cylinder, 23-mm 1440 99.4 Not measured
diameter
Cylinder, 32-mm 1430 99.0 9.9
diameter
Roll press |c) Pillow, volume of 1320 99,5 Not measured
5 cm
Pillow, volume of 1290 99.3 Not measured
20 cm3
Ram-extrusion Cylinder, 50-mm 1320 99.6 22.6
press [d| diameter

laI
A hydraulic press and a cylindrical closed-end die and plunger were used. Moisture content = 12%
(w.b.); Preheating temperature = 80C; Pressure = 147 MPa; Compression speed = 20 mm/s.
[b]
Flat-die pellet mill was used for the study. Two die sizes were tested: 23-mm diameter x 45-mm
length, and 32-mm diameter x 75-mm length. No preheating was used. Pellet temperature after the
pellet mill was about 80C due to frictional heat developed in the pellet mill. The specific energy input
to the pellet mill was about 180 MJ/t.
[cl
Roll diameter = 1000 mm; Roll width = 300 mm; and roll speed = 1 rpm. No preheating was used.
[dl
Ram-extrusion press was operated at a stroke length of 110 mm and a speed of 200 strokes per min.
No preheating was used. Briquette temperature after exiting the press was about 80C due to frictional
heat developed in the press. The specific energy input to the press was about 180 MJ/t.
[e]
Bulk density of powdered water hyacinth was 100 kg m"3. In this study, for all the densification
equipment, a compaction ratio of 1:14 was the target (i.e., briquette density of 1400 kg m"3) at a
moisture content of about 12% (w.b.).
[f|
Briquettes or pellets were tumbled at 25 rpm for 1 min. The mass percentage of original briquettes or
pellets that remained on a 10-mm sieve after tumbling was taken as the durability of briquettes.

129
Table 3.10. Effect of pressure on the physical quality of densified products.
Feed material Forming Durability of Compressive Reference
pressure (MPa) densified strength of
products (%) densified
products (MPa)
Oak sawdust 34 Not measured 25 Li and Liu (2000)
(a typical
hardwood; MC
= 8.9% w.b.)
69 93.3 28
103 94.0 45
138 98.3 49
Pine sawdust 34 Not measured 25 Li and Liu (2000)
(a typical
softwood; MC
= 8.4% w.b.)
69 71.2 35
103 91.7 44
108 93.2 45
Corn stover 30 0.0 Not measured Kaliyan and
(MC= 10.0% Morey (2005)
w.b.)
60 7.1 Not measured
100 49.5 Not measured
130 51.7 Not measured
150 61.6 Not measured

Table 3.11. Effect of pellet mill die size on starch gelatinization (measured after
pelleting) and pellet durability for a poultry layer ration (Heffner and
Pfost, 1973).
Die size Trial 1 Trial 2
(diameter Gelatinization (%) Pellet Gelatinization (%) Pellet
x length durability (%) durability (%)
mm)
4.8 x 44.5 24.1 93.6 27.4 88.6
6.4 x 57.2 21.1 92.4 24.1 88.2
9.5 x 76.2 20.6 87.2 21.3 87.2

130
Table 3.12. Effect of expander conditioning on the pellet quality (van der Poel et al.,
1997).

Pelleting condition Steam conditioning Kahl pellet hardness Holmen pellet


temperature (C) (N) durability (%)
Without steam 24.0 39.2 2.8
conditioning
With steam 86.0 47.1 41.5
conditioning
With steam 100.0 77.5 87.4
conditioning and
expander
conditioning

Table 3.13. Minimum feed pellet quality recommended for feed mills (Major, 1984)
[a]

Pellet Pellet Pellet durability based on tumbling Pellet durability based on


diameter hardness can method (ASABE Standards, Holmen tester (1 min test)
(mm) (N) 2003) (%) (%)
6.0 to 8.6 63?7 96\6 95^0
4.0 to 5.0 39.2 96.0 85.0
3.0 to 3.2 Not 96.0 60.0
applicable
[b]

|a|
These are only guidelines, applicable to the so-called "average" mill delivering to the average
"farmer" (both of which do not exist). Each mill must set its own minimum standards based on (i) the
amount of handling the pellets get between manufacturing and feeding, and (ii) the demands of the local
market. The Holmen pellet tester gives lower results for smaller pellets. It is a pneumatic method and
the small pellets move faster in the air-stream. This mirrors what happens when pellets are delivered in
bulk.
[bl
Small pellets (3.0 to 3.2 mm diameter) can be properly tested for durability, but not hardness.

131
Table 3.14. Generalized optimum densification conditions 'a\

Factor Range of optimum Strategy to produce high quality


conditions/values densified products [bl
Feed constituents
Starch Conditions to aid gelatinization of Starch acts as a binder. High
starch such as high temperature and amount of starch would
high moisture (i.e., low pressure positively affect the quality of
steam such as a steam pressure of the densified products.
103 kPag is used to provide a feed Adding pre-gelatinized starch is
temperature of 85C) better than raw-starch.
Protein Conditions to aid denaturation of Protein acts as a binder. High
protein such as high temperature amount of protein would
and limited moisture (i.e., high positively affect the quality of
pressure steam such as a steam the densified products.
pressure of 448 kPag is used to Adding raw protein is better
provide a feed temperature of 85C) than denatured protein.
Fiber 4 to 27% of crude fiber content High percentage of water-
(limited moisture addition during insoluble fiber is deleterious to
steam conditioning) the quality of the densified
products.
Reduce the resilience
characteristics of the fiber by
increasing moisture content up
to 25% (w.b.) (for densification
under room temperature) or by
adding chemicals such as CaO.
Additives such as
lignosulfonate or lime can help
increase the quality of the
densified products.
Providing high maceration or
high shear to the high fiber feed
by double pelleting or expander
conditioning may improve the
densification behavior of the
feed.
Fat / oil 1.5 to 6.5% fat added to the feed High amount of fat would
before densification negatively affect the quality of
the densified products.
For high-fat feeds, use thick
dies in pellet mill or double
pelleting.
Adding chemical binders would
help pellet high-fat feeds.
Lignin plus < 34% Lignin acts as a binder. High
extractives amount of lignin and extractives
content up to 34% would positively
affect the quality of the densified
products.

132
Table 3.14. Continued.
Factor Range of optimum Strategy to produce high quality
conditions/values densified products [bl
Feed moisture content 8 to 20% (w.b.) Moisture contents from 8 to
20% (w.b.) may help the
densification process under
room temperature. Generally,
at moisture contents more than
20% (w.b.), densification may
not be possible.
Moisture contents more than
16% (w.b.) might choke the
pellet mill. However, larger
die diameter (e.g., 7.8-mm
diameter) can be used for
pelleting high moisture feed
(>10%w.b.).
Roll press can handle high
moistures (up to 20% w.b.).
Feed particle size Geometric mean diameter of Smaller particle size would
particles: 0.5 to 1.0 mm help the densification process.
Hammer mill screen size: 2.4 to Also, wider particle size
3.2 mm. distribution might aid the
densification.
Hammer mill screen sizes of
up to 9.53 mm can be tried.
Steam conditioning Steam pressure: 103 to 448 Using a differential scanning
/ preheating of feed kPag(15 to65psig). calorimeter (DSC) apparatus,
Steam quality: 70 to 80%, or temperatures related to the
higher. starch gelatinization, protein
Conditioning temperature: 65 denaturation, and glass
to 100C. transition can be obtained.
Conditioning time: 15 to 250 s. One of these thermal transition
temperatures can be used as
starting temperature for the
conditioning process.
Temperatures higher than
100C and up to 300C can be
used if desired.

133
Table 3.14. Continued.
Factor Range of optimum Strategy to produce high quality
conditions/values densified products [b|
Binders (additives) Chemical binders: 0.5 to 5% (by Chemical binders such as
weight). lignosulfonate and lime would
Biological binders: 20% (by positively affect the quality of
weight) and more. the densified products.
(Cost and emission Biological binders such as
characteristics of the binders molasses, starch, waste paper,
might limit the amount of and sawdust can help improve
binders added.) the quality of the densified
products.
Densification
equipment variables
Pressure 100 to 150 MPa, or more High pressure helps the
densification. The densification
equipment and mechanical
properties of the feed material
may have limitation on the
amount of pressure applied to
the feed.
Pellet mill and roll press can
provide pressures in the order of
100-150 MPa.
Pellet mill Die dimension: length to Each feed material requires
diameter ratio of 8 to 10. (die specific optimum pellet mill
diameter = 4.8 to 7.8 mm) operating conditions. The
Die speed: 150 to 250 rpm. (low optimum conditions can be
speed for low bulk density established by conducting
materials) preliminary tests.
Gap between the roller and die: Double pelleting, or
2 to 4 mm conditioning using an
expander/ripener can help
improve the quality of the
pelleted feeds.
Roll press Optimum conditions could not Speed of screw feeder, roll speed,
be established from the roll force, roll torque, and gap
information found in the between the rollers should be
literature. The effects of several optimized by conducting
roll press variables are discussed preliminary tests.
in the text.

134
Table 3.14. Continued.
Factor Range of optimum Strategy to produce high quality
conditions/values densified products tbl
Post-production
conditions
Time of measurement Immediately after production Measuring green and cured strength
(i.e., green strength) or after one and durability of the densified
week of curing (i.e., cured products can help understand the
strength). destructive (e.g., expansion) or
binding (e.g., solid bridges)
mechanisms. The time of
measurement should be reported
along with the strength and
durability values.
Cooling/drying Final temperature of the Forced air cooling systems must be
products should be within 5C of operated at predetermined optimum
ambient. cooling conditions to avoid stress
Final moisture content of the cracks on the densified products.
products should be less than
13%(w.b.).
Storage conditions Relative humidity: 60 to 70%. Exposure of densified products
Temperature: room temperature to rain or high humidity
(25C) and below. environments should be
avoided.
During storage, increase in
moisture content to more than
13% (w.b.) would be
deleterious to the quality of the
densified products.

w
From the optimum conditions given in table 3.14, starting conditions for pelleting, briquetting, or
cubing of biomass materials or animal feeds can be chosen.
[b] High quality of densified products means high strength and high durability of the densified products.

135
Table 3.15. Tested strategies to improve pellet durability (Winowiski, 2006).
Strategy Pellet durability (%)
Corn-soybean meal based feed (temperature of 70.0 (control)
feed mash, steam conditioned = 80C; pellet
production rate = 26 t/h)
Increased temperature by 5C 75.1
C
Increased temperature by 10 C 79.4
Reduced fat by 0.5% (by weight) 75.0
Added 1.25% (by weight) lignosulfonate 82.5
(additive)
Added 10% (by weight) wheat (additive) 75.4
Added 23% (by weight) wheat (additive) 82.5
Decreased production rate (t/h) by 20% 71.3

136
Chapter 4
DENSIFICATION CHARACTERISTICS OF CORN STOVER AND SWITCHGRASS

- LABORATORY SCALE STUDY

4.1 Abstract

Densification characteristics of corn stover and switchgrass were studied under


uniaxial compression. The effects of pressure (100 to 150 MPa), moisture content (10
to 20%), particle size (0.56 to 0.80 mm), and preheating temperature (75 to 150C) on
the densification characteristics of corn stover and switchgrass were studied. Under
room temperature (about 25C), corn stover briquettes (about 19.2 mm diameter) with
relaxed densities of 825 to 1013 kg m"3, and switchgrass briquettes (about 19.4 mm
diameter) with relaxed densities of 417 to 825 kg m"3 were produced. Densification at
25C resulted in briquette durability of 50 to 96% for corn stover, and zero percent for
switchgrass. Preheating corn stover and switchgrass to 75, 100 or 150C increased the
density further, and resulted in briquettes with maximum durability of 97% for corn
stover, and 67% for switchgrass. Specific energy required for densification including
the preheating energy for both biomass materials was about 0.12 to 1.72% of the
energy content of the biomass material. This study suggests that more research is
required to improve the durability of switchgrass briquettes.

4.2 Introduction

Using biomass feedstocks to produce fuels, chemicals, heat and power has
recently received greater attention in the U.S. About 80% of the agricultural residues
in the U.S. are corn (Zea mays L.) stover (Walsh et al., 2005). Corn stover is the
aboveground portion of the corn plant, including stalks, leaves, cobs and husks, but
excluding the corn kernels. Annually about 230 million dry tons of corn stover are
available in the U.S. (Walsh et al., 2005), and a major portion of this amount of corn
stover can be used for producing ethanol, chemicals, biomaterials, process heat, and

137
electricity. Switchgrass (Panicum virgatum L.), a native perennial warm-season grass,
has been identified as a dedicated energy crop for producing bio-fuels and renewable
energy (McLaughlin and Kszos, 2005; ORNL, 2005). Use of biomass such as corn
stover and switchgrass in place of fossil fuels could reduce the nation's dependence on
foreign oil, reduce greenhouse gas emissions, increase farm income, and create job
opportunities in the rural areas (DOE, 2005).
Corn stover and switchgrass can be collected only during a limited harvest
season and, therefore, they should be stored to feed the processing facilities year-
round. Corn stover and switchgrass are often stored in baled forms, which involve a
lot of handling, storage and transportation cost because of low bulk density. One of the
solutions to reducing handling, storage, and transportation costs is densification of the
biomass material into pellets, briquettes, or cubes. Densified biomass will be in more
consistent form, and easier to handle and feed into the processing equipment.
Densification (pelleting, briquetting, cubing, or compaction) of particulate matter is
achieved by forcing the particles together by applying a mechanical force to create
inter-particle bonding, which makes well-defined shapes and sizes such as pellets,
briquettes, and cubes. Densification characteristics of several biomass residues such as
alfalfa, wheat straw, barley straw, rice straw, rice husk, and sawdust have been studied
in the past (Mani et al., 2003). However, only limited studies have been conducted to
study the densification behaviors of corn stover and switchgrass because of the recent
interest on these biomass materials (Jannasch et al., 2004; Mani et al., 2004a; Colley et
al., 2005). Therefore, extensive research is required to study the densification
characteristics of corn stover and switchgrass in order to produce highly durable
pellets/briquettes from corn stover and switchgrass.

4.3 Objectives

The objective of this research was to study the effects of pressure, moisture
content, particle size, and preheating temperature on the densification characteristics
of corn stover and switchgrass.

138
4.4 Materials and Methods

4.4-1 Bio mass Samples

Corn stover (cv. Pioneer 38W21) used for the study was harvested during
November 2004 and stored outdoors in round bales at the West Central Research and
Outreach Center (WCROC), University of Minnesota, Morris, MN. About two weeks
after harvest, corn stover was collected from these round bales and transported to the
lab and stored in a -10C freezer until used. To obtain switchgrass samples, whole
switchgrass plant (about 50-mm above the ground) was manually harvested during
August 2005 from a field at WCROC.
A hammer mill (J.B. Sedberry Inc., Franklin, TN), which operated at 3600
rpm, and was powered by a 5.6-kW motor, was used to grind the corn stover and
switchgrass. Corn stover had an initial moisture content of about 42% (wet basis).
Therefore, corn stover was dried at room temperature for 48 h before grinding. The
initial moisture content of the switchgrass was about 48% (wet basis). Before
grinding, the switchgrass was sun-dried to about 10% moisture content and then it was
chopped into lengths of about 100 to 150-mm. Two hammer mill screens (with 3.0-
and 4.6-mm round holes) were used to grind the corn stover and switchgrass to obtain
corn stover and switchgrass grinds with two different particle sizes. After grinding,
both grinds had about 10% (wet basis) moisture content. To increase the moisture
content of the grinds to 15% (wet basis), a predetermined amount of distilled water
was added to the grinds, thoroughly mixed and stored in zip-lock plastic bags at 5C
for 48 h for tempering.
Moisture content of the grinds was measured using the procedure given in
ASAE Standard S358.2 (ASAE Standards, 2003a). The moisture content values
reported in this paper are on wet basis. Bulk density of grind was calculated from the
mass of grind that occupied in a 250-mL glass container. Particle size distribution of
the grinds was determined based on ASAE Standard S319.3 (ASAE Standards,
2003b).

139
4.4-2 Briquette Compression and Ejection Apparatuses

An INSTRON model 4206 universal testing machine (Instron Corporation,


Canton, MA) was used for applying the mechanical pressure required for briquetting
the biomass grind. Figures 4.1a and 4.1b show the piston-cylinder (die) apparatus used
for compressing the biomass grind. The inner diameter and height of the steel cylinder
were 18.8 mm and 300 mm, respectively. The piston was made of brass and was 50-
mm longer than the cylinder to help eject briquettes from the cylinder.
A separate apparatus was made to allow ejection of briquettes from the
cylinder after compression (figs. 4.2a and 4.2b). The ejection apparatus was a steel
cylinder of 60-mm ID and 100-mm high. Circular plates (20-mm width) were welded
at the top and bottom edges of the cylinder. During briquette ejection, the top plate
with pin connections would support the cylinder (die), and the bottom plate would be
bolted on the base plate of the INSTRON.
Figures 4.3a - 4.3d show the modifications to be made to the densification
cylinder to allow preheating of the biomass grind before compression.

4.4-3 Briquetting Procedure

All briquetting experiments were conducted under uniaxial compression using


the INSTRON. The INSTRON was controlled by a computer loaded with TestWorks
4.0 software (MTS Systems Corp., Eden Prairie, MN). The top of the piston was
connected to the crosshead of the INSTRON by a pin connection. The bottom of the
cylinder was mounted on the base plate of the INSTRON. About 5.0-g of biomass
grind was added to the cylinder with a funnel. Using a steel rod, the grind was stirred
to help the flow of grind from the funnel. The TestWorks 4.0 software was
programmed to actuate the crosshead to compress the grind to a set maximum pressure
at a constant speed of 25.4 mm min"1 (1.0 in. min"1). The crosshead speed was kept
constant for both loading and unloading processes. The set maximum pressure was
achieved within 0.001 to 0.01%. During compression, the piston goes down during
loading and compresses the grind, and goes up during unloading and the briquette
140
relaxes. After completion of compression, the piston is taken out of the cylinder. A
typical compression curve is shown in figure 4.4. During compression, TestWorks 4.0
software recorded the time, force, and distance traveled by the crosshead at a data
collection rate of 40 Hz. To calculate the specific energy consumption for compression
(i.e., compression energy), the area under the force-displacement curve was estimated
using the trapezoidal rule (Cheney and Kincaid, 1985).

After compaction of biomass grind into briquettes, the cylinder was removed
from the base of the INSTRON. The ejection apparatus was attached to the base of the
INSTRON and the cylinder (die) was placed on top of the ejection apparatus (figs.
4.2a and 4.2b). Then, a separate TestWorks 4.0 program was used to operate the
crosshead-piston to push the briquette out of the cylinder. A constant crosshead speed
of 25.4 mm min"1 was used for briquette ejection. Figure 4.4 shows a typical briquette
ejection cycle. Also, time, force and distance traveled by the crosshead were recorded
during the briquette ejection cycle to estimate the specific energy consumption for the
briquette ejection (i.e., ejection energy). In this study, the cylinder was not lubricated
during briquetting but periodically cleaned using a vacuum cleaner.

4.4-4 Preheating of Biomass Grinds

To preheat the biomass grinds, the bottom opening of the briquetting cylinder
was closed tightly with a steel base that had an O-ring seal, and about 5.0 g of sample
was filled into the cylinder. The top opening of the cylinder was closed with a
specially designed cap. On the cap, a pressure relief valve (RL3 Series Relief Valve;
Swagelok Company, Solon, OH) and a 450-mm long K-type thermocouple were
installed. The pressure relief valve was set constantly at 506.6 kPa for all preheating
temperatures. This allowed preheating the grind to 150C because water boils at 475.8
kPa pressure when the temperature is 150C. This procedure also reduced the amount
of moisture loss from the grind during preheating. On the outside of the cylinder wall
and bottom of steel base, heating tapes (volt = 120, amps = 2.58, watts = 310, and
phase = 1; BH Thermal Corporation, Columbus, OH) were covered to heat the

141
biomass grind along with the cylinder. About 50-mm thick fiberglass insulation was
used to cover the heating tapes to avoid heat loss. Temperature of the grind was
measured inside of the cylinder at the center and about 50-mm above the bottom of the
cylinder. Figures 4.3a - 4.3d depict the preheating arrangement including the above
mentioned accessories.
A hand-held thermocouple-temperature sensor (DIGI-SENSE, Model No.
91100-40; Cole-Parmer Instruments Co., Vernon Hills, IL) was used to monitor the
temperature of the grind. When the biomass grind reached the predetermined
temperature, the power to the heating tapes was shut-off, heating tape and insulation at
the bottom of the cylinder were removed, the cylinder was mounted on the INSTRON,
pressure inside the cylinder was released through the pressure relief valve, the cap was
removed from the cylinder, and the grind was compressed immediately. Since the
cylinder was air- and vapor-tight during preheating of the biomass grind, temperature
of the grind inside the cylinder was considered to be uniform everywhere. The specific
energy required for preheating biomass grind (i.e., preheating energy) was
theoretically calculated by assuming a specific heat value of 2.0 kJ kg"1 K"1 for both
corn stover and switchgrass as:

Preheating energy (MJ/t) = Specific heat of biomass x [preheating temperature - room


temperature].

In some of the briquetting cases at 100C and all of the cases at 150C, it was
observed that moisture was deposited on the grind and at the bottom of the briquetting
cylinder. Movement of water from the grind either as vapor or liquid to the bottom of
the sample was probably a result of the inability to perfectly insulate the cylinder
during preheating. Also, the bottom of the cylinder was difficult to insulate during
compression resulting in lower temperature at the bottom of the cylinder, which
caused moisture to migrate and accumulate on the bottom of the sample. The amount
of water that was deposited was more at 150C than at 100C. Due to the liquid water
at the bottom of the briquetting cylinder, about 0.5 to 1.0-g of grind at the bottom of

142
the briquette did not get compressed well, and sometimes stayed as loose grind. These
loose particles were scraped off using a sharp knife, and the remaining part of the
briquette was used for the study. It was observed that these loose particles lost the
moisture quickly after removal from the briquettes. This may be due to their small
particle size. However, these loose particles were stored immediately in zip-lock
plastic bags, and their moisture contents were measured after one week of storage at
room temperature. At the time of moisture measurement, it was observed that these
loose particles were somewhat drier than the condition observed immediately after
removal from the briquettes.

4.4-5 Differential Scanning Calorimeter Experiments

To select the preheating temperatures, tests were conducted in a Differential


Scanning Calorimeter (DSC) (Pyris-1, PerkinElmer Life and Analytical Sciences,
Shelton, CT) to determine the glass transition temperatures of corn stover and
switchgrass. The DSC was used to obtain thermograms in the temperature range of 20
to 150C. The DSC was calibrated with indium and zinc standards before the tests. All
DSC tests were conducted using aluminum sample pans under nitrogen environment.
An empty aluminum pan was used as a reference. About 5 to 20 mg of biomass
samples were heated from 20 to 150C at a heating rate of 10C/min, and the
thermogram showing heat flow (W/g) versus temperature (C) was recorded. The DSC
thermograms for corn stover and switchgrass were obtained at about 10, 15, and 20%
moisture contents. Three replications were made at each moisture content. For
comparison, DSC thermogram for lignin (Lignin-hydrolytic; Aldrich, www.sigma-
aldrich.com) at about 10% moisture content was also obtained. The DSC thermograms
were analyzed to determine the temperatures of onset, mid-point, and end of the glass
transition phase (fig. 4.5). In this study, glass transition temperature (Tg) was
measured as the temperature at the midpoint of the change in the slope of the DSC
thermogram change [i.e., the midpoint of the change in heat capacity (Cp)] (Zhong and
Sun, 2005). The overall average of glass transition temperature (i.e., 75C) and the

143
temperature at the end of the glass transition phase (i.e., 100C) for corn stover and
switchgrass at the three measured moisture contents, and the maximum temperature
used for the DSC tests (i.e., 150C) were selected as preheating temperatures.

4.4-6 Briquetting Experiments

Using corn stover and switchgrass grinds obtained from the hammer mill
screen that had openings of 3.0-mm and 4.6-mm, briquettes were made for the
following combination of conditions: (i) maximum pressures of 100 and 150 MPa, (ii)
grind moisture contents of 10, 15, and 20% (w.b.), and (iii) preheating temperatures of
25 (i.e., room temperature), 75, 100, and 150C. To determine the densification
conditions that would result in high density and durable briquettes, densification
experiments were conducted progressively from one set of conditions to the next
conditions. The boundary limits on the pressure, moisture content, or preheating
temperature were reached within a few sets of experiments. Therefore, a complete
factorial experimental design of experiments was not necessary, especially considering
the time required to conduct the experiments.
Additional experiments were conducted using corn stover and switchgrass
grinds obtained from a hammer mill screen size of 2.4 mm to study the effect of
preheating to 40, 60, and 80C at 15 and 20% (w.b.) moisture contents at 150 MPa.
Also, one test was conducted at a very high preheating temperature of 200C to study
the densification behaviors of switchgrass. At each briquetting condition, ten
briquettes were made. In this study, no binders/lubricants (i.e., additives) were used
for the briquetting.

4.4-7Briquette Properties

Immediately after ejection from the die, unit density of the briquettes (i.e.,
density of individual briquettes) was measured. Then, the briquettes were transferred
to zip-lock plastic bags and stored for one week at room temperature. Durability,

144
percentage expansion in axial and radial directions and volume, and moisture content
of briquettes were measured after one week of storage.
Durability of briquettes was measured according to ASAE Standard S269.4
(ASAE Standards, 2003c). Durability was calculated as the percentage of briquette
mass retained on 16.0-mm screen after tumbling in a durability tester (Continental-
Agra Equipment, Inc., Newton, KS) at 50 rpm for 10 min (fig. 4.6). Only two
replications were done for the durability measurement because of lack of samples. For
each replication, five briquettes were used. Unit density of briquettes was calculated
from the mass, diameter and height of the briquettes (ASAE Standards, 2003c). Axial,
radial, and volume expansions of briquettes were calculated as the percentage increase
in height, diameter, and volume, respectively. Moisture content of the briquettes was
measured based on ASAE Standard S358.2 (ASAE Standards, 2003a).

4.4-8 Statistical Analysis

The means of unit density and durability of the corn stover and switchgrass
briquettes measured after one week of storage at room temperature were statistically
analyzed to study the significant effects of densification variables. All statistical
analyses were performed using the software SPSS 16.0 for Windows (SPSS Inc.,
Chicago, IL) at 5% significance level. Tests on the analysis of variance (ANOVA)
using a univariate general linear model (a = 0.05) were performed on the density or
durability for both corn stover and switchgrass to understand the main and interaction
effects of densification variables (tables 4.A1-4.A4 in section 4.8 Appendix). To
conduct the above analyses, all data on briquette density or durability measured in this
study (i.e., data for the densification conditions listed in tables 4.3a, 4.3b, and 4.6a)
were compiled.
To separate the effects of different levels of treatment factors (temperature or
moisture content) on the means of briquette density or durability, Tukey's test was
performed after the significance confirmation by the one-way ANOVA at 5% level

145
(table 4.A5 in section 4.8 Appendix). This test was conducted for the data given in
tables 4.3a, 4.3b, and 4.6a.
To investigate the difference in the effect of moisture conditioning (15 to 20%
w.b.) versus preheating (75 to 100C) for corn stover and switchgrass, "independent
samples T-test" was conducted (tables 4.A6 and 4.A7 in section 4.8 Appendix). To
obtain data for the preheating treatment, data on density or durability were pooled for
preheating temperatures of 75 and 100C [other conditions were pressure of 150 MPa,
moisture content of 10% (w.b.), and particle sizes of 3.0 and 4.6 mm hammer mill
screen sizes]. For the moisture conditioning treatment, data for the moisture contents
of 15 and 20% (w.b.) (other conditions were pressure of 150 MPa, particle size of 2.4,
3.0 and 4.6 mm hammer mill screen sizes, and preheating temperature of 25C) were
pooled. A control set of data was obtained from the densification conditions involving
moisture content of 10% (w.b), preheating temperature of 25C, pressure of 150 MPa,
and particle sizes of 3.0 and 4.6 mm hammer mill screen sizes.
To study the significant differences between particle sizes obtained from 3.0-
mm and 4.6-mm hammer mill screen sizes, data (pooled set) obtained for the moisture
contents of 15 and 20% (w.b.) (other conditions were pressure of 150 MPa and
preheating temperature of 25C) at these two hammer mill screen sizes were analyzed
using the "independent samples T-test" (tables 4.A8 in section 4.8 Appendix).
Using a one-way ANOVA followed by T-test, the significance of the particle
size or moisture content when preheating switchgrass to 100C was investigated for
the pooled data from the densification conditions of pressure of 150 MPa, particle
sizes of 2.4, 3.0, and 4.6 mm hammer mill screen sizes, moisture contents of 8.3, 10,
and 15% (w.b.), and preheating temperature of 100C (tables 4.A9 and 4.A10 in
section 4.8 Appendix).

146
4.5 Results and Discussion

4.5-1 Biomass Grinds

Corn stover grind used for briquetting had initial bulk density of about 100 to
130 kg m"3, and geometric mean particle diameter of 0.66 to 0.80 mm (fig. 4.7a and
table 4.1). Switchgrass grind used for briquetting had initial bulk density of about 180
to 200 kg m"3, and geometric mean particle diameter of 0.56 to 0.64 mm (fig. 4.7b and
table 4.1). On average, the length of switchgrass particles was longer than that of corn
stover particles for both hammer mill screen sizes of 3.0 and 4.6 mm. The shape of the
corn stover particles was almost like spheres whereas the shape of the switchgrass
particles was like needles/flakes.

4.5-2 Glass Transition Temperature

Glass transition temperature is defined as the temperature at which the material


softens due to the onset of long-range coordinated molecular motion (Roos, 1995).
During glass transition, amorphous materials change their state from a hard glassy to a
soft rubbery sate. Within the glass transition region, many macroscopic properties of
the materials such as viscosity and mechanical properties (e.g., modulus of elasticity)
change their values dramatically (Irvine, 1984). Using DSC, the glass transition region
is detectable by a step change in the heat capacity of the material that is being heated
or cooled under a constant rate in the DSC equipment. Figure 4.5 shows the typical
DSC thermograms for corn stover, switchgrass, and lignin. In table 4.2, the glass
transition temperatures of corn stover, switchgrass, and lignin are provided.
The glass transition occurred from 50 to 115C for both corn stover and
switchgrass (table 4.2). The average glass transition temperature for both biomass
species at the three measured moisture contents was 75C. Also, the average end point
of the glass transition temperature was 100C for both biomass materials. In general,
increasing moisture content from 10 to 20% decreased the glass transition temperature

147
(table 4.2). This is because water acts as a plasticizer which would help decrease the
glass transition temperature (Irvine, 1984; Zhong and Sun, 2005). The glass transition
temperatures of cellulose and hemicellulose (Back, 1987), and grains such as corn,
pea, and soybean (Sun and Leopold, 1994) were found to be below 0C at moisture
contents of more than 10% (w.b.). Therefore, it is expected that the glass transition in
corn stover and switchgrass at the nominal grind moisture contents of 15 and 20%
(w.b.) may occur below room temperature (i.e., < 25C). Because of possible drying
of the biomass grind during the DSC experiments, the measured glass transition
temperatures at the nominal grind moisture contents of 15 and 20% (w.b.) were above
50C. In addition, the DSC experiments were conducted for the temperature range of
20 to 150C, which could have prevented the determination of the correct glass
transition temperatures at high grind moistures if the glass transition temperatures of
high moisture grinds were below 20C. Therefore, in the future, DSC experiments
may be conducted for temperatures starting from below 0C to help measure the exact
glass transition temperatures of high moisture biomass grinds.

Irvine (1984) found that the glass transition temperature of lignin ranged from
60 to 90C. In this study, glass transition of lignin occurred from 62 to 101C. This
result is comparable to the values given by Irvine (1984). There were no previous
studies on glass transition temperatures of corn stover and switchgrass. However,
Zhong and Sun (2005) reported that the glass transition temperature of corn starch was
60C at 11.9% (w.b.) moisture content. This result is well within the value measured
for corn stover in this study. Since the glass transition temperature represents the
softening of the amorphous components in the materials such as lignin and hemi-
cellulose in biomass, the average temperatures corresponding to the mid-point and end
point of the glass transition region (i.e., 75 and 100C) for both corn stover and
switchgrass were used as preheating temperatures to study their densification
behaviors at these elevated temperatures. A preheating temperature of 150C was also
used to observe any improvements on the densification behaviors of the biomass
materials when the preheating temperature was beyond the glass transition
temperature region.
148
4.5-3 Densijication Process Variables

Figures 4.8-4.12 present the pictures of corn stover and switchgrass briquettes
before and after the durability test.
The ANOVA analyses (tables 4.A1-4.A4 in section 4.8 Appendix) showed
that the unit density of corn stover briquettes was significantly affected by pressure,
particle size, moisture content, and preheating temperature (P < 0.05). The interactions
x
of pressure particle size, and pressure x temperature on the corn stover briquette
density were not significant (P > 0.05). The interactions of pressure x moisture,
particle size x moisture, particle size x temperature, and moisture x temperature were
significant on the corn stover briquette density at the 5% level. Corn stover briquette
durability was significantly affected by pressure, moisture content, and preheating
temperature (P < 0.05); however, the briquette durability was not significantly affected
by particle size (P > 0.05). The interaction of pressure x particle size on the corn
stover briquette durability was not significant (P > 0.05). The interactions of pressure
x moisture, pressure x temperature, particle size x moisture, particle size x
x
temperature, and moisture temperature were significant on corn stover briquette
durability at the 5% level.
The unit density of switchgrass briquettes was significantly affected by
pressure, particle size, moisture content, and preheating temperature (P < 0.05). The
interactions of pressure x particle size, and pressure x temperature on the switchgrass
briquette density were not significant (P > 0.05). The interactions of pressure x
moisture, particle size x moisture, particle size x temperature, moisture x temperature,
and particle size x moisture x temperature were significant on switchgrass briquette
density at the 5% level. Switchgrass briquette durability was significantly affected by
pressure, moisture content, and preheating temperature (P < 0.05); however, briquette
durability was not significantly affected by particle size (P > 0.05). The interactions of
pressure x particle size, pressure x moisture, particle size x moisture, and particle size
x moisture x temperature on the switchgrass briquette durability were not significant

149
(P > 0.05). The interactions of pressure x temperature, particle size * temperature, and
moisture x temperature were significant on the switchgrass briquette durability at the
5% level.

4.5.3-1 Effect of pressure

From preliminary studies conducted at compression pressures of 30 to 150


MPa, it was found that compression pressure of at least 100 MPa was required to
produce briquettes (19.2 mm in diameter and 19.1 to 20.6 mm in length) with unit
densities of 829 to 905 kg m" and durability of more than 50% from corn stover with
geometric mean particle diameters of 0.66 to 0.80 mm at 10% (w.b.) moisture content
(Kaliyan and Morey, 2005). Therefore, pressures of 100 and 150 MPa were selected
for the laboratory densification experiments. Moreover, compaction pressures in
commercial scale roll press briquetting and pelleting machines are on the order of 100
to 150 MPa [Chapter 3; Kaliyan and Morey, 2006].
The average time of compression was about 340 and 215 s for corn stover and
switchgrass, respectively, when compressing the grind obtained from the hammer mill
screen size of 4.6 mm at about 10% moisture content at the maximum pressure of 100
or 150 MPa under room temperature (about 25C). Switchgrass grind required lower
compression time than corn stover grind because switchgrass grind had higher initial
bulk density than that of corn stover grind. The average time required for ejecting the
briquettes from the die was about 35 and 40 s for corn stover and switchgrass
briquettes, respectively. The relaxed diameter and height of both corn stover and
switchgrass briquettes ranged from 19.0 to 20.0 mm, and 14.0 to 36.1 mm,
respectively, depending on the pressure, moisture content, particle size, and
temperature.
Increasing the applied pressure from 100 to 150 MPa at a temperature of 25C
increased the unit density of corn stover and switchgrass briquettes (tables 4.3a and
4.3b). Mani et al. (2002 and 2004a) also observed this pressure effect on density while
pelleting wheat straw, barley straw, corn stover, and switchgrass. At 25C

150
temperature, increasing the pressure from 100 to 150 MPa increased the durability of
corn stover briquettes from 50 to 62% for geometric mean particle size of 0.80 mm at
10% moisture content. The durability of switchgrass briquettes was zero at both 100
and 150 MPa pressures when densified at 25C. For briquettes made at 25C, due to
relaxation of briquettes after ejection from the die, and surface particle loss and
moisture loss during storage, about 11 to 15% reduction in the unit density of corn
stover briquettes and about 24% reduction in the unit density of switchgrass briquettes
were observed after one week of storage at room temperature (tables 4.3a and 4.3b).
Tables 4.4a and 4.4b give the percentage expansion in axial and radial directions, and
volume of the briquettes after one week of storage at room temperature. In general,
switchgrass briquettes relaxed more than corn stover briquettes when briquettes were
made at 100 or 150 MPa and at 25C (tables 4.4a and 4.4b).
Increasing the pressure from 100 to 150 MPa at the temperature of 25C
increased the specific energy required for briquetting corn stover and switchgrass.
Switchgrass consumed slightly less energy than corn stover at either pressure
condition. For pressures of 100 to 150 MPa for the grind obtained from hammer mill
screen sizes of 3.0 and 4.6 mm, 10% moisture content, and 25C temperature, the
specific energy required for briquetting corn stover (24.8 to 45.4 MJ/t) was estimated
at about 0.12 to 0.22% of the energy content of the corn stover if the energy content of
the corn stover was taken as 20 MJ/kg (Pordesimo et al., 2005). Whereas for
briquetting switchgrass, the specific energy required (26.7 to 40.6 MJ/t) was estimated
at about 0.16 to 0.24% of the energy content of the switchgrass if the energy content
of the switchgrass was taken as 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004b). The
briquette ejection energy ranged from 0.1 to 1.3 MJ/t for all briquetting cases (tables
4.5a and 4.5b).

151
4.5.3-2 Effect of moisture content

For the corn stover grind obtained from the hammer mill screen size of 4.6
mm, relaxed briquette densities for 10 and 15% moisture contents were similar at an
applied pressure of 100 MPa and at the temperature of 25C (P > 0.05) (table 4.3a).
However, at 150 MPa, corn stover grind with 15% moisture content resulted in about
9% less density briquettes than that for 10% moisture content (P < 0.05). At 25C
temperature, increasing the moisture content from 10 to 15%o resulted in 30 to 40%
decrease in the relaxed densities of switchgrass briquettes (table 4.3b). Chancellor
(1962) reported that the higher the moisture content of alfalfa hay, the less the
compacted density. Mani et al. (2002) also reported that increasing the moisture
content decreased the density of pellets made from wheat straw, barley straw, corn
stover, and switchgrass.
At 25C temperature, increasing the moisture content from 10 to 15%
increased the durability of corn stover briquettes from 50 to 80% at 100 MPa pressure,
and from 62 to 84% at 150 MPa pressure. Further increase in moisture content from
15 to 20% (w.b.) at 150 MPa increased the corn stover briquette durability to 96%.
Switchgrass briquettes made at 25C were not stable at 15 and 20% (w.b.) moisture
contents and resulted in zero percent durability. Mohsenin and Zaske (1976), Smith et
al. (1977), and Coates (2000) found that higher moisture content (15 to 20%) produced
more durable compacts for alfalfa, wheat straw, and cotton plant residues,
respectively. The axial expansion of briquettes was accelerated by an increase in
moisture content (tables 4.4a and 4.4b). Mohsenin and Zaske (1976) reported that the
expansions of pellets are due to the residual stress in the pellets after ejection from the
die, and due to water loss from the pellets during storage. Also, Mohsenin and Zaske
(1976) reported that lower moisture content of hay pellets resulted in lower residual
stress and thus less expansion. Faborode (1989) found that increasing moisture content
increased the axial expansions of compacts made from barley straw.
For conditions with or without preheating, increasing moisture content from 10
to 15% or 20% (w.b.) generally decreased the specific energy required for briquetting

152
(tables 4.5a and 4.5b). This may be due to the fact that at high moisture content the
wall friction is less. Also, due to high resistance created by the high moisture grind
during compression, the maximum forming pressure was reached in shorter time than
for low moisture grind. It was noted that increasing the moisture content from 10 to
15% reduced the compression time by about 20 and 10% for corn stover and
switchgrass densified at 25C, respectively.

4.5.3-3 Effect of particle size

Decreasing the geometric mean particle size of corn stover grind from 0.8 to
0.66 mm resulted in about 5 to 10% increase in relaxed density of briquettes made at
25C (table 4.3a). Mani et al. (2002 and 2004a) found similar trends when pelleting
corn stover grind obtained from the hammer mill screen sizes of 3.2, 1.6 and 0.8 mm.
They reported that corn stover pellets made from the grind obtained from 1.6 mm
hammer mill screen was 5 to 16% more dense than the pellets made from the grind
obtained from 3.2 mm hammer mill screen. For switchgrass, reducing geometric mean
particle size from 0.64 to 0.56 mm did not show much impact on the relaxed density
of briquettes made at 25C (table 4.3b). Mani et al. (2002) also showed that the density
of switchgrass pellets had no trends with geometric mean particle size in the range of
0.25 to 0.46 mm.
At the briquetting temperature of 25C, decreasing the particle size of corn
stover grind from 0.8 to 0.66 mm increased the durability of briquettes from 50 to 58%
at 100 MPa pressure, and from 62 to 75% at 150 MPa pressure at 10% moisture
content. This may be due to the fact that more particle surface area is available for
bonding at smaller particle size than at larger particle size. From table 4.3a, it can be
concluded that moisture content is a more important factor than particle size in
determining the durability of corn stover briquettes made at 25C, although high
moisture resulted in less dense briquettes. For corn stover, statistical analyses
revealed that at the moisture contents of 15 to 20% (w.b.), the particle size obtained
from the hammer mill screen size of 3.0 mm would result in significantly higher unit

153
briquette density (about 93.4 kg m"3 higher) than that for the particle size obtained
from the hammer mill screen size of 4.6 mm (P < 0.05). However, the durability of the
briquettes would be similar at these two particle sizes (P > 0.05), and the average
briquette durability for these two particle size was about 92% (table 4.A8 in section
4.8 Appendix).
Switchgrass briquettes made with the particle sizes of 0.56 and 0.64 mm
resulted in zero percent durability of briquettes at 100 and 150 MPa pressures, 10%
moisture content, and 25C temperature. At 25C, decreasing the particle size of corn
stover grind from 0.8 to 0.66 mm increased the specific energy consumption by 0.8 to
1.3 MJ/t (table 4.5a). Whereas for switchgrass briquetting, decreasing the particle size
from 0.64 to 0.56 mm decreased the specific energy consumption by 2.5 to 4.3 MJ/t
(table 4.5b).

4.5.3-4 Effect of preheating temperature

To study the effect of preheating temperatures in the range of 75 to 150C, all


briquetting experiments were conducted at 150 MPa because 150 MPa resulted in
higher density and durability of briquettes made under 25C. In general, preheating the
biomass grinds from 25 to 75, 100 or 150C resulted in briquette densities of more
than 1000 kg m"3. Relaxed briquette densities at 75C were about 16 to 24%
significantly higher than those at 25C for corn stover (P < 0.05) (table 4.3a). For
switchgrass, the relaxed densities at 75C were about 50 to 100% higher than those at
25C (P < 0.05) (table 4.3b).
Our goal was to preheat at elevated pressures to allow heating without
moisture loss. However, after preheating, pressure was released before the piston was
inserted in the densification cylinder. Some moisture was lost during the compression
process as the material was exposed to atmospheric conditions. The higher the
temperature the greater the moisture loss or the lower the final moisture content of the
briquettes (tables 4.4a and 4.4b). For the initial moisture contents of 10 and 15%,
about 2 to 6 percentage points, and 5 to 11 percentage points of moisture loss,

154
respectively, occurred for both biomass materials due to preheating. Also, due to
preheating of both biomass materials, at the small particle size about 0.5 to 1% higher
moisture loss occurred than at the large particle size. Unfortunately, we did not
measure the moisture content of the briquettes made under room temperature. It
appeared that the moisture loss from the briquettes made under room temperature was
very small compared to the preheating cases.
At nominal 10 and 15% initial moisture contents, preheating the corn stover
grind with geometric mean particle size of 0.8 mm from 25 to 75C, the durability
values increased by 35.7, and 13.1 percentage points, respectively (table 4.3a). At
about 10% moisture content, preheating the corn stover grind with geometric mean
particle size of 0.66 mm from 25 to 75C, the durability of briquettes increased by
21.6 percentage points. Durability of corn stover briquettes was about 97% for the
preheating temperature of 75 or 100C. At equal preheating temperatures, the low
moisture content (10%) and small particle size (geometric mean particle size of 0.66
mm) of corn stover resulted in numerically higher density than for the high moisture
content (15%) and large particle size (geometric mean particle size of 0.8 mm).
Increasing the preheating temperature from 25 to 150C resulted in briquette durability
of 92 to 94%. Also, if the moisture content of the corn stover is 15 to 20% (w.b.),
preheating from room temperature (i.e., 25C) to 40 or 60C is not necessary to
produce corn stover briquettes with durability of more than 90% (tables 4.6a and
4.6b).
Due to the higher moisture migration problem encountered during the
preheating process at 150C than at 100C, the durability values measured at 150C
were slightly lower than those at 100C. However, the difference in the durability
values between 100 and 150C for most of the cases was not statistically significant at
5% level (tables 4.3a and 4.3b). As previously mentioned, at 100 or 150C, it was
observed that moisture was deposited in thin layer of the grind at the bottom of the
briquetting cylinder, which corresponded to about 0.5 to 1.0-g of the grind. It appeared
that there were zones of low moisture grind (i.e., top part of the briquette) and high
moisture grind (i.e., bottom part of the briquette). The particles in top part of the

155
briquette compressed better than those in the bottom part of the briquette. The
presence of a high amount of free moisture in the bottom layer of the grind could have
hindered effective particle bonding during compression. This moisture migration
problem and its effect on the amount of particles that did not get compressed were
higher at 150C than at 100C, which could have increased the part of the briquette
that was less durable at 150C than at 100C. So, the net effect was less durability of
briquettes at 150C than that at 100C. Also, tables 4.4a and 4.4b present the moisture
content of the loose particles that were scraped off from the briquettes made at the
preheating temperatures of 100 and 150C. Clearly, the moisture contents of the loose
particles were higher than those of the briquettes. This confirms that during preheating
and compression, moisture migrated to the bottom of the sample.
Moisture conditioning to 15 to 20% (w.b.) resulted in significantly different
unit briquette density (58.7 kg m"3 lower density) and durability (22.9% points higher
durability) of corn stover briquettes than for the control tests at 10% (w.b.) moisture
content (P < 0.05). Preheating of corn stover to 75 to 100C produced significantly
higher unit briquette density (178.5 kg m"3 higher density) and durability (28.5%
points higher durability) of briquettes than for the control tests at 25C (P < 0.05).
Preheating resulted in corn stover briquettes with significantly higher density (237.3
kg m"3 higher density) and durability values (5.6% points higher durability) than for
moisture conditioning (P < 0.05) (table 4.A6 in section 4.8 Appendix).
The durability of switchgrass briquettes made without preheating (i.e., 25C)
was zero percent for the range of pressure, particle size, and moisture content tested in
this study. At nominal 10 and 15% initial moisture contents, preheating the
switchgrass grind with geometric mean particle size of 0.64 mm from 25 to 75C, the
durability values were increased by 54.7, and 46.5 percentage points, respectively
(table 4.3b). At nominal 10% initial moisture content, preheating the switchgrass grind
with geometric mean particle size of 0.56 mm from 25 to 75C, the durability was
increased by 62.9 percentage points. Preheating the switchgrass from 25 to 75 or
100C resulted in briquette durability of 63 to 67% at both moisture contents and
particle sizes tested. However, durability of switchgrass briquettes made at 100C was

156
about 4.4 to 16 percentage points higher than the briquettes made at 75C. Increasing
the preheating temperature from 75 to 150C increased the briquette durability by 2.8
to 14.6 percentage points. The maximum durability of switchgrass briquettes obtained
was 67.3%. At equal preheating temperatures, the low moisture content (10%) and
small particle size (geometric mean particle size of 0.56 mm) of switchgrass resulted
in numerically higher density and durability than for the high moisture content (15%)
and large particle size (geometric mean particle size of 0.64 mm).
Switchgrass moisture conditioning to 15 to 20% (w.b.) resulted in significantly
lower unit briquette density (72.1 kg m"3 lower density) of briquettes than for the
control tests at 10% (w.b.) moisture content (P < 0.05). Preheating to 75 to 100C
resulted in significantly higher unit briquette density (365.5 kg m"3 higher density) and
durability (62.1% points higher durability) of switchgrass briquettes than for the
control at 25C (P < 0.05). Preheating treatment resulted in switchgrass briquettes with
significantly higher density (437.6 kg m" higher density) and durability (62.2% points
higher durability) values compared to the moisture conditioning treatment (P < 0.05)
(table 4.A7 in section 4.8 Appendix).
Moisture contents of 8 and 10% (w.b.) resulted in similar switchgrass briquette
densities (P > 0.05), but significantly different briquette densities compared to 15%
(w.b.) moisture content with preheating to 100C (P < 0.05). However, when
preheating switchgrass to 100C, the nominal grind moisture contents of 8, 10, and
15% (w.b.) resulted in similar briquette durabilities (P > 0.05) (table 4.A9 in section
4.8 Appendix). Table 4.6a indicates that preheating of switchgrass grind with nominal
moisture content of 15 or 20% (w.b.) from room temperature (i.e., 25C) to 40, 60, or
80C resulted in briquette durabilities of 0.0 to 38.8% (table 4.6a). Furthermore, at a
preheating temperature of 80C, increasing moisture content of the grind beyond the
moisture content range of 8 to 12% (w.b.) significantly decreased (P < 0.05) the
briquette density and durability values for switchgrass (table 4.6a). Also, at low
moisture of 8.3% (w.b.), preheating from 80 to 100 or 200C resulted in similar
density and durability values for switchgrass briquettes (P > 0.05) (table 4.6a).
Therefore, for producing good quality switchgrass briquettes with preheating to 75 to

157
100C, the moisture content of the grind should be above 8% (w.b.) and below 15%
(w.b.). In addition, statistical analyses indicated that with preheating to 100C, the
particle sizes obtained from the hammer mill screen sizes of 2.4, 3.0, and 4.6 mm
resulted in similar switchgrass briquette densities and durabilities (P > 0.05) (tables
4.A10 in section 4.8 Appendix).
Tables 4.4a and 4.4b show that briquetting at elevated temperatures of 75 to
150C resulted in less volume expansion for corn stover and switchgrass briquettes
than briquetting at room temperature. This shows that elevated temperatures may have
caused more plastic deformation of the particles than room temperature. The specific
(compression plus ejection) energy consumption for briquetting corn stover and
switchgrass at 75 to 150C were similar to that at 25C. Including the preheating
energy required in the temperature range of 75 to 150C, the specific energy required
for corn stover was 0.6 to 1.5% of the energy content of the corn stover (table 4.5a).
For switchgrass, the specific energy required was 0.7 to 1.7% of the energy content of
the switchgrass (table 4.5b) for the preheating temperatures of 75 to 150C.
This study shows that preheating corn stover and switchgrass to 75 to 100C
has potential to activate the natural binding components in corn stover (protein = 8.7%
of dry matter [DM], fat = 1.33% DM, lignin = 3.12% DM, hemi-cellulose = 21.08%
DM; Mani et al., 2002) and switchgrass (protein = 1.59% DM, fat = 1.87%> DM, lignin
= 7.43%o DM, hemi-cellulose = 30.0% DM; Mani et al., 2002) because the temperature
range of 75 to 100C corresponds to the glass transition (i.e., softening) temperature of
both of these biomass materials. Lignin and hemi-cellulose were found to be
amorphous thermoplastic materials which would undergo plastic deformation at
temperatures in the range of their glass transition regions (Back and Salmen, 1982). In
addition, during compression at these elevated temperatures, the natural binding
components may come out of the particles and make solid bridges between particles.
After cooling, these solid bridges will be hardened. This would make the briquettes
strong and durable (Rumpf, 1962). This may explain why corn stover and switchgrass
briquettes made at elevated temperatures (i.e., 75 to 150C) had higher durability than

158
at room temperature (i.e., 25C). [A detailed study conducted on the binding
mechanisms of corn stover and switchgrass is given in Chapter 6.]
In a commercial scale pelleting of switchgrass, Jannasch et al. (2004) found
that switchgrass lacked natural binding properties compared to alfalfa and resulted in
poor durability pellets. Assuming that all natural binders in the switchgrass were
effectively activated/softened due to elevated temperatures used in this study, it can be
concluded that there is a need to further increase the durability of switchgrass
briquettes by some other means such as by adding external binders (additives).
[Chapter 5 extends the laboratory scale densification study to determine the cost
effective ways of improving the durability of switchgrass briquettes.]

4.6 Summary and Conclusions

The following summary and conclusions are drawn from this study.

For briquetting under room temperature (about 25C), corn stover briquettes with
relaxed densities of 830 to 1013 kg m"3 and durability of 50 to 96% can be made at
pressures of 100 to 150 MPa, geometric mean particle sizes of grind from 0.8 to
0.66 mm, and moisture contents of 10 to 20%. For briquetting at elevated
temperatures of 75 to 150C and at 150 MPa, corn stover briquettes with relaxed
densities of 985 to 1162 kg m"3 and durability of 92 to 97% can be produced.
Specific energy required for briquetting corn stover ranged from 0.12 to 0.22%,
and 0.63 to 1.48% of the energy content of the corn stover for briquetting at room
temperature, and at elevated temperatures of 75 to 150C, respectively.

For briquetting under room temperature (about 25C), switchgrass briquettes with
relaxed densities of 417 to 825 kg m"3 can be produced at pressures of 100 to 150
MPa, geometric mean particle sizes of grind from 0.56 and 0.64 mm, and moisture
contents of 10 to 20%. However, durability of switchgrass briquettes made under
room temperature was zero percent. For briquetting at elevated temperatures of 75

159
to 150C and at 150 MPa, switchgrass briquettes with relaxed densities of 834 to
1065 kg m'3 and durability of 55 to 67% can be produced. Specific energy required
for briquetting switchgrass ranged from 0.15 to 0.24%, and 0.74 to 1.72% of the
energy content of the switchgrass for briquetting at room temperature, and at
elevated temperatures of 75 to 150C, respectively. Clearly, more research work
needs to be done to improve the durability of switchgrass briquettes.

4.7 References

ASAE Standards. 2003a. S358.2: Moisture measurement - Forages. St. Joseph, Mich.:

ASABE.

ASAE Standards. 2003b. S319.3: Method of determining and expressing fineness of

feed materials by sieving. St. Joseph, Mich.: ASABE.

ASAE Standards. 2003c. S269.4: Cubes, pellets, and crumbles - Definitions and

methods for determining density, durability, and moisture content. St. Joseph,

Mich.: ASABE.

Back, E.L. 1987. The bonding mechanism in hardboard manufacture. Holzforschung

41(4): 247-258.

Back, E.L., and N.L. Salmen. 1982. Glass transitions of wood components hold

implications for molding and pulping processes. TappiJ. 65(7): 107-110.

Chancellor, W.J. 1962. Formation of hay wafers with impact loads. Agricultural

Engineering 43(3): 136-138,149.

Cheney, W., and D. Kincaid. 1985. Numerical Mathematics and Computing.

Monterey, CA: Brooks/Cole Publishing Company.

160
Coates, W. 2000. Using cotton plant residue to produce briquettes. Biomass and

Bioenergy 18(3): 201-208.

Colley, Z., O.O. Fasina, D. Bransby, and C.W. Wood. 2005. Compaction behavior of

poultry litter and switchgrass. ASAE Paper No. 056053. St. Joseph, Mich.:

ASABE.

DOE. 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The

Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.

Department of Energy (DOE), Office of Scientific and Technical Information.

Available at: http://www/osti.gov/bridge. Accessed 18 June 2005.

Faborode, M.O. 1989. Moisture effects in the compaction of fibrous agricultural

residues. Biological Wastes 28(1): 61-71.

Irvine, G.M. 1984. The glass transitions of lignin and hemicellulose and their

measurement by differential thermal analysis. TappiJ. 67(5): 118-121.

Jannasch, R., Y. Quan, and R. Samson. 2004. A Process and Energy Analysis of

Pelletizing Switchgrass. Ste. Anne de Bellevue, QC: Resource Efficient

Agricultural Production (REAP-Canada). Available at: http://www.reap-

canada.com/library.htm. Accessed 26 August 2004.

Kaliyan, N., and R.V. Morey. 2005. Densification of corn stover. ASAE Paper No.

056134. St. Joseph, Mich.: ASABE.

Kaliyan, N., and R.V. Morey. 2006. Factors affecting strength and durability of

densified products. ASABE Paper No. 066077. St. Joseph, Mich.: ASABE.

161
Lemus, R., E.C. Brummer, K.J. Moore, N.E. Molstad, C.L. Burras, and M.F. Barker.

2002. Biomass yield and quality of 20 switchgrass populations in southern

Iowa, USA. Biomass and Bioenergy 23(6): 433-442.

