Anda di halaman 1dari 103

SPRINGER BRIEFS IN

APPLIED SCIENCES AND TECHNOLOGY

AngeloFreni BelalDawoud
LucioBonaccorsi StefanieChmielewski
AndreaFrazzica LuigiCalabrese
GiovanniRestuccia

Characterization
of Zeolite-Based
Coatings for
Adsorption
Heat Pumps

123
SpringerBriefs in Applied Sciences
and Technology
More information about this series at http://www.springer.com/series/8884
Angelo Freni Belal Dawoud

Lucio Bonaccorsi Stefanie Chmielewski


Andrea Frazzica Luigi Calabrese


Giovanni Restuccia

Characterization
of Zeolite-Based Coatings
for Adsorption Heat Pumps

123
Angelo Freni Andrea Frazzica
Consiglio Nazionale delle Ricerche, Istituto Consiglio Nazionale delle Ricerche, Istituto
di Tecnologie Avanzate per lEnergia di Tecnologie Avanzate per lEnergia
Nicola Giordano Nicola Giordano
Messina Messina
Italy Italy

Belal Dawoud Luigi Calabrese


East Bavarian Technical University Department of Electronic Engineering,
(OTH-Regensburg) Industrial Chemistry and Engineering
Regensburg University of Messina, Contrada di Dio
Germany Messina
Italy
Lucio Bonaccorsi
Department of Electronic Engineering, Giovanni Restuccia
Industrial Chemistry and Engineering Consiglio Nazionale delle Ricerche, Istituto
University of Messina, Contrada di Dio di Tecnologie Avanzate per lEnergia
Messina Nicola Giordano
Italy Messina
Italy
Stefanie Chmielewski
Viessmann Werke Allendorf GmbH
Allendorf (Eder)
Germany

ISSN 2191-530X ISSN 2191-5318 (electronic)


SpringerBriefs in Applied Sciences and Technology
ISBN 978-3-319-09326-0 ISBN 978-3-319-09327-7 (eBook)
DOI 10.1007/978-3-319-09327-7

Library of Congress Control Number: 2015934678

Springer Cham Heidelberg New York Dordrecht London


The Author(s) 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media


(www.springer.com)
Contents

1 Basics of Adsorption Heat Pump Processes . . . . . . . . . . . . . . . . . . 1


1.1 Thermodynamics of Thermally Driven Heat Pumps . . . . . . . . . . 2
1.2 Working Pairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2.1 Silica Gels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.2.2 Composite Materials . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.3 Activated Carbons . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2.4 Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.2.5 New Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3 The Ideal Intermittent Adsorption Heat Pump Cycle. . . . . . . . . . 15
1.4 Energetic Evaluation of the Real Intermittent Adsorption
Heat Pump Cycle. . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 21
1.5 Basic Features and Design Requirements of Intermittent
Adsorption Heat Pumps . . . . . . . . . . . . . . . . . . . . . . . ...... 25
1.6 Short Overview on Advanced Cycles . . . . . . . . . . . . . . ...... 28
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...... 30

2 Adsorption Heat Exchangers . . . . . . . . . . . . . . . . . . . . . ....... 35


2.1 Ideal Adsorber Design Requirements . . . . . . . . . . . . . ....... 35
2.2 Overview on Adsorber Developments for Adsorption
Heat Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... 36
2.3 Introduction to the Concept of Coated Adsorbers . . . . . ....... 42
2.4 Methodologies for the Experimental Characterization
of Zeolites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... 44
2.5 Experimental Methodologies to Verify the Performance
of Adsorbers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... 47
2.5.1 Small Scale Large Pressure Jump (LPJ) . . . . . . ....... 47
2.5.2 Small Scale Large Temperature Jump (LTJ) . . . ....... 48
2.5.3 Full Scale Experimental Setup for Cycling
Characterization. . . . . . . . . . . . . . . . . . . . . . . ....... 50
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....... 51

v
vi Contents

3 Hydrothermal Stability of Adsorbent Coatings . . . . . . ......... 55


3.1 Hydrothermal Aging of Zeolites . . . . . . . . . . . . . . ......... 55
3.2 Real-Time or Long-Term Heat Pump Aging Cycling
Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2.1 Description of the Set-up . . . . . . . . . . . . . . . . . . . . . . . 59
3.2.2 Experimental Testing . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3 Accelerated Aging Cycling Procedure. . . . . . . . . . . . . . . . . . . . 65
3.3.1 Description of the Set-up . . . . . . . . . . . . . . . . . . . . . . . 65
3.3.2 Estimation of the Operating Conditions
for AQSOA- FAMZ02 Zeolite . . . . . . . . . . . . . . . . . . . 67
3.3.3 Experimental Testing . . . . . . . . . . . . . . . . . . . . . . . . . . 69
3.4 Very-Quick Aging Cycling Procedure . . . . . . . . . . . . . . . . . . . 71
3.4.1 Description of the Set-up . . . . . . . . . . . . . . . . . . . . . . . 71
3.4.2 Experimental Testing . . . . . . . . . . . . . . . . . . . . . . . . . . 72
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

4 Mechanical Stability of Adsorbent Coatings. . . . . . . . . . . . . . . . . . 81


4.1 Static Mechanical Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.1.1 Flexural Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.1.2 Shear Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.1.3 Micro-hardness Test. . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.1.4 Scratch Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.1.5 Pull-off Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.1.6 Peel Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.2 Dynamic Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2.1 Impact Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
4.2.2 Vibrational Test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.3 Methods Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.4 Experimental Case History: Adhesion Quality Control
on a Coated Adsorber. . . . . . . . . . . . . . . . . . . . . . . ........ 92
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ........ 95
Introduction

The present book illustrates novel methods for the complete characterization of
advanced coated adsorbers to be used in thermally efcient adsorption heat pumps.
It provides a deep analysis of the different procedures necessary for the evaluation
of the achievable performance of the adsorber as well as for its stability against both
hydrothermal and mechanical stresses.
In Chap. 1 we treat the thermodynamic fundamentals of the basic adsorption heat
pump process. The most common working pairs, as well as the new research trends,
are explained with emphasis on the zeolitewater pair, which is the specic
adsorbent/adsorbate system addressed in this book. We provided basic information
about the structural characteristics of the zeolites (NaA, 13X, Y, SAPO34), which
are commonly employed in adsorption heat pumps. Afterwards, the ideal and real
intermittent adsorption heat pump cycles are precisely described and the corre-
spondent energetic evaluations are developed. These formulations permit to
investigate the influence of design parameters on the performance of an adsorption
heat pump module, focusing on the adsorber design. This study leads, in essence, to
identify the adsorption heat exchanger basic design requirements, which will be
treated in the following chapter. This chapter ends with a short overview of other
more sophisticated adsorption heat pump processes.
In Chap. 2, we initially report the evolution of adsorber concepts over the years,
starting from the earlier congurations employing loose adsorbent grains, to the
modern concept of coated adsorber, which is the focus of this book. Then we
explain the most common coating procedures that, essentially, consist of the
deposition, by means of a binder or by direct crystallization, of the adsorbent
material over a metallic heat exchanger, aiming at the improvement of the heat
transfer efciency. After having introduced the coated adsorber concept, we deal
with the fundamental task of the experimental characterization of coated adsorbers.
Initially, standard methodologies for the experimental characterization of pure
zeolites are given as fundamental background to better understand the more
sophisticated techniques developed to assess the performance and stability of coated
adsorbers. Subsequently, we provide the description of the recent experimental
methodologies for adsorber dynamic efciency evaluation. Two main approaches

vii
viii Introduction

are distinguished considering the scale of the tested adsorber: small scale, which
allows to test small but representative piece of adsorber concepts, and full scale,
where the entire adsorber can be tested under almost-realistic boundary conditions
Chapters 3 and 4 are focused on the crucial subject of the verication of
mechanical and hydrothermal cycling stability of adsorbent coatings. In Chap. 3 we
initially discuss the effects of hydrothermal aging of pure zeolites due to repeated
cyclic adsorption and desorption of water, which can result in reduced adsorption
capacity and adsorption rate. Afterwards, we propose three original methodologies
for hydrothermal aging treatment suitable for adsorbent coatings and pelletized
adsorbers: (a) real-time or long-term heat pump cycling under realistic vacuum
adsorption heat pumping conditions, (b) accelerated aging cycling under wet/dry air
flux, and (c) very quick aging cycling under severe hot air blow/cold liquid water
treatment. Such techniques have been applied on samples of pelletized adsorbent
and adsorbent coatings prepared according to different methods having different
maturity levels: from laboratory scale to pre-commercial pilot scale. For each aging
method, advantages and limits are explained and some results of aging tests are
provided and compared.
In Chap. 4, we address the mechanical characterization of the same adsorbent
coating samples characterized in the previous chapter. In this chapter, the experi-
mental protocols under static and dynamic testing setup were developed, in order to
ensure a broad, comprehensive, and reliable identication of the mechanical per-
formance of the coatings used in the heat exchangers. In particular, the results of
tests carried out on coatings at laboratory-scale level have shown a quite brittle
behavior that influences its mechanical stability especially under dynamic loading.
Instead, the pilot-scale coating, characterized by high hardness, showed good
adhesion with the metallic substrate and very good performance in static and
dynamic testing conditions. This chapter ends with an experimental case history
dealing with the adhesion quality control on a coated adsorber.
Combined results of the testing protocol presented in Chaps. 3 and 4 indicate
that some coating techniques still at laboratory level of development need further
improvement to reach sufciently high lifetime to meet durability requirements of
commercial applications. Good hydrothermal and mechanical stabilities were
exhibited by the pilot-scale coated adsorber heat exchangers. In a large-scale serial
production process, it is a fundamental requirement to guarantee or even to further
improve both mechanical and hydrothermal stabilities of the coated heat exchangers
at moderate costs. Full commercial state of adsorber heat exchangers should be
approved according to a quality assurance test protocol based on the experimental
methods introduced in this book, which could be a subject of the future work.
We hope that this book will give a hand to researchers and developing teams of
adsorption heat pumps for heating and cooling applications, helping to guide and
intensify both research and development activities, most specically for developing
and approving novel adsorbents and novel adsorbent beds.
Introduction ix

Finally, the authors are highly grateful to the German Federal Ministry of
Economy and Technology (BMWi) for partial funding of the research and devel-
opment activities between July 2007 and March 2010 (Frderkennzeichen:
0327435A) as well as between June 2010 and November 2012 (Frderkennzeichen:
0327435B).
Chapter 1
Basics of Adsorption Heat Pump Processes

Nomenclature
16 State points in Fig. 1.9
3-WV Three way valve
AdHX Adsorber heat exchanger
AHS Ambient heat source
DHS Driving heat source
COP Coefcient of performance []
c Specic heat capacity [kJ=(kg K)]
E Exergy [J]
EvHX Evaporator heat exchanger
HN Heating net
hfg Latent heat of vaporization [kJ/kg]
K Dimensionless parameter
m Mass [kg]
P Pump
Q Heat [J]
S Entropy [J/K]
T Thermodynamic temperature (K)
TDHP Thermally driven heat pump
Wt Technical work [J]
x Refrigerant uptake [kg/kg]

Greek Letters
IIV State points in Fig. 1.11
Dhads Average isosteric heat of adsorption [kJ/kg]
Differential operator
Temperature [C]
Efciency []

The Author(s) 2015 1


A. Freni et al., Characterization of Zeolite-based Coatings
for Adsorption Heat Pumps, SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-09327-7_1
2 1 Basics of Adsorption

Subscripts
AdHX Adsorber heat exchanger
AHS Ambient heat source
Ads Adsorbent, adsorption and adsorption end state
Adsref Adsorbentrefrigerant interaction
C Carnot
Con Condenser
cool Precool loss of the evaporator
DHS Driving heat source
Des Desorption & desorption end state
El Electrical
Ev Evaporator
HN Heating net
heat Preheat losses of the condenser
min Minimum value
max Maximum value
PC Pre-cooling
PH Pre-heating
ref Refrigerant
r Return
s Supply

Superscripts
HE Heat engine
HP Heat pump
ic Ideal cycle
rev Reversible

Overbar
Averaged value

1.1 Thermodynamics of Thermally Driven Heat Pumps

A heat pump process is a thermodynamic process having the main target to pump
heat from a heat reservoir at a low temperature (ambient heat source) to a heat sink at
a higher temperature (heating net). According to the second law of thermodynamics,
this target can only be realized if a driving energy is applied. Contrary to the vapor
compression heat pump process, where mechanical work is applied as a driving
energy to run the compressor, thermally driven heat pumps (TDHP) make use of heat
at a higher temperature (driving heat source), compared to the heat sink temperature,
1.1 Thermodynamics of Thermally Driven Heat Pumps 3

Fig. 1.1 Basic schematic of a


thermally driven heat pump

as a driving energy. Figure 1.1 illustrates the very basic schematic of a TDHP along
with the temperature allocation of both heat sources as well as the heat sink (heating
net).
The energy balance on the TDHP depicted in Fig. 1.1 requires that the sum of
the amounts of heat transferred from both driving and ambient heat sources shall be
rejected to the heating net assuming no heat losses taking place to the surroundings.
This can be mathematically formulated as

jQHN j QDHS QAHS . 1:1

The coefcient of performance (COP) of a TDHP process is dened as the ratio


between the magnitude of the heat rejected to the heating net and the driving heat
received from the driving heat source.

jQHN j
COP 1:2
QDHS

Applying the energy balance formulated in Eq. (1.1), the COP of the TDHP can
be expressed as

QDHS QAHS QAHS


COP 1 . 1:3
QDHS QDHS

Figure 1.2 presents the amounts of heat exchanged between both heat sources
and the TDHP as well as between the TDHP and the heating net on a temperature-
entropy-diagram. The exergy (E) of a heat flow (Q) is expressed as the maximum
amount of work, which can be generated out of such heat flow. In case the heat flow
has a higher temperature than the ambient, the maximum amount of work can be
expressed by the Carnot efciency of the heat engine working between the heat
source and the ambient temperatures [1]. As the heat transferred from the ambient
4 1 Basics of Adsorption

Fig. 1.2 Exchanged heat and


exergy amounts between a
reversible TDHP and the
different heat sources and
sinks on a T-S, diagram

heat source takes place at the ambient temperature, it has no potential to produce
work and its exergy amounts, therefore, to zero.
This implies for a reversible TDHP that the exergy of the heat flow from the
driving heat source shall be equal to the exergy of the rejected heat to the heating
net as depicted in Fig. 1.2. The exergies of both amounts of heat exchanged
between the driving heat source and the TDHP as well as between the TDHP and
the heating net can be expressed as

EDHS QDHS  1  TAHS =TDHS


1:4
EHN jQHN j  1  TAHS =THN

where, TAHS, TDHS and THN are the absolute thermodynamic temperatures in degree
Kelvin [K] of the driving and ambient heat sources as well as of the heating net heat
sink, respectively. For a reversible TDHP (no exergy losses) both exergies
expressed in Eq. (1.4) shall be equal. Accordingly, the COP of a reversible TDHP
can be obtained by combining Eqs. 1.1 and 1.4 and expressed as
   
jQHN j 1  TAHS =TDHS TAHS THN
COPrev 1 
TDHP
QDHS 1  TAHS =THN TDHS THN  TAHS 1:5
 THN
gC TTDHS
AHS
 COPrev
HP TAHS .

Equation 1.5 shows that the COP of a reversible TDHP can be expressed as the
product of the Carnot efciency of a heat engine working between the ambient and
the driving heat source temperatures and the Carnot COP of a reversible heat pump
process working between the ambient heat source and the heating net temperatures
[2]. Figure 1.3 shows how the internal process of a thermally driven heat pump does
incorporate rst a heat engine to internally produce the reversible technical work
(Wtrev ), which is then applied to drive the reversible heat pump process between the
ambient heat source and the heating net temperatures, a fact which can be simply
demonstrated by rewriting Eq. (1.5) as follows:
1.1 Thermodynamics of Thermally Driven Heat Pumps 5

Fig. 1.3 The reversible


TDHP process as a
combination between a
reversible heat engine and a
reversible heat pump
processes

jQNH j Wtrev jQHN j  THN


TDHP
COPrev  gC TTDHS  COPrev
WP TAHS . 1:6
QDHS QDHS Wtrev AHS

For an ambient heat source at 0 C, a heating net at 45 C and a driving heat


source at 200 C, the maximum attainable COP of the reversible TDHP can be
estimated by Eq. (1.5) to be equal 2.99.
The relevant values of the amounts of heat and work exchanged in such a
reversible process are illustrated in Fig. 1.4.
The Carnot efciency of the heat engine (HE) is estimated to 0.423 and the
reversible COP of the HP process to 7.07. The net heat extracted out of the ambient
AHS  QAHS ) amounts to 1.99 kWh in order to reject 2.99 kWh of heat
heat source (QHP HE

into the heating net for one kWh heat to be applied by the driving heat source. The
COP of real single stage thermally driven heat pumps can reach typical values
between 1.2 and 1.7 depending among other parameters on the ambient and heating
net temperatures as well as on the applied adsorbent-refrigerant pair.
Ground heat, ground water, ambient air as well as heat out of solar collectors are
the typical ambient heat sources for thermally driven heat pumps. The reference

Fig. 1.4 Exchanged amounts


of heat and work in the
reversible TDHP process
6 1 Basics of Adsorption

Table 1.1 Design supply (s)


Reference heat sink condition Dry bulb
and return (r) temperatures of
temperature
the different heat sink (heating
conditions
net) conditions considered in
the FprEN12309:2014 [3, 4] s [C] r [C]
Low temperature application 35 28
Medium temperature application 45 35
High temperature application 55 43
Very high temperature application 65 48

heating net conditions for heating systems in Europe have been dened in parts 3
and 7 of the European Standard FprEN12309:2014 [3, 4] as given in Table 1.1.
The design supply (to the heating net) and return (back from the heating net to
the heating appliance) temperatures have to be dealt with at the on-site coldest
ambient (heating-design) temperature. For example, the heating-design temperature
for Strasburg as the average climate condition in Europe is set to 10 C, while the
heating-design temperatures of +2 and 22 C are set for both Athens and Helsinki
as the warmer and colder climate conditions, respectively [3, 4]. The driving heat
source temperature depends on both ambient heat source and heating net heat sink
conditions as well as on the applied adsorbent-refrigerant working pair.

1.2 Working Pairs

The working fluid in a vapor compression heat pump is in general a pure substance
or a mixture of pure substances, the so-called refrigerant. Within the single-stage
vapor compression heat pump cycle, the refrigerant exhibits the following four
processes:
1. Liquid to vapor phase change upon receiving the heat of vaporization from the
ambient heat source in the evaporator.
2. The refrigerant vapor is sucked from the evaporator and compressed to the
condenser pressure by a mechanical compressor. This results in a superheated
vapor at high pressure and temperature entering the condenser.
3. De-superheating and condensation of the refrigerant vapor in the condenser
upon giving this useful heat to the heating net.
4. Pressure reduction of the liquid refrigerant in the expansion valve, resulting in
evaporating a certain portion of the refrigerant and, consequently reducing the
refrigerant temperature below the temperature of the ambient heat source. Doing
so, process 1 can follow again and a continuous heat pumping process is
realized.
In a TDHP, the mechanically driven compressor is replaced by a thermally
driven compression process. The basic idea behind is to make use of a working pair
1.2 Working Pairs 7

out of a refrigerant and a suitable sorbent for the refrigerant. Depending on the state
of the sorbent, TDHPs can be classied in two categories; absorption HPs in case of
liquid sorbents (Absorbents) and adsorption HPs in case of solid sorbents
(Adsorbents) are utilized. In this book we deal only with solid sorbents and their
application in thermally driven adsorption heat pumps (TDAdHPs).
The refrigerant vapor produced in the evaporator is adsorbed by the adsorbent.
This adsorption process is an exothermic process, it means that the so called heat of
adsorption (sum of heat of condensation of the refrigerant and heat of binding the
refrigerant into the porous structure of the adsorbent) is to be removed, in order for
the adsorbent to be able to adsorb more refrigerant. This adsorption process is
equivalent to the suction process of the vapor generated in the evaporator by the
mechanical compressor in a vapor compression heat pump. If the adsorbent is
saturated with the refrigerant depending on the working conditions of the TDAdHP,
the adsorption process shall be stopped and the so called desorption process has to
start, which is simply the reverse process of the adsorption. This means that heat is
applied to the adsorbent, resulting in giving more energy to the bonded refrigerant
molecules. If the heat energy transferred from the driving heat source to the
refrigerant molecules inside the adsorbent equals the sum of energies required to
overcome the binding energy between the refrigerant and adsorbent as well as to
evaporate the refrigerant, the refrigerant molecules leave the adsorbent as vapor at
higher temperature and pressure (desorption process). This process is equivalent to
the compression process of the mechanically operated compressor.
It is widely recognized that one of the most important tasks in designing an
AdHP is the choice of the suitable adsorbent/refrigerant pair [5]. The criteria for
choosing a suitable refrigerant, when considering the refrigerant alone, are large
enthalpy of vaporization, reasonable working pressure, low freezing temperature,
small molecular dimensions, chemical stability, low toxicity and low cost. The
suitability of different refrigerants was compared by Critoph [5] and updated later
on by Critoph [6] as well as Srivastava and Eames [7]. In general, the most widely
used refrigerants are water, ammonia and methanol with a recent trend to make use
of Ethanol [8, 9]. Water was found to be the most favorable refrigerant, although it
was acknowledged that its application is restricted to temperatures over 0 C when
used in refrigeration, due to the formation of ice in the evaporator, unless this is a
desired result.