Mani, S., L.G. Tabil, and S. Sokhansanj. 2002. Compaction behavior of some biomass

grinds. AIC Paper No. 02-305. Saskatoon, Saskatchewan: AIC 2002 Meeting,

CSAE/SCGR Program.

Mani, S., L.G. Tabil, and S. Sokhansanj. 2003. An overview of compaction of biomass

grinds. Powder handling & Processing 15(3): 160-168.

Mani, S., L.G. Tabil, and S. Sokhansanj. 2004a. Evaluation of compaction equations

applied to four biomass species. Canadian Biosystems Engineering 46(3):

3.55-3.61.

Mani, S., L.G. Tabil, and S. Sokhansanj. 2004b. Grinding performance and physical

properties of wheat and barley straws, corn stover and switchgrass. Biomass

and Bioenergy 27(4): 339-352.

McLaughlin, S.B., and L.A. Kszos. 2005. Development of switchgrass {Panicum

virgatum) as a bioenergy feedstock in the United States. Biomass and

Bioenergy 28(6): 515-535.

Mohsenin, N., and J. Zaske. 1976. Stress relaxation and energy requirements in

compaction of unconsolidated materials. J. Agric. Engng Res. 21(2): 193-205.

ORNL. 2005. Biofuels from Switchgrass: Greener Energy Pastures. Oak Ridge, TN:

Oak Ridge National Laboratory (ORNL). Available at:

http://bioenergy.ornl.gov/papers/misc/switgrs.html. Accessed 18 June 2005.

162
Pordesimo, L.O., B.R. Hames, S. Sokhansanj, and W.C. Edens. 2005. Variation in

corn stover composition and energy content with crop maturity. Biomass and

Bioenergy 28(4): 366-374.

Roos, Y.H. 1995. Phase Transitions in Foods. San Diego: Academic Press.

Rumpf, H. 1962. The strength of granules and agglomerates. In Agglomeration, W.A.

Knepper, ed., 379-418. New York, NY: John Wiley and Sons.

Smith, I.E., S.D. Probert, R.E. Stokes, and R.J. Hansford. 1977. The briquetting of

wheat straw. J. Agric. Engng Res. 22(2): 105-111.

Sun, W.Q., and A.C. Leopold. 1994. Glassy state and seed storage stability: a viability

equation analysis. Annals of Botany 74(6): 601-604.

Walsh, M.E., R.L. Perlack, A. Turhollow, D.T. Ugarte, D.A. Becker, R.L. Graham,

S.E. Slinsky, and D.E. Ray. 2005. Biomass Feedstock Availability in the

United States: 1999 State Level Analysis. Oak Ridge, TN: Oak Ridge National

Laboratory. Available at: http://bioenergy.ornl.gov/resourcedata/. Accessed 18

June 2005.

Zhong, Z., and X.S. Sun. 2005. Thermal characterization and phase behavior of

cornstarch studied by differential scanning calorimetry. J. Food Eng. 69(4):

453-459.

163
* Pin connection for connecting the load cell-
j| crosshead of the INSTRON and the piston

350 mm
1 Piston

1. . Gap between the piston and


the cylinder wall = 0.01 mm

6.0-mm thick wall


300 mm

-18.8-mm ID cylinder (Die), which was filled with


5.0-g of biomass grind before compression

RgP
Connection to the base of the INSTRON

Figure 4.1a. Schematic of briquette compression apparatus (figure not to scale).

Figure 4.1b. Uniaxial, piston-cylinder densification (compression) apparatus. Pressure


is applied by the Instron universal testing machine.

164
Pin connection to the
crosshead of the INSTRON

Pin connection

40-mm wide x 90-mm high slot (opening) for


easy collection of briquettes after ejection

60-mm ID x 100-mm high cylinder


(briquette ejection apparatus)

Bolted connection to the


+ base of the INSTRON

Figure 4.2a. Schematic of briquette ejection apparatus in operation (figure not to


scale).

Figure 4.2b. Briquette ejection apparatus located below the piston-cylinder to eject
the briquette. Pressure is applied by the Instron universal testing
machine.
165
H Pressure relief valve
!- Gas out
Removable cap

Cylinder (die)

Heating tape
Fiberglass insulation
Thermocouple (temperature of the grind was
measured at center, half the height of the grind from
the bottom of the cylinder)
O-ring
4 Heating tape

Fiberglass insulation

Figure 4.3a. Modifications to the densification cylinder during preheating of biomass


grind.

Figure 4.3b. Preheating of biomass grind using heating tapes (covered with
insulation).

166
Figure 4.3c. Preheating of biomass grind using heating tapes (insulation is removed to
show the heating tape).

(B) -^

Figure 4.3d. (A): A specially designed cap (consisting of a pressure relief valve, an O-
ring seal, and a K-type thermocouple) used to cover the top of the
cylinder (die) during preheating; (B): Inclusion of an O-ring seal in
the bottom steel plate of the compression apparatus during
preheating and compression.

167
A "5" B
ja.
ra 150 - | 1.5 -
E

Pressure
"g 100 |
3
in

o
jb 50
a.
0\ , r**

0-
100 200 300 10 20 30 40
Time (s) Time (s)

Figure 4.4. Typical briquette compression (A), and ejection (B) curves.

Corn stover, MC = 7.9% wb


Q.
Switchgrass, MC = 10.9% wb
Lignin, MC = 9.7% wb

O
C
LU

03 Mid-point = Tg
0)
X X J Onset

10 30 50 70 90 110 130 150


Temperature (C)

Figure 4.5. DSC thermograms for corn stover, switchgrass, and lignin.

168
Figure 4.6. Durability tester (tumbler or tumbling can) used for measuring the
durability of briquettes based on ASAE Standards S269.4 (2003c).

169
Figure 4.7a. Corn stover grind obtained using a hammer mill screen size of 4.6 mm
(geometric mean particle diameter = 0.80 mm).

Figure 4.7b. Switchgrass grind obtained using a hammer mill screen size of 4.6 mm
(geometric mean particle diameter = 0.64 mm).

170
Figure 4.8. Corn stover briquettes made at pressures of 30, 60, 100, 130 and 150 MPa
(left to right) (geometric mean particle diameter = 0.80 mm; grind
moisture content = 10% w.b.; preheating temperature = 25C).

Figure 4.9. Corn stover briquettes made at 100 MPa pressure for five replications
(geometric mean particle diameter = 0.66 mm; moisture content = 10%
w.b.; preheating temperature = 25C; durability = 58.3%).

Figure 4.10. Switchgrass briquettes made at 100 MPa pressure for five replications
(geometric mean particle diameter = 0.56 mm; moisture content = 10%
w.b.; preheating temperature = 25C; durability = 0.0%).

171
he
^fafl$y$[] _"_ jb,.

fel Ivl f-a M M


Figure 4.11-1. Whole briquettes collected after tumbling in the durability tester.

Figure 4.11-2. Dust generated due to tumbling in the durability tester.

Figure 4.11. Corn stover briquettes after durability test (pressure = 150 MPa;
geometric mean particle diameter = 0.66 mm; moisture content = 10%
w.b.; preheating temperature = 75C; durability = 96.8%).

Figure 4.12-1. Whole briquettes collected after tumbling in the durability tester.

jsgialtoL

"^Kr
Figure 4.12-2. Dust generated due to tumbling in the durability tester.

Figure 4.12. Switchgrass briquettes after durability test (pressure = 150 MPa;
geometric mean particle diameter = 0.56 mm; moisture content = 10%
w.b.; preheating temperature = 100C; durability = 67.3%).

172
Table 4.1. Properties of corn stover and switchgrass grinds.
* Hammer mill ** Grind moisture *** Particle size ** Bulk density
screen used for content (n=3) (kg m 3) (n = 3)
grinding (mm) (%, w.b.) (n = 3)
Corn Stover
3.0 10.1 0.3 0.66 0.32 130.6 2.0
4.6 10.0 0.3 0.80 0.41 108.7 0.5
4.6 15.0 0.5 0.80 0.41 103.1 1.4
Switchgrass
3.0 9.7 + 0.2 0.56 0.29 200.5 3.6
4.6 9.8 0.3 0.64 0.30 181.3 + 3.8
4.6 15.1+0.5 0.64 + 0.30 190.1 6.2
Diameter of the holes on the screen.
** Mean standard deviation.
*** Geometric mean particle diameter (mm) geometric standard deviation.

Table 4.2. Glass transition temperatures of corn stover, switchgrass, and lignin.
Sample * Moisture content n * Glass transition temperature (C)
(%, w.b.) (n = 3) Onset End (\ \

Corn stover 7.9 + 0.2 3 51.4 + 6.6 102.1+7.2 79.6 + 0.5


14.1+0.2 3 55.0 + 5.7 101.2 + 9.6 77.5 + 5.9
19.8 + 0.1 3 54.0+9.5 100.0+8.4 74.7 + 7.9
Switchgrass 10.9 + 0.1 3 57.8+10.5 113.2 + 9.7 78.7+8.2
14.5 + 0.3 3 54.6+12.1 100.7 + 6.3 75.3 + 9.8
20.2 + 0.2 3 50.2 + 4.6 91.6 + 4.6 63.4 + 3.6
Lignin 9.7 + 0.1 2 61.5 + 9.2 100.7+16.5 81.7+10.8
* Mean standard deviation.

173
Table 4.3a. Unit density and durability of corn stover briquettes.

Maximum Preheating * Grind Immediately after After one week of storage


pressure temperature moisture ejection from the at room temperature
(MPa) (C) content die
(%, w.b.) * Unit briquette *Unit * Durability
(n = 3 ) # density briquette (%)
(kg m"3) (n = 5) density (n = 2 ) *
(kg m 3 )

Corn stover grind from hammer mill screen size of 3.0 mm (** Particle size = 0.66 mm 0.32)
100 25 10.1 0.3 1020.9 9.6 904.7 10,4 58.3 2.6
51 1

100 75 10.2 0.3 1125.4 6.8 1087.6 8.3 91.80.6


bl bl
100 100 10.2 + 0.3 1115.6+ 11.4 1068.0 + 91.4 + 0.3
60.5 bl bl
150 25 10.1 0.3 1117.9 9.1 997.4 11.1 75.2 0.6
a2 a2
150 75 10.1 0.3 1191.2 7.0 1153.9 8.8 96.8 0.04
b2 b2
150 100 10.1 0.3 1197.6 14.9 1161.7 96.6 0.04
25.6 b2 b2
150 150 10.1 +0.3 1161.1 19.8 1114.8 + 94.2 + 0.1
21.6 c2 c2
150 25 10.1 0.3 1117.9 + 9.1 997.4 11.1 75.2 0.6
a3 a3
150 25 14.9 0.3 1118.7 7.2 1012.7 + 9.2 93.1 0.5
a3 b3
150 25 19.4 0.6 1074.1 13.4 995.5 11.2 95.5 0.7
a3 b3

174
Table 4.3a. Continued.
Maximum Preheating * Grind Immediately after After one week of storage
pressure temperature moisture ejection from the at room temperature
(MPa) (C) content die
(%, w.b.) * Unit briquette *Unit * Durability
(n = 3 ) # density briquette (%)
(kgm- 3 )(n = 5) density (n = 2 r
(kg m"3)

Corn stover grind from hlammer mill screen size of 4.6 mm (** Particle size = 0.80 mm 0.41)
100 25 10.0 0.3 972.5 6.8 828.5 7.4 49.5 4.3
a4 a4
100 25 15.0 0.5 981.1 16.1 825.4 10.8 80.2 7.3
a4 b4
150 25 10.0 0.3 1074.5 12.2 954.7 17.0 61.62.9
a5 a5
150 75 10.3 0.2 1189.2 12.2 1153.9 97.3 0.1
17.5 b5 b5
150 100 10.3 0.2 1191.7 9.4 1148.8 96.8 0.01
32.2 b5 b5
150 150 10.3 + 0.2 1105.7 10.2 984.6 30.5 91.8 4.9
a5 b5
150 25 15.0 0.5 1020.0 11.7 873.0 7.4 83.5 2.0
a6 a6
150 75 15.4 0.3 1111.0 17.4 1084.4 96.6 0.5
10.1 b6 b6
150 100 15.4 0.3 1129.8 18.8 1078.6 95.7 0.6
22.7 b6 b6c6
150 150 15.4 0.3 1146.0 21.4 1106.4 92.0 0.7
44.1 b6 c6
150 25 10.0 0.3 1074.5 12.2 954.7 17.0 61.6 2.9
a7 a7
150 25 15.0 0.5 1020.0 11.7 873.0 7.4 83.5 2.0
b7 b7
150 25 20.4 0.2 1050.4 30.6 948.4 30.9 96.4 0.02
a7 c7
* Mean + standard deviation.
** Geometric mean particle diameter geometric standard deviation (n = 3).
" Moisture content of the grind before densification and preheating (if applicable).
* Means for unit briquette density or durability of briquettes are not significantly different if they are
followed by similar lowercase letters for a particular treatment (P > 0.05). The number following the
lower case letter refers to particular treatment grouping for a comparison. For example, in "al", "a"
represents significance, and " 1 " refers to the treatment group, in this case particle size of 0.66 mm,
maximum pressure of 100 MPa, and grind moisture content of 10% (w.b.). The treatment is either
preheating at a constant moisture content or moisture conditioning at a particular preheating
temperature.

175
Table 4.3b. Unit density and durability of switchgrass briquettes.

Maximum Preheating * Grind Immediately after After one week of storage


pressure temperature moisture ejection from the at room temperature
(MPa) (C) content die
(%, w.b.) * Unit briquette *Unit * Durability
(n = 3)* density briquette (%)
(kgm- 3 )(n = 5) density (n = 2 ) v
(kg m"3)
(n=sr
Switchgrass grind from hammer mill screen size of 3.0 mm (** Particle size = 0.56 mm 0.29)
100 25 9.7 + 0.2 849.8+ 10.9 645.1 + 14.5 0.0 +0.0 al
n1
100 75 10.4 0.1 1043.7 4.3 1027.0 9.5 54.7 1.9
bl bl
100 100 10.4 + 0.1 1045.8 14.2 1015.6 61.0 1.9
22.9 bl bl
150 25 9.7 0.2 889.0 7.4 678.9 10.9 0.00.0a2
a2
150 75 9.8 0.2 1093.8 7.4 1045.2 62.9 0.2
15.8 b2 b2
150 100 9.8 0.2 1099.7 12.5 1053.5 67.3 1.5
61.7 b2 b2
150 150 9.8 0.2 1095.5 25.2 1065.1 65.7 2.2
33.0 b2 b2
150 25 15.1 0.3 951.1 3.3 824.6 4.0 0.0 0.0 a3
a3
150 75 15.1 0.3 994.7 7.5 968.6 5.2 51.8 0.6
b3 b3
150 100 15.1 0.3 1025.2 + 20.2 997.7 24.9 57.6 0.6
c3 c3
150 25 9.7 + 0.2 889.0 7.4 678.9 10.9 0.0 0.0
a4
150 25 15.1 0.3 951.1 3.3 824.6 4.0 0.0 0.0
b4
150 25 20.4 + 0.2 740.3 14.0 572.2+ 12.0 0.0 + 0.0
c4
150 25 20.4 0.2 740.3 14.0 572.2 12.0 0.00.0a5
a5
150 75 20.4 0.2 839.4 19.9 788.5 23.5 35.6 2.5
b5 b5

176
Table 4.3b. Continued.

Maximum Preheating * Grind Immediately after After one week of storage


pressure temperature moisture ejection from the at room temperature
(MPa) (C) content die
(%, w.b.) * Unit briquette *Unit * Durability
(n = 3) # density briquette (%)
(kg m"3) (n = 5) density
(kg m"3)

Switchgrass grind from lilammer mill screen size of 4.6 mm (** Particle size = 0.64 mm 0.30)
100 25 9.80.3 824.2 1 11.6 631.1 15.4 0.0 0.0
a6
100 25 15.1 +0.5 652.7 23.5 437.3 + 23.6 0.0 + 0.0
b6
150 25 9.8 + 0.3 901.2 + 6.9 688.1 7.7 0.0 0.0 a7
a7
150 75 9.6 + 0.1 1066.8 12.8 1034.0 54.7 0.4
22.6 b7 c7 b7
150 100 9.6 0.1 1094.5 14.7 1063.4 63.8 2.4
57.1 c7 c7
150 150 9.6 + 0.1 1061.2 17.2 1003.6 65.1 3.2
19.5 b7 c7
150 25 15.1 0.5 664.5 10.9 417.2 18.3 0.00.0a8
a8
150 75 14.9 0.6 901.7+ 14.1 833.7 17.1 46.5 + 2.2
b8 b8
150 100 14.9 + 0.6 1017.6 40.8 961.4 76.6 62.5 7.4
c8 b8
150 150 14.9 0.6 1085.4 27.5 1043.8 61.1 2.6
24.3 d8 b8
150 25 9.8 0.3 901.216.9 688.1 7.7 0.010.0
a9
150 25 15.1 0.5 664.5 10.9 417.2 18.3 0.0 + 0.0
b9
150 25 20.3 0.2 693.3 20.9 548.0 7,1 0.010.0
c9
* Mean + standard deviation.
** Geometric mean particle diameter geometric standard deviation (n = 3).
* Moisture content of the grind before densification and preheating (if applicable).
''' Means for unit briquette density or durability of briquettes are not significantly different if they are followed by
similar lowercase letters for a particular treatment (P > 0.05). The number following the lower case letter refers to
particular treatment grouping for a comparison. For example, in "al", "a" represents significance, and " 1 " refers to
the treatment group, in this case particle size of 0.56 mm, maximum pressure of 100 MPa, and grind moisture
content of 10% (w.b.). The treatment is either preheating at a constant moisture content or moisture conditioning at
a particular preheating temperature.

177
Table 4.4a. Stability and moisture content of corn stover briquettes measured after one
week of storage at room temperature.
Maximum Preheating * Briquette * Moisture Mean expansioni (%)
pressure temperature moisture content of (n = 5)
(MPa) (C) content the briquette Axial Radial Briquette
(%, w.b.) particles direction direction volume
(n = 3) (%, w.b.)
(n = 2)*
Corn stover grind from hai[Timer mill screen size of 3.0 mm (** Particle size = 0.66 mm 0.32)
Nominal initial moisture content of the grind = 10% w.b.
100 25 *** NA - 11.6 0.4 11.6
100 75 7.9 0.1 - 2.5 0.2 2.8
100 100 6.1+0.3 10.1 3.0 0.4 3.8
150 25 NA - 10.7 0.6 11.9
150 75 7.7 0.1 - 2.3 0.2 2.6
150 100 7.0 0.2 - 2.6 0.1 2.7
150 150 4.1 +0.2 10.4 3.5 0.7 4.8
Nominal initial moisture content of the grind : 15% w.b.
150 25 10.4 0.1 6.3 0.5 7.4
Nominal initial moisture content of the grind = 20% w.b.
150 25 12.4 0.4 5.2 0.2 5.6
Corn stover grind from hammer mill screen size of 4.6 mm (** Particle size = 0.80 mm 0.41)
Nominal initial moisture content of the grind = 10% w.b.
100 25 NA - 14.1 1.1 16.7
150 25 NA - 11.2 0.6 12.6
150 75 8.4 0.2 - 2.7 0.1 2.9
150 100 7.9 + 0.1 - 3.0 0.3 3.7
150 150 5.8 0.3 10.2 0.7 12.0 1.2 14.7
Nominal initial moisture content of the grind = % w.b.
100 25 NA - 15.8 0.7 17.4
150 25 NA - 13.4 0.8 15.2
150 75 10.3 0.1 - 0.7 0.04 0.8
150 100 8.6 0.2 10.4 3.2 0.2 3.7
150 150 4.1 0.2 10.2 3.9 0.6 5.2
Nominal initial moisture content of the grind = 20% w.b.
150 25 9.9 0.5 4.0 -0.1 3.8
* Mean standard deviation.
** Geometric mean particle diameter geometric standard deviation (n = 3).
*** NA = Data were not taken. However, moisture loss from the briquettes made at room temperature
was small.
Moisture content of the loose particles that were scraped off from the bottom of the briquettes for
preheating at 100 and 150C. When the amount of collected loose particles was small, one replication of
moisture measurement was made for those cases, and thus, there were no standard deviations.
* Shrinkage of briquettes due to moisture loss may have caused a decrease in briquette diameter.

178
Table 4.4b. Stability and moisture content of switchgrass briquettes measured after
one week of storage at room temperature.
Maximum Preheating * Briquette * Moisture Mean expansion (%)
pressure temperature moisture content of (n = 5)
(MPa) (C) content the briquette Axial Radial Briquette
(%, w.b.) particles direction direction volume
(n = 3) (%, w.b.)
("=2)#
Switchgrass grind from hammer mill screen size of 3.0 mm (*** Particle size = 0.56 mm 0.29)
Nominal initial moisture content of the grind = 10% w.b.
100 25 NA - 26.0 1.0 28.6
100 75 8.1 0.1 - 0.7 0.02 0.8
100 100 5.8 0.5 9.0 0.02 2.1 0.2 2.5
150 25 NA - 26.0 0.9 28.2
150 75 7.0 0.2 - 4.0 0.3 4.5
150 100 6.1 0.4 10.4 3.3 0.5 4.8
150 150 3.7 0.1 10.4 0.3 3.6 0.5 4.4
Nominal initial moisture content of the grind =- 15% w.b.
150 25 9.9 0.04 - 11.2 0.5 12.3
150 75 8.5 0.3 - -0.1 & 0.2 0.3
150 100 6.8 0.6 - 0.1 0.3 0.8
Nominal initial moisture content of the grind = 20% w.b.
150 25 9.4 0.1 - 20.1 0.4 21.1
150 75 10.7 + 0.1 : 2J 03 2.6
Switchgrass grind from hammer mill screen size of 4.6 mm (*** Particle size = 0.64 mm 0.30)
Nominal initial moisture content of the grind = 10% w.b.
100 25 *** NA - 26.1 0.9 28.4
150 25 NA - 26.1 1.2 29.1
150 75 7.2 0.1 - 2.3 0.09 2.5
150 100 6.8 0.3 9.0 0.2 2.5 0.3 3.2
150 150 4.7 0.2 8.9 0.2 6.2 0.7 7.9
Nominal initial moisture content of the grind == 15% w.b.
100 25 NA - 36.0 2.0 41.6
150 25 NA - 41.9 3.2 51.1
150 75 10.0 0.2 - 4.4 0.6 5.8
150 100 7.4 0.2 10.0 + 0.1 5.4 0.2 6.0
150 150 4.8 0.5 8.9 0.9 3.5 0.6 4.7
Nominal initial moisture content of the grind = 20% w.b.
150 25 9.9 0.1 15.2 -0.2' 14.8
* Mean + standard deviation.
** Geometric mean particle diameter geometric standard deviation (n = 3).
*** NA == Data were not taken. However, moisture loss from the briquettes made at room temperature was small.
* Moisture content of the loose particles that were scraped off from the bottom of the briquettes for preheating at
100 and 150C. When the amount of collected loose particles was small, one replication of moisture measurement
was made for those cases, and thus, there were no standard deviations.
&
Particle lost from the bottom of the briquettes due to inadequate bonding which caused a decrease in briquette
length. Shrinkage of briquettes due to moisture loss may have caused a decrease in briquette diameter.

179
Table 4.5a. Specific energy required for briquetting corn stover.

Maximum Preheating Compression Ejection Preheating Specific energy


pressure temperature energy (MJ/t) energy energy consumption *
(MPa) (C) (n = 10)* (MJ/t) (MJ/t)
** MJ/t Percentage
(n = 10)*
of energy in
material
(%)*
Corn stover grind from hiiimmer mill screen size of 3.0 mm (** Particle size = 0.66 imm 0.32)
Nominal initial moisture content of the grind = 10%w.b.
100 25 32.60.5 1.0 0.1 0.0 33.6 0.17
100 75 24.9 0.5 0.1 100.0 125.0 0.63
0.02
100 100 26.5 + 0.8 0.2 + 150.0 176.7 0.88
0.02
150 25 44.3 0.4 1.1 0.1 0.0 45.4 0.23
150 75 40.0 0.7 1.1 0.1 100.0 141.1 0.71
150 100 38.6 1.1 1.0 0 . 2 150.0 189.6 0.95
150 150 42.6 1.1 0.70.2 250.0 293.3 1.47
Nominal initial moisture content of the grind = 15% w.b.
150 25 34.2 0 . 5 0.2 0 . 2 0.0 34.4 0.17
Nominal initial moisture content of the grind = 20% w.b.
150 25 37.1 2.1 1.1 0 . 2 0.0 38.2 0.19
Corn stover grind from hammer mill screen size of 4.6 mm (** Particle size = 0.80 mm 0.41)
Nominal initial moisture content of the grind = 10% w.b.
100 25 32.3 0 . 3 0.5 0.1 0.0 32.8 0.16
150 25 43.6 0 . 6 0.5 0.1 0.0 44.1 0.22
150 75 38.9 0 . 8 1.2 0 . 2 100.0 140.1 0.70
150 100 38.3 1 . 0 1.0 0 . 2 150.0 189.3 0.95
150 150 44.0+ 1.2 1.3 + 0.3 250.0 295.3 1.48
Nominal initial moisture content of the grind == 15% w.b.
100 25 24.4 0.8 0.4 0 . 1 0.0 24.8 0.12
150 25 33.4 0 . 4 0.6 0.1 0.0 34.0 0.17
150 75 30.0 0 . 7 0.7 0.1 100.0 130.7 0.65
150 100 33.5 1 . 5 1.0 0 . 2 150.0 184.5 0.92
150 150 41.1 0 . 5 0.7 0.2 250.0 291.8 1.46
Nominal initial moisture content of the grind = 20% w.b.
150 25 34.6 1.1 1.0 0.2 0.0 35.6 0.18
* Mean standard deviation.
** Preheating energy (MJ/t) = Specific heat of corn stover x [preheating temp. - room temp.]
Assumed specific heat of corn stover = 2.0 kJ kg"' K"1; room temperature = 25C.
*** Geometric mean particle diameter geometric standard deviation (n = 3).
* Specific energy consumption = Compression energy + ejection energy + preheating energy.
&
Percentage of energy in material (%) = Specific energy consumption (MJ/t) * 100 / Energy content of
corn stover (MJ/t).
Energy content of corn stover = 20 MJ/kg (Pordesimo et al., 2005).

180
Table 4.5b. Specific energy required for briquetting switchgrass.
Maximum Preheating Compression Ejection Preheating Specific energy
pressure temperature energy (MJ/t) energy energy consumption"
(MPa) (C) (n = 1 0 ) * (MJ/t) (MJ/t)
(n = 10)* ** MJ/t Percentage
of energy in
material
(%) &
Switchg rass grind from hammer mill screen size of 3.0 mm (*** Particle size = 0.56 mm 0.29)
Nominal initial moisture content of the grind = 10% w.b.
100 25 26.4 0 . 3 0.310.1 0.0 26.7 0.16
100 75 26.210.5 0.3 1 0 . 1 100.0 126.5 0.74
100 100 26.610.8 0.210.1 150.0 176.8 1.04
150 25 36.010.3 0.310.1 0.0 36.3 0.21
150 75 38.511.1 0.410.1 100.0 138.9 0.82
150 100 3 7 . 5 1 1.4 0.210.1 150.0 187.7 1.10
150 150 40.010.9 0.210.1 250.0 290.2 1.71
Nominal iniitial moisture content of the grind = 15%w.b.
150 25 36.5 1 0 . 4 0.4 1 0 . 2 0.0 36.9 0.22
150 75 31.1 1 0 . 6 0.21 100.0 131.3 0.77
0.03
150 100 34.011.6 0.210.1 150.0 184.2 1.08
Nominal iniitial moisture content of the grind = 20% w.b.
150 25 32.510.6 0.710.1 0.0 33.2 0.20
150 75 27.04 1 0.5 0.21 100.0 127.2 0.75
0.04
Switchg;rass grind from ha mmer mill screen size of 4.6 mm (*** Particle size = 0.64 mm 0.30)
Nominal initial moisture content of the grind = 10% w.b.
100 25 28.910.4 0.310.1 0.0 29.2 0.17
150 25 40.3 1 0 . 7 0.3 1 0 . 1 0.0 40.6 0.24
150 75 39.910.6 0.410.1 100.0 140.3 0.83
150 100 39.311.2 0.310.1 150.0 189.6 1.12
150 150 4 1 . 6 1 1.0 0.3 1 0 . 1 250.0 291.9 1.72
Nominal initial moisture content of the grind =: 15% w.b.
100 25 24.5 1 0.4 0.510.1 0.0 25.0 0.15
150 25 32.910.6 0.610.1 0.0 33.5 0.20
150 75 32.010.7 0.3 1 0 . 1 100.0 132.3 0.78
150 100 3 5 . 8 1 1.8 0.410.2 150.0 186.2 1.10
150 150 37.511.3 0.110.1 250.0 287.6 1.69
Nominal initial moisture content of the grind = 2 0 % w.b.
150 25 33.010.5 0,6 1 0 . 2 0.0 33.6 0.20
* Mean 1 standard deviation
** Preheating energy (MJ/t) Specific heat of switchgrass x [preheating temp. - room temp.]
Assumed specific heat of switchgrass = 2.0 kJ kg"1 K~'; room temperature = 25C.
*** Geometric mean particle diameter geometric standard deviation (n = 3).
* Specific energy consumption = Compression energy + ejection energy + preheating energy.
&
Percentage of energy in material (%) = Specific energy consumption (MJ/t) x 100 / Energy content of
switchgrass (MJ/t).
Energy content of switchgrass = 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004b).

181
Table 4.6a. Properties of corn stover and switchgrass briquettes (pressure = 150 MPa).

* Grind Preheating Immediately after After one week of storage at room temperature
moisture temperature ejection from the
content ("C) die
(%, w.b.) * Unit briquette * Briquette *Unit * Durability * Mean
(n-3)* density moisture briquette of briquettes briquette
content density (%) volume
(kgm- 3 )(n = 5)
(%, w.b.) expansion
(kg m"3) (n-2)*
(n = 3) (n = 5) (%)
(n = 5 ) s
Corn stover grind from hammer mill screen size of 2.4 mm (** Particle !size = 0.34 mm 0.33)
14.9 25 1009.5 10.3 13.5 928.9 91.40.8 5.9
0.1 0.1 11.2 al al
14.9 + 40 1082.2 + 3.2 9.1 0.4 1040.4 94.50.6 -0.4
0.1 11.6bl al
14.9 60 1088.9+14.6 8.6 0.2 1068.9 95.1 1.4 -2.5
0.1 9.9 cl al
21.7 25 875.4 10.4 16.1 745.2 9.5 88.0+1.1 14.0
0.1 0.2 a2 a2
21.7 40 863.1111.4 16.1 776.5 4.0 93.8 0.5 5.8
0.1 0.4 b2 b2
21.7 + 60 881.5+15.9 14.0 834.7 90.9 0.4 0.9
0.1 0.4 15.8 c2 a2b2
Switchgrass grind fro m hammer mill screen size of 2.4 mm (** Particle :size = 0.49 mm 0.31)
14.8 25 865.6 6.7 11.3 739.2 0.00.0 13.3
0.1 0.2 10.2 a3 a3
14.8 40 866.0 6.5 10.2 801.2 0.0 0.0 3.4
0.1 0.3 11.6 b3 a3
14.8 60 870.0 7.7 10.2 849.0 5.0 23.5 2.2 -1.7
0.1 0.3 c3 b3
15.3 80 810.7 11.5 12.6 804.4 4.7 38.8 1.2 -1.1
0.2 0.4 b3 c3
19.9 + 25 709.3+ 11.1 14.6 567.4 6.0 0.0 0.0 20.4
0.2 0.4 a4 a4
19.9 40 683.6 17.0 13.5 599.7 0.0 0.0 7.6
0.2 0.2 23.7 b4 a4
19.9 60 680.3 15.0 13.1 636.7 0.0 0.0 0.5
0.2 0.4 17.6 c4 a4
20.4 + 80 599.0 7.3 16.2 579.3 7.8 19.0 0.4 0.01
0.1 0.4 a4b4 b4

182
Table 4.6a. Continued.
* Grind Preheating Immediately after After one week of storage at room temperature
moisture temperature ejection from the
content (C) die
(%, w.b.) * Unit briquette * Briquette "Unit * Durability * Mean
(n = 3)* density moisture briquette of briquettes briquette
content density (%) volume
(kgm^Xn^S)
(%, w.b.) (kg m') (n-2)* expansion
(%)
(n = 3) (n-5)*
(n-5)*
Switchgrass grind fro m h a m m e r mill screen size of 2.4 mm (** Particle size = 0.49 mm 0.31)
8.3 0 . 2 80 1110.1 11.3 7.9 0.2 1098.6 + 56.3 2 . 6 0.7
14.9 a5 a5
8.3 0.2 100 1086.3 29.8 7.4 0.1 1074.5 + 61.9 1.0 1.0
35.1 a5 a5
8.3 0.2 200 1109.3 20.0 2.8 0.1 1108.1 54.2 6.2 0.1
13.5 a5 a5
8.3 0.2 80 1110.1 11.3 7.9 0.2 1098.6 56.3 2.6 0.7
14.9 a6 a6
11.9 80 989.3 7.8 10.8 972.5 + 52.4 5.8 0.9
0.5 0.1 13.7 b 6 a6
15.3 80 810.7+ 11.5 12.6 + 804.4 + 4.7 38.8 1.2 -1.1
0.2 0.4 c6 b6
20.4 80 599.0 7.3 16.2 579.3 7 . 8 19.0 0.4 0.01
0.1 0.4 d6 c6
* Mean standard deviation.

** Geometric mean particle diameter + geometric standard deviation (n = 3).

* Moisture content of the grind before densification and preheating (if applicable).
&
Briquette volume reduction may have caused by shrinkage of briquettes either due to loss of moisture
from briquettes or due to loss of particles from the bottom of the briquettes because of inadequate
bonding.
v
Means for unit briquette density or durability of briquettes are not significantly different if they are
followed by similar lowercase letters for a particular treatment (P > 0.05). The number of rows related
to a particular treatment comparison is indicated by the numbers added with the lowercase letters such
as al, a2, and a6. For example, in "al", "a" represents significance, and " 1 " groups the treatment levels
(or rows) that should be compared at a time. The treatment is either preheating at a constant moisture
content or moisture conditioning at a particular preheating temperature.

183
Table 4.6b. Specific energy required for briquetting corn stover and switchgrass
(pressure = 150 MPa).
* Grind Preheating Compression Ejection Preheating Specific energy
moisture temperature energy (MJ/t) energy energy (MJ/t) consumption *
content (C) (n = 5) * (MJ/t)
(%, w.b.) (n = 5) * MJ/t Percentage of
(n = 3)* energy in
material (%)

Corn stover grind from hammer mill screen size of 2.4 mm (***! Particle size = 0.34 mm 0.33)
14.9 + 0.1 25 26.5 + 0.7 0.4 0.1 0.0 26.9 0.14
14.9 + 0.1 40 28.2 + 0.4 0.5 + 0.1 30.0 58.7 0.29
14.9 + 0.1 60 27.0 + 0.3 0.4 + 0.1 70.0 97.4 0.49
21.7 + 0.1 25 24.3 0.5 0.5 0.1 0.0 24.8 0.12
21.7 + 0.1 40 21.3 + 0.5 0.3 0.1 30.0 51.6 0.26
21.70.1 60 20.8 0.3 0.2 0.1 70.0 91.0 0.46
Switchgrass grind from hammer mill screen size of 2.4 mm (*** Particle size = 0.49 mm 0.31)
14.8 + 0.1 25 33.2 + 0.2 0.5 + 0.1 0.0 33.7 0.20
14.8 + 0.1 40 30.5 + 0.4 0.3+0.1 30.0 60.8 0.36
14.8 + 0.1 60 28.5 + 0.2 0.1+0.01 70.0 98.6 0.58
15.3 + 0.2 80 26.4 + 0.3 0.1 0.01 110.0 136.5 0.80
19.9 0.2 25 29.2 0.3 0.4 0.04 0.0 29.6 0.17
19.9 0.2 40 25.5 0.1 0.1 0.01 30.0 55.6 0.33
19.9 0.2 60 24.8 0.2 0.1 0.03 70.0 94.9 0.56
20.4 0.1 80 22.0 0.3 0.04 110.0 132.0 0.78
0.01
8.3 0.2 80 38.9 0.7 0.2 0.03 110.0 149.1 0.88
11.9 + 0.5 80 30.9 0.3 0.2 0.02 110.0 141.1 0.83
8.3 + 0.2 100 38.3 0.4 0.2 0.04 150.0 188.5 1.11
8.3 0.2 200 38.9 0.3 0.1 0.02 350.0 389.0 2.29
* Mean + standard deviation.
** Preheating energy (MJ/t) = Specific heat of biomass x [preheating temp. - room temp.]
Assumed specific heat of biomass = 2.0 kJ kg"' K"'; room temperature = 25C.
*** Geometric mean particle diameter + geometric standard deviation (n = 3).
* Specific energy consumption = Compression energy + ejection energy + preheating energy.
&
Percentage of energy in material (%) = Specific energy consumption (MJ/t) x 100 / Energy content of
biomass (MJ/t).
Energy content of corn stover = 20 MJ/kg (Pordesimo et al., 2005).
Energy content of switchgrass = 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004b).

184
4.8 Appendix

In this section, the results from the statistical analyses are given.

Table 4.A1. ANOVA using univariate general linear model (GLM) on the unit
density of corn stover briquettes.

[P = pressure (MPa); S = hammer mill screen size (mm); M = moisture content (%


w.b.); T = preheating temperature (C)]

Table 4.A.l-a. Between-Subjects Factors.

N
p 100 25
150 105
s 2.4 30
3 45
4.6 55
M 10 59
15 46
20 25
T 25 55
40 10
60 10
75 21
100 20
150 14

185
Table 4.Al-b. Tests of Between-Subjects Effects.
Dependent Variable: Density

Source Type III Sum of Squares df Mean Square F Sig.

Corrected Model 1.944E6 25 77777.075 160.110 .000


Intercept 7.382E7 1 7.382E7 151971.172 .000
P 66120.974 1 66120.974 136.115 .000
S 150077.706 2 75038.853 154.473 .000
M 133970.077 2 66985.038 137.894 .000
T 451951.714 5 90390.343 186.075 .000
P*S 1401.880 1 1401.880 2.886 .092
P*M 7727.781 1 7727.781 15.908 .000
p* j 1200.762 2 600.381 1.236 .295
S*M 90497.400 3 30165.800 62.099 .000
S*T 27798.241 3 9266.080 19.075 .000
M*T 83767.320 5 16753.464 34.488 .000
Error 50520.359 104 485.773
Total 1.308E8 130
Corrected Total 1994947.236 129
a. R Squared = .975 (Adjusted R Squared = .969)

The unit density of corn stover briquettes was significantly affected by


pressure, particle size, moisture content, and preheating temperature (P < 0.05). The
interactions of pressure x particle size, and pressure * temperature on corn stover
briquette density were not significant (P > 0.05). The interactions of pressure x
moisture, particle size x moisture, particle size x temperature, and moisture *
temperature were significant on corn stover briquette density at the 5% level.

186
Table 4.A2. ANOVA using univariate GLM on the durability of corn stover
briquettes.

[P = pressure (MPa); S = hammer mill screen size (mm); M = moisture content (%


w.b.); T = preheating temperature (C)]

Table 4.A2-a. Between-Subjects Factors.

N
p 100 10
150 42
s 2.4 12
3 18
4.6 22
M 10 24
15 18
20 10
T 25 22
40 4
60 4
75 8
100 8
150 6

187
Table 4.A2-b. Tests of Between-Subjects Effects.
Dependent Variable: Durability

Type III Sum of


Source Squares df Mean Square F Sig.
Corrected Model 8300.7248 25 332.029 70.960 .000
Intercept 240865.256 1 240865.256 51476.573 .000
P 183.443 1 183.443 39.205 .000
S 31.312 2 15.656 3.346 .051
M 1028.877 2 514.438 109.943 .000
T 2999.213 5 599.843 128.195 .000
P*S 11.763 1 11.763 2.514 .125
P*M 38.953 1 38.953 8.325 .008
p* j 93.531 2 46.765 9.994 .001
S*M 193.767 3 64.589 13.804 .000
S*T 133.615 3 44.538 9.518 .000
M*T 385.148 5 77.030 16.462 .000
Error 121.657 26 4.679
Total 411003.685 52
Corrected Total 8422.381 51
a. R Squared = .986 (Adjusted R Squared = .972)

Corn stover briquette durability was significantly affected by pressure,


moisture content, and preheating temperature (P < 0.05); however, briquette durability
was not significantly affected by particle size (P > 0.05). The interaction of pressure x
particle size on corn stover briquette durability was not significant (P > 0.05). The
interactions of pressure x moisture, pressure x temperature, particle size x moisture,
particle size x temperature, and moisture x temperature were significant on corn
stover briquette durability at the 5% level.

188
Table 4.A3. ANOVA using univariate GLM on the unit density of switchgrass
briquettes.

[P = pressure (MPa); S = hammer mill screen size (mm); M = moisture content (%


w.b.); T = preheating temperature (C)]

Table 4.A3-a. Between-Subjects Factors.

N
p 100 25
150 145
s 2.4 55
3 60
4.6 55
M 8 10
10 60
12 5
15 60
20 35
T 25 55
40 10
60 10
75 30
80 20
100 30
150 15

189
Table 4.A3-b. Tests of Between-Subjects Effects.
Dependent Variable: Density

Type III Sum of


Source Squares df Mean Square F Sig.

Corrected Model 7.171E6 33 217312.302 328.698 .000


Intercept 6.459E7 1 6.459E7 97690.474 .000
P 3565.848 1 3565.848 5.394 .022
S 85086.509 2 42543.254 64.349 .000
M 1061261.795 4 265315.449 401.305 .000
T 1810599.859 6 301766.643 456.440 .000
P*S 673.871 1 673.871 1.019 .314
P*M 7427.415 1 7427.415 11.234 .001
p * i" 539.889 2 269.944 .408 .666
S*M 214861.458 3 71620.486 108.330 .000
S*T 97831.739 3 32610.580 49.325 .000
M*T 122632.578 7 17518.940 26.498 .000
S*M *T 95394.265 2 47697.133 72.145 .000
Error 89913.810 136 661.131
Total 1.231E8 170
Corrected Total 7261219.775 169
a. R Squared = .988 (Adjusted R Squared = .985)

The unit density of switchgrass briquettes was significantly affected by


pressure, particle size, moisture content, and preheating temperature (P < 0.05). The
interactions of pressure x particle size and pressure x temperature on the switchgrass
briquette density were not significant (P > 0.05). The interactions of pressure x
moisture, particle size x moisture, particle size x temperature, moisture x temperature,
and particle size x moisture x temperature were significant on switchgrass briquette
density at the 5% level.

190
Table 4.A4. ANOVA using univariate GLM on the durability of switchgrass
briquettes.

[P = pressure (MPa); S = hammer mill screen size (mm); M = moisture content (%


w.b.); T = preheating temperature (C)]

Table 4.A4-a. Between-Subjects Factors.

N
p 100 10
150 58
s 2.4 22
3 24
4.6 22
M 8 4
10 24
12 2
15 24
20 14
T 25 22
40 4
60 4
75 12
80 8
100 12
150 6

191
Table 4.A4-b. Tests of Between-Subjects Effects.
Dependent Variable: Durability

Type III Sum of


Source Squares df Mean Square F Sig.

Corrected Model 53717.469" 33 1627.802 364.406 .000


Intercept 42625.780 1 42625.780 9542.381 .000
P 48.602 1 48.602 10.880 .002
S 7.271 2 3.635 .814 .452
M 2143.094 4 535.773 119.940 .000
T 20181.298 6 3363.550 752.978 .000
P*S .000 1 .000 .000 1.000
P*M .000 1 .000 .000 1.000
p * TT 2 4.134
36.936 18.468 .025
S*M 21.140 3 7.047 1.577 .213
S *T 76.487 3 25.496 5.708 .003
M*T 882.475 7 126.068 28.222 .000
S*M*T 17.748 2 8.874 1.987 .153
Error 151.878 34 4.467
Total 120218.426 68
Corrected Total 53869.347 67
a. R Squared = .997 (Adjusted R Squared = .994)

Switchgrass briquette durability was significantly affected by pressure,


moisture content, and preheating temperature (P < 0.05); however, briquette durability
was not significantly affected by particle size (P > 0.05). The interactions of pressure
x particle size, pressure x moisture, particle size x moisture, and particle size x
moisture x temperature on switchgrass briquette durability were not significant (P >
0.05). The interactions of pressure x temperature, particle size x temperature, and
moisture x temperature were significant on switchgrass briquette durability at the 5%
level.

192
Table 4.A5. One-way ANOVA followed by Tukey's procedure (T-test) to
investigate the significant effect of treatment factors within a group of
tests.

These statistical analyses were used to separate the means of density or


durability to show the significant differences at the 5% level (tables 4.3a, 4.3b, and
4.6a). Below are the statistical test results for one example case of densification
conditions:

Biomass = corn stover; pressure = 150 MPa; particle size = 3.0 mm; moisture content
of the grind = 10% (w.b.); and preheating temperature = 25, 75, and 100C.

Table 4.A5-a. One-way ANOVA.


Density

Sum of Squares df Mean Square F Sig.

Between Groups 86327.423 3 28775.808 87.089 .000


Within Groups 5286.673 16 330.417
Total 91614.096 19

193
Table 4.A5-b. Post Hoc Tests: Multiple Comparisons.
Density
Tukey HSD

Mean Difference 95% Confidence Interval


(I)T (J)T (I-J) Std. Error Sig. Lower Bound Upper Bound

25 75 -156.54208* 11.49638 .000 -189.4335 -123.6507

100 -164.26291* 11.49638 .000 -197.1543 -131.3715

150 -117.45731* 11.49638 .000 -150.3487 -84.5659

75 25 156.54208* 11.49638 .000 123.6507 189.4335


100 -7.72082 11.49638 .906 -40.6122 25.1706
150 39.08477* 11.49638 .017 6.1934 71.9762

100 25 164.26291* 11.49638 .000 131.3715 197.1543


75 7.72082 11.49638 .906 -25.1706 40.6122

150 46.80560* 11.49638 .004 13.9142 79.6970

150 25 117.45731* 11.49638 .000 84.5659 150.3487


75 -39.08477* 11.49638 .017 -71.9762 -6.1934
100 -46.80560* 11.49638 .004 -79.6970 -13.9142
* The mean difference is significant at the 0.05 level.

194
Table 4.A5-C. Homogeneous Subsets for Density.
Tukey HSD

Subset for alpha = 0.05


T N 1 2 3

25 5 997.3908
150 5 1114.8481
75 5 1153.9329
100 5 1161.6537
Sig. 1.000 1.000 .906
Means for groups in homogeneous subsets are displayed.

The above table (table 4.A7-c) suggests that for the densification conditions
tested, the unit density of corn stover briquettes at the preheating temperatures of 75
and 100C were similar (P > 0.05). However, the briquette densities were significantly
different from each other when comparing for the preheating temperatures of 25, 75 to
100, and 150C(P< 0.05).

Table 4.A5-al. One-way ANOVA.


Durability

Sum of Squares df Mean Square F Sig.

Between Groups 647.426 3 215.809 2187.948 .000


Within Groups .395 4 .099
Total 647.821 7

195
Table 4.A5-bl. Post Hoc Tests: Multiple Comparisons.
Durability
Tukey HSD

Mean Difference 95% Confidence Interval


(I)T (J)T (I-J) Std. Error Sig. Lower Bound Upper Bound

25 75 -21.60422* .31406 .000 -22.8827 -20.3257

100 -21.35041* .31406 .000 -22.6289 -20.0719

150 -18.96066* .31406 .000 -20.2392 -17.6822

75 25 21.60422* .31406 .000 20.3257 22.8827


100 .25381 .31406 .848 -1.0247 1.5323
150 2.64356* .31406 .004 1.3651 3.9221

100 25 21.35041* .31406 .000 20.0719 22.6289


75 -.25381 .31406 .848 -1.5323 1.0247
150 2.38975* .31406 .006 1.1112 3.6683

150 25 18.96066* .31406 .000 17.6822 20.2392

75 -2.64356* .31406 .004 -3.9221 -1.3651

100 -2.38975* .31406 .006 -3.6683 -1.1112


* The mean difference is significant at the 0.05 level.

Table 4.A5-cl. Homogeneous Subsets for Durability.


Tukey HSD

Subset for alpha = 0.05


T N 1 2 3

25 2 75.2266
150 2 94.1873
100 2 96.5771
75 2 96.8309
Sig. 1.000 1.000 .848
Means for groups in homogeneous subsets are displayed.

196
One can conclude from the above table (table 4.A7-cl) that for the
densification conditions tested, the durability of corn stover briquettes at the
preheating temperatures of 75 and 100C were similar (P > 0.05). However, the
briquette durabilities were significantly different from each other when comparing the
preheating temperatures of 25, 75 to 100, and 150C (P < 0.05).

Table 4.A6. Independent samples T-test on the unit density or durability of corn
stover briquettes (moisture conditioning versus preheating).

Control conditions:

Particle size: 3.0 and 4.6 mm hammer mill screen sizes


Moisture content of the grind: 10% (w.b.)
Preheating temperature: 25C

Moisture conditioning (moist) conditions:

Particle size: 2.4, 3.0, and 4.6 mm hammer mill screen sizes
Moisture of the grind: 15.0 and 20% (w.b.)
Preheating temperature: 25C

Preheating (heat) conditions:

Particle size: 3.0 and 4.6 mm hammer mill screen sizes


Moisture content of the grind: 10% (w.b.)
Preheating temperature: 75 and 100C

Tests below are for control versus moisture conditioning.

Table 4.A6-al. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Density 1 control 10 976.0507 26.26041 8.30427

2 moist 30 917.2880 91.95950 16.78943

197
Table 4.A6-a2. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Density Equal
variances 7.975 .008 1.98 38 .055 58.76261 29.70300 -1.36797 118.89
assumed

Equal
variances 3.14 37.66 .003 58.76261 18.73088 20.8327 96.692
not assumed

Table 4.A6-a3. Group Statistics.


group N Mean Std. Deviation Std. Error Mean

Durability 1 control 4 68.3870 8.08209 4.04105

2 moist 12 91.2910 4.70804 1.35909

198
Table 4.A6-a4. Independent Samples Test.
Levene's Test for
Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Durability Equal variances


4.879 .044 -7.08 14 .000 -22.90399 3.23589 -29.8443 -15.96
assumed

Equal
variances not -5.37 3.70 .007 -22.90399 4.26347 -35.1237 -10.68
assumed

Moisture conditioning resulted in significantly different unit briquette density


(58.7 kg m"3 lower density) and durability (22.9% points higher durability) for corn
stover briquettes than for the control (P < 0.05).

199
Tests below are for control versus preheating.

Table 4.A6-bl. Group Statistics.


group N Mean Std. Deviation Std. Error Mean

Density 1 control 10 976.0507 26.26041 8.30427

3 heat 20 1154.5878 21.43759 4.79359

Table 4.A6-b2. Independent Samples Test.

Levene's Test
for Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Density Equal
variances 1.008 .324 -19.96 28 .000 -178.5372 8.94576 -196.8617 -160.22
assumed

Equal
variances
-18.62 15.198 .000 -178.5372 9.58850 -198.9514 -158.13
not
assumed

200
Table 4.A6-b3. Group Statistics.
group N Mean Std. Deviation Std. Error Mean

Durability 1 control 4 68.3870 8.08209 4.04105

3 heat 8 96.8693 .28011 .09903

Table 4.A6-b4. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the

Sig. (2- Mean Std. Error Difference


F Sig. t df tailed) Difference Difference Lower Upper

Durability Equal
variances 129.689 .000 -10.49 10 .000 -28.48226 2.71461 -34.5308 -22.44
assumed

Equal
variances -7.046 3.00 .006 -28.48226 4.04226 -41.3378 -15.63
not assumed

Preheating resulted in significantly different unit briquette density (178.5 kg m


3
higher density) and durability (28.5% points higher durability) for corn stover
briquettes than for the control (P < 0.05).

201
Tests below are for moisture conditioning versus preheating.

Table 4.A6-cl. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Density 2 moist 30 917.2880 91.95950 16.78943

3 heat 20 1154.5878 21.43759 4.79359

Table 4.A6-c2. Independent Samples Test.

Levene's Test
for Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Density Equal
variances 19.101 .000 -11.30 48 .000 -237.2998 20.99817 -279.5194 -195.08
assumed

Equal
variances not -13.59 33.580 .000 -237.2998 17.46034 -272.7998 -201.80
assumed

Table 4.A6-c3. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Durability 2 moist 12 91.2910 4.70804 1.35909

3 heat 8 96.8693 .28011 .09903

202
Table 4.A6-c4. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95%
Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Durability Equal variances


15.661 .001 -3.32 18 .004 -5.57827 1.68178 -9.1116 -2.05
assumed

Equal
variances not -4.09 11.117 .002 -5.57827 1.36270 -8.5737 -2.58
assumed

Preheating resulted in corn stover briquettes with significantly different density


(237.3 kg m"3 higher density) and durability values (5.6% points higher durability)
than for moisture conditioning (P < 0.05).

Table 4.A7. Independent samples T-test on the unit density or durability of


switchgrass briquettes (moisture conditioning versus preheating).

Control conditions:

Particle size: 3.0 and 4.6 mm hammer mill screen sizes


Moisture content of the grind: 10% (w.b.)
Preheating temperature: 25C

Moisture conditioning (moist) conditions:

Particle size: 2.4, 3.0, and 4.6 mm hammer mill screen sizes
Moisture of the grind: 15.0 and 20% (w.b.)
Preheating temperature: 25C

203
Preheating (heat) conditions:

Particle size: 3.0 and 4.6 mm hammer mill screen sizes


Moisture content of the grind: 10% (w.b.)
Preheating temperature: 75 and 100C

Tests below are for control versus moisture conditioning.

Table 4.A7-al. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Density 1 control 10 683.4837 10.13879 3.20617

2 moist 30 611.4347 136.17601 24.86222

Table 4.A7-a2. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the

Sig. (2- Mean Std. Error Difference


F Sig. t df tailed) Difference Difference Lower Upper

Density Equal variances


21.185 .000 1.66 38 .106 72.04898 43.47608 -15.9638 160.06
assumed

Equal
variances not 2.87 29.946 .007 72.04898 25.06810 20.84921 123.25
assumed

204
Table 4.A7-a3. Group Statistics.
group N Mean Std. Deviation Std. Error Mean

Durability 1 control 4 .0000 .00000" .00000

2 moist 12 .0000 .00000" .00000


a. t cannot be computed because the standard deviations of both groups are 0.

Moisture conditioning resulted in significantly lower unit briquette density


(72.1 kg m"3 lower density) for switchgrass briquettes than for the control (P < 0.05).

Tests below are for control versus preheating.

Table 4.A7-bl. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Density 1 control 10 683.4837 10.13879 3.20617

3 heat 20 1049.0157 42.07358 9.40794

Table 4.A7-b2. Independent Samples Test.


Levene's Test
for Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Density Equal
variances 8.834 .006 -26.87 28 .000 -365.532 13.60646 -393.4035 -337.66
assumed

Equal
variances not -36.78 23.014 .000 -365.532 9.93926 -386.0922 -344.97
assumed

205
Table 4.A7-b3. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Durability 1 control 4 .0000 .00000 .00000

3 heat 8 62.1664 5.05729 1.78802

Table 4.A7-b4. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Durability Equal
variances 5.658 .039 -23.99 10 .000 -62.16645 2.59109 -67.9398 -56.39
assumed

Equal
variances not -34.77 7.000 .000 -62.16645 1.78802 -66.3945 -57.94
assumed

Preheating resulted in significantly higher unit briquette density (365.5 kg m'


higher density) and durability (62.1% points higher durability) for switchgrass
briquettes than for the control (P < 0.05).

Tests below are for moisture conditioning versus preheating.

Table 4.A7-cl. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Density 2 moist 30 611.4347 136.17601 24.86222

3 heat 20 1049.0157 42.07358 9.40794

206
Table 4.A7-c2. Independent Samples Test.
Levene's Test
for Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Density Equal
variances 23.306 .000 -13.89 48 .000 -437.5809 31.49644 -500.9088 -374.25
assumed

Equal
variances not -16.46 36.749 .000 -437.5809 26.58269 -491.455 -383.71
assumed

Table 4.A7-c3. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Durability 2 moist 12 .0000 .00000 .00000

3 heat 8 62.1664 5.05729 1.78802

207
Table 4.A7-c4. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Durability Equal variances


18.332 .000 -43.2 18 .000 -62.16645 1.43950 -65.1907 -59.14
assumed

Equal
variances not -34.8 7.0 .000 -62.16645 1.78802 -66.3945 -57.94
assumed

Preheating resulted in switchgrass briquettes with significantly different


density (437.6 kg m"3 higher density) and durability (62.2% points higher durability)
values than for moisture conditioning (P < 0.05).

208
Table 4.A8. Independent samples T-test to determine the effect of particle size on
corn stover briquette density or durability when moisture conditioning to
15 to 20% (w.b.).

Data were pooled for the densification conditions of pressure of 150 MPa,
particle sizes from 3.0- and 4.6-mm hammer mill screen sizes, moisture contents of 15
and 20% (w.b.), and preheating temperature of 25C.

Table 4.A8-a. Group Statistics.


group N Mean Std. Deviation Std. Error Mean

Density 3.0 mm 10 1004.1159 13.25972 4.19309

4.6 mm 10 910.6925 45.04745 14.24525

Table 4,A8-b. Independent Samples Test.


Levene's Test for
Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Density Equal variances


15.378 .001 6.29 18 .000 93.42341 14.84956 62.22565 124.62
assumed

Equal
variances not 6.29 10.548 .000 93.42341 14.84956 60.56804 126.28
assumed

The particle size obtained from 3.0-mm hammer mill screen size resulted in
significantly higher briquette density (93.4 kg m"3 higher) than for 4.6-mm hammer
mill screen size (P < 0.05).

209
Table 4.A8-al. Group Statistics.

group N Mean Std. Deviation Std. Error Mean

Durability 3.0 mm 4 94.2810 1.46808 .73404

4.6 mm 4 89.9133 7.54014 3.77007

Table 4.A8-bl. Independent Samples Test.

Levene's Test for


Equality of
Variances t-test for Equality of Means

95% Confidence
Interval of the
Sig. (2- Mean Std. Error Difference
F Sig. t df tailed) Difference Difference Lower Upper

Durability Equal variances


73.891 .000 1.137 6 .299 4.36771 3.84087 -5.0306 13.77
assumed

Equal
variances not 1.14 3.22 .334 4.36771 3.84087 -7.3831 16.12
assumed

The particle sizes obtained from 3.0- and 4.6-mm hammer mill screen sizes
resulted in similar durability values for corn stover briquettes (P > 0.05).

210
Table 4.A9. Statistical analyses (one-way ANOVA followed by T-test) to
determine the effect of moisture content on switchgrass briquette density
or durability at the preheating temperature of 100C.

Data were pooled for the densification conditions of pressure of 150 MPa,
particle sizes from 2.4-, 3.0-, and 4.6-mm hammer mill screen sizes, moisture contents
of 8.3, 10, and 15% (w.b.), and preheating temperature of 100C.

Table 4.A9-a. One-way ANOVA.


Density

Sum of Squares df Mean Square F Sig.

Between Groups 43439.260 2 21719.630 7.625 .003


Within Groups 62666.211 22 2848.464
Total 106105.471 24

Table 4.A9-b. Post Hoc Tests: Multiple Comparisons.


Density
Tukey HSD

Mean Difference 95% Confidence Interval


(I) M (J) M (I-J) Std. Error Sig. Lower Bound Upper Bound

8 10 16.02735 29.23250 .848 -57.4065 89.4613

15 94.92959* 29.23250 .010 21.4957 168.3635


10 8 -16.02735 29.23250 .848 -89.4613 57.4065
15 78.90224* 23.86824 .009 18.9437 138.8608
15 8 -94.92959* 29.23250 .010 -168.3635 -21.4957

10 -78.90224* 23.86824 .009 -138.8608 -18.9437


* The mean difference is significant at the 0.05 level.

211
Table 4.A9-C. Homogeneous Subsets for Density.
Tukey HSD

Subset for alpha = 0.05


M N 1 2

15 10 979.5369
10 10 1058.4392
8 5 1074.4665
Sig. 1.000 .831
Means for groups in homogeneous subsets are displayed.

The above table (table 4. A9-c) indicates that the moisture contents of 8 and
10% (w.b.) resulted in similar switchgrass briquette density compared to that at 15%
(w.b.) moisture content when the switchgrass was preheated to 100C (P < 0.05).

Table 4.A9-al. One-way ANOVA.


Durability

Sum of Squares df Mean Square F Sig.

Between Groups 82.075 2 41.037 1.605 .267


Within Groups 178.931 7 25.562
Total 261.006 9

212
Table 4.A9-bl. Post Hoc Tests: Multiple Comparisons.
Durability
Tukey HSD

Mean Difference 95% Confidence Interval


(I) M (J) M (I-J) Std. Error Sig. Lower Bound Upper Bound

8 10 -1.49286 4.37849 .938 -14.3878 11.4020

15 4.74582 4.37849 .553 -8.1491 17.6407


10 8 1.49286 4.37849 .938 -11.4020 14.3878
15 6.23868 3.57502 .255 -4.2900 16.7673

15 8 -4.74582 4.37849 .553 -17.6407 8.1491

10 -6.23868 3.57502 .255 -16.7673 4.2900

Table 4.A9-cl. Homogeneous Subsets for Durability.


Tukey HSD

Subset for alpha = 0.05


M N 1

15 4 57.1322
8 2 61.8780
10 4 63.3709
Sig. .342
Means for groups in homogeneous subsets are displayed.

The above table (table 4.A9-cl) indicates that the moisture contents of 8, 10,
and 15% (w.b.) resulted in similar briquette durabilities for switchgrass preheated to
100C (P > 0.05). This conclusion can be derived from the one-way ANOVA itself (P
> 0.05).

213
Table 4.A10. Statistical analyses (one-way ANOVA followed by T-test) to
determine the effect of particle size on switchgrass briquette density or
durability at the preheating temperature of 100C.

Data were pooled for the conditions of pressure of 150 MPa, particle sizes
from 2.4-, 3.0-, and 4.6-mm hammer mill screen sizes, moisture contents of 8.3, 10,
and 15% (w.b.), and preheating temperature of 100C.

Table 4.A10-a. One-way ANOVA.


Density

Sum of Squares df Mean Square F Sig.


Between Groups 13178.474 2 6589.237 1.560 .233
Within Groups 92926.997 22 4223.954
Total 106105.471 24

Table 4.A10-b. Post Hoc Tests: Multiple Comparisons.


Density
Tukey HSD

Mean Difference 95% Confidence Interval


(I) S (J) S (I-J) Std. Error Sig. Lower Bound Upper Bound

2.4 3 48.89429 35.59756 .372 -40.5290 138.3176

4.6 62.06266 35.59756 .212 -27.3607 151.4860

3 2.4 -48.89429 35.59756 .372 -138.3176 40.5290


4.6 13.16837 29.06529 .894 -59.8455 86.1822
4.6 2.4 -62.06266 35.59756 .212 -151.4860 27.3607
3 -13.16837 29.06529 .894 -86.1822 59.8455

214
Table 4.A10-C. Homogeneous Subsets for Density.
Tukey HSD

Subset for alpha = 0.05


s N 1

4.6 10 1012.4038
3 10 1025.5722
2.4 5 1074.4665
Sig. .177
Means for groups in homogeneous subsets are displayed.

The above table (table 4.A10-c) confirms that with preheating to 100C, the
particle sizes obtained from the hammer mill screen sizes of 2.4, 3.0, and 4.6 mm
resulted in similar switchgrass briquette densities (P > 0.05). This conclusion can be
obtained from the one-way ANOVA itself (P > 0.05).

Table 4.A10-al. One-way ANOVA.


Durability

Sum of Squares df Mean Square F Sig.

Between Groups 72.811 2 36.405 1.354 .318


Within Groups 188.195 7 26.885
Total 261.006 9

215
Table 4.A10-bl. Post Hoc Tests: Multiple Comparisons.
Durability
Tukey HSD

Mean Difference 95% Confidence Interval


(I) S (J) S (I-J) Std. Error Sig. Lower Bound Upper Bound

2.4 3 4.55432 4.49040 .592 -8.6702 17.7788

4.6 -1.30136 4.49040 .955 -14.5259 11.9231


3 2.4 -4.55432 4.49040 .592 -17.7788 8.6702
4.6 -5.85568 3.66640 .308 -16.6534 4.9421
4.6 2.4 1.30136 4.49040 .955 -11.9231 14.5259
3 5.85568 3.66640 .308 -4.9421 16.6534

Table 4.A10-cl. Homogeneous Subsets for Durability.


Tukey HSD

Subset for alpha = 0.05


s N 1

3 4 57.3237
2.4 2 61.8780
4.6 4 63.1794
Sig. .399
Means for groups in homogeneous subsets are displayed.

The above table (table 4.A10-cl) confirms that with preheating to 100C, the
particle sizes obtained from the hammer mill screen sizes of 2.4, 3.0, and 4.6 mm
resulted in similar briquette durabilities for switchgrass (P > 0.05). This conclusion
can be deduced from the one-way ANOVA itself (P > 0.05).

216
Chapter 5
STRATEGIES TO IMPROVE DURABILITY OF SWITCHGRASS BRIQUETTES

- LABORATORY SCALE STUDY

5.1 Abstract

The effects of fine grinding of switchgrass, adding biomass based binders


(corn stover, corn Distillers Dried Grains with Solubles (DDGS), and starch), and
mixing commercial chemical feed binders (lignin-sulfonate, lime, and sodium
bentonite) on the switchgrass briquette durability were investigated using a uniaxial,
piston-cylinder densification apparatus. At a forming pressure of 150 MPa, preheating
of switchgrass to 100C without adding any binders improved the briquette durability
to 67% compared to the briquette durability of 0% without preheating. In addition to
preheating to 100C, either mixing of 20% (wt.) of corn stover or 2% (wt.) of lignin-
sulfonate based binders or 5% (wt.) lime produced switchgrass briquettes with
durability of about 80%. Preheating was essential to fully activate the binding agents
added to the feedstock. In commercial-scale pelleting/briquetting of switchgrass, at the
optimized particle size (0.49 to 0.89 mm), moisture content (>10% wet basis), and
preheating temperature (>75C), switchgrass pellets/briquettes with durability of
>80% could be produced by activating the natural binders in the switchgrass, without
adding any binding agents.

5.2 Introduction

Switchgrass (Panicum virgatum L.), a native perennial warm-season grass, has


been identified as a dedicated energy crop for producing bio-fuels, chemicals, and
renewable energy (ORNL, 2005; McLaughlin and Kszos, 2005). Renewable energy
derived from the biomass materials such as switchgrass could reduce the nation's
dependence on foreign oil, reduce greenhouse gas emissions, increase farm income,
and create job opportunities in rural areas (DOE, 2005). Switchgrass can be collected

217
as bales during a limited harvest season with a typical bulk density of about 200 kg m"
3
(Sanderson et al., 1997; Shinners et al., 2006). Thus, densification of switchgrass into
pellets and briquettes is important to help handling, transportation, and storage of
switchgrass for bio-based and bio-energy industrial applications. Also, densified
switchgrass would be in more consistent form, and easier to handle and feed into the
processing equipment.
Using uniaxial compression apparatuses, Mani et al. (2004a), Colley et al.
(2005), and Kaliyan and Morey (2006a) [Chapter 4] studied the densification
characteristics of switchgrass in the laboratory. These laboratory studies concluded
that switchgrass was difficult to densify compared to other biomass materials such as
corn stover. Samson et al. (2000) and Jannasch et al. (2004) studied the pelleting
characteristics of switchgrass in 19-kW (25-HP) and 149-kW (200-HP) pellet mills,
respectively. These two pelleting studies concluded that switchgrass pellets had bulk
density of 593 to 641 kg m"3 with hardness of >30 on the Pfizer tablet hardness tester.
However, they found that switchgrass pellets created large quantities of fines (about
43% of initial feedstock dry matter) during the process of handling, screening, and
shaking before bagging of pellets. Therefore, they recommended finding ways to
improve the durability of switchgrass pellets for successful commercialization.
A literature review by Kaliyan and Morey (2006b) [Chapter 3] concluded that
strength and durability of densified biomass products could be increased by: (i)
activating the natural binders such as lignin, starch, and protein in the feedstock using
moisture and temperature, (ii) adding 0.5 to 5% (wt.) chemical binding agents such as
lignin-sulfonate and lime, and (iii) mixing >20% (wt.) biomass based binders such as
starch, molasses and sawdust.

218
5.3 Objectives

In a previous laboratory densification study (Chapter 4; Kaliyan and Morey,


2006a), the effects of pressure (100 and 150 MPa), particle size (0.56 and 0.64 mm),
moisture content (10, 15 and 20% wet basis), and preheating temperature (75, 100 and
150C) on the densification characteristics of switchgrass were studied. It was found
that at room temperature (about 25C), switchgrass briquettes (about 19.4 mm
diameter) with relaxed densities of 417 to 825 kg m"3 could be produced at pressures
of 100 to 150 MPa, geometric mean particle sizes of grind from 0.56 to 0.64 mm, and
moisture contents of 10 to 20% (wet basis). However, durability of switchgrass
briquettes made at room temperature was zero percent. At elevated temperatures of 75
to 150C and at 150 MPa, switchgrass briquettes with relaxed densities of 834 to 1065
kg m"3 and durability of 55 to 67%) were produced.
Currently, there is no standard available on the acceptance levels of durability
of densified biomass products (Chapter 3; Kaliyan and Morey, 2006b). Adapa et al.
(2003) classified alfalfa pellet quality as high, medium, and low when the pellet
durability was above 80%>, between 70 and 80%>, and below 70%>, respectively. It
seems that further research is required to determine the densification conditions that
would improve the durability of switchgrass briquettes to >80%.

The main objective of this study was to develop strategies to improve the
durability of switchgrass briquettes. The sub-objectives were to:

Determine the effect of fine grinding on the durability of switchgrass briquettes.

Determine the effect of adding biomass-based binders on the durability of


switchgrass briquettes.

Determine the effect of adding chemical binders on the durability of switchgrass


briquettes.

219
5.4 Materials and Methods

5.4-1 Biomass and Binders

To obtain switchgrass samples, whole switchgrass plants (about 50-mm above


the ground) were manually harvested during August 2005 from a field at the West
Central Research and Outreach Center (WCROC), University of Minnesota, Morris,
MN. The initial moisture content of the switchgrass was about 48% (wet basis).
Before grinding, the switchgrass was sun-dried to about 10% moisture content and
then it was chopped into lengths of about 100 to 150-mm. Corn stover used for the
study was collected from freshly harvested bales stored outdoors in WCROC, Morris,
MN during November 2004.
A 5.6-kW hammer mill (J.B. Sedberry Inc., Franklin, TN) with a screen having
3.0 mm round holes was used to obtain corn stover and switchgrass grinds used for the
study. Switchgrass and corn stover grinds were stored in a refrigerator maintained at
5C until used for the tests. To obtain switchgrass grinds to study the effect of fine
grinding, a Thomas-Wiley Mill (Model Ed-5; 0.25 kW (0.33 hp); Arthur H. Thomas
Company, Philadelphia, PA) with 1.0- and 2.0-mm screens was used to reduce the
particle size of the hammer milled switchgrass grind. To adjust the moisture content of
the grinds to 10% (wet basis), a predetermined amount of distilled water was added to
the grinds, thoroughly mixed and stored in zip-lock plastic bags at 5C for 48 h for
tempering.
The biomass-based binders used for the study were corn stover, corn Distillers
Dried Grains with Solubles (DDGS), and Starch. The chemical binders used for the
study were lime, sodium bentonite, Super-bind, Pel-stik, and Xtra-bond. These
chemical binders are commonly used for animal feed pelleting (Tabil et al., 1997;
Kaliyan and Morey, 2006b [Chapter 3]). The corn DDGS were collected in July 2005
from Chippewa Valley Ethanol Company LLC, Benson, MN. The chemical binders
and starch were provided by Uniscope Inc., Johnstown, CO (www.uniscope-inc.com).

220
Moisture content of the grinds was measured using the procedure given in
ASAE Standard S358.2 (ASAE Standards, 2003a). The moisture content values
reported in this paper are on wet basis (w.b.). Bulk density of grind was calculated
from the mass of grind that occupied a 250-mL glass container. Particle size
distribution of the grinds was determined based on ASAE Standard S319.3 (ASAE
Standards, 2003b). Samples of corn stover grind, switchgrass grind and corn DDGS
were sent to a forage analysis laboratory (Dairy One, Ithaca, NY) to determine the
compositions of these materials using Near Infrared Reflectance (NIR) spectroscopy.

5.4-2 Briquetting Procedure

Briquettes were made using a uniaxial, piston-cylinder densification apparatus


in the laboratory. The details of the uniaxial piston-cylinder compression
(densification) apparatus and briquette ejection apparatus used for this study, and the
briquetting procedure involved are given in Kaliyan and Morey (2006a) [Chapter 4]. A
brief description of the briquetting procedure follows. An INSTRON model 4206
universal testing machine (Instron Corporation, Canton, MA) was used to apply a
mechanical pressure of 150 MPa for briquetting the switchgrass grind. The inner
diameter and height of the steel cylinder were 18.8 mm and 300 mm, respectively. The
piston was made of brass and was 50-mm longer than the cylinder to help eject
briquettes from the cylinder. A separate apparatus was used to allow ejection of
briquettes from the cylinder after compression.
About 5.0-g of switchgrass grind or switchgrass grind mixed with a binder was
added to the cylinder through a funnel. Using a steel rod, the grind was stirred to help
the flow of grind from the funnel. Before compression of switchgrass grind to form
briquettes, the grind along with the cylinder was heated to achieve a preheating
temperature of 100C. During preheating, the top opening of the cylinder was sealed
with a specially designed vapor-tight cover. The outside of the compression cylinder
wall and bottom steel base of the cylinder were covered with heating tapes (volt = 120,
amps = 2.58, watts = 310, and phase = 1; BH Thermal Corporation, Columbus, OH) to

221
heat the biomass grind along with the cylinder. About 50-mm thick fiberglass
insulation was used to cover the heating tapes to reduce heat loss. Temperature of the
grind was measured inside of the cylinder at the center and about 50-mm above the
bottom of the cylinder. Immediately after reaching the grind temperature of 100C, the
power to the heating tapes was turned off, the cylinder was attached to the base of the
INSTRON, and the vapor-tight cover on the top of the cylinder was removed. The
piston attached to the crosshead of the INSTRON was actuated to compress the grind
to a maximum pressure of 150 MPa at a constant speed of 25.4 mm min"1 (1.0 in. min"
'). After completion of compression, the piston was taken out of the cylinder. Typical
compression curves are shown in figure 5.1.
After compaction of switchgrass grind into briquettes, the cylinder was
removed from the base of the INSTRON and placed on the briquette ejection
apparatus. A constant crosshead speed of 25.4 mm min"1 was applied to the piston to
eject the briquette from the cylinder. Figure 5.2 shows selected briquette ejection
curves. The time, force, and distance traveled by the crosshead were recorded during
the briquette compression and ejection to estimate the specific energy consumption for
briquette compression (i.e., compression energy) and ejection (i.e., ejection energy).
To calculate the specific energy consumption for compression or ejection, the area
under the force-displacement curve was estimated using the trapezoidal rule (Cheney
and Kincaid, 1985). In this study, the cylinder was not lubricated during briquetting
but periodically cleaned using a vacuum cleaner.
In most of the briquetting cases at 100C, it was observed that moisture was
deposited on the grind and at the bottom of the briquetting cylinder. Movement of
water from the grind either as vapor or liquid to the bottom of the sample was
probably a result of the inability to perfectly insulate the cylinder during preheating.
Also, the bottom of the cylinder was difficult to insulate during compression resulting
in lower temperature at the bottom of the cylinder, which caused moisture to migrate
and accumulate on the bottom of the sample. Due to the liquid water at the bottom of
the briquetting cylinder, about 0.5 to 1.0-g of grind at the bottom of the briquette was
not compressed well, and sometimes stayed as loose grind. These loose particles were

222
scraped off using a sharp knife, and the remaining part of the briquette was used for
the study. The loose particles from the briquettes were stored immediately in zip-lock
plastic bags, and their moisture contents were measured after one week of storage at
room temperature. It was observed that these loose particles lost moisture quickly after
removal from the briquettes. This may be due to their small particle size.

5.4-3 Briquetting Experiments

The densification conditions that resulted in the highest durability of


switchgrass briquettes (i.e., 67%) in the study by Kaliyan and Morey (2006a) [Chapter
4] were compression pressure of 150 MPa, particle size of 0.56 mm (i.e., grind
obtained from hammer mill with a screen size of 3.0 mm), grind moisture content of
10% (w.b.), and preheating temperature of 100C. For the current study, the above
densification conditions were used and kept constant for all of the briquetting cases. In
addition, the briquetting results from these conditions (without adding any binders)
served as control for comparison with the results from this study.
To study the effect of fine grinding on the switchgrass briquettes, briquetting
experiments were conducted using the grinds obtained from Thomas-Wiley mill with
two different screens having 1.0 and 2.0 mm openings. To study the effect of biomass
based binders, briquetting experiments were conducted using corn stover and
switchgrass grinds obtained from the hammer mill screen that had 3.0-mm openings.
The specific experiments were (i) switchgrass mixed with 10, 20, 40, 60, and 80%
(wt.) corn stover obtained from the hammer mill screen with 3.0-mm openings; (ii)
switchgrass mixed with 20% (wt.) corn DDGS; and (iii) switchgrass mixed with 2%
(wt.) starch powder. To study the effect of adding chemical binders, briquetting
experiments were conducted using switchgrass grind (obtained from the hammer mill
screen that had 3.0-mm openings) mixed with 2% (wt.) lime powder, 5% (wt.) lime
powder, 2% (wt.) lime paste, 2% lime powder plus sprinkled water, 2% (wt.) sodium
bentonite powder, 2% (wt.) Super-bind powder, 2% (wt.) Pel-stik powder, and 2%
(wt.) Xtra-bond powder.

223
In commercial pelleting, chemical binders are usually mixed in a steam
conditioning chamber where high temperature steam is sprayed at the same time when
the chemical binder powders are added to the feed. This would help the powdered
binders effectively adhere to the surface of feed particles, and the high temperature
activates the binders (Chapter 3; Kaliyan and Morey, 2006b). In this study, heat was
added to the grind by conduction through the wall of the compression cylinder, and
then conduction-convection inside the cylinder. Thus, before preheating, the
switchgrass grind and the predetermined quantity of biomass-based binders or
chemical binder powders were thoroughly mixed in a plastic container using a steel
rod, and then the mixture was filled into the compression cylinder.
Also, during steam conditioning in commercial pelleting, some amount of
water from the steam would be added to the grind. Therefore, we studied this steam
conditioning effect by mixing a small amount of distilled water in two cases with 2%
(wt.) lime. In the first case, lime and distilled water at a ratio of 1.0 g of lime powder:
2.0 g of distilled water was mixed first to make lime paste, and then the lime paste was
mixed to the switchgrass grind before filling the mixture into the compression
cylinder. In the second case, switchgrass grind, 2% (wt.) lime power, and a known
quantity of distilled water were simultaneously mixed in a plastic container with a help
of a steel rod before filling the mixture into the compression cylinder. The quantity of
distilled water added for the above case was based on a ratio of 1.0 g of lime powder:
2.0 g of distilled water. To ensure effective mixing, the lime powder was added to the
grind in small amounts at different small time intervals, when the distilled water was
sprinkled on the grind in small quantities.
Although the review of literature by Kaliyan and Morey (2006b) [Chapter 3]
found that preheating/steam conditioning is essential to activate the binding agents, we
also studied the effect of adding 2% (wt.) Super-bind with the switchgrass grind
(obtained from the hammer mill with the screen size of 3.0 mm) without preheating to
100C (i.e., briquettes were made at room temperature of about 25C). At each
briquetting condition, ten briquettes were made.

224
5.4-4 Briquette Properties

Immediately after ejection from the die, unit density of the briquettes (i.e.,
density of individual briquettes) was measured. Then, the briquettes were transferred
to zip-lock plastic bags and stored for one week at room temperature. Durability;
percentage expansion in the axial direction, radial direction, and volume; and moisture
content of briquettes were measured after one week of storage.
Durability of briquettes was measured according to ASAE Standard S269.4
(ASAE Standards, 2003c). Durability was calculated as the percentage of the original
mass of the briquettes retained on a 16.0-mm screen after tumbling in a durability
tester (Continental-Agra Equipment, Inc., Newton, KS) at 50 rpm for 10 min. Only
two replications were used for the durability measurement because of lack of samples.
For each replication, five briquettes were used. Unit density of briquettes was
calculated from the mass, diameter, and height of the briquettes (ASAE Standards,
2003c). Axial, radial, and volume expansions of briquettes were calculated as the
percentage increase in height, diameter, and volume, respectively. Moisture content of
the briquettes was measured based on ASAE Standard S358.2 (ASAE Standards,
2003c).
Statistical analyses (independent samples T-tests at a = 0.05) using the
software SPSS 16.0 for Windows (SPSS Inc., Chicago, IL) were conducted to
investigate the significant difference between the means of unit briquette density or
durability values measured for the "control briquetting conditions" and for the
individual briquetting condition that was considered as a strategy to improve the
durability of the switchgrass briquettes (i.e., treatments) (table 5.4). The control
briquetting conditions were 100% (by wt.) switchgrass, compression pressure of 150
MPa, particle size of 0.56 mm, nominal grind moisture content of 10% (w.b.), and
preheating temperature of 100C.

225
5.5 Results and Discussion

5.5-1 Effect of Fine Grinding

Table 5.1 provides the three particle sizes of the switchgrass grind tested along
with their bulk density values. With preheating to 100C at 150 MPa, decreasing the
geometric mean particle diameter of the switchgrass grind from 0.64 mm to 0.56 mm
increased the durability of switchgrass briquettes from 63.8 to 67.3% (Chapter 4;
Kaliyan and Morey, 2006a). Further decrease in the particle size from 0.56 to 0.38 mm
increased the briquette density by 49 kg m" (which was not significant at the 5%
level), but there was no improvement in the briquette durability (table 5.4) (P > 0.05).
Decreasing the geometric mean particle diameter by about half the size of the control
(i.e., decreasing the particle size from 0.56 to 0.26 mm) increased the briquette density
from 1054 to 1136 kg m"3 (significant at 5% level), and briquette durability from 67.3
to 74.1%; however, the difference in the durability values was not significant at 5%
level (table 5.4). Although there was a modest numerical increase in the briquette
durability, the cost of fine grinding to a particle size of 0.26 mm may be a limiting
factor for using the strategy of fine grinding the switchgrass to increase the briquette
durability.

5.5-2 Effect ofBiomass Based Binders

Some components in the biomass materials such as lignin, protein, starch, fat,
and water soluble carbohydrates can act as binders in situ (Chapter 3; Kaliyan and
Morey, 2006b). Table 5.2 provides the compositions of the switchgrass, corn stover,
and corn DDGS used for the study. Because corn stover had briquette durability of
96.6%o, about 10 to 80% (wt.) of corn stover was mixed with switchgrass to determine
its effect on switchgrass briquette durability. It was found that increasing the
percentage of corn stover mixed increased briquette durability (table 5.4). With 20%o
(wt.) corn stover, the switchgrass briquette durability was about 11%) points higher

226
than that of the control (significant at the 5% probability level). With 80% (wt.) corn
stover, the switchgrass briquette durability was about 22% points higher than that of
the control (significant at the 5% level). When biomass materials are used as binders,
usually the amount of biomass materials added to the feedstock is limited to about
20% (wt.) for economic and environmental (e.g., limitation on the ash content of the
briquettes produced) reasons (Chapter 3; Kaliyan and Morey, 2006b).
Mixing 20% (wt.) corn DDGS with switchgrass resulted in about 6% points
decrease in the durability of switchgrass (P > 0.05). The high percentages of starch
(5.8% d.b.) and protein (30.7%o d.b.) in the DDGS should have helped improve the
durability of the switchgrass briquettes. But, because of high fat content (13.6% d.b.)
in the DDGS, the binding effects of starch and protein may have been decreased.
According to Kaliyan and Morey (2006b) [Chapter 3], the amount of fat/oil in the
feedstock should be less than 6.5% (d.b.) to produce highly durable briquettes.
Therefore, defatted DDGS might help improve the durability of feedstock and this
should be considered in the future.
Adding 2% (wt.) starch did not improve the switchgrass briquette durability (P
> 0.05); however, starch resulted in higher briquette density than that of the control
(not significant at 5% level) (table 5.4). Mixing biomass-based binders with the
feedstock may not be a problem if the briquettes are used for
combustion/gasification/pyrolysis. If the briquettes are used for biochemical-based
conversions (e.g., ethanol production), the downstream processes may be affected by
different biomass materials added as binders with the feedstock.

227
5.5-3 Effect of Chemical Binders

For switchgrass grind with a nominal moisture content of 10% (w.b.),


densification at a preheating temperature of 100C, pressure of 150 MPa, and
inclusion of either 2% (wt.) lime powder or 2% (wt.) sodium bentonite powder did not
increase switchgrass briquette durability compared to that of the control (table 5.4) (P
> 0.05). Addition of 5% (wt.) lime increased the durability of the briquettes from 67.3
to 76.2%, and this increase in durability was found to be statistically significant at the
5% level. Mixing of water with the switchgrass grind either in the form of lime paste
or sprinkled water along with 2% (wt.) lime powder was found to reduce the briquette
durability compared to the case where 2% (wt.) lime was added as powder. This may
be due to the availability of free water between particles during compression, which
may have reduced the compressibility of the particles. Also, addition of water along
with lime resulted in higher volume expansion of the briquettes than for the case with
no water addition along with lime (table 5.5).
Mixing of 2% (wt.) Super-bind or Pel-stik or Xtra-bond powder with the
switchgrass grind resulted in improved durability (73.2 to 76.0%). However, the
increase in durability for Super-bind or Pel-stik or Xtra-bond compared to the control
was not significant at the 5% probability level. This statistical conclusion should be
tested with more replications in the future because only two replications were made
for testing the briquette durability in this study. Comparing briquette durability and
volume expansion values, Super-bind performed slightly better than Pel-stik or Xtra-
bond. The durability of briquettes for inclusion of 2% (wt.) Super-bind powder or 5%
(wt.) lime powder was about 76%.
Table 5.3 suggests that addition of 2% (wt.) of chemical binders to increase
switchgrass briquette durability would involve additional cost for the densification due
to the binder cost per tonne of briquettes produced. On the other hand, addition of 20%
(wt.) corn stover to increase the switchgrass briquette durability would involve no
additional cost because an equal amount of biomass with similar cost and heat energy
is substituted in the feedstock to make good quality briquettes (table 5.3).

228
The chemical binders added to the feedstock could increase the net ash content
of the feedstock; however, addition of biomass-based binders may not increase the net
ash content of the feedstock. Even if there is an increase in ash content of the
feedstock due to the added biomass-based binders, the increase in ash content in the
feedstock should be a small amount because an equal amount of biomass with similar
chemical composition is replaced for making the briquettes. Adding lime to the
feedstock may be beneficial because lime could control sulfur dioxide produced
during combustion/gasification of the biomass briquettes (Morey et al., 2006). Also,
lime would improve the enzymic hydrolysis of biomass materials used for producing
ethanol (Kaar and Holtzapple, 2000). Adding Super-bind or Pel-stik with the feedstock
may be a problem because this would increase the net sulfur content of the feedstock.

5.5-4 Strategy to Improve Durability of Switchgrass Briquettes

Without preheating and without addition of any binding agents, the durability
of switchgrass briquettes was zero percent (Chapter 4; Kaliyan and Morey, 2006a).
With preheating to 100C, the durability of switchgrass briquettes increased from 0%
to 67.3%. This dramatic improvement in briquette durability may probably be due to
the activation of "natural binders" in the switchgrass such as lignin, starch, protein, fat,
and water soluble carbohydrates by the preheating process (table 5.2). Increasing the
preheating temperature from 100 to 150 or 200C decreased briquette durability to
65.7% (at 150C) or 54.0% (at 200C) [Chapter 4]. This is probably due to the loss of
moisture from the grind due to preheating to more than 100C. The moisture contents
of the briquettes (measured after one week of storage of briquettes in zip-lock plastic
bags at room temperature) were 6.1, 3.7, and 2.8% (w.b.) for preheating at 100, 150,
and 200C, respectively. Therefore, during compression, the moisture content of the
switchgrass grind should be at least 10% (w.b.) to produce highly durable briquettes.
With preheating to 100C, addition of either 20% (wt.) corn stover or 2% (wt.)
Super-bind produced briquettes with about 10% points higher durability than that of
the control. Without preheating, inclusion of 2% (wt.) Super-bind powder produced

229
switchgrass briquettes with durability of 39.2%. This suggests that preheating is a
more dominant factor influencing the durability of switchgrass briquettes than addition
of chemical binders at 2% (wt.) inclusion rate. This result also indicates that
preheating is important to enhance the functionality of the added binders.
Addition of 20% (wt.) of corn stover with switchgrass to improve the
durability of briquettes from about 67% to about 80% may be an attractive strategy
because there would be no additional cost involved for the densification, assuming that
the cost of switchgrass and corn stover are almost the same. On the other hand, adding
2% (wt.) of chemical binders such as Super-bind with the switchgrass would also
improve the durability of briquettes from about 67% to about 80%; however, this may
not be a good strategy considering the added cost for densification due to the binder
cost per tonne of briquettes produced.
As discussed previously, it was difficult to control the migration of moisture
from the switchgrass grind due to preheating to 100C in the uniaxial compression
apparatus used for the study. Table 5.5 shows that there existed two separate moisture
zones inside the compression cylinder after preheating. The high moisture grind at the
bottom of the cylinder was not compressed very well and sometimes remained as
loose particles. Therefore, future work is required to make some modifications to the
densification apparatus to reduce/avoid moisture movement from the biomass grind
when preheating to >100C. This moisture movement effect may have affected the
durability of switchgrass briquettes to some extent.
Colley et al. (2006) and Kaliyan and Morey (2007) [Chapter 6] studied the
pelleting characteristics of switchgrass. Colley et al. (2006) produced switchgrass
pellets with durability of 78.4 to 96.7% using switchgrass grind having a geometric
mean diameter of 0.89 mm (i.e., grind obtained from a hammer mill with a screen size
of 3.2 mm). The temperature of the pellets was 85C. The durability was found to be
affected by change in moisture content of the pellets during storage (tested pellet
moisture content range was 6.3 to 17.0%) w.b.). Kaliyan and Morey (2007) [Chapter 6]
made switchgrass pellets with durability of 72.6 to 86.5% using switchgrass grind
having a geometric mean diameter of 0.49 mm (i.e., grind obtained from a hammer

230
mill with a screen size of 2.4 mm). The temperature of the pellets ranged from 70 to
81C and the moisture content of the pellets ranged from 10.7 to 12.2% (w.b.).
Results from the laboratory densification studies and the pelleting studies
suggest that highly durable pellets/briquettes (>80% durability) could be produced
from the switchgrass at suitable particle size (0.49 to 0.89 mm), moisture content
(>10% w.b.), and preheating temperature (>75C). The glass transition temperature of
the switchgrass is about 75C (Chapter 4; Kaliyan and Morey, 2006a), and thus, the
preheating temperature should be >75C to soften the natural binders in the
switchgrass. It is concluded that the ultimate strategy to improve the durability of
switchgrass briquettes is to fully activate the natural binders in the switchgrass using
the right particle size, moisture content, and preheating temperature.
Furthermore, the results from this laboratory study on binders could be used to
understand the binding potential of various binders tested and to identify the best
binders for other biomass feedstocks similar to switchgrass when pellet/briquette
durability needs to be increased.

5.5-5 Implications of Briquette Ejection Curves

During uniaxial compression of powders, both elastic and plastic strains are
induced within the powder compact. Upon unloading the applied stress, a small elastic
component is partially relieved, but a part remains in the compact. The remaining
elastic strains and the plastic strains cause a residual stress normal to the die (i.e.,
densification cylinder) wall. As a result, the compact is held inside the cylinder
because of the residual stress dependent shear stress at the wall. Therefore, the
compact should be ejected forcibly. The ejection stress required (for lubricated or
unlubricated, wall-compact frictional condition) was found to be affected by the
applied compaction pressure and aspect ratio of the compact (Briscoe and Rough,
1998). The ejection stress also was found to be independent of the ejection speed
(Briscoe and Rough, 1998). Briscoe and Rough (1998) observed that ejection stress
for spray dried alumina compact was considerably higher for the unlubricated die than

231
for the die lubricated with zinc stearate powder. They also observed a "stick-slip" type
of mechanism in the plot of ejection stress versus displacement for the unlubricated
die. This "stick-slip" behavior was attributed to the stress relaxation induced by the
higher levels of traction produced in the unlubricated compact systems.
In this study, the densification cylinder was not lubricated. During ejection of
switchgrass briquettes, the applied stress sharply reached a maximum value (i.e., static
ejection stress) after which the briquette started to emerge out of the die. Then, there
was almost a linear decrease in the ejection stress (i.e., dynamic ejection stress) with
the reduction in the briquette-cylinder wall contact area (fig. 5.2). This linear decrease
in ejection stress was observed for switchgrass briquettes made with 100% switchgrass
grinds (particle sizes of 0.26, 0.38, and 0.56 mm), mixtures of switchgrass and corn
stover grinds, 2% (wt.) lime with and without added water, and 2% (wt.) sodium
bentonite. Switchgrass briquettes required lower static ejection pressure (about 1.0
MPa) than corn stover briquettes (about 2.5 MPa). This may be related to the stronger
particle-particle bonds created inside the briquette matrix, and higher adhesion and
bonding at the wall created by corn stover than by switchgrass. The particle-particle
and particle-wall bonding may have been formed by the "natural binders". The natural
binding components such as protein and water soluble carbohydrates from the corn
stover or switchgrass could be squeezed out of cells during compression and then
hardened before ejection from the die.
The chemical binders, Super-bind, Pel-stik, and Xtra-bond, showed "stick-slip"
type of ejection behavior (fig. 5.2C). The variations in the ejection stress during
"stick-slip" behavior were larger for Super-bind than for Pel-stik and Xtra-bond.
However, the mean ejection stress (or ejection energy) for Pel-stik was similar to that
for Super-bind (table 5.6). This may be due to the fact that both Pel-stik and Super-
bind were formulated from lignin-sulfonate (table 5.3). Figure 5.3 compares the
ejection curves for Super-bind obtained for conditions with (i.e., 100C) and without
(i.e., 25C) preheating. For the case with preheating, there was "stick-slip" type of
behavior in the ejection stress curve, whereas without preheating, this behavior was
absent. We believe that preheating at 100C could have activated the Super-bind

232
binder much better than at 25C, which could have helped to make stronger particle-
particle bonding inside the briquette, and particle-wall bonding. During ejection of
briquettes made with Super-bind preheated at 100C, after overcoming the static
ejection stress, the dynamic ejection stress appeared to cycle (i.e., fluctuate) with time.
The cyclic behavior in the dynamic ejection stress may be due to the response of the
"intrinsic solid matrix system created by the particle-particle bonding in the briquette"
to the relaxation of the residual radial stress acting normal to the die wall during
ejection at a given ejection speed (i.e., 25.4 mm min"1). It may be concluded that if the
intrinsic solid matrix system created by the particle-particle bonding in the briquette is
stronger and more durable, the briquette can store and maintain higher residual radial
stress, which may require higher static and dynamic ejection stresses.
The briquette ejection curves showed that the cohesion and adhesion
conditions inside the briquette matrix and at the wall could be affected by the type of
feedstock and the type of binders used. It should be noted that the chemical binders (or
the densification conditions) that had the "stick-slip" behavior produced higher
briquette durability than the chemical binders (or the densification conditions) that did
not have this behavior. The energy required for ejection of briquettes ranged from 0.1
to 1.0 MJ/t (table 5.6). The lower ejection energy values were obtained for the
cohesion and adhesion conditions that could have been developed by water or oil (e.g.,
lime paste and corn DDGS cases). The higher ejection energy values were obtained
for the cohesion and adhesion conditions that could have been developed by the
natural binders or chemical binders (e.g., 100% corn stover and Super-bind cases).
When densifying biomass feedstocks with binders that have "stick-slip" ejection
behavior in extrusion type of densification equipment such as pelleting mills, the
energy consumption for the densification process would be expected to increase, at the
same time the quality (durability) of the densified products would be increased.

233
5.6 Conclusions

The following conclusions were drawn from this study.

1. Preheating (>75C) is the primary factor influencing switchgrass briquette


durability.

2. With preheating to 100C at 150 MPa pressure, either fine grinding of switchgrass
to a particle size of 0.26 mm, or addition of 5% (wt.) lime powder, or 2% (wt.)
chemical binders derived from lignin-sulfonate, or mixing 20% (wt.) corn stover
increased the briquette durability by about 10% points compared to the
densification condition of preheating to 100C without addition of any binders.
Although this is a modest increase in briquette durability, mixing of 20% (wt.)
corn stover with 80% (wt.) switchgrass appears to be the best strategy to improve
the switchgrass briquette durability because addition of 20% (wt.) corn stover
would involve almost no additional cost for densification compared to the
strategies of fine grinding of switchgrass to a particle size of 0.26 mm, and
addition of 2 to 5% (wt.) of chemical binders.

3. At the optimum particle size, moisture content, and preheating temperature (either
by steam conditioning or by shear/friction), switchgrass pellets/briquettes with
durability of >80% could be produced without adding any binding agents in
commercial densification equipment such as a pelleting machine or a roll press
briquetting machine.

4. By analyzing the briquette ejection curves, binding performance of different


commercial binders could be compared.

234
5.7 References

Adapa, P.K., G.J. Schoenau, L.G. Tabil, S. Sokhansanj, and B. Crerar. 2003. Pelleting
of fractionated alfalfa products. ASAE Paper No. 036069. St. Joseph, Mich.:
ASABE.
ASAE Standards. 2003a. S358.2: Moisture measurement - Forages. St. Joseph, Mich.:
ASABE.
ASAE Standards. 2003b. S319.3: Method of determining and expressing fineness of
feed materials by sieving. St. Joseph, Mich.: ASABE.
ASAE Standards. 2003c. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: ASABE.
Briscoe, B.J., and S.L. Rough. 1998. The effects of wall friction on the ejection of
pressed ceramic parts. Pow. Technol. 99(3): 228-233.

Cheney, W., and D. Kincaid. 1985. Numerical Mathematics and Computing.


Monterey, CA: Brooks/Cole Publishing Company.
Colley, Z., O.O. Fasina, D. Bransby, and C.W. Wood. 2005. Compaction behavior of
poultry litter and switchgrass. ASAE Paper No. 056053. St. Joseph, Mich.:
ASABE.
Colley, Z., O.O. Fasiba, D. Bransby, and Y.Y. Lee. 2006. Moisture effect on the
physical characteristics of switchgrass pellets. Trans. ASABE 49(6): 1845-
1851.

DOE. 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The
Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.
Department of Energy (DOE), Office of Scientific and Technical Information.
Available at: http://www/osti.gov/bridge. Accessed 18 June 2005.
Jannasch, R., Y. Quan, and R. Samson. 2004. A Process and Energy Analysis of
Pelletizing Switchgrass. Ste. Anne de Bellevue, QC: Resource Efficient

235
Agricultural Production (REAP-Canada). Available at: http://www.reap-
canada.com/library.htm. Accessed 27 August 2004.
Kaar, W.E., and M.T. Holtzapple. 2000. Using lime pretreatment to facilitate the
enzymatic hydrolysis of corn stover. Biomass and Bioenergy 18(3): 189-199.

Kaliyan, N., and R.V. Morey. 2006a. Densification characteristics of corn stover and
switchgrass. ASABE Paper No. 066174. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2006b. Factors affecting strength and durability of
densified products. ASABE Paper No. 066077. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2007. Roll press briquetting of corn stover and
switchgrass: a pilot scale continuous briquetting study. ASABE Paper No.
076044. St. Joseph, Mich.: ASABE.
Lemus, R., E.C. Brummer, K.J. Moore, N.E. Molstad, C.L. Burras, and M.F. Barker.
2002. Biomass yield and quality of 20 switchgrass populations in southern
Iowa, USA. Biomass and Bioenergy 23(6): 433-442.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004a. Evaluation of compaction equations
applied to four biomass species. Can. Biosys. Eng. 46(3): 3.55-3.61.

Mani, S., L.G. Tabil, and S. Sokhansanj. 2004b. Grinding performance and physical
properties of wheat and barley straws, corn stover and switchgrass. Biomass
and Bioenergy 27(4): 339-352.
McLaughlin, S.B., and L.A. Kszos. 2005. Development of switchgrass (Panicum
virgatum) as a bioenergy feedstock in the United States. Biomass and
Bioenergy 28(6): 515-535.
Morey, R.V., D.L. Hatfield, R. Sears, and D.G. Tiffany. 2006. Characterization of feed
streams and emissions from biomass gasification/combustion at fuel ethanol
plants. ASABE Paper No. 064180. St. Joseph, Mich.: ASABE.
ORNL. 2005. Biofuels from Switchgrass: Greener Energy Pastures. Oak Ridge, TN:
Oak Ridge National Laboratory (ORNL). Available at:
http://bioenergy.ornl.gov/papers/misc/switgrs.html. Accessed 18 June 2005.

236
Samson, R., P. Duxbury, M, Drisdelle, and C. Lapointe. 2000. Assessment of
Pelletized Biofuels. Ste. Anne de Bellevue, QC: Resource Efficient
Agricultural Production (REAP-Canada). Available at: http://www.reap-
canada.com/library.htm. Accessed 27 August 2004.
Sanderson, M.A., R.P. Egg, and A.E. Wiselogel. 1997. Biomass losses during harvest
and storage of switchgrass. Biomass and Bioenergy 12(2): 107-114.
Shinners, K.J., G.C. Boettcher, R.E. Muck, P.J. Wiemer, and M.D. Casler. 2006.
Drying, harvesting and storage characteristics of perennial grasses as biomass
feedstocks. ASABE Paper No. 061012. St. Joseph, Mich.: ASABE.
Tabil, L.G., Jr., S. Sokhansanj, and R.T. Tyler. 1997. Performance of different binders
during alfalfa pelleting. Can. Agric. Eng. 39(1): 17-23.

237
200
100% switchgrass (particle size
of 0.56 mm)

150
-20% corn stover plus 80%
(0 switchgrass
Q.

a 100 2% Super-bind
3
to
tn

-100% switchgrass (particle size
50 of 0.26 mm)

0
0 50 100 150 200
Time (s)

ure 5.1. Compression curve obtained while forming switchgrass briquettes. Compression pressure of 150 MPa was reached earlier
at the particle size of 0.26 mm than at the particle size of 0.56 mm. This is because the bulk density of the switchgrass
grind for the particle size of 0.26 mm was higher than for the particle size of 0.56 mm (table 5.1).

238
A. Fine grinding. Particle size == 0.56mm
2.5 -+- Particle size = 0.38mm

w 2 Particle size : = 0.26mm


Q.

1.5
3
W
(0

0.5

10 15 20 25 30 35
Time (s)

B. Biomass based binders. 1 0 0 % switchgrass

2.5 , 100% corn stover

-*-20% corn stover plus


TO 2 - 80% switchgrass
Q.
- * - 20% corn DDGS plus
2 80% switchgrass
1.5 2% starch powder
3
V)
V)

1 1

0.5 j ~~~ ~ ^ ^ L
^ > l > > n>i

H|JK 1 1 I I ^^r^^*"^^*' i

10 15 20 25 30 35
Time (s)

ure 5.2. Briquette ejection curves (A, B, and C) for switchgrass briquettes made
using binding agents at different inclusion rates (% wt), and at a
compression pressure of 150 MPa and 100C.

239
3

0 5 10 15 20 25 30 35
Time (s)
Figure 5.2. Briquette ejection curves (A, B, and C) for switchgrass briquettes made using binding agents at different inclusion rates (%
wt), and at a compression pressure of 150 MPa and 100C.

240
3

0 10 20 30 40
Time (s)

Figure 5.3. Effect of temperature on the briquette ejection pressure for Super-bind
binder mixed with switchgrass.

Table 5.1. Properties of switchgrass and corn stover grinds.

Grinder and screen Grind moisture content Particle size Bulk density
size * (% w.b.) (n = 3) ** (n = 3) *** (kg m 3 ) (n = 3) **
Switchgrass
Thomas-Wiley milled 10.8 + 0.3 0.26 0.24 246.3 1 0.6
with a screen size of
1.0 mm
Thomas-Wiley milled 10.40.5 0.3810.27 205.212.3
with a screen size of
2.0 mm
Hammer milled with a 9.710.2 0.56 1 0.29 200.513.6
screen size of 3.0 mm
Corn Stover
Hammer milled with a 10.1 0.3 0.66 + 0.32 130.612.0
screen size of 3.0 mm
* The screen size represents the diameter of the holes on the screen.
** Mean 1 standard deviation.
*** Geometric mean particle diameter (mm) 1 geometric standard deviation.

241
Table 5.2. Compositions of corn stover, switchgrass, and corn Distillers Dried Grains
with Solubles (DDGS).

Component Corn stover Switchgrass Corn DDGS


(% of dry matter) (% of dry (% of dry matter)
matter)
Dry matter (%) 94.90 94.60 90.40
Acid detergent fiber 58.20 53.00 15.60
(ADF)
Neutral detergent fiber 84.40 81.80 30.30
(NDF)
Lignin ' 8.80 9.20 5.60
Cellulose2 49.40 43.80 10.00
3
Hemicellulose 26.20 28.80 14.70
Non-fiber carbohydrates 2.10 10.70 31.30
Water soluble 7.90 2.20 -
carbohydrates
Ethanol soluble 1.20 1.20 -
carbohydrates
Crude protein 3.60 3.90 30.70
Starch 0.40 1.00 5.80
Crude fat 0.70 0.90 13.60
Ash 11.23 4.99 5.88
Calcium 0.34 0.37 0.06
Phosphorus 0.05 0.08 1.03
Magnesium 0.13 0.11 0.31
Potassium 0.84 0.23 1.05
Sulfur 0.06 0.06 0.77

Acid insoluble lignin values.


2
Cellulose = ADF-lignin.
3
Hemicellulose = NDF - ADF.

242
Table 5.3. Cost of the binders used for the study.

Trade Ingredients Bulk density Cost of the Binder Cost of


name of of the binder binder ($/kg inclusion binder
3
the binder (kg m" ) of binder) rate (% wt.) per tonne
of
briquettes
($)
Biomass based binders
Corn stover Table 5.2 Table 5.1 0.06 to 0.08 20 12 to 16*
(assumed)
Corn Table 5.2 585.8 11.8 0.08 to 0.1 20 16 to 20
DDGS (MC = 9.8 (assumed) **
0.2% w.b.)
Starch Food starch, 650.3 084 2 16.8
modified

243
Table 5.3. Continued.
Trade Ingredients Bulk density Cost of the Binder Cost of
name of of the binder binder ($/kg inclusion binder
the binder (kg m"3) of binder) rate (% wt.) per tonne
of
briquettes
1&
Chemical binders ***
Lime Calcium Not available Not available 2 to 5 Not
hydroxide 96% available
(lab grade)
Sodium Sodium bentonite Not available 0.15 2 3.0
bentonite
Super- Condensed 480.6 0.57 to 0.62 2 11.4 to
bind lignin-sulfonate 12.4
Pel-stik Condensed 480.6 to 496.6 1.21 2 24.2
lignin-sulfonate,
propylene glycol,
sodium acetate,
and sodium
sulfate
Xtra-bond Urea 784.9 1.87 2 37.4
formaldehyde
condensation
polymer, and
calcium sulfate
* Assuming that the cost of corn stover and switchgrass are about the same, if 20% (wt.) corn stover is
added with 80% (wt.) switchgrass to increase the briquette durability, the added cost for the
densification due to the cost of binder per tonne of briquettes produced would be about zero because
20% (wt.) switchgrass is replaced in the feedstock by 20% (wt.) corn stover. Also, the heat energy of
the mixture of 20% (wt.) corn stover and 80% (wt.) switchgrass would be about the same as that of
the 100%o (wt.) switchgrass.

** Assuming that the cost of corn DDGS is slightly higher than that of switchgrass, if 20% (wt.) corn
DDGS is mixed with 80% (wt.) switchgrass to increase the briquette durability, the small difference in
the costs is expended as the cost of binder per tonne of briquettes produced because 20% (wt.) of
switchgrass is replaced in the feedstock by 20% (wt.) corn DDGS. Also, the heat energy of the mixture
of 20% (wt.) corn DDGS and 80% (wt.) switchgrass would be slightly higher than that of the 100%>
switchgrass because of the higher heat content of corn DDGS (Morey et al., 2006).

*** The data for starch, sodium bentonite, Super-bind, Pel-stik, and Xtra-bond were provided by
Uniscope Inc., Johnstown, CO. Adding 2 to 5% (wt.) of chemical binders would involve additional cost
for the densification due to the cost of binder per tonne of briquettes produced. This is because the cost
of 2 to 5% (wt.) of chemical binders would be higher than the cost of 2 to 5% (wt.) of switchgrass.
Also, the 2 to 5% (wt.) of chemical binders added may not substitute for the heat energy lost due to the
reduction of 2 to 5% (wt.) of switchgrass in the briquettes.

244
Table 5.4. Properties of switchgrass briquettes (maximum compression pressure = 150
MPa; preheating temperature = 100C).

Briquetting; conditions Immediately After one week of


after ejection storage at room
from the die temperature
Unit briquette Unit Durability
density briquette of
Binder type/condition Particle size Grind (kg m"3) density briquettes
of moisture (n = 5) ** (kg m"3) (%)
switchgrass content (n = 5) (n = 2)
(mm) * **,A
(% w.b.)
0 = 3)
**.

Fine grinding
Thomas-Wiley milled with 0.26 0.24 10.8 0.3 1135.03.5 1135.8 74.1
1.0 mm screen 10.2 b 3.3 a
Thomas-Wiley milled with 0.380.27 10.4 0.5 1128.6 + 1102.0 66.0 +
2.0 mm screen 12.2 12.9 a 4.2 a
Biomass based binders (added as % wt.)
100% switchgrass (control) 0.56 0.29 9.8 0.2 1099.7 1053.5 67.3 +
12.5 61.7 a 1.5 a
100% corn stover - 10.1 0.3 1197.6 + 1161.7 96.6 +
14.9 25.6 b 0.04 b
10% corn stover+ 90% 0.56 0.29 10.4 0.1 1096.8 1058.7 68.3 +
switchgrass & 14.6 29.2 a 0.8 a
20% corn stover + 80% 0.56 0.29 10.4 0.1 1135.5 1096.5 77.8 +
switchgrass & 21.9 20.8 a 2.1 b
40% corn stover + 60% 0.56 0.29 10.4 0.1 1124.6 1088.4 78.6 +
switchgrass & 10.0 31.4a 1.4 b
60% corn stover + 40% 0.56 0.29 10.4 0.1 1136.0 1098.3 81.9 +
switchgrass & 15.2 47.1 a 2.3 b
80% corn stover + 20% 0.56 0.29 10.4 0.1 1163.6 7.1 1139.6 + 89.3 +
switchgrass & 15.4 b 0.6 b
20% corn DDGS + 80% 0.56 0.29 10.4 + 0.1 1093.7 8.9 1082.4 61.4 +
switchgrass & 3.8 a 0.8 a
2% starch added as powder 0.56 0.29 10.7 0.2 1119.5 1112.8 65.7 +
12.2 7.9 a 1.5 a

245
Table 5.4. Continued.

Briquetting iconditions Immediately After one week of


after ejection storage at room
from the die temperature
Unit briquette Unit Durability
density briquette of
Binder type/condition Particle size Grind (kg m"3) density briquettes
of moisture (n = 5) ** (kg m) (%)
switchgrass content (n = 5) (n = 2)
(mm)* (% w.b.) **.A **.A

(n = 3)

Chemical binders (added as % wt.)


2% lime added as powder 0.56 0.29 10.7 1087.6 1061.6 66.3
0.2 10.6 19.7 a 1.9 a
5% lime added as powder 0.56 0.29 10.4 1122.8 8.9 1105.8 76.2
0.1 14.0 a 1.0 b
2% lime added as paste 0.56 0.29 9.8 0.3 1057.8 1014.8 62.3
15.5 63.2 a 5.7 a
2% lime powder plus 0.56 0.29 12.3 1068.2 1038.9 61.5
sprinkled water 0.5 21.1 18.6 a 5.1a
2% sodium bentonite added 0.56 0.29 10.7 1126.4 9.8 1104.1 68.8
as powder 0.2 11.3a 1.1 a
2% Super-bind added as 0.56 0.29 10.7 1119.9 1092.7 76.0
powder v 0.2 14.8 24.2 a 3.6 a
2% Pel-stik added as powder 0.56 0.29 10.7 1133.2 1109.6 75.2
0.2 14.5 14.6 a 2.8 a
2% Xtra-bond added as 0.56 0.29 10.7 1122.1 4.9 1108.5 73.2
powder 0.2 9.2 a 0.4 a
* Geometric mean particle diameter geometric standard deviation (n = 3).
** Mean standard deviation.
#
Moisture content of the switchgrass grind before densification and preheating.
&
Moisture content of the corn stover grind, and corn DDGS was 10.2 0.1, and 9.8 0.2% (w.b.),
respectively.
w
When briquetting switchgrass with 2% (wt.) Super-bind powder without preheating (i.e., at room
temperature), the unit briquette density measured immediately after ejection from the die, unit briquette
density measured after one week of storage at room temperature, and durability of the briquettes were
967.9 8.0 kg m"3, 812.8 6.9 kg m"3, and 39.2 0.2%, respectively. The moisture content of the
switchgrass grind (0.56 0.29 mm) was 11.6 0.2% (w.b.).
A
When comparing the control (i.e., 100% switchgrass) with each individual densification condition
(i.e., treatment), the means of unit briquette density or durability followed by similar lowercase letters
for control and treatment are not significantly different from each other (P > 0.05). Significance
comparisons between treatments (e.g., 20% corn stover versus 40% corn stover) should not be made
because the statistical tests were conducted only to compare control versus individual treatment.