1.2.1 Silica Gels

The most common adsorbents include silica gel, composite salt in porous matrix
materials, activated carbon and zeolite [10]. Pure Silica, SiO2, is naturally a
chemically inactive non-polar material like quartz but when it has a hydroxyl
functional group (silanol group), the surface becomes very polar and hydrophilic.
Silica gel is the adsorbent particle prepared by coagulation of a colloidal solution of
silicic acid followed by controlled dehydration. Liquid sodium silicates are
8 1 Basics of Adsorption

neutralized by sulfuric acid and the mixture is then coagulated to form hydrogel.
Spherical silica gel particles are prepared by spray drying of the hydrogel in hot air
[11]. Silica gel presents an amorphous structure and thus is generally considered
less stable than crystalline zeolites to repeated cycles of adsorption/desorption of
water vapor. Silica gel is best suited for adsorbing water as a refrigerant. It should
be noted, however, that the degeneration of silica gel may begin at temperatures
higher than 100 C [12]. The use of microporous silica gel (e.g. Fuji-Davison RD)
in adsorption chillers [13] is a typical low temperature application eld of silica
gels. One of the most important drawbacks of silica gel is its lower differential
refrigerant uptake at higher temperature lifts (temperature difference between
condenser and evaporator) compared, for example, with zeolite. In such case, the
adsorbent amount needed can be around three times as large as that of zeolite to
produce the same heat pump effect [14]. However, these materials have the obvious
advantage of the low cost and should therefore be considered in applications where
working conditions are not stressful.

1.2.2 Composite Materials

Aristov et al. [1517], Restuccia et al. [18, 19] and Freni et al. [20] have reported on
a range of composite salt in porous matrix materials (Selective Water Sorbents or
SWSs) in which either CaCl2 or LiBr is impregnated into mesoporous silica gel.
These materials adsorb water in an intermediate way between that of solid adsor-
bents, salt hydrates and liquid absorbents. A large variety of both host matrices
(such as meso-porous or micro-porous silica gels, aluminas, porous carbons,
polymers, etc.) and salts (CaCl2, LiBr, MgCl2, LiCl, etc.) gives the possibility of
changing the sorbent properties to match a wide range of applications among which
are sorption heat pumps and refrigerators. SWS-1L (meso-porous silica gel
impregnated with CaCl2) showed up to 75 % concentration of water and the
majority of desorption is achieved at low temperatures (70120 C). Recently, the
composite sorbent SWS-8L silica modied by calcium nitrate proved to be
attractive for utilization in adsorption chillers driven by low grade heat [21, 22].
Possible issues that must be further investigated for practical utilization of such
composite materials are the risk of leakage of the corrosive salt and the long-term
stability against hydrothermal cycling.

1.2.3 Activated Carbons

Activated carbons are the microporous carbonaceous adsorbents whose history can
be tracked back to 1600 B.C. when wood chars were used for medicinal purposes in
Egypt. In Japan, a well for underground water equipped with a charcoal lter at the
bottom was found at an old shrine (Kashiwara Jingu, Nara) constructed in the 13th
1.2 Working Pairs 9

century A.D. In Europe wood char and later bone char were used for rening beet
sugar, a practice started in France because of the blockade against the continent
during the Napoleonic era. In the 20th century, during the world wars, the need to
develop gas masks stimulated rapid growth in adsorption research [11]. Carbons
may be used with most refrigerants, with the exception of water, and are available
in granular, extruded and consolidated forms as well as bers, felts and cloths. Most
carbons used for gas adsorption will adsorb up to perhaps 30 % of ammonia or
methanol, but so-called super-activated carbons can adsorb up to twice that quantity
at the expense of a drop in the temperature lift. The benets of new adsorbents
materials (rather than conductivity enhancers) are to reduce the adsorbent mass
required and/or to result in an increased differential refrigerant uptake during a
cycle. Both can result in higher coefcients of performance (COP). It is well known
that the physical adsorption process within active carbon is a function of pore size
distribution and of active sites. Within certain limits these may be controlled and
depend on the precursor carbon source and the manufacturing route. Activated
carbon (AC) is highly utilized for solar ice making purposes with methanol as a
refrigerant [2325]. The latent heat of evaporation of methanol is about half that of
water, but its low freezing point offers the possibility to obtain subzero evaporation
temperatures without freezing problems. However, above 125 C, AC becomes a
catalyst for the reaction: methanol water + dimethyl ether, which would stop the
adsorption process [26]. Vasiliev [27, 28] has successfully experimented with
mixtures of metallic chlorides impregnated into active carbon bers. Since CaCl2
for example has such a large concentration change (1, 2 or 4 mol of ammonia per
mole of CaCl2, depending upon the reaction) it can signicantly enhance the per-
formance. However, the well-known features of systems that use only metallic salts
and ammonia are that there is a large volume change in the salt upon adsorption or
desorption and that the reaction rate is limited by chemical kinetics (in addition to
the heat and mass transfer limitations experienced in physical adsorption). It would
seem reasonable that the use of one or more salts in combination with an active
carbon would be advantageous, but determining the optimum mix is a subtle and
complex task.

1.2.4 Zeolites

The history of zeolites began in 1756 when the Swedish mineralogist Cornstedt
discovered the rst zeolite mineral; stilbite [29]. He recognized zeolites as a new
class of minerals consisting of hydrated aluminosilicates of the alkali and alkaline
earths. Because the crystals exhibited intumescence when heated in a blowpipe
flame, Cornstedt called the mineral a zeolite from two Greek words zeo and
lithos meaning to boil and a stone. In 1840, Damour observed that crystals of
zeolites could be reversibly hydrated and dehydrated with no apparent change in
their transparency or morphology [30]. More than 30 kinds of zeolite crystals have
been found in natural mines [11]. Barrer began his pioneering work in zeolite
10 1 Basics of Adsorption

adsorption and synthesis in the mid-1930s1940s [31]. He presented the rst


classication of the then known zeolites based on molecular size considerations in
1945 [32] and in 1948 reported the rst denitive synthesis of zeolites including the
synthetic analogue of the zeolite material mordenite [33]. Barrers work inspired
Milton of the Linde Division of Union Carbide Corporation in the mid to late 1940s
to initiate studies in zeolite synthesis in search for new approaches for separation
and purication of air [31].
Zeolite may be used with the three major sorption refrigerants, but most of
research has been done on zeolite-water. It is nontoxic and environmentally
friendly. The commercial products commonly used in several areas of the process
chemistry, catalysis, etc. are the alumino-silicate zeolites of type 4A,13X and Y [34,
35]. The afnity of these zeolites to the water molecule is due to their strongly
hydrophilic character that results in a high adsorption capacity for water vapor (in
terms of weight of water over the weight of the adsorbent material), high adsorp-
tion/desorption heat and a high structural stability under a high number of operating
cycles. Such alumino-silicate zeolites, however, have the disadvantage of requiring
high regeneration temperature (and high expenditure of thermal energy) to effec-
tively enable the desorption of water from the zeolite, limiting the use of these
materials for systems where a high temperature driving heat source (150300 C) is
available, and where higher temperature lifts in the order of 40 C or more are
required [6].
Accordingly, R&D on adsorbent materials for low temperature applications
(drying, air conditioning, thermal storage, refrigeration) focused on new classes of
zeolite materials, denominated AlPO and SAPO, able to combine a moderate
hydrophilicity with a high capacity of adsorption of water vapor, resulting in
moderately low regeneration temperature (60100 C) and a reduced desorption
heat, while maintaining high performance [36, 37]. A pioneering study on AlPO
zeolites carried out by Jaenchen et al. [38], demonstrated that the partially hydro-
phobic framework of such materials allows regeneration temperature signicantly
lower than that typical for strongly hydrophilic zeolites. A number of studies were
carried out on different types of (silico)aluminophosphates (AlPO5, AlPO18,
SAPO34, etc.) [3941], conrming the interesting properties of such materials for
adsorption cycles. Originally synthesized for catalytic interest, only few SAPO
zeolites have been utilized in industrial processes for their moderate Bronsted
acidity, like in lube oil dewaxing by Chevron and in methanol-to-olens conversion
by UOP.
Novel synthetic functionalized materials AQSOA, belonging to the class of
SAPO zeolites, were industrially developed and produced in a pilot scale by
Mitsubishi Plastic Incorporation (MPI) [42]. Furthermore, adsorption equilibrium
and durability of Functional Adsorbent Material-Zeolites; (AQSOA-FAM-Z01 and
AQSOA-FAM-Z02) have been tested [43, 44]. The water vapor adsorption iso-
therms of both AQSOA-zeolites were found to be S-shaped and highly dependent
on temperature. A small hysteresis has been observed and there has been almost no
change observed in the adsorption equilibrium after 100,000 cycles of water vapor
adsorption-desorption indicating that AQSOA-zeolites are durable enough for
1.2 Working Pairs 11

practical use in adsorption heat pumps. Further experimental testing of the adsor-
bent AQSOA FAM-Z02 showed that this material can provide attractive perfor-
mance in the typical operating conditions of adsorption chillers [45, 46].
A general issue of AlPO and SAPO zeolites is the high capital cost, as the
synthesis process is more expensive than standard commercial zeolites and no
industrial mass production has been established until now. Even the Y type zeolites
have partially hydrophilic character, and may therefore represent a cheaper alter-
native to AlPO and SAPO. Some years ago, UOP developed a novel zeolite
DDZ70, which is a high silica content Y zeolite exchanged with cerium to reduce
its hydrophilicity [47, 48]. Some studies demonstrated that utilization of DDZ70
can be advantageous for low temperature adsorption cycles, due to the lower
regeneration temperature (<150 C) and lower water adsorption heat [49]. How-
ever, as far as the authors know, DDZ70 is no longer available on the market.

1.2.4.1 Structural Characteristics of Zeolites

The most important characteristic of zeolite materials is their crystalline structure


made of primary structural units assembled into secondary polyhedral building
units such as cubes, hexagonal prisms, octahedral and truncated octahedral orga-
nized in a tridimensional framework forming pores and channels where the
adsorbate molecules can be accommodated. The largest family of zeolite materials
is the alumino-silicate zeolites where the primary units are tetrahedra at the center of
which a silicon (SiIV) or an aluminum (AlIII) atom is located with four oxygen
anions (O2) surrounding it. Each oxygen atom is shared with the adjacent tetra-
hedron forming the secondary building units. Unit cells of type A and type X
zeolites are shown in Fig. 1.5b, c. The aluminate tetrahedra (AlO4) result in a
decit of positive valence, requiring the addition of cations such as alkaline or
alkaline earth ions corresponding to the number of Al atoms. These cations are
positioned extra-framework and balance the negative charge of the framework, so
that are easily exchangeable. The size and properties of theses cations modify the
adsorption characteristics of zeolites since they affect the size of the window
between the cells [34, 35].
Adsorbate molecules are contained in the channels and cavities formed by the
tetrahedra spatial arrangement. The unit cell of zeolite 4A, for example, contains 24
tetrahedra, 12AlO4 and 12SiO4 and when fully hydrated, 27 water molecules are
contained in the central cavity of the unit cell and in the eight smaller sodalite cages
(Fig. 1.5a, b) [50]. The commercial 4A is synthesized in the sodium form, i.e. with
12 sodium cations occupying all eight sites in I, three sites in II and one site in III
(Fig. 1.5d) with an effective aperture size of 3.8 . The sodium cation can be
replaced by various other cations or by hydrogen ion causing a variation of the
effective aperture of the central cavity from a minimum of 3.0 (in the potassium
form) to a maximum of 4.3 (in the calcium form).
12 1 Basics of Adsorption

Fig. 1.5 Zeolite structure:


a sodalite cage, or beta cage
or truncated octahedron;
b type A zeolite unit cell;
c unit cell of types X and Y,
or faujasite; d cation sites in
type A; e cation sites in types
X and Y

The structure of type X and Y zeolites is similar to type 4A and occurs in the
natural mineral zeolite faujasite (Fig. 1.5c). Each cell contains 192 (Al, Si)O4
tetrahedra, however the number of aluminum ions per unit cell varies from 96 to 77
for X zeolite, and from 76 to 48 for Y zeolite. The framework of faujasite has the
largest central cavity pore volume of any known zeolite, amounting to about 50 %
void fraction in the dehydrated form (Fig. 1.5c, e). The free diameter of the central
cavity is 13.7 . The unit cell, when fully hydrated, contains approximately 235
water molecules, primarily in the central cavity. The BET surface area measured
with N2 for zeolites falls in the range between 500 and 800 m2/g [50].
The strong anionic nature of the framework in alumino-silicate zeolites and the
correspondingly strong cations that are bonded to the framework zeolite are
determinant in the interaction with the adsorbate so that zeolites A, X and Y show
evident afnity to polar molecules and are classied as high hydrophilic materials.
A different family of porous materials with similar characteristics of alumino-
silicate zeolites are the crystalline microporous alumino-phosphates. In these zeo-
lites, denominated AlPO4, the crystal structure is basically formed by tetrahedra of
aluminum (AlIII) and phosphor (PV) sharing oxygen atoms and forming a tridi-
mensional network with a porosity in the order of nanometers. In the framework,
phosphate has a valency of ve (PV) and a coordination number of four so that the
tridimensional crystalline structure is perfectly electroneutral (i.e., without unbal-
anced electrical charges) [37]. In the regular microporous crystalline alumino-
phosphate structures, no oxo bridges between two Al atoms or between two P
atoms are present with a strict alternation of Al and P in the tetrahedral nodes of the
framework. This framework concept is reflected in the name for crystalline
1.2 Working Pairs 13

microporous alumino-phosphates, which is based on the idealized chemical com-


position formula of the framework, AlPO4. The positive electronic charge on the Al
atoms of the framework is higher than on the P atoms, owing to the lower elec-
tronegativity. Each pair of adjacent AI, P atoms, then, represents a permanent
dipole that may play an important role in many processes, including adsorption and
catalytic conversion reactions with polar molecules [51].
These molecular sieves have a very narrow range of chemical composition
compared with the wide range of Si/Al ratio in alumino-silicate zeolites, but exhibit
a rich diversity of framework structures. They are formed in both cage-type and
channel-type framework structures, and the sizes of their pore apertures vary from
2.8 in AlPO-20 to 7.3 in AlPO-5. To create the porosity, these zeolites require
the addition of a template molecule during the synthesis, which is normally an
amine or a quaternary ammonium ion. The template remains entrapped in the voids
of the zeolite and is removed by calcination [51].
Many types of isomorphic substitutions are possible in microporous crystalline
alumino-phosphates. Per denition, isomorphic substitution corresponds to a
replacement of an element in a crystalline lattice by another element with similar
cation radius and coordination requirements. Isomorphic substitution is currently
achieved during crystallization, by adding the element to be incorporated to the
synthesis mixture. SAPO zeolites, for example, are obtained by isomorphic sub-
stitution of some P atoms with Si atoms. The replacement of (PV) with (SiIV)
generates an imbalance of charges in the tridimensional framework. After removal
of the organic template from the micropores, these negatively charged frameworks
have the potential of Brnsted acidity and cation exchange, provided the incor-
porated elements are stable in the framework. The existence of a net framework
charge influence the adsorption behavior of SAPO zeolites offering promising
opportunities for their use as new sorbent materials.
The material that has received the most study is SAPO-34. SAPO-34 has cages
containing six 8-membered rings and bounded by 36 T-atoms (T is any tetrahe-
drally coordinated cation) (Fig. 1.6). Because of the high symmetry of the structure

Fig. 1.6 Structure of SAPO-


34 cage (CHA)
14 1 Basics of Adsorption

Table 1.2 Properties comparison for different zeolites


Zeolite Unit cell (T-atoms) Structure Water capacity (g/g) Commercial availability
4A(-Na)a 24 (Al, Si)O4 LTA 0.29 High
X(-Na)a 192 (Al, Si)O4 FAU 0.34 High
Y(-Na)a 192 (Al, Si)O4 FAU 0.34 High
AlPO-18 48 (Al, P)O4 AEI 0.260.28 n.a.
SAPO-34 36 (Al, P, Si)O4 CHA 0.300.32 Scarce
a
Sodium form

each T-atom is actually shared by 3 adjacent cages, so there are effectively 12


T-atoms per cage. This structure is observed in the mineral chabasite so that to such
a framework type is assigned the code CHA.
The surface selectivity of AlPO-based molecular sieves is reported to be weakly
to moderately hydrophilic. Adsorption of permanent gases and hydrocarbons on
AlPOs follows a micropore volume-lling mechanism resulting in type I iso-
therms, which is the common isotherm shape for microporous sorbents. Unusual
sorption properties regarding polar molecules have been found with many AlPO
materials, in particular, adsorption of water in AlPOs gives rise to a type V iso-
therm shape. The introduction of framework charges in substituted analogues
generates strong adsorption centra for polar molecules. The analogues SAPOs
encompass a range of moderately to highly hydrophilic surface properties,
depending on the degree of isomorphic substitution. SAPO-34 zeolite, for example,
shown a partial hydrophilic behavior with a type V isotherm towards the water
vapor adsorption.
In Table 1.2 are resumed some characteristics of the most commonly used
zeolites for adsorption systems.

1.2.5 New Trends

In addition, evaluation of super-molecular Metal-organic frameworks (MOFS) for


application in adsorption chiller is a major trend of the research in this eld. MOFS
attracted attention due to the open crystalline structure and very high surface area.
Some studies revealed a large capacity of water adsorption under practically useful
operating conditions [5254]. Possible obstacles for practical use are a signicant
ad/desorption hysteresis and insufcient stability.
Other authors proposed other promising adsorbents such as attapulgite clay
sorbents impregnated with inorganic salts [55, 56], silicalite MCM-41 [57] or
zeolite/active carbon compounds [58], that showed attractive sorption properties for
utilization of low grade heat.
Figure 1.7 shows the water adsorption isobar (pH2 O 11 mbar) measured at
CNR ITAE for different adsorbents by the thermo-gravimetric technique based on a
1.2 Working Pairs 15

30

Silica Gel
25
Y DDZ 70
Water uptake, wt.%

AQSOA Z02
20
Zeolite 13X

15

10

0
20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Temperature, C

Fig. 1.7 Water adsorption isobar (pH2 O 11 mbar) measured at equilibrium for different
adsorbents by thermo-gravimetric technique

Cahn microbalance. The equilibrium curves comparison demonstrate that the


adsorbents Y-zeolite DDZ70 from UOP and SAPO-zeolite AQSOA-FAM-Z02
from MPI, allow a water uptake variation much higher than standard adsorbents.
In general, the choice of an optimal adsorbent material for a certain application,
should be made taking into account various factors both of thermodynamic type
(water adsorption capacity under given operating conditions, regeneration tem-
perature level, ad/desorption heat) and dynamic (permeability to water vapour
diffusion, intraparticle water adsorption kinetics). Elevated stability against
mechanical stresses and hydrothermal cycling must be ensured. Other decisive
parameters for the nal choice are the cost and the availability on the market.
In this book we concentrate on zeolite-water adsorption heat pumps. However,
the methodologies described are valid for other working pairs.

1.3 The Ideal Intermittent Adsorption Heat Pump Cycle

The simplest conguration of an adsorption heat pump is the intermittent, or


periodical, cycle. In its very basic conguration, the heat pump module consists of
two heat exchangers incorporated in a tight Vessel (two-heat-exchanger concept) as
depicted in Fig. 1.8. These are namely, adsorber/desorber heat exchanger and
16 1 Basics of Adsorption

Fig. 1.8 The two-heat-exchanger adsorption heat pump concept

evaporator/condenser heat exchanger. The adsorption heat pump cycle consists of 4


processes, which are preheating, desorption-condensation, precooling and adsorp-
tion-evaporation. Refrigerant vapor can flow inside the heat pump module in the
directions given in Fig. 1.8 between the adsorber/desorber and evaporator/con-
denser heat exchangers (black during the desorption-condensation phase and red
during the adsorption-evaporation phase).
Figure 1.9 illustrates the ideal intermittent adsorption heat pump cycle on a
schematic vapor pressure diagram. The refrigerant uptake (x) is dened as the ratio
between the mass of the adsorbed refrigerant and the mass of dry adsorbent. Starting
with the state point 1, in which the adsorbent is cold and saturated with the refrig-
erant (x = xmax), heat is applied to the adsorber/desorber heat exchanger. This results
in heating up the adsorbent, which consequently results in desorbing a certain
amount of the refrigerant. Accordingly, the pressure inside the heat pump module
increases, ideally without changing the refrigerant uptake (so-called isosteric
preheating), until the minimum desorption temperature is reached (state point 2).
The preheating can only take place at a constant uptake, if the volume of the vapor
phase inside the heat pump module is negligibly small. An additional, but quite
important requirement for the ideal adsorption heat pump cycle is that the evapo-
rator/condenser heat exchanger shall have a negligible heat capacity (product of
mass and specic heat capacity (m c) of both heat exchanger material and its heat
transfer mediums content or holdup). In such a case, the evaporator/condenser heat
exchanger can be heated from the evaporator temperature (6) up to the condenser
temperature (5) with a negligible amount of heat or, in other words, vapor leaving
the adsorbent during the preheating process (12) and being condensed over the
condenser/evaporator heat exchanger in order to provide the required heat.
At the minimum desorption temperature (2) the system pressure becomes equal
to the saturation pressure corresponding to the temperature (con) of the heat sink
available for receiving the heat of condensation (Heating Net). The desorption-
condensation process starts from this point on and the refrigerant is desorbed and
1.3 The Ideal Intermittent Adsorption Heat Pump Cycle 17

H2O xmax xmin

100

Condenser Desorption
50 3
5 2
Pressure [mbar]

Preheating Precooling

Evaporator
4
10 6 1 Adsorption max

Indices
HN: Heating Net
AHS: Ambient Heat Source
AHS = Ev HN = 1 = Con
1
10 20 30 40
0 50 100 150 200 300
Temperature [ o C]