246
Table 5.5. Stability and moisture content of switchgrass briquettes measured after one
week of storage at room temperature (nominal grind moisture content =
10% (w.b.); maximum compression pressure = 150 MPa; preheating
temperature = 100C).

Briquetting conditions Briquette Moisture Meari expansion (%)


moisture content of the (n = 5)
content briquette
(% w.b.) particles Axial Radial Briquette
(n = 3 ) * (%w.b.)*'* direction direction volume

Fine |grinding
Thomas-Wiley milled with 1.0 7.2 0.1 7.3 ( n = l ) -0.5 -0.2 -0.8 **
mm screen
Thomas-Wiley milled with 2.0 7.9 0.2 8.6 ( n = l ) 1.3 0.3 1.8
mm screen
Biomass based binders (added as % wt.)
100% switchgrass (control) 6.1 0.4 10.4 (n= 1) 3.6 0.5 4.8
100% corn stover 7.0 0.2 - 2.6 0.1 2.7
10% corn stover+ 90% 7.3 0.3 8.9 ( n = l ) 2.9 0.2 3.3
switchgrass
20% corn stover + 80% 7.9 0.2 11.20.6(n 2.8 0.3 3.5
switchgrass = 2)
40% corn stover + 60% 6.7 0.1 9.7 1.1 (n = 3.1 0.0 3.1
switchgrass 2)
60% corn stover + 40% 7.8 0.2 9.3 ( n = 1) 2.7 0.1 3.0
switchgrass
80% corn stover + 20% 7.3 0.4 10.2 ( n = l ) 1.4 0.1 1.6
switchgrass
20% corn DDGS + 80% 8.0 0.2 11.2 (n=1) 0.8 0.01 0.9
switchgrass
2% starch added as powder 5.8 0.1 5.9 (n = l) -0.4 0.0 -0.4

247
Table 5.5. Continued.
Briquetting conditions Briquette Moisture Meai ii expansion (%)
moisture content of the (n = S)
content briquette
(% w.b.) particles Axial Radial Briquette
(n = 3)* (% w.b.) * * direction direction volume

Chemical binders (added as % 'Wt.)


2% lime added as powder 8.5 1.1 11.5 0.1 (n 2.2 -0.1 2.0
= 2)
5% lime added as powder 7.2 0.2 12.6 (n= 1) 0.3 0.5 1.3
2% lime added as paste 8.4 0.4 9.6 0.5 (n = 2.1 0.4 3.1
3)
2% lime powder plus sprinkled 8.4 0.2 10.8 0.5 (n 2.2 0.01 2.2
water = 2)
2% sodium bentonite added as 5.0 0.2 6.4 (n= 1) 1.2 0.1 1.5
powder
2% Super-bind added as powderl|' 5.4 0.1 5.60.2(n = 0.1 0.2 0.4
2)
2% Pel-stik added as powder 5.4 0.2 6.80.6(n = 1.1 0.2 1.5
2)
2% Xtra-bond added as powder 5.3 0.1 6.6 0.01 (n 0.1 0.2 0.6
= 2)
* Mean + standard deviation.
* Moisture content of the loose particles that were scraped off from the bottom of the briquettes
immediately after ejection of briquettes from the die.
** Particles lost from the bottom of the briquettes due to inadequate bonding which caused a decrease
in briquette length. Shrinkage of briquettes due to moisture loss from the briquettes may have caused a
decrease in briquette diameter.
V|
' When briquetting switchgrass with 2% (wt.) Super-bind powder without preheating (i.e., at room
temperature), the briquette moisture content was 10.4 0.1% (w.b.), and the mean expansions in axial
direction, radial direction, and volume were 16.1, 1.0, and 18.4%, respectively.

248
Table 5.6. Specific energy required for briquetting switchgrass (nominal grind
moisture content = 10% (w.b.); maximum compression pressure = 150
MPa; preheating temperature = 100C).
Briquetting conditions Compression Ejection energy Specific energy
energy (MJ/t) (MJ/t) consumption
(n = 10) * (n = 10) *
MJ/t* Percentage
of energy in
switchgrass
(%)*
Fine grinding
Thomas-Wiley milled with 1.0 mm 32.9 0.2 0.2 0.03 181] L08
screen
Thomas-Wiley milled with 2.0 mm 32.3 0.6 0.2 0.02 182.5 1.07
screen
Biomass based binders (added as % wt.)
100% switchgrass (control) 37.5 1.4 0.2 0.1 187.7 1.10
100% corn stover 38.6 1.1 1.0 0.2 189.6 1.12
10% corn stover + 90% switchgrass 36.5 0.6 0.3 0.1 186.8 1.10

20% corn stover + 80% switchgrass 36.5 1.2 0.4 0.1 186.9 1.10
40% corn stover + 60% swichgrass 33.7 1.6 0.3 0.04 184.0 1.08
60% corn stover + 40% swichgrass 34.9 1.5 0.2 0.01 185.1 1.09

80%) corn stover + 20% swichgrass 34.8 1.5 0.2 0.01 185.0 1.09

20% DDGS + 80%) swichgrass 30.8 0.5 0.1 0.02 180.9 1.06
2% starch added as powder 37.2 0.5 0.2 0.04 187.4 1.10

249
Table 5.6. Continued.

Briquetting conditions Compression Ejection energy Specific energy


energy (MJ/t) (MJ/f) consumption
(n = 10) * (n = 10) *
MJ/t* Percentage
of energy
in
switchgrass
(%) *
Chemical binders (added as % wt.)
2% lime added as powder 35.7 0.8 0.3 0.04 186.0 1.09
5% lime added as powder 37.1 1.1 0.3 0.1 187.4 1.10
2% lime added as paste 32.0 1.4 0.1 0.03 182.1 1.07
2% lime powder plus sprinkled water 30.9 1.7 0.2 0.04 181.1 1.07
2% sodium bentonite added as powder 39.0 0.9 0.3 0.1 189.3 1.11
2% Super-bind added as powder v|/ 40.5 1.5 0.8 0.3 191.3 1.13
2% Pel-stik added as powder 39.7 1.0 0.8 0.3 190.5 1.12

2% Xtra-bond added as powder 38.5 0.6 0.3 0.04 188.8 1.11


* Mean + standard deviation.
* Specific energy consumption (MJ/t) = Compression energy + ejection energy + preheating energy.
Preheating energy (MJ/t) = Specific heat of switchgrass x [preheating temp. - room temp.]
Assumed specific heat of switchgrass = 2.0 kJ kg"1 K"'; room temperature = 25C.
For all of the briquetting conditions, preheating energy = 150.0 MJ/t.
&
Percentage of energy in switchgrass (%) = Specific energy consumption (MJ/t) * 100 / Energy
content of switchgrass (MJ/t).
Energy content of switchgrass = 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004b).
v
When briquetting switchgrass with 2% (wt.) Super-bind powder without preheating (i.e., at room
temperature), the compression energy, ejection energy, specific energy consumption in MJ/t, and
specific energy consumption in percentage of energy in switchgrass were 39.6 0.4 MJ/t, 0.7 0.2
MJ/t, 40.3 MJ/t, and 0.24%, respectively.

250
Chapter 6
ROLL PRESS BRIQUETTING AND PELLETING - PILOT SCALE STUDY

6.1 Abstract

Experiments on briquetting of corn stover and switchgrass were conducted in a


pilot-scale roll press briquetting machine. Briquettes were of almond shape - 28.7 to
31.3 mm in length. The effects of briquetting machine variables such as screw feeder
and roll speeds, and biomass variables such as particle size, moisture content, and
steam conditioning temperature on the bulk density, durability, and crushing strength
of the briquettes were investigated. It was found that no steam conditioning was
necessary to produce good quality corn stover briquettes. For switchgrass, a grind
temperature of about 75C obtained by steam conditioning was necessary to produce
good quality briquettes. At the optimum roll press briquetting conditions determined,
corn stover and switchgrass briquettes with bulk density of 453 and 420 kg m"3,
durability of 88 and 70%, and crushing strength of 169 and 171 N, respectively, could
be produced. Briquettes produced with the roll press briquetting machine had bulk
densities, durabilities, and crushing strengths that were somewhat less than but in a
range comparable to the pellets (9.6- to 9.8- mm diameter) produced with a
conventional ring-die pelleting machine. Micro-structural studies (chemical
composition analyses, scanning electron microscopy imaging, and UV auto-
fluorescence imaging) on biomass grinds, pellets, and briquettes were conducted to
understand and optimize the densification behaviors of the corn stover and
switchgrass. Roll press briquetting appears to be a promising low-cost, low-energy,
high-capacity approach for commercial production of densified biomass.

6.2 Introduction

To reduce dependency on fossil fuels, there is tremendous interest in using


biomass materials such as corn stover and switchgrass in the U.S. for producing liquid

251
transportation fuels, combined heat and power, chemicals, and bio-products (DOE,
2005). In addition to numerous advantages, use of biomass materials in place of fossil
fuels would result in low emissions of greenhouse and acid gases. In order to make
biomass materials available for a variety of applications, challenges with the use of
biomass materials in their original form must be resolved. Because of high moisture
content, irregular shape and size, and low bulk density, biomass is very difficult to
handle, transport, store, and utilize in its original form. One solution to these problems
is densification of biomass materials into pellets or briquettes.
Densification increases the bulk density of biomass from an initial bulk density
(including baled density) of 40 to 200 kg m"3 to a final bulk density of 600 to 800 kg
m"3 (Holley, 1983; Mani et al., 2003; Obernberger and Thek, 2004; McMullen et al.,
2005). Densification of biomass materials could reduce the costs of transportation,
handling, and storage. In addition, because of uniform shape and size, densified
products can be more easily handled using existing handling and storage equipment,
and can be easily adopted in direct-combustion or co-firing with coal, gasification,
pyrolysis, and in other biomass-based conversions.
In the U.S., pelleting and roll press briquetting technologies are widely used
for densifying particulate materials for various end uses (Li and Liu, 2000; Pietsch,
2002). In pelleting equipment, the feed material is pressed through open-ended
cylindrical holes (dies) made in the periphery of a ring (fig. 6.1 A). One to three small
rotating rolls push the feed material into the die holes from inside of the ring towards
the outside of the ring. The skin friction between the feed particles and the wall of the
die resists the free flow of feed and thus the particles are compressed against each
other inside the die to form pellets. One or two adjustable knives placed outside the
ring cut the pellets into desired lengths. The diameter of the pellets may range from
4.8 to 19.0-mm, and the length of the pellets may range from 12.7 to 25.4 mm.
Pelleting technology is commonly used in the U.S. and Europe for producing animal
feed(Sitkei, 1986).

252
In a roll press briquetter/compactor, material is densified by compression
between two counter-rotating rolls (fig. 6. IB). Initial densification of the material may
occur through compressing the material with a tapered auger in the feed mechanism.
The primary purpose of this initial densification phase is to remove air from low bulk
density material. The final compaction occurs because of high pressures created as the
material flows between the two rolls. The roll surfaces have pockets to form briquettes
of desired size and shape when the material passes between the rolls. The briquettes
are easily separated and handled after leaving the machine. The densified products are
mostly of pillow-shape with a size of 10 to 40 mm or larger. The roll press is generally
used for densifying coals, fertilizers, minerals, and metals in the U.S. and Europe
(Pietsch, 2002).
We believe that these two densification approaches (pelleting and roll press
briquetting) would be readily adapted to densifying biomass materials to use for
renewable energy applications. Pelleting would produce densified biomass (i.e.,
pellets) suitable for home, business, and institutional heating where high bulk density
and high quality (i.e., high strength and durability) of densified biomass are required.
On the other hand, roll press briquetting would produce densified biomass (i.e.,
briquettes) suitable for applications requiring a large amount of biomass materials
such as combustion or gasification at ethanol plants where moderate bulk density and
moderate quality of densified biomass are sufficient to facilitate the transportation,
handling, and storage of biomass with minimum costs. Based on this rationale, we
have attempted to study the densification behaviors of biomass materials (i.e., corn
stover and switchgrass) in pilot-scale roll press briquetting and conventional pelleting
machines. More effort was placed on roll press briquetting than on pelleting because
very limited fundamental studies on roll press briquetting of biomass materials are
available in the literature (e.g., Koser et al., 1982; Holley, 1983; Dec, 2002).
Conversely, several studies on the performance of conventional pelleting of biomass
materials have been conducted in the past (e.g., Hill and Pulkinen, 1988; Tabil and
Sokhansanj, 1996; Briggs et al., 1999; Samson et al, 2000; Gilpin et al, 2002;
Jannasch et al., 2004; Colley et al., 2006; Opoku et al., 2007).

253
6.3 Objectives

The objectives of this study were to:

Study the roll press briquetting characteristics of corn stover and switchgrass.
Determine optimum roll press briquetting conditions (machine variables and
biomass variables) for densifying corn stover and switchgrass.
Compare performances of the roll press briquetting machine and a conventional
pelleting machine at selected densification conditions for corn stover and
switchgrass.
Investigate the binding mechanisms of corn stover and switchgrass by micro-
structural analyses.

6.4 Materials and Methods

6.4-1 Biomass Feedstock

Corn stover was purchased from Mat Ag-Fiber LLC (Floodwood, MN) in
October 2006 as coarsely ground (i.e., hammer milled) corn stover packed in plastic
bags each weighing 22.5 kg (50.0 lb). Further size reduction (i.e., grinding) of corn
stover was done at Agricultural Utilization Research Institute (AURI), Waseca to use
in the briquetting and pelleting experiments. Two particle sizes were obtained by
grinding the coarse corn stover in an 18.7-kW (25-HP) hammer mill (Jacobson Quality
Machinery, Minneapolis, MN; Motor speed = 1770 rpm; Volt = 480 volt; Current = 30
A; 3 phase) with two different screens with 2.4- and 4.0-mm (3/32- and 5/32-in.)
openings.
Switchgrass was harvested as 1.2 m x 1.2 m x 2.4 m (4 ft x 4 ft x 8 ft) square
bales in September 2006 from a field in Owatonna, MN. The bulk density of
switchgrass bales was measured at about 79.4 kg m"3. Switchgrass bales were
transported to AURI, Waseca for size reduction. Size reduction of switchgrass was

254
done in two steps. First, switchgrass bales were ground in a 5.6-kW (7.5-HP) tub
grinder (Agri-Val, DC Atlas CO., Loyal, WI) to a length of 101.6 to 127.0 mm (4 to 5
in.). Then, fine grinding of the tub-ground switchgrass was done using the hammer
mill with two different screens with 2.4- and 4.0-mm (3/32- and 5/32-in.) openings.
Moisture content of the biomass grinds was determined using the procedure
given in AS ABE Standard S3 5 8.2 (AS ABE Standards, 2003a). The moisture content
values reported in this paper are on a wet basis. Bulk density of the biomass grinds
was calculated from the mass of grind that occupied a 250-mL glass container (table
6.1). Particle size distribution of the biomass grinds was determined based on AS ABE
Standard S319.3 (ASABE Standards, 2003b) (table 6.1). Corn stover and switchgrass
grind samples were sent to a forage analysis laboratory (Dairy One, Ithaca, NY) to
determine the compositions of these two biomass materials using Near Infrared
Reflectance (NIK) spectroscopy (table 6.2).

6.4-2 Roll Press Briquetting

Roll press briquetting experiments were conducted during January 2007 at


Bepex International LLC (Minneapolis, MN) using a pilot-scale roll press briquetting
machine (Bepex International LLC, Minneapolis, MN; Model: CS-25
Compactor/Briquetter). Figure 6.2 shows the picture of the roll press briquetting
machine used for the study. The specifications of the pilot-scale roll press briquetting
machine used for the study are given in table 6.3.
The roll press briquetting process is depicted in fig. 6.3. First, the moisture
content of the biomass grind was adjusted by mixing a predetermined quantity of
distilled water with the grind in a mechanical mixer. Before briquetting, the biomass
grinds were left in plastic containers for about 2 to 3 h for moisture equilibration. For
the briquetting conditions where the temperature of the grind should be higher than
room temperature, the grind was steam conditioned using a Turbulizer (Bepex
International LLC, Minneapolis, MN) steam conditioner. The Turbulizer was a steam-
jacketed cylindrical container with mixing paddles at its center.

255
A feeder metered the moisture conditioned grind to an airlock. The airlock was
required to heat the grind to 100C; however, the airlock was not required to heat the
grind to 75C. From the airlock, the grind was fed to the inlet of the Turbulizer. In the
Turbulizer, paddles moved the grind from the inlet to the outlet while increasing the
temperature of the grind to the required levels (within 30 s). Steam at 0.35 to 0.83
MPag (50 to 120 psig) and 150 to 178C was input to the steam-jacket of the
Turbulizer to indirectly raise the temperature of the grind. The steam was generated in
a cylindrical heat exchanger before introducing it to the Turbulizer. The paddles in the
Turbulizer were operated at 1200 rpm to obtain a grind temperature of 75C or at 1650
to 1800 rpm to obtain 100C or above. At higher paddle rpm, the grind has greater
contact with the steam-jacketed wall of the Turbulizer resulting in higher grind
temperature. Also, some amount of steam at 0.10 to 0.14 MPag (15 to 25 psig) was
also sprayed directly inside the Turbulizer to avoid any possible moisture loss from the
grind during heating. The steam sprayed inside the Turbulizer was produced by
evaporating 7.6 to 22.7 L/h of water in another cylindrical heat exchanger. Samples
were collected from the steam conditioned grind to estimate the moisture content,
which served as the initial moisture content of the grind for the tests involving steam
conditioning.
The steam conditioned grind was fed to the hopper of the roll press where a
screw feeder moved the grind to the rolls. The grind was densified between the rolls to
form briquettes continuously. The briquettes were collected in a plastic container after
reaching a steady state production of briquettes (i.e., after about 1.0 to 2.0 min of
feeding the grind). Briquettes were allowed to naturally cool to room temperature
(about 20C). Figure 6.4 shows the pictures of corn stover and switchgrass briquettes.
During briquetting, some amount of fines, although agglomerated to some
extent, leaked through the rolls due to inadequate sealing between the end of the screw
feeder (i.e., hopper) and the rolls. The leaked fines were separated from the briquettes
by sieving in a vibratory screen with 5.2-mm openings. Then, properties of the
briquettes were immediately measured. Some briquette samples were stored in plastic

256
bags at room temperature (about 23C) to measure their properties after one week of
curing.
The temperature of the moisture conditioned or steam conditioned grind, and
the temperature of the briquettes immediately after coming out of the rolls were
measured using a hand-held K-type thermocouple-temperature sensor (DIGI-SENSE,
Model No. 91100-40; Cole-Parmer Instruments Co., Vernon Hills, IL). To estimate
the specific energy consumed for the roll press briquetting process, the current
consumption by the roll motor and the screw feeder motor was recorded using
ammeters. The specific energy consumption per unit mass of briquettes produced
(MJ/t), and the specific energy consumption per unit decrease in volume of the
biomass feedstock (MJ/m ) were estimated by subtracting the energy consumption for
running the empty roll press briquetting machine (i.e., no load condition) from the
energy consumption during biomass briquetting. To estimate the throughput (kg/h) of
the roll press briquetting machine, the mass of briquettes collected for about 4 to 10
min was weighed.

6.4-3 Pelleting

Pelleting experiments were conducted during February 2007 at AURI, Waseca


using a pilot-scale conventional ring-die pelleting machine (California Pellet Mill
[CPM] Co., San Francisco, CA; CPM Master Model No. 818806). Table 6.4 provides
the specifications of the pelleting machine used for the study. Before pelleting corn
stover and switchgrass, the ring-die was cleaned to remove rusts and any previous
plugging in the holes of the ring-die. To achieve effective cleaning of the ring-die, a
mixture of oat grain, fine sand, and vegetable oil was extruded through the ring-die for
about two to three minutes. This start-up procedure also preheated the ring-die. In
addition, the same mixture was used to clean the ring-die whenever there was a change
in the pelleting condition such as pelleting switchgrass after pelleting corn stover.
The pelleting process involves moisture conditioning the biomass grind, and
then feeding the moisture conditioned grind to the pelleting mill to make pellets. The

257
moisture content of the biomass grinds was adjusted by mixing predetermined
quantities of water with the biomass grinds in a mechanical mixer. The biomass grinds
were left in plastic containers for about 2 to 3 h for moisture equilibration before
pelleting. Then, the grind was fed to the feed hopper of the pelleting mill. A screw
conveyor located below the feed hopper directed the grind to the conditioning
chamber with paddles of the pelleting mill, and then the grind was fed directly to the
pelleting chamber where two corrugated rolls forced the grind into the holes in the
ring-die to make pellets. No steam conditioning was done either for corn stover or
switchgrass grind at the conditioning chamber for any of the pelleting tests. In the
pelleting mill, both the corrugated rolls and the ring-die rotated in a clockwise
direction.
A knife fixed at about 25.4 mm (1.0 in.) from the outer surface of the ring-die
cut the pellets. The pellets were collected in a plastic container after achieving a steady
state pelleting condition. The steady state pelleting condition occurred after about two
to three minutes of feeding the grind. The temperature of the pellets was immediately
measured using a non-contact thermometer (MITCHELL, San Marcos, CA; Model
No. PLMITST2). Hot pellets were spread in a thin-layer on a screen with 6.4 mm (1/4
in.) diameter holes to naturally cool the pellets to room temperature (about 12.8C).
After cooling, the screen was electrically shaken to collect fines less than 6.4 mm, and
the cleaned pellets were used for evaluating immediate properties of the pellets. Some
pellet samples were stored in plastic bags for measuring pellet properties after one
week of storage at room temperature (about 23C). Figure 6.5 shows the pictures of
corn stover and switchgrass pellets.
During pelleting, the current input to the overall operation of the pelleting
machine was recorded using an ammeter to estimate the specific energy consumption
for the pelleting process. The specific energy consumption per unit mass of pellets
produced (MJ/t), and the specific energy consumption per unit decrease in volume of
the biomass feedstock (MJ/m3) were estimated by subtracting the energy consumption
for running the empty pelleting machine (i.e., no load condition) from the energy

258
consumption during biomass pelleting. To measure the throughput (kg/h) of the
pelleting machine, the mass of pellets collected for about 1.5 to 2.0 min was measured.

6.4-4 Machine and Biomass Feedstock Variables Studied

In the briquetting industry, for example, at Bepex International LLC


(Minneapolis, MN), prior to conducting experiments in pilot-scale roll press
briquetting machines, preliminary densification tests are conducted in a bench-scale
uniaxial compression apparatus to know the suitability of the powder for briquetting,
and to determine the range of powder variables that would help manufacture good
quality briquettes. Kaliyan and Morey (2006a) [Chapter 4] conducted preliminary
densification studies on corn stover and switchgrass using a lab-scale piston-cylinder
uniaxial compression apparatus to select the range of biomass variables such as
particle size, moisture content, and preheating temperature which could produce good
quality densified products in the pilot-scale roll press briquetting and pelleting
machines. They concluded that particle size obtained from a hammer mill screen size
of 3.0 mm, grind moisture content of 15% (w.b.) for corn stover and 10% (w.b.) for
switchgrass, and preheating temperature of 75C or above would produce strong and
high durability briquettes. They also concluded that when the moisture content of corn
stover grind was 15% (w.b.), there was no need to preheat the corn stover grind in
order to produce high quality corn stover briquettes. For switchgrass, preheating to at
least 75C with a grind moisture content of 10% (w.b.) was required to produce good
quality switchgrass briquettes.

For roll press briquetting experiments, the effects of two particle sizes (grinds
obtained from hammer mill screen sizes of 2.4 and 4.0 mm), two moisture contents
(10 and 15% w.b.), and three preheating (i.e., steam conditioning) temperatures (room
temperature, 75, and 100C) on briquetting both corn stover and switchgrass were
studied. Although we aimed to study the effect of two moisture contents (10 and 15%
w.b.), the moisture contents of the moisture conditioned grinds or steam conditioned
grinds deviated from these two values and ranged from 7 to 20% (w.b.). The roll press

259
briquetting machine variables such as roll force per unit width of the rolls, and size
and shape of the pockets on the rolls were kept constant due to the pilot-scale testing
nature of the study (table 6.3). The speeds of the rolls and screw feeder were variables,
and these speeds were recorded for each briquetting condition.
During briquetting, we had to adjust the roll speed and screw feeder speed
simultaneously to match (i.e., synchronize) them for each briquetting condition to
produce consistent briquettes. If the speed of the screw feeder was too fast or the
speed of the rolls was too slow, the biomass grinds plugged the hopper just above the
rolls, and thus no briquettes were formed. We then needed to stop the roll press and
remove the materials that caused plugging, before further operation. If the speed of the
rolls was too fast, biomass grinds came out through the rolls without forming
briquettes. For the roll press briquetting machine used in this study, the lowest
mechanically achievable roll speed was 4 rpm. The speed of the rolls was controlled
using a frequency controller to achieve roll speeds of less than 4 rpm.
To compare the densification performance of the roll press briquetting machine
with the conventional ring-die pelleting machine, pelleting experiments were
conducted for selected conditions. For pelleting experiments, the effects of particle
size (grinds obtained from 2.4- and 4.0-mm hammer mill screen sizes for corn stover,
and grinds obtained only from 2.4-mm hammer mill screen size for switchgrass), and
two die (i.e., hole in the ring-die) length to diameter (L/D) ratios (5.3 and 6.0) were
studied. The diameter of the die was kept constant at 9.5 mm (3/8 in.) for all of the
pelleting tests. Preliminary pelleting trials showed that at a moisture content of 15%
(w.b.) or below, consistent pellets were not formed; pellets appeared to be soft with
weak particle-particle bonding. Therefore, a constant moisture content of about 20%
(w.b.) was used for both corn stover and switchgrass for all of the pelleting tests. Also,
no steam conditioning was used for any of the pelleting tests.
No binding agents (i.e., additives) were used for any of the briquetting or
pelleting experiments. However, in this study, the "natural" binding components such
as lignin and protein in the biomass materials were activated mainly by the moisture

260
and preheating temperature - either due to steam conditioning or shear/frictional
heating or both, and, to some extent, due to the particle sizes used in this study.

6.4-5 Properties of Briquettes and Pellets

Properties of briquettes and pellets were measured immediately after forming,


and after one week of storage at room temperature to allow more curing of particle-
particle bonds. Bulk density, durability, and crushing strength (i.e., compressive
strength) of the roll press briquettes were measured immediately after making the
briquettes. After one week of storage of briquettes in plastic bags at room temperature,
bulk density, durability, crushing strength, individual briquette dimensions, angle of
repose, and moisture content of the briquettes were determined. Immediately after
making the pellets, bulk density and durability of the pellets were measured. After one
week of storage of pellets in plastic bags at room temperature, bulk density, durability,
hardness, individual pellet dimensions, angle of repose, and moisture content of the
pellets were estimated. It is noted that the bulk density and durability are properties of
a bulk sample of briquettes or pellets whereas the crushing strength or hardness is a
property of individual briquettes or pellets. Also, the crushing strength test for
briquettes is equivalent to the hardness test for pellets because compressive stress is
applied to fail the briquettes and the pellets.
Bulk density of the briquettes and pellets was measured from the mass of
briquettes or pellets occupying a 1.0-L glass container (ASABE Standards, 2003c).
The moisture content of the briquettes and pellets was quantified based on ASABE
Standard S269.4 (ASABE Standards, 2003c). Individual briquette or pellet dimensions
were measured using a digital Vernier caliper.
Durability of briquettes and pellets was measured using the tumbling can
method given in ASABE Standard S269.4 (ASABE Standards, 2003c). About 500-g
of briquettes or pellets were tumbled in a tumbler at 50 rpm for 10 min (fig. 6.6).
Then, the percentage of original weight of briquettes or pellets retained on a 3.2-mm
(1/8 in.) screen was calculated as the durability. The screen size of 3.2-mm used for
removing the fines from the "good" quality briquettes or pellets after tumbling was
261
based on the recommendation by U.S. Pellet Fuels Institute (PFI) (Arlington, VA:
http://www.pelletheat.org). The U.S. Pellet Fuels Institute (Arlington, VA) classifies
particles which are less than 3.2-mm as fines in packaged biomass pellets used for
heating purposes. Figures 6.7 and 6.8 show the whole briquettes and broken pieces of
briquettes retained on the 3.2-mm screen after the durability test for two typical cases.
Crushing strength of the briquettes was measured using a hand operated
crushing strength tester (Process Measurement Company, Minneapolis, MN).
Individual briquettes were compressed along their length or width (fig. 6.9A). The
briquettes failed mostly as two halves, and in some cases, as small individual broken
pieces. The force at failure of briquettes was taken to be the crushing strength of the
briquettes (Chapter 3; Kaliyan and Morey, 2006b). The crushing strength along the
depth of the briquettes was not measured because the biomass grinds were compressed
along the depth of the briquettes in the roll press briquetting machine to form
briquettes.
Pellet hardness was measured using a hand operated Dillon Quantrol Basic
Force Gauge (Model: BFG 500 N; Itin Scale Co., Inc., Brooklyn, NY) (fig. 6.9B). The
pellet hardness tester measured the force applied on individual pellets (along the
diameter) to a value where there was a reduction of force or resistance to compression
due to the fracturing or breaking of pellet into two halves (i.e., pellet failure point).
The maximum force at pellet failure was taken as pellet hardness (Chapter 3; Kaliyan
and Morey, 2006b).
Angle of repose of briquettes and pellets (about 250-g sample) on a galvanized
steel surface was measured using a tilted-plane angle of repose apparatus (Mohsenin,
1986).
Scanning Electron Microscopy (SEM) (Hitachi S3500N) images were taken
for corn stover and switchgrass grinds, and cross sections of the briquettes and pellets.
The SEM images were analyzed to understand the binding mechanisms of corn stover
and switchgrass.
Ultraviolet (UV) auto-fluorescence images (Olympus 1X70 Inverted
Microscope; UV excitation at 330-385 nm; dichroic mirror at 410 nm; emission at

262
420-700 run) of corn stover and switchgrass grinds, and cross sections (i.e., fractured
surfaces) of the briquettes and pellets were obtained to identify the natural binders.
UV auto-fluorescence in plant tissues is caused by the presence of aromatic molecules
(e.g., pigments) (Rost, 1995). According to Rost (1995), the color interpretation of UV
auto-fluorescence is: deep red for chloroplasts; green or yellow-green for protein
compounds; brilliant blue or bluish-white for lignin, cutin, suberin, or phenolic acids
such as ferulic acid; and whitish fluorescence for cutin (cuticle). Also, pure
carbohydrates (cellulose, hemicellulose, starch, and lipid/fat molecules) do not
fluoresce (Rost, 1995).

6.5 Results and Discussion

6.5-1 Characteristics ofBiomass Feedstock

Currently, corn stover and switchgrass are collected from the field as bales
with bulk density of 100 to 200 kg m"3 (Shinners et al., 2003 and 2006). Table 6.1
provides the particle size and bulk density of the corn stover and switchgrass grinds
used for the briquetting and pelleting studies. Corn stover grind had initial bulk
densities of 139 to 161 kg m"3 and geometric mean particle diameter of 0.34 to 0.36
mm, whereas switchgrass grind had bulk densities of 184 to 220 kg m" and geometric
mean particle diameter of 0.49 to 0.59 mm (table 6.1). Therefore, compared to the
bulk density of the bales, the increase in the bulk density of the biomass feedstocks
due to the size reduction (i.e., tub grinding and then hammer milling) was small.
Table 6.2 gives the compositions of the corn stover and switchgrass grinds
used for this study. Also, in table 6.2, the composition values provided by Mani et al.
(2006) and DOE (2007) are compared. Differences between the composition values of
the biomass materials provided by Mani et al. (2006) and DOE (2007) and the
composition values of the biomass materials measured in this study may be due to the
differences in the biomass variety and maturity, and other biological and ecological
factors. The constituents such as lignin, crude protein, starch, crude fat, and water
soluble carbohydrates are "natural binders" in the biomass materials. These natural
binders can be activated (softened or melted locally) either by high moisture or heat or
263
steam to use their binding functionality. Lignin and hemi-cellulose were found to be
amorphous thermoplastic materials which would undergo plastic deformation at
temperatures in the range of their glass transition temperatures (i.e., softening
temperatures) (Back and Salmen, 1982). In addition, during briquetting or pelleting at
temperatures close to the glass transition range, the natural binding components can be
squeezed out of the particles, which can then make solid bridges between particles.
After cooling, these solid bridges harden (i.e., curing process). This would make the
briquettes or pellets strong and durable (Rumpf, 1962; Tabil, 1996).
During briquetting or pelleting, the lignin (7.4 to 8.8% d.b.), crude protein (3.6
to 3.9% d.b.), starch (0.4 to 1% d.b.), and crude fat (0.7 to 0.9% d.b.) in the corn
stover and switchgrass could be activated when there is enough moisture, and
temperature above the glass transition temperature of the corn stover and switchgrass.
The glass transition occurs in corn stover and switchgrass from 50 to 113C (Chapter
4; Kaliyan and Morey, 2006a). The mean glass transition temperature for both corn
stover and switchgrass is 75C (Chapter 4; Kaliyan and Morey, 2006a). The water
soluble carbohydrates (2.2 to 7.9% d.b.) in the corn stover and switchgrass would be
activated when there is high enough moisture in the biomass during briquetting or
pelleting (Chapter 3; Kaliyan and Morey, 2006b). Compared to switchgrass, corn
stover has higher water soluble carbohydrates (table 6.2). Therefore, briquetting or
pelleting corn stover with high enough moisture to completely activate the water
soluble carbohydrates could result in stronger and more durable briquettes than for
switchgrass. The ash contents (i.e., mineral contents) in corn stover (11.2% d.b.) and
switchgrass (5.0% d.b.) show their relative abrasiveness to densification equipment
when there is high friction/shear during densification such as in pelleting. The higher
the mineral content, the higher the abrasion (Chapter 3; Kaliyan and Morey, 2006b).

264
6.5-2 Roll Press Briquetting

6.5.2-1 Effect of roll press briquetting machine variables

As mentioned previously, the force per unit width of rolls was set constant at
5.1 kN/mm for briquetting corn stover and switchgrass. The speed of the rolls ranged
from 1.3 to 2.5 rpm, and the speed of the screw feeder ranged from 41 to 52 rpm
(tables 6.5 and 6.7). The combination of screw feeder speed and roll speed that created
consistent briquettes was different for each briquetting condition; however, the
differences in the respective speeds were small (tables 6.5 and 6.7). The roll speed and
screw feeder speed obtained from this study could serve as starting values for making
briquettes of these biomass materials in commercial-scale roll press briquetting
machines (e.g., roll press with a maximum roll force of 250 tons) although these
speeds can be adjusted on the go. The biomass briquettes produced were of almond
shape with length of 28.7 to 31.3 mm (tables 6.6 and 6.8).
For briquetting cases without steam conditioning, the temperature of the corn
stover and switchgrass briquettes ranged from 51 to 63 C due to the frictional heating
of the grinds during briquetting (tables 6.5 and 6.7). Because this temperature range is
well within the range of glass transition temperature of corn stover and switchgrass
(i.e., 50 to 113C), the natural binders in the biomass materials would have been
activated to some extent to help produce durable particle-particle bonding.
Throughput of the roll press briquetting machine ranged from 16.8 to 32.6 kg
of briquettes per hour for corn stover (table 6.5), and 9.6 to 30.6 kg of briquettes per
hour for switchgrass (table 6.7). The throughput was mostly influenced by the roll
speed rather than the screw feeder speed. In general, the faster the roll speed, the
higher the throughput of briquettes (tables 6.5 and 6.7). For corn stover, about 4.2 to
20.7% of the product from the roll press briquetting machine was fines (<5.2 mm)
with bulk densities of 231 to 357 kg m"3. For switchgrass, about 10.7 to 51.2% of the
product was fines (<5.2 mm) with bulk densities of 252 to 392 kg m"3. It was found
that when the percentage of fines through the rolls was lower, the durability and

265
crashing strength of the briquettes were higher (tables 6.5 to 6.8). During commercial-
scale production of briquettes, the fines can be recycled to produce good quality
briquettes.
Table 6.9 provides the specific energy consumption for the roll press
briquetting process for each briquetting condition (without including the energy
consumed for running the roll press briquetting machine at no load condition, and the
energy involved for the steam conditioning). For corn stover, the specific energy
consumption for the roll press briquetting process was 1.69 to 3.04% of the energy
available in the corn stover. For switchgrass, about 1.74 to 4.38% of the energy in the
switchgrass was utilized for the roll press briquetting process.
The bulk density, durability, and crushing strength of the corn stover and
switchgrass briquettes measured after one week of curing at room temperature (tables
6.6 and 6.8 ) were slightly higher than those measured immediately after making the
briquettes (table 6.5 and 6.7) for most of the briquetting conditions. For the rest of the
discussion on roll press briquetting, the data measured after one week of curing of
briquettes at room temperature were involved (tables 6.6 and 6.8).

266
6.5.2-2 Effect of particle size

For roll press briquetting of corn stover at about 15 to 17% (w.b.) nominal
moisture content of the grind and at room temperature, increasing the geometric mean
particle size from 0.34 mm to 0.36 mm increased the bulk density, durability, and
crushing strength of the briquettes; however, the differences between the properties at
these two particle sizes were small because of a small difference between the two
particle sizes tested (table 6.6). Increasing the particle size of the corn stover grind
increased the specific energy required to compress a unit volume of the corn stover
(table 6.9). This may be due to the lower bulk density of the corn stover grind at the
larger particle size (table 6.1).
For briquetting switchgrass with a nominal grind moisture content of 9 to 10%
(w.b.) and at room temperature, increasing the geometric mean particle size from 0.49
mm to 0.59 mm decreased the bulk density of the briquettes from 467 to 433 kg m"3
and durability of the briquettes from 40 to 39% (table 6.8). Also, for switchgrass
briquettes obtained at nominal grind moisture content of 10 to 12% (w.b.) and grind
temperature of 68 to 85C, increasing the particle size from 0.49 mm to 0.59 mm
decreased the bulk density, durability, and crushing strength of the briquettes (table
6.8).
Therefore, for commercial-scale roll press briquetting of switchgrass, grind
obtained from a hammer mill screen size of 2.4 mm (geometric mean particle diameter
of 0.49 mm) could be used to produce good quality briquettes. For commercial-scale
roll press briquetting of corn stover, grind obtained from a hammer mill screen size of
4.0 mm (geometric mean particle diameter of 0.36 mm) could be used to make high
quality briquettes.

267
6.5.2-3 Effect of moisture content

For roll press briquetting of corn stover with a geometric mean particle
diameter of 0.34 mm and at room temperature, increasing the moisture content of the
grind from 7 to 15% decreased the bulk density of the briquettes from 481 to 422 kg
m"3, but increased the durability of briquettes from 67 to 87% (table 6.6). For roll press
briquetting of switchgrass with a geometric mean particle diameter of 0.49 mm and at
room temperature, increasing the moisture content of the grind from 9.2 to 19.8%
decreased the bulk density of the briquettes from 467 to 351 kg m"3, but increased the
durability of briquettes from 40 to 50%> (table 6.6).
The lower bulk density at the higher moisture content of the grind may have
been due to loss of mass due to the evaporation of water from the briquettes, and due
to the possible expansion of the briquettes after leaving the roll press briquetting
machine. The higher briquette durability at the higher moisture content may have been
due to the activation of the water soluble carbohydrates in the biomass materials. The
binding functionality of the water soluble carbohydrates could have been activated
much more at the higher moisture content than at the lower moisture content. In
addition, when briquetting without steam conditioning, the frictional heating from the
roll press briquetting machine resulted in briquette temperatures of 51 to 57C for corn
stover, and 54 to 63C for switchgrass. This temperature range is within the range of
the glass transition temperatures of these biomass materials, and thus, the other natural
binding components such as lignin, protein, starch, and fat could have been activated
to some extent. Also, due to the elevated temperatures, the resiliency of the grind
could have been decreased.
Furthermore, when briquetting without steam conditioning, at the higher
moisture content, corn stover resulted in higher durability of briquettes (briquette
durability of 87% at the moisture content of 15% w.b.) than for switchgrass (briquette
durability of 50% at the moisture content of 20% w.b.). This may be due to the
presence of larger amount of water soluble carbohydrates (i.e., natural binding
components) in corn stover (7.9% d.b.) than for switchgrass (2.2% d.b.). Also, the

268
specific energy required to compress a unit volume of the corn stover (MJ/m3)
decreased as the moisture content of the grind increased, whereas for switchgrass, the
opposite trend was observed (table 6.9).

6.5.2-4 Effect of steam conditioning temperature

For the roll press briquetting of corn stover with a geometric mean particle size
of 0.34 mm and nominal grind moisture content of 7 to 10% (w.b.), increasing the
temperature of the grind to 77C by steam conditioning from room temperature (i.e.,
22C) increased the durability (by 23% points) and crushing strength of briquettes (by
76-154 N), and decreased the bulk density of the briquettes (by 9 kg m"3) (table 6.6).
Similarly, for briquetting of corn stover with a particle size of 0.36 mm and nominal
moisture content of the grind of 15 to 17% (w.b.), increasing the temperature of the
grind to 75C by steam conditioning from room temperature (i.e., 18C) increased the
durability (by 0.2% points) and crushing strength of briquettes (by 51 to 58 N), and
decreased the bulk density of the briquettes (by 26 kg m"3) (table 6.6). The above
results suggest that if the corn stover grind moisture content is 15%) (w.b.) or above,
steam conditioning to increase the temperature of the grind to about 75C is not
necessary in order to produce strong and durable corn stover briquettes. As discussed
earlier, the binding functionality of the water soluble carbohydrates could be activated
much more at the higher moisture content than at the lower moisture content. In
addition, steam conditioning of corn stover grind appeared to activate the binding
functionalities of the additional natural binding components such as lignin, protein,
starch, and fat, which need a grind temperature close to the glass transition
temperature (i.e., 75C) for complete activation. This improved binding effect due to
the elevated temperature of the grind was reflected as increased durability and
crushing strength values of the corn stover briquettes.

With no steam conditioning (i.e., briquetting at room temperature), a consistent


briquette shape was not formed during roll press briquetting of switchgrass grind at a
nominal moisture content of 9 to 10% (w.b.) and geometric mean particle size of 0.49

269
mm or 0.59 mm. However, for these densification conditions, the switchgrass grind
was agglomerated, and the agglomerated product had durability of 39 to 40% (table
6.8). Steam conditioning of switchgrass grind to 68C (for the geometric mean particle
size of 0.49 mm) or 85C (for the geometric mean particle size of 0.59 mm) at a
nominal grind moisture content of 10 to 12% (w.b.) resulted in briquettes with bulk
density of 410 to 420 kg m"3, durability of 58 to 70%, and crushing strength of 53 to
171 N. With steam conditioning the switchgrass grind to a temperature close to 75C,
higher values of bulk density, durability, and crushing strength of briquettes were
obtained for a geometric mean particle size of 0.49 mm than for 0.59 mm. The
improved strength and durability of the switchgrass briquettes for the steam
conditioning cases was probably due to the activation of the natural binding
components in the switchgrass (table 6.2).
Steam conditioning of switchgrass grind to 109C resulted in lower durability
and crushing strength of briquettes than those values measured at 68C (table 6.8).
This may be due to the loss of moisture from the switchgrass when heating the grind
to more than 100C. The moisture content of the briquettes made at 109C was 6%
(w.b.) whereas moisture content of the briquettes made at 68C was 10% (w.b.).
Because enough moisture (at least 10% w.b.) was not available to soften the natural
binding components that require water for activation (softening), the net natural
binding effect could have been reduced at the elevated temperature of 109C.
Increasing the temperature of the corn stover grind to about 75C decreased the
specific energy required to compress a unit volume of the corn stover, whereas for
switchgrass, the opposite trend was observed (table 6.9).
Therefore, it can be concluded that switchgrass grind with a geometric mean
particle size of 0.49 mm, nominal grind moisture content of 10% (w.b.), and grind
temperature of about 75C would produce good quality briquettes. Also, in
commercial-scale roll press briquetting of switchgrass, steam conditioning can be
omitted if the frictional heating from the briquetting machine could raise the
temperature of the grind to about 75C.

270
6.5-3 Pelleting

6.5.3-1 Effect of pelleting machine variables

In the literature, several studies identified the pelleting machine variables that
affect the quality of pellets (Chapter 3; Kaliyan and Morey, 2006b). A review by
Kaliyan and Morey (2006b) [Chapter 3] found that following pellet machine variables
influence the strength and durability of the pellets: length to diameter (L/D) ratio of
the die, speed of the ring-die, gap between the rolls and ring-die, the specific energy
input to the pellet mill, and steam conditioning/high shear processing before pelleting
such as feed conditioning in an expander. In this study, only the effect of the L/D ratio
of the die was tested, and other pellet machine variables were kept constant (table 6.4).
The results of the pelleting experiments are given in tables 6.10 to 6.12. The bulk
density and durability of the pellets measured after one week of curing at room
temperature (table 6.11) were slightly higher than those measured immediately after
making the pellets (table 6.10) for most of the pelleting conditions.
In this study, a constant die diameter of 9.5 mm was used to pellet corn stover
and switchgrass. This resulted in a pellet diameter of 9.6 to 9.8 mm with individual
pellet densities of 1029 to 1111 kg m"3 (table 6.11). The length of the pellets ranged
from 20.7 to 24.0 mm (table 6.11). The switchgrass pellets had durability of 73 to 87%
(table 6.11). Colley et al. (2006) found that a die diameter of 4.76 mm produced
switchgrass pellets with durability of 78 to 97%. Also, for pelleting alfalfa, a die
diameter of 6.1 to 6.4 mm was found to produce good quality pellets (i.e., pellet
durability of up to 80%) (Hill and Pulkinen, 1988; Tabil and Sokhansanj, 1996). In
this study, we used a larger diameter of the die than has been used in the past for
pelleting biomass materials. Selection of a larger diameter of the die results in higher
throughput of pellets, and thus, a die diameter of 9.5 mm could be used for
commercial production of corn stover and switchgrass pellets.

271
Increasing the L/D ratio of the die from 5.3 to 6.0 increased the temperature of
the pellets, bulk density of the pellets, durability of the pellets, and hardness of the
pellets (tables 6.10 and 6.11). Also, increasing the L/D ratio of the die decreased the
specific energy consumption per unit mass of pellets produced (MJ/t), but increased
the specific energy consumption per unit decrease in the volume of the biomass
feedstock (MJ/m3) (tables 6.10 and 6.12). Although the two L/D ratios of the die
resulted in small difference in the pellet durability values for corn stover, they resulted
in a big difference in switchgrass pellet durability. Results in tables 6.10 and 6.11
suggest that L/D ratio of the die of 6.0 should be used to produce highly durable
switchgrass pellets. Hill and Pulkinen (1988) and Tabil and Sokhansanj (1996)
concluded that larger L/D ratios of the die (L/D ratios of 7.3 to 9.0 with D of 6.1 to 6.4
mm) resulted in higher durability of alfalfa pellets. Therefore, we conclude that the
L/D ratio of the die of 6.0 would produce high quality pellets with minimum energy
consumption in commercial pelleting of corn stover and switchgrass.
A peripheral ring-die speed of 4 to 5 m s"1 has been suggested to expel a large
volume of air during pelleting low bulk density feed (Heinemans, 1991; Tabil and
Sokhansanj, 1996; Kaliyan and Morey, 2006b [Chapter 3]). Tabil and Sokhansanj
(1996) produced highly durable alfalfa pellets at a ring-die speed of 250 rpm (2.8 m s"
'). In this study, the speed of the ring-die used was 232 rpm, which is equivalent to a
peripheral ring-die speed of 4.3 to 4.4 m s"1. With this speed of the ring-die (i.e., 232
rpm), highly durable pellets were produced from corn stover (pellet durability of up to
96%) and switchgrass (pellet durability of up to 87%).
Although no steam conditioning was used for pelleting corn stover and
switchgrass, the temperature of the pellets was measured at 67 to 81C due to the
frictional heating of the grinds during pelleting (table 6.10). This suggests that the
increase in temperature of the grinds during pelleting could have created a temperature
in the range of glass transition of the biomass grinds, and thus, the natural binding
components in the biomass materials would have been activated.

272
Throughput of pellets ranged from 86 to 226 kg h"1. Also, about 1.1 to 4.4% of
fines (<6.4 mm) along with pellets were collected as products. The fines were
produced mainly by the action of the knife while cutting the pellets, and due to the
abrasion of pellets in the collection container, and handling immediately after
production before sufficient curing could take place. In commercial pellet production,
fines can be recycled to make good quality pellets. The specific energy consumption
for the pelleting process was 0.95 to 1.31% of the energy in the corn stover, and 2.37
to 2.43% of the energy in the switchgrass (table 6.12). The requirement of higher
energy input for pelleting switchgrass suggests that switchgrass is a "hard to pellet"
category of feedstock.

6.5.3-2 Effect of biomass feedstock variables

The results showed that corn stover pellets with bulk density of up to 616 kg
3
m" , durability of up to 96%, and hardness of up to 224 N could be produced. Also,
switchgrass pellets with bulk density of up to 571 kg m"3, durability of up to 87%, and
hardness of up to 216 N could be made (table 6.11).
No previous pelleting study on corn stover was found in the literature;
however, Samson et al. (2000), Jannasch et al. (2004), and Colley et al. (2006) studied
the pelleting of switchgrass. Samson et al. (2000) and Jannasch et al. (2004) produced
switchgrass pellets with bulk density of 593 to 641 kg m"3 and hardness of >30 on the
Pfizer tablet hardness tester. They found that switchgrass pellets created large
quantities of fines (about 43% of initial feedstock dry matter) during the process of
handling, screening, and shaking before bagging of pellets; however, no pellet
durability data were reported. Colley et al. (2006) reported that switchgrass grind with
a moisture content of about 20% (w.b.) produced high quality pellets. We also found
that a grind moisture content of about 20% (w.b.) would result in good quality corn
stover and switchgrass pellets. Also, Colley et al. (2006) manufactured switchgrass
pellets (die diameter = 4.76 mm) with bulk density of 687 kg m"3, durability of 95.8%,
and pellet hardness of 27 N at a pellet temperature of 85C and pellet moisture content

273
of 11% (w.b.). Similarly, in this study, we produced switchgrass pellets (die diameter
= 9.5 mm) with bulk density of 571 kg m" , durability of 86.5%, and pellet hardness of
216 N at a pellet temperature of 81C and pellet moisture content of 11% (w.b.) (table
6.11). The main difference between this study and Colley et al. (2006) is that Colley et
al. (2006) steam conditioned the switchgrass grind to raise the temperature of the grind
to 75C before pelleting in a lab-scale, low-power pelleting machine (1.5 kW),
whereas in this study, steam conditioning was not done before pelleting the
switchgrass in a pilot-scale, high-power pelleting machine (29.8 kW).
At the L/D ratio of the die of 6.0, increasing geometric mean diameter of the
corn stover grind from 0.34 mm to 0.36 mm decreased the bulk density of the pellets
from 616 to 543 kg m" , durability from 96 to 95%, and pellet hardness from 224 to
197 N (table 6.11). However, the larger particle size of the corn stover consumed more
specific energy to produce a unit mass of pellets (i.e., MJ/t) and to compress a unit
volume of the corn stover (i.e., MJ/m3) (table 6.12). Considering the cost of grinding,
corn stover grind with a geometric mean diameter of 0.36 mm (i.e., hammer mill
screen size of 4.0 mm) could be chosen for producing high quality corn stover pellets
in commercial-scale pelleting machine. For switchgrass, results from this study and
from Colley et al. (2006) suggest that switchgrass grind with geometric mean
diameters of 0.49 mm to 0.87 mm (i.e., hammer mill screen sizes of 2.4 to 3.2 mm)
could be used to produce highly durable switchgrass pellets in a commercial-scale
pelleting machine.

274
6.5-4 Comparison of Roll Press Briquetting with Pelleting

The briquettes produced were of almond shape with a maximum size of 31.3
mm length x 23.3 mm width x 17.9 mm depth, and the briquettes had a maximum of
5.0-mm flashings. The pellets made were of cylindrical shape with a maximum size of
9.8 mm diameter x 24.0 mm length. Higher bulk density, durability, and strength
values were obtained for pellets than for briquettes. Corn stover briquettes had bulk
density of 422 to 481 kg m"3, durability of 67 to 90%, and crushing strength of 88 to
277 N, whereas corn stover pellets had bulk density of 555 to 616 kg m'3, durability of
94 to 96%, and hardness of 197 to 224 N. Also, switchgrass briquettes had bulk
density of 351 to 527 kg m"3, durability of 39 to 70%, and crushing strength of 28 to
171 N, whereas switchgrass pellets had bulk density of 534 to 571 kg m"3, durability of
73 to 87%, and hardness of 148 to 216 N. For comparison, the bulk density of shelled
corn is 721 kg m"3 (Mohsenin, 1986). Furthermore, the briquettes produced had bulk
densities from a minimum of 1.8 to 2.6 times to a maximum of 3.5 to 5.3 times the
bulk density of the bales (i.e., 100 to 200 kg m"3). The pellets produced had bulk
densities from a minimum of 2.8 to 3.1 times to a maximum of 5.6 to 6.2 times the
bulk density of the bales.

With no steam conditioning, the temperature of the briquettes was 51 to 63C,


and the temperature of the pellets was 67 to 81C. This rise in temperature of the
products was purely due to the heating provided by the friction/shear between the
biomass particles, and between the biomass particles and the machine. The higher
temperatures of the pellets may be due to the higher power of the pelleting machine
(33.5 kW) compared to that of the roll press briquetting machine (5.9 kW).
For the roll press briquetting process, the specific energy consumption per unit
decrease in volume of the biomass feedstock was 105.6 to 190.8 MJ/m3 for corn
stover, and 73.0 to 228.6 MJ/m3 for switchgrass (table 6.9). For the conventional ring-
die pelleting process, the specific energy consumption per unit decrease in volume of
the biomass feedstock was 77.4 to 105.9 MJ/m3 for corn stover, and 129.8 to 141.3
MJ/m3 for switchgrass (table 6.12). Thus, roll press briquetting and pelleting processes

275
would consume similar specific energy to decrease a unit volume of the biomass
feedstocks. Furthermore, the specific energy consumption for the roll press briquetting
process was 1.69 to 3.04% of energy in corn stover, and 1.74 to 4.38% of energy in
switchgrass. The specific energy consumption for the pelleting process was 0.95 to
1.31% of energy in corn stover, and 2.37 to 2.43% of energy in switchgrass. Specific
energy consumption is likely to be less for both briquettes and pellets produced in
commercial-scale systems optimized for production. Since roll press briquetting relies
primarily on normal compression with a small amount of friction, while ring-die
pelleting requires some normal compression and relatively high friction/shear, it is
likely that lower specific energy consumption may be possible in briquetting
compared to pelleting in commercial-scale systems optimized for production.
The angle of repose values for corn stover and switchgrass briquettes ranged
from 21 to 24. The angle of repose values for corn stover and switchgrass pellets
ranged from 19 to 21. For comparison, the angle of repose of shelled corn ranges
from 16 to 27 (Mohsenin, 1986). The angle of repose values provide an indication of
flow-ability of the briquettes and pellets (Mohsenin, 1986).
The storage stability of the briquettes and pellets is determined by their
moisture contents (Maier et al., 1992; Colley et al., 2006). The moisture content of the
corn stover and switchgrass briquettes ranged from 7 to 14% (w.b.). The moisture
content of the corn stover and switchgrass pellets ranged from 11 to 15% (w.b.). For
safe long-term storage of briquettes and pellets (to avoid molding and to avoid loss of
strength and durability), their moisture contents should be <13% (w.b.) (Maier et al.,
1992). During commercial-scale production, the excess moisture from the briquettes
and pellets can be safely removed using commercial coolers (Maier et al., 1992;
Thomas et al., 1997; Kaliyan and Morey, 2006b [Chapter 3]).

276
6.5-5 Binding Mechanisms of Corn Stover and Switchgrass

According to Rumpf (1962) and Pietsch (2002), bonding between particles


could take place through a solid bridge or through inter-particle attraction forces when
there is no solid bridge. Solid bridges are developed due to partial melting of
components, crystallization of soluble substances, chemical reaction, hardening of
binders, adsorption layers due to highly viscous binders (adhesion and cohesion forces
due to the binder layer between particles), and mechanical interlocking of particles.
When there is no solid bridge between particles, but the particles are brought close
enough together, short range attraction forces such as molecular (valence forces, for
instance, due to free chemical bonds, hydrogen bridges, and van der Waals' forces),
electrostatic, and magnetic forces can cause the particles to adhere to each other.
In the roll press briquetting and pelleting machines, biomass particles could
have experienced pressures of 100 to 200 MPa (Dec, 2002; Kaliyan and Morey, 2006b
[Chapter 3]). Because of the application of high pressures, particles were brought
close together causing inter-particle attraction forces, and the natural binding
components in the corn stover and switchgrass were squeezed out of the cells, which
made solid bridges between the particles. After cooling, these solid bridges hardened
(i.e., curing process). This caused the briquettes and pellets to become strong and
durable.
This postulate on the binding mechanisms was verified through SEM images.
Figures 6.10al and 6.10a2 show that most of the corn stover particles appear to be
bare without any sign of coating of natural binding components; however, on the
switchgrass particles, expression of some natural binding components (due to size
reduction) can be observed. The SEM images of the cross sections (i.e., fractured
surfaces) of briquettes and pellets show that the particles are covered with a layer of
natural binders (fig. 6.10). When viewed with a light microscopy (Nikon SMZ 1500),
these coatings appeared as glassy/white sugar like coatings on the particles. At the
junction of particles, high accumulation of these binding components was observed.
The UV auto-fluorescence images of briquettes and pellets indicate that the natural

277
binders that are coated on the particles in the briquettes or pellets are primarily lignin
and protein compounds (fig. 6.11). The SEM images for briquettes and pellets shown
in fig. 6.10 are for the briquetting or pelleting conditions that resulted in the highest
durability and strength values. The SEM images of briquettes and pellets made under
conditions that resulted in lower durability and strength values also appeared to have
similar coatings of natural binding components (see 6.8 Appendix). Therefore, the
SEM images can only explain the squeezing of natural binders from the particles due
to the high pressure, but not the differences in the durability and strength values. The
differences in the durability and strength values can be explained by the extent of
activation of the expressed natural binders through moisture or temperature or both.
The amount of moisture required depends on the amount of natural binder (e.g., water
soluble carbohydrates) available that requires moisture for activation. For example, as
discussed previously, at high moistures (15 to 20% w.b.), corn stover resulted in
stronger and more durable briquettes than switchgrass because of the presence of a
higher amount of water soluble carbohydrates than switchgrass.
Glass transition temperature is the minimum temperature required for
activation of natural binders. The glass transition in corn stover and switchgrass starts
at 50C (Chapter 4; Kaliyan and Morey, 2006a). Therefore, in addition to extracting
the natural binding components from the particles, the natural binding components
should be fully activated to use their binding functionality by providing enough
moisture and a temperature in the range of the glass transition for the biomass
materials.

6.6 Summary and Conclusions

In this research, pilot-scale roll press briquetting and pelleting behaviors of


corn stover and switchgrass were studied. The optimum machine and biomass
feedstock variables that could produce strong and durable briquettes and pellets in
commercial-scale briquetting and pelleting machines were determined. The briquettes
made were of almond shape with a maximum size of 31.3 mm length x 23.3 mm

278
width x 17.9 mm depth, and the briquettes had a maximum of 5.0-mm flashings. The
pellets made were of cylindrical shape with a maximum size of 9.8 mm diameter x
24.0 mm length. This study proved that good quality briquettes and pellets could be
produced from the corn stover and switchgrass without adding chemical binders (i.e.,
additives) by activating (softening) the natural binding components such as water
soluble carbohydrates, lignin, protein, starch, and fat in the corn stover and
switchgrass through moisture and temperature. Also, this study showed that if the
temperature rise of the biomass grinds due to frictional heating in the briquetting and
pelleting machines is in the range of glass transition (i.e., softening) temperature of the
biomass materials (i.e., >75C), strong and durable briquettes and pellets could be
produced without steam conditioning.

From the pilot-scale roll press briquetting experiments, the following


conclusions were drawn.

The speeds of screw feeder and rolls need to be adjusted simultaneously to


produce consistent briquettes. The ranges of screw feeder speed and roll speed
that produced high quality briquettes in the pilot-scale roll press briquetting
machine were 1.5 to 2.4 rpm, and 41 to 44 rpm, respectively.

To produce corn stover briquettes with bulk density of 453 kg m" , durability
of 88%, and crushing strength of 169 N, corn stover with a geometric mean
particle diameter of 0.36 mm (i.e., grind obtained from a hammer mill screen
size of 4.0 mm), and a grind moisture content of 15% (w.b.) or above should
be used. No preheating or steam conditioning of corn stover is needed to
produce strong and durable briquettes.

To produce switchgrass briquettes with bulk density of 420 kg m"3, durability


of 70%o, and crushing strength of 171 N, switchgrass with a geometric mean
particle diameter of 0.49 mm (i.e., grind obtained from a hammer mill screen
size of 2.4 mm), a grind moisture content of 10% (w.b.), and a grind

279
temperature of about 75C (i.e., glass transition temperature of switchgrass)
should be used. The temperature of the switchgrass grind of about 75C may
be increased either by frictional heating in the briquetting machine or by steam
conditioning.

For the above roll press briquetting conditions, the bulk densities of corn stover
and switchgrass briquettes manufactured were from a minimum of 2.1 to 2.3
times to a maximum of 4.2 to 4.5 times the bulk density of bales (i.e., 100 to
200 kg m"3).

From the pilot-scale pelleting experiments, the following conclusions were


drawn.

To produce corn stover pellets with bulk density of 543 kg m'3, durability of
95%, and hardness of 197 N, corn stover grind with a geometric mean particle
diameter of 0.36 mm (i.e., grind obtained from a hammer mill screen size of
4.0 mm), a grind moisture content of 20% (w.b.), and a pellet mill die length to
diameter (L/D) ratio of 6.0 with a die diameter (D) of 9.5 mm should be used.
No preheating or steam conditioning of corn stover is needed to produce strong
and durable pellets.

To produce switchgrass pellets with bulk density of 571 kg m"3, durability of


87%, and hardness of 216 N, switchgrass grind with a geometric mean particle
diameter of 0.49 mm (i.e., grind obtained from a hammer mill screen size of
2.4 mm), a grind moisture content of 20% (w.b.), and a pellet mill die length to
diameter (L/D) ratio of 6.0 with a die diameter (D) of 9.5 mm should be used.
No preheating or steam conditioning of switchgrass is needed to produce
strong and durable pellets because the frictional heating in the pelleting mill
could raise the temperature of the grind to 75C or more.

For the above pelleting conditions, the bulk densities of corn stover and
switchgrass pellets manufactured were from a minimum of 2.7 to 2.9 times to a
maximum of 5.4 to 5.7 times the bulk density of bales.

280
Briquettes produced with the roll press briquetting machine had bulk densities,
durabilities, and crushing strengths that were somewhat less than but in a range
comparable to the pellets produced with the conventional pelleting machine. Future
work is required to compare the cost and energy involved for commercial-scale
briquetting and pelleting of corn stover and switchgrass. Moreover, roll press
briquetting appears to be a promising low-cost, low-energy, high-capacity approach
for densifying corn stover and switchgrass for renewable energy applications.

6.7 References

ASABE Standards. 2003a. S358.2: Moisture measurement - Forages. St. Joseph,


Mich.: ASABE.
ASABE Standards. 2003b. S319.3: Method of determining and expressing fineness of
feed materials by sieving. St. Joseph, Mich.: ASABE.
ASABE Standards. 2003c. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: ASABE.
Back, E.L., and N.L. Salmen. 1982. Glass transitions of wood components hold
implications for molding and pulping processes. TappiJ. 65(7): 107-110.
Briggs, J.L., D.E. Maier, B.A. Watkins, and K.C. Behnke. 1999. Effects of ingredients
and processing parameters on pellet quality. Poultry Sci. 78(10): 1464-1471.
Colley, Z., O.O. Fasina, D. Bransby, and Y.Y. Lee. 2006. Moisture effect on the
physical characteristics of switchgrass pellets. Trans. ASABE 49(6): 1845-
1851.
Dec, R.T. 2002. Optimization and control of roller press operating parameters. Powder
Handling & Process. 14(3): 222-225.
DOE. 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The
Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.

281
Department of Energy (DOE), Office of Scientific and Technical Information.
Available at: http://www.osti.gov/bridge. Accessed 18 June 2005.
DOE. 2007. Biomass Feedstock Composition and Property Database. Washington,
DC: U.S. Department of Energy (DOE). Available at:
http://wwwl .eere.energy.gov/biomass/feedstock_databases.html. Accessed 25
April 2007.

Gilpin, A.S., T.J. Herrman, K.C. Behnke, and F.J. Fairchild. 2002. Feed moisture,
retention time, and steam as quality and energy utilization determinants in the
pelleting process. AppliedEng. inAgric. 18(3): 331-338.
Heinemans, H. 1991. The interaction of practical experience and the construction of
new pelleting and cooling machinery. Advances in Feed Tech. 6: 24-38.
Hill, B., and D.A. Pulkinen. 1988. A Study of the Factors Affecting Pellet Durability
and Pelleting Efficiency in the Production of Dehydrated Alfalfa Pellets.
Saskatchewan, Canada: Saskatchewan Dehydrators Association.

Holley, C.A. 1983. The densification of biomass by roll briquetting. Proc. of the
Institute for Briquetting and Agglomeration (IBA), Vol. 18. pp. 95-102.
Jannasch, R., Y. Quan, and R. Samson. 2004. A Process and Energy Analysis of
Pelletizing Switchgrass. Ste. Anne de Bellevue, QC: Resource Efficient
Agricultural Production (REAP-Canada). Available at: http://www.reap-
canada.com/library.htm. Accessed 27 August 2004.
Kaliyan, N., and R.V. Morey. 2006a. Densification characteristics of corn stover and
switchgrass. ASABE Paper No. 066174. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2006b. Factors affecting strength and durability of
densified products. ASABE Paper No. 066077. St. Joseph, Mich.: ASABE.
Koser, H.J.K., G. Schmalstieg, and W. Siemers. 1982. Densification of water
hyacinth-basic data. Fuel 61(9): 791-798.
Lemus, R., E.C. Brummer, K.J. Moore, N.E. Molstad, C.L. Burras, and M.F. Barker.
2002. Biomass yield and quality of 20 switchgrass populations in southern
Iowa, USA. Biomass and Bioenergy 23(6): 433-442.
282
Li, Y., and H. Liu. 2000. High-pressure densification of wood residues to form an
upgraded fuel. Biomass and Bioenergy 19(3): 177-186.

Maier, D.E., R.L. Kelley, and F.W. Bakker-Arkema. 1992. In-line, chilled air pellet
cooling. FeedMgmt. 43(1): 28-32.

Mani, S., L.G. Tabil, and S. Sokhansanj. 2003. An overview of compaction of biomass
grinds. Powder Handling & Process. 15(3): 160-168.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004. Grinding performance and physical
properties of wheat and barley straws, corn stover and switchgrass. Biomass
and Bioenergy 27(4): 339-352.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2006. Effects of compressive force, particle
size and moisture content on mechanical properties of biomass pellets from
grasses. Biomass and Bioenergy 30(7): 648-654.
McMullen, J., O.O. Fasina, C.W. Wood, and Y. Feng. 2005. Storage and handling
characteristics of pellets from poultry litter. Applied Eng. inAgric. 21(4): 645-
651.
Mohsenin, N.N. 1986. Physical Properties of Plant and Animal Materials. New
York, NY: Gordon and Breach Science Publishers.
Obernberger, I., and G. Thek. 2004. Physical characterisation and chemical
composition of densified biomass fuels with regard to their combustion
behavior. Biomass and Bioenergy 27(6): 653-669.
Opoku, A., M.D. Shaw, L.G. Tabil, D. Pulkinen, and B.J. Crerar. 2007. Effect of
whole canola inclusion on physical quality characteristics of feed pellets
processed from reground alfalfa, barley, and canola grains. App. Eng. in Agric.
23(1): 77-82.
Pietsch, W. 2002. Agglomeration Processes - Phenomena, Technologies, Equipment.
Weinheim: Wiley-VCH.
Pordesimo, L.O., B.R. Hames, S. Sokhansanj, and W.C. Edens. 2005. Variation in
corn stover composition and energy content with crop maturity. Biomass and
Bioenergy 28(4): 366-374.

283
Rost, F.W.D. 1995. Fluorescence Microscopy. Vol II. New York, NY: Cambridge
University Press.
Rumpf, H. 1962. The strength of granules and agglomerates. In Agglomeration, W.A.
Knepper, ed., 379-418. New York, NY: John Wiley and Sons.
Samson, R., P. Duxbury, M. Drisdelle, and C. Lapointe. 2000. Assessment of
Pelletized Biofuels. Available at: http://www.reap-
canada.com/Reports/pelletaug2000.html. Accessed 27 August 2004.

Shinners, K.J., B.N. Binversie, and P. Savoie. 2003. Harvest and storage of wet and
dry corn stover as a biomass feedstock. ASAE Paper No. 036088. St. Joseph,
Mich.: AS ABE.
Shinners, K.J., G.C. Boettcher, R.E. Muck, P.J. Wiemer, and M.D. Casler. 2006.
Drying, harvesting and storage characteristics of perennial grasses as biomass
feedstocks. ASABE Paper No. 061012. St. Joseph, Mich.: ASABE.

Sitkei, G. 1986. Mechanics of Agricultural Materials. New York, NY: Elsevier.


Tabil, L.G., Jr. 1996. Binding and pelleting characteristics of alfalfa. Ph.D. diss.
Saskatoon, Saskatchewan, CA: University of Saskatchewan, Department of
Agricultural and Bioresource Engineering.
Tabil, L., Jr., and S. Sokhansanj. 1996. Process conditions affecting the physical
quality of alfalfa pellets. Applied Eng. in Agric. 12(3): 345-350.
Thomas, M., D.J. van Zuilichem, and A.F.B. van der Poel. 1997. Physical quality of
pelleted animal feed. 2. Contribution of processes and its conditions. Animal
FeedSci. Tech. 64(2-4): 173-192.

284
Figure 6.1. Schematics of conventional ring-die pelleting (A), and roll press
briquetting (B) machines.

285
Figure 6.2. Roll press briquetting machine (A: Front view of the roll press where one
half of the conical cover for the screw feeder is removed, and the arrows
indicate the direction of operation. B: Side view of one of the rolls
showing the briquette pockets.).

286
Size reduction of Moisture conditioning Roll Press Briquetting of
biomass (i.e., tub and/or steam biomass grind
grinding and hammer conditioning of biomass
milling) grind

Briquettes
+ Fines

Measuring properties of Screening of briquettes Cooling of briquettes


the briquettes to remove fines that plus fines to room
leaked through the rolls temperature

Figure 6.3. Flow diagram of roll press briquetting of corn stover and switchgrass.

287
Corn stover briquettes Switchgrass briquettes

Figure 6.4. Corn stover and switchgrass briquettes produced.

Corn stover pellets Switchgrass pellets

Figure 6.5. Corn stover and switchgrass pellets produced.

288
Figure 6.6. Durability tester (tumbler or tumbling can) used for measuring the
durability of briquettes and pellets based on ASABE Standards S269.4
(2003c).

289
f

Figure 6.7-1. Whole briquettes retained on a 3.2 mm screen after the durability test.

Figure 6.7-2. Broken pieces of briquettes retained on a 3.2 mm screen after the durability test.

Figure 6.7. Corn stover briquettes (whole briquettes plus broken pieces of briquettes)
retained on a 3.2 mm (1/8 in.) screen after a durability test.

290
$*&'&&&

mmmmM

Figure 6.8. Switchgrass briquettes (whole briquettes plus broken pieces of briquettes)
retained on a 3.2 mm (1/8 in.) screen after a durability test.

Figure 6.9. Crushing strength testers for briquettes (A), and pellets (B).

291
m
W/i

r a

al. Corn stover grind (particle size = 0.34 mm) a2. Switchgrass grind (particle size = 0.49 mm)
before briquetting or pelleting (no coatings of before briquetting or pelleting (some natural
natural binders on most of the particles). binders are expressed on the particles due to size
reduction).

.-*._ **w i

\ Y , \ , s ? / . - . - f j ' y
" . V-iT-*'' - - \ ' " -

bl. Cross section of a corn stover briquette b2. Cross section of a switchgrass briquette made
made with a particle size of 0.34 mm, grind with a particle size of 0.49 mm, grind moisture
moisture content of 15% (w.b.), and no steam content of 12% (w.b.), and steam conditioning to
conditioning (see coatings of natural binders). 68C (see coatings of natural binders).

H&Jj '*

5i ^*#*
cl. Cross section of a corn stover pellet made c2. Cross section of a switchgrass pellet made
with a particle size of 0.34 mm, grind moisture with a particle size of 0.49 mm, grind moisture
content of 19% (w.b.), die length to diameter content of 20% (w.b.), die length to diameter
(L/D) ratio of 6, and no steam conditioning (IVD) ratio of 6, and no steam conditioning (see
(see coatings of natural binders). coatings of natural binders).

Figure 6.10. SEM micrographs of corn stover and switchgrass grinds, briquettes, and
pellets (magnification at X600).
292
al. Corn stover grind (particle size = 0.34 mm) a2. Switchgrass grind (particle size = 0.49 mm)
before briquetting or pelleting. before briquetting or pelleting.

bl. Cross section of a corn stover briquette b2. Cross section of a switchgrass briquette made
made with a particle size of 0.34 mm, grind with a particle size of 0.49 mm, grind moisture
moisture content of 15% (w.b.), and no steam content of 12% (w.b.), and steam conditioning to
conditioning. 68C.

cl. Cross section of a corn stover pellet made c2. Cross section of a switchgrass pellet made
with a particle size of 0.34 mm, grind moisture with a particle size of 0.49 mm, grind moisture
content of 19% (w.b.), die length to diameter content of 20% (w.b.), die length to diameter
(L/D) ratio of 6, and no steam conditioning. (L/D) ratio of 6, and no steam conditioning.

Figure 6.11. UV auto-fluorescence images of corn stover and switchgrass grinds,


briquettes, and pellets (magnification at X145). Auto-fluorescence color
interpretation (Rost, 1995): green or yellow-green for protein
compounds; brilliant blue or bluish-white for lignin, cutin, suberin, or
phenolic acids such as ferulic acid; and whitish fluorescence for cutin
(cuticle). Pure carbohydrates (cellulose, hemicellulose, starch, and
lipid/fat molecules) do not fluoresce (Rost, 1995).

293
Table 6.1. Properties of corn stover and switchgrass grinds used for roll press
briquetting and pelleting experiments.

Hammer mill screen Initial grind moisture Particle size *** Bulk density **
used for grinding content** (n=3) (kgm- 3 )(n=3)
(mm) * (% w.b.) (n = 3)
Corn Stover
2.4 (3/32 in.) 7.3 0.1 0.34 0.33 160.5 6.8
4.0 (5/32 in.) 8.5 0.04 0.36 0.38 139.1 7.8
Switchgrass
2.4 (3/32 in.) 9.2 0.01 0.49 + 0.31 219.9 2.7
4.0 (5/32 in.) 10.4 + 0.1 0.59 + 0.30 184.2+1.2
* Diameter of the holes in the screen.
** Mean standard deviation.
*** Geometric mean particle diameter (mm) geometric standard deviation.

294
Table 6.2. Compositions of corn stover and switchgrass grinds.

Component Corn stover Switchgrass


Corn Mani et DOE Switchgrass Mani et al. DOE
stover used al. (2006) (2007) used for this (2006) (% (2007)
for this (%ofdry (range, % study (% of of dry (range,
study (% matter) of mass) dry matter) matter) %of
of dry (n = l) mass)
matter)
(n = l)
Dry matter (%) 94.90 NA a NA 94.60 NA NA
Acid detergent 58.20 34.44 NA 53.00 51.77 NA
fiber (ADF)
Neutral 84.40 55.52 NA 81.80 81.77 NA
detergent fiber
(NDF)
Lignin 8.80" 3.12 b 17.13- 9.20" 7.43" 13.15-
21.25 b 22.49 b
Cellulose 49.40 c 31.32 30.61- 43.80 c 44.34 c 27.81-
38.12 37.14
Hemicellulose 26.20 d 21.08" 19.13- 28.80" 30.00 d 22.43-
25.29 28.63
Non fiber 2.10 NA NA 10.70 NA NA
carbohydrates
(NFC)
Water soluble 7.90 NA NA 2.20 NA NA
carbohydrates
(WSC)
Ethanol 1.20 NA NA .20 NA NA
soluble
carbohydrates
(ESC)
Crude protein 3.60 8.70 NA 3.90 1.59 NA
Starch 0.40 NA NA 1.00 NA NA
Crude fat 0.70 1.33 NA 0.90 1.87 NA
Ash 11.23 7.46 9.82-13.51 4.99 5.49 2.54-7.55
Calcium 0.34 NA NA 0.37 NA NA
Phosphorus 0.05 NA NA 0.08 NA NA
Magnesium 0.13 NA NA 0.11 NA NA
Potassium 0.84 NA NA 0.23 NA NA
Sulfur 0.06 NA 0.06-0.10 0.06 NA 0.07-0.11
NA = Data not available.
b
Lignin values estimated for the biomass materials used in this study and in Mani et al. (2006) were acid insoluble
lignin contents, whereas the lignin contents obtained from DOE (2007) were total lignin in the biomass materials.
c
Cellulose = ADF - Lignin.
d
Hemicellulose = NDF - ADF.

295
Table 6.3. Specifications of the roll press briquetting machine.

Roll press component Value


Roll motor
Power 3.7kW(5.0hp)
Voltage 460.0 V (3 phase)
Peak load current 8.0 A
Maximum speed of rolls 15 rpm
Roll dimension, and roll force
Diameter of rolls 228.6 mm (9.0 in.)
Width of rolls 38.1 mm (1.5 in.)
Maximum available roll force 222.4 kN (25 tons) (i.e., 5.84 kN/mm width of
rolls)
Maximum pressure of the hydraulic system 20.7 MPa (3000 psi)
corresponding to the maximum roll force (the
rolls were pressurized by the hydraulic system)
Force applied to the rolls during briquetting of 17.9 MPa (2600 psi) of hydraulic system
corn stover and switchgrass (kept constant) pressure (i.e., 5.1 kN/mm width of rolls)
Gap between the rolls 1.0 mm (0.04 in.)
Shape and size of pockets on rolls
Shape of the briquettes Large almond
Length of the briquette (almond) pocket 31.75 mm (1.25 in.)
Width of the briquette (almond) pocket 20.32 mm (0.8 in.)
Depth of the briquette pocket on one roll 5.08 mm (0.2 in.)
Approximate briquette dimension (Length * 31.75 x 20.32 x 11.16 mm
width x depth)
Screw feeder motor
Power 2.2 kW (3.0 hp)
Voltage 460.0 V (3 phase)
Peak load current 3.7 A
Maximum screw feeder speed 70 rpm

296
Table 6.4. Specifications of the pelleting machine.

Pelleting machine component Value


Pelleting mill motor (this motor operated the corrugated rolls and ring-die)
Power 29.8 kW (40.0 hp)
Voltage 240.0 V (3 phase)
Peak load current 111.0 A
Speed of the motor 1775 rpm
Ring-die, and corrugated rolls
Inner diameter of the ring-die 304.8 mm (12.0 in.)
Outer diameter of the ring-die Inner diameter of the ring-die +
thickness of the ring-die (2.0 or 2.25
in.)
Speed of the ring-die 232 rpm
Number of corrugated rolls inside the ring-die 2
Gap between the corrugated rolls and the ring-die 0.4 mm (0.02 in.)
Number of knives used for cutting the pellets 1
Gap between the knife and the outer surface of the ring-die 25.4 mm (1.0 in.)
Dimensions of the holes in the ring-die
Diameter of the holes in the ring-die 9.5 mm (3/8 in.)
Length-to-diameter (L/D) ratio of the holes in the ring-die 5.3 and 6.0
(two L/D ratios (i.e., two ring-dies) were tested)
Feeder motor (Feeder motor operated a screw conveyor located below the feed hopper, and
paddles in the conditioning chamber)
Power 3.7kW(5.0hp)
Voltage 240.0 V (3 phase)
Peak load current 15.7 A

297
Table 6.5. Properties of corn stover briquettes measured immediately after making the briquettes (mean standard deviation, if given).

Briquetting conditions Data measured immediately after making the briquettes


Moisture Temperature of Roll Screw Fines leakage Throughput of Temperature Bulk density of Bulk Durability Crushing strength
content of conditioned speed feeder through rolls briquettes of briquettes fines density of of briquettes of briquettes (N)
3
conditioned grind (C) (rpm) speed (%) (kg h 1 ) CQ * (kg m ) briquettes (%) (n = 3 ) *
grind (% (n = 3) (rpm) (n = l ) # (n=l) (n=l) (kg m 3 ) (n = l) Along Along
w.b.)(n=3) (n = D length width
Corn stover (Hammer mill screen size = 2.4 mm (3/32 in.); particle size of grind = 0.34 mm 0.33)
7.3 0.1 22.00.0 1.5 42.0 20.7 19.2 57.0 231.3 417.6 66.6 80.1 151.2
40.8 16.0
15.0 0.0 20.3 0.0 1.6 41.0 7.9 20.0 51.0 286.7 415.5 83.6 183.9 115.7
80.2 4.4
9.9 0.2 77.3 5.9 2.3 52.0 4.2 32.6 >70.0 320.9 429.0 87.5 317.3 286.2
18.5 40.4
Corn stover (Hammer mill screen size = 4.0 mm (5/32 in.); particle size of grind = 0.36 mm 0.38)
16.6 0 . 6 18.0 0 . 0 1.5 44.0 7.1 16.8 54.0 356.5 397.7 85.9 120.1 176.4
27.8 20.1
14.8 0 . 3 75.4 0.6 1.5 49.0 4.6 20.5 >52.0 277.7 463.4 86.9 271.3 203.1
38.0 43.2
* Fines leakage through rolls (%) = (kg of fines * 100) / (kg of fines + kg of briquettes).
* For cases with steam conditioning, temperature of the briquettes was less than the temperature of the grinds because of faster cooling of briquettes exiting the roll press
briquetting machine.
&
Crushing strength of the briquettes was measured using a hand-operated crushing strength tester.

298
Table 6.6. Properties of corn stover briquettes measured after one week of storage at room temperature (mean standard deviation, if
given).

Briquetting; conditions Data measured after one week ol' storage of briquettes at room temperature
Moisture Temperature Roll Screw Temperature Moisture Briquette dimensions * Bulk Durability Crushing Angle
content of of conditioned speed feeder of briquettes content of (n = 5) density of of strength of of
conditioned grind CO (rpm) speed (C)* briquettes briquettes briquettes briquettes (N) repose
grind (% (n = 3) (rpm) (% w.b.) (kg m 3 ) (%) (n == 3) e (deg)
w.b.)(n=3) (n = 3) Length Width Depth (n = l) (n = l) Along Along (n = 3)
(mm) (mm) (mm) length width
Corn stover (Hammer mill screen size = 2.4 mm (3/32 in.); particle size of grind = 0.34 mm 0.33)
7.3 0 . 1 22.0 0.0 1.5 42.0 57.0 7.1 0 . 1 29.0 21.0 15.1 480.8 66.7 174.7 122.6 23.7
0.5 0.4 0.4 61.9 65.1 1.5
15.0 0 . 0 20.3 0.0 1.6 41.0 51.0 13.9 0 . 1 29.7 21.1 15.6 422.4 86.5 88.1 138.5 23.3
0.6 0.7 0.8 17.9 35.5 0.6
9.9 0.2 77.3 5.9 2.3 52.0 >70.0 9.0 0 . 1 28.7 20.7 14.8 472.3 89.8 250.3 276.7 21.3
0.5 0.1 0.1 27.0 22.8 1.5
Corn stover (Hammer mill screen size = 4.0 mm (5/32 in.); particle size of grind = 0.36 mm 0.38)
16.6 0 . 6 18.0 0 . 0 1.5 44.0 54.0 12.3 0 . 1 29.7 22.2 15.5 452.5 87.5 128.2 169.3 24.0
0.4 0.4 0.4 28.3 40.1 1.0
14.8 0.3 75.4 0.6 1.5 49.0 >52.0 9.5 0.03 28.8 20.7 14.8 478.4 87.7 179.2 227.0 24.0
0.5 0.6 0.3 35.6 25.6 1.0
* For cases with steam conditioning, temperature of the briquettes was less than the temperature of the grinds because of faster cooling of briquettes exiting the roll press
briquetting machine.
&
Dimensions of the almond shape of the briquettes were measured using a digital Vernier caliper. Also, the briquettes had a maximum of 5-mm flashings around the almond
shape.
* Crushing strength of the briquettes was measured using Instron universal testing machine (Instron Corporation, Canton, N4A; Model 4206) at a compression speed of 2.54 mm

299
Table 6.7. Properties of switchgrass briquettes measured immediately after making the briquettes (mean standard deviation, if
given).

Briquetting conditions Data measured immediately after making the briquettes


Moisture Temperature of Roll Screw Fines leakage Throughput of Temperature of Bulk Bulk density Durability of Crushing strength
content of conditioned speed feeder through rolls briquettes briquettes (C) * density ol of briquettes briquettes of briquettes (N)
3
conditioned grind (C) (rpm) speed (%) (kgh- 1 ) fines (kg m" ) (%) (n = 3 ) *
grind (% (n = 3) (rpm) (n = l) (n = l) (kg m 3 ) (n = l) (n = l) Along Along
w.b.)(n = 3) (n = l) length width
Switchgrass (Hammer mill screen size = 2.4 mm (3/32 in.); particle size of grind = 0.49 mm 0.31)
9.2 0.01 19.0 0 . 0 1.3 46.0 47.1 9.6 58.0 330.5 438.7 NA ** NA NA
19.8 0 . 5 19.0 0 . 0 1.3 42.0 24.8 21.2 63.0 266.6 335.9 NA NA NA
11.7 0 . 2 68.3 7.6 2.35 41.0 10.7 30.6 75.0 251.9 364.2 64.7 62.3 117.1
11.8
51.6
10.1 0 . 1 109.4 1 0 . 7 2.5 48.0 18.9 26.9 >85.0 392.4 510.7 53.9 133.4 164.6
58.8
55.7
Switchgrass (Hammer mill screen size = 4.0 mm (5/32 in.); particle size of grind = 0.59 mm 0.30)
10.4 0 . 1 19.0 0 . 0 1.75 43.0 51.2 NA 54.0 300.0 384.8 NA NA NA
10.0 0 . 0 85.0 0 . 0 1.92 41.0 23.2 NA 75.0 274.1 414.3 54.7 NA NA
Fines leakage through rolls (%) = (kg of fines * 100) / (kg of fines + kg of briquettes).
* For cases with steam conditioning, temperature of the briquettes was less than the temperature of the grinds because of faster cooling of briquettes exiting the roll press
briquetting machine.
&
Crushing strength of the briquettes was measured using a hand-operated crushing strength tester.
** NA = Data not available.

300
Table 6.8. Properties of switchgrass briquettes measured after one week of storage at room temperature (mean standard deviation, if
given).

Briquetting conditions Data measured after one week of storage of 1briquettes at room temperature
Moisture Temperature Roll Screw Temperature Moisture Briquette dimensions* Bulk Durability Crushing Angle
content of of conditioned speed feeder of briquettes content of (n = 5) density of of strength of of
conditioned grind (C) (rpm) speed CO* briquettes briquettes briquettes briquettes (N) repose
grind (% (n = 3) (rpm) (% w.b.) (kg m3) (%) (n == 3) (deg)
w.b.)(n = 3) (n = 3) Length Width Depth (n=l) (n = l) Along Along
(mm) (mm) (mm) length width
Switchgrass (Hammer mill screen size = 2.4 mm (3/32 in.); particle size of grind = 0.49 mm 0.31)
9.2 0 . 0 1 19.0 0 . 0 1.3 46.0 58.0 8.0 0.04 NA** NA NA 467.3 39.5 NA NA NA
19.8 0 . 5 19.0 0 . 0 1.3 42.0 63.0 12.0 0 . 1 31.3 23.3 17.9 351.1 50.2 27.5 28.5 22.3
0.9 1.6 0.6 7.3 7.3 1.5
11.7 0 . 2 68.3 7.6 2.35 41.0 75.0 9.8 0 . 1 29.0 22.0 16.6 419.7 70.0 58.9 170.8 23.0
0.2 0.6 0.6 4.2 58.2 1.0
10.1 0 . 1 109.4 10.7 2.5 48.0 >85.0 5.5 0.02 29.0 21.1 14.8 526.7 61.6 95.6 142.3 21.3
0.5 0.8 0.5 14.9 62.5 0.6
Switchgrass (Hammer mill screen size = 4.0 mm (5/32 in.); particle size of grind = 0.59 mm 0.30)
10.4 0 . 1 19.0 0 . 0 1.75 43.0 54.0 7.6 0 . 1 NA NA NA 433.1 39.1 NA NA NA
10.0 0 . 0 85.0 0 . 0 1.92 41.0 75.0 7.3 0 . 1 29.2 21.4 15.9 409.6 58.1 52.9 80.1 21.3
0.7 0.8 0.3 3.6 6.3 0.6
* For cases with steam conditioning, temperature of the briquettes was less than the temperature of the grinds because of faster cooling of briquettes exiting the roll press
briquetting machine.
* Dimensions of the almond shape of the briquettes were measured using a digital Vernier caliper. Also, the briquettes had about a maximum of 5-mm flashings around the almond
shape.
* Crushing strength of the briquettes was measured using Instron universal testing machine (Instron Corporation, Canton, MA; Model 4206) at a compression speed of 2.54 mm
min"1.
** NA = Data not available. For these briquetting conditions, consistent briquette shape was not formed although the switchgrass particles agglomerated.

301
Table 6.9. Specific energy consumption for the roll press briquetting process (mean standard deviation, if given).

Briquetting conditions Current drawn Specific energy consumption


during briquetting
Moisture Temperature of Roll Screw Roll Screw Energy Energy Specific energy for roll press
content of conditioned speed feeder motor feeder consumed by consumed by briquetting process
conditioned grind (C) (rpm) speed (A) motor roll motor screw feeder MJ/t H~ MJ/m J Percentage
grind (n = 3) (rpm) (A) (MJ/t)' motor (MJ/t) Idl of energy in
[bl
(% w.b.) material
(n = 3)
Corn stover (Hammer mill screen size == 2.4 mm (3/32 in.); particle size of grind = 0.34 mm 0.33)
7.3 0.1 22.0 0.0 1.5 42.0 6.0 2.0 373.3 14.2 387.5 124.1 1.94
15.0 0.0 20.3 0.0 1.6 41.0 6.3 2.1 401.8 27.3 429.1 112.4 2.14
9.9 0.2 77.3 5.9 2.3 52.0 6.3 3.0 246.5 92.0 338.5 105.6 1.69
Corn stover (Hammer mill screen size == 4.0 mm (5/32 in.); particle size of grind = 0.36 mm 0.38)
16.6 0.6 18.0 0.0 1.5 44.0 6.5 2.5 511.5 97.2 608.7 190.8 3.04
14.8 0.3 75.4 0.6 1.5 49.0 6.0 2.4 349.1 66.3 415.5 141.0 2.08
Switchgrass (Hammer mill screen size == 2.4 mm (3/32 in.); particle size of grind - 0.49 mm 0.31)
9.2 0.01 19.0 0.0 1.3 46.0 4.2 2.2 209.7 85.4 295.1 73.0 1.74
19.8 0.5 19.0 0.0 1.3 42.0 8.5 2.3 676.4 45.0 721.4 94.7 4.24
11.7 0.2 68.3 7.6 2.35 41.0 9.0 2.6 515.8 62.4 578.2 115.5 3.40
10.1 0.1 109.4 10.7 2.5 48.0 8.5 4.0 532.7 212.6 745.3 228.6 4.38

302
Table 6.9. Continued.

Briquetting conditions Current drawn Specific energy consumption


during briquetting
Moisture Temperature of Roll Screw Roll Screw Energy Energy
Specific energy for roll press
content of conditioned speed feeder motor feeder consumed by consumed by briquetting process
conditioned grind (C) (rpm) speed (A) motor roll motor screw feeder
MJ/t'c| MJ/m3 Percentage
|d|
grind (% (n = 3) (rpm) (A) (MJ/t) w motor (MJ/t) of energy in
w.b.) material
(n = 3) (%) !e|
Switchgrass (Hammer mill screen size = 4.0 mm (5/32 in.); particle size of grind = 0.59 mm 0.30)
10.4 0.1 19.0 0.0 1.75 43.0 5.0 2.0 NA1" NA NA NA NA
10.0 0.0 85.0 0.0 1.92 41.0 5.0 2.2 NA NA NA NA NA
w
Energy consumed by roll motor (MJ/t) = [1.732 x 460 V x (current to roll motor during briquetting in A - current to roll motor during no load condition in A) * power factor *
3.6] / kg of briquettes per hour.
Current to roll motor during no load condition was 3.5 A. The power factor for roll motor was 1.0. We used a frequency controller to adjust the roll speed during briquetting. This
could have reduced the actual voltage in the line well below 460 V. Therefore, due to the reduction in the voltage and the possible energy loss in the frequency controller drive
system, we believe that the energy consumed by the roll motor would have been lower by about 30% of the energy values given in the above table.
w
Energy consumed by screw feeder motor (MJ/t) = [1.732 * 460 V * (current to screw feeder motor during briquetting in A - current to screw feeder motor during no load
condition in A) x power factor x 3.6] / kg of briquettes per hour.
Current to screw feeder motor during no load condition was 1.9 A. A power factor of 0.95 was assumed for screw feeder motor.
[cI
Specific energy for roll press briquetting process per unit mass of briquettes (MJ/t) = Energy consumed by roll motor (MJ/t) + energy consumed by screw feeder motor (MJ/t).
Note that the energy involved for the steam conditioning was not included in the specific energy for roll press briquetting process.
idl
Specific energy for roll press briquetting process per unit decrease in volume of the biomass feedstock (MJ/m3) = Specific energy for roll press briquetting process (MJ/t) x
(bulk density of the briquettes measured after one week in t/m3 - bulk density of the grind in t/m3).
w
Percentage of energy in material (%) = Specific energy for roll press briquetting process (MJ/t) x 100 / energy content of the biomass material (MJ/t).
Energy content of corn stover = 20 MJ/kg (Pordesimo et al., 2005). Energy content of switchgrass = 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004).
[f]
NA = Data not available.

303
Table 6.10. Properties of corn stover and switchgrass pellets measured immediately after making the pellets (mean standard
deviation, if given).

Pelleting conditions Data measured immediately after making the pellets


Moisture Temperature L/D ratio of the Fines output Throughput of Temperature of Bulk density of Durability of
content of of the grind die along with pellets pellets (C) pellets (kg m"3) pellets (%)
1
the grind (C)(n=l) (D = 9.5mm)* pellets (%) (kg h" ) (n=l) (n=2) (n=2)
(% w.b.) (n (n = 1) # (n = 1)
-3)
Corn stover (Hammer mill screen size = 2.4 mm (3/32 in.); Particle size of the grind = 0.34 mm 0.33)
19.8 0.2 12.2 5.3 1.1 180.9 67.2 554.6 8.9 94.6 0.2
19.3 0.1 1U 6^0 4.4 225.5 75J6 603.6 4.2 94.5 0.02
Corn stover (Hammer mill screen size = 4.0 mm (5/32 in.); Particle size of the grind = 0.36 mm 0.38)
21.80.3 1U M L2 162.6 75U6 552.5 13.7 94.0 0.8
Switchgrass (Hammer mill screen size =2.4 mm (3/32 in.); Particle size of the grind = 0.49 mm 0.31)
20.8 0.3 11.1 5.3 3.9 85.9 70.0 522.4 3.8 78.1 0.3
20.0 0.9 1_U 6J) 43 176.3 8U 569.4 10.1 84.8 0.3
* L = length of the holes in the ring-die; D = diameter of the holes in the ring-die.
* Fines output along with pellets (%) = (kg of fines * 100) / (kg of fines + kg of pellets).

304
Table 6.11. Properties of corn stover and switchgrass pellets measured after one week of storage at room temperature (mean:
standard deviation, if given).
Pelleting <
conditions Data measured after one week of storage of pellets at room temperature
Moisture Temperature L/D Temperature Moisture Properties of individual pellets Bulk density Durability of Pellet Angle of
content of of the grind ratio of of pellets (C) content of (n = 5) of pellets pellets (%) hardness repose
the grind fC) the die (n = 1) pellets (%Diameter Length Density (kgm3) (n = 2) (along (deg)
(% w.b.) (n = l) (D = 9.5 w.b.) (mm) (mm) (kg m"3) (n = 2) diameter) (n = 3)
(n = 3) mm)* (n = 3) (N)
(n = 10)
Corn stover (Hammer mill screen size = 2.4 mm (3/32 in.); Particle size of the grind = 0.34 mm 0.33)
19.8 0 . 2 12.2 5.3 67.2 15.4 0 . 1 9.8 0 . 3 22.7 1029.4 554.6 94.3 0.5 200.9 21.3
3.7 161.4 15.6 52.5 0.6
19.3 0 . 1 11.1 6.0 75.6 14.2 0 . 2 9.7 0 . 1 24.0 1110.7 616.2 95.8 1 . 4 224.2 19.3
4.4 32.1 7.8 99.3 0.6
Corn stover (Hammer mill screen size = 4.0 mm (5/32 in.); Particle size of the grind =0.36 mm 0.38)
21.8 0 . 3 11.1 6.0 75.6 15.1 0 . 2 9.6 0 . 2 22.9 1065.5 542.7 94.9 0.3 196.5 22.0
3.2 111.4 7.4 52.4 1.7
Switchgrass (Hammer mill screen size =2.4 mm (3/32 in.); Particle size of the grind = 0.49 mm 0.31)
20.8 0 . 3 11.1 5.3 70.0 12.2 0 . 1 9.6 0 . 1 20.7 1060.1 533.5 72.6 0 . 9 148.1 21.0
3.7 115.7 4.8 81.8 1.0
20.0 0.9 11.1 6.0 81.1 10.7 0 . 2 9.6 0 . 1 22.2 1107.5 570.5 86.5 1 . 4 216.3 20.7
3.7 46.2 7.0 61.6 1.2
* L = length of the holes in the ring-die; D = diameter of the holes in the ring-die.

305
Table 6.12. Specific energy consumption for the pelleting process (mean standard deviation, if given).

Pelleting conditions Current drawn during Specific energy consumption for pelleting
pelleting by pelleting mill process
Moisture Temperature of the L/D ratio of the die motor and feeder motor (A) MJ/t MJ/m3 Percentage of
content of grind (C)(n = l) (D = 9.5 mm) & energy in
the grind (% material ( % )
w.b.)
(n = 3)

Corn stover (Hammer mill screen size = 2.4 mm (3/32 in.); Particle size of the grind = 0.34 mm 0.33)
19.8 0.2 12.2 5.3 70.0 196.5 77.4 0.98
19.3 0.1 11.1 6.0 75.0 189.1 86.2 0.95
Corn stover (Hammer mill screen size = 4.0 mm (5/32 in.); Particle size of the grind 0.36 mm 0.38)
21.8 0.3 11.1 6.0 75.0 262.3 105.9 1.31
Switchgrass (Hammer mill screen size = 2.4 mm (3/32 in.); Particle size of the grind = 0.49 mm 0.31)
20.8 0.3 11.1 5.3 70.0 413.8 129.8 2.43
20.0 0.9 11.1 6.0 95.0 403.2 141.3 2.37
L = length of the holes in the ring-die; D = diameter of the holes in the ring-die.
* Specific energy consumption per unit mass of pellets (MJ/t) = [1.732 * 240 V * (current to pelleting mill motor and feeder motor during pelleting in A -
current to pellet mill motor and feeder motor during no load condition in A) * power factor * 3.6 ] / kg of pellets per hour.
Current to pellet mill motor and feeder motor during no load condition was 45.0 A. A power factor of 0.95 was assumed. Also, note that no steam conditioning
was involved for any of the pelleting tests.
** Specific energy consumption per unit decrease in volume of the biomass feedstock (MJ/m3) = Specific energy consumption in MJ/t * (bulk density of the
pellets measured after one week in t/m3 - bulk density of the grind in t/m3).
v
Percentage of energy in material (%) = Specific energy consumption for pelleting process (MJ/t) * 100 / energy content of the biomass material (MJ/t).
Energy content of corn stover = 20 MJ/kg (Pordesimo et al., 2005).
Energy content of switchgrass = 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004).

306
6.8 Appendix

In this section, additional scanning electron microscopy (SEM) images


obtained for corn stover and switchgrass grinds (before briquetting or pelleting), and
cross sections (i.e., fractured surfaces) of briquettes and pellets are given. The SEM
observations were made at magnifications of X40, X250, and X600. Table 6.A1 shows
that the SEM images of the corn stover particles appear to be bare without any sign of
coating of natural binding components; however, on the switchgrass particles,
expression of some natural binding components (due to size reduction) can be
observed. The SEM images of the briquettes and pellets show that the particles are
covered with a layer of natural binders (tables 6.A2-6.A5). The natural binders are
squeezed out of the particles during roll press briquetting or pelleting due to the
application of mechanical pressure.

Table 6.A1. SEM images of corn stover and switchgrass grinds (before roll press
briquetting or pelleting).

Magnification at X40 Magnification at X250 Magnification at X600

307
Table 6.A2. SEM images of com stover briquettes made using roll press briquetting
machine.

Magnification at X40 Magnification at X250 Magnification at X600


Corn stover briquette, cross section (particle size of the grind = 0.34 mm; moisture content of the grind =
7.3% w.b.; preheating temperature = 22.0C; briquette durability = 66.7%)

K W^^^J

'jf tr L^t-
Corn stover briquette, cross section (particle size of the grind = 0.34 mm; moisture content of the grind =
15.0% w.b.; preheating temperature = 20.3C; briquette durability = 86.5%)

V/Xjr " ' *'i * * fW f ":-:."'.

c. $'~4*:
. V. " % V * v. ., , &
, .. I" -

Corn stover briquette, cross section (particle size of the grind = 0.34 mm; moisture content of the grind =
9.9% w.b.; preheating temperature = 77.3C; briquette durability = 89.8%)

4 * -

I irt^f*

Corn stover briquette, outer surface (particle size of the grind = 0.34 mm; moisture content of the grind =
9.9% w.b.; preheating temperature = 77.3GC; briquette durability = 89.8%)
A-
" / . - *

*V
' .' .. : /.?>,,'- ?

'**' <?k'Jt

i+sffat tfeKIKMKiv

308
Table 6.A3. SEM images of switchgrass briquettes made using roll press briquetting
machine.

Magnification at X40 Magnification at X250 X600


Switchgrass briquette, cross section (particle size of the grind = 0.49 mm; moisture content of the
grind = 9.2% w.b.; preheating temperature = 19.0C; briquette durability = 39.5%)

Switchgrass briquette, cross section (particle size of the grind = 0.49 mm; moisture content of the
grind = 11.7% w.b.; preheating temperature = 68.3C; briquette durability = 70.0%)

Switchgrass briquette, outer surface (particle size of the grind = 0.49 mm; moisture content of the
grind = 11.7% w.b.; preheating temperature = 68.3C; briquette durability = 70.0%)

309
Table 6. A4. SEM images of corn stover pellets made using ring-die pelleting machine
(no preheating/steam conditioning was involved).

Magnification at X40 Magnification at X250 Magnification at X600


Corn stover pellet, cross section (particle size of the grind = 0.34 mm; moisture content of the
grind = 19.3% w.b.; die hole diameter = 9.5 mm; die 1/d ratio = 6; pellet durability = 95.8%)

Corn stover pellet, outer surface (circumference) (particle size of the grind = 0.34 mm; moisture
content of the grind = 19.3% w.b.; die hole diameter = 9.5 mm; die 1/d ratio = 6; pellet durability =
95.8%)
i
v. r
. / I-
>
. . " A v.- - ; , y

310
Table 6.A5. SEM images of switchgrass pellets made using ring-die pelleting
machine (no preheating/steam conditioning was involved).

Magnification at X40 Magnification at X250 Magnification at X600


Switchgrass pellet, cross section (particle size of the grind = 0.49 mm; moisture content of the
grind = 20.0% w.b., die hole diameter = 9.5 mm; die 1/d ratio = 6; pellet durability = 86.5%)

Switchgrass pellet, outer surface (circumference) (particle size of the grind = 0.49 mm; moisture
content of the grind = 20.0% w.b.; die hole diameter = 9.5 mm; die 1/d ratio = 6; pellet durability =
86.5%)

311
Chapter 7
CONSTITUTIVE MODELS FOR DENSIFICATION OF BIOMASS

7.1 Abstract

A constitutive model is an important input to the computer simulation of


densification of powders. An elasto-visco-plastic solid model was developed for
characterizing the compression behaviors of corn stover and switchgrass grinds. The
constitutive model characterized the biomass grinds through five parameters: elastic
modulus, strength coefficient, strain hardening exponent, viscous coefficient, and
frictional loss factor. The constitutive model parameters were found to be affected by
the densification conditions (i.e., applied stress, particle size, moisture content, and
preheating temperature). During compression of biomass grinds, the development of
the structure of the compact inside the die was captured by the elastic modulus. Also,
the elastic modulus and viscous coefficient correlated with the durability and
compressive strength of briquettes made from corn stover and switchgrass at various
densification conditions.

7.2 Introduction

Corn stover and switchgrass are two important biomass materials identified for
the sustainable production of renewable energy, fuels, and chemicals in the U.S.
(DOE, 2005; McLaughlin and Kszos, 2005). For successful use of these renewable
biomass materials in various applications, densification is one of the essential pre-
processing steps considered in the biomass conversion process. Densification of low
bulk density corn stover and switchgrass into pellets/briquettes would increase their
bulk density by more than three-fold as well as provide more efficient handling,
storage, transportation, and use of these biomass materials (Chapter 6; Kaliyan and
Morey, 2007). Therefore, studying the densification behaviors of corn stover and
switchgrass through experiments and modeling would help to understand the

312
densification mechanisms, to produce high quality densified products such as
pellets/briquettes, and to design energy efficient densification machines.

7.2-7 Densification Mechanism

The evolution of pressure with time during the compression of ceramic,


metallurgical, and pharmaceutical powders is similar to the curve shown in fig. 7.1
(Rumpf, 1962; Ghebre-Sellassie, 1989; Pietsch, 2002). During the initial stage of
compression, particles rearrange themselves under low pressure to form close packing
and inter-particle air is removed from the bulk material. During this phase, the original
particles retain most of their properties, although energy is dissipated due to inter-
particle and particle-to-wall friction.
At high pressures, the particles are forced against each other even more and
undergo elastic and plastic deformation, thereby increasing inter-particle contact.
Because the particles approach each other closely enough, short range bonding forces
like van der Waal's forces, electrostatic forces and sorption layers become effective.
Under yield stress, brittle particles may fracture leading to mechanical interlocking of
particles. At still higher pressures, reduction in volume continues until the density of
the compacted material approaches the true density of the particles. If the pressure
applied is high enough to generate heat, some of the components of the particles will
melt locally. Once cooled, the molten material forms very strong solid bridges. After
removal from the die, the compacted material relaxes (elastic spring-back) due to
residual stresses and release of compressed air.
The above densification mechanism has been postulated by Mani et al. (2004)
for densification of wheat straw, barley straw, corn stover, and switchgrass, and by
Tabil and Sokhansanj (1996) and Adapa et al. (2005) for densification of alfalfa.
Bilanski and Graham (1984) and O'Dogherty (1989) reported that at high compaction
pressures, biomass particles would be flattened/crushed leading to damage to the cell
structure which would release cell contents such as protein and pectin. These
compounds would act as natural binders and aid the adhesion of biomass particles.

313
Presence of natural binding compounds in the biomass particles is a major
difference between biomass particles and ceramic or metallic or pharmaceutical
powders. Other differences are biomass particles are porous; biomass particles contain
multi-compounds (e.g., stem and leaf) with complex mechanical properties; and
biomass particles are readily compressible.
According to Mohsenin (1986), for biomass materials, even at very small
strains, there is always a part of the deformation that is irrecoverable. The major part
of the residual deformation is due to the presence of pores or air spaces, weak ruptured
cells on the surface, microscopic cracks, and other discontinuities which may exist in
the structure of the material. This can be viewed as an analog to the phenomenon of
slip and dislocation in metals due to imperfections in their crystal structures. These
defects in crystal structure are believed to be responsible for plastic or permanent
deformation which results from slip, or glide, of one part of the body over the other.

7.2-2 Constitutive Models

Several researchers have demonstrated that the finite element method (FEM) is
a powerful technique for simulating densification and predicting mechanical behaviors
(i.e., stress-strain, or stress-density response) of powders (Huang and Puri, 2000).
FEM modeling of densification of powders would involve equilibrium equations
(material independent), compatibility equations (material independent), constitutive
equations for the powder material (material dependent), geometry of the problem, and
initial and boundary conditions. The equilibrium and compatibility equations are built-
in in any FEM software. Geometry, and initial and boundary conditions are known to
the specific densification problem. The constitutive equations need to be input to the
FEM software. Constitutive equations are usually modeled from uniaxial or triaxial
compression data measured from laboratory experiments (Huang and Puri, 2000;
Mittal and Puri, 2002; Sinka et al., 2004).
What are constitutive models? Different materials of the same geometry may
respond differently under identical external effects. Such a difference in response is
often attributed to the inherent constitution of the material. The response of a
314
particular material is described mathematically by so-called 'constitutive models'
(Macosko, 1994; Haddad, 1995). In general terms, constitutive models establish the
connection between the stimuli (e.g., stress) acting on the material specimen and the
evolution of the occurring response (e.g., strain). An engineering approach to
predicting the mechanical behavior of powders is to develop a constitutive model.
Constitutive models are important because they allow a designer to quantitatively
predict the mechanical behavior. Constitutive models can be divided into two groups
(Tripodi et al., 1992; Aydin et al., 1997): empirical models and rational models.

7.2.2-1 Empirical models

Empirical models are developed by collecting experimental data of a


particulate material under specific loading conditions in the laboratory using uniaxial
compression devices such as a piston-cylinder apparatus, and then statistically
determining the equations that will closely match the data. Usually, the axial pressure
(stress) and relative density data obtained from the compression test will be fitted to
the following form (Heckel, 1961):

(
In AP + B (1)
\X~PrJ

Where, p r is the relative density of the compacted material, P is the applied pressure,
and A and B are constants. More than 20 different empirical models similar to the form
of equation 1 have been developed for different particulate materials (Panelli and
Filho, 2001). The model constants (A and B) are interpreted to identify the
deformation characteristics of the powders. For example, Heckel (1961) postulated
that the constant A gave a measure of the ability of the compacted material to densify
by plastic deformation, and constant B represented the degree of packing achieved at
low pressures as a result of rearrangement before appreciable amounts of inter-particle
bonding occurred.

315
Compaction equations relating stress and compact density or compact volume
were developed for rice, barely, and wheat straws by Faborode and O'Callaghan
(1986) and Ferrero et al. (1991); for alfalfa by Tabil and Sokhansanj (1996); and for
wheat straw, barley straw, corn stover, and switchgrass by Mani et al. (2004).
Empirical models relating density and energy consumed for the compression of
biomass materials have also been developed (e.g., Abd-Elrahim et al., 1981).
There are several limitations with the empirical models: (i) no single equation
can perfectly represent the densification phenomenon over the entire range of
pressures (Panelli and Filho, 2001); (ii) the model ignores radial stress transmission
and wall friction (Aydin et al., 1997). In a few pressure-density models, the wall
friction effect was included (e.g., Aydin et al., 1997). To include the wall friction
effect, the die has to be instrumented to measure stress on the wall, and at the bottom
of the compact; (iii) the model cannot explain the densification mechanisms; (iv) the
model cannot characterize the powder material such as elastic or plastic or viscoelastic
material; and (v) the model cannot be used for predicting temporal and spatial
distribution of stress and density inside the compact.
In addition, several authors defined indices to characterize the elasticity and
plasticity properties of the powder (e.g., Malamataris et al., 1984; Antikainen and
Yliruusi, 2003). These indices are calculated from data obtained by measurement of
macroscopic dimensional variations in compacts during compression, during in-die
relaxation, and after ejection from the die. Compression behavior of powders has also
been related to the elastic and plastic energy values calculated from force-deformation
curves (e.g., Garekani et al., 2000). However, these indices do not give any
information on the stress gradients and heterogeneous density changes in the volume
of compact during die compaction (Michrafy et al., 2002).

316
7.2.2-2 Rational models

Rational models are developed by applying physical laws to describe the


stress-strain behavior of the material. These models are based on either microscopic or
macroscopic scale parameters. Microscopic models consider each particle in a bulk
solid as a distinct entity and then predict the stress-strain behavior based on a
distribution of inter-particle forces. However, microscopic models are still under
development (Huang and Puri, 2000; Zavaliangos, 2002; Mani et al., 2003).
Macroscopic models treat the bulk solid as a continuum or interacting continua
and describe force-deformation characteristics of the materials as a whole. The
macroscopic models include various endochronic models and rheological models.
Endochronic models are time-dependent constitutive equations developed based on
thermodynamic theory instead of plasticity. The endochronic theory is still in
developmental stages for bulk solids (Tripodi et al., 1992).
There is a collection of rheological models found in the literature, such as the
Drucker-Prager-Cap model (Drucker and Prager, 1952), the Cam-Clay model
(Schofield and Wroth, 1968), and the Dimaggio-Sandler model (Dimaggio and
Sandler, 1971). All of these models have been developed for application in soil
mechanics and geotechniques. Also, these models have been adopted to model the
compaction of metal powders (Zavaliangos, 2002), ceramic powders (Aydin et al.,
1996), and pharmaceutical powders (Michrafy et al., 2002; Sinka et al., 2003;
Cunningham et al., 2004; Wu et al., 2005). All of these models are based on the elasto-
plastic theory (time independent models) and are essentially the same, even though
their yield criteria, which define the admissible stresses, are expressed in different
forms and can be defined with one or more yield surfaces (Michrafy et al., 2002).
Also, elasto-viscoplasticity theory has been used to model time-dependent
compression behaviors of powders (Adachi and Oka, 1982; Huang and Puri, 2000;
Mittal and Puri, 2002).