Fig. 1.9 The ideal adsorption heat pump process on a schematic vapor pressure diagram

then condensed, ideally at a constant pressure (process 23 in Fig. 1.9). The con-
stant pressure desorption-condensation process requires that the condenser heat
exchanger may have an innite heat transfer capacity, despite its negligible heat
capacity. In other words, the condenser shall attain an innite overall heat transfer
coefcient, so that every desorbed refrigerant vapor molecule shall be condensed
immediately. The desorption process proceeds until the adsorbent temperature
reaches the maximum available desorption temperature and the refrigerant uptake
reaches the cycle minimum uptake (xmin) (state point 3). The ideal heat pump
process requires, moreover, that the temperature gradient between the adsorbent bed
and the heat transfer medium flowing from the driving heat source is negligible, in
order that the adsorbent temperature may reach the maximum available driving
temperature (3 = max). During this working phase a refrigerant amount equals to
(xmax xmin) per kg of dry adsorbent is condensed and collected in the evaporator
space.
The second working phase begins as the adsober/desorber heat exchanger is
cooled down by rejecting its heat to the heating net by switching the upper 3-way
valve in the red direction (Fig. 1.8). The adsorbent is pre-cooled and becomes able
to adsorb refrigerant vapor. This results in decreasing the system pressure, ideally
without changing the refrigerant uptake within the adsorbent (vanishing volume for
the vapor phase as well as the negligible heat capacity of the evaporator-condenser
heat exchanger). As the adsorber/desorber heat exchanger is pre-cooled a portion of
the previously desorbed and condensed refrigerant is adsorbed and the latent heat of
vaporization is drawn from the remaining liquid refrigerant on the evaporator. This
18 1 Basics of Adsorption

results in decreasing the refrigerant temperature from state point 5 to state point 6 at
the end of the pre-cooling process (34). This process is equivalent to the throttling
process of the refrigerant in vapor compression heat pumps, which takes place in
the expansion valve.
In other words, the adsorberdesorber in an intermittent adsorption heat pump is
not only taking the role of the mechanical compressor, but also the role of the
expansion valve in a vapor compression cycle. The adsorption process, within
which the cooling effect is produced in the evaporator, starts from state point 4 and
proceeds by further cooling the adsorber/desorber heat exchanger until the whole
amount of refrigerant is evaporated upon receiving the required heat from the
ambient heat source and adsorbed by the adsorbent. The heat of adsorption gen-
erated in the adsorbent is then rejected as useful heat to the heating net along with
the heat of further cooling the adsorbent and the adsorber heat exchanger. During
this process, the lower 3-way valve depicted in Fig. 1.8 shall connect the ambient
heat source with the evaporator along the red direction. During this working phase
of the idealized intermittent heat pump cycle, a refrigerant amount equals to
(xmax xmin) per kg of dry adsorbent is evaporated upon receiving the required heat
from the ambient heat source and is adsorbed by the adsorbent to increase its
refrigerant uptake from xmin (state point 4) to xmax (state point 1). The evaporation-
adsorption process takes place under constant pressure conditions as the evaporator
heat exchanger is supposed to have an innite heat transfer capacity (ideal cycle
assumption). The evolved heat of adsorption is rejected to the heating net, so that
the adsorbent temperature may reach the minimum heating net temperature at state
point (1) and the heat pump cycle can restart again. The temperature gradient
between the adsorbent and the heating net is also neglected (1 = HN) as a
requirement of the idealized cycle.
Figure 1.10 illustrates a possible design of a two-heat-exchanger adsorption heat
pump module [59] applying zeolite in form of loose pellets between the ns of a
nned-tube adsorber heat exchanger. The evaporator-condenser heat exchanger is
constructed as a snail out of a nned tube located on the bottom of the heat pump
module. Between both heat exchangers, a radiation guarding plate is located to
reduce the direct radiation losses from the adsorber-desorber to the evaporator-
condenser heat exchangers.
The coefcient of performance (COP) of an intermittent adsorption heat pump is
dened as the ratio between the useful heat gained and the input energy required for
driving the process. Considering the aforementioned description of the ideal
intermittent adsorption heat pump cycle, the useful heat gained by the heating net
shall equal the sum of heats gained during the condensation, precooling and
adsorption phases. The driving energies of the process are the heat added during the
preheating and desorption phases as well as the electrical energy required for
driving the circulating pumps of the heat transfer mediums as well as the incor-
porated valves (the two 3-WV in Fig. 1.8). Accordingly, the COP can be mathe-
matically formulated as
1.3 The Ideal Intermittent Adsorption Heat Pump Cycle 19

Fig. 1.10 An adsorption heat


pump module design [59]

Quseful Q QPC QAds


COP Con . 1:7
Qinput QPH QDes WEEl

A complete energy balance on the heat pump module, neglecting its heat losses
to the surroundings, requires

QCon QPC QAds QPH QDdes QEv WEl . 1:8

Neglecting the electrical energy required for driving the pumps and valves and
applying Eq. (1.8) in Eq. (1.7) result in the following formulation for the COP
(compare with Eq. (1.3) before).

QEv
COP 1 1:9
QPH QDes

It is evident from Eq. (1.9) that in order to maximize the COP, the amount of
heat received by the evaporator out of the ambient heat source shall be maximized
against the amount of heat flowing from the driving heat source to realize the
preheating and desorption processes.
The main assumptions of the idealized intermittent adsorption heat pump cycle
can be summarized as follows:
20 1 Basics of Adsorption

1. All heat exchangers are assumed ideal or of an innite heat transfer capacity, (no
temperature gradient is required for the heat transfer processes).
2. All heat exchangers have a negligible heat capacity.
3. The volume of the vapor phase within the heat pump module is negligibly small.
Under these conditions, the evaporator cooling effect attains its maximum value
and the least amount of driving heat to realize both preheating and desorption
processes is required, resulting in the COP of the ideal intermittent cycle. The
maximum evaporator cooling effect is dened in Eq. (1.10) as the product of
the maximum exchangeable refrigerant mass between the adsorber/desorber and the
evaporator/condenser heat exchangers and the latent heat of vaporization of the
refrigerant (hfg,ref) at the evaporator temperature.

QicEv;max mref;max  hfg;ref mAds  xmax  xmin  hfg;ref mAds  Dxmax  hfg;ref
1:10

The rst part of the least amount of heat required to realize both preheating (12)
and desorption processes (23) under the aforementioned assumptions can be
estimated as the sensible heat required to heat up only the adsorbent mass (mAds)
having the average specic heat capacity (cAds ) from the temperature (1) up to the
maximum available heat source temperature (max or 3). The specic heat capacity
is in general a function of both temperature and refrigerant uptake. Accordingly, it
has to be averaged between both temperatures and refrigerant uptakes between state
points (1 & 3) in Fig. 1.9. The second part of the least driving heat QPH QDes icmin
is required to overcome the binding energy (qB ) applied by the adsorbent on the
adsorbed refrigerant and to evaporate the refrigerant (hfg,ref). The sum of both heats
is termed as the heat of adsorption (qAds ). The heat of adsorption is mainly
dependent on the refrigerant uptake; therefore, its value at the mean refrigerant
uptake between (xmin and xmax) shall be applied as given in Eq. (1.11).

QPH QDes icmin mAds  cAds  #3  #1 mref ; max  qAds


mAds  cAds  #3  #1 mAds  Dxmax  qAds

mAds  cAds  #3  #1 mAds  Dxmax  hfg;ref qB
1:11

The COP of the ideal intermittent adsorption heat pump cycle (COP)ic can then
be estimated by Eq. (1.12) below combining both Eqs. (1.10) and (1.11).

QicEv; max
COPic 1 1:12
QPH QDes icmin
1.4 Energetic Evaluation of the Real Intermittent Adsorption Heat Pump Cycle 21

1.4 Energetic Evaluation of the Real Intermittent


Adsorption Heat Pump Cycle

In his PhD-study, Westerfeld [60] has derived a set of dimensionless parameters to


identify the most relevant influencing parameters on the COP of an intermittent
zeolite-water adsorption heat pump cycle. This methodology will be generalized
within the remaining course of this chapter, in order to make it applicable for other
possible adsorbent-refrigerant working pairs of intermittent adsorption heat pumps.
A thorough discussion on the effect and the order of magnitude of each relevant
influencing parameter leading to specic conclusions on how to maximize the COP
of intermittent adsorption heat pumps will follow.
Figure 1.11 illustrates both idealized and real intermittent adsorption heat pump
cycles on a schematic vapor pressure diagram.
As a certain temperature gradient is required to realize the heat transfer between
the adsorbent and the heat transfer mediums of both driving and heating net res-
ervoir loops, the minimum and maximum refrigerant uptakes of the ideal cycle
(xmin, xmax) cannot be reached in the real cycle. As depicted in Fig. 1.11, the
maximum achievable uptake is reduced to (xAds at state point I of the real cycle). In
the same time, the minimum uptake xmin of the ideal cycle is increased to xDes in the

xmax xAds xDes xmin

100 H2O xheat

Con 2 II III
50 3
Pressure [mbar]

Ev 1
4
10 I IV

5
xcool
x = xAds - xDes

xmax = xmax - xmin


1
10 20 30 40
0 50 100 150 200 300
o
Temperature [ C]

Fig. 1.11 Ideal and real adsorption heat pump cycles on a schematic vapor pressure diagram [60,
61]
22 1 Basics of Adsorption

real cycle (state point III). Accordingly, the maximum evaporator cooling effect of
the real cycle can be expressed by Eq. (1.13) as:

QEv;max mref  hfg;ref mAds  xAds  xDes  hfg;ref mAds  Dx  hfg;ref 1:13

In the real cycle, the heat capacity (product of mass and specic heat capacity) of
the evaporator/condenser heat exchanger as well as of the heat transfer medium
included inside this heat exchanger cannot be neglected. At the end of the
desorption-condensation process, the condenser heat exchanger (its own mass and
the mass of the heat transfer medium of the ambient heat source loop inside the
condenser) as well as the refrigerant mass just condensed over it shall have
the condenser temperature. All those heat capacities must be cooled down to the
evaporator temperature, during the precooling phase of the adsorber heat exchange,
before the effective evaporation process upon receiving the heat of vaporization
from the ambient heat source may start. The heat required for cooling those heat
capacities (QEv,cool) is simply realized by evaporating a certain amount of the
refrigerant, which is previously condensed over the condenser heat exchanger,
under adiabatic conditions, in order for the heat capacities to be cooled down to the
evaporator temperature. In other words, the heat required to vaporize this amount of
refrigerant is going to be absorbed from the remaining refrigerant and consequently
from the condenser/evaporator heat exchanger and its heat transfer mediums
content. The amount of refrigerant, which has to be evaporated in order to realize
the required precooling effect, shall be estimated iteratively. This amount is then
adsorbed by the adsorbent during the precooling process IIIIV, which cannot be
considered isosteric any more (compared to 34 of the ideal cycle in Fig. 1.11). In
other words, the net evaporated refrigerant mass upon receiving its heat from the
AHS shall be reduced by the amount of refrigerant required to realize the preco-
oling of the evaporator heat exchanger, its heat transfer medium content as well as
the refrigerant itself from the condenser down to the evaporator temperature.
Equation 1.14 gives the mathematical expression for the heat required to cool down
the evaporator heat exchanger and the refrigerant.

QEv; cool m  cEvHX m  cref  0Con  0Ev


m  c)EvHX m  c)ref  D 0EvCon 1:14

Herein, the term (m c)EvHX shall include both heat capacities of the evaporator/
condenser heat exchanger and its heat transfer mediums content. The change in the
differential refrigerant uptake required to realize this precooling effect can be
estimated by Eq. (1.15) as

QEv; cool m  c)EvHX m  c)ref


D xcool  D 0EvCon : 1:15
mAds  hfg;ref mAds  hfg;ref
1.4 Energetic Evaluation of the Real Intermittent Adsorption Heat Pump Cycle 23

The net cooling effect of the evaporator can be estimated by subtracting


Eq. (1.14) from Eq. (1.13) as given in Eq. (1.16).

QEv QEv;max  QEv; cool mAds  Dx  hfg;ref


 m  c)EvHX m  c)ref  D 0EvCon
mAds  Dx  Dxcool  hfg;ref 1:16

In a real cycle, the evaporator shall not have an innite heat transfer capacity,
simply in order to minimize its heat capacity and, consequently, to minimize its
precooling loss just described before. Accordingly, the evaporator, if undersized,
may not be able to generate the required vapor mass flow rate required by the
adsorber. This results in reducing the pressure down to a certain limit, dependent on
the relative effectiveness of both evaporator and adsorber heat exchangers. After
reaching a minimum value, the evaporator starts to satisfy the adsorber require-
ments and the pressure shall increase once more to approach the saturation pressure
at the end of the adsorption process (State point I in Fig. 1.11).
At the end of the adsorption-evaporation process, the Adsorbent may reach the
state point (I) and the evaporator/condenser heat exchanger the evaporator tem-
perature. The preheating process starts by connecting the adsorber heat exchanger
with the DHS, resulting in driving a certain amount of refrigerant out of the
adsorbent. The condenser heat exchanger is, however, still cold from the previous
process and must be heated up before the useful heat of condensation can be
supplied to the heating net. In this time, only the heat capacity of the heat exchanger
and its heat transfer medium content shall be heated from the evaporator up to the
condenser temperature. This requires also that a certain amount of the desorbed
refrigerant vapor shall be condensed on the heat exchanger under adiabatic con-
ditions in order to achieve this preheating process of the condenser heat exchanger,
which represents a further difference between ideal and real cycles. Accordingly,
the preheating process of the adsorbent in a real cycle cannot be considered isosteric
as depicted in Fig. 1.11 (compare III with 12 of the ideal cycle).
In order to reduce the preheating loss of the condenser heat exchanger and to
keep the cost of such heat exchanger in a reasonable range, its heat transfer capacity
shall be nite. Figure 1.11 illustrates the course of pressure change during the
desorption-condensation process with an undersized condenser heat exchanger. It
follows, that the rate of desorption of refrigerant vapor shall exceed the rate of
condensation, resulting in a pressure increase over the saturated pressure. By
approaching the end of the desorption-condensation process, the condenser starts to
be able to condense all refrigerant vapor being desorbed, which results in reducing
the pressure once more until it approaches the saturation pressure by the end of the
desorption process (state point III on Fig. 1.11).
The heat required to drive the real adsorption heat pump cycle shall differ from
that expressed by Eq. (1.11) by rst adding the heat capacity of the adsorber heat
24 1 Basics of Adsorption

exchanger including its heat transfer mediums content to the rst sensible heat
part. In addition, the minimum and maximum adsorbent temperatures (1 & 3)
shall be replaced by the real temperatures (I & III) or (Ads & Des) reached by the
adsorbent. Moreover, the maximum differential refrigerant uptake shall be replaced
by the real differential refrigerant uptake (x). Equation 1.17 gives the mathe-
matical expression for the driving heat required for the real intermittent adsorption
heat pump.

QPH QDes mAds  cAds m  cAdHX  0III  0I mref  qAds



(mAds  cAds m  cAdHX  D#AdsDes mAds  Dx  hfg;ref qB
1:17

The COP of the real intermittent adsorption heat pump cycle (COP) can then be
estimated by Eq. (1.18).

QEv
COP 1
QPH QDes
mAds  Dx  Dxcool  hfg;ref
1 
mAds  Dx  hfg;ref qB (mAds  cAds m  cAdHX )  D#AdsDes
1:18

Restructuring Eq. (1.18) and applying Eq. (1.15) for xcool result in the fol-
lowing Eq. (1.19) for the coefcient of performance of real intermittent adsorption
heat pump cycles.

1  cref D# cref D#EvCon mc)EvHX


hfg;ref  hfg;ref  mcref
EvCon

COP 1 1:19
q mc)
1 h B cAdshD#AdsDes cAdshD#AdsDes  mcAdHX
fg;ref fg;ref Dx fg;ref Dx Ads

Equation (1.19) can be rewritten to represent the COP of a real intermittent


adsorption heat pump cycle as a function of ve dimensionless parameters as given
in Eq. (1.20).

1  Kref  Kref  KEvHX


COP 1
1 K1; Adsref K2; Adsref K2; Adsref  KAdHX
1  Kref  1 KEvHX
1 1:20
1 K1; Adsref K2; Adsref K2; Adsref  KAdHX
1.5 Basic Features and Design Requirements 25

1.5 Basic Features and Design Requirements


of Intermittent Adsorption Heat Pumps

It is evident that all given dimensionless parameters in Eq. (1.20) shall be mini-
mized in order to maximize COP. It is also evident that the maximum possible
value attainable for the COP of an intermittent adsorption heat pump cycle is two
(only in case all dimensionless parameters may vanish).
The rst dimensionless parameter (Kref) combines two caloric properties of the
refrigerant with the temperature lift of the heat pump process (Con Ev). Indeed
refrigerants with lower specic heat capacity to the latent heat of vaporization ratio
are desired for obtaining higher COPs. In addition, higher latent heat of vapori-
zation is advantageous if minimizing the size of the heat pump is a desired goal.
Similar to vapor compression heat pumps, increasing the temperature lift results in
decreasing the COP. In most cases, the temperature lift is dictated by the application
eld of the heat pump and cannot, therefore, be left as a design parameter. For
zeolite-water working pairs, the magnitude of Kref lies in the range of 0.002 because
of the quite high latent heat of vaporization of water.
The evaporator heat exchanger dimensionless parameter (KEvHX) represents the
ratio between the heat capacity of the heat exchanger itself as well as its heat
transfer medium content (or holdup) and that of the refrigerant exchanged between
adsorber/desorber and condenser/evaporator within the real adsorption cycle.
Reducing this parameter in order to increase COP implies the reduction of the heat
capacity of the evaporator/condenser heat exchanger and that of its heat transfer
medium content as well. In addition, maximizing the amount of refrigerant
exchanged within the heat pump cycle leads also to increasing COP for a certain
evaporator heat exchanger design. Increasing the driving heat source temperature is
almost the only operation parameter remaining to increase the refrigerant amount
exchanged. Increasing the driving heat source temperature shall, however, follow
the recommendation given afterwards upon discussing the second adsorbent-
refrigerant dimensionless parameter (K2,Adsref).
The rst adsorbent-refrigerant dimensionless interaction parameter (K1,Adsref)
represents the ratio between the binding energy acted by the adsorbent on the
adsorbed refrigerant molecules and the latent heat of vaporization of pure refrig-
erant. Dependent on the adsorbent-refrigerant pair and on the refrigerant uptake
range, this ratio may vary between 0.1 and 0.75. For most of zeolite-water systems,
the average value of K1,Adsref lies in the range between 0.3 and 0.6.
The second adsorbent-refrigerant dimensionless interaction parameter (K2,Ads-ref)
combines the sensible heat required to heat up the adsorbent from the adsorber-end
(I) to the desorber-end temperature (III) and the maximum evaporator cooling
effect (QEv,max) dened by Eq. (1.13). The specic heat capacity of the adsorbent
depends on both adsorbent temperature and refrigerant uptake. Besides, the latent
heat of vaporization is almost constant within the temperature range between
evaporator and condenser temperatures. In its reduced form given in Eqs. (1.19) and
(1.20), the second adsorbent-refrigerant dimensionless interaction parameter
26 1 Basics of Adsorption

Table 1.3 Design parameters of the two-heat exchanger adsorption heat pump module [59]
Heating net 40/30 C
Working pair Zeolite MgNaA-Water [62]
Temperature gradient for heat transfer in all heat exchangers 5K
Construction Material of all components Aluminum
Mass of adsorbent 2.56 kg
Mass of AdHX 9 kg
Water content of AdHX 1.1 kg
Mass of EvHX 1.32 kg
Brine content of EvHX 0.35 kg
Specic heat capacity of brine 3.55 J/(kg K)

depends, moreover, on the derivative d/dx, which depends mainly on the adsor-
bent structure and solely on the adsorbent-refrigerant energetic interaction as a
function of the refrigerant uptake or adsorbent temperature during the desorption
process. Reducing (K2,Adsref) requires then the knowledge of the vapor-adsorbent
equilibrium data, in order to nd out the optimum driving heat source temperature,
for a certain set of the other operating temperatures, at which the derivative dx/d
reaches its maximum value or d/dx its minimum value leading to the minimum
value of (K2,Adsref).
The last dimensionless parameter (KAdHX) represents the ratio between the heat
capacity of the adsorber heat exchanger itself as well as its heat transfer medium
content and the heat capacity of the adsorbent incorporated. Reducing this
parameter results not only in increasing COP but also in enhancing the dynamics of
the heat pump cycle, leading to more heating power.
The previous analysis has been applied on the two-heat exchanger adsorption
heat pump module depicted in Fig. 1.10 and having the design parameters given in
Table 1.3. Figure 1.12 illustrates the different values of the dimensionless param-
eters discussed before along with the COP obtained at an exemplary ambient heat
source temperature of 5 C as a function of the driving heat source temperature in
the range between 140 and 220 C.
The order of magnitude of the dimensionless group Kref  1 KEvHX  amounts
to ca. 0.1 because of the superior properties of water as a refrigerant. The rst
interaction parameter between adsorbent and adsorbate (K1,Ads-ref) increases slightly
from 0.32 to 0.34 by increasing the driving heat source temperature from 140 to
220 C.
It is evident from Fig. 1.12 that (K2,Ads-ref) attains a minimum at a driving heat
source temperature of 175 C. In the last combination between the dimensionless
parameters (K2,Adsref) and (KAdHX), the effect of the specic heat capacity variation
with both refrigerant uptake and adsorbent temperature is eliminated, as it appears
in both numerator and denominator of this combination. In other words, the course
of variation of the dimensionless parameter combination (K2;Adsref  KAdHX ) shall
follow that of the derivative /x, which is simply weighted by the remaining
1.5 Basic Features and Design Requirements 27

0.6 1.50

K2,Ads-ref
1.45
0.5

1.40

K-Parameter & COP


0.4
K1,Ads-ref
K-Parameter [-]

1.35
0.3
COP
1.30

0.2
1.25
K2,Ads-ref.KAdHX
0.1
1.20
.
Kref [1+KEvHX]
0 1.15
130 140 150 160 170 180 190 200 210 220 230
Driving Heat Source Temperature [oC]

Fig. 1.12 Dimensionless parameters and COP of the two heat exchanger adsorption heat pump
module given in Table 1.3 at an ambient heat source temperature of 5 C

constant parameters. The main weighting factor in this combination is the adsorber
heat exchanger heat capacity, which shall be minimized in order to increase the
COP. It is evident from Fig. 1.12 that /x and, consequently (K2;Adsref  KAdHX )
attain a minimum at the driving heat source temperature of 170 C. The maximum
of the COP takes place at the driving heat source temperature, at which the mini-
mum of /x or (K2;Adsref  KAdHX ) occurs; namely, at 170 C.
The optimum driving heat source temperature, at which the minimum of /x
or the maximum of the COP takes place, at different sets of operating conditions
can be estimated, if the vapor equilibrium as well as the caloric data of the
incorporated adsorbent-refrigerant pair are available.
The courses of variation of both COP and the dimensionless parameter combi-
nation (K2;Adsref  KAdHX ) with the driving heat source temperature for different
ambient heat source temperatures are illustrated in Fig. 1.13. The increase in
(K2;Adsref  KAdHX ) with decreasing the ambient heat source temperature can be
referred to the decrease in the absolute differential refrigerant uptake (x), which
means a reduction in the evaporator cooling effect. Accordingly, the COP decreases
as the ambient heat source temperature decreases. It is evident that the optimum
driving heat source temperature increases from 165 to 185 C as the ambient tem-
perature decreases from 10 C down to 10 C. It is also evident from Fig. 1.13 that
the maximum of COP occurs at the driving heat source temperature, at which the
minimum of /x or the dimensionless parameter combination (K2; Adsref  KAdHX )
takes place.
28 1 Basics of Adsorption

1.40 2.6
AHS = -10 oC
o
AHS = -5 C 2.4
o
1.35 AHS = 0 C
o
AHS = 5 C
o
AHS = 10 C 2.2

1.30

K2,Ads-ref.KAdHX [-]
2.0
COP [-]

1.25 1.8

1.6
1.20

1.4

1.15
1.2

1.10 1.0
130 140 150 160 170 180 190 200 210 220 230
Driving Heat Source Temperature [oC]

Fig. 1.13 COP and a dimensionless parameter combination of the two heat exchanger adsorption
heat pump module at different ambient heat source temperatures

In summary, the following design requirements shall be met in order to increase


the coefcient of performance of intermittent adsorption heat pump cycles:
1. Utilize refrigerants with lower specic heat capacity to latent heat of vapori-
zation ratio.
2. Choose refrigerants with possibly higher latent heat of vaporization.
3. The heat capacity of the evaporator heat exchanger as well as its heat transfer
medium shall be kept as low as possible.
4. Refrigerant-adsorbent pairs with low heat of binding are favorites.
5. Refrigerant-adsorbent pairs, with which the minimum of /x or the maximum
of x/ occurs at lower driving heat source temperatures are advantageous.
6. The heat capacity of the adsorber heat exchanger and its heat transfer medium
content shall be minimized against that of the adsorbent encountered.

1.6 Short Overview on Advanced Cycles

Special focus has been given to the research of advanced sorption cycle, trying to
overcome two critical issues of the basic cycle: the discontinuity in the production
of useful effect and the need to signicantly reduce the amount of heat needed for
1.6 Short Overview on Advanced Cycles 29

the desorption phase. Some of the main advanced cycles, which have been pro-
posed, are shortly described below.
The heat recovery cycle represents the rst historical attempt to improve the
COP of an adsorption machine [63]. It is realized with two beds working in counter
phase, so that the thermal energy needed during the isosteric phase of the bed
undergoing heating is not provided by the heat source but instead from the thermal
flux of the bed being cooled. Such heat recovery is possible until the temperatures
of the two beds are almost equal, so theoretically even during the initial part of the
desorption phase.
Thermal wave or flow reversal cycles are based on the division of adsorption
chambers into several modular units [64], so that the fluid in the adsorber heat
exchangers follows a multi-step path, which maximizes the difference between inlet
and outlet temperatures. In this way, the thermal flux is used efciently and the
recovery of up to 80 % of desorption heat is possible. Variants were proposed [65]
where inert gases are used for the regeneration of beds. However, a complicated
layout of the system and low efciency in heat transfer between the various units
have limited the practical use of thermal wave cycles.
Advanced mass recovery cycles have also been proposed in order to reduce the
temperature of the heat source needed [66] and consist in dividing the pressure drop
between condenser and evaporator in more stages, by connecting the adsorption
beds through the refrigerant pipe and at the same time heating and cooling the
adsorbing and desorbing bed, respectively.
In addition, a reallocation of adsorption/desorption phases is possible [45, 67]:
usually adsorption and desorption phase are set to an equal time. Nevertheless, it
has been demonstrated that the kinetic of the desorption process is faster than that of
adsorption, the amount depending partly on the granulate geometry of the adsorbent
material used [68]. It is then possible to close the connection between the hot bed
and the heat source while the cold bed is still adsorbing. Consequently, qualitative
criteria might be dened in order to maximize both COP and specic cooling power
(SCP) of adsorption chillers by simply fractionizing the adsorption cycle in a
different way. Reallocation is possible even by using multiple beds machines [69].
Figure 1.14 demonstrates the influence of the cycle time and of the relative duration
of the adsorption and desorption phase on the performance of a laboratory scale
single bed water/zeolite adsorption chiller tested at ITAE [45].
A thorough analysis of the different advanced cycles may be the topic of the
upcoming publications by the authors.
30 1 Basics of Adsorption

Fig. 1.14 Performances map (cooling COP vs. specic cooling power) for different cycle times
(7, 10, 15, 20 min) and Ads-/Desorption Duration Ratio (R = 1, 1.5, 2, 2.5, 3) at Tdes = 90 C,
Tads = Tcond = 35 C, Tev = 15C

References

1. Bosnjakovic F, Knoche KF (1997) Technische Thermodynamic Teil 2, 6.Edition. Steinkopf


Verlag, Darmstadt, Germany
2. Lucas K (2003) Thermodynamik, Die Grundgesetze der Energie- und Stoffumwandlungen,
Vierte edn. Springer Verlag, Berlin-Heidelberg, Germany
3. FprEN12309-3, 2014 E, Gas-red sorption appliances for heating and/or cooling with a net
heat input not exceeding 70 kWPart 3: Test conditions, CEN-CENELEC Management
Centre: Avenue Marnix 17, B-1000 Brussels
4. FprEN12309-7, 2014 E, Gas-red sorption appliances for heating and/or cooling with a
net heat input not exceeding 70 kWPart 7: Specic provisions for hybrid appliances,
CEN-CENELEC Management Centre: Avenue Marnix 17, B-1000 Brussels
5. Critoph RE (1988) Performance limitations of adsorption cycles for solar cooling. Sol Energy
41(1):2131
6. Critoph RE (1996) Evaluation of alternative refrigerantadsorbent pairs for refrigeration
cycles. Appl Therm Eng 16(11):891900
7. Srivastava NC, Eames IW (1998) A review of adsorbents and adsorbates in solidvapour
adsorption heat pump systems. Appl Therm Eng 18(910):707714
8. El-Sharkawy II, Kuwahara K, Saha BB, Koyama S, Ng KC (2006) Experimental investigation
of activated carbon bers/ethanol pairs for adsorption cooling system application. Appl Therm
Eng 26:859865
9. El-Sharkawy II, Uddin Kutub, Miyazaki T, Saha BB, Koyama S, Kil HS, Yoon S-H,
Miyawaki J (2015) Adsorption of ethanol onto phenol resin based adsorbents for developing
next generation cooling systems. Int J Heat Mass Transfer 81:171178
10. Breck DW (1964) Zeolite molecular sieves. Wiley, New York
11. Suzuki M (1990) Adsorption engineering. Elsevier Science Publications B.V, Amsterdam
12. Gassel A (1998) Die Adsorptionsklteanlage Betriebserfahrungen und thermodynamische
Berechnung. Klima Luft- und Kltetechnik 8:380384
13. Aristov YuI, Tokarev M, Freni A, Glaznev IS, Restuccia G (2006) Kinetics of water
adsorption on silica Fuji Davison RD. Microporous Mesoporous Mater 96(13):6571
14. Zhu R, Han B, Lin M, Yu Y (1992) Experimental investigation on an adsorption system for
producing chilled water. Int J Refrig 15(1):3134
References 31

15. Aristov YI, Tokarev MM, Cacciola G, Restuccia G (1996) Selective water sorbents for
multiple applications, 1. CaCl2 conned in mesopores of silica gel: sorption properties. React
Kinet Catal Lett 59(2):325333
16. Aristov YI, Tokarev MM, Parmon VN, Restuccia G et al (1999) New working materials for
sorption cooling/heating driven by low temperature heat: properties. In: Proceedings of
international sorption heat pump conference, Munich, Germany, pp 247254
17. Aristov YI, Restuccia G, Cacciola G, Parmon VN (2002) A family of new working materials
for solid sorption air conditioning systems. Appl Therm Eng 22:191204
18. Restuccia G, Freni A, Vasta S, Tokarev MM, Aristov YI (2002) Lab-scale tests on Selective
Water Sorbents for solid sorption air conditioning. In: Proceedings of the international sorption
heat pumps conference (ISHPC02), Shanghai, China, 2327Sept, pp 645649
19. Restuccia G, Freni A, Vasta S, Aristov YuI (2004) Selective Water Sorbents for solid sorption
chiller: experimental results and modelling. Int J Refrig 27(3):284293
20. Freni A, Russo F, Vasta S, Tokarev M, Aristov YI, Restuccia G (2004) An advanced solid
sorption chiller using SWS-1L. In: Proceeding of the 3rd international conference on heat
powered cycles, Cyprus, Oct 2004, Paper Number: 1412, pp 16
21. Simonova IA, Freni A, Restuccia G, Aristov YuI (2009) Composite silica modied by calcium
nitrate: water sorption equilibrium. Microporous Mesoporous Mater 122(13):223228
22. Freni A, Sapienza A, Glaznev IS, Aristov YuI, Restuccia G (2012) Experimental testing of a
lab-scale adsorption chiller using a novel selective water sorbent silica modied by calcium
nitrate. Int J Refrig 35:518524
23. Pons M, Grenier Ph (1987) Experimental data on a solar powered ice maker using activated
carbon and methanol adsorption pair. J Sol Energy Eng 109:303310
24. Douss N, Meunier F (1988) Effect of operating temperatures on the coefcient of performance
of active carbon-methanol systems. Heat Recovery Syst CHP 8(5):383392
25. Leite APF, Chaya MR, Horn M, Grilo MB (2003) Description and operation of two prototypes
of adsorptive solar refrigerators. In: Proceedings of the international conference on energy and
the environment, Shanghai, China, 1113 Dec 2003
26. Hu E (1998) A study of thermal decomposition of methanol in solar powered adsorption
refrigeration systems. Sol Energy 62(5):325329
27. Vasiliev LL, et al (1999) Multisalt-carbon chemical cooler for space applications. Proceedings
of international sorption heat pump conference, Munich, Germany, pp 579583
28. Vasiliev LL, Mishkinis DA, Antukh AA, Vasiliev LL Jr. (2001) Solar-gas solid sorption
refrigerator. Heat Powered Cycles 01 Paris, France, pp 7380
29. Cornstedt AF (1756) Akad Handl Stockholm 18:120130
30. Damour A (1840) Ann Mines 17:191
31. Bekkum HV, Flanigen EM, Jansen JC (1991) Introduction to zeolite science and practice.
Elsevier, Amsterdam
32. Barrer RM (1945) J Soc Chem Ind 64:130
33. Barrer RM (1948) J Chem Soc 2158
34. Ruthven DM (1984) Principles of adsorption and adsorption processes. Wiley, New York
35. Kast W (1988) Adsorption aus der Gasphase; Ingenierwissenschaftliche Grundlagen und
tecnische Verfaheren, VCH-Verlagsgesellschaft mbH
36. Wang LW, Wang RZ, Oliveira RG (2009) A review on adsorption working pairs for
refrigeration. Renew Sustain Energy Rev 13:518534
37. Ng EP, Mintova S (2008) Nanoporous materials with enhanced hydrophilicity and high water
sorption capacity. Microporous Mesoporous Mater 114:126
38. Jaenchen J, Ackermann D, Stach H (2002) Adsorption properties of aluminophosphate
molecular sievespotential applications for low temperature heat utilization. In: International
sorption heat pump conference, Shanghai, P.R. China, 2427 Sept 2002, pp 635638
39. van Heyden H, Munz G, Schnabel L, Schmidt F, Mintova S (2009) Kinetics of water
adsorption in microporous aluminophosphates for regenerative heat exchangers. Appl Therm
Eng 29:15141522
32 1 Basics of Adsorption

40. Henninger SK, Schmidt FP, Henning H-M (2010) Water adsorption characteristics of novel
materials for heat transformation applications. Appl Therm Eng 30(13):16921702
41. Newalkar BL, Jasra RV, Kamath V, Bhat SGT (1998) Sorption of water in aluminophosphate
molecular sieve AlPO4-5. Microporous Mesoporous Mater 20(13):129137
42. Kakiuchi H, Takewaki T, Takumi H, Yamazaki M, Watanabe H (2003) Adsorption heat pump
and use of adsorption material as adsorption material for adsorption heat pump, WO 02/
066910 A1
43. Kakiuchi H, Shimooka S, Iwade M, Oshima K, Yamazaki M, Terada S, Watanabe H,
Takewaki T (2005) Novel water vapor adsorbent FAM-Z01 and its applicability to an
adsorption heat pump. Kagaku Kogaku Ronbunshu 31:361364
44. Kakiuchi H, Shimooka S, Iwade M, Oshima K, Yamazaki M, Terada S, Watanabe H,
Takewaki T (2005) Water vapor adsorbent FAM-Z02 and its applicability to adsorption heat
pump. J Chem Eng Japan 31(4):273277
45. Sapienza A, Santamaria S, Frazzica A, Freni A (2011) Influence of the management strategy
and operating conditions on the performance of an adsorption chiller. Energy 36(9):55325538
46. Kohlenbach P, Rossington D, Weigand A (2008) Experimental performance results of a
FAMZ01 dessiccant wheel. In: International sorption heat pump conference, Seoul, South
Korea, 2326 Sept 2008
47. U.S. Patent 5-512-083
48. U.S. Patent 5-667-560
49. Dawoud B, Vedder U, Amer E-H, Dunne S (2007) Non-isothermal adsorption kinetics of
water vapour into a consolidated zeolite layer. Int J Heat Mass Transf 50(1112):21902199
50. Yang RT (2003) Adsorbents: fundamentals and applications. Wiley, New York
51. Jansen JC, Stbcker M, Karge HG, Weitkamp J (1994) Advanced zeolite science and
applications. In: Studies in surface science and catalysis, vol 85
52. Ksgens P, Rose M, Senkovska I, Frode H, Henschel A, Siegle S, Kaskel S (2009)
Characterization of metal-organic frameworks by water adsorption. Microporous Mesoporous
Mater 120(3):325330
53. Aristov YuI (2007) Novel materials for adsorptive heat pumping and storage: screening and
nanotailoring of sorption properties. J Chem Eng Jpn 40(13):12421251
54. Henninger SK, Jeremias F, Kummer H, Janiak C (2012) MOFs for use in adsorption heat
pump processes. Eur J Inorg Chem 16:26252634
55. Jaenchen J, Ackermann D, Weiler E, Stach H, Brosick W (2005) Calorimetric investigation on
zeolites, AlPO4s and CaCl2 impregnated attapulgite for thermochemical storage of heat.
Thermochim Acta 434:3741
56. Chen H, Cui Q, Tang Y, Chen X, Yao H (2008) Attapulgite based LiCl composite adsorbents
for cooling and air conditioning applications. Appl Therm Eng 28:21872193
57. Tokarev M, Gordeeva L, Romannikov V, Glaznev I, Aristov YI (2002) New composite
sorbent CaCl2 in mesopores for sorption cooling/heating. Int J Therm Sci 41(5):470474
58. Zhenyan L, Yunzhuang L, Jiaxin Z (1998) Zeolite-active carbon compound adsorbent and its
use in adsorption solar cooling tube. Solar En Mat solar Cells 52:4553
59. Lang R, Dawoud B, Miltkau T, Stricker M (2002) Sorption heat pump. European Patent ofce,
EP1 178 269 B1
60. Westerfeld T (1996) Numerische Untersuchung einer periodisch arbeitenden
Adsorptionswrmepumpe. Ph.D. thesis, RWTH Aachen University, Germany, Shaker Verlag
61. Lang R, Roth M, Stricker M, Westerfekd T (1999) Development of a modular zeolite-water
heat pump, vol 35. In: Heat and mass transfer. Springer, Berlin, pp 229234
62. Dawoud B (1999) Thermische und Kalorische Stoffdaten des Stoffsystems Zeolith MgNaA-
Wasser. Ph.D. thesis, Aachen University of Technology (RWTH-Aachen), Germany
63. Meunier F (1999) Adsorption heat pump technology: possibilities and limits. International
sorption heat pump conference (ISHPC99), Munich, p. 2535
64. Pons M, Poyelle F (2000) Adsorptive machines with advanced cycles for heat pumping or
cooling applications. Int J Refrig 22(1):2737
References 33

65. Critoph RE (1999) Forced convection adsorption cycle with packed bed heat regeneration. Int
J Refrig 22(1):3846
66. Alam KCA, Khan MZI, Uyun AS, Hamamoto Y, Akisawa A, Kashiwagi T (2007)
Experimental study of a low temperature heat driven re-heat two-stage adsorption chiller. Appl
Therm Eng 27(10):16861692
67. Aristov YuI, Sapienza A, Ovoshchnikov DS, Freni A, Restuccia G (2012) Reallocation of
adsorption and desorption times for optimisation of cooling cycles. Int J Refrig 35(3):525531
68. Glaznev IS, Aristov YI (2010) The effect of cycle boundary conditions and adsorbent grain
size on the water sorption dynamics in adsorption chillers. Int J Heat Mass Transfer 53
69. Uyun AS, Miyazaki T, Ueda Y, Akisawa A (2009) Experimental investigation of a three-bed
adsorption refrigeration chiller employing an advanced mass recovery cycle. Energies 2
(4):11701191
Chapter 2
Adsorption Heat Exchangers

2.1 Ideal Adsorber Design Requirements

As described in Chap. 1, one of the major adsorber heat exchanger design


requirements is the reduction of the ratio between the heat capacity of the adsorber
heat exchanger material as well as its heat transfer mediums holdup and the heat
capacity of the applied adsorbent. This results in enhancing the coefcient of
performance of the adsorption heat pump and increases the obtainable specic
heating or cooling power.
Beside low cost, manufacturability including a reasonable quality assurance
procedure, there are two more sets of adsorber design requirements; namely,
dynamic performance and reliability design requirements.
The dynamic performance design requirements include
(a) High overall heat transfer coefcient between the heat transfer medium and
the adsorbent.
(b) High overall (inter- and intra-particle) mass diffusion coefcient.
(c) Keeping the refrigerant vapour velocities as low as possible (<50 m/s) in the
connecting ducts between adsorber and evaporator or condenser. This implies
the reduction of pressure losses, most specically during the adsorption-
evaporation phase.
(d) In case non-return valves with floating elements are incorporated, the required
force to open those valves and, consequently, the required opening pressure
shall be minimized.
The reliability design requirements are:
1. High stability against hydrothermal aging.
2. High mechanical stability against loading, transport and operating conditions.
3. No corrosion potential under loading, transport and operating conditions.
4. No release of inert gases upon repeated adsorptiondesorption cycling.