317
7.2-3 Rheological Models

Rheology is defined as "a science devoted to the study of deformation and


flow", and more recently as "the study of those properties of materials that govern the
relationship between stress and strain" (Peleg and Bagley, 1983; Macosko, 1994).
Stress is defined as the intensity of force components acting on a body and is
expressed in units of force per unit area. Strain is the change in size or shape of a body
in response to the applied force. It is expressed as the change in relation to the original
size or shape. There are two modes of strain expression: engineering strain, which
expresses the change in shape at any given time with respect to the original shape at
time = 0, and true (natural) strain, which expresses the continuous change in shape as
the stress is being applied.
Rheologically, a material may deform in three ways: elastic, plastic, or viscous
(Peleg and Bagley, 1983; Macosko, 1994). In an ideal elastic body, deformation (or
strain) occurs instantly the moment stress is applied, is directly proportional to stress,
and disappears instantly and completely when stress is removed. In an ideal plastic
body, deformation does not begin until a certain value of stress (called yield stress) is
reached. Deformation is permanent and no recovery occurs when stress is removed. In
an ideal viscous body, deformation occurs instantly the moment stress is applied. The
stress is proportional to the rate of strain. The deformation is not recovered when
stress is removed. These three types of behaviors are denoted by a spring element, a
friction element, and a dashpot element, respectively, in rheological models (Peleg and
Bagley, 1983; Macosko, 1994).
Developing rational models using rheological principles involves the use of
combinations of discrete spring and dashpot elements (Peleg and Bagley, 1983;
Mohsenin, 1986; Sitkei, 1986; Macosko, 1994; Haddad, 1995). The arrangement of
the spring and dashpot elements within the models depends on the characteristics of
the particulate material and whether the material is a viscoelastic fluid or solid. For
example, the coupling of a spring and dashpot in series is termed a Maxwell
viscoelastic fluid (fig. 7.2). The coupling of a spring and dashpot in parallel is termed
a Kelvin-Voigt viscoelastic solid (fig. 7.2). In the Maxwell model, the total strain is
318
additive of strains in the spring and dashpot elements, and same stress is carried
through the spring and the dashpot. In the Kelvin-Voigt model, the total stress is
divided between the spring and the dashpot, and the dashpot and the spring take the
same strain.
The rheological model is developed such that the model should behave
qualitatively, to some degree of approximation, in a manner similar to that of an actual
material. The rheological behavior expressed as a function of stress, strain, and time
can be used to predict the mechanical behavior of the material under various loading
conditions. When the stress-strain relationship is a function of time alone and not of
the stress magnitude, the material is said to be linearly viscoelastic. However, if the
viscoleastic behavior is a function of time and stress, the material is said to be non-
linearly viscoelastic, or viscoplastic (Macosko, 1994).
The mechanical behavior of biological materials depends on stress, strain,
strain rate, moisture content, temperature, and size and shape of the biological
materials (Mohsenin, 1986; Sitkei, 1986). Biological materials behave in complex
ways when loaded, and exhibit time-dependent, force-deformation characteristics.
Therefore, rheological models have been used to simulate stress relaxation, elastic
recovery, and creep in biological materials such as apple, tomato, potato, egg, cheese,
butter, corn, wheat, barley, and forage stalks (alfalfa, clover, timothy, and corn)
(Mohsenin, 1986). However, only a few rheological models have been developed for
the compression of biological materials (Munoz and Herrera, 2002).
Rehkugler and Buchele (1969) proposed a five-element model (two spring
elements, one dashpot element, and two Coulomb friction elements) to predict the
compression and relaxation densities of forage in a closed-end die to simulate the
wafering process, where a step input of axial stress was involved. In this study, they
did not obtain the coefficients of their rheological model. Mohsenin and Zaske (1976)
used the generalized Maxwell model to study the stress relaxation of the compacted
materials produced from grass, hay, and wood powders. Rippie and Danielson (1981)
used linear viscoelastic theory to study the unloading and in-die relaxation behaviors
of tablets produced in a tablet press. They used a spring element model and Kelvin
solid model to predict these two behaviors.
319
Peleg (1983) developed a rheological model to predict mechanical behaviors
such as compression, creep, stress relaxation, and expansion of materials such as
fruits, vegetables, cardboard, and plastic foams. Compression was modeled as two
separate regions: elastic stage and elasto-plastic stage. The model consisted of two
non-linear springs, a viscous damper (dashpot), and a Coulomb friction-damper.
Bilanski and Graham (1984) developed a five-element non-linear rheological
model to simulate the wafering process where bulk forage materials underwent
instantaneous compression and creep during loading, and elastic rebound and creep
recovery during unloading. The model involved two spring elements, one rigid body
element, and two dashpot elements. A creep compliance model was derived as a
function of time and stress to represent the non-linear viscoelastic behaviors of alfalfa,
grass, and a mixture of alfalfa and grass materials during the wafering process.
Graham and Bilanski (1984) applied the non-linear viscoelastic model developed by
Bilanski and Graham (1984) to study the behavior of alfalfa and grass during wafering
in the pressure range of 39 to 51 MPa. Wafering was accomplished in a closed-end die
by applying an impulsive axial stress of constant magnitude and holding this stress for
a predetermined period of time (about 5 s). The creep compliance model parameters
were found to be affected by stress applied, the type of forage material, length of cut
of the forage material (1.0 to 15.0 cm), and moisture content of the forage (13.0 to
29.0% w.b.).
Faborode and O'Callagan (1989) developed a five-element non-linear
rheological model for analyzing the compression and relaxation behavior of ground
and unchopped barley straw samples. The compression curve was divided into three
stages: inertial stage up to an initial density (critical density), elastic stage up to an
elastoplastic point, and thereafter elastoplastic stage. The compression was modeled
from the creep response of the material. Two stages were identified for the
compression process, namely, the dispersed and the dense stages, which met at critical
density. The critical density corresponds to the compressed density at the point where
the air voids have become insignificant, and only then does elastic resistance of the
material particles begin to dominate the compression response. The Faborode and
O'Callagan (1989) model was an extension of the Peleg (1983) model. Faborode and
320
O'Callagan (1989) introduced the concept of inertial deformation (critical density),
and after the critical density, the Peleg (1983) model was applied to model the dense
phase compression. Faborode and O'Callagan (1989) used a dashpot element to
simulate the instantaneous irreversible deformation during the inertial deformation
stage. The model constants were evaluated using data obtained from wafering of
barley straw in a closed-end die with step loads to 50 MPa.
Munoz and Herrera (2002) used rheological models to predict the compression
behavior of feed containing 75% (wt.) corn stalks and leaves. The compression curve
was divided into two regions: plateau region (characterized by Maxwell fluid) and
densification region (characterized by elastic spring). One limitation of this study was
that the friction on the cylinder walls was not considered. Ren (1992) and Thoemen et
al. (2006) developed rheological models to describe the development of vertical
density profile and of internal stresses within wood-furnish mats during hot pressing.
The coefficients of rheological models were obtained as a function of temperature,
moisture, and density. Thoemen et al. (2006) incorporated the rheological model with
a three-dimensional simulation model that accounted for those mechanisms most
important during the hot pressing process, including heat and mass transfer inside the
mat and adhesive cure.

321
7.3 Objectives

The objectives of this study were:

1. To develop constitutive models for the uniaxial compression of corn stover and
switchgrass grinds.

2. To study the effect of compression conditions (particle size, moisture content, and
preheating temperature) on the constitutive model parameters of corn stover and
switchgrass.

3. To correlate constitutive model parameters to compressive strength and durability


of the briquettes made from corn stover and switchgrass.

7.4 Materials and Methods

7.4-1 Constitutive Model Development

A constitutive model (i.e., stress-strain relationship) for the compression stage


of the densification of corn stover and switchgrass was developed based on the
mechanical analogy concept developed in the field of rheology. The compression
stage was viewed as a combination of two distinct deformation processes: inertial
deformation, and elasto-visco-plastic deformation.
Initially, at low pressures, air is expelled from the die, and the biomass
particles re-arrange due to inertial forces. This stage of compression was called the
inertial deformation stage. The inertial deformation stage of the compression process
was not modeled in this study. This is because the inertial deformation data obtained
from the laboratory compression study would be affected by several variables, such as
biomass grind filling method into the die (Mittal and Puri, 1999). Also, with an
industrial type of densification machine (e.g., roll press briquetting machine), air
removal from the biomass grind would take place in a separate location (e.g., screw
feeder) well ahead of the actual place of compression (e.g., rolls) where biomass
particles are compacted to form durable, densified products (Pietsch, 2002). Therefore,
322
for computer simulation of densification of corn stover and switchgrass, the
constitutive model for the elasto-visco-plastic deformation stage would be sufficient.
After a certain stress (o\), high inter-particle friction between the particles
would prevent any further inter-particle movement. The subsequent reduction of
volume is therefore accomplished by elastic deformation of the biomass particles.
Because of the complex and diverse mechanical behavior of individual particles in the
biomass grind, at any instant during compression, some particles may undergo elastic
deformation and some particles may undergo plastic deformation. Examples of plastic
deformation are fragmentation of particles, rearrangement of fractured particles, and
pressing of particles beyond the elastic limit/yield point causing release of natural
binders from the particles. If the applied stress is adequate, squeezing of natural
binders from some biomass particles would occur instantly. The expressed natural
binders may soften/melt if the prevailing moisture and temperature are favorable. This
would lead to a solid bridge type of bonding between particles. The squeezing of cell
compounds (i.e., natural binders) is equivalent to viscous flow (or viscous dissipation).
In wet pressing, squeezing of intra-fiber water out of the fiber was identified as a
viscous phenomenon by Lobosco and Kaul (2001). Also, frictional forces are always
acting between the particles and the wall of the die.

Therefore, after the inertial deformation stage, the total deformation of the
biomass bed is caused by simultaneous cumulative action of elastic deformation,
plastic deformation, viscous dissipation, and frictional loss. At any instant, the total
stress measured is equivalent to the sum of the stress involved for elastic deformation,
plastic deformation, viscous dissipation, and frictional loss. To depict the simultaneous
action of these individual deformation mechanisms, an elasto-visco-plastic solid
mechanical analog was created to model the deformation of the biomass particles after
the inertial deformation.
The elasto-visco-plastic solid mechanical analog was made of parallel
connection of a strain hardening spring element, a dashpot element, and a Coulomb
friction element (fig. 7.3) The strain hardening spring element represents the elastic
deformation and plastic deformation of biomass particles, the dashpot element
represents the viscous dissipation due to squeezing of cell components (i.e., time
323
dependent loss during compression), and the Coulomb friction element represents the
frictional loss (i.e., time independent loss during compression). This stage of
compression was called elasto-visco-plastic deformation stage. The derivation of the
constitutive model for this stage of compression is given below.

The constitutive behavior of the strain hardening spring element was modeled
as (Peleg, 1983; Faborode and O'Callagan, 1989):

a - E s + R E" (2)

Where, a = stress, E = elastic modulus, s = natural strain, R = strength coefficient


(plastic modulus), and n = strain hardening exponent.

The constitutive behavior of the dashpot element was modeled as (Peleg, 1983;
Faborode and O'Callagan, 1989; Macosko, 1994):

= n^r (3)
at

Where, a = stress, n = viscous coefficient, and ds/dt = natural strain rate.

The constitutive behavior of the Coulomb friction element was taken as a


constant (Peleg, 1983; Faborode and O'Callagan, 1989):

a = af (4)

Therefore, the total stress in the elasto-visco-plastic solid mechanical analog


material (fig. 7.3) is the sum of the stress in the spring element, dashpot element, and
the Coulomb friction element:

<T = Ee + Rn+r1 + af (5)

324
Equation 5 is applicable for s > Sj and a > o\. Where, a\ = stress corresponding
to the end of the inertial deformation, and Sj = natural strain corresponding to the end
of the inertial deformation. Equation 5 is a non-linear elasto-visco-plastic solid model.
The constitutive model given by eq. 5 is a modified version of the constitutive models
developed by Peleg (1983), and Faborode and O' Callagan (1989) for compression of
biological materials such as apples and ground barley straw.

7.4-2 Constitutive Model Parameter Estimation

The compression curve (i.e., stress-natural strain curve) for corn stover and
switchgrass grinds was separated into inertial deformation stage, and elasto-visco-
plastic deformation stage (fig. 7.4). The stress (CTJ) at which the change in inertial to
elasto-visco-plastic deformation would occur was found by analyzing the measured
pressure-time curve.
During the initial stage of the loading, due to air removal and rearrangement of
the particles, the pressure-time curve has increasing and decreasing patterns of
pressure values. That is, the pressure gradients (dP/dt) change from positive to
negative values until the elastic force dominates, which is indicated by an increasing
positive pressure gradient with time. The end point of the inertial stage was identified
as the stress and natural strain corresponding to the last negative pressure gradient.
After this point, the pressure gradient would always be positive. A typical inertial
deformation stage is shown in fig. 7.4B. The stress (cij) and natural strain (si) values at
the end of the inertial deformation were obtained from the experimental stress-natural
strain curve (or data).
The data on stress-natural strain for the elasto-visco-plastic deformation stage
(fig. 7.4C) were used to estimate the constitutive model parameters (E, R, n, r\, and
Of). These parameters were estimated as a function of stress. Estimating model
parameters for the entire range of the stress (i.e., 0 to 150 MPa) is not correct because
the structure of the compact is changing with applied stress, which would in turn
change the mechanical properties of the compact (Ren, 1992; Thoemen et al., 2006).

325
The measured stress-natural strain data were divided into several stress range
segments such as <jj to 5 MPa, 5 to 25 MPa, 25 to 50 MPa, 50 to 100 MPa, and 100 to
150 MPa to fit the model parameters by piecewise nonlinear regression. The model
parameters were assumed to be constant at each stress range involved for the
parameter estimation. An example calculation is given in the section 7.8 Appendix to
show the procedure used to estimate the model parameters.
For each stress range, corresponding data on stress-natural strain were fitted to
the constitutive model (eq. 5) using SPSS software (SPSS 13.0 for Windows; SPSS
Inc., Chicago, IL). In the SPSS software, Levenberg-Marquardt non-linear multiple
regression algorithm with sum-of-squares convergence of 10"8 and parameter
convergence of 10"8 was used to estimate the model parameters.

7.4-3 Bio mass Compression Experimental Procedure

The data on time, force, and deformation of corn stover and switchgrass grinds
during compression were obtained using a uniaxial, piston-cylinder compression
(densification) apparatus. The preparation including hammer milling, particle size
analysis, and moisture adjustment of corn stover and switchgrass grinds used for this
study is given in Kaliyan and Morey (2006) [Chapter 4],
The compression cylinder (i.e., die) used for the study had an inner diameter of
18.85 mm. An INSTRON model 4206 universal testing machine (Instron Corporation,
Canton, MA) was used to apply a compression pressure of 150 MPa. The INSTRON
was controlled by a computer loaded with TestWorks 4.0 software (MTS Systems
Corp., Eden Prairie, MN). The details of the uniaxial, piston-cylinder compression
(densification) apparatus used for the study can be found in Kaliyan and Morey (2006)
[Chapter 4].

About 5.0-g of corn stover or switchgrass grind was added to the cylinder
through a funnel. Using a steel rod, the grind was stirred to help the flow of grind from
the funnel. For the cases with preheating to 75C, before compression of biomass
grind, the grind along with the cylinder was heated using heating tapes to achieve a
326
preheating temperature of 75C (Chapter 4; Kaliyan and Morey, 2006). The piston
attached to the crosshead of the INSTRON was actuated to compress the grind to a
maximum pressure of 150 MPa at a constant speed of 25.4 mm min"1 (1.0 in. min"1). A
compression speed of 25.4 mm min"1 was chosen because the Instron was able to stop
the application of pressure close to 150 MPa with this speed of the crosshead. At
higher speeds, the Instron stopped the application of pressure at much higher pressure
than 150 MPa because of the higher inertia of the Instron crosshead. The time, force,
and the distance traveled by the crosshead were recorded during compression at a data
collection rate of 40 Hz.
The compression of the biomass grinds in the die resulted in briquettes. The
briquettes were ejected from the die using a briquette ejection apparatus (Chapter 4;
Kaliyan and Morey, 2006). In this study, the cylinder was not lubricated during
compression, but periodically cleaned using a vacuum cleaner. Also, no binding
agents (i.e., additives) were used for the study.
Immediately after ejection from the die, unit density of the briquettes (i.e.,
density of individual briquettes) was measured. Then, the briquettes were transferred
to zip-lock plastic bags and stored for one week at room temperature. Diameter and
height of the briquettes, percentage expansion in volume of the briquettes, and
moisture content of the briquettes (ASAE Standards, 2003) were measured after one
week of storage.
In addition, diametrical compressive strength of the briquettes was measured
using the Instron universal testing machine (Instron Corporation, Canton, MA; Model
4206) at a compression speed of 2.54 mm min"1. The compressive strength was
measured as the maximum force at which the briquette failed when loading the
briquette along the radial direction (Raghavan and Conkle, 1991). The briquette failure
load was measured from the force-time curve. During compression, the load increased
from zero to a maximum force, and then, the force declined suddenly due to briquette
failure. At the failure load, cracks were initiated at several locations on the briquettes
along the radial direction. With further increase in load beyond the failure load,
lamination of briquettes occurred.

327
7.4-4 Biomass Compression Conditions

Compression data of corn stover and switchgrass grinds were measured for the
following combination of conditions: (i) maximum compression pressure of 150 MPa
applied at a constant speed of 25.4 mm min"1 (1.0 in. min"1); (ii) grind obtained from
the hammer mill with a screen size of 3.0 mm (i.e., geometric mean particle diameter
of corn stover and switchgrass grind was 0.66 mm 0.32, and 0.56 mm 0.29,
respectively.), or grind obtained from the hammer mill with a screen size of 4.6 mm
(i.e., geometric mean particle diameter of corn stover and switchgrass grind was 0.80
mm 0.41, and 0.64 mm 0.30, respectively.); (iii) moisture content of the grind of
10 or 20% (w.b.); and (iv) grind temperature of 25C (i.e., room temperature) or 75C
(i.e., preheating temperature). Three replications were made at each compression
condition.
The preheating temperature of 75C was selected because the glass transition
temperature of the corn stover and switchgrass is about 75C (Chapter 4; Kaliyan and
Morey, 2006). Preheating of biomass close to its glass transition temperature would
activate the natural binders in the biomass such as lignin, starch, protein, fat, and water
soluble carbohydrates, which would in turn help produce strong and durable briquettes
(Chapter 6; Kaliyan and Morey, 2007).

328
7.5 Results and Discussion

Corn stover briquettes made for the compression conditions involved in this
study had diameter of 19.1 to 19.3 mm, and height of 14.9 to 23.8 mm, unit briquette
density of 696 to 1164 kg m" , diametrical compressive strength of 412 to 946 N, and
durability of 62 to 97% (table 7.1). Switchgrass briquettes made for the compression
conditions studied had diameter of 19.1 to 19.6 mm, and height of 16.3 to 38.1 mm,
unit briquette density of 419 to 1057 kg m~ , diametrical compressive strength of 18 to
516 N, and durability of 0 to 63% (table 7.1). At similar compression conditions, corn
stover produced better quality briquettes than switchgrass. This could be due to
differences in their cell structures, compositions, and mechanical properties. Kaliyan
and Morey (2007) [Chapter 6] reported that corn stover had higher amounts of natural
binding components than switchgrass. The natural binding components in these
biomass materials are lignin, starch, protein, fat, and water soluble carbohydrates and
they could be activated by providing sufficient moisture and temperature (Chapter 6;
Kaliyan and Morey, 2007).
The constitutive model (elasto-visco-plastic solid model) fitted the measured
compression data of corn stover and switchgrass grinds with sum of squared residuals
(SSR) of 0.03 to 8.6, mean square residuals (MSR) of 10"4 to 0.4, and coefficient of
determination (R ) of 1.0. Table 7.2 gives the constitutive model parameters estimated
for corn stover and switchgrass at various compression conditions. Figure 7.5 shows
an example comparison between the measured stress values and the values predicted
by the constitutive model for corn stover and switchgrass at about 20% (w.b) moisture
content. Excellent agreement between the measured stress and predicted stress values
shows that the developed constitutive model is capable of capturing the mechanical
behaviors of the biomass grinds during compression.
The inertial deformation (i.e., air removal and particle rearrangement) occurred
until about 1.3 0.3 MPa and 1.1 0.3 MPa for corn stover and switchgrass,
respectively (table 7.2). Corn stover required higher energy for inertial deformation
than switchgrass because the bulk density of the corn stover grinds (149 to 185 kg m")
was lower than for switchgrass grinds (161 to 204 kg m"3) (table 7.1). After the inertial
329
deformation stage, the elasto-visco-plastic deformation of the biomass bed was
assumed to take place. During the elasto-visco-plastic deformation stage, the
remaining inter-particle air would be continually removed from the biomass grind.
Also, it is possible that some amount of air could be trapped inside the compact. In
addition to considering the mechanical behaviors of the inter-particle air and intra-
particle air, which would be small and negligible at higher pressures, the elasto-visco-
plastic deformation stage considers the deformation of the biomass particles. The
viscous coefficient in the constitutive model may capture this additional inter- and
intra-particle air removal effects.
Tables 7.2, 7.3 and 7.4 clearly show that the constitutive model parameters are
a function of stress. This is because the structure of the biomass bed would be
changing from a loose powder bed to a solid compact with the application of
compressive stress. For both corn stover and switchgrass, the elastic modulus
increased with stress applied (tables 7.3 and 7.4). The evolution of elastic modulus
from about 4 to 930 MPa showed that the structure of the compact inside the
densification cylinder (i.e., die) changed from a state of soft rubber (typical elastic
modulus = 3.4 MPa) to a state of hard rubber (typical elastic modulus = 1379 MPa).
For comparison, typical elastic modulus values of wood and steel are 8963 MPa and
100000 MPa, respectively (Mohsenin, 1986). The higher the elastic modulus, the
lower the volume expansion of the briquettes (table 7.1). This is because higher elastic
modulus would result in lower elastic deformation of particles, which would result in
reduced expansion (i.e., elastic spring-back) of briquettes. The elastic spring-back of
briquettes may also be affected by the stress relaxation and release of compressed air
from the briquettes. In addition, the higher the elastic modulus, the higher the unit
briquette density and diametrical compressive strength of briquettes (tables 7.1 and
7.2).
For both corn stover and switchgrass, smaller particle size (i.e., grind obtained
using hammer mill screen size of 3.0 mm) resulted in higher elastic modulus values
(tables 7.3 and 7.4). At higher temperature and moisture content, lower elastic
modulus values were predicted for corn stover (table 7.3). This may be due to the
plasticization effect of high moisture or high temperature on corn stover. For
330
switchgrass, increasing the moisture content from 8.5 to 19.9% (w.b.) slightly reduced
the elastic modulus values; however, preheating to 75C increased the elastic modulus
of the compact. Similarly, increasing the moisture content from 8.5 to 19.9% (w.b.)
reduced the compressive strength of switchgrass briquettes, but increasing the
temperature from 25 to 75C increased the compressive strength of switchgrass
briquettes (table 7.4). Also, from table 7.4, it can be seen that preheating is necessary
to produce durable switchgrass briquettes.
The strength coefficient (R) and strain hardening exponent (n) can be
combined as one parameter R1/n, because the plastic strain (ss) and stress (<rs) during
strain hardening may be related as, a s l/n = Rl/n s s (Koval'chenko, 1990). Therefore, at
equal stress, the lower the value of R1 n, the higher the amount of plastic strain induced
in the biomass bed. Lower values of Rl/n were obtained for the conditions of larger
particle size (i.e., grind obtained from the hammer mill screen size of 4.6 mm), lower
moisture content (i.e., about 10% w.b.), and lower temperature (i.e., 25C) (tables 7.3
and 7.4). These conditions may favor brittle failure of particles (i.e., fracturing of
original particles into smaller pieces). Increasing the stress, the strain hardening
exponent decreased from about 3 to 2 or lower (table 7.2). This showed that increasing
stress increased the plastic strain in the biomass compact.
The values for the viscous coefficient were negative because the dashpot
element modeled the time dependent energy loss during compression. That is, the
viscous coefficient represented the squeezing of natural binding components from the
biomass cells. For both corn stover and switchgrass, higher (absolute) values of
viscous coefficient predicted for smaller particle size (i.e., grind obtained from the
hammer mill screen size of 3.0 mm) and preheating to 75C indicate that these
conditions are favorable for expression of more binding components from the biomass
particles. Also, for the compression conditions that resulted in higher (absolute) values
of viscous coefficient, the compressive strength and durability of the briquettes were
higher (tables 7.3 and 7.4). For corn stover and switchgrass, lower (absolute) values of
viscous coefficient were predicted at about 20% (w.b.) moisture content than that at
about 10% (w.b.) moisture content. Similarly, for both of these biomass materials, at

331
the higher moisture content, compressive strengths of the briquettes were lower.
However, for corn stover, higher briquette durability was measured at about 20%
(w.b.) moisture content than that at about 10% (w.b.) moisture content. This may be
due to the enhanced activation of the water soluble carbohydrates (i.e., natural binders)
in the corn stover at the higher moisture content (Chapter 6; Kaliyan and Morey,
2007). This suggests that the constitutive model could not effectively account for the
curing or bonding effect due to the activation of natural binders. However, in general,
the densification conditions that resulted in higher (absolute) values of viscous
coefficient produced higher compressive strength of briquettes.
The frictional loss factor indicates the energy loss due to the frictional effects
such as inter-particle friction, particle-wall friction, and particle-wall adhesion. During
compression (i.e., loading), a simple force analysis would show that the applied force
is equal to the sum of the force used for densification of the powder bed (i.e., elastic,
plastic, and viscous deformations) and the force involved for the frictional effects
(Macleod and Marshall, 1977; Li et al., 1996). The force involved for the frictional
effects is usually related to or calculated from the measured radial force at the die-wall
(Macleod and Marshall, 1977). During loading, the radial force on the die-wall would
increase with an increase in applied force (i.e., accumulation of stress at the interface
of particle-die wall due to the reaction force from the die-wall), and during unloading,
the radial wall force would gradually reduce to zero (i.e., dissipation) or to a small
value which is stored at the interface of particle-die wall as the residual stress (Li et
al., 1996; Takeuchi et al., 2004). For both corn stover and switchgrass, the sign of the
frictional loss factor was negative until 25 MPa (tables 7.3 and 7.4). This suggests the
energy (or stress) involved for the frictional effects was dissipated during loading up
to 25 MPa. For pressures above 25 MPa, the frictional loss factor was positive for both
corn stover and switchgrass, and increased with increasing pressure. This indicates
that all or some of the energy (or stress) involved for the frictional effects was stored
at the interface of particle-die wall. No clear conclusions could be derived from the
effects of particle size, moisture content, and preheating temperature on the frictional
loss factor (tables 7.3 and 7.4). However, the stress involved for the frictional loss was
small compared to the stress involved for the deformation of the biomass bed.
332
Measuring the die-wall force may help understand the frictional loss effects. Also, it
appears that the frictional loss factor could be affected by adding lubricants (frictional
loss factor would be reduced) or binding agents (frictional loss factor would be
increased) to the powder mass.
Ren (1992) reported that if the behavior of the rheological elements could be
controlled in a desired way, the compression process and the quality of the densified
products may be improved. In this study, it was demonstrated that the elastic modulus
and viscous coefficient could be related to the durability and compressive strength of
the briquettes. In general, for the compression conditions that resulted in higher
(absolute) values of elastic modulus and viscous coefficient also resulted in higher
compressive strength and durability of the briquettes. Thus, conditioning of biomass
materials such that it would increase the (absolute) values of elastic modulus and
viscous coefficient may be a good strategy to increase the strength and durability of
the briquettes.
It should be noted that the constitutive model is capable of capturing the
particle level behaviors because the model parameters predicted for the two particle
sizes studied (i.e., grinds from hammer mill screen sizes of 3.0 and 4.6 mm) were
distinct (table 7.2). The maximum values of elastic modulus predicted for corn stover
and switchgrass were 930 and 734 MPa, respectively. The maximum (absolute) values
of viscous coefficient predicted for corn stover and switchgrass were 31 and 16 MPa-
s, respectively. The above comparison of elastic modulus and viscous coefficient
suggests that corn stover is easier to compress and would produce stronger and more
durable briquettes than switchgrass at similar densification conditions. The same
conclusion was derived, as mentioned at the start of the discussion (see above), from
the measured values of density, compressive strength, and durability of the corn stover
and switchgrass briquettes (table 7.1). Thus, the constitutive model would help to
understand the densification mechanisms as well as to find optimum densification
conditions to produce high quality pellets/briquettes.
Further study is required to determine the interaction of particle size, moisture
content, and preheating temperature on the constitutive model parameters. Also,
statistical relationships can be derived to estimate the model parameters as functions
333
of stress, particle size, moisture content, and temperature. Extension of the
predictability of the constitutive model in two- or three-dimensions and the model
parameters predicted as functions of densification variables would be required for
computer simulation of densification of corn stover and switchgrass.

7.6 Conclusions

The following conclusions were drawn from this study.

1. An elasto-visco-plastic solid model (i.e., constitutive model) was developed to


predict the mechanical behaviors of corn stover and switchgrass grinds during
compression/densification.

2. The constitutive model would help understand the structure development with
the applied pressure during compression of biomass grinds in a closed-end die.

3. The constitutive model parameters were found to be functions of stress,


particle size, moisture content, and temperature.

4. Elastic modulus and viscous coefficient of the constitutive model were found
to be correlated to the durability and compressive strength of biomass
briquettes. Thus, the constitutive model may help engineer the densification
process.

7.7 References

Abd-Elrahim, Y.M., A.S. Huzayyin, and I.S. Taha. 1981. Dimensional analysis and
wafering cotton stalks. Trans. ASAE 24(4): 829-832.
Adachi, T., and F. Oka. 1982. Constitutive equations for normally consolidated clay
based on elasto-viscoplasticity. Soils and Foundations 22(4): 57-70.
Adapa, P., G. Schoenau, L. Tabil, S. Sokhansanj, and A. Singh. 2005. Compression of
fractionated sun-cured and dehydrated alfalfa chops into cubes: pressure and
density models. Can. Biosys. Eng. 47(3): 3.33-3.39.
334
Antikainen, O., and J. Yliruusi. 2003. Determining the compression behavior of
pharmaceutical powders from the force-deformation profile. Int. J. Pharm.
252(1-2): 253-261.
ASAE Standards. 2003. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: AS ABE.
Aydin, I., B.J. Briscoe, and K.Y. Sanliturk. 1996. The internal form of compacted
ceramic components: a comparison of a finite element modeling with
experiment. Pow. Technol. 89(3): 239-254.
Aydin, I., B.J. Briscoe, and N. Ozkan. 1997. Modeling of powder compaction: a
review. MRS Bulletin 22(12): 45-51.
Bilanski, W.K., and V.A. Graham. 1984. A viscoelastic model for forage wafering.
Trans. CSME 8(2): 70-76.
Cunningham, J.C., I.C. Sinka, and A. Zavaliangos. 2004. Analysis of tablet
compaction. I. Characterization of mechanical behavior of powder and
powder/tooling friction. J. Pharm. Sci. 93(8): 2022-2039.
DiMaggio, F.L., and I.S. Sandler. 1971. Material model for granular soils. J. Eng.
Mechanics ASCE 96: 935-950.
DOE, 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The
Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.
Department of Energy (DOE), Office of Scientific and Technical Information.
Available at: http://www.osti.gov/bridge. Accessed 18 June 2005.
Drucker, D.C., and W. Prager. 1952. Soil mechanics and plastic analysis or limit
design. Quarterly of Applied Mathematics 10(2): 157-165.
Faborode, M.O., and J.R. O'Callaghan. 1986. Theoretical analysis of compression of
fibrous agricultural materials. J. Agric. Eng. Res. 35(3): 175-191.
Faborode, M.O., and J.R. O'Callaghan. 1989. A rheological model for the compaction
of fibrous agricultural materials. J. Agric. Eng. Res. 42(3): 165-178.

335
Ferrero, A., J. Horabik, and M. Molenda. 1991. Density-pressure relationships in
compaction of straw. Can. Agri. Eng. 33(1): 107-111.
Garekani, H.A., J.L. Ford, M.H. Rubinstein, and A.R. Rajabi-Siahboomi. 2000.
Highly compressible paracetamol - II. Compression properties. Int. J. Pharm.
208(1-2): 101-110.

Ghebre-Sellassie, I. 1989. Mechanism of pellet formation and growth. In


Pharmaceutical Pelletization Technology, 123-143.1. Ghebre-Sellassie, ed.
New York, NY: Marcel Dekker, Inc.
Graham, V.A., and W.K. Bilanski. 1984. Non-linear viscoelastic behavior during
forage wafering. Trans. ASAE 27(6): 1661-1665.
Haddad, Y.M. 1995. Viscoelasticity of Engineering Materials. New York, NY:
Chapman & Hall.

Heckel, R.W. 1961. An analysis of powder compaction phenomena. Trans.


Metallurgical Society ofAIME 221 (October): 1001 -1008.
Huang, L., and V.M. Puri. 2000. A three-dimensional finite element model for powder
compaction - time-dependent formulation and validation. Particulate Sci. and
Technol. 18(1): 257-274.
Kaliyan, N., and R.V. Morey. 2006. Densification characteristics of corn stover and
switchgrass. ASABE Paper No. 066174. St. Joseph, Mich.: ASABE.

Kaliyan, N., and R.V. Morey. 2007. Roll press briquetting of corn stover and
switchgrass: a pilot scale continuous briquetting study. ASABE Paper No.
076044. St. Joseph, Mich.: ASABE.
Koval'chenko, M.S. 1990. A rheological model of pressing of powders. Poroshkovaya
Metallurgiya 333(9): 100-104.
Li, Y., H. Liu, and A. Rockabrand. 1996. Wall friction and lubrication during
compaction of coal logs. Pow. Technol. 87(3): 259-267.
Lobosco, V., and V. Kaul. 2001. An elastic/viscoplastic model of the fiber network
stress in wet pressing: part 1. Nordic Pulp and Paper Res. J. 16(1): 12-17.

336
Macleod, H.M., and K. Marshall. 1977. The determination of density distribution in
ceramic compacts using autoradiography. Pow. Technol. 16(1): 107-122.
Macosko, C.W. 1994. Rheology: Principles, Measurements and Applications.
Poughkeepsie, NY: Wiley/VCH.

Malamataris, S., S.B. Baie, and N. Pilpel. 1984. Plasto-elasticity and tableting of
paracetamol, Avicel and other powders. J. Pharm. Pharmacol. 36: 616-617.
Mani, S., M. Roberge, L.G. Tabil, Jr., and S. Sokhansanj. 2003. Modeling of
densification of biomass grinds using discrete element method by PFC3D.
CSAE/SCGR 2003 Meeting Paper No. 03-207. Montreal, Quebec, CA:
CSAE/SCGR 2003 Meeting.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004. Evaluation of compaction equations
applied to four biomass species. Can. Biosys. Eng. 46(3): 3.55-3.61.

McLaughlin, S.B., and L.A. Kszos. 2005. Development of switchgrass {Panicum


virgatum) as a bioenergy feedstock in the United States. Biomass and
Bioenergy 28(6): 515-535.
Michrafy, A., D. Ringenbacher, and P. Tschoreloff. 2002. Modelling the compaction
behavior of powders: application to pharmaceutical powders. Pow. Technol.
127(3): 257-266.

Mittal, B., and V.M. Puri. 1999. Correlations between powder deposition methods and
green compact quality: Part II Compact quality and correlations. Particulate
Sci. and Technol. 17(4): 301-315.
Mittal, B., and V.M. Puri. 2002. Determination of visco-elastoplastic properties of a
powder using cubical triaxial tester. ASAE Meeting Paper No. 024035. St.
Joseph, Mich.: AS ABE.
Mohsenin, N., and J. Zaske. 1976. Stress relaxation and energy requirements in
compaction of unconsolidated materials. J. Agric. Eng. Res. 21(2): 193-205.
Mohsenin, N.N. 1986. Physical Properties of Plant and Animal Materials. New York,
NY: Gordon and Breach Science Publishers.

337
Munoz, G., and P. Herrera. 2002. Multidimensional modeling of agricultural fibrous
materials in densification: compression stage. ASAE Paper No. 023151. St.
Joseph, Mich.: AS ABE.
O'Dogherty, M.J. 1989. A review of the mechanical behavior of straw when
compressed to high densities. J. Agric. Eng. Res. 44: 241-265.
Panelli, R., and F.A. Filho. 2001. A study of a new phenomenological compacting
equation. Pow. Technol. 114(1-3): 255-261.
Peleg, K. 1983. A rheological model of nonlinear viscoplastic solids. J. Rheology
27(5): 411-431.
Peleg, M., and E.B. Bagley. 1983. Physical Properties of Foods. Westport,
Connecticut: AVI Publishing Company, Inc.
Pietsch, W. 2002. Agglomeration Processes: Phenomena, Technologies, Equipment.
Verlag GmbH, Weinheim: Wiley-VCH.
Raghavan, J.K., and H.N. Conkle. 1991. Physical characteristic measurements for
reconstituted coal pellets. Proc. of the Institute for Briquetting and
Agglomeration (IBA), Vol. 22. pp. 85-96.

Rehkugler, G.E., and W.F. Buchele. 1969. Biomechanics of forage wafering. Trans.
ASAE 12(1): 1-8.
Ren, S. 1992. Thermo-hygro rheological behavior of materials used in the
manufacture of wood-based composites. Ph.D. diss. Corvallis, OR: Oregon
State University.
Rippie, E.G., and W. Danielson. 1981. Viscoelastic stress/strain behavior of
pharmaceutical tablets: analysis during unloading and postcompression
periods. J. Pharm. Sci. 70(5): 476-482.
Rumpf, H. 1962. The strength of granules and agglomerates. In Agglomeration, 379-
419. W.A. Knepper, ed. New York, NY: John Wiley and Sons.
Schofield, A.N., and C.P. Wroth. 1968. Critical State Solid Mechanics. New York,
NY: McGraw Hill.

338
Sinka, I.C., J.C. Cunningham, and A. Zavaliangos. 2003. The effect of wall friction in
the compaction of pharmaceutical tablets with curved faces: a validation study
oftheDrucker-Prager Cap model. Pow. Technol. 133(1-3): 33-43.
Sinka, I.C., J.C. Cunningham, and A. Zavaliangos. 2004. Analysis of tablet
compaction. II. Finite element analysis of density distributions in convex
tablets. J. Pharm. Sci. 93(8): 2040-2053.
Sitkei, G. 1986. Mechanics of Agricultural Materials. New York, NY: Elsevier.
Tabil, L.G., and S. Sokhansanj. 1996. Compression and compaction behavior of
alfalfa grinds -1. Compression behavior. Powder Handling and Processing.
The Int. J. Storing, Handling, and Processing Powder 8(1): 17-23.
Takeuchi, H., S. Nagira, H. Yamamoto, and Y. Kawashima. 2004. Die wall pressure
measurement for evaluation of compaction property of pharmaceutical
materials. Int. J. Pharm. 274(1-2): 131-138.
Thoemen, H., C.R. Haselein, and P.E. Humphrey. 2006. Modeling the physical
processes relevant during hot pressing of wood-based composites. Part II.
Rheology. Holz als Roh- und Werkstofj'64(2): 125-133.

Tripodi, M.A., V.M. Puri, H.B. Manbeck, and G.L. Messing. 1992. Constitutive
models for cohesive particulate materials. J. Agric. Eng. Res. 53(1): 1-21.

Wu, C.Y., O.M. Ruddy, A.C. Bentham, B.C. Hancock, S.M. Best, and J.A. Elliot.
2005. Modelling the mechanical behavior of pharmaceutical powders during
compaction. Pow. Technol. 152(1-3): 107-117.
Zavaliangos, A. 2002. Constitutive models for the simulation of P/M processes. The
Int. J. Powder Metallurgy 38(2): 27-39.

339
CfQ

PRESSURE

H
13.

1/5

3
a>
o
C
-

<
O
-!

o
o

en
en
5'
3 (Plastic)

O Compacted
o
o ffl
en CO
+

o
m
en
H "D
O
p* 3
m (O
tv> CT
o 0)
o O
to
4

H-hi Tll-Ll

H- hi

t
a
Spring Dashpot Maxwell Material Kelvin- Voigt Material
Element Element

Figure 7.2. Mechanical elements used to describe stress-strain behavior of materials.


Spring element represents an elastic solid, and the dashpot element
represents a viscous fluid, (a = stress; E = modulus of elasticity; and r\ =
coefficient of viscosity.)

341
a = stress
I

Tl
8 = strain

ure 7.3. Mechanical analogy of biomass grind for the development of constitutive
model for the compression process. This mechanical analogy system
represents an elasto-visco-plastic solid. The dashpot viscous element (r\)
represents the time dependent loss, the strain hardening spring element (E,
R, and n) represents the elastic and plastic deformation, and the Coulomb
friction element (<jf) represents the time independent loss during uniaxial
compression of biomass grind.

342
200 30
Pressure (P)
25
150
Pressure gradient (dP/dt) 20
a. <D
TJ
15 2-
2? 100 CD D _
3
(/> 10 3 5
in </>

w
(U
50
+0
-5
50 100 150 200 250
Time (s)
A. Biomass compression curve from 0 to 150 MPa. The
compression curve was separated into inertial deformation stage,
and elasto-visco-plastic deformation stage.

0.2
1.5 Pressure (P)
Pressure gradient (dP/dt) w
Q.
CO 2
o.

3
in

50 100 150
Time (s)
B. Inertial deformation (i.e., air removal process) during biomass
compression. This part of the compression curve was not modeled.

Figure 7.4. Identification of inertial deformation and elasto-visco-plastic deformation


stages from the compression curve of corn stover or switchgrass grind.

343
30

1
Pressure (P)

152 -*- Pressure gradient (dP/dt) 20


TO
Q.

102
(/>
m
(D

51.5 infif-*l&*^^ '


10
f
1.5 -10
150 200 250
Time (s)
C. Elasto-visco-plastic deformation of biomass during
compression. This part of the compression curve was modeled.

Figure 7.4. Identification of inertial deformation and elasto-visco-plastic deformation


stages from the compression curve of corn stover or switchgrass grind.

344
A1. Stress range of 0 to 150 MPa. A2. Stress range of 1.1 to 5 MPa.
200
x Data

J
_ 150
is
Model
E
100
A End of inertial strain at 1.1 MPa
m 50

1 1.6
Natural strain

A3. Stress range of 5 to 25 MPa. A4. Stress range of 25 to 50 MPa.

1.9 2 2.2
Natural strain Natural strain

A5. Stress range of 50 to 100 MPa. A6. Stress range of 100 to 150 MPa.

2.3 2.35 2.4 2.42


Natural strain Natural strain

A. Corn stover (particle size = grind from hammer mill screen size of 3.0 mm; moisture content =
18.8% (w.b.); preheating temperature = 25C).

Figure 7.5. A comparison of constitutive model versus measured data. In addition to


the total compression curve for 0 to 150 MPa, the compression curves are
shown for different stress range segments. The model curve represents the
( H ^
Eq. 5 with the natural strain substituted as In , natural strain
{H-(Vxt)
rate substituted as , and the model parameters (E, R, n, r\,
[
H-(Vxt)
and Of) estimated for one replication data on the measured stress, natural
strain, and natural strain rate. [H = measured initial height of the grind at
time t = 0 (i.e., at the stress of 0 MPa) inside the die (m); V = the set speed
of the INSTRON crosshead (m s"1); t = elapsed time since the start of
compression (s)]
345
B 1 . Stress range of 0 to 150 MPa. B2. Stress range of 1.1 to 5 MPa.

u
200
x Data
x Data
5 Model
1 5 0
S ra
Model
a.
| 100 A End of inertial strain at 1.1 MPa i 3
in
M
50 2-

1 -
0.5 1 1.5 1 1.1 1.2
Natural strain Natural strain

B3. Stress range of 5 to 25 MPa. B4. Stress range of 25 to 50 MPa.

1.5 1.6 1.7


Natural strain Natural strain

B5. Stress range of 50 to 100 MPa. B6. Stress range of 100 to 150 MPa.
160
x Data
150
Model
140
130
120
110
100
2,05 2.1 2.1 2.14 2.16
Natural strain Natural strain

B. Switchgrass (particle size = grind from hammer mill screen size of 3.0 mm; moisture content =
19.9% (w.b.); preheating temperature = 25C).

Figure 7.5. A comparison of constitutive model versus measured data. In addition to


the total compression curve for 0 to 150 MPa, the compression curves are
shown for different stress range segments. The model curve represents the
Eq. 5 with the natural strain substituted as In
f H
, natural strain
H-(Vxt)

rate substituted as , and the model parameters (E, R, n, ri,


KH-(Vxt)J
and Gf) estimated for one replication data on the measured stress, natural
strain, and natural strain rate. [H = measured initial height of the grind at
time t = 0 (i.e., at the stress of 0 MPa) inside the die (m); V = the set speed
of the INSTRON crosshead (m s"1); t = elapsed time since the start of
compression (s)]
346
Table 7.1. Properties of corn stover and switchgrass briquettes.

Densification conditions Unit density of


Particle Moisture Preheating Height of Bulk the briquettes
size of the content of temperature of the grind density of immediately
grind the grind (% the grind (C) inside the the grind after ejection
(Hammer w.b.) cylinder inside the from the die
(N = 3)
mill (N = 3) (mm) cylinder (kg m 3 )
screen (N = 3) (kg m"3) (N = 3)
size in (N = 3)
mm)
Corn Stover
4.6 8.5 0.3 25.0 0.0 119.6 5.6 149.8 5.9 1074.8 7.6

3.0 8.7 0.02 25.0 0.0 112.5 5.7 158.3 8.8 1120.8 12.3
3.0 18.80.3 25.0 0.0 120.4 5.6 149.1 6.3 853.9 7.0

3.0 8.7 0.02 75.5 0.4 96.1 1.7 184.93.1 1169.9 14.8

Switchgrass
4.6 8.4 0.1 25.0 0.0 113.03.9 160.5 5.2 971.6 22.6

3.0 8.5 0.2 25.0 0.0 91.60.9 195.4 1.3 1006.6 0.5

3.0 19.9 0.4 25.0 0.0 89.2 1.2 201.3 2.9 573.7 6.4

3.0 8.5 0.2 75.1 0.1 87.3 2.8 204.1 6.3 1077.8 3.4

347
Table 7.1. Continued.

Densification conditions Data measured after one week of storage of briquettes at room temperature
Particle Moisture Preheating Briquette Briquette Moisture content Unit density of Mean expansion Diametrical Durability
size of the content of temperature of diameter (mm) height (mm) of the briquettes the briquettes in briquette compressive of briquettes
grind the grind the grind (C) (N = 3) (% w.b.) (kg m3) volume (%) strength of (%)
(Hammer (% w.b.) (N = 3) (N = 3) (N = 3) briquettes (N) (N = 2) *
mill screen (N = 2)
(N = 3) (N = 3)
size in
mm)
Corn Stover
4.6 8.5 0 . 3 25.0 0 . 0 19.2 0 . 0 1 18.2 0 . 2 8.5 0 . 0 2 939.7 1 1 . 5 14.2 411.8 4 6 . 9 61.6 2 . 9

3.0 8.7 0.02 25.0 0 . 0 19.2 0 . 0 4 17.2 0 . 3 7.3 0 . 0 1 992.7 1 8 . 5 12.2 502.8 3 8 . 8 75.2 0.6

3.0 18.8 0 . 3 25.0 0 . 0 19.3 0 . 0 3 23.8 0 . 2 16.4 0 . 0 695.5 4 . 0 19.7 287.0 1 0 . 8 95.5 0 . 7

3.0 8.7 0.02 75.5 0 . 4 19.1 0 . 0 3 14.9 0.2 6.7 0.2 1164.2 7 . 9 0.3 945.7 24.0 96.8 0.04

Switchgrass
4.6 8.4 0 . 1 25.0 0 . 0 19.3 0 . 0 4 21.3 0 . 4 8.3 0 . 3 804.8 17.7 20.3 177.6 1 7 . 0 0.0 0.0

3.0 8.5 0 . 2 25.0 0 . 0 19.2 0.01 19.8 0 . 1 7.8 0 . 2 863.0 6 . 3 15.6 235.2 3 . 5 0.0 0.0

3.0 19.9 0 . 4 25.0 0 . 0 19.6 0 . 1 38.1 0 . 6 16.7 0 . 4 418.8 7 . 8 32.2 18.1 1 1 . 4 0.0 0 . 0

3.0 8.5 0 . 2 75.1 0 . 1 19.1 0 . 0 3 16.3 0 . 2 7.3 0.04 1057.1 12.9 1.2 516.4 3 3 . 5 62.9 0 . 2

* Data on the durability of briquettes were measured by Kaliyan and Morey (2006) [Chapter 4].

348
Table 7.2. Constitutive model parameters for uniaxial compression of corn stover and switchgrass.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural Stress Elastic Strength Strain Viscous Frictional loss Sum of Mean square R!
applied strain at the range for modulus, E coefficient, R hardening coefficient, T\ factor, af (MPa) squared residuals
stress, a maximum which the (MPa) (MPa) exponent, n (MPa-s) residuals (MSR)
(MPa) applied model was (-) (SSR)
stress, E fitted
t) (MPa)
Biomass = Corn stover; Hammer mill screen size = 4.6 mm; Moisture content of the grind = 8.5 0.3% (w. p.); Preheating temperature of the grind = 25.0 0.0 C
1.597 0.988 0.0-1.597 * * * * * * * *
0.235 0.061
5.008 1.406 1.597- 4.927 10.130 2.3160.133 -0.161 0.052 -3.33 * 10"4 0.0210.006 1.0 * l O ^ i 1.00
0.005 0.032 5.008 0.661 0.377 0.006 0.000
0.00
25.038 1.989 5.008 - 16.071 55.692 3.0600.044 -0.6260.063 -0.0840.020 0.9530.121 0.001 0.001 1.00
0.015 0.040 25.038 0.446 3.768
0.00
50.190 2.201 25.038 - 79.684 i 350.037 2.408 0.068 -2.014 0.410 0.046 0.028 0.475 0.063 0.003 0.001 1.00
0.052 0.045 50.190 2.199 39.168
0.00
100.497 2.393 50.190- 181.013
181.013: 675.091 2.2690.060 -4.5220.603 0.1740.065 2.191 0.146 0.0210.002 1.00
0.170 0.048 100.497 6.915 97.278
0.00
152.771 2.509 100.497- 350.158 1190.129 2.0450.639 -8.006 1.184 0.4630.135 3.0760.266 0.0590.005 1.00
0.356 0.052 152.771 61.602 1080.222
0.00

349
Measured data Constitutive model parameters, mean standard deviation (N = 3)
Maximum Natural Stress Elastic Strength Strain Viscous Frictional loss Sum of Mean square R2
applied strain at the range for modulus, E coefficient, R hardening coefficient, TI factor, o r (MPa) squared residuals
stress, a maximum which the (MPa) (MPa) exponent, n (MPa-s) residuals (MSR)
(MPa) applied model was (-) (SSR)
stress, 8 fitted
t) (MPa)
Biomass = Corn stover; Hammer mill screen size = 3.0 mm; Moisture content of the grind = 8.7 0.02% (w. b.); Preheating temperature of the grind = 25.0 0.0 C
0.992 0.521 0.0-0.992 * * * * * * * *
0.134 0.038
5.005 0.870 0.992- 5.641 40.259 2.801 0.078 -0.2140.021 -0.0040.009 0.0420.017 1.0 * 10"4 1.00
0.005 0.025 5.005 0.885 10.479 0.000
0.00
25.049 1.171 5.005- 29.449 658.079 3.3890.027 -1.117 0131 -0.113 0.017 1.369 0.141 0.0020.001 1.00
0.033 0.031 25.049 0.905 100.029
0.00
50.193 1.263 25.049- 177.562 3852.611 2.5300.027 -5.0980.802 0.0050.038 0.4440.060 0.0030.001 1.00
0.065 0.033 50.193 8.434 513.415
0.00
100.329 1.335 50.193- 454.725 10079.660 2.4190.054 -12.618 0.1180.027 2.303 0.290 0.023 0.003 1.00
0.116 0.034 100.329 21.447 1870.663 2.055
0.00
152.909 1.375 100.329- 929.462 14271.940 l'.909 0.562 -22.727 0.1860.573 4.1930.716 0.0820.013 1.00
0.635 0.036 152.909 227.478 19664.880 4.274
0.00

350
Table 7.2. Continued.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural strain Stress range Elastic Strength Strain Viscous Frictional loss Sum of Mean square R2
applied at the for which the modulus, coefficient, R hardening coefficient, r| factor, crf (MPa) squared residuals
stress, cr maximum model was E (MPa) (MPa) exponent, n (-) (MPa-s) residuals (MSR)
(MPa) applied stress, fitted (MPa) (SSR)
EJ
Biomass = Corn stover; Hammer mill screen size = 3.0 mm; Moisture content of the grind = 18.8 0.3% (w. p.); Preheating temperature of the grind = 25.0 0.0 C
1.458 1.280 0.0-1.458 * * * * * * * *
0.516 0.124
5.006 1.727 1.458- 4.707 16.105 2.738 0.104 -0.174 -0.0070.004 0.0600.013 1.0 * lO^i 1.00
0.003 0.054 5.006 1.885 7.613 0.050 0.000
0.00
25.102 2.144 5.006- 19.563 246.175 3.4270.030 -1.096 -0.0880.007 1.0070.150 0.003 1.00
0.083 0.074 25.102 0.878 66.876 0.113 0.001
0.00
50.241 2.270 25.102- 133.767 1675.308 2.5370.048 -4.266 0.0070.033 0.6430.146 0.009 1.00
0.210 0.082 50.241 12.284 556.599 0.767 0.002
0.00
101.060 2.371 50.241- 326.881 7504.002 2.5990.073 -11.009 0.2300.056 4.3990.589 0.086 1.00
0.365 0.092 101.060 43.560 3410.388 0.916 0.013
0.00
8
153.337 2.426 101.060- 857.524 1.354* 10 3.3292.541 -13.174 0.0250.739 8.613 1.059 0.349 1.00
8
0.791 0.097 153.337 177.170 2.346 * 10 7.408 0.038
0.00

351
Table 7.2. Continued.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural strain Stress range Elastic Strength Strain Viscous Frictional loss Sum of Mean square R
applied at the for which the modulus, coefficient, R hardening coefficient, r\ factor, cf (MPa) squared residuals
stress, CT maximum model was E (MPa) (MPa) exponent, n (-) (MPa-s) residuals (MSR)
(MPa) applied stress, fitted (MPa) (SSR)
ejH
Biomass = Corn stover; Hammer mill screen size = 3.0 mm; Moisture content of the grind = 8.7 0.02% (w. b.); Preheating temperature of the grind = 75.5 0.4 C
0.963 0.496 0.0-0.963 * * * * * * * *
0.286 0.057
5.005 0.882 0.963 - 5.326 34.049 2.915 0.143 -0.216 -0.0I80.015 0.0810.074 1.0 * 10"4 1.00
0.001 0.043 5.005 1.189 9.982 0.025 0.000
0.00
25.031 1.192 5.005- 28.307 675.212 3.4690.016 -1.623 -0.0170.004 1.2420.057 0.002 1.00
0.013 0.068 25.031 2.042 170.161 0.126 0.001
0.00
0.107 1.284 25.031- 178.822 3722.818 2.5270.019 -6.464 0.011 0.042 0.5280.049 0.005 1.00
0.058 0.076 50.107 16.022 893.433 0.798 0.001
0.00
)0.305 1.358 50.107- 450.178 10664.870 2.4570.078 -15.085 0.1240.142 2.5030.409 0.029 1.00
0.154 0.083 100.305 43.326 3169.080 2.040 0.005
0.00
153.217 1.399 100.305- 846.065 3386.025 1.621 0.101 -30.488 0.6660.056 4.7350.207 0.111 1.00
0.782 0.088 153.217 165.681 866.151 5.821 0.005
0.00

352
Table 7.2. Continued.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural strain Stress range Elastic Strength Strain Viscous Frictional loss Sum of Mean square R2
applied at the for which the modulus, coefficient, R hardening coefficient, r| factor, crf (MPa) squared residuals
stress, a maximum model was E (MPa) (MPa) exponent, n (-) (MPa-s) residuals (MSR)
(MPa) applied stress, fitted (MPa) (SSR)
m
Biomass = Switchgrass; Hammer mill screen size = 4.6 mm; Moisture content of the grind = 8.4 0.1 % (w. b.); Preheating temperature of the grind = 25.0 0.0 C
1.215 0.821 0.0-1.215 * * * * * * * *
0.361 0.121
5.007 1.299 1.215- 4.131 10.800 2.414 0.224 -0.148 0.001 0.011 0.0300.017 1.0><10"4 1.00
0.006 0.049 5.007 0.947 0.729 0.007 0.000
0.00
25.062 1.835 5.007- 17.328 73.454 3.068 0.012 -0.750 -0.074 0.007 0.855 0.045 0.001 1.00
0.065 0.068 25.062 0.312 8.859 0.129 0.001
0.00
50.113 2.029 25.062- 87.356 425.905 2.4040.085 -2.363 0.041 0.031 0.421 0.037 0.003 1.00
0.089 0.078 50.113 5.440 103.614 0.474 0.000
0.00
100.229 2.209 50.113- 192.329 676.144 2.1870.068 -5.074 0.051 0.090 2.0050.229 0.018 1.00
0.220 0.089 100.229 11.810 185.073 0.570 0.002
0.00
152.753 2.322 100.229- 365.057 588.629 1.808 0.149 -7.984 0.253 0.112 3.502 0.337 0.060 1.00
0.315 0.099 152.753 13.614 50.943 1.862 0.006
0.00

353
Table 7.2. Continued.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural strain Stress range Elastic Strength Strain Viscous Frictional loss Sum of Mean square R2
applied at the for which the modulus, coefficient, R hardening coefficient, r\ factor, crf (MPa) squared residuals
stress, a maximum model was E (MPa) (MPa) exponent, n (-) (MPa-s) residuals (MSR)
(MPa) applied stress, fitted (MPa) (SSR)
H
Biomass = Switchgrass; Hammer mill screen size = 3.0 mm; Moisture content of the grind = 8.5 0.2% (w. b.); Preheating temperature of the grind = 25.0 0.0 C
1.045 0.547 0.0-1.045 * * * * * * * *
0.206 0.107
5.005 0.960 1.045- 4.698 18.023 2.473 0.216 -0.203 -0.004 0.003 0.050 0.017 1.0 x 10'4 1.00
0.001 0.096 5.005 0.246 4.791 0.045 0.000
0.00
25.047 1.370 5.005- 22.628 203.148 3.2080.039 -1.074 -0.0950.007 1.041 0.130 0.002 1.00
0.049 0.144 25.047 2.305 90.622 0.180 0.000
0.00
50.140 1.510 25.047- 119.793 1100.096 2.4160.074 -3.841 0.0520.086 0.4560.031 0.003 1.00
0.086 0.163 50.140 18.694 555.979 1.318 0.001
0.00
100.501 1.632 50.140- 281.414 2317.134 2.2900.052 -7.831 0.1080.020 2.1620.300 0.021 1.00
0.190 0.184 100.501 50.162 1224.441 1.797 0.003
0.00
152.760 1.702 100.501- 630.978 4969.942 2.3460.194 -14.405 0.3260.339 4.675 1.209 0.089 1.00
0.434 0.196 152.750 108.190 2056.386 4.773 0.022
0.00

354
Table 7.2. Continued.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural strain Stress range Elastic Strength Strain Viscous Frictional loss Sum of Mean square R
applied at the for which the modulus, coefficient, R hardening coefficient, r) factor, ay (MPa) squared residuals
stress, or maximum model was E (MPa) (MPa) exponent, n (-) (MPa-s) residuals (MSR)
(MPa) applied stress, fitted (MPa) (SSR)
IB
Biomass = Switchgrass; Hammer mill screen size = 3.0 mm; Moisture content of the grind = 19.9 0.4% (w. b.); Preheating temperature of the grind = 25.0 0.0 C
1.420 0.956 0.0-1.420 * * * * * * * *
0.335 0.091
5.015 1.393 1.420- 4.552 13.806 2.585 0.220 -0.187 -0.005 0.004 0.035 0.018 1.0 * 10"4 1.00
0.002 0.010 5.015 0.755 1.909 0.016 0.000
0.00
25.078 1.854 5.015- 18.880 141.725 3.250 0.031 -1.010 -0.081 0.012 0.742 0.037 0.002 1.00
0.049 0.029 25.078 0.815 24.698 0.093 0.000
0.00
50.174 2.008 25.078- 110.162 828.934 2.4560.042 -3.590 0.0540.040 0.5920.052 0.007 1.00
0.042 0.037 50.174 7.071 117.360 0.180 0.001
0.00
101.094 2.137 50.174- 268.830 5466.422 2.743 0.312 -7.482 0.246 0.232 3.358 0.374 0.054 1.00
0.263 0.044 101.094 23.628 3718.655 1.640 0.005
0.00
153.426 2.202 101.094- 639.772 4909.126 2.0950.493 -15.831 0.9350.486 7.692 1.832 0.292 1.00
0.327 0.048 153.426 89.690 4761.876 6.302 0.070
0.00

355
Table 7.2. Continued.

Measured data Constitutive model parameters, mean standard deviation (N = 3)


Maximum Natural strain Stress range Elastic Strength Strain Viscous Frictional loss Sum of Mean square R2
applied at the for which the modulus, coefficient, R hardening coefficient, r\ factor, <jf (MPa) squared residuals
stress, CT maximum model was E(MPa) (MPa) exponent, n (-) (MPa-s) residuals (MSR)
(MPa) applied stress, fitted (MPa) (SSR)
eH
Biomass = Switchgrass; Hammer mill screen size = 3.0 mm; Moisture content of the grind = 8.5 0.2% (w. b.); Preheating temperature of the grind = 75.1 0.1 C
0.722 0.471 0.0-0.722 * * * * * * * *
0.058 0.082
5.003 0.961 0.722 - 3.905 20.926 2.855 0.123 -0.202 -0.011 0.004 0.075 0 . 0 2 7 1.0 * 10"4 1.00
0.003 0.169 5.003 0.669 9.489 0.019 0.000
0.00
25.066 1.336 5.003 - 24.466 378.027 3.271 0.125 -1.343
1.343 -0.090 0 . 0 1 9 0.975 0.222 0.002 LOO
0.062 0.265 25.066 4.887 237.960 0.332 0.000
0.00
50.281 1.461 25.066- 142.884 1986.556 2.458 0.080 -4.620 0.015 0.039 0.516 0.013 0.004 1.00
0.051 0.308 50.281 40.098 1264.146 1.549 0.001
0.00
100.335 1.568 50.281- 340.032 4182.072 2.367 0 . 0 8 8 -11.008 0.082 0 . 1 3 0 2.326 0 . 4 1 7 0.026 1.00
0.253 0.349 100.335 99.121 2619.910 3.314 0.005
0.00
153.043 1.631 100.335- 733.606 6327.183 2.141 0.309 -15.365 0.2820.283 5.5430.738 0.122 1.00
0.594 0.376 153.043 239.115 6720.910 5.204 0.019
0.00

Only inertial deformation of the biomass grind (i.e., no elasto-visco-plastic deformation) takes place during the corresponding stress range.