The Author(s) 2015 35


A. Freni et al., Characterization of Zeolite-based Coatings
for Adsorption Heat Pumps, SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-09327-7_2
36 2 Adsorption Heat Exchangers

2.2 Overview on Adsorber Developments for Adsorption


Heat Pumps

Considering the methodology of bringing the adsorbent on the outer surface of the
adsorber heat exchanger, the following four congurations have been identied so
far:
Consolidated
Loose grains
In situ crystallized coatings
Binder-based coatings
This section will focus on the rst three technologies, while the binder-based
technique, to whom this book is dedicated, will be described in more details in the
section.
In general, the main drawbacks to overcome in the development of more efcient
adsorbers are the intrinsic low heat and mass transfer efciencies showed by these
components. This is always true when the vapour pressure inside the chamber is
noticeably lower than the atmospheric pressure (e.g. refrigerants like water, ethanol
and methanol), while the mass transfer resistance becomes negligible where
working fluids at high pressure are employed (e.g. ammonia and HFCs). Since the
present book is focused on zeolite-water as a working pair, the state of the art of its
adsorber heat exchanger designs will be reviewed.
Basically, the development of thermally efcient adsorbent beds has been carried
out at two levels of activities:
(i) small-scale adsorber concepts, whose thermo-physical parameters as well as
dynamic performance are experimentally evaluated to verify the achievable
advancement, beyond the state-of-the-art, in heat and mass transfer efciency.
(ii) full scale adsorbers, whose performance, in terms of specic power (massive
and volumetric) and COP, are experimentally evaluated, in order to identify
the most efcient congurations.
In the following, the evolution of adsorber concepts over the years is reported,
by analyzing some of the most innovative researches carried out.
The rst proposed idea to enhance the poor heat transfer properties of an adsorber
was the development of the consolidated adsorbent bed. Early studies proposed the
realization of consolidated thick beds with enhanced thermal conductivity.
Guilleminot et al. [1] realized consolidated cylindrical adsorbents made of
compressed zeolite powder (NaX or NaA) and metal foams (copper or nickel). The
resulting equivalent thermal conductivity was in the range between 1.7 (Ni foam) and
8.3 W m1 K1 (Cu Foam), which is clearly superior to that of a granular adsorbent
bed. However, with this arrangement the mass transfer becomes the limiting factor,
due to the high thickness and the low gas permeability of the bed (see Table 2.1).
In order to reduce the mass transfer resistance, Lang et al. [2] prepared aconsolidated
adsorbent consisting of zeolite powder integrated into an aluminium honeycomb-like
2.2 Overview on Adsorber Developments for Adsorption Heat Pumps 37

structure, adding pore forming materials such as melanine or tartaric acid. Indeed, the
organic components were completely removed during the drying process of the
zeolite increasing the macro-pore fraction of the material. The same concept of using
pore-forming additives to improve mass transfer was adopted by Gordeeva et al. in
[3], that realized consolidated layers consisting of hydrophobic zeolite CBV901 as
methanol adsorbent, aluminium hydroxide as binder and ammonium bicarbonate as
pore-forming additive.
Generally, the concept of forming a consolidated layer of adsorbent was
potentially interesting, but the following issues were found to negatively affect the
adsorption heat pump performance:
(i) high amount of binder/additive in the adsorbent formulation, affecting the
overall adsorption capacity
(ii) too high density of the adsorbent layer due to the compression process,
affecting the mass transfer efciency
(iii) too thick adsorbent layers (typically 210 mm) affecting the vapour diffusion
through the layer
(iv) shaping of the adsorbent restricted to planar or cylindrical conguration,
which limited the practical utilization in real heat exchangers
Table 2.1 summarizes some of the main activities carried out in the eld of
consolidated bed developments.
The loose grains adsorbent bed represents the opposite concept of the consoli-
dated one. In this case, grains or pellets are embedded inside a heat exchanger
without any kind of additional binder. Such a conguration allows to reach optimal
mass transfer properties, thanks to the high void fraction existing between the
grains. Clearly, on the opposite side, the heat transfer results to be strongly limited,
due to the point contacts between heat exchanger and grains as well as to the
thermal insulating properties of the adsorbent material itself. Activities devoted to
the development of this adsorber concept have been mainly oriented towards the
optimization of the adsorbent grains integration within the heat exchanger by the
identication of optimal grain sizes as well as the employment of highly efcient
heat exchangers. Figure 2.1 reports some concepts of loose grains adsorbers.
Aristov et al. [12] have performed a deep analysis on the heat and mass transfer
phenomena affecting the performance of loose grains adsorbers. They analysed the
influence of heat transfer surface to adsorbent mass ratio, as well as the grain size,
on the dynamic behaviour of adsorber congurations. Starting from the simple flat
plate adsorber, they veried that, under certain conditions, the kinetics of isobaric
ad-/desorption is invariant with respect to the ratio (S/m) heat transfer surface/
adsorbent mass. In particular, they identied two different regimes: lumped
regime and grain sensitive regime. The former one is established under certain
combinations of grain sizes and S/m ratios. Under these conditions the dynamic
performance is affected by the heat transfer efciency, while the mass transfer effect
is negligible. The latter one, on the contrary, is realized for large adsorbent grain
size. In these conditions, the kinetic is mainly affected by intra-grain diffusion
resistance, while the heat transfer plays a minor role. Similar conclusions were
38

Table 2.1 Summarizes some of the main activities carried out on in the eld of the consolidated bed development
Authors Adsorbent conguration/ Eq. thermal Thickness, Wall heat Binder content or Gas permeability, m2
adsorbate conductivity, mm transfer coeff. mads/madditive
W m1 K1 W m2 K1
Guilleminot Consolidated NaA or 13X 1.7 (Ni foam) 20 110 (Ni foam) 35 wt% 1013
et al. [1] zeolite powder in metal foam/ 8.3 (Cu foam) 180 (Cu foam)
water
Guilleminot Consolidated 13X zeolite 417 75 2003000 30 wt% 1013
et al. [4] powder in expanded natural
graphite/water
Wang et al. [5] Consolidated activated carbon 2.5 12 n.a. mads/mGNE = 0.52 1012
in expanded natural graphite/
ammonia
Tamainot- Consolidated monolithic 0.340.44 10 350800 0 1014
Telto and activated carbons/ammonia
Critoph [6]
Pino et al. [7] Consolidated 4A zeolite 0.120.43 30 n.a. 2040 wt% n.a.
powder with different binders
and additives/water
Restuccia Consolidated 4A zeolite 0.43 2 120 40 wt% 1012
et al. [8] powder and Al(OH3) on
aluminium sheets/water
Lang et al. [2] Consolidated zeolite powder in 0.4 3 1000 25 wt% 109
aluminium honeycomb matrix/
water
Gordeeva Consolidated hydrophobic 0.4 5 55 2030 wt% 1012
et al. [3] zeolite powder and Al(OH3)/
methanol
2 Adsorption Heat Exchangers
2.2 Overview on Adsorber Developments for Adsorption Heat Pumps 39

Fig. 2.1 Schematic of the loose grains arrangement within a nned flat-tube heat exchanger,
proposed in [9], on the left hand side. Example of the integration of loose grains within a nned
tube heat exchanger [10], on the right hand side. Schematic of flat plate loose grains adsorber
concept, with different layers of adsorbent grains [11]

recently drawn for more complex congurations by Santamaria et al. [9] and
Gordeeva et al. [13].
In general, the main drawbacks related to the employment of loose grains
congurations are the following:
(i) low efciency of secondary ns in extended surface heat exchangers, which
limit the overall heat transfer coefcient.
(ii) possible pressure drops along the packed grains, which limits the vapour
diffusion and so the kinetic performance
(iii) need of metallic net to keep the material inside the heat exchanger, which
increases the complexity of the realization and can add signicant mass
transfer resistance.
Table 2.2 summarizes some of the most recent activities on loose grains ad-
sorbers development, with particular attention paid to their kinetic performance.
The direct synthesis of thin zeolite coatings on heat exchanger walls has been
proposed as another way to realize efcient adsorbers. Some authors consider this
technology as a possible solution to achieve high heat and mass transfer efciency
through the adsorber. Indeed, the heat transfer across the interface strongly increases
due to the nearly perfect contact between the heat exchanger surface and the zeolite,
40 2 Adsorption Heat Exchangers

Table 2.2 Relevant activities on the development of loose grains adsorbers


Authors Adsorber conguration Working pair Main evaluated
parameters
(kW/kgads)
Aristov et al. [12] Loose adsorbent grains over SWS-1 Lwater SCP = 0.33.7
flat plate HEX
Gordeeva and Loose adsorbent grains over Composite LiCl/ SCP = 2.57.9
Aristov [14] flat plate HEX SiO2methanol
Gordeeva and Loose adsorbent grains over Activated carbon SCP = 0.211.5
Aristov [11] flat plate HEX ACM-35.4methanol
Santamaria Loose adsorbent grains within AQSOA FAMZ02 SCP = 1.32.3
et al. [9] nned flat-tube HEX water
Gordeeva Loose adsorbent grains within Composite LiBr/ SCP = 0.40.9
et al. [13] nned flat-tube HEX SiO2ethanol

while the mass transfer resistance is substantially reduced due to the very thin zeolite
layer. Tatlier and Erdem-enatalar [15] and Bonaccorsi and Proverbio [16] have
developed methods for direct hydrothermal synthesis of zeolite NaA on stainless
steel plates. Bonaccorsi et al. [17] improved the synthesis method to Y and SAPO
zeolites over stainless steel and aluminum plates. Generally, homogeneous zeolite
coating layers in the range of 20150 m in thickness rmly bonded over the metal
substrate were obtained. More recently, this technique has been applied to directly
crystallize the zeolite over metal or graphite foams and metal bers, giving a relevant
increase of surface area and more convenient zeolite-to-metal-mass ratio. Bonaccorsi
et al. [18] obtained growth of zeolite NaA layers on open-cell copper foam by in situ
hydrothermal synthesis processes. The copper foam surface was totally coated with a
compact layer of zeolite NaA crystals rmly bonded to the substrate. The quantity of
zeolite coating deposited was valuated to be 17 % in (dry) weight. Fuldner et al. [19],
proposed and numerically studied adsorbent layers (24 mm thick) consisting of
SAPO34 directly synthesized on sintered aluminium bers by a partial support
transformation method developed in [20]. This structure presented high porosity
(approx. 70 %) and enhanced equivalent adsorbent thermal conductivity
(8 W m1 K1).
In general, even though zeolite direct crystallization offers the clear advantage of
a high coating stability and a perfect thermal contact between zeolitic crystals and
metal surface, it is characterized by some drawbacks:
(i) production process complex and expensive, with synthesis conditions that can
be rather drastic (e.g. 20 bar and 200 C in autoclave for SAPOs zeolite)
(ii) necessity to increase the heat exchanger surface in order to reach the needed
adsorbent mass, leading to the decrease of the zeolite/metal mass ratio and a
consequent reduction of the adsorption machine efciency, COP
Table 2.3 summarizes some of the main features of the developed directly
synthesized adsorber concepts.
Table 2.3 Relevant activities on the development of directly synthesized adsorbers
Authors Adsorbent conguration/ Eq. thermal Thickness, Wall heat Binder content or Gas permeability, m2
adsorbate conductivity, mm transfer mads/madditive
W m1 K1 coeff.
W m2 K1
Tatlier and Erdem- Synthesized NaA on 0.2 0.070.15 >1000 0 108
enatalar [15] stainless steel plates/
water
Bonaccorsi and Synthesized NaA, Y, 0.15 0.020.08 >1000 0 108
Proverbio [16], SAPO34 on SS and
Bonaccorsi et al. [17] Aluminum plates/water
Bonaccorsi et al. [18] Synthesized NaA on 27 5 >1000 mads/mmetal = n.a. 108
copper foam/water
Fuldner et al. [19] Synthesized SAPO34 on 8 24 >1000 mads/mmetal = 1.31.5 n.a.
aluminum bers/water
Bonaccorsi et al. [21] Synthesized SAPO on 14 25 >1000 mads/mmetal = n.a. 108
2.2 Overview on Adsorber Developments for Adsorption Heat Pumps

aluminum foam/water
41
42 2 Adsorption Heat Exchangers

2.3 Introduction to the Concept of Coated Adsorbers

The idea of developing binder-based coated adsorbers originated by the need to


overcome the drawbacks showed by the techniques reported in the previous
paragraph:
high heat and mass transfer resistance, respectively for loose grains and con-
solidated congurations,
complex and expensive preparation procedure in the case of directly synthesized
adsorbers.
Generally, coating methods (e.g. dip coating, spray coating) offer the potential
benet of mild reaction conditions, easily variable coating thickness (up to 0.5 mm)
and, in principle, easier implementation in serial production lines. Adsorbent
coatings can be realized by gluing the adsorbent grains on the metal surface or by
dip coating technique where the metal substrate is immersed into a liquid solution
made of active powder and an organic (e.g. resins) or inorganic binder (e.g. clays,
etc.). Usually, a proper thermal treatment is carried out to remove the excess
solvent, so obtaining a compact adsorbent layer.
According to the former approach, Jaeschke and Wolf from Sortech AG, pat-
ented a method for realizing a silica gel coating on the heat exchanger surface using
a resin as gluing agent [22], as showed in Fig. 2.2.
Dawoud et al. [23] realized and numerically studied an adsorbent coating based
on the special Y-type zeolite DDZ70, attached on the metal substrate by a proper
adhesive. A wall heat transfer coefcient of 230 W m2 K1 has been estimated.
Following the dip-coating approach, Restuccia et al. [24], developed and
numerically studied a NaA zeolite coating (4 mm thick) on a stainless steel tube.
Alumina gel precipitated in situ was used as binder. This conguration of the
adsorbent bed allowed to obtain a slight increase of the thermal conductivity of the
adsorbent and a strong rise of the metal/adsorbent wall heat transfer coefcient.
Van Hayden et al. [25] developed ALPO18-coated aluminum plates using
Polyvinyl alcohol as binder. An optimal layer thickness of 0.2 mm was found.
Waszkiewicz et al. [26] developed a heat exchanger with coated anular ns of
CBV901 hydrophobic zeolite for methanol adsorption. Cellulose methyl ether was

Fig. 2.2 Three examples of dip coated adsorbers, employing a binder: silica gel coating, by
Sortech [22], on the left hand side, AQSOA Z01 coating, by Mitsubishi [27], central, SAPO 34
coating, by CNR ITAE and University of Messina [28], on the right hand side
2.3 Introduction to the Concept of Coated Adsorbers 43

used as binder and a noticeable increase in thermal conductivity and wall heat
transfer coefcient was found.
Recently, Okamoto et al. [27] presented a coating technique based on an organic
binder for the AQSOA materials (Z01, Z02, Z05), focusing mainly on the perfor-
mance achievable by the AQSOA Z01. Figure 2.2 reports a picture of the coating
realized. The procedure allowed to coat lamellas of a heat exchanger with an average
coating thickness of 0.3 mm. Interestingly, they found a clear increasing in thermal
conductivity passing from powders (0.113 W/m K) to coating (0.36 W/m K).
Moreover they proved the physical durability of the coating itself by performing up
to 200,000 ad/desorption cycles.
In [28] Freni et al. introduced a new coating composition, employing SAPO 34
powder as adsorbent and silane as binder, see Fig. 2.2. The complete thermo-
physical characterization was presented, conrming the goodness of the developed
technique. Moreover, kinetic characterization of small scale coated samples with a
thickness of 0.3 mm, by means of the Isothermal Differential Steps Method, was
performed. The results showed only slight lowering in kinetic performance, if
compared to pure adsorbent material, which conrmed that the presence of the
binder does not heavily affect the water vapour transport through the layer.
Another coating methodology proposed by Freni et al. [29] consists in the
employment of an inorganic clay as binder for AQSOA Z02 adsorbent material.
Also in this case, the coated samples were prepared by dip coating technique,
reaching thicknesses variable between 0.2 and 0.8 mm. Their hydrothermal as well
as mechanical stability were tested, proving sufcient performance, even if the layer
resulted to be not strong enough.
A patent by Sauer et al. [30], dened a procedure for realization of coated
adsorbers. In this case, they proposed to employ inorganic binders, in order to
reduce the thermal degradation of the layer during operation as well as to avoid
releasing of non-condensable gases. Moreover, they described the benecial effects
related to the employment of bres (carbon, carbon nanotubes), which allows to
increase the elasticity and strength of the layer, thus avoiding shrinkage cracks and
increasing the overall thermal conductivity of the coating itself.
In general, the advantages related to the employment of dip coating methods are:
(i) possibility of easily coat complex heat exchanger geometries with an adsor-
bent layer, maintaining a uniform thickness
(ii) large coating thicknesses, up to 0.5 mm, obtainable by properly changing the
preparative parameters (e.g. slurry viscosity, temperature)
(iii) reduced heat and mass transfer resistances, thanks to the good contact between
adsorbent layer and heat exchanger surface and to the controlled thickness
Nevertheless, still some issues have to be systematically analyzed and solved in
order to obtain reliable and marketable products, like:
(i) the poor mechanical strength of the coating layer,
(ii) possible adsorbent pores occlusion by the employed binder,
(iii) possible production of volatile compounds from organic binders.
44 2 Adsorption Heat Exchangers

2.4 Methodologies for the Experimental Characterization


of Zeolites

Structural and chemiophysical properties of zeolites are commonly investigated by


a variety of characterization techniques normally adopted for porous solids.
A general description of most common characterization technique is reported
elsewhere [31]. Many of such techniques can be also applied for chemiophysical
characterization of adsorbent coatings based on zeolite as active material, however
taking into account that organic compounds are often employed as binding agent.
Herein below, a short overview of the most common techniques specically applied
for zeolite characterization is reported.
Infrared and Raman molecular spectroscopy are major tools for obtaining
information regarding the zeolite structure. Indeed, each chemical structure has
specic frequencies associated with internal vibrations of group of atoms, which
yield unique IR spectra, i.e. ngerprints for each zeolite.
Magnetic resonance (NMR) spectroscopy is a useful technique to provide
information on the state of the water molecules in hydrated zeolites and can be
applied to measure diffusion mechanisms in the porous media, as widely discussed
in [32].
X-Ray Analysis (XRD) has been widely used to obtain information about the
structure, composition and phase purity of crystalline zeolites. Each zeolite exhibits
a typical XRD pattern, so that several zeolites that belong to various structural
groups have been precisely identify and classied. The International Zeolite
Association (IZA) published online a wide database, which provides structural
information on all of the Zeolite Framework Types [33].
X-Ray Fluorescence (XRF) is another useful and non-destructive spectroscopic
technique for quantitative analysis of elemental composition of zeolites.
Atomic Absorption (AA) and Inductively Coupled Plasma (ICP) spectroscopy are
other powerful tools for elemental analysis in zeolites.
Scanning Electron Microscopy (SEM) is the preferred microscopy method to
study the habit, crystal growth and surface topography of zeolitic crystals. SEM are
usually equipped with electron probe microanalysis for X-ray spectroscopy.
Differential Thermal Analysis (DTA) coupled with Thermogravimetric Analysis
(TGA) are useful methods for thermal characterization of zeolites and are usually
used to study the dehydration behaviour of the zeolite which may involve major
structural changes, as reported in [34].
X-Ray Photoelectron Spectroscopy (XPS) is a surface-sensitive spectroscopic
tool that can be used to collect information about the chemical state and the con-
centration of elements comprising the surface layers of the zeolite.
Physical gas adsorption techniques (Langmuir, BET methods) are used to
determine the pore volume, pore size distribution and surface area of solids by
physical gas (e.g. nitrogen, helium, argon) adsorption at low temperature. No
universal adsorption equations exists for zeolites. Since many zeolites exhibit the
2.4 Methodologies for the Experimental Characterization of Zeolites 45

classical type I isotherm according to the IUPAC classication, the monolayer


adsorption Langmuir model has been successfully applied. The classical BET
theory, which was developed for multilayer adsorption, can hardly be applied on
zeolites, as reported in [35].
Equilibrium adsorption curves. This is the most important characterization for
the evaluation of the most suitable adsorbent for a practical application. The amount
of a given gas or vapour which is adsorbed by a dehydrated crystalline zeolite
depends in the equilibrium pressure, on the temperature, the nature of gas or vapour
and the nature of pores in the zeolite crystal. Considerable experimental data for
adsorption of gases and vapours on several zeolites are widely available. Usually,
equilibrium data are presented in tabular form or graphically summarized with
isothermal, isobaric or isosteric representation. Adsorption equilibrium curves are
generally determined by volumetric or gravimetric method. The volumetric method
determines the quantity of gas or vapour present in the system by measuring the
pressure, volume and temperature. After exposing the dry adsorbent to a quantity of
gas in the closed system, the adsorbed amount is indirectly determined from
pressure, temperature and volume when equilibrium is reached. The gravimetric
method directly measure the amount of gas or vapour adsorbed by weighing the
sample in a closed system on a balance. McBain adsorption balance using a quarz
spring type and Cahn balance using a force-to-current converter are the most
common types of microbalance adopted. In the gravimetric method, the buoyancy
correction must normally be applied, which involves the determination of the
volume occupied by sample.
Recently, Henninger et al. proposed in [36] a standardized procedure for
determination of the water adsorption equilibrium on adsorption materials to be
used in an adsorption heat pump or chiller. The procedure consists of a pre-treat-
ment of the sample under continuous evacuation (vacuum level: 1e4 kPa). The
optimal sample pre-treatment temperature should be selected according to the
following classication.
1. Strongly hydrophilic zeolites (4A, 13X): pre-treatment T = 300 C.
2. Hydrophilic aluminosilicates (NaY): pre-treatment T = 200 C
3. Hydrophobic aluminosilicates (silicalites, ZSM5): T = 150 C
4. Aluminophospates (AlPO, SAPO): T = 150 C
5. Others (silica gels, activated carbons): T = 150 C
The sample is heated starting from ambient conditions with a heating rate of 1 K/
min followed by an isothermal drying step for another 8 h. In the following step,
isobar measurement at a water vapour pressure of 1.2 and 5.6 kPa takes place. The
selection of the two pressure levels is motivated with respect to the possible
applications. The pressure level of 1.2 kPa corresponds to an evaporation tem-
perature of 10 C, which marks a useful temperature level for cooling applications.
The second pressure level of 5.6 kPa corresponds to 35 C, which either marks the
temperature where heat can be rejected (cooling application) or can be used for low
46 2 Adsorption Heat Exchangers

Fig. 2.3 Proposed measurement procedure, including sample pre-treatment and the rst isobaric
step

temperature heating (heat pumping application). For each pressure level the sample
temperature is varied in 5 or 10 K steps between 150 and 40 C (for 5.6 kPa 9 or
20C (for 1.2 kPa) respectively. The water uptake is calculated as w
mpH2 O ; T =m0 where m0 is the dry weight of the sample. In addition at least one
adsorption and desorption measurements should be performed in order to detect
possible hysteresis effects. Schematic representation of the procedure is reported in
Fig. 2.3.
Heat of adsorption. Adsorption process is exothermic, involving heat produc-
tion. The heat of adsorption is commonly derived from the isosters by the use of the
well-known ClausiusClapeyron equation, as reported in [35].
Adsorption rate. Isothermal Differential Step (IDS) method is considered as the
most reliable for the kinetic characterization of the adsorbent material itself. Indeed,
such thermogravimetric approach allows the identication of several kinetic
parameters like the diffusion constant and the adsorbate diffusivity, as the analytical
solution under these testing conditions is well known since long time [37]. An
example of application of the IDS method for studying the kinetics of water
adsorption on silica Fuji Davison RD under typical adsorption cooling operating
conditions is reported in [38].
Hydrothermal stability. Plenty of studies about thermal stability of zeolites for
various applications (gas separation, drying, etc.) are available in literature, while
only a limited number of studies were dedicated to heat transformation applications.
As far as the authors know, rst systematic study of hydrothermal stability on 22
commercial zeolites for heat pumping applications was carried out by Rothmeyer
[39] in his PhD thesis (1985). Other authors determined the stability of adsorbents
for heat storage or dessicant open-cycles systems [4042].
Mechanical stability. Generally, few information are available concerning the
mechanical properties of adsorbents. A hardness of crystals of zeolites of
100400 HV (Vickers hardness) and a crush strength of 2.510 kg for commercial
pellets or extrudates are reported in [35].
2.5 Experimental Methodologies to Verify the Performance of Adsorbers 47

2.5 Experimental Methodologies to Verify the Performance


of Adsorbers

One of the key points in the development of advanced adsorber concepts is related
to the experimental methodologies to be employed to verify their performance. In
fact, in the past, several studies were oriented towards the adsorbent material
characterization and full scale prototype testing, while, only during last years a
systematic experimental study of adsorber concepts has been developed. Such an
evolution has been guaranteed by the design and realization of proper setups.
Usually, they are mainly focused on the analysis of the adsorber dynamic behav-
iour, thus allowing the evaluation of the achievable specic power. Nevertheless,
also test rigs dedicated to the characterization of full-scale adsorbers, both in terms
of efciency (COP) and cooling power, have been developed.
In general, the experimental methodologies for adsorber performance evaluation
can be distinguished considering the scale of the tested adsorber: small scale, which
allow to test small but representative piece of adsorber concepts, and full scale,
where the entire adsorber can be tested under almost-realistic boundary conditions.
In the following, the main apparatuses realized for this purpose are described,
putting in evidence the working principles and peculiarities of each one.

2.5.1 Small Scale Large Pressure Jump (LPJ)

The rst version of Large Pressure Jump (LPJ) kinetic characterization of small
scale adsorbers was proposed by Dawoud et al. in [43]. The adsorption/desorption
process is triggered by a sudden jump/drop in absolute vapour pressure over the
adsorber kept at almost constant temperature.
The adsorption kinetics test-rig is depicted in Fig. 2.4. It comprises mainly two
compartments. The rst compartment is the measuring cell, in which the small scale
adsorber is placed .The temperature of the holding surface of the sample is adjusted
and controlled using an oil circuit coupled to the circulating thermal bath 2. The
second compartment is a constant volume vapor vessel. A water circuit coupled to
the circulating thermal bath 1 allows the management of the temperature of this
vessel.
The typical test procedure is the following: regeneration of the adsorbent mate-
rial, heating up to 150 C the measuring cell under continuous evacuation for 2 h.
Subsequently, establishing of the adsorption temperature inside the measuring cell
and charging of the water vapour inside the vapour vessel, by means of the steam
generator. Once reached the initial conditions, the connection between vapour vessel
and measuring chamber is opened and the adsorption phase starts, followed by the
pressure decreasing inside the system. The pressure decreasing recorded is then used
to determine the amount of vapour adsorbed.
48 2 Adsorption Heat Exchangers

Fig. 2.4 Test-rig for measuring the kinetics of small scale adsorbers, according to the LPJ
procedure, as reported in [43]

Such an approach, is useful to give a comparative estimation of the achievable


dynamic performance of adsorber congurations, in terms of characteristic times of
ad/desorption. Nevertheless, since the working boundary conditions are completely
different from the ones typical of adsorption machines (i.e. temperature jump/drop
over the adsorber), it doesnt allow to directly evaluate achievable performance, like
specic powers.

2.5.2 Small Scale Large Temperature Jump (LTJ)

In order to overcome the main limitation of the LPJ approach, the Large Tem-
perature Jump (LTJ) technique, in its volumetric version, was developed by Aristov
et al. [44]. Clearly, in this case the adsorption/desorption stages are promoted by a
sudden drop/jump in temperature over the adsorber under almost-isobaric condi-
tions, thus closely replicating the active phases of an adsorption machine.
The kinetic setup, depicted in Fig. 2.5, has a structure similar to the LPJ one. The
main difference is related to the introduction of a second heat source in the oil loop
connected to the measuring chamber, which allows to x the nal temperature for
the adsorption or desorption phases.
2.5 Experimental Methodologies to Verify the Performance of Adsorbers 49

Fig. 2.5 Test-rig for measuring the kinetics of small scale adsorbers, according to the volumetric
LTJ procedure, as reported in [44]

In this case, after the regeneration of the adsorber at high temperature and under
continuous evacuation, the vapour vessel and measuring cell are charged with water
vapour, from the steam generator, up to the required starting pressure for the
adsorption process (i.e. evaporator pressure). Once that the system is in thermal
equilibrium, a sudden drop in temperature is applied to the adsorber. The adsorption
process driven by the sample cooling started, results in reducing the vapour pres-
sure inside the system with time. This decrease did not exceed 23 mbar, which is
quite typical for adsorption heat pumps, as the evaporator could not be designed to
keep the pressure perfectly constant, most specically, at the beginning of the
adsorption process. Accordingly, the adsorption process can be considered as a
quasi-isobaric one. The pressure variation, being measured with two pressure
transducers P1 and P2 is used to determine the amount of the adsorbed water
vapour on the sorbent sample.
This experimental setup allows to evaluate achievable dynamic performance, in
terms of specic power, of small scale adsorber congurations. Still, some limitations
are present, namely, the possibility of testing only flat plate adsorber congurations
and the reduced quantity of adsorbent that can be loaded inside the measuring cell, in
order to keep the pressure decreasing within the range of 23 mbar, to consider the
test almost isobaric.
50 2 Adsorption Heat Exchangers

Accordingly, more recently, a new version of the LTJ approach has been devel-
oped by Sapienza et al. [45]. In this case, the evolutions of the adsorption/desorption
phases are directly followed by measuring the weight of the adsorber. This is why it is
referred as Gravimetric Large Temperature Jump (G-LTJ). It allows to overcome the
limits showed by the volumetric LTJ version. Indeed, it can be employed to test more
complex small scale adsorber congurations, as reported in [9] and in [13], while,
virtually, there is no limitation in adsorbent amount to be tested.
A LTJ-like approach has been developed also by Wittstadt et al. [46] to test full
scale adsorbers. Also in this case the adsorption/desorption phases are followed
directly weighting the adsorber by means of a proper identied balance. The
pressure inside the testing chamber is kept constant by means of an oversized
evaporator/condenser.

2.5.3 Full Scale Experimental Setup for Cycling


Characterization

A different approach for the characterization of full scale adsorbers have been
developed at the CNR ITAE laboratory [10]. It consists in the evaluation of both
efciency (i.e. COP) and specic powers by the energy balance of each component
of the test rig: evaporator, condenser and the adsorber itself. In particular, the
amount of heating and cooling energy provided or removed from a single com-
ponent is calculated considering the energy transferred between the external heat
transfer fluids and the components during the thermodynamic cycle described.
The experimental test facility as well as the layout of the test rig are represented
in Fig. 2.6. The system consists of two separated and interconnected units: an
intermittent single bed adsorption chiller able to deliver a maximum cooling power
of about 1 kW and a test bench. The lab-scale adsorption chiller consists of the

Fig. 2.6 Experimental test-rig for full scale cycling characterization, on the left-hand side, and
schematic layout, on the right hand side [10]
2.5 Experimental Methodologies to Verify the Performance of Adsorbers 51

single adsorber connected to a single evaporator and a single condenser operating


with an intermittent mode. The vacuum chamber, where the adsorber is placed, was
designed to allow the test of different kinds of adsorbers. The evaporator and the
condenser consist of two vacuum vessels, in which the thermal energy is supplied/
removed by the heat transfer fluid flowing into nned-tubes copper coils. In
Fig. 2.6, also the test bench, which supplies the external heating and cooling energy
to the adsorption chiller and guarantees the automatic control and management of
the system is represented. A similar approach has been employed by Dawoud [47]
to test adsorbers for heat pumping applications.
The present technique is of primary interest, as it allows to completely charac-
terize an adsorber, evaluating both efciency as well as achievable power density.
Nevertheless, since it relies on the indirect measure of external heat carrier fluids,
a high accuracy in temperatures and flow rates measurements is requested, in order
to keep the uncertainty as low as possible.

References

1. Guilleminot JJ, Choisier A, Chaflen JB, Nicolas S, Reymoney JL (1993) Heat transfer
intensication in xed bed adsorbers. Heat Recovery Syst CHP 13:297300
2. Lang R, Westerfeld T, Gerlich A, Knoche KF (1996) Enhancement of heat and mass transfer
in compact zeolite layers. Adsorption 2:121132
3. Gordeeva LG, Aristov YuI, Freni A, Restuccia G (2002) Preparation of zeolite layers with
enhanced mass transfer properties for adsorption air conditioning. Int Sorp Heat Pump Conf
Shanghai 2427:625629
4. Guilleminot JJ, Chaflen JB, Choisie A (1994) Heat and mass transfer characteristics of
composites for adsorption heat pumps. In: Proceedings of the international absorption heat
pump conference, AES-vol 31, pp. 401406. New Orleans, Louisiana
5. Wang LW, Tamainot-Telto Z, Thorpe R, Critoph RE, Metcalf SJ, Wang RZ (2011) Study of
thermal conductivity, permeability, and adsorption performance of consolidated composite
activated carbon adsorbent for refrigeration. Renew Energy 36:20622066
6. Tamainot-Telto Z, Critoph RE (2001) Monolithic carbon for sorption refrigeration and heat
pump applications. Appl Therm Eng 21:3752
7. Pino L, Aristov Yu, Cacciola G, Restuccia G (1996) Composite materials based on zeolite 4A
for adsorption heat pumps. Adsorption 3:3340
8. Restuccia G, Freni A, Cacciola G (1999) Adsorption beds of zeolite on aluminium sheets. In:
International sorption heat pump conference, 2426 March, Munich, Germany, pp. 343347.
ISBN 3-7846-1220-2
9. Santamaria S, Sapienza A, Frazzica A, Freni A, Girnik IS, Aristov YuI (2014) Water
adsorption dynamics on representative pieces of real adsorbers for adsorptive chillers. Appl
Energy 134:1119
10. Sapienza A, Santamaria S, Frazzica A, Freni A (2011) Influence of the management strategy
and operating conditions on the performance of an adsorption chiller. Energy 36:55325538
11. Gordeeva L, Aristov YuI (2014) Dynamic study of methanol adsorption on activated carbon
ACM-35.4 for enhancing the specic cooling power of adsorptive chillers. Appl Energy
117:127133
12. Aristov YI, Glaznev IS, Girnik IS (2012) Optimization of adsorption dynamics in adsorptive
chillers: loose grains conguration. Energy 46:484492
52 2 Adsorption Heat Exchangers

13. Gordeeva L, Frazzica A, Sapienza A, Aristov YuI, Freni A (2014) Adsorption cooling
utilizing the LiBr/silicaethanol working pair: dynamic optimization of the adsorber/heat
exchanger unit. Energy 75:390399
14. Gordeeva LG, Aristov YuI (2011) Composite sorbent of methanol LiCl in mesoporous silica
gel for adsorption cooling: dynamic optimization. Energy 36:12731279
15. Tatlier M, Erdem-enatalar A (1999) the effects of thermal and mass diffusivities on the
performance of adsorption heat pumps employing zeolite synthesized on metal supports.
Microporous Mesoporous Mater 28:195203
16. Bonaccorsi L, Proverbio E (2004) Synthesis of thick zeolite 4A coatings on stainless steel.
Microporous Mesoporous Mater 74:221229
17. Bonaccorsi L, Calabrese L, Freni A, Proverbio E, Restuccia G (2013) Zeolites direct synthesis
on heat exchangers for adsorption heat pumps. Appl Therm Eng 50:15901595
18. Bonaccorsi L, Freni A, Proverbio E, Restuccia G, Russo F (2006) Zeolite coated copper foams
for heat pumping applications. Microporous Mesoporous Materials 91:714
19. Fldner G, Schnabel L, Wittstadt U, Henning H-M, Schmidt FP (2011) Numerical layer
optimization of aluminium bre/SAPO-34 composites for the application in adsorptive heat
exchangers. In: Proceedings of the international sorption heat pump conference 2011, 0608
April, 2011, Padova, Italy, pp 533542
20. Bauer J, Herrmann R, Mittelbach W, Schwieger W (2009) Zeolite/aluminum composite
adsorbents for application in adsorption refrigeration. Int J Energy Res 33:12331249
21. Bonaccorsi L, Calabrese L, Freni A, Proverbio E (2013) Hydrothermal and microwave
synthesis of SAPO (CHA) zeolites on aluminium foams for heat pumping applications.
Microporous Mesoporous Mater 167:3037
22. Jaeschke S, Wolf M (2007) Layer composite and production thereof, Patent WO 2007017015
A3
23. Dawoud B, Vedder U, Amer E-H, Dunne S (2007) Non-isothermal adsorption kinetics of
water vapour into a consolidated zeolite layer. Int J Heat Mass Transf 50:21902199
24. Restuccia G, Freni A, Maggio G (2002) A zeolite-coated bed for air conditioning adsorption
systems: parametric study of heat and mass transfer by dynamic simulation. Appl Therm Eng
22:619630
25. van Heyden H, Munz G, Schnabel L, Schmidt F, Mintova S (2009) Kinetics of water
adsorption in microporous aluminophosphates for regenerative heat exchangers. Appl Therm
Eng 29:15141522
26. Waszkiewicz SD, Tierney MJ, Saidani Scot H (2009) Development of coated, annular ns for
adsorption chillers. Appl Therm Eng 29:22222227
27. Okamoto K, Teduka M, Nakano T, Kubokawa S, Kakiuchi H (2010) The development of
AQSOA water adsorbent and AQSOA coated heat exchanger. In: IMPRES conference,
pp. 2732
28. Freni A, Bonaccorsi L, Calabrese L, Capr A, Frazzica A, Sapienza A (Accepted) Sapo-34
coated adsorbent heat exchanger for adsorption chillers. Appl Therm Eng
29. Freni A, Frazzica A, Dawoud B, Chmielewski S, Bonaccorsi L, Calabrese L (2013) Adsorbent
coatings for heat pumping applications: verication of hydrothermal and mechanical
stabilities. Appl Therm Eng 50:16581663
30. Sauer J, Dawoud B, van Heyden H, Klaschinsky H, Lohmoller S (2008) Adsorberelement und
Verfahren zur Herstellung eines Adsorberelements, Patent DE102008050926 A1
31. Sibilia JP et al (1996) Materials characterization and chemical analysis. Wiley, New York.
ISBN 1-56081-922-7
32. Karger J, Ruthven DM (1992) Diffusion in zeolites and other microporous solids. Wiley,
London
33. http://www.iza-structure.org/databases/
34. Mackenzie RC (1970) Differential thermal analysis. Academic press, London
35. Breck DW (1964) Zeolite molecular sieves. Wiley, New York
References 53

36. Henninger SK, Freni A, Schossig P, Restuccia G (2011) Unied water adsorption measurement
procedure for sorption materials. In: Proceedings of international sorption heat pump conference,
68 April 2011, Padova, Italy, pp. 513522. ISBN 978-2-913149-87-8
37. Crank J (1975) Mathematics of diffusion. Oxford University Press, London
38. Aristov YI, Tokarev MM, Freni A, Glaznev IS, Restuccia G (2006) Kinetics of water
adsorption on silica Fuji Davison RD. Microporous Mesoporous Mater 96:6571
39. Rothmeyer M (1985) Warmetransformation mit dem stoffpaar zeolith-wasser. PhD thesis,
Institut fur Festkorperphysik, TU Munich
40. Storch G, Reichenauer G, Scheffler F, Hauer A (2008) Hydrothermal stability of pelletized
zeolite 13X for energy storage applications. Adsorption 14:2752841
41. Belding WA, Delmast MPF, Holeman W (1996) Dessiccant aging and its effect on desiccant
cooling system performance. Appl Therm Eng 16:447459
42. Lourenco JP, Ribeiro MF, Ribeiro FR, Rocha J, Gabelica Z, Derouane EG (1995) Thermal and
hydrothermal stability of SAPO-40. Microporous Mater 4:445453
43. Dawoud B, Dunne S, Lang R (2002) Experimental investigation of the kinetics of water vapor
adsorption into Molsivtm DDZ-70 under typical operating conditions of adsorption heat
pumps. In: International sorption heat pump conference
44. Aristov YuI, Dawoud B, Glaznev IS, Elyas A (2008) A new methodology of studying the
dynamics of water sorption/desorption under real operating conditions of adsorption heat
pumps: experiment. Int J Heat Mass Transf 51:49664972
45. Sapienza A, Santamaria S, Frazzica A, Freni A, Aristov YuI (2014) Dynamic study of
adsorbers by a new gravimetric version of the large temperature jump method. Appl Energy
113:12441251
46. Wittstadt U, Jahnke A, Schnabel L, Sosnowski M, Schmidt FP, Ziegler F (2008) Test facility
for small-scale adsorbers. In: International sorption heat pump conference
47. Dawoud B (2013) Water vapor adsorption kinetics on small and full scale zeolite coated
adsorbers: a comparison. Appl Therm Eng 50:16451651
Chapter 3
Hydrothermal Stability of Adsorbent
Coatings

3.1 Hydrothermal Aging of Zeolites

Hydrothermal aging of zeolites is an irreversible change of adsorbent surface


characteristics due to repeated cyclic adsorption and desorption of water resulting
in, among others, reduced adsorption capacity and kinetics of adsorption.
Depending on their type, dry zeolites can be heated in dry air up to 500800 C
without exhibiting a permanent damage in their porous structure. Such thermal
structure stability increases with increasing the Si/Al-ratio. Heating a dry zeolite
beyond its structure stability temperature limit results in a partial or a total loss of
the porous structure and, consequently, of the adsorption capability. The recrys-
tallization temperature depends on both lattice and cation types. Puppe [1] has
reported that Na-A zeolite is stable if heated in dry air up to 700 C, Ca-A zeolite up
to 800 C, Na-X up to 760 C and Na-Y zeolite up to 800 C. Indeed the Si/Al-ratio
of Na-A zeolite is one, of Na-X zeolite 2 to 3 and of Na-Y zeolite between 3 and 6.
High silicate structures having a Si/Al-ratio between 20 and 100 have a structure
stability temperature of up to 1100 C [1].
Lcke [2] has carried out a differential-scanning-calorimetry-analysis on almost
dry zeolite Na-A at a heating rate of 10 K/min and identied one endothermic and
two exothermic peaks. The endothermic peak at 216 C has been attributed to
desorption of water out of the zeolites porous structure. The exothermic peak at
826 C designated a collapse of the lattice structure, which has been associated with
a volume reduction and an increase in the electrical conductivity of the tested
zeolite. The amorphous transient phase has been recrystallized at 938 C to a pore
free structure giving the second exothermic peak.
Piguzova [3] has investigated the stability of zeolites 4A, 5A and 13X against
heating in dry air and came to the conclusion that all investigated types are stable up
to 600 C. Beyond this temperature limit, 4A zeolite has been destroyed faster than
the other two types. To judge on the stability, Piguzova has measured the

The Author(s) 2015 55


A. Freni et al., Characterization of Zeolite-based Coatings
for Adsorption Heat Pumps, SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-09327-7_3
56 3 Hydrothermal Stability of Adsorbent Coatings

adsorption capacity (water uptake) of the samples before and after the thermal
treatment in dry air.
Heating a zeolite in the presence of water vapor can reduce the structure stability
temperature limit remarkably, depending on zeolite lattice structure, type of cation
or cations encountered, degree of cation exchange, zeolite form (powder, pellets,
granulates or a consolidated layer) as well as on the binder type and amount. The
number of adsorption/desorption cycles, the temperature level of each and the water
vapour pressure during cycling play a crucial rule in the degree of aging of a certain
zeolite. One of the crucial tasks in developing a zeolite water adsorption heat pump
or a chiller is, therefore, the approval of the stability of the applied zeolite, its form,
type of binder and binder content on the same metal substrate, if any, as in the nal
product against hydrothermal aging cycling, in order to guarantee a stable operation
over the life-time of the developed system.
Wyart and Sabatier [4] have investigated water diffusion through zeolite Na-A
and observed a change in the crystal structure upon treating the zeolite with water
vapor at the test temperatures of 350 and 750 C. Moreover, they have documented
that the cations exchange their places within the zeolite lattice at a temperature of
300 C in a self-diffusion mode. Such cation movements does create a stress on the
oxygen atoms, which tie the elementary tetrahedral units together. Accordingly,
the tetrahedral units can be separated from each other. In the presence of water
molecules inside the zeolite structure, the resulting stress on the oxygen bridges
between the tetrahedral units increases.
Wolf et al. [5] have carried out a thorough investigation on the stability of
A-zeolite with different cations and cation-mixtures with different exchange degrees
against hydrothermal aging at different temperatures, water vapor pressures and
durations of water treatment. Warm air at different temperatures have been con-
ditioned to different relative humidities of water before flowing over the different
zeolite samples. The zeolite temperature has been changed between 300 and
600 C, while the water vapor pressure between 16 and 610 mm Hg. The duration
of treating each sample amounted to 67 h. In addition to the original Na-A zeolite,
Wolf et al. have investigated partially exchanged Na-cations against alkaline ions
[Lithium (Li), Potassium (K), Cesium (Cs)] as well as alkaline earth ions [Mag-
nesium (Mg), Calcium (Ca), Strontium (Sr) and Barium (Ba) cations] with different
degrees of cation exchange. Upon exchanging Na against K, Li or Cs, they
observed a reduction in pore dimension from 0.42 to 0.3 nm. In addition, they
observed more or less a total collapse of the porous structure of K-, Sr- and
Ba-exchanged A-zeolites after the drying (activation) treatment at higher degrees of
ion exchange. The adsorption capacity of the original Na-A zeolite has been sig-
nicantly reduced after one treatment at 600 C and a water vapor partial pressure
of 16 mm Hg. They observed the same result for A-Zeolite in Li-form. Increasing
the partial pressure of water vapor to the level between 100 and 200 mm Hg
resulted in a dramatic reduction in the adsorption capacity of both zeolites. The
other forms of the investigated A-zeolites showed clearly higher stability. One of
the major conclusions of Wolf et al. is that Mg-Na-A zeolite has shown the highest
stability against the explained hydrothermal treatment.
3.1 Hydrothermal Aging of Zeolites 57