356
Table 7.3. Effects of densification conditions on constitutive model parameters for corn stover. [The model parameters (mean of three
replications) are plotted as a function of maximum stress involved for estimating the model parameters at various stress
ranges.]

Effect of particle size (hammer mill screen Effect of moisture content Effect of preheating temperature (C)
size used for grinding, mm) (% w.b.) (Hammer mill screen size = 3.0 mm; moisture
(Moisture content = 8.5 to 8.7% w.b.; (Hammer mill screen size = 3.0 mm; preheating content = 8.5 to 8.7% w.b.)
preheating temperature = 25C) temperature = 25C)
Elastic modulus (E)

3 600
3 -4.6 mm -8.7%
"O
- 3 . 0 mm1 -18.8%
I 400
u
1 200
uj

0
50 100 150 200
Stress (MPa)

Strength coefficient (R) to the power of (1/n). (n is the strain hardening exponent)

- 4 . 6 mm |
-3.0 mm

100 150
Stress (MPa)

357

/
Table 7.3. Continued.

Effect of particle size (hammer mill screen Effect of moisture content Effect of preheating temperature (C)
size used for grinding, mm) (% w.b.) (Hammer mill screen size = 3.0 mm; moisture
(Moisture content = 8.5 to 8.7% w.b.; (Hammer mill screen size = 3.0 mm; preheating content = 8.5 to 8.7% w.b.)
preheating temperature = 25C) temperature = 25C)
Strain hardening exponent (n)

-4.6 mm
-3.0 mm

100
Stress (MPa)

Viscous coefficient (y\)

*
1-10
I -4.6 mm -8.7%
I8 - -3.0 mm -18.8%

I -30

100 50 100 150 200


Stress (MPa) Stress (MPa)

358
Table 7.3 Continued.
Effect of particle size (hammer mill screen Effect of moisture content Effect of preheating temperature (C)
size used for grinding, mm) (% w.b.) (Hammer mill screen size = 3.0 mm; moisture
(Moisture content = 8.5 to 8.7% w.b.; (Hammer mill screen size = 3.0 nun; preheating content = 8.5 to 8.7% w.b.)
preheating temperature = 25C) temperature = 25C)
Frictional loss factor (<Xf)

-4.6 mm -8.7%
-3.0 mm -18.8%

100 100
Stress (MPa) Stress (MPa)

Durability (%), and diametrical compressive strength (N) of briquettes


100

80

60
D 4.6 mm 8.7%
D 3.0 mm D18.8%

Durability (%) Compressive strength/10 Durability PQ Compressive strength/10 Durability (%) Compressive strength/10
(N) (N) (N)

359
Table 7.4. Effects of densification conditions on constitutive model parameters for switchgrass. [The model parameters (mean of three
replications) are plotted as a function of maximum stress involved for estimating the model parameters at various stress
ranges.]

Effect of particle size (hammer mill screen Effect of moisture content Effect of preheating temperature (C)
size used for grinding, mm) (% w.b.) (Hammer mill screen size = 3.0 mm; moisture
(Moisture content = 8.4 to 8.5% w.b.; (Hammer mill screen size = 3.0 mm; preheating content = 8.4 to 8.5% w.b.)
preheating temperature = 25C) temperature = 25C)
Elastic modulus (E)

= 600
-4.6mm; -8.5%
- 3 . 0 mm -19.9%

50 100 100 150


Stress (MPa) Stress (MPa)

Strength coefficient (R) to the power of (1/n). (n is the strain hardening exponent)
300 | 300

wer
JM S. 250 s. P50

the
the

a . 200
i 200
-4.6 mm a 2 -8.5% $ .
150 .5 C 150 160
f(1/n
cient

- 3 . 0 mm I) . -19-9%

of (1/n
100 I 100 100 3
s Ii 50

100
Stress (MPa)

360
Table 7.4. Continued.

Effect of particle size (hammer mill screen Effect of moisture content Effect of preheating temperature (C)
size used for grinding, mm) (% w.b.) (Hammer mill screen size = 3.0 nun; moisture
(Moisture content = 8.4 to 8.5% w.b.; (Hammer mill screen size = 3.0 mm; preheating content = 8.4 to 8.5% w.b.)
preheating temperature = 25C) temperature = 25C)
Strain hardening exponent (n)

-4.6 mm
- 3 . 0 mm

Viscous coefficient (T|)

- 4 . 6 mm i
i -20 - 3 . 0 mm
8
W
g -30

100
Stress (MPa)

361
Table 7.4. Continued.

Effect of particle size (hammer mill screen Effect of moisture content Effect of preheating temperature (C)
size used for grinding, mm) (% w.b.) (Hammer mill screen size = 3.0 mm; moisture
(Moisture content = 8.4 to 8.5% w.b.; (Hammer mill screen size = 3.0 mm; preheating content = 8.4 to 8.5% w.b.)
preheating temperature = 25C) temperature = 25C)
Frictional loss factor (<yf)

-4.6 mm -8.5%
-3.0 mm -19.9%

Stress (MPa)

Durability (%), and diametrical compressive strength (N) of briquettes


100 100

30 80
75.1C 75.1C
a 4.6 mm
8 60 8.5% a 25.0oC
a 3.0 mm O 19.9% S75.1C
5 40
25.0C
20
4.6 mm 3-0 mm 25.0'C

Durability (%) Compressive strength/10 Durability (%) Compressive strength/10 Durability (%) Compressive strength/10
(N) (N)

362
7.8 Appendix

Table 7.A1. Example calculation showing estimation of model parameters for the
measured compression data in the stress range of 100 to 150 MPa.
Measured data * Measured data involved for model parameter estimation
using Eq. 4a **
Time Stress Natural (CT - CTP) (e - ep)
(t,s) (a, MPa) strain (e)
dt (WH)
152.750 100.624 1.356 0 0 0
152.800 101.511 1.357 0.886 0.001 0.023
152.850 102.333 1.358 1.708 0.002 0.023
152.900 103.067 1.359 2.442 0.003 0.023
152.950 103.993 1.361 3.368 0.005 0.023
153.000 104.954 1.362 4.329 0.006 0.023
153.050 105.898 1.362 5.274 0.006 0.000

(Note that the intermediate data were cut intentionally to save space)

154.875 145.688 1.398 45.063 0.042 0.024


154.925 146.941 1.399 46.316 0.043 0.024
154.975 148.021 1.400 47.397 0.044 0.024
155.025 149.326 1.401 48.701 0.045 0.024
155.085 150.565 1.402 49.941 0.046 0.020
155.100 150.991 1.402 50.366 0.046 0.000
155.150 152.261 1.402 51.637 0.046 0.000
155.195 153.304 1.404 52.679 0.048 0.026

The stress and natural strain should be calculated for the entire data set starting from 0 to 150
MPa before fitting the model parameter for individual ranges of stress data. The stress ranges
selected for piecewise nonlinear regressions were ~1 to 5, 5 to 25, 25 to 50, 50 to 100, and 100
to 150 MPa.

Measured force (N)


Measured stress (a, Pa) = (la)
Cross sectional area of the densification cylinder bore (m )

H
Measured natural strain at time t (e) - In (2a)
H-AH

Where, H = initial height of the grind at time t = 0 (i.e., at the stress of 0 MPa) inside the die,
and AH = total change in the grind height due to compression until time t (i.e., at the stress of CT
MPa at time t). Also, AH is the same as the total displacement of the crosshead of the Instron
recorded (i.e., measured) until time t. Note that compressive stress and strain were considered
to be positive.

The proposed constitutive model is:

363
ds
a = EE + RE" +n + erf (3a)
f
dt

To estimate the constitutive model parameters for a selected range of measured stress
and natural strain data, the constitutive model would be modified as:

d{s-s )
(a~ap) = E(-ep) + R(e-py +ij +07 (4a)

Where, a p and sp are the measured stress and measured natural strain values corresponding to
the first data points in the selected stress range (i.e., break points or knots for the piecewise
nonlinear regression). For instance, for the stress range of 100 to 150 MPa, a p is 100 MPa and
ep is the natural strain corresponding to 100 MPa. For the example calculation (table 7.A1), a p
= 100.624 MPa, and ep = 1.356. Note that the a p and ep data would be included while
estimating the model parameters for the previous stress range (e.g., 50 to 100 MPa for the
example calculation given in table 7.A1).

This modification for the constitutive model was done assuming that the compacted
material inside the die would have zero stress and zero strain before applying the incremental
stress in the selected range of the stress data for which the model parameters are to be
estimated. Also, the stress and strain "history" would be assumed to be involved for making
the compacted structure. Thus, for each subsequent stress range data used for the model
parameter estimation, the mechanical behavior of the compacted material would be different.

For each stress range, the data on (a - a p ), (s - sp), and d(s - ep)/dt were fitted to
equation 4a using SPSS software to estimate the model parameters.

The fitted model parameters for the example calculation in table 7.Alare:
E = 815.931 MPa, R = 3227.71 MPa, n = 1.794, r] = -21.301 MPa-s, and a f = 0.225 MPa, with
sum of squares of residuals (SSR) = 6.172, mean square residuals (MSR) = 0.140, and R2 =
0.999.

364
Chapter 8
SUMMARY AND OVERALL CONCLUSIONS

In this dissertation research, densification mechanisms of corn stover and


switchgrass were studied by: experiments conducted using a laboratory scale
densification apparatus; experiments conducted using pilot-scale ring-die pelleting and
roll press briquetting machines; micro-structural analyses of grinds, pellets and
briquettes; and modeling of the stress-strain behaviors of the compression process.

In the laboratory scale densification studies, densification characteristics of


corn stover and switchgrass were studied using a uniaxial, piston-cylinder
densification apparatus. The effects of compression pressure (100 to 150 MPa),
moisture content (10 to 20% w.b.), particle size (0.56 to 0.80 mm), and preheating
temperature (25 to 150C) on the densification characteristics (density, durability,
stability, and specific energy consumption) of corn stover and switchgrass were
studied. Also, experiments were conducted to improve the durability of switchgrass
briquettes by fine grinding of switchgrass (particle size of 0.38 and 0.26 mm), adding
biomass based binders (corn stover, corn Distillers Dried Grains with Solubles
(DDGS), and starch), and mixing commercial chemical feed binders (lignin-sulfonate,
lime, and sodium bentonite) at the forming pressure of 150 MPa and preheating
temperature of 100C.

In the pilot-scale densification studies, experiments on briquetting of corn


stover and switchgrass were conducted in a pilot-scale roll press briquetting machine.
The effects of briquetting machine variables such as screw feeder (1 to 70 rpm) and
roll speeds (1 to 15 rpm), and biomass variables such as particle size (0.34 to 0.59
mm), moisture content (7 to 20% w.b.), and steam conditioning temperature (20 to
100C) on the bulk density, durability, and crushing strength of the briquettes were
investigated. Also, limited experiments were conducted using a pilot-scale ring-die

365
pelleting machine to compare the densification performance with the roll press
briquetting machine.

For characterizing the compression behaviors of corn stover and switchgrass


grinds, a constitutive model referred to as an elasto-visco-plastic solid model was
developed. The constitutive model characterized the compression behaviors of the
biomass grinds through five parameters: elastic modulus, strength coefficient, strain
hardening exponent, viscous coefficient, and frictional loss factor.

The following conclusions were drawn from the laboratory scale densification
studies:

A compression pressure of at least 100 MPa was required to produce


durable briquettes from corn stover (i.e., briquette durability of > 50%).

For briquetting under room temperature (about 25C), corn stover


briquettes (about 19.2 mm diameter) with relaxed densities of 825 to 1013
kg m"3 and durabilities of 50 to 96% can be made at pressures of 100 to 150
MPa, geometric mean particle sizes of grind from 0.66 to 0.80 mm, and
moisture contents of 10 to 20% (w.b.). For briquetting at elevated
temperatures of 75 to 150C and at 150 MPa, corn stover briquettes with
relaxed densities of 985 to 1162 kg m"3 and durabilities of 92 to 97% can
be produced. Specific energy required for briquetting corn stover ranged
from 0.12 to 0.22%), and 0.63 to 1.48%) of the energy content of the corn
stover for briquetting at room temperature, and at elevated temperatures of
75 to 150C, respectively.

For briquetting under room temperature (about 25C), switchgrass


briquettes (about 19.4 mm diameter) with relaxed densities of 417 to 825
kg m" can be produced at pressures of 100 to 150 MPa, geometric mean

366
particle sizes of grind from 0.56 to 0.64 mm, and moisture contents of 10 to
20% (w.b.). However, durability of siwtchgrass briquettes made under
room temperature was zero percent. For briquetting at elevated
temperatures of 75 to 150C and at 150 MPa, switchgrass briquettes with
relaxed densities of 834 to 1065 kg m" and durabilities of 55 to 67% can
be produced. Specific energy required for briquetting switchgrass ranged
from 0.15 to 0.24%, and 0.74 to 1.72% of the energy content of the
switchgrass for briquetting at room temperature, and at elevated
temperatures of 75 to 150C, respectively.

With preheating to 100C at 150 MPa pressure, either fine grinding of


switchgrass to a particle size of 0.26 mm, or addition of 5% (wt.) lime
powder, or 2% (wt.) chemical binders derived from lignin-sulfonate, or
mixing 20% (wt.) corn stover increased the briquette durability by about 10
percentage points compared to the densification condition of preheating to
100C without addition of any binders. Although this is a modest increase
in briquette durability, mixing of 20% (wt.) corn stover with 80% (wt.)
switchgrass appears to be the best strategy to improve the switchgrass
briquette durability because addition of 20% (wt.) corn stover would
involve almost no additional cost for densification compared to the
strategies of fine grinding of switchgrass to a particle size of 0.26 mm, and
addition of 2 to 5% (wt.) of chemical binders.

It can be concluded that at moisture contents of 15 to 20% (w.b.), high


quality corn stover briquettes can be produced without requiring
preheating. With a moisture content of at least 10% (w.b.), preheating to
>75C is required to produce high quality switchgrass briquettes. The glass
transition temperatures of both corn stover and switchgrass measured using
a differential scanning calorimeter averaged about 75 C.

367
The following conclusions were deduced from the pilot-scale roll press
briquetting and pelleting studies:

The briquettes made were of almond shape with a maximum size of 31.3
mm length x 23.3 mm width x 17.9 mm depth, and the briquettes had a
maximum of 5.0-mm flashings. The pellets made were of cylindrical shape
with a maximum size of 9.8 mm diameter x 24.0 mm length. The bulk
densities of corn stover and switchgrass briquettes ranged from 1.8 to 5.3
times the bulk density of bales (bulk density of bales is 100 to 200 kg m"3).
The bulk densities of corn stover and switchgrass pellets ranged from 2.8 to
6.2 times the bulk density of bales.

Briquettes produced with the roll press briquetting machine had bulk
densities, durabilities, and crushing strengths that were somewhat less than
but in a range comparable to the pellets produced with the conventional
pelleting machine. The specific energy consumption for roll press
briquetting and pelleting ranged from 1 to 4% of the energy in the corn
stover and switchgrass.

Roll press briquetting appears to be a promising low-cost, low-energy,


high-capacity approach for densifying corn stover and switchgrass for
renewable energy applications.

It was shown that high quality briquettes and pellets could be produced
from corn stover and switchgrass without adding chemical binders (i.e.,
additives) by activating (softening) the natural binding components such as
water soluble carbohydrates, lignin, protein, starch, and fat in the corn
stover and switchgrass through moisture and temperature. Micrographs
obtained using scanning electron microscopy for grinds, and cross sections
of the pellets and briquettes supported that the expressed natural binders

368
would help make strong bonding between biomass particles. If the
temperature rise of the biomass grinds due to shear/frictional heating in the
briquetting and pelleting machines is in the range of glass transition (i.e.,
softening) temperature of the biomass materials (i.e., >75C), strong and
durable briquettes and pellets could be produced without steam
conditioning.

The following conclusions were obtained from the constitutive modeling of the
compression behaviors of corn stover and switchgrass:

An elasto-visco-plastic solid model (i.e., constitutive model) could be used


to describe and predict the mechanical behaviors of corn stover and
switchgrass grinds during compression/densification.

The constitutive model parameters were found to be functions of stress,


particle size, moisture content, and temperature. The constitutive model
would help understand the structure development with the applied pressure
during compression of biomass grinds in a closed-end die.

Elastic modulus and viscous coefficient of the constitutive model were


found to be correlated to the durability and compressive strength of
biomass briquettes. Therefore, the constitutive model could help optimize
the densification process variables such as particle size, moisture content,
and preheating temperature.

The results from this dissertation research contribute to the production of high
density, strong, and durable densified products such as pellets, briquettes, or
compacted-granulated materials from corn stover and switchgrass using existing
commercial scale densification machines in the U.S. such as ring-die pelleting and roll
press briquetting/compaction machines.

369
Chapter 9
RECOMMENDATIONS FOR FUTURE WORK

The following recommendations are made for the future work.

1. Modification of the laboratory scale uniaxial, piston-cylinder densification


apparatus is required to control moisture movement from the biomass grinds
during preheating of corn stover and switchgrass when preheating temperatures
are 100-150C or higher. One suggestion would be adding a steam line to
preheat the biomass grinds instead of using conduction-based electrical heating
system (i.e., heating tapes).
2. Determine the upper limit on the particle size (i.e., hammer mill screen size
used for grinding) of corn stover and switchgrass which will allow for the
production of high quality (density and durability) briquettes. Producing high
quality briquettes with larger particle size can reduce the cost of densification.
3. The corn stover used for the current densification studies consisted mostly of
stalks, leaves, and husks with a small amount of cobs. Therefore, densification
of corn cobs may be studied in the future. Also, densification characteristics of
individual plant components (e.g., stems, stalks, or leaves alone) or mixtures of
different percentages of various plant components (e.g., a mixture of leaves,
husks and cobs excluding stalks from the corn stover) may be studied in the
future.
4. The effect of variety of the corn stover and switchgrass on their densification
characteristics may be studied. The amount of natural binders in corn stover or
switchgrass may vary between varieties, and thus, the varietal differences may
alter the optimum densification conditions required for producing high quality
briquettes and pellets.
5. In this research, only limited experiments were conducted on pelleting of corn
stover and switchgrass using a pilot-scale pelleting machine. Extended studies
on pelleting are required for determining the energy and cost involved for

370
making pellets with different sizes (diameters) and at various densification
conditions such as die speed, die length-to-diameter ratio, and capacity of the
pelleting machine.
6. A few commercial scale trials on pelleting and roll press briquetting are
required to determine the actual energy and cost involved for the densification
of corn stover and switchgrass.
7. In this research, one-dimensional constitutive models for the compression
process of corn stover and switchgrass were developed using laboratory
densification data. Constitutive models may be extended to two- or three-
dimensional cases by collecting compression stress-strain data in two- or three-
dimensions of the laboratory scale densification apparatus. Furthermore, the
constitutive modeling concepts presented in this thesis are useful to model the
compression behaviors of biomass consisting of either individual plant
components or mixture of plant components.
8. Computer simulation of densification of corn stover and switchgrass involving
their constitutive models would help to design energy efficient densification
machines or to optimize existing densification machines.
9. Moisture sorption/desorption studies on corn stover and switchgrass grinds are
required at temperatures used for steam conditioning process (up to 100C).
This would help understand and optimize the preheating/steam conditioning
process. During the pilot-scale roll press briquetting experiments, it was
observed that switchgrass lost moisture faster than corn stover when steam
conditioned to 100C or higher.
10. Drying or cooling behaviors of the "hot" corn stover and switchgrass pellets
and briquettes should be studied. Optimizing the drying or cooling systems is
essential to avoid loss of strength and durability of pellets and briquettes due to
improper drying or cooling process.
11. The effect of relative humidity on the strength and durability of corn stover and
switchgrass pellets and briquettes needs to be studied. High relative humidity

371
may negatively affect the quality of biomass pellets and briquettes during
transportation and storage.
12. Densification behaviors of corn stover or switchgrass mixed with coal powders
may be studied if these biomass materials are considered for co-firing with
coal.
13. Densification characteristics of corn stover and switchgrass with a pre-process
(such as torrefaction to increase the energy density, or adding chemicals to
solve the emission or ash-fusion problem) to upgrade the pellet/briquette fuel
quality may be studied.
14. Granulation (forming clumps of compacted particles not as large as pellets or
briquettes) of corn stover and switchgrass using roll press compaction
machines may be studied in the future. Compaction, granulation of corn stover
and switchgrass may result in densified products similar to pellets with less
specific energy consumption for densification.
15. Studies on the mechanical/pneumatic handling characteristics of pellets and
briquettes made from corn stover and switchgrass would be useful.
16. Standards on the acceptable levels on the strength and durability values of the
biomass pellets and briquettes need to be established.
17. The re-grinding characteristics and the specific energy required for re-grinding
of biomass pellets and briquettes may be determined in the future.
18. Studies on the cost of producing briquettes using roll press briquetting
machines or pellets using pelleting machines are required. The densification
cost data could be incorporated with the cost estimation for supplying corn
stover and switchgrass from the field to a conversion facility in a systems
approach.

372
Appendix A
EFFECT OF ENHANCED DEAERATION DURING DENSIFICATION OF CORN STOVER

AND SWITCHGRASS - LABORATORY SCALE STUDY

A.l Abstract

Densification characteristics of the powders are usually studied in the


laboratory using a closed-end, piston-cylinder densification apparatus. During
compression of powders in this type of apparatus, some amount of air can be
entrapped in the powder mass due to inadequate deaeration, which may result in high
expansion of pellets/briquettes after ejection from the cylinder. In this study, a closed-
end, piston-cylinder densification apparatus was modified to include a steel filter at the
bottom of the cylinder to increase the open area available for deaeration. This
modified densification apparatus was used to study the effect of enhanced deaeration
during densification of corn stover and switchgrass at a compression pressure of 150
MPa, compression speed of 25.4 mm min"1, and temperatures of 25 and 75C. The
modified densification apparatus (12% open area for deaeration) resulted in almost no
changes in the density and durability of the briquettes, but reduced the volume
expansion of briquettes than for the conventional closed-end, piston-cylinder
densification apparatus (0.11% open area for deaeration). It is concluded that effective
deaeration would improve the densification behaviors of the biomass materials in
commercial densification equipment. However, for laboratory densification studies
using a closed-end, piston-cylinder densification apparatus, the open area available for
deaeration should be > 0.11% of the cross-sectional area of the cylinder-bore and the
speed of compression should be < 25.4 mm min"1.

A.2 Introduction

Corn stover and switchgrass are two important biomass feedstocks in the U.S.
that have been considered for the production of liquid transportation fuels (ethanol),

373
chemicals, bio-products, process heat, and electricity to reduce/replace the use of
fossil fuels aiming mainly to curb the emission of greenhouse and acid gases (DOE,
2005; ORNL, 2005). Currently, corn stover and switchgrass are collected from the
field as round or square bales with bulk density of 100 to 200 kg m"3. Due to the low
bulk density of baled biomass materials, direct use of baled biomass materials is
limited. In order to use these biomass materials in a wide range of renewable energy
applications, these biomass materials should be pre-processed to obtain products of
uniform size and shapes with high density through pre-processing such as
densification. Therefore, studying the densification characteristics of corn stover and
switchgrass are important to produce good quality densified products such as pellets
and briquettes from these biomass materials.
Densification characteristics of powders are usually studied in the laboratory
using closed-end (i.e., single-acting or uniaxial), piston-cylinder densification
apparatus. In this type of apparatus, during compression of powders, at low pressures,
inter-particle air is removed, and the densification occurs due to particle re-
arrangement. At high pressures, further densification and bonding of particles occur by
elastic deformation, plastic deformation, and fragmentation of particles. Particles that
are brittle in nature consolidate predominantly by fragmentation, whereas particles that
have plastic behavior deform by plastic (ductile) flow.
During compression of powders, some amount of air may be entrapped in the
powder mass due to inadequate deaeration. This would result in high expansion of
pellets/briquettes after ejection from the densification cylinder. The deaeration process
could be affected by the speed of compression, and the amount of openings/provisions
available for escape of air. Complete and reliable removal of air or gas from the
densification equipment is important to make good quality densified products (Pietsch,
1991). A literature review on the effect of compression speed on the densification of
powders is presented below.
The speed of compression has been reported to affect the performances of the
commercial densification machines used for the production of briquettes, pellets, and
pharmaceutical tablets. In roll press compaction/briquetting machine, the feeding is

374
done by gravity feeder (for free flowing particles), or force feeder (powder is pushed
towards the rolls by one or more screws). The air fed with the powder can only escape
by two paths: axially through the powder, counter-currently to the feed; and through
the gap between rolls and cheek plate (Pietsch, 1991; Guigon and Simon, 2003).
Oftentimes, because of inefficient air removal from the feed, some air can be
compressed inside the compact. This is a key factor limiting compaction production
throughput and compact quality. One solution to enhance deaeration in roll press
machines is vacuum deaeration before the roll nip region. Vacuum deaeration has been
found to be efficient in optimizing roll press compaction and minimizing un-
compacted powder leakage through rolls (Pietsch, 1991; Guigon and Simon, 2003).
Also, according to Pietsch (1991), better deaeration can be achieved using increased
roll diameter and/or decreased roll speed. This would help reduce the expansion of
compacted material due to a more complete conversion of elastic into permanent,
plastic deformation of powder particles. Typical screw feeder speeds are 7.5 to 300
rpm, and roll speeds are 1 to 16 rpm (0.01 to 1.5 m s"1) (Pietsch, 1991; Guigon and
Simon, 2003).

A review by Kaliyan and Morey (2006a) [Chapter 3] found that pelleting of


animal feed in conventional ring-die pelleting machines required optimum ring-die
speeds of 150 to 250 rpm (4 to 10 m s"1). Higher speeds of >250 rpm (up to 500 rpm)
were found to be unsatisfactory for pelleting of several animal feeds and alfalfa.
Lower speeds (4 to 7 m s"1) were used for larger diameter pellets or for low bulk
density feed.
Marshall et al. (1993) reported that for plastically deforming materials, a
decrease in compaction speed produced an increase in the compact density at a given
pressure, but for brittle materials, the density of the compacts was independent of
compaction speed. At lower speeds, higher amount of plastic deformation would occur
due to the time dependency of the plastic deformation. They found that axial recovery
(i.e., expansion) of the compacts increased with increased compaction speed as the
opportunity for the plastic deformation was reduced, as a result of decreased time
available for the plastic deformation. Akande et al. (1997) found that tensile strength

375
of the tablets decreased and the elastic recovery of the tablets increased as the
compression speed was increased. These effects were assumed to be caused by the
reduction in the stress relaxation and bond formation at higher speeds. The
compression speeds involved in Marshall et al. (1993) and Akande et al. (1997) ranged
from 15 to 390 mm s"1. Tye et al. (2005) also found that tablet density and tensile
strength were dependent on compaction speed. They mentioned that it is a challenge
during scale up and technology transfer when tabletting speeds are significantly
increased, compared to the speeds used for studying the compaction behaviors of the
pharmaceutical powders in the laboratory. Compressed pockets of air can be an
important reason for tablet failure such as capping (Pietsch, 1991; Mittal and Puri,
2003).

A.3 Objectives

Following the literature (e.g., Mani et al., 2004a), we fabricated a uniaxial,


piston-cylinder densification apparatus to study the densification characteristics of the
corn stover and switchgrass in the laboratory (Chapter 4; Kaliyan and Morey, 2006b).
We found that a constant compression speed of 25.4 mm min"1 (1.0 in. min"1) helped
to achieve a maximum compression pressure of 150 MPa in an Instron universal
testing machine. Higher speeds could be achieved in the Instron; however, it was
difficult to stop the application of pressure at the exact set pressure because of higher
momentum of the Instron at higher speeds. The speed of 25.4 mm min" may be
considered to be very slow compared to the speeds used in commercial densification
machines. Therefore, our densification apparatus may not impose any compression
speed related effects on the quality of the compacts (i.e., briquettes) produced.
However, the amount of open area available for deaeration may affect the quality of
the briquettes produced. Therefore, the objective of this study was to determine the
effect of enhanced deaeration during densification of corn stover and switchgrass in
the uniaxial, piston-cylinder densification apparatus.

376
A.4 Materials and Methods

A.4-1 Biomass Samples

Corn stover used for the study was collected from freshly harvested bales
stored outdoors in West Central Research and Outreach Center (WCROC), University
of Minnesota, Morris, MN during November 2004. Switchgrass samples were
obtained by manually harvesting the whole switchgrass plants (about 50-mm above
the ground) during August 2005 from a field at WCROC, Morris, MN. The initial
moisture content of the switchgrass was about 48% (wet basis). The switchgrass was
sun-dried to about 10% moisture content and then it was chopped into lengths of about
100 to 150-mm before grinding in a hammer mill.
A 5.6-kW hammer mill (J.B. Sedberry Inc., Franklin, TN) with a screen having
3.0 mm round holes was used to obtain corn stover and switchgrass grinds. Corn
stover and switchgrass grinds were stored in a refrigerator maintained at 5C until used
for the tests. To adjust the moisture content of the grinds to 10% (wet basis),
predetermined quantities of distilled water were added to the grinds, thoroughly mixed
and stored in zip-lock plastic bags at 5C for 48 h for moisture equilibration. Moisture
content of the grinds was measured using the procedure given in ASAE Standard
S358.2 (ASAE Standards, 2003a). The moisture content values reported in this paper
are on wet basis (w.b.). Particle size distribution of the grinds was determined based
on ASAE Standard S319.3 (ASAE Standards, 2003b).

A.4-2 Briquetting Procedure

Briquettes were made in the laboratory using a uniaxial, piston-cylinder


densification apparatus. The inner diameter and height of the steel cylinder (i.e., die)
were 18.85 mm and 300.0 mm, respectively. The piston was made of brass (18.84-mm
diameter) and was 50.0-mm longer than the cylinder to help eject briquettes from the
cylinder. An INSTRON model 4206 universal testing machine (Instron Corporation,

377
Canton, MA) was used to apply a mechanical pressure of 150 MPa for briquetting the
corn stover and switchgrass grinds. A separate apparatus was used to allow ejection of
briquettes from the cylinder after compression. The details of the uniaxial, piston-
cylinder compression (densification) apparatus and briquette ejection apparatus can be
found in Kaliyan and Morey (2006b) [Chapter 4].
About 5.0-g of corn stover or switchgrass grind was added to the cylinder
through a funnel. Using a steel rod, the grind was stirred to help the flow of grind from
the funnel. For the cases with preheating to 75C, before compression of biomass
grind to form briquettes, the grind along with the cylinder was heated to achieve a
preheating temperature of 75C. During preheating, the top opening of the cylinder
was covered with a specially designed vapor-tight cover. The outside of the
compression cylinder wall and base steel plate were covered with heating tapes (volt =
120, amps = 2.58, watts - 310, and phase = 1; BH Thermal Corporation, Columbus,
OH) to heat the biomass grind along with the cylinder. About 50.0-mm thick
fiberglass insulation was used to cover the heating tapes to avoid heat loss.
Temperature of the grind was measured inside of the cylinder at the center and about
half the height of the biomass grinds in the cylinder. The distance from the top surface
of the base plate where the tip of the temperature sensor (K-type thermocouple) was
positioned (i.e., the half the height of the biomass grinds) was 95.0 mm for corn stover
grind, and 50.0 mm for switchgrass grind. Immediately after reaching the grind
temperature of 75C, the power to the heating tapes was turned off, the cylinder was
attached to the base of the INSTRON, and the vapor-tight cover on the top of the
cylinder was removed from the cylinder. The piston attached to the crosshead of the
INSTRON was actuated to compress the grind to a maximum pressure of 150 MPa at
a constant speed of 25.4 mm min"1 (1.0 in. min"1).

After completion of compression, the piston was taken out of the cylinder by
unloading the piston to 0 MPa at a speed of 25.4 mm min"1. The cylinder was removed
from the base of the INSTRON, the base steel plate was detached from the bottom of
the cylinder to expose the briquettes, and the cylinder placed on the briquette ejection
apparatus. A constant crosshead speed of 25.4 mm min"1 was applied to the piston to

378
eject the briquette from the cylinder. The time, force, and distance traveled by the
crosshead were recorded during the briquette compression and ejection to estimate the
specific energy consumption for the briquette compression (i.e., compression energy)
and ejection (i.e., ejection energy). To calculate the specific energy consumption for
compression or ejection, the area under the force-displacement curve was estimated
using the trapezoidal rule (Cheney and Kincaid, 1985).
In this study, the cylinder was not lubricated during briquetting, but
periodically cleaned using a vacuum cleaner. Also, no binding agents (i.e., additives)
were used for the study.

A.4-3 Deaeration Without Steel Filter

During compression of biomass grind in the conventional uniaxial, piston-


cylinder densification apparatus, the cylinder is placed on top of a steel (solid) base
plate. Therefore, during compression of biomass grind, air escapes through the annular
gap between the cylinder and piston (figs. A.l and A.2). For the apparatus used in this
study, the annular gap between the piston and cylinder was 0.01 mm. The cross
sectional area of the gap between the cylinder and piston was 2.96 x 10"7m2. The
cross sectional area of the cylinder-bore was 2.79 x 10"4m2. The open area available
for deaeration can be scaled to the cross-sectional area of the cylinder-bore as:

Percentage open area without filter on the base plate (%) =

Cross sectional area of the gap between


the cylinder and piston (m 2 )
, _ y^ | \J\J
Cross sectional area of the cylinder bore (m )

Thus, the percentage of open area available for the case without steel filter on
the base plate was 0.11% of the cross-sectional area of the cylinder-bore.

379
A.4-4 Deaeration With Steel Filter

Enhanced deaeration from the cylinder was achieved through the bottom of the
cylinder by incorporating a steel filter (21.0-mm diameter x 1.5-mm thick) with 20
micron pores (Pall Trinity Micro Corp., Cortland, NY; www. pall.com) at the center of
the steel base plate (figs. A.l and A.2). Four holes of 3.25-mm diameter and 10.24-
mm deep were made in the base plate right below the filter to help the air that passed
through the filter escape (fig. A.2). Therefore, during compression of biomass grind,
air escaped through the steel filter (i.e., deaeration at the bottom of the cylinder) as
well as through the annular gap between the cylinder and the piston (i.e., deaeration at
the top of the cylinder) (fig. A.l).
After making ten briquettes at each test condition, high pressure air (from an
air compressor) was passed through the filter to clean the filter as well as to ensure
that the pores in the filter were not closed due to the application of the densification
pressure of 150 MPa. At the end of the study, it was found that the pores in the filter
allowed the air to pass through at a pressure of about 5.0 psig (0.035 MPa). It was also
noted that there was a small annular gap between the steel filter (outer edge) and the
base plate because the filter was not welded to the base plate. It appeared that this gap
also helped the air to escape through the four holes in the base plate. Thus, the
effective open area available for deaeration in the base plate was equal to the cross
sectional areas of the four holes in the base plate (i.e., 3.35 x 10 m ).
The open area available for deaeration can be scaled to the cross-sectional area
of the cylinder-bore as:

Percentage open area with filter on the base plate (%) =

[ Cross sectional area of the gap between


the cylinder and piston (m 2 ) +
cross sectional area of the four holes in the base plate (m 2 ) 1 ,
-xlOO
Cross sectional area of the cylinder bore (m )

380
It was calculated that the percentage of effective open area available for
deaeration with steel filter on the base plate was 12.0% of the cross-sectional area of
the cylinder-bore.

A.4-5 Briquetting Experiments

With and without inclusion of the steel filter on the base plate (placed below
the cylinder during compression of biomass grind), corn stover and switchgrass
briquettes were made for the following combination of conditions: (i) maximum
compression pressure of 150 MPa applied at a constant speed of 25.4 mm min"1 (1.0
in. min"1), (ii) grind obtained from the hammer mill with a screen size of 3.0 mm (i.e.,
geometric mean particle diameter of corn stover and switchgrass grind was 0.66 mm
0.32, and 0.56 mm 0.29, respectively.), (iii) moisture content of the grind of 10%
(w.b.), and (iv) grind temperature of 25C (i.e., room temperature) or 75C (i.e.,
preheating temperature).
The preheating temperature of 75C was selected because the glass transition
temperature of the corn stover and switchgrass is about 75C (Chapter 4; Kaliyan and
Morey, 2006b). Preheating of biomass close to its glass transition temperature would
activate the natural binders in the biomass such as lignin, starch, protein, fat, and water
soluble carbohydrates, which would in turn help produce strong and durable briquettes
(Chapter 4; Kaliyan and Morey, 2006a). At each briquetting condition, ten briquettes
were made.

381
A.4-6 Briquette Properties

Immediately after ejection from the die, unit density of the briquettes (i.e.,
density of individual briquettes) was measured. Then, the briquettes were transferred
to zip-lock plastic bags and stored for one week at room temperature. Durability,
percentage expansions in axial direction, radial direction and volume, and moisture
content of briquettes were measured after one week of storage.
Durability of briquettes was measured according to ASAE Standard S269.4
(ASAE Standards, 2003c). Durability was calculated as the percentage of the original
mass of the briquettes retained on a 16.0-mm screen after tumbling in a durability
tester (Continental-Agra Equipment, Inc., Newton, KS) at 50 rpm for 10 min. Only
two replications were used for the durability measurement because of lack of samples.
For each replication, five briquettes were used. Unit density of briquettes was
calculated from the mass, diameter, and height of the briquettes (ASAE Standards,
2003c). Axial, radial, and volume expansions of briquettes were calculated as the
percentage increase in height, diameter, and volume, respectively. Moisture content of
the briquettes was measured based on ASAE Standard S358.2 (ASAE Standards,
2003c).
Statistical analyses (independent samples T-test at a = 0.05) using the software
SPSS 16.0 for Windows (SPSS Inc., Chicago, IL) were conducted on the means of
density or durability values measured after one week of storage of briquettes at room
temperature to investigate the significance of enhanced deaeration (i.e., with steel
filter) compared to the deaeration condition without steel filter at 25C or 75C (table
A.l).

A.5 Results and Discussion

A. 5-1 Corn Stover

For with (i.e., 75C) or without preheating (i.e., 25C) conditions, the density
and durability of corn stover briquettes made without a steel filter on the base plate
382
(i.e., 0.11% open area for deaeration) were similar to those values obtained for the
case with the steel filter on the base plate (i.e., 12% open area for deaeration) (P >
0.05) (table A.l). At 75C, the slight decrease in the briquette density values for the
case with the steel filter may be due to the slightly higher initial moisture content of
the grind used for the test. In addition, excessive moisture loss due to preheating to
75C or during storage may have affected the density and durability values (table
A.2).
The volume expansion of corn stover briquettes showed greater differences
between the cases with and without the steel filter (table A.2). For briquetting at 25C,
the volume expansion of briquettes was 5.2% points less for the case with the steel
filter than for the case without the steel filter. Similarly, for briquetting at 75C, the
volume expansion of briquettes was 2.4% points less for the case with the steel filter
than for without the steel filter. The possible reason for higher volume expansion of
briquettes for the case without the steel filter may be due to the compression of some
entrapped air in the grind mass due to lower open area available for escape of air/gas.
The data also suggest that preheating to 75C also could help increase the briquette
density and durability, and reduce the volume expansion of the briquettes (tables A.l
and A.2). This may be due to the enhanced activation of the natural binders in the corn
stover at 75C than at 25C (Chapters 3 and 4; Kaliyan and Morey, 2006a and 2006b).
Comparing the effect of preheating to 75C with the effect of enhanced deaeration due
to the availability of larger open area (i.e., 12% open area), the effect of preheating
was dominant in producing better quality briquettes.

A. 5-2 Switchgrass

Increasing the open area available for deaeration from 0.11% to 12%
significantly increased the density of switchgrass briquettes made at 25C or 75C (P
< 0.05) (table A.l). However, the durability of switchgrass briquettes was not
significantly affected by the enhanced deaeration (P > 0.05). Preheating to 75C had a
dramatic effect on the switchgrass briquette durability. With preheating to 75C, the
natural binders in the switchgrass could have been activated, which may have helped
383
to make durable particle-particle bonding (Chapters 3 and 4; Kaliyan and Morey,
2006a and 2006b).
For briquetting at 25C, the volume expansion of switchgrass briquettes was
16.9 percentage points higher for the case with 0.11% open area available for
deaeration than for the case with 12% open area available for deaeration. The
difference between the volume expansions of briquettes measured for the case with
and without the steel filter was greatly reduced when preheating the switchgrass to
75C (table A.2). Furthermore, the effect of preheating was dominant in producing
better quality briquettes than the effect of enhanced deaeration due to the provision of
larger open area.
In conclusion, for both corn stover and switchgrass, the briquette durability
values were not affected by the enhanced deaeration condition, but there was slight
reduction in the volume expansion of the briquettes due to the enhanced deaeration.
However, the difference in the volume expansion of the briquettes made at 0.11% and
12% open area was small for the densification conditions that resulted in better quality
briquettes (i.e., preheating to 75C). Therefore, the piston-cylinder densification
apparatus made by Kaliyan and Morey (2006b) [Chapter 4], which had about 0.11%
open area for deaeration, could be used for determining the optimum densification
conditions such as particle size, moisture content, and preheating temperature for corn
stover and switchgrass, provided that the speed of compression was < 25.4 mm min"1.
Tables A. 1 and A.2 indicate that corn stover produced better quality briquettes
than switchgrass. This may be due to the inherent differences in the cell structure,
chemical composition, and mechanical properties between these two biomass
materials. For both corn stover and switchgrass, the specific energy consumption for
briquetting with and without steel filter on the base plate was similar (table A.3). For
briquetting at room temperature (i.e., 25C), the specific energy consumption ranged
from 0.20 to 0.25% of the energy in the biomass materials. With preheating to 75C,
the specific energy consumption ranged from 0.69 to 0.82% of the energy in the
biomass materials (table A.3).

384
A.6 Conclusions

From the results of this study, the following conclusions could be drawn.

In a closed-end, piston-cylinder densification apparatus, increasing the open


area available for deaeration during compression from 0.11% to 12% of the
cross sectional area of the cylinder-bore did not appear to affect the density and
durability of the corn stover and switchgrass briquettes made at 25 and 75C.
However, the volume expansion of the briquettes (measured after one week)
was 2.4 to 16.9% points less for the case with the steel filter (i.e., 12% open
area) than for the case without the steel filter (i.e., 0.11% open area) mounted
at the bottom of the densification cylinder to increase the open area available
for deaeration during compression.

The literature review showed that inadequate deaeration due to higher speeds
of compression might decrease the quality of the briquettes produced in
commercial-scale densification equipment. With a compression speed of <
25.4 mm min"1, a laboratory-scale, closed-end, piston-cylinder densification
apparatus having an open area of > 0.11% of cross-sectional area of the
cylinder-bore for deaeration could be used for determining the optimum
densification conditions such as particle size, moisture content, and preheating
temperature for biomass materials.

A.7 References

Akande, O.F., M.H. Rubinstein, P.H. Rowe, and J.L. Ford. 1997. Effect of
compression speeds on the compaction properties of a 1:1 paracetamol-
microcrystalline cellulose mixture prepared by single compression and by
combinations of pre-compression and main-compression. Int. J. Pharmaceutics
157(2): 127-136.
ASAE Standards. 2003a. S358.2: Moisture measurement - Forages. St. Joseph, Mich.:
ASABE.
385
ASAE Standards. 2003b. S319.3: Method of determining and expressing fineness of
feed materials by sieving. St. Joseph, Mich.: AS ABE.
ASAE Standards. 2003c. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: AS ABE.
Cheney, W., and D. Kincaid. 1985. Numerical Mathematics and Computing.
Monterey, CA: Brooks/Cole Publishing Company.

DOE. 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry: The
Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.
Department of Energy (DOE), Office of Scientific and Technical Information.
Available at: http://www/osti.gov/bridge. Accessed 18 June 2005.

Guigon, P., and O. Simon. 2003. Roll press design - influence of force feed systems
on compaction. Pow. Technol 130(1-3): 41-48.
Kaliyan, N., and R.V. Morey. 2006a. Factors affecting strength and durability of
densified products. ASABE Paper No. 066077. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2006b. Densification characteristics of corn stover and
switchgrass. ASABE Paper No. 066174. St. Joseph, Mich.: ASABE.
Lemus, R., E.C. Brummer, K.J. Moore, N.E. Molstad, C.L. Burras, and M.F. Barker.
2002. Biomass yield and quality of 20 switchgrass populations in southern
Iowa, USA. Biomass and Bioenergy 23(6): 433-442.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004a. Evaluation of compaction equations
applied to four biomass species. Can. Biosys. Eng. 46(3): 3.55-3.61.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004b. Grinding performance and physical
properties of wheat and barley straws, corn stover and switchgrass. Biomass
and Bioenergy 27(4): 339-352.

Marshall, P.V., P. York, and J.Q. Maclaine. 1993. An investigation of the effect of the
punch velocity on the compaction properties of ibuprofen. Pow. Technol.
74(2): 171-177.

386
Mittal, B., and V.M. Puri. 2003. An elasto-viscoplastic constitutive model
incorporating pore air compressibility during powder compression process.
Particulate Sci. and Technol. 21(2): 131-155.
ORNL. 2005. Biofuels from Switchgrass: Greener Energy Pastures. Oak Ridge, TN:
Oak Ridge National Laboratory (ORNL). Available at:
http://bioenergy.ornl.gov/papers/misc/switgrs.html. Accessed 18 June 2005.
Pietsch, W. 1991. Size Enlargement by Agglomeration. New York, NY: John Wiley &
Sons.
Pordesimo, L.O., B.R. Hames, S. Sokhansanj, and W.C. Edens. 2005. Variation in
corn stover composition and energy content with crop maturity. Biomass and
Bioenergy 28(4): 366-374.
Tye, C.K., C.C. Sun, and G.E. Amidon. 2005. Evaluation of the effects of tableting
speed on the relationships between compaction pressure, tablet tensile strength,
and tablet solid fraction. J. Pharm. Sci. 94(3): 465-472.

387
Pin connection for connecting the load
cell-crosshead of the INSTRON and the
piston

350 mm
Piston

Deaeration when base plate did have or did


not have steel filter (gap = 0.01 mm)

18.8-mm ID cylinder (Die), which is filled with


300 mm 5.0-g of biomass grind before compression

Connection to the base of the INSTRON


lnminSa -d-

Deaeration when base plate had steel filter

Figure A.l. Schematic of uniaxial densification apparatus showing deaeration (i.e., air
removal) paths during compression of biomass grind.

+ '. t*~ r

(A) * ' :.:.?*

Figure A.2. Pictures of base plates (114.29-mm diameter x 11.74-mm thick) with and without
steel filter (21.01-mm diameter x 1.50-mm thick) at their centers. (A) Solid base
plate (no steel filter) for no deaeration through the bottom of the die; (B) Base
plate with steel filter (top side) for deaeration through the bottom of the die; (C)
Base plate with steel filter (bottom side) showing four 3.25-mm diameter holes to
allow air to escape.

388
Table A. 1. Properties of corn stover and switchgrass briquettes.
Deaeration Preheating Grind Immediately after ejection After one week of storage at
method temperature moisture from the die room temperature
("C) content (%, Unit briquette density Unit briquette Durability
w.b.) (kg m'3) (n = 5) * density (%)
(n = 3)*' (kg m'3) (n = 2)*'''
(n = 5 ) * v
Corn stover grind from hammer mill screen size of 3.0 mm
(Particle size = 0.66 mm 0.32 **)
Without 25 10.1 0.3 1117.9 9.1 997.4 11.1 75.2 0.6
steel filter a a
With steel 25 10.2 0.3 1088.7 10.0 1002.7 74.9 0.5
filter 11.2a a
Without 75 10.1 0.3 1191.2 7.0 1153.9 8.8 96.8
steel filter a 0.04 a
With steel 75 10.2 0.3 1162.6 10.6 1144.8 91.5 1.8
filter 14.9 a a
Switchgrass grind from hammer mill screen size of 3.0 mm
(Particle size = 0.56 mm 0.29 **)
Without 25 9.7 0.2 889.0 7.4 678.9 10.9 0.0 0.0
steel filter a
With steel 25 10.6 0.3 989.8 3.7 874.5 12.4 0.0 0.0
filter b
Without 75 9.8 0.2 1093.8 7.4 1045.2 62.9 0.2
steel filter 15.8 a a
With steel 75 10.6 0.3 1102.7 15.5 1077.0 61.4 2.1
filter 10.6 b a
* Mean standard deviation.
* Moisture content of the grind before densification and preheating (if applicable).
** Geometric mean particle diameter geometric standard deviation (n = 3).
l|
' When comparing deaeration method (without versus with steel filter) at 25 or 75C, means of unit
briquette density or durability followed by similar lowercase letters are not significantly different (P >
0.05). Comparison between temperatures or between biomass materials should not be made because the
statistical tests were conducted only to compare the difference between without and with the steel filter
for one biomass material (corn stover or switchgrass) and one temperature (25 or 75C) at a time.

389
Table A.2. Moisture content and stability of corn stover and switchgrass briquettes
measured after one week of storage at room temperature.

Deaeration Preheating Grind Briquette moisture Mean expansion (%) (n = 5)


method temperature moisture content (%, w.b.) Axial Radial Briquette
(C) content (%, (n = 3)* direction direction volume
w.b.) (n = 3)

Corn stover grind from hammer mill screen size of 3.0 mm


(Particle size = 0.66 mm 0.32 **)
Without 25 10.1 0.3 *** NA 10.7 0.6 11.9
steel filter
With 25 10.2 0.3 8.8 0.2 7.0 -0.14& 6.7
steel filter
Without 75 10.1 0.3 7.7 0.1 2.3 0.2 2.6
steel filter
With 75 10.2 0.3 6.6 0.1 0.5 -0.16 0.24
steel filter
Switchgrass grind from hammer mill screen size of 3.0 mm
(Particle size = 0.56 mm 0.29 **)
Without 25 9.7 0.2 NA 26.0 0.9 28.2
steel filter
With 25 10.6 0.3 8.5 0.1 10.0 0.6 11.3
steel filter
Without 75 9.8 0.2 7.0 0.2 4.0 0.3 4.5
steel filter
With 75 10.6 0.3 6.5 0.2 1.7 0.0 1.7
steel filter
* Mean standard deviation.
#
Moisture content of the grind before densification and preheating (if applicable).
** Geometric mean particle diameter geometric standard deviation (n = 3).
*** NA = Data were not taken. However, the moisture loss from the briquettes made at room
temperature appeared to be small.
&
Decrease in briquette diameter may have been due to the shrinkage of briquettes due to the moisture
loss during storage.

390
Table A.3. Specific energy required for briquetting corn stover and switchgrass.

Deaeration Preheating Grind Compression Ejection Preheating Specific energy


method temperature moisture energy (MJ/t) energy energy consumption *
(C) content (n = 10)* (MJ/t) (MJ/t) ** MJ/t Percentage
(%, w.b.) (n = 10) * of energy
(n = 3) in material
*.*
Com stover grind from hammer mill screen size of 3.0 mm
(Particle size = 0.66 mm 0.32 ***)
Without 25 10.1 44.3 0.4 1.1 0.0 45.4 0.23
steel 0.3 0.1
filter
With 25 10.2 40.1 0.2 0.2 0.0 40.3 0.20
steel 0.3 0.1
filter
Without 75 10.1 40.0 0.7 1.1 100.0 141.1 0.71
steel 0.3 0.1
filter
With 75 10.2 36.7 0.6 0.2 100.0 136.9 0.69
steel 0.3 0.02
filter
Switchgrass grind from hammer mill screen size of 3.0 mm
(Particle size = 0.56 mm 0.29 ***)
Without 25 9.7 0.2 36.0 0.3 0.3 0.0 36.3 0.21
steel 0.1
filter
With 25 10.6 42.7 0.3 0.3 0.0 43.0 0.25
steel 0.3 0.02
filter
Without 75 9.8 0.2 38.5 1.1 0.4 100.0 138.9 0.82
steel 0.1
filter
With 75 10.6 37.6 0.7 0.3 100.0 137.9 0.81
steel 0.3 0.04
filter
* Mean standard deviation.
* Moisture content of the grind before densification and preheating (if applicable).
** Preheating energy (MJ/t) = Specific heat of biomass x [preheating temp. - room temp.].
Assumed specific heat of biomass (corn stover or switchgrass) = 2.0 kJ kg"1 K"'; room temperature =
25C.
*** Geometric mean particle diameter geometric standard deviation (n = 3).
* Specific energy consumption = Compression energy + ejection energy + preheating energy.
&
Percentage of energy in material (%) = Specific energy consumption (MJ/t) x 100 / Energy content of
biomass (MJ/t).
Energy content of corn stover = 20 MJ/kg (Pordesimo et al., 2005).
Energy content of switchgrass = 17 MJ/kg (Lemus et al., 2002; Mani et al., 2004b).

391
Bibliography

Aarseth, K.A. 2004. Attrition of feed pellets during pneumatic conveying: the
influence of velocity and bend radius. Biosys. Eng. 89(2): 197-213.
Abd-Elrahim, Y.M., A.S. Huzayyin, and I.S. Taha. 1981. Dimensional analysis and
wafering cotton stalks. Trans. ASAE 24(4): 829-832.
Adachi, T., and F. Oka. 1982. Constitutive equations for normally consolidated clay
based on elasto-viscoplasticity. Soils and Foundations 22(4): 57-70.
Adapa, P., G. Schoenau, L. Tabil, S. Sokhansanj, and A. Singh. 2005. Compression of
fractionated sun-cured and dehydrated alfalfa chops into cubes: pressure and
density models. Can. Biosys. Eng. 47(3): 3.33-3.39.
Adler, P.R., M.A. Sanderson, A.A. Boateng, P.J. Weimer, and H.G. Jung. 2006.
Biomass yield and biofuel quality of switchgrass harvested in fall or spring.
Agron.J. 98(6): 1518-1525.
Akande, O.F., M.H. Rubinstein, P.H. Rowe, and J.L. Ford. 1997. Effect of
compression speeds on the compaction properties of a 1:1 paracetamol-
microcrystalline cellulose mixture prepared by single compression and by
combinations of pre-compression and main-compression. Int. J. Pharmaceutics
157(2): 127-136.
Al-Widyan, ML, and H.F. Al-Jalil. 2001. Stress-density relationship and energy
requirement of compressed olive cake. Applied Eng. inAgric. 17(6): 749-753.
Angulo, E., J. Brufau, and E. Esteve-Garcia. 1995. Effect of sepiolite on pellet
durability in feeds differing in fat and fibre content. Animal Feed Sci. Tech.
53(3-4): 233-241.
Angulo, E., J. Brufau, and E. Esteve-Garcia. 1996. Effect of a sepiolite product on
pellet durability in pig diets differing particle size and in broiler starter and
finisher diets. Animal Feed Sci. Tech. 63(1-4): 25-34.
Antikainen, O., and J. Yliruusi. 2003. Determining the compression behavior of
pharmaceutical powders from the force-deformation profile. Int. J. Pharm.
252(1-2): 253-261.
Aqa, S., and S.C. Bhattacharya. 1992. Densification of preheated sawdust for energy
conservation. Energy 17(6): 575-578.
Arthur, J.F., R.A. Kepner, J.B. Dobie, G.E. Miller, and P.S. Parsons. 1982. Tub
grinder performance with crop and forest residues. Trans. ASAE 25(6): 1488-
1494.
AS ABE Standards. 2003 a. S3 5 8.2: Moisture measurement - Forages. St. Joseph,
Mich.: ASABE.
AS ABE Standards. 2003b. S319.3: Method of determining and expressing fineness of
feed materials by sieving. St. Joseph, Mich.: ASABE.
ASABE Standards. 2003c. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: ASABE.
392
ASAE Standards. 2003a. S358.2: Moisture measurement - Forages. St. Joseph, Mich.:
ASABE.
ASAE Standards. 2003b. S319.3: Method of determining and expressing fineness of
feed materials by sieving. St. Joseph, Mich.: ASABE.
ASAE Standards. 2003c. S269.4: Cubes, pellets, and crumbles - Definitions and
methods for determining density, durability, and moisture content. St. Joseph,
Mich.: ASABE.
ASTM. 1998a. C39-96: Standard test method of compressive strength of cylindrical
concrete specimens. Annual book of ASTM Standards, vol. 04.02. West
Conshohocken, PA: American Society for Testing and Materials, pp. 17-21.
ASTM. 1998b. D441-86: Standard test method of tumbler test for coal. Annual book
of ASTM Standards, vol. 05.05. West Conshohocken, PA: American Society
for Testing and Materials, pp. 192-194.
ASTM. 1998c. D440-86: Standard test method of drop shatter test for coal. Annual
book of ASTM Standards, vol. 05.05. West Conshohocken, PA: American
Society for Testing and Materials, pp. 188-191.
Aydin, I., B.J. Briscoe, and K.Y. Sanliturk. 1996. The internal form of compacted
ceramic components: a comparison of a finite element modeling with
experiment. Pow. Technol. 89(3): 239-254.
Aydin, I., B.J. Briscoe, and N. Ozkan. 1997. Modeling of powder compaction: a
review. MRS Bulletin 22(12): 45-51.
Back, E.L. 1987. The bonding mechanism in hardboard manufacture. Holzforschung
41(4): 247-258.
Back, E.L., andN.L. Salmen. 1982. Glass transitions of wood components hold
implications for molding and pulping processes. TappiJ. 65(7): 107-110.
Baker, J.L., and G.C. Shove. 1978. Solar drying of commercially produced large
round hay bales. ASAE Paper No. 78-3066. St. Joseph, Mich.: ASABE.
Balatinecz, J.J. 1983. The potential role of densification in biomass utilization. In
Biomass Utilization, 181-190. W.A. Cote, ed. New York, NY: Plenum Press.
Balk, W.A. 1964. Energy requirements for dehydrating and pelleting coastal
bermudagrass. Trans. ASAE 7(3): 349-351, 355.
Behnke, K.C. 1994. Factors affecting pellet quality. Proc. of Maryland Nutrition
Conference. Department of Poultry Science and Animal Science, College of
Agriculture, University of Maryland, College Park.
Behnke, K.C. 2006. The art (science) of pelleting. Feed Technology Technical Report
Series, American Soybean Association International Marketing Southeast Asia,
Singapore, pp. 5-9.
Bellinger, P.L., and H.H. McColly. 1961. Energy requirements for forming hay
pellets. Agri. Eng. 42(5): 244-247.
Beven, R.R. 1977. A Review of Industrial Size Reduction Equipment Used in the
Processing of Coal. Report prepared for the United States Energy Research and
Development Administration under contract EX-76-C-01-2475. Danville, PA:
Kennedy Van Saun Corporation.

393
Bhattacharya, S.C., G.Y. Saunier, N. Shah, and N. Islam. 1985a. Densification of
biomass residues in Asia. In Bioenergy 84. Vol III Biomass Conversion, 559-
563. H. Egneus and A. Ellegard, eds. London: Elsevier Applied Science
Publishers.
Bhattacharya, S.C., R. Bhatia, M.N. Islam, and N. Shah. 1985b. Densified biomass in
Thailand: potential, status and problems. Biomass 8(4): 255-266.
Bika, D., G.I. Tardos, S. Panmai, L. Farber, and J. Michaels. 2005. Strength and
morphology of solid bridges in dry granules of pharmaceutical powders.
Powder Technol. 150(2): 104-116.
Bilanski, W.K., and V.A. Graham. 1984. A viscoelastic model for forage wafering.
Trans. CSME 8(2): 70-76.
Bodig, J., and B.A. Jayne. 1982. Mechanics of Wood and Wood Composites. New
York, NY: Van Nostrand Reinhold Company.
Bradfield, J., and M.P. Levi. 1984. Effect of species and wood to bark ratio on
pelleting of southern woods. Forest Products J. 34(1): 61-63.
Briggs, J.L., D.E. Maier, B.A. Watkins, and K.C. Behnke. 1999. Effects of ingredients
and processing parameters on pellet quality. Poultry Sci. 78(10): 1464-1471.
Briscoe, B.J., and S.L. Rough. 1998. The effects of wall friction on the ejection of
pressed ceramic parts. Pow. Technol. 99(3): 228-233.
Brown, R.C. 2003. Biorenewable Resources, Engineering New Products from
Agriculture. Ames, IA: Iowa State Press.
Bruhn, H.D., A. Zimmerman, and R.P. Niedermeier. 1959. Developments in pelleting
forage crops. Agric. Eng. 40(2): 204-207.
Butler, J.L., and H.F. McColly. 1959. Factors affecting the pelleting of hay. Agric.
Eng. 40(8): 442-446.
Carson, J.M., and B.G. Kreider. 1988. Forced heated air drying of hay. ASAE Winter
Meeting Paper No. 88-6583. St. Joseph, Mich.: ASABE.
Cavalcanti, W.B. 2004. The effect of ingredient composition on the physical quality of
pelleted feeds: a mixture experimental approach. Ph.D. diss. Manhattan, KS:
Kansas State University.
Chancellor, W.J. 1962. Formation of hay wafers with impact loads. Agric. Eng. 43(3):
136-138, 149.
Chen, P.Y.S., J.G. Haygreen, and M.A. Graham. 1989. An evaluation of wood/coal
pellets made in a laboratory pelletizer. Forest Products J. 39(7/8): 53-58.
Cheney, W., and D. Kincaid. 1985. Numerical Mathematics and Computing.
Monterey, CA: Brooks/Cole Publishing Company.
Chin, O.C., and K.M. Siddiqui. 2000. Characteristics of some biomass briquettes
prepared under modest die pressures. Biomass and Bioenergy 18(3): 223-228.
Clarke, D.E., and H. Marsh. 1989. Factors influencing properties of coal briquettes.
Fwe/68(8): 1031-1038.
Colley, Z., O.O. Fasina, D. Bransby, and C.W. Wood. 2005. Compaction behavior of
poultry litter and switchgrass. ASAE Paper No. 056053. St. Joseph, Mich.:
ASABE.
394
Colley, Z., O.O. Fasiba, D. Bransby, and Y.Y. Lee. 2006. Moisture effect on the
physical characteristics of switchgrass pellets. Trans. ASABE 49(6): 1845-
1851.
Cook, J., and J. Beyea. 2000. Bioenergy in the United States: progress and
possibilities. Biomass and Bioenergy 18(6): 441-455.
Cundiff, J.S., and L.S. Marsh. 1996. Harvest and storage costs for bales of switchgrass
in the southeastern United States. Bioresource Technol. 56(1): 95-101.
Cundiff, S.J., P.P. Ravula, and R.D. Grisso. 2004. Management system for biomass
delivery at a plant conversion. ASAE Paper No. 046169. St. Joseph, Mich.:
ASABE.
Cunningham, J.C., I.C. Sinka, and A. Zavaliangos. 2004. Analysis of tablet
compaction. I. Characterization of mechanical behavior of powder and
powder/tooling friction. J. Pharm. Sci. 93(8): 2022-2039.
Curley, R.G., J.B. Dobie, and P.S. Parsons. 1973. Comparison of stationary and field
cubing of forage. Trans. ASAE 16(2): 361-364, 366.
De Kam, M.J., R.V. Morey, and D.G. Tiffany. 2007. Integrating biomass to produce
heat and power at ethanol plants. ASABE Paper No. 076232. St. Joseph,
Mich.: ASABE.
Dec, R.T. 1999. Selection of proper roll press operating parameters. Proc. of the
Institute for Briquetting and Agglomeration (IBA), Vol. 26. pp. 29-36.
Dec, R.T. 2002. Optimization and control of roller press operating parameters. Powder
Handling & Process. 14(3): 222-225.
Dec, R.T. 2004. Processing of industrial wastes in the roller press for recovery, recycle
or safe disposal. REWAS'2004 Global symposium on recycling, waste
treatment and clean technology, Madrid, Spain, 26-29 September, 2004.
Dec, R.T., and R.K. Komarek. 1997. Testing and scaling up compaction process using
small laboratory roll press. The Institute for Briquetting and Agglomeration
25th Biennial Conference, Charleston, SC, November 3-5, 1997.
Dec, R.T., A. Zavaliangos, and J.C. Cunningham. 2003. Comparison of various
methods for analysis of powder compaction in roller press. Powder Tech.
130(1-3): 265-271.
DeFrain, J.M., J.E. Shirley, K.C. Behnke, E.C. Titgemeyer, and R.T. Ethington. 2003.
Development and evaluation of a pelleted feedstuff containing condensed corn
steep liquor and raw soybean hulls for dairy cattle diets. Animal Feed Sci.
Tech. 107(1-4): 75-86.
Demirbas, A. 1999. Physical properties of briquettes from waste paper and wheat
straw mixtures. Energy Conversion and Mgmt. 40(4): 437-445.
DiMaggio, F.L., and I.S. Sandler. 1971. Material model for granular soils. J. Eng.
Mechanics ASCE 96: 935-950.
Dobie, J.B. 1961. Materials-handling systems for hay wafers. Agric. Eng. 42(12): 692-
697.
Dobie, J.B. 1975. Cubing tests with grass forages and similar roughages sources.
Trans. ASAE 18(5): 864-866.
395
DOE, 2005. Biomass as Feedstock for a Bioenergy and Bioproducts Industry; The
Technical Feasibility of a Billion-Ton Annual Supply. Oak Ridge, TN: U.S.
Department of Energy (DOE), Office of Scientific and Technical Information.
Available at: http://www.osti.gov/bridge. Accessed 18 June 2005.
DOE. 2007. Biomass Feedstock Composition and Property Database. Washington,
DC: U.S. Department of Energy (DOE). Available at:
http://wwwl .eere.energy.gov/biomass/feedstock_databases.html. Accessed 25
April 2007.
Drucker, D.C., and W. Prager. 1952. Soil mechanics and plastic analysis or limit
design. Quarterly of Applied Mathematics 10(2): 157-165.
Drzymala, Z. 1993. Industrial Briquetting - Fundamental and Methods. Studies in
mechanical engineering. Vol. 13. Warszawa: PWN-Polish Scientific
Publishers.
Edens, W.C., L.O. Pordesimo, and S. Sokhansanj. 2002. Field drying characteristics
and mass relationships of corn stover fractions. ASAE Paper No. 026015. St.
Joseph, Mich.: AS ABE.
Edwards, W.C., and P. Tam. 1987. Equipment for production of a water resistant
densified fuel from forest biomass. In Proc. of the Sixth Canadian Bioenergy R
& D Seminar, 214-218. London: Elsevier Applied Science.
EIA. 2005. Renewable Energy Trends 2004. Washington, DC: Energy Information
Administration. Available at: http://www.eia.doe.gov/. Accessed 20 October
2005.
Faborode, M.O., and J.R. O'Callaghan. 1986. Theoretical analysis of compression of
fibrous agricultural materials. J. Agric. Eng. Res. 35(3): 175-191.
Faborode, M.O., and J.R. O'Callaghan. 1989. A rheological model for the compaction
of fibrous agricultural materials. J. Agric. Eng. Res. 42(3): 165-178.
Fairchild, F., and D. Greer. 1999. Pelleting with precise mixture moisture control.
Feed Int. 20(8): 32-36.
Fasina, O.O., and S. Sokhansanj. 1996. Storage and handling characteristics of alfalfa
pellets. Powder Handling & Processing 8(4): 361-365.
Ferrero, A., J. Horabik, and M. Molenda. 1991. Density-pressure relationships in
compaction of straw. Can. Agri. Eng. 33(1): 107-111.
Franke, M , and A. Rey. 2006. Pelleting quality. World Grain May 2006: 78-79.
Friedrich, W., and K.F. Robohm. 1968. Die Abriebfestigkeit von Pellets und ihre
Abhangigheit vom Pressprozess insbesondere der Kilhlung. Kraftfutter 52: 59-
64.
Garekani, H.A., J.L. Ford, M.H. Rubinstein, and A.R. Rajabi-Siahboomi. 2000.
Highly compressible paracetamol - II. Compression properties. Int. J. Pharm.
208(1-2): 101-110.
Ghebre-Sellassie, I. 1989. Pharmaceutical Pelletization Technology. New York, NY:
Marcel Dekker.

396
Gilpin, A.S., T.J. Herrman, K.C. Behnke, and F.J. Fairchild. 2002. Feed moisture,
retention time, and steam as quality and energy utilization determinants in the
pelleting process. Applied Eng. inAgric. 18(3): 331-338.
Graham, V.A., and W.K. Bilanski. 1984. Non-linear viscoelastic behavior during
forage wafering. Tram. ASAE 27(6): 1661-1665.
Grover, P.D., and S.K. Mishra. 1996a. Biomass Briquetting: Technology and
Practices. Regional Wood Energy Development Program in Asia, Field
Document No. 46. Bangkok, Thailand: Food and Agriculture Organization of
the United Nations.
Grover, P.D., and S.K. Mishra. 1996b. Proceedings of the International Workshop on
Biomass Briquetting. Regional Wood Energy Development Program in Asia,
RWEDP Report No. 23. Bangkok, Thailand: Food and Agriculture
Organization of the United Nations.
Guigon, P., and O. Simon. 2003. Roll press design - influence of force feed systems
on compaction. Pow. Technol. 130(1-3): 41-48.
Gustafson, A.S., and W.L. Kjelgaard. 1963. Hay pellet geometry and stability. Agric.
Eng. 44(8): 442 - 445.
Haddad, Y.M. 1995. Viscoelasticity of Engineering Materials. New York, NY:
Chapman & Hall.
Hall, G.E., and C.W. Hall. 1968. Heated die wafer formation of alfalfa and bermuda
grass. Trans. ASAE 11(4): 578-581.
Hartman, H. 1996. The self-propelled briquetting machine for biofuels - features and
chances of the Haimer-Biotruck 2000. In Biomass for Energy and the
Environment. Proceedings of the 9' European Bioenergy Conf, Copenhagen,
Denmar, 24-27 June 1996, 839-845. Amsterdam: Pergamon Press, Elsevier.
Haussmann, F. 1975. Briquetting wood waste by the Fred Haussmann method.
Institute for Briquetting and Agglomeration Proceedings, Vol. 14. pp. 75-90.
Heckel, R.W. 1961. An analysis of powder compaction phenomena. Trans.
Metallurgical Society ofAIME 221 (October): 1001-1008.
Heffner, L.E., and H.B. Pfost. 1973. Gelatinization during pelleting. Feedstuff '45(23):
33.
Heinemans, H. 1991. The interaction of practical experience and the construction of
new pelleting and cooling machinery. Advances in Feed Tech. 6: 24-38.
Hill, B., and D.A. Pulkinen. 1988. A Study of the Factors Affecting Pellet Durability
and Pelleting Efficiency in the Production of Dehydrated Alfalfa Pellets.
Saskatchewan, Canada: Saskatchewan Dehydrators Association.
Holley, C.A. 1983. The densification of biomass by roll briquetting. Proceedings of
the IBA 18th Biennial Conference, 95-102.
Holt, G.A., J.L. Simonton, M.G. Beruvides, and A. Canto. 2004. Utilization of cotton
gin by-products for the manufacturing of fuel pellets: an economic perspective.
Applied Eng. inAgric. 20(4): 423-430.
Hoque, M , S. Sokhansanj, T. Bi, S. Mani, L. Jafari, L. Lim, P. Zaini, S. Melin, T.
Sowlati, and M. Afzal. 2006. Economics of pellet production for export
397
market. CSBE/SCGAB Paper No. 06-103. The Canadian Society for
Bioengineering, CSBE/SCGAB 2006 Annual Conference, Edmonton, Alberta,
Canada.
Hoskinson, R.L., D.L. Karlen, S.J. Birrell, C.W. Radtke, and W.W. Wilhelm. 2007.
Engineering, nutrient removal, and feedstock conversion evaluations of four
corn stover harvest scenarios. Biomass and Bioenergy 31(2-3): 126-136.
Huang, B.K., and R.R. Yoerger. 1961. Maceration as a pretreatment in hay wafering.
Trans. ASAE 4(1): 69-71.
Huang, L., and V.M. Puri. 2000. A three-dimensional finite element model for powder
compaction - time-dependent formulation and validation. Particulate Sci. and
Technol. 18(1): 257-274.
Inghelbrecht, S., and J.P. Remon. 1998. The roller compaction of different types of
lactose. Int. J. Pharm. 166(2): 135-144.
Israelsen, M., J. Busk, and J. Jensen. 1981. Pelleting properties of dairy compounds
with molasses, alkali-treated straw and other byproducts. Feedstuffs 7: 26-28.
Iyengar, M. 1959. The problem of briquetting slack coal in India. Proc. of 6th Biennial
Briquetting Conf. of the International Briquetting Association, August 24-26,
Glacier Park, MT. pp. 20-29.
Jannasch, R., Y. Quan, and R. Samson. 2001. A Process and Energy Analysis of
Pelletizing Switchgrass. Final Report. Ste. Anne de Bellevue, QC: Resource
Efficient Agricultural Production (REAP-Canada, www.reap-canada.com).
Jones, D., and J. Jones. 1980. Wood chips vs. densified biomass: an economic
comparison. Symposium papers "Energy from Biomass and Wastes IV". Hotel
Royal Plaza, Lake Buena Vista, Florida, Jan 21-25, 1980. Sponsored by
Institute of Gas Technology, Chicago, pp. 223-249.
Kaar, W.E., and M.T. Holtzapple. 2000. Using lime pretreatment to facilitate the
enzymatic hydrolysis of corn stover. Biomass and Bioenergy 18(3): 189-199.
Kadam, K.L., and J.D. McMillan. 2003. Availability of corn stover as a sustainable
feedstock for bioethanol production. Bioresource Technol. 88(1): 17-25.
Kaliyan, N., and R.V. Morey. 2005. Densification of corn stover. ASAE Paper No.
056134. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2006a. Densification characteristics of corn stover and
switchgrass. ASABE Paper No. 066174. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2006b. Factors affecting strength and durability of
densified products. ASABE Paper No. 066077. St. Joseph, Mich.: ASABE.
Kaliyan, N., and R.V. Morey. 2007. Roll press briquetting of corn stover and
switchgrass: a pilot scale continuous briquetting study. ASABE Paper No.
076044. St. Joseph, Mich.: ASABE.
Kaltschmitt, M., and M. Weber. 2006. Markets for solid biofuels within the EU-15.
Biomass and Bioenergy 30(11): 897-907.
Khankari, K.K., M. Shrivastava, and R.V. Morey. 1989a. Densification characteristics
of rice hulls. ASAE Paper No. 89-6093. St. Joseph, Mich.: ASABE.

398
Khankari, K.K., M. Shrivastava, and R. Weggel. 1989b. Densification characteristics
of rice hulls under hot uniaxial compression. Institute for Briquetting and
Agglomeration Proceedings, Vol. 21. pp. 143-147.
Khankari, K.K., R.V. Morey, J.E. Fruin, and D.W. Halbach. 1989c. Engineering cost
estimates for pelleting Minnesota soybean meal. Department of Agricultural
and Applied Economics Staff Paper No. P89-42. St. Paul, MN: University of
Minnesota.
Khoshtaghaza, M.H., S. Sokhansanj, and B.D. Gossen. 1999. Quality of alfalfa cubes
during shipping and storage. Applied Eng. inAgric. 15(6): 671-676.
Kim, H., G. Lu, T. Li, and M. Sadakata. 2002. Binding and desulfurization
characteristics of pulp black liquor in biocoalbriquettes. Environ. Sci. Tech.
36(7): 1607-1612.
Klass, D.L. 1998. Biomassfor Renewable Energy, Fuels, and Chemicals. San Diego,
CA: Academic Press, Elsevier.
Kleinebudde, P. 2004. Roll compaction/dry granulation: pharmaceutical applications.
Eur. J. Pharm. Biopharm. 58(2): 317-326.
Komarek, K.R. 1963. The roll type briquette machine. Proceedings of the IBA 8th
Biennial Briquetting Conference, 35-37.
Komarek, R.K. 1991. Binderless briquetting of peat, lignite, sub-bituminous and
bituminous coals in roll presses. Proc. of the Institute for Briquetting and
Agglomeration (IBA), Vol. 22. pp. 223-242.
Koser, H.J.K., G. Schmalstieg, and W. Siemers. 1982. Densification of water
hyacinth-basic data. Fuel 61(9): 791-798.
Koval'chenko, M.S. 1990. A rheological model of pressing of powders. Poroshkovaya
Metallurgiya 333(9): 100-104.
Kumar, A., and S. Sokhansanj. 2007. Switchgrass (Panicum virgatum, L.) delivery to
a biorefinery using integrated biomass supply analysis and logistics (IBSAL)
model. Bioresource Technol. 98(5): 1033-1044.
Lemus, R., E.C. Brummer, K.J. Moore, N.E. Molstad, C.L. Burras, and M.F. Barker.
2002. Biomass yield and quality of 20 switchgrass populations in southern
Iowa, USA. Biomass and Bioenergy 23(6): 433-442.
Li, Y., and H. Liu. 2000. High-pressure densification of wood residues to form an
upgraded fuel. Biomass and Bioenergy 19(3): 177-186.
Li, Y., H. Liu, and A. Rockabrand. 1996. Wall friction and lubrication during
compaction of coal logs. Pow. Technol. 87(3): 259-267.
Lindley, J. A., and M. Vossoughi. 1989. Physical properties of biomass briquets.
Trans. ASAE 32(2): 361-366.
Lobo, P. 2002. The right grinding solution for you: roll, horizontal or vertical. Feed
Management 53(3): 23-26.
Lobosco, V., and V. Kaul. 2001. An elastic/viscoplastic model of the fiber network
stress in wet pressing: part I.Nordic Pulp and Paper Res. J. 16(1): 12-17.
Lund, D. 1984. Influence of time, temperature, moisture, ingredients and processing
conditions on starch gelatinization. CRC Crit. Rev. Food Sci. Nut. 20: 249-273.
399
MacBain, R. 1966. Pelleting Animal Feed. Chicago, IL: American Feed
Manufacturing Association.
Macleod, H.M., and K. Marshall. 1977. The determination of density distribution in
ceramic compacts using autoradiography. Pow. Technol. 16(1): 107-122.
Macosko, C.W. 1994. Rheology: Principles, Measurements and Applications.
Poughkeepsie, NY: Wiley/VCH.
Maier, D.E., and F.W. Bakker-Arkema. 1992. The counterflow cooling of feed pellets.
J. Agric. Eng. Res. 53: 305-319.
Maier, D.E., and J. Gardecki. 1992. Feed mash conditioning field case studies. ASAE
Paper No. 92-1541. St. Joseph, Mich.: AS ABE.
Maier, D.E., R.L. Kelley, and F.W. Bakker-Arkema. 1992. In-line, chilled air pellet
cooling. FeedMgmt. 43(1): 28-32.
Major, R. 1984. The pneumatic method. FeedMgmt. 35(6): 20-26.
Malamataris, S., S.B. Baie, and N. Pilpel. 1984. Plasto-elasticity and tableting of
paracetamol, Avicel and other powders. J. Pharm. Pharmacol. 36: 616-617.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2002a. Grinding performance and physical
properties of selected biomass. ASAE Paper No. 026175. St. Joseph, Mich.:
ASABE.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2002b. Compaction behavior of some
biomass grinds. AIC Paper No. 02-305. Saskatoon, Saskatchewan: AIC 2002
Meeting, CSAE/SCGR Program.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2003a. An overview of compaction of
biomass grinds. Powder Handling & Processing 15(3): 160-168.
Mani, S., M. Roberge, L.G. Tabil, Jr., and S. Sokhansanj. 2003b. Modeling of
densification of biomass grinds using discrete element method by PFC D.
CSAE/SCGR 2003 Meeting Paper No. 03-207. Montreal, Quebec, CA:
CSAE/SCGR 2003 Meeting.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004a. Evaluation of compaction equations
applied to four biomass species. Can. Biosys. Eng. 46(3): 3.55-3.61.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2004b. Grinding performance and physical
properties of wheat and barley straws, corn stover and switchgrass. Biomass
andBioenergy 27(4): 339-352.
Mani, S., S. Sokhansanj, and X. Bi. 2005a. Modeling of forage drying in single and
triple pass rotary drum dryers. ASAE Paper No. 056082. St. Joseph, Mich.:
ASABE.
Mani, S., L.G. Tabil, and S. Sokhansanj. 2006a. Effects of compressive force, particle
size and moisture content on mechanical properties of biomass pellets from
grasses. Biomass and Bioenergy 30(7): 648-654.
Mani, S., S. Sokhansanj, X. Bi, and A. Turhollow. 2006b. Economics of producing
fuel pellets from biomass. Applied Eng. in Agric. 22(3): 421-426.
Mani, S., S. Sokhansanj, X. Bi, and L.G. Tabil. 2005b. Modeling of biomass drying
and densification processes. ASAE Paper No. 056144. St. Joseph, Mich.:
ASABE.
400
Marrero, T.R. 1999. Theory and application of binders: an update. Proc. of the
Institute for Briquetting and Agglomeration (IBA), Vol. 26. pp. 103-109.
Marshall, P.V., P. York, and J.Q. Maclaine. 1993. An investigation of the effect of the
punch velocity on the compaction properties of ibuprofen. Pow. Technol.
74(2): 171-177.
McEllhiney, R.R. 1987. What's new in pelleting. Feed Mgmt. 38(2): 32-34.
McEllhiney, R.R. 1988. Mill management feedback. Feed Mgmt. 39(6): 37-39.
McLaughlin, S.B., and L.A. Kszos. 2005a. Development of switchgrass (Panicum
virgatum) as a bioenergy feedstock in the United States. Biomass and
Bioenergy 28(6): 515-535.
McMullen, J., O.O. Fasina, C.W. Wood, and Y. Feng. 2005b. Storage and handling
characteristics of pellets from poultry litter. Applied Eng. inAgric. 21(4): 645-
651.
Michrafy, A., D. Ringenbacher, and P. Tschoreloff. 2002. Modelling the compaction
behavior of powders: application to pharmaceutical powders. Pow. Technol.
127(3): 257-266.
Miles, T.R. 1984. Biomass preparation for thermochemical conversion. In
Thermochemical Processing of Biomass, 69-90. A.V. Bridgwater, ed. London:
Butterworths.
Mina-Boac, J., R.G. Maghirang, and M.E. Casada. 2006. Durability and breakage of
feed pellets during repeated elevator handling. ASABE Paper No. 066044. St.
Joseph, Mich.: ASABE.
Mittal, B., and V.M. Puri. 1999. Correlations between powder deposition methods and
green compact quality: Part II Compact quality and correlations. Particulate
Sci. and Technol. 17(4): 301-315.
Mittal, B., and V.M. Puri. 2002. Determination of visco-elastoplastic properties of a
powder using cubical triaxial tester. ASAE Meeting Paper No. 024035. St.
Joseph, Mich.: ASABE.
Mittal, B., and V.M. Puri. 2003. An elasto-viscoplastic constitutive model
incorporating pore air compressibility during powder compression process.
Particulate Sci. and Technol. 21(2): 131-155.
Mohsenin, N.N. 1986. Physical Properties of Plant and Animal Materials. New
York, NY: Gordon and Breach Science Publishers.
Mohsenin, N., and J. Zaske. 1976. Stress relaxation and energy requirements in
compaction of unconsolidated materials. J. Agric. Eng. Res. 21(2): 193-205.
Moore, J. 1965. Processing applications for roll-type briquetting-compacting
machines. In Proc. 9' Biennial Briquetting Conf, 2-16. Denver, CO:
International Briquetting Association.
Moran, E.T., Jr. 1989. Effect of pellet quality on the performance of meat birds. In
Recent Advances in Animal Nutrition, 87-108. W. Haresign and D.J.A. Cole,
eds. London, England: Butterworths.

401
Morey, R.V., and D.R. Thimsen. 1980. Two-stage combustion to provide heat for
drying corn. ASAE Winter Meeting Paper No. 80-3507. St. Joseph, Mich.:
ASABE.
Morey, R.V., D.G. Tiffany, and D.L. Hatfield. 2006a. Biomass for electricity and
process heat at ethanol plants. Applied Eng. inAgric. 22(5): 723-728.
Morey, R.V., D.L. Hatfield, R. Sears, and D.G. Tiffany. 2006b. Characterization of
feed streams and emissions from biomass gasification/combustion at fuel
ethanol plants. ASABE Paper No. 064180. St. Joseph, Mich.: ASABE.
Moritz, J.S., K.J. Wilson, K.R. Cramer, R.S. Beyer, L.J. McKinney, W.B. Cavalcanti,
and X. Mo. 2002. Effect of formulation density, moisture, and surfactant on
feed manufacturing, pellet quality, and broiler performance. J. Appl Poult.
Res. 11: 155-163.
Muck, R.E., and K.J. Shinners. 2006. Effect of inoculants on the ensiling of corn
stover. ASABE Paper No. 061013. St. Joseph, Mich.: ASABE.
Mukunda, A., K.E. Ileleji, and H. Wan. 2006. Simulation of corn stover logistics from
on-farm storage to an ethanol plant. ASABE Paper No. 066177. St. Joseph,
Mich.: ASABE.
Mufioz, G., and P. Herrera. 2002. Multidimensional modeling of agricultural fibrous
materials in densification: compression stage. ASAE Paper No. 023151. St.
Joseph, Mich.: ASABE.
Munoz-Hernandez, G., J. Dominguez-Dominguez, and O. Alvarado-Mancilla. 2006.
An easy laboratory method for optimizing the parameters for the mechanical
densification process: an evaluation with an extruder. Agricultural Engineering
International: the CIGR Ejournal, Manuscript PM 06 015. Vol. VIII. July,
2006.
Naylor, J.L., and R. Smith. 1981. Roller mills vs hammermills on corn. ASAE Paper
No. 81-3028. St. Joseph, Mich.: ASABE.
O'Dogherty, M.J. 1989. A review of the mechanical behavior of straw when
compressed to high densities. J. Agric. Eng. Res. 44: 241-265.
O'Dogherty, M.J., and J.A. Wheeler. 1984. Compression of straw to high densities in
closed cylindrical dies. J. Agric. Eng. Res. 29(1): 61-72.
Obernberger, I., and G. Thek. 2004. Physical characterisation and chemical
composition of densified biomass fuels with regard to their combustion
behavior. Biomass and Bioenergy 27(6): 653-669.
Olorunnisola, A. 2007. Production of fuel briquettes from waste paper and coconut
husk admixtures. Agricultural Engineering International: the CIGR Ejournal,
Manuscript EE 06 006. Vol. IX. February, 2007.
Olsson, M., J. Kjallstrand, and G. Petersson. 2003. Specific chimney emissions and
biofuel characteristics of softwood pellets for residential heating in Sweden.
Biomass and Bioenergy 24(1): 51-57.
ONORM M 7135. 2000. Compressed wood or compressed bark in natural state -
pellets and briquettes, requirements and test specifications. Vienna, Austria:
OsterreichischesNormungsinstitut.

402
Opoku, A., M.D. Shaw, L.G. Tabil, D. Pulkinen, and B.J. Crerar. 2007. Effect of
whole canola inclusion on physical quality characteristics of feed pellets
processed from reground alfalfa, barley, and canola grains. App. Eng. in Agric.
23(1): 77-82.
ORNL. 2005. Biofuels from Switchgrass: Greener Energy Pastures. Oak Ridge, TN:
Oak Ridge National Laboratory (ORNL). Available at:
http://bioenergy.ornl.gov/papers/misc/switgrs.html. Accessed 18 June 2005.
Orth, H.W., and R. Lowe. 1977. Influence of temperature on wafering in a continuous
extrusion process. J. Agric. Eng. Res. 22(3): 283-289.
Panelli, R., and F.A. Filho. 2001. A study of a new phenomenological compacting
equation. Pow. Technol. 114(1-3): 255-261.
Parker, B.F., G.M. White, M.R. Lindley, R.S. Gates, M. Collins, S. Lowry, and T.C.
Bridges. 1992. Forced-air drying of baled alfalfa hay. Trans. ASAE 35(2): 607-
615.
Patil, R.T., S. Sokhansanj, E.A. Arinze, and G.J. Schoenau. 1993. Methods of
expediting drying rates of chopped alfalfa. Trans. ASAE 36(6): 1799-1803.
Payne, F.A., I.J. Ross, H.E. Hamilton, and J.D. Fox. 1973. Short time, high
temperature extrusion of chicken excreta. ASAE Paper No. 72-450. St. Joseph,
Mich.: ASAE.
Payne, J.D. 1978. Improving quality of pelleted feeds. Milling Feed and Fertilizer,
161(5): 34-41.
Payne, J.D. 2004. Predicting pellet quality and production efficiency. World Grain
August 2004: 68-70.
Payne, J.D. 2006. Troubleshooting the pelleting process. Feed Technology Technical
Report Series, American Soybean Association International Marketing
Southeast Asia, Singapore, pp. 17-23.
Peisker, M. 1992. Improving feed quality by expansion. Int. Milling Flour Feed March
1992: 15-20.
Peleg, K. 1983. A rheological model of nonlinear viscoplastic solids. J. Rheology
27(5): 411-431.
Peleg, M., and E.B. Bagley. 1983. Physical Properties of Foods. Westport,
Connecticut: AVI Publishing Company, Inc.
Pfost, H.B. 1964. The effect of lignin binders, die thickness and temperature on the
pelleting process. Feedstuffs 36(22): 20, 54.
Pfost, H.B., and L.R. Young. 1973. Effect of colloidal binder and other factors on
pelleting. Feedstuffs 45(49): 21-22.
Pickard, G.E., W.M. Roll, and J.H. Ramser. 1961. Fundamentals of hay wafering.
Trans. ASAE 4(1): 65-68.
Pietsch, W. 1991. Size Enlargement by Agglomeration. New York, NY: John Wiley &
Sons.
Pietsch, W. 1997. Granulate dry particulate solids by compaction and retain key
powder particle properties. Chemical Eng. Progress April 1997: 24-46.

403
Pietsch, W. 2002. Agglomeration Processes - Phenomena, Technologies, Equipment.
Weinheim: Wiley-VCH.
Plue, P.S., and W.K. Bilanski. 1990. On-farm drying of large round bales. Applied
Eng. inAgric. 6(4): 418-421.
Pordesimo, L.O., B.R. Hames, S. Sokhansanj, and W.C. Edens. 2005. Variation in
corn stover composition and energy content with crop maturity. Biomass and
Bioenergy 28(4): 366-374.
Raghavan, J.K., and H.N. Conkle. 1991. Physical characteristic measurements for
reconstituted coal pellets. Proc. of the Institute for Briquetting and
Agglomeration (IBA), Vol. 22. pp. 85-96.
Rambali, B., L. Baert, E. Jans, and D.L. Massart. 2001. Influence of the roll compactor
parameter settings and the compression pressure on the buccal bio-adhesive
tablet properties. Int. J. Pharm. 220(1-2): 129-140.
Reece, F.N. 1966. Temperature, pressure, and time relationships in forming dense hay
wafers. Trans. ASAE 9(6): 749-751.
Reece, F.N., B.D. Lott, and J.W. Deaton. 1985. The effects of hammer mill screen size
on ground corn particle size, pellet durability, and broiler performance. Poultry
Sci. 65: 1257-1261.
Reed, T.B., G. Trezek, and L. Diaz. 1980. Biomass densification energy requirements.
In Thermochemical Conversion of Solid Wastes and Biomass, 169-178. J.L.
Jones, S.B. Radding, S. Takaoka, A.G. Buekens, M.Hiraoka, and R. Overend,
eds. Washington, DC: ASC Symposium Series 130, American Chemical
Society.
Rehkugler, G.E., and W.F. Buchele. 1969. Biomechanics of forage wafering. Trans.
ASAE 12(1): 1-8.
Ren, S. 1992. Thermo-hygro rheological behavior of materials used in the
manufacture of wood-based composites. Ph.D. diss. Corvallis, OR: Oregon
State University.
Richards, S.R. 1990. Physical testing of fuel briquettes. Fuel Process. Tech. 25(2): 89-
100.
Richardson, W., and E.J. Day. 1976. Effect of varying levels of added fat in broiler
diets on pellet quality. Feedstuffs 48(20): 24.
Rippie, E.G., and W. Danielson. 1981. Viscoelastic stress/strain behavior of
pharmaceutical tablets: analysis during unloading and postcompression
periods. J. Pharm. Sci. 70(5): 476-482.
Robohm, K.F. 1991. Erfahrungen zum Einsatz des Druckkonditioneurs in Verbindung
mit Futtermittelpressen. Int. ZDS-Fachtagung SIS-11, Extrusions-Tagung '91,
23-25 September, Solingen, Germany, pp: 1-10.
Robohm, K.F. 1992. Adjustable roll gap: benefits to energy demand, throughput and
pellet durability. Feed Int. 13(3): 30-35.
Robohm, K.F., and J. Apelt. 1989a. Die automatische Spaltweitenverstellung. Die
Muhle+ Mischfuttertechnik 126: 271-275.

404
Robohm, K.F., and J. Apelt. 1989b. Verhoging van de flexibiliteit door gebruik van
het perssysteem met voorverdichting. De Molenaar 92: 615-625.
Rost, F.W.D. 1995. Fluorescence Microscopy. Vol. II. New York, NY: Cambridge
University Press.
Rotz, C.A., and Y. Chen. 1985. Alfalfa drying model for the field environment. Trans.
ASAE 28(5): 1686-1691.
Rumpf, H. 1962. The strength of granules and agglomerates. In Agglomeration, W.A.
Knepper, ed 379-418. New York, NY: John Wiley and Sons.
Sah, P., B. Singh, and U. Agrawal. 1980. Compaction behavior of straw. J. Agric.
Eng.-India 18(1): 89-96.
Sah, P.C., B.P.N. Singh, U. Chandra, and B.N. Sachchan. 1977. Grinding of wheat
straw in a hammer mill. J. Agric. Eng. (India) XIV (3): 108-112.
Salmon, R.E. 1985. Effects of pelleting, added sodium bentonite and fat in a wheat-
based diet on performance and carcass characteristics of small turkeys. Animal
FeedSci. Tech. 12(3): 223-232.
Samson, R., P. Duxbury, M. Drisdelle, and C. Lapointe. 2000. Assessment of
Pelletized Biofuels. A Report. Ste. Anne de Bellevue, QC: Resource Efficient
Agricultural Production (REAP-Canada, www.reap-canada.com).
Sanderson, M.A., R.P. Egg, and A.E. Wiselogel. 1997. Biomass losses during harvest
and storage of switchgrass. Biomass and Bioenergy 12(2): 107-114.
Savoie, P., and S. Descoteaux. 2004. Artificial drying of corn stover in mid-size bales.
ASAE Paper No. 048009. St. Joseph, Mich.: ASABE.
Savoie, P., S. Beauregard, and C. Lague. 1990. Rapid hay drying with super hay
conditioning. CSAE Paper No. 90-502. Saskatoon: CSAE.
Schell, D.J., and C. Harwood. 1994. Milling of lignocellulosic biomass: results of
pilot-scale testing. Applied Biochemistry and Biotechnology 45/46: 159-168.
Schofield, A.N., and C.P. Wroth. 1968. Critical State Solid Mechanics. New York,
NY: McGraw Hill.
Shen, K.C. 1987. Development of a waterproof densified solid fuel pellet from
forestry residues. In Vroc. of the Sixth Canadian Bioenergy R&D Seminar,
209 - 213. London, UK: Elsevier Applied Science.
Shinners, K.J., and B.N. Binversie. 2004. Harvest and storage of wet corn stover
biomass. ASAE Paper No. 041159. St. Joseph, Mich.: ASABE.
Shinners, K.J., B.N. Binversie, and P. Savoie. 2003. Harvest and storage of wet and
dry corn stover as a biomass feedstock. ASAE Paper No. 036088. St. Joseph,
MI: ASABE.
Shinners, K.J., D.S. Hoffman, G.C. Boettcher, and J.T. Munk. 2007. Harvesting rate,
power requirements and fuel use for single-pass harvesting of corn stover.
ASABE Paper No. 071018. St. Joseph, Mich.: ASABE.
Shinners, K.J., G.C. Boettcher, J.T. Munk, M.F. Digman, R.E. Muck, and P.J.
Weimer. 2006a. Single-pass, split-stream of corn grain and stover:
characteristic performance of three harvester configurations. ASABE Paper
No. 061015. St. Joseph, Mich.: ASABE.

405
Shinners, K.J., G.C. Boettcher, R.E. Muck, P.J. Wiemer, and M.D. Casler. 2006b.
Drying, harvesting and storage characteristics of perennial grasses as biomass
feedstocks. AS ABE Paper No. 061012. St. Joseph, Mich.: AS ABE.
Shinners, K.J., R.G. Koegel, and R.J. Straub. 1989. Rapid field drying of forages
utilizing shredded mat technology. In Proc. Eleventh International Congress
on Agricultural Engineering, Dublin, 4-8 September 1989, 2063-2070. V.A.
Dodd and P.M. Grace, eds.
Shrivastava, M., P. Shrivastava, and K.K. Khankari. 1989. Densification
characteristics of rice husk under cold and hot compression. In Agricultural
Engineering: Proceedings of the 11th International Congress on Agricultural
Engineering, Dublin, 4-8 September 1989, 2441-2443. V.A. Dodd and P. M.
Grace, eds. Rotterdam: A.A. Balkema Pub.
Siegerink, A. 2007. Molasses as a binder in agglomeration processes. Proceedings of
the Institute for Briquetting and Agglomeration 30th Biennial Conference.
Manitowish Waters, WI: IBA.
Singh, A., and Y. Singh. 1982. Briquetting of paddy straw. Agric. Meek in Asia,
Africa and Latin America 13(4): 42-44.
Singh, D., and M.M. Kashyap. 1985. Mechanical and combustion characteristics of
paddy husk briquettes. Agric. Wastes 13(3): 189-196.
Sinka, I.C., J.C. Cunningham, and A. Zavaliangos. 2003. The effect of wall friction in
the compaction of pharmaceutical tablets with curved faces: a validation study
of the Drucker-Prager Cap model. Pow. Technol. 133(1-3): 33-43.
Sinka, I.C., J.C. Cunningham, and A. Zavaliangos. 2004. Analysis of tablet
compaction. II. Finite element analysis of density distributions in convex
tablets. J. Pharm. Sci. 93(8): 2040-2053.
Sitkei, G. 1986. Mechanics of Agricultural Materials. New York, NY: Elsevier.
Skoch, E.R., K.C. Behnke, C.W. Deyoe, and S.F. Binder. 1981. The effect of steam-
conditioning rate on the pelleting process. Animal Feed Sci. Tech. 6(1): 83-90.
Smith, I.E., S.D. Probert, R.E. Stokes, and R.J. Hansford. 1977. The briquetting of
wheat straw. J. Agric. Eng. Res. 22(2): 105-111.
Sokhansanj, S., A. Kumar, and A.F. Turhollow. 2006b. Development and
implementation of integrated biomass supply analysis and logistics model
(IBSAL). Biomass andBioenergy 30(10): 838-847.
Sokhansanj, S., and A. Turhollow. 2002. Baseline cost for corn stover collection.
Applied Eng. in Agric. 18(5): 38-43.
Sokhansanj, S., and A.F. Turhollow. 2004. Biomass densification - cubing operations
and costs for corn stover. Applied Eng. in Agric. 20(4): 495-499.
Sokhansanj, S., A. Turhollow, J. Cushman, and J. Cundiff. 2002. Engineering aspects
of collecting corn stover for bioenergy. Biomass and Bioenergy 23(5): 347-
355.
Sokhansanj, S., A. Turhollow, S. Tagore, and S. Mani. 2006a. Integrating biomass
feedstock with an existing grain handling system for biofuels. ASABE Paper
No. 066189. St. Joseph, Mich.: ASABE.

406
Sokhansanj, S., and H.C. Wood. 1991. Engineering aspects of forage processing for
pellets, cubes, dense chops and bales. Advances in Feed Technology 5: 6-23.
Sokhansanj, S., J. Cushman, and L. Wright. 2003. Collection and delivery of feedstock
biomass for fuel and power production. Agricultural Engineering
International: the CIGR Journal of Scientific Research and Development.
Invited Overview Paper. Vol. V. February 2003.
Sokhansanj, S., L. Tabil, Jr., and W. Yang. 1999. Characteristics of plant tissue to
form pellets. Powder Handling & Process. 11(2): 149-159.
Sokhansanj, S., R.T. Patil, G. Ahmadnia, O.O. Fasina, and J. Irudayaraj. 1991.
Procedures for evaluating durability and density of forage cubes and pellets.
CASE Paper No. 91-402. Canadian Society of Agricultural Engineering-
Agricultural Institute of Canada Annual Conference, July 29-31, 1991 -
Fredericton, New Brunswick.
Srivastava, A.C., W.K. Bilanski, and V.A. Graham. 1981. Feasibility of producing
large-size hay wafers. Can. Agric. Eng. 23(2): 109-112.
Stahl, M , K. Granstrom, J. Berghel, and R. Renstorm. 2004. Industrial processes for
biomass for drying and their effects on the quality properties of wood pellets.
Biomass and Bioenergy 27(6): 621-628.
Stark, C.R. 1994. Pellet quality. I. Pellet quality and its effect on swine performance.
II. Functional characteristics of ingredients in the formation of quality pellets.
Ph.D. diss. Manhattan, KS: Kansas State University.
Stevens, C.A. 1987. Starch gelatinization and the influence of particle size, steam
pressure and die speed on the pelleting process. Ph.D. diss. Manhattan, KS:
Kansas State University.
Svenningson, P.J., and R. Hosier. 1987. Biomass briquettes in the Dominican
Republic. Part II: technical analyses. Biomass 13(4): 275-291.
Tabil, L.G., Jr. 1996. Binding and pelleting characteristics of alfalfa. Ph.D. diss.
Saskatoon, Saskatchewan, CA: University of Saskatchewan, Department of
Agricultural and Bioresource Engineering.
Tabil, L., Jr., and S. Sokhansanj. 1996a. Process conditions affecting the physical
quality of alfalfa pellets. Applied Eng. in Agric. 12(3): 345-350.
Tabil, L.G., and S. Sokhansanj. 1996b. Compression and compaction behavior of
alfalfa grinds -1. Compression behavior. Powder Handling and Processing.
The Int. J. Storing, Handling, and Processing Powder 8(1): 17-23.
Tabil, L.G., Jr., S. Sokhansanj, and R.T. Tyler. 1997. Performance of different binders
during alfalfa pelleting. Can. Agric. Eng. 39(1): 17-23.
Takeuchi, H., S. Nagira, H. Yamamoto, and Y. Kawashima. 2004. Die wall pressure
measurement for evaluation of compaction property of pharmaceutical
materials. Int. J. Pharm. 274(1-2): 131-138.
Temmerman, M., F. Rabier, P.D. Jensen, H. Hartmann, and T. Bohm. 2006.
Comparative study of durability test methods for pellets and briquettes.
Biomass and Bioenergy 30(11): 964-972.

407
Thek, G., and I. Obernberger. 2004. Wood pellet production costs under Austrian and
in comparison to Swedish framework conditions. Biomass and Bioenergy
27(6): 671-693.
Thoemen, H., C.R. Haselein, and P.E. Humphrey. 2006. Modeling the physical
processes relevant during hot pressing of wood-based composites. Part II.
Rheology. Holz als Roh- und Werkstoff64(2): 125-133.
Thomas, M , and A.F.B. van der Poel. 1996. Physical quality of pelleted animal feed.
1. Criteria for pellet quality. Animal Feed Sci. Tech. 61(1-4): 89-112.
Thomas, M., D.J. van Zuilichem, and A.F.B. van der Poel. 1997. Physical quality of
pelleted animal feed. 2. Contribution of processes and its conditions. Animal
Feed Sci. Tech. 64(2-4): 173-192.
Thomas, M., T. van Vliet, and A.F.B. van der Poel. 1998. Physical quality of pelleted
animal feed 3. Contribution of feedstuff components. Animal Feed Sci. Tech.
70(1-2): 59-78.
Tripodi, M.A., V.M. Puri, H.B. Manbeck, and G.L. Messing. 1992. Constitutive
models for cohesive particulate materials. J. Agric. Eng. Res. 53(1): 1-21.
Turner, R. 1995. Bottomline in feed processing: achieving optimum pellet quality.
FeedMgmt. 46(12): 30-33.
Tye, C.K., C.C. Sun, and G.E. Amidon. 2005. Evaluation of the effects of tableting
speed on the relationships between compaction pressure, tablet tensile strength,
and tablet solid fraction. J. Pharm. Sci. 94(3): 465-472.
USDA. 2002. Crop production - annual summary (PCP-BB). Washington, DC:
National Agricultural Statistical Service, U.S. Department of Agriculture
(USDA).
van der Poel, A.F.B., H.M.P. Fransen, and M.W. Bosch. 1997. Effect of expander
conditioning and/or pelleting of a diet containing tapioca, pea and soybean
meal on the total tract digestibility in growing pigs. Animal Feed Sci. Tech.
66(1-4): 289-295.
Varadharaju, N., and L. Gothandapani. 1998. Design and development of equipment
for pelleting decomposed coir pith. Agric. Mech. in Asia, Africa and Latin
America 29(2): 33, 34, 38.
Vest, L. 1993. Southeastern survey: factors which influence pellet production and
quality. FeedMgmt. 44(5): 60-68.
Von Eggelkraut-Gottanka, S.G., S.A. Abed, W. Miiller, and P.C. Schmidt. 2002.
Roller compaction and tabletting of St. John's wort plant dry extract using a
gap width and force controlled roller compactor. I. Granulation and tabletting
of eight different extract batches. Pharm. Dev. Tech. 7(4): 433-445.
Waelti, H., and J.B. Dobie .1973. Cubability of rice straw as affected by various
binders. Trans. ASAE 16(2): 380-383.
Walsh, M.E., R.L. Perlack, A. Turhollow, D.T. Ugarte, D.A. Becker, R.L. Graham,
S.E. Slinsky, and D.E. Ray. 2005. Biomass Feedstock Availability in the
United States: 1999 State Level Analysis. Oak Ridge, TN: Oak Ridge National

408
Laboratory. Available at: http://bioenergy.ornl.gov/resourcedata/. Accessed 18
June 2005.
Wamukonya, L., and B. Jenkins. 1995. Durability and relaxation of sawdust and
wheat-straw briquettes as possible fuels for Kenya. Biomass and Bioenergy
8(3): 175-179.
Weaver, J.W. Jr., CD. Grinnells, and R.L. Lovvorn. 1947. Drying baled hay with
forced air. Agri. Eng. 28 (7): 301-304.
Wilen, C , P.Stahlberg, K. Siplila, and J. Ahokas. 1987. Pelletization and combustion
of straw. In Energy from Biomass and Wastes X, 469-484. D.L. Klass, ed.
London: Elsevier Applied Science Publishers.
Winowiski, T. 1988. Wheat and pellet quality. FeedMgmt. 39(9): 58-64.
Winowiski, T. 1998. Examining a new concept in measuring pellet quality: which test
is best? FeedMgmt. 49(1): 23-26.
Winowiski, T.S. 2006. Factors that affect pellet quality and trouble-shooting the
pelleting process. Technical Bulletin on Feed Technology, Vol. FT23-1995.
Singapore: American Soybean Association. Available at: www.asasea.com.
Accessed 8 August 2006.
Winowiski, T. 1985. Optimizing pelleting temperature. FeedMgmt. 36(7): 28-33.
Wiselogel, A.E., F.A. Agblevor, D.K. Johnson, S. Deutch, J.A. Fennell, and M.A.
Sanderson. 1996. Compositional changes during storage of large round
switchgrass bales. Bioresource Technol. 56(1): 103-109.
Wood, J.F. 1987. The functional properties of feed raw materials and the effect on the
production and quality of feed pellets. Animal Feed Sci. Tech. 18(1):1-17.
Wu, C.Y., O.M. Ruddy, A.C. Bentham, B.C. Hancock, S.M. Best, and J.A. Elliot.
2005. Modelling the mechanical behavior of pharmaceutical powders during
compaction. Pow. Technol. 152(1-3): 107-117.
Yaman, S., M. Sahan, H. Haykiri-Acma, K. Sesen, and S. Kilciikbayrak. 2001. Fuel
briquettes from biomass-lignite blends. Fuel Process. Tech. 72(1): 1-8.
Young, L.R., H.B. Pfost, and A.M. Feyerherm. 1963. Mechanical durability of feed
pellets. Trans. ASAE 6(2): 145-147, 150.
Yu, M., A.R. Womac, and L.O. Pordesimo. 2003. Review of biomass size reduction
technology. ASAE Paper No. 036077. St. Joseph, Mich.: ASABE.
Zavaliangos, A. 2002. Constitutive models for the simulation of P/M processes. The
Int. J. Powder Metallurgy 38(2): 27-39.

409

Anda mungkin juga menyukai