Moreover, Wolf et al. have observed a reduction in the adsorption capacity of


Li-Na-A zeolites at an exchange level of 510 %. At an ion exchange degree
between 30 and 40 % the adsorption capacity reaches a minimum. The minimum in
the adsorption capacity could be observed upon exchanging 10 % of the Na-ions
against K-ions. Increasing the degree of K-ion-exchange to 40 % resulted in the
maximum stability against the hydrothermal treatment. However, the stability of K-
Na-A zeolite has been dramatically reduced if the K-ion exchange degree have been
increased beyond 40 %. Contrary to that, the stability of Cs-exchanged A-zeolite
has been increased continuously upon enhancing the degree of ion-exchange up to
30 %. Furthermore, Wolf et al. have identied the maximum stability of A-zeolites
with alkaline earth ions [Mg, Ca, Sr and Ba] against their hydrothermal treatment
process at an ion exchange degree between 20 and 40 %. The observed results have
been explained as follows: as not all the Na-ions within the zeolite lattice have the
same energy level (three different locations have been identied within the Na-A-
zeolite lattice as explained in Chap. 1 (Sect. 1.2.4.1), the maximum stability may be
reached if the three type-II ions and the one type-III ion are exchanged. The
exchange of further Na-ions out of the 8 type-I ions results in a decrease in the
adsorption capacity depending mainly on the type of the exchanged cation.
Finally, Wolf et al. [5] have concluded that the stability of Na-A-zeolite lattice
against the described hydrothermal treatment process decreases with increasing the
ion radius of the exchanged alkaline earth ions. Contrary to that, the lattice stability
of Cs-Na-A-zeolite is much higher than that of Li-Na-A zeolite. Despite the very
high value of the results obtained by Wolf et al. the utilized hydrothermal treatment
process is not matching with a heat pump process, where other temperature and
water vapor pressure levels are valid. Moreover, it is quite important to include the
number of aging cycles as one of the most important influencing factors on the
stability of the investigated zeolite lattice.
Rothmeyer [6] has investigated the aging effect resulting from multi cycling
different zeolite types at two different sets of boundary conditions. First, he
investigated the effect of heating and cooling zeolite samples at a constant water
uptake between the temperature levels of 150 and 50 C. This isosteric heating and
cooling cycle incorporates both temperature and vapor pressure changes. Each half
cycle lasted 2 h and a total of 2000 cycles have been performed on each zeolite
sample. In order to judge on whether and how much degradation in the water
uptake took place in the different zeolite samples, the equilibrium uptake of fresh
and aged samples has been measured at 20 C and a vapor pressure of 12 mbar and
the results have been compared.
The results of investigations according to the rst set of boundary conditions
showed that granulated zeolite Mg-Na-A exhibits no water uptake degradation after
the 2000 cycles. Contrary to that, granulated zeolite Na-X and powder zeolite Na-Y
lost almost 25 % of their adsorption capacities compared to the fresh samples. In
addition, Rothmeyer has observed that cycled powder zeolites Ca-A and Na-A
showed a vanishing small water uptake reduction compared to their fresh samples.
On the other hand, aged powder zeolite Na-X lost ca. two thirds of their initial water
uptake while aged powder Mg-Na-A lost only one third of its initial uptake.
58 3 Hydrothermal Stability of Adsorbent Coatings

The second set of operating conditions was chosen to simulate an isobaric


desorption/adsorption process at a constant vapor pressure of 500 mbar. The tem-
perature has been altered between 50 and 150 C resulting in changing the water
uptake. Each half cycle lasted also 2 h and a total of 1200 cycles have been carried
out on each sample. Also here, the equilibrium uptake of fresh and aged samples
has been measured and compared to evaluate the degradation effect of the aging
cycles, if any.
It came out that cycled granulated Mg-Na-A zeolite has lost 4 g/100 g (from 31
to 27 g/100 g) over the 1200 cycles. This loss in water uptake have been realized
after the rst 370 cycles. Afterwards no further degradation has been observed.
Contrary to that, all other cycled zeolites (Ca-A, Na-X, Na-Y, Mg-Y and Ca-Y)
have lost almost half of their initial water uptake after the 1200 cycles. One further
investigation (1200 cycles) has been carried out on the already cycled Mg-Na-A
sample, however with a desorption temperature of 310 C, showing no further
degradation effect on this sample.
Rothmeyer [6] has concluded that the granulation process including the type and
amount of binder did have a great effect on the degradation resistance of zeolites
against hydrothermal cycling. He observed that granulated Mg-Na-A zeolite is
clearly more stable than in powder form. Same effect has been observed by
Rothmeyer [6] as well as Nefedow and Valujskaja [7] for Na-X zeolite. Moreover,
both references have concluded that Na-X zeolite is not suitable for heat pump
processes, where the desorption temperature is higher than 150 C. Nefedow and
Valujskaja [7] have added that enhancing the Si-content of Na-X zeolite up to
MSiO2 4:8 increases its stability limit temperature up to 550 C. Our own results
on the stability of granulated zeolites against hydrothermal aging, which are pre-
sented at the end of this chapter, are in phase with the observations of Wolf et al. [5]
and Rothmeyer [6] concerning Mg-Na-A and 13X zeolites. However, specially
granulated Na-Y zeolites have shown also premium stabilities against hydrothermal
cycling. The main focus of this chapter remains, however, to explore the stability of
zeolite layers against hydrothermal cycling.
Plenty of studies about thermal stability of zeolites for various applications (gas
separation, drying, etc.) are available in literature, while only a few studies about
the stability of adsorbent coatings have been carried out. Henninger et al. [8]
developed an aging method suitable for adsorbent coatings, demonstrating that
some novel SAPO structures may have dramatic decay of adsorption capacity due
to the aging process. Restuccia et al. [9] investigated the hydrothermal stability of a
coated adsorbent bed based on the composite sorbent SWS-1L (CaCl2 in meso-
porous silica gel) after a large number of sorption/desorption cycles in a laboratory
scale adsorption heat pump. To verify the coating stability, small samples of
adsorbent coating were extracted from different parts of the adsorber after about
1000 working cycles and were characterized in comparison with the fresh state.
Measurement of the XRD pattern and of the water adsorption capacity of the aged
samples demonstrated that no degradation of the sorbent occurred under the
imposed testing conditions.
3.1 Hydrothermal Aging of Zeolites 59

In this book, we are going to introduce and focus on three different methodologies
to produce aging due to a hydrothermal treatment of zeolites. In other words, the
stability of different zeolite forms against hydrothermal aging has been investigated
according to three different aging procedures. The following aging cycling proce-
dures will be described:
(a) real-time or long term heat pump cycling,
(b) accelerated aging cycling,
(c) very quick aging cycling.

3.2 Real-Time or Long-Term Heat Pump Aging Cycling


Procedure

3.2.1 Description of the Set-up

This method allows long-term aging of coated tubes under real operating conditions
(pressure, temperature, cycle time) of an adsorption heat pump. Possibility to test
samples which replicate realistic adsorption heat pumps operating conditions under
vacuum is a clear advantage. However, this method is time consuming and only a
limited number of samples can be tested simultaneously. Figure 3.1a, b shows the
scheme and practical realization of the dedicated set-up, able to automatically test
up to four coated tubes (length * 30 cm). The coated tubes to be tested are placed
inside a cylindrical vacuum chamber, which was equipped with two viewing ports
to check the samples under test during the aging process. The same chamber holds
two blank nned tubes which act as condenser/evaporator. The heating/cooling of
coated tubes is realized by means of an external hydraulic circuit. The test rig is
equipped with a set of thermocouples, pressure sensors and a flow meter in order to
control the evolution of the process. A set of solenoid valves, connected to a control
and management unit, allows to periodically switch between cooling/heating phase.
The set up allows automatic cycling of the sample tested. An auxiliary heating
system was installed to prevent condensation over the internal walls of the testing
chamber. Concerning the DAQ system, a specic control panel was realized in
LabVIEW environment allowing to follow the evolution of the process and to
record data with the desired time-step. Specically, the following data are acquired:
temperature of the external sources, temperature of the external heating/cooling
fluid at the inlet of coated tubes, temperature of a coated tube inside the chamber,
temperature of a blank tube inside chamber, pressure of the chamber.
Figure 3.2a, b show further details of the practical realization of the vacuum
chamber. The chamber was designed in order to limit the dead volume as much as
possible. The testing tubes are inserted inside (DN 25 mm) flanges and blocked by
special vacuum tight feed-through. Two specular viewing ports (DN 200 mm)
permit the optical analysis of the samples under test. Additional flanges placed on
60 3 Hydrothermal Stability of Adsorbent Coatings

(a)

(b)

Fig. 3.1 a, b Scheme and practical realization of the whole set up for long term aging

the chamber side are used for connection to the vacuum pump, thermocouples,
feed-through etc.
Table 3.1 reports the selected testing operating conditions, which are similar to
those of a real adsorption heat pump.
3.2 Real-Time or Long-Term Heat Pump Aging Cycling Procedure 61

Fig. 3.2 a, b Conceptual


scheme and the practical
realization of the vacuum
chamber

3.2.2 Experimental Testing

Experimental testing was carried out on coated tubes prepared according to the
methods reported in:
(a) Freni et al. 2011, laboratory scale development [10]
(b) Sauer et al. 2008, Pre-development scale [11]
(c) Okamoto et al. 2010, Pilot scale [12]
All coatings employed adsorbent AQSOA FAMZ02 from Mitsubishi Plastic
Incorporation as active material. Figure 3.3ac shows the fresh coated tubes real-
ized by the above mentioned techniques, as installed inside the vacuum chamber.
The coated tubes were subjected to a relevant number of continuous ad/
desorption cycles under the conditions reported in the previous Table 3.1. Typical
testing procedure included a preliminary degassing of the coated tubes. With this
aim, the coated tubes were heated up to 90 C for 24 h and connected to a vacuum
pump. After this process, the connection with the vacuum pump was closed and the
coated tubes were cooled down to ambient temperature. The pressure measured at
ambient temperature was completely constant and equal to 0.1 mbar. Afterwards,
connection to the thermostated evaporator was opened allowing the water vapour to
enter the testing chamber. Once the pressure inside the chamber was stabilized,

Table 3.1 Operating conditions of the long-term stability setup


Adsorption phase Desorption phase
Duration of the phase (min) 20 20
Min/max adsorbent temperatures (C) 15 90
Evap./cond. temperatures (C) 15 15
Flow rate of external HT fluid (l/min) 25 25
62 3 Hydrothermal Stability of Adsorbent Coatings

Fig. 3.3 Fresh coated tubes, as installed in the vacuum chamber, realized according to: a Freni
et al. [10], b Sauer et al. [11], c Okamoto et al. [12]

periodical heating/cooling of the coated tubes started by switching the external


hydraulic circuit, thus allowing hydrothermal cycling of the samples under test.
Figure 3.4 reports the evolution of a typical single ad/desorption cycle, allowing to
better appreciate the evolution of the most important testing parameters.
From the gure, it is evident that the tubes are efciently heated and cooled with
a large temperature lift (9015 C; *75 C), which ensures a consistent water
loading variation in the range 2025 wt% for SAPO34 zeolites. The blank nned
tubes serving as evaporator/condenser are kept at quasi-constant temperature of
around 15 C, apart an initial slight temperature rise due to the phase-switching.
The operating pressure was in the range 18 mbar (adsorption) and 24 mbar

100 45
90 40
80 35
70
Temperature [C]

30
Pressure [mbar]

60
25
50
20
40
15
30
20 10

10 5

0 0
27500 28000 28500 29000 29500 30000
Time [s]
T-in coated tubes T-in evap/cond P chamber

Fig. 3.4 Typical evolution of the testing parameters for single ad/desorption cycle
3.2 Real-Time or Long-Term Heat Pump Aging Cycling Procedure 63

Fig. 3.5 Water adsorption isobars at 11 mbar measured for fresh and aged samples

(desorption), due to the effect of the anti-condensation system on the warm walls of
the testing chamber. Stability of coated tubes under test was visually controlled by
the viewing ports of the testing chamber. Moreover, periodically the setup was
stopped and small portions of coating were extracted from the chamber and sub-
jected to XRD analysis and water adsorption capacity measurement by thermo-
gravimetric technique.
Figure 3.5 reports the water adsorption isobars measured at 11 mbar for the fresh
and aged coated tubes. Tubes coated prepared according to the methodology
developed by Freni et al. [10] showed poor mechanical stability already after 1000
cycles, which was probably due to the brittle behavior of the clay-based binder.
Accordingly, the aging experiment was interrupted after about 1 month of cycling.
Coated tubes realized according to Sauer et al. [11], were tested for about one year.
Adsorption capacity of aged tubes was practically unchanged with respect to the
fresh sample. XRD crystalline pattern was regular as well, indicating that the zeolite
structure was not damaged. However, the coating over the nned tube appeared
slightly damaged (in the mechanical sense) already after 3000 cycles. This is
demonstrated in Fig. 3.6a, which shows that the aged tubes inside the testing
chamber released some small parts of coating due to the aging treatment. The weak
points in Sauer et al. [11] procedure have been identied and thoroughly discussed
in [13], while the optimization measures taken to remarkably enhance the
mechanical stability, reduce the binder content and to achieve a better and uniform
coating process have been presented in [14]. Coated tubes realized according to
Okamoto et al. [12] were successfully tested for more than one year without
64 3 Hydrothermal Stability of Adsorbent Coatings

Fig. 3.6 a, b Aged coated tubes realized according to: a Sauer et al. (3000 cycles) [11],
b Okamoto et al. (16,000 cycles) [12]

showing a mechanical instability. Figure 3.6b demonstrated the good status of the
tubes after 16,000 ad/desorption cycles, which was also conrmed by the regular
XRD pattern and water adsorption capacity measured on the aged samples at the
end of the experimentation. Overall aging results obtained for the tested coated
tubes are collected in Table 3.2.
Another but similar version of the previously described long-term aging cycling
setup has been developed to investigate the long time stability of loose pellets of
Zeolite 13X (Baylith, WE-G592 of Bayer AG) against heat pump cycling. As
depicted in Fig. 3.7, the loose zeolite pellets have been integrated between the ns
of a nned-tube heat exchanger. Another nned-tube has been applied to work as
evaporator/condenser. Heating the zeolite pellets has been realized by an electrical
heating wire, which is inserted inside the tube. The maximum desorption temper-
ature amounted to 200 C. During this process, the condenser nned-tube has been
cooled by cold water (at 15 C) from a cooling thermal bath.
After reaching the maximum desorption temperature (200 C), the desorption
phase is stopped by switching off the power supply to the heating wire and
switching on the cooling air blower. Air flows rst through a plate heat exchanger
to be cooled by the cold water (15 C) out of the cooling thermal bath and then
flows through the adsorber tube from the bottom to top. In this phase, the

Table 3.2 Results obtained for the aged coated tubes


Total Water adsorption capacity, wt% XRD Mechanical
no. (T = 40 C pH2O = 11 mbar) crystalline stability
cycles pattern
Freni 1000 19.5 (fresh) Regular Poor
et al. [10] 19.2 (aged)
Sauer 12,000 19.0 (fresh) Regular Slight
et al. [11] 19.0 (aged) damage
Okamoto 16,000 21.5 (fresh) Regular Good
et al. [12] 21.2 (aged)
3.2 Real-Time or Long-Term Heat Pump Aging Cycling Procedure 65

Electrically insulated Water to the


Heating Wires Thermal Bath

200C
Vacuum Flange
oooo oooo
oooo oooo
oooo oooo
oooo oooo
Control Unit oooo oooo
oooo oooo
oooo oooo
oooo oooo
Finned-Tube oooo oooo Finned-Tube
Adsorber with oooo oooo Condenser/Evaporator
oooo oooo
loose pellets oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo Glas Cylinder with
oooo oooo
oooo oooo Vacuum Flanges
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
oooo oooo
Cooling Blower

.
Q
Cold Water from the
Thermal Bath (15 C)
Plate Heat Exchanger
For Air Cooling before
Entering the Adsorber

Fig. 3.7 Long-term adsorption heat pump cycling setup for loose zeolite pellets

evaporator tube is heated by water (at 15 C) from the cooling thermal bath. The
adsorption phase continued until the zeolite temperature falls below 60 C.
A differential water uptake of 20 g/100 g could be realized and a daily number of
cycles of 58 could be carried out. A total of 40,000 heat pump cycles could be
realized. The detailed analysis of the so-aged loose pellets compared to aged pellets
according to the very quick aging procedure will be presented at the end of this
chapter.

3.3 Accelerated Aging Cycling Procedure

3.3.1 Description of the Set-up

The aim of this test method is to allow accelerated aging of multiple small-scale
samples by thermal ad/desorption cycling under wet/dry air flux. This method has
the advantage of allowing many cycles per day over several samples, but typical
vacuum conditions of adsorption heat pumps are not achieved. Moreover, due to the
66 3 Hydrothermal Stability of Adsorbent Coatings

presence of oxygen in the air flux, there may be an interaction with the coating
affecting its stability. Figure 3.8 shows the scheme of the dedicated test-rig, which
was specically designed to test multiple samples and to achieve a higher number
of ad/desorption cycles in a reduced time-scale. The set-up allows simultaneously
cycling of 15 samples, displaced in three testing chambers operating in parallel.
Alternate cooling and heating of samples was achieved by means of thermoelectric
elements placed over an air-cooled aluminium heat dissipater. Ad/desorption of
water vapour was achieved by passing through the samples a flux of slightly
compressed wet/dry air. Humidity of the air was accurately regulated by tube-in-
shell Naon humidiers and drying columns. Wet air was obtained by a membrane
tube-in-shell moisture exchanger, that allows the transfer of water vapor between a
liquid water supply and the air stream. Water is diffused through the wall of the
Naon tube and is then vaporized into the air stream. This vaporization process is
driven by the difference in the partial pressure of water vapor on both sides of the
membrane. Dry air was efciently obtained by a common silica gel drying column.
Figure 3.9 shows the overall view of the practical realization of the setup,
depicting the three testing chambers, the external air circuit, the control units and
the set of solenoid valves used to automatically perform the aging cycles under the
imposed operating conditions. Each testing chamber is made of 5 thermoelectric-
plates which lay over a heat dissipator, which accumulates the heat coming from the
sample cooling. The samples to be tested are directly placed over the thermo-
electric plates. A transparent cover protects the samples and directs the dry/wet air
to them. An automatic management system makes it possible to set the cycle time,
to control the solenoid valves of the air circuit and to control the thermo-electric
plates. Furthermore, the maximum and minimum temperature of the cycle are also

Fig. 3.8 Scheme of the test-rig for accelerated adsorption/desorption cycles


3.3 Accelerated Aging Cycling Procedure 67

Fig. 3.9 Overall view of the


practical realization of the
setup

set and controlled. Each air circuit contains two automatic solenoid valves for each
testing chamber (in order to switch the air flux to be wetted or dried), a hygrometer
and a bubble flow meter to control the air humidity and flow rate, respectively.
Figure 3.10 shows the testing chambers and the dissipators. It is easy to rec-
ognize the samples placed over 5 thermoelectric plates. A thermo-resistance is xed
over a plate in order to control the temperature of the cycle. A special heat sink
compound is applied between the above mentioned elements to maximize the heat
transfer efciency. During the heating phase, the thermoelectric plates heat-up the
samples allowing desorption. Meanwhile, a flux of dry air is introduced by the
plastic tubes on the left. The air exits the testing chamber by flowing through small
holes done on the opposite side. During the cooling phase the thermoelectric plates
cool-down the samples and transfer heat to the heat dissipator. The heat dissipator
used is an aluminium nned heat sink (5 kg30 cm long) usually used for elec-
tronic applications. To maximize the efciency of the heat sink system, the alu-
minium block is continuously cooled by a fan (the black boxes on the head of the
dissipators).

3.3.2 Estimation of the Operating Conditions


for AQSOA- FAMZ02 Zeolite

Figure 3.11 presents the adsorption isotherms measured for the FAMZ02 zeolite,
permitting to identify the two limits of the cycle (R.H. < 1 %, T = 90 C
desorption; R.H. > 70 %, T = 20 Cadsorption). It is evident that, theoretically,
68 3 Hydrothermal Stability of Adsorbent Coatings

Fig. 3.10 Detailed view of


the testing chambers

the imposed conditions of pressure and temperature make it possible to obtain a


large variation of uptake (about 0.3 g/g at equilibrium). Obviously, a system for
quick cycles must operate with relatively short cycle duration, so that the real
variation of uptake will be lower than that one estimated for the ideal thermody-
namic case. In agreement with the sorption kinetics measurement presented in
literature for FAMZ02 by Dawoud [15], a cycle time of 10 min (5 adsorption + 5
desorption) can be reasonably selected in order to reach 5080 % of the ideal water
loading.
In order to verify if the incoming wet air can provide the proper amount of
vapour to the 5 samples, the following data can be assumed: wet air flow = 5 l/min,
mass of a single zeolite sample 0.2 g, R.H during adsorption 70 %. The total mass
of the 5 zeolite samples is 1 g, so that, assuming a real variation of uptake
W = 0.2 g/g, the total mass of vapour to be provided is 0.2 g. From simple
calculations it can be derived that, in the dened conditions, the wet air provides
such amount of water vapour in almost 3 min, which is coherent with the 5 min
selected for the adsorption phase. The above mentioned estimation were conrmed
by preliminary experimental tests.
Accordingly, a typical aging cycle was carried out as follows: during the cooling
step (Tsample = 20 C), a flux of wet air (5 l/min, R.H. 70 %) was introduced at the
inlet of the testing chambers providing water vapour for samples adsorption. During
the heating step (Tsample = 125 C), a flux of dry air (5 l/min, R.H. < 2 %) was
passed to achieve samples desorption. Duration of a single aging cycle was 10 min
(5 min cooling/adsorption plus 5 min heating/desorption). The variation of water
uptake due to the imposed conditions, was estimated from the equilibrium adsorption
curves as w = 20 wt%. Testing conditions are resumed in the following Table 3.3.
3.3 Accelerated Aging Cycling Procedure 69

0.4

0.3

T=25C
uptke, g/g

T=40C
0.2
T=55C
W~0.3 g/g
T=80C

0.1

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
P/Ps

Fig. 3.11 Identication of the cycle limits over the AQSOA FAM-Z02 zeolite adsorption isotherms

Table 3.3 Testing conditions applied for the accelerated aging set up
Adsorption phase Desorption phase
Duration of the phase (min) 5 5
Min/max temperature (C) 20 125
Relative humidity of the air (%) >70 <3
Air flow (l/min) 5 5

3.3.3 Experimental Testing

Similarly to the long-term stability testing activity presented in the previous para-
graph, accelerated aging was carried out on several different coated samples pre-
pared according to the methods reported in:
(a) Freni et al. 2011, laboratory scale development [10]
(b) Sauer et al. 2008, Pre-development scale [11]
(c) Okamoto et al. 2010, Pilot scale [12]
Even in this case, all coated samples employed the novel adsorbent AQSOA
FAMZ02 from Mitsubishi Plastic Incorporation as active material. Characterization
of samples (adsorption capacity measurement and X-ray analysis), was periodically
carried out on small portions of sample, in order to verify if the aging treatment did
not affect the water adsorption capacity or the zeolite stability with respect to the
fresh state of the coating. Figure 3.12 shows the status of some samples prepared
according to the above listed techniques after 35,000 aging cycles.
As clearly shown in Fig. 3.12a, b, the coated samples prepared according to
Freni et al. and Sauer et al. appeared to be mechanically partially damaged, which is
70 3 Hydrothermal Stability of Adsorbent Coatings

Fig. 3.12 Aged coatings (35,000 cycles) realized according to: a Freni et al. [10], b Sauer et al.
[11], c Okamoto et al. [12]

coherent with the results found in the previously reported long- term stability
testing results. Differently, Fig. 3.12c indicates that samples prepared according to
Okamoto et al. did not show any mechanical failure after 35,000 cycle. Measure-
ment of the water adsorption capacity demonstrated that all cycled samples have
practically unchanged adsorption ability within the limits of error. As an example,
next Fig. 3.13 reports the adsorption isobar measured at 11 mbar for a sample
prepared according to Freni A. et al. which was aged 20,000 cycles in comparison
with the fresh state. Measurement of the X-ray diffraction pattern of all aged
samples did not show any relevant decay of the zeolitic crystalline structure. The
latter result conrmed that the failure of the coatings showed in Fig. 3.12a, b was
due to the binder formulation not optimized for this type of hydrothermal stress,
while the zeolite structure itself was not affected by the aging treatment.
Table 3.4 resumes the adsorption capacities measured at T = 40 C and
pH2 O 11 mbar for the tested coatings after 35,000 aging cycles.

Fig. 3.13 Adsorption isobars at 11 mbar measured for fresh and aged 20,000 cycles samples
prepared according to a sample prepared according to [10]
3.4 Very-Quick Aging Cycling Procedure 71

Table 3.4 Adsorption capacity (T = 40 C and pH2 O 11 mbar) of different aged coatings
Number of cycles Freni et al. [10] Sauer et al. [11] Okamoto et al. [12]
0 (wt%) 19.5 19.0 21.5
35,000 (wt%) 19.5 18.9 21.4

3.4 Very-Quick Aging Cycling Procedure

3.4.1 Description of the Set-up

The components of the very quick aging cycling test bench and their intercon-
nections are illustrated in Fig. 3.14. The operation of this setup comprises the
following two alternating phases:
(a) Regeneration or Desorption Phase
In the desorption phase, air is sucked by means of an air handling pump with a
constant flow rate of two norm cubic meter per hour from the environment and
forwarded to drying in a vessel which is lled with dry silica gel. Afterwards, air is
passed and heated by a tubular furnace. This hot and dry air is fed on the zeolite
sample, which is accommodated in a sample container with a sieve-like bottom, so
that the air can flow through the pellets and pass into the surroundings again. In this
cycle, the zeolite sample is heated by hot and dry air to a desorption temperature of
1.5 times the maximum desorption temperature during the normal heat pump
operation. The adsorbed water molecules are then expelled from the zeolite lattice
structure. The dry air is able to receive the desorbed water vapor and transport it
into the environment.

Time Controller

Oven Control Unit


Air Flow Meter

Dosing Pump

Air Handeling Pump


30 ml/min
3
2 nm /h

Tube Oven Metal Wool


Temperature
Sensor Zeolite Sample
Silica Gel

Water Reservoir at Room Temperature

Fig. 3.14 Schematic of the experimental setup to produce very quick hydrothermal aging in
zeolite granulates or beads
72 3 Hydrothermal Stability of Adsorbent Coatings

In total, the desorption phase occurs over a period of 35 min, after which the
zeolite becomes hot and dry.
(b) Charging or Adsorption Phase
The adsorption phase follows directly after nishing the desorption phase by
switching off the air pump and setting the water dosing pump into operation. Water
at ambient temperature is sucked out of a water reservoir and is fed at a volume flow
rate of 30 cm3/min over the hot and dry zeolite sample. Such a thermal treatment is
simply a sudden quenching of the zeolite pellets with cold liquid water. The
adsorption phase takes place over 6 min, during which the water uptake increases to
its maximum possible value. After this period a desorption or a regeneration phase
starts again.
After a certain number of aging cycles, both the adsorption uptake according to
the large pressure jump kinetic methodology and the broken zeolite mass per-
centage (<1.25 mm size) related to the initial dry zeolite sample mass are measured
and compared with the results obtained with non-cycled or fresh zeolite samples.

3.4.2 Experimental Testing

During the early stage of developing a multi-modular adsorption heat pump,


Stricker [16] has decided to utilize zeolite Na-X in form of pellets (1.82.5 mm
size) as an adsorbent. The applied zeolite has been produced by Bayer AG under
the commercial name: Baylith WE-G 592 and has been granulated using zeolite
Na-A as a binding material. In order to approve the stability of the zeolite pellets
against hydrothermal aging, some samples of those pellets have been aged
according to both real-time (or long term) and very quick aging cycling procedures.
As the maximum desorption temperature applied in the heat pump operation
amounted to 200 C, it has been decided to set the maximum temperature during the
very quick aging to 300 C (1.5 times the maximum desorption temperature).
A total of 200 very quick aging cycles have been carried out. Figure 3.15 illustrates
the accumulated percentage of the broken zeolite pellets, which have a dimension
less than 1.25 mm with the number of very quick aging cycles. Indeed, almost a
linear increase in the broken percentage of the zeolite sample is obvious indicating a
very poor mechanical stability of such pellets, which has been attributed to the bad
binding characteristics of the applied zeolite A as a binding material.
Samples out of the broken pellets have been taken to determine the adsorption
kinetics according to the large-pressure-jump-kinetic methodology under the con-
ditions of 50 C adsorption temperature and a starting pressure of 60 mbar. The
adsorption kinetic measurements have been carried out on non-cycled (reference)
samples as well as on cycled samples after 10, 30, 50, 100, 150 and 200 very quick
aging cycles. Table 3.5 presents the summary of the obtained nal and differential
water uptakes as well as the variation in the loss in the nal differential water uptake
under the given test conditions. Figure 3.16 illustrates the obtained adsorption
3.4 Very-Quick Aging Cycling Procedure 73

70

60.7
60
Accumulated brocken zeolite pellets [%]

Original Pellet Size: 1.8 - 2.5 mm

50 Brocken pellets < 1.25 mm size 47.8

40
33.2

30

20 16.1
10.9
10
3.2

0
0 20 40 60 80 100 120 140 160 180 200
Number of very quick aging cycles [-]

Fig. 3.15 Accumulated broken percentage of the zeolite sample (Baylith WE-G 592) over 200
very quick aging cycles

Table 3.5 Final, differential and loss of nal water uptake variation with the number of very quick
aging cycles on Baylith WE-G 592 pellets
Number of very quick Xo (g/100 g) X (g/100 g) X Xo (g/ Loss of nal
aging cycles (-) 100 g) uptake (g/100 g)
0 1.6 22.6 21.2 0
10 1.6 21.3 19.7 1.3
30 1.6 19.2 17.6 3.4
50 1.6 18.5 16.9 4.1
100 1.6 18.1 16.5 4.5
150 1.6 16.7 15.1 5.9
200 1.6 15.7 14.1 6.9

kinetic curve for each sample. The observed continuous loss of nal water uptake
indicates a very poor adsorption stability of the tested zeolite against the very quick
hydrothermal aging cycling. This may be referred to the severity of such aging
cycles. Another explanation may be the poor stability of the zeolite structure, which
may be affected by the instability of the utilized binder (here Zeolite A) during the
production of the pellets, as can be deducted from the quite high broken portion of
the samples. The obtained results are very much in phase with the results by Wolf
et al. [5] as well as Rothmeyer [6]; namely, that Zeolite Na-X is not stable against
hydrothermal cycling.
74 3 Hydrothermal Stability of Adsorbent Coatings

Refrigerant Uptake [gW/100gZeo, dry] 25

20

15

10 Zeolite 13X Reference


Baylith WE-G 592 10 cycles
Binder: Zeolite A 30 cycles
mZeo, dry = 3g
xo = 1.6 g/100g 50 cycles
5 po = 60 mbar 100 cycles
TZeo = 50 oC 150 cycles
200 cycles

0
0 400 800 1200 1600 2000 2400 2800 3200 3600
Time [s]

Fig. 3.16 Adsorption kinetic curves of the reference and aged zeolite 13X (Baylith WE-G 592)
samples in the very quick aging setup

In parallel to the above-mentioned investigations a set of real-time or long-term


heat pump cycling has been carried out on the same zeolite pellets, in order to nd
out, if any, a correlation between the effects of both types of aging procedures. The
experimental work has been carried out under the following conditions:

Maximum zeolite temperature 200 C


Minimum zeolite temperature 60 C
Differential water uptake >20 g/100 g
No. of cycles /day 58
Total number of cycles 40,000

In almost regular intervals, the setup depicted in Fig. 3.14 has been opened to
extract a zeolite sample for investigating its adsorption kinetic performance.
Table 3.6 presents the nal and the loss in the nal water uptake with the number of
long-term heat pump cycles.
Figure 3.17 presents the obtained nal water uptake out of each adsorption
kinetic investigation for the different samples aged by both long-term and very
quick cycling procedures. It becomes obvious, that the very quick aging cycling has
a quite severe effect on accelerating the aging effect on the tested zeolite pellets. For
example, the loss of nal water uptake of 1.3 g/100 g is realized after only 10 very
quick cycles compared to almost 8000 real time or long-term heat pump cycles.
Applying a linear interpolation between the obtained results, Fig. 3.18 can be
3.4 Very-Quick Aging Cycling Procedure 75

Table 3.6 Final and loss of nal water uptake variation with the number of long-term cycles on
Baylith WE-G 592 pellets
Number of long-term heat pump cycles [-] X (g/100 g) Loss of nal uptake (g/100 g)
0 22.5 0
6000 21.4 1.1
20,000 20 2.5
30,000 18.9 3.6
40,000 17.9 4.6

Number of real-time heat pump cycles [x1000]


0 4 8 12 16 20 24 28 32 36 40
8
Loss an final water uptake [g/100g]

200
7
Very quick aging
Real-time Heat Pump Cycles 150
6

5 40
100
50
4 30
30

3 20

2
10 6
1
0 0
0
0 20 40 60 80 100 120 140 160 180 200
Number of very quick aging cycles [-]
Fig. 3.17 Loss of nal water uptake for both long-term and very quick aging cycling procedures

deduced to show the ratio between the number of real-time heat pump cycles and
the very quick aging cycles required to realize the same loss in the nal water
uptake (degree of aging). The correlation factor lies between 400 and 920 with a
maximum of 1000 between a number of 20 very quick aging cycling (can be
tackled in 24 h) and 20,000 long-term heat pump cycles.
Knowing that performing the 20,000 real time heat pump cycles required almost
one year with the setup explained in Fig. 3.7, the very quick aging cycling procedure
becomes a very efcient tool to screen out and verify the stability of different adsor-
bents against hydrothermal aging at least during the early stages of development.
The obtained results with the zeolite 13X (Baylith WE-G 592) have led to the
decision to exclude it in the further development stages of the adsorption heat
pump. The stability of different zeolite types in form of pellets against hydrothermal
76 3 Hydrothermal Stability of Adsorbent Coatings

1400
Ratio between the numbers of aging cycles [-]

No. of real-time heat pump cycles / No. of very quick aging cycles
to realize same loss of final water uptake
1200

1000

800

600

400

200
1 1.8 2.6 3.4 4.2 5
Loss of final water uptake [g/100g]

Fig. 3.18 Ratio between the number of long-term and the number of very quick aging cycles to
realize same loss in the nal water uptake

Table 3.7 Loss of nal water uptake and broken portion of different zeolite-type-pellets after 200
very quick aging cycles
Producer Zeolite type Broken portion (%) Loss in nal water uptake (g/100 g)
Grace Na-Y-1 0.2 0.4
Grace Na-Y-2 14.6 0.2
Grace Na-Y-3 1.4 0.9
Mitsubishi H-Y 5 12.3
Mitsubishi Na-Y 0.8 1.5
Bayer AG Na-A 2.3 1.5
Bayer AG MgNa-A 0.6 0.8
Bayer AG CaNa-A 1.4 1.2

aging has been thoroughly investigated. The broken Portion of the investigated
samples along with the loss in the nal water uptake of some of the tested zeolite
candidates are presented in Table 3.7 after 200 cycles according to the very quick
aging procedure.
It is evident that Na-Y, MgNa-A and CaNa-A zeolites are very good candidates
for developing adsorption heat pumps, when the binder type and content are
optimized. Figures 3.19 and 3.20 illustrates the X-Ray diffraction results carried out
on fresh and aged samples of both zeolites H-Y and Na-Y Zeolites of Mitsubishi
Plastic Incorporation.
3.4 Very-Quick Aging Cycling Procedure 77

3500
Smp 1 - fresh
Smp 2 - aged
3000
Y zeolite - literature reference

2500

2000
Intensity

1500

1000

500

0
5 10 15 20 25 30 35 40 45
Angle, 2 Theta

Fig. 3.19 X-ray diffraction results on fresh and aged samples of H-Y zeolite of Mitsubishi Plastics
Incorporation

4500
Smp 3 - fresh
4000 Smp 4 - aged
Y zeolite - literature reference
3500

3000

2500
Intensity

2000

1500

1000

500

0
5 10 15 20 25 30 35 40 45
Angle, 2 Theta

Fig. 3.20 X-ray diffraction results on fresh and aged samples of Na-Y zeolite of Mitsubishi
Plastics Incorporation
78 3 Hydrothermal Stability of Adsorbent Coatings

The XRD pattern detected for the aged H-Y zeolite sample (Fig. 3.19), showed
that the typical zeolitic crystalline peaks almost totally decayed to an amorphous
phase, clearly indicating that the aging treatment completely destroyed the zeolite
structure. Differently, the aged Na-Y zeolite sample (Fig. 3.20) presented almost
unchanged XRD pattern with respect to the fresh state, indicating that the aging
treatment did not alter the zeolite crystalline structure.

References

1. Puppe L (1986) Zeolithe Eigenschaften und praktische Anwendungen Chemie in unserer


Zeit, vol 20(4), pp 117127
2. Lcke L (1989) Entwicklung zeolithischer Membranen zur Trennung von Gasgemischen
Dissertation, RWTH Aachen
3. Piguzova LI (1962) Thermal stability of the structure of sorbents of the molecular sieve. Int
Chem Eng 2(2):189192
4. Wyart J, Sabatier G (1959) Zeitschrift fr Kristallographie-Cryst Mater 112:161168. doi:10.
1524/zkri.1959.112.jg.161
5. Wolf F, Frtig H, Nemitz G (1967) ber die thermische Stabilitt von Molekursieben des
Typs A in Gegenwart von Wasserdampf. Chem Technol 19(2):8387
6. Rothmeyer M (1985) Wrmetransformation mit dem Stoffpaar Zeolith-Wasser. Dissertation,
TU Mnchen
7. Nefedow BK, Valujskaja OM (1986) Die Untersuchung der thermischen und
Wasserdampfstabilitt synthitischer Faujasit-Zeolithe, granuliert mit Bindemittelzustzen,
vol 50, pp 8288
8. Henninger SK, Munz G (2009) Hydrothermal stability of sorption materials and composites
for the use in heat pumps and cooling machines. In: Heat powered cycles conference 2009,
79 September, CD-proceedings. Springer, Berlin. ISBN 978-0-9563329-0-5
9. Restuccia G, Vasta S, Freni A, Russo F, Aristov YI (2005) An advanced solid sorption chiller
using SWS-1L: performance analysis and hydrothermal cycling stability of the sorbent bed.
In: CD-proceedings of the international sorption heat pump conference, paper no. 78, Denver,
2224 June 2005
10. Freni A, Frazzica A, Dawoud B, Chmielewski S, Calabrese L, Bonaccorsi L (2013) Adsorbent
coatings for heat pumping applications: verication of hydrothermal and mechanical
stabilities. Appl Therm Eng 50:16581663
11. Sauer J, Dawoud B, van Heyden H, Klaschinsky N, Lohmoller S (2008) Adsorberelement und
Verfahren zur Herstellung eines Adsorberelements, Patent DE102008050926 A1
12. Okamoto K, Teduka M, Nakano T, Kubokawa S, Kakiuchi H (2010) The development of
AQSOA water vapor adsorbent and AQSOA coated heat exchanger, IMPRES2010. In:
International symposium on innovative materials for processes in energy systems, 29 Nov1
Dec 2010, Singapore. ISBN-10:981-08-7614-9
13. Dawoud B, Chmielewski S (2011) Entwicklung eines innovativen gasbetriebenen Zeolith
Heizgertes; Ein Beitrag zur Steigerung der Energieumwandlungsefzienz und
Emmisionsminderung. VDI Publishing Ltd, Dsseldorf
14. Dawoud B (2013) Entwicklung eines innovativen gasbetriebenen Zeolith-Kompaktgertes;
Ein Beitrag zur Steigerung der Energieumwandlungsefzienz und Emmisionsminderung
(Phase II). VDI Publishing Ltd, Dsseldorf
References 79

15. Dawoud B (2007) On the effect of grain size on the kinetics of water vapor adsorption and
desorption into/from loose pellets of FAM-Z02 under a typical operating condition of
adsorption heat pumps. J Chem Eng Jpn 40(13):12981306
16. Stricker M (2003) Entwicklung einer mehrmodularen Zeolith-Wasser-Adsorptions-
wrmepumpe. PhD-Thesis, RWTH Aachen University, VDI Verlag, Germany
Chapter 4
Mechanical Stability of Adsorbent
Coatings

An adsorbent material consolidated in form of coating could present poor


mechanical stability if a proper binder and/or an optimized binder content are not
adopted in the coating formulation. Mechanical stability of the adsorbent coating
may be affected by different typologies of mechanical stresses during its lifetime
(vibrations, impacts, etc.). For this reason, a single mechanical test is not sufcient
to prove completely the overall mechanical stability of the coated layer [1].
Accordingly, it is preferred to dene a protocol of various tests to measure different
mechanical properties of the adsorbent coating deposited over the metal substrate
[2, 3]. Table 4.1 lists specic static and dynamic mechanical tests that can be used
for the mechanical characterisation of the adsorbent coatings.

4.1 Static Mechanical Tests

The term static refers to mechanical tests where the load rate is low enough to
avoid stress effects due to the rate of application of the load on the coating. For this
reason, in addition to the generic bending and static shear tests other more specic
tests for the coating characterization, like scratch and pull off tests, are considered
(Fig. 4.1).

4.1.1 Flexural Test

The typical three points bending test can be applied to characterize the mechanical
properties of the adsorption coatings [4]. The three points bend test is a technique
normally used to evaluate the resistance of the material to the bending stress. In our
case, the test was used to determine the maximum beam deflection reached without
cracks being observed on the outer coating surface. In Fig. 4.2 is shown the bending
test of an aluminium foil coated with a SAPO-34 layer. The coated side considered
during the test was the tensile side of the bent sample where the coating layer is

The Author(s) 2015 81


A. Freni et al., Characterization of Zeolite-based Coatings
for Adsorption Heat Pumps, SpringerBriefs in Applied Sciences and Technology,
DOI 10.1007/978-3-319-09327-7_4
82 4 Mechanical Stability of Adsorbent Coatings

Table 4.1 Mechanical tests for the characterization of an adsorbent coating


Mechanical test Quantitative parameter measured
Static tests Flexural test Max deflection (mm)
Shear test Max stress applied (N/cm2)
Micro-hardness test Vickers hardness number (HV)
Scratch test Critical load (N)
Pull off test Strength (N)
Peel test Amount of detached coating (qualitative scale)
Dynamic tests Impact test Impact energy (J)
Vibration test Damage time (s), frequency (Hz)

Fig. 4.1 Scheme of main static mechanical tests

Fig. 4.2 Three points bending test setup

subjected to stretching stresses. The deformation rate applied was of 1 mm/min and
the maximum deflection observed before detection of a coating fracture was in the
range of few millimetres. Even if the maximum deflection cannot be considered an
absolute measure of the coating elastic properties because the intrinsic heterogeneity
of the adsorption coatings, however can be used as useful parameter in comparative
tests [5].
4.1 Static Mechanical Tests 83

Fig. 4.3 Shear test setup for


adsorption coated aluminium
foils

4.1.2 Shear Test

Shear test is typically used to evaluate mechanical resistance to tangentially applied


stresses of a glued joint between two thin sheet materials (metals, composites etc.)
[6]. The technique can be transferred to the coated supports by gluing two coated
samples by a high strength bonder (epoxy resin, cianoacrilate adhesive, ) [7]. The
bonder is used to transfer the shear stress to the coating so to verify the adhesive
force at the interface between the adsorbent coating and the metal support by
evaluating the maximum shear stress that the coating can resist without rupture. In
Fig. 4.3 is shown the tensile machine and the experimental setup during a shear test.
A minimum shear strength value of 75 N/cm2 is considered acceptable for an
adequate adhesion of the coating to the support [5].

4.1.3 Micro-hardness Test

The micro-hardness test consists of a static indentation test made with loads not
exceeding 1 kgf. The typical indenter used is Vickers, a square base diamond
pyramid with an apical angle of 136 (Fig. 4.4). The resulting indentation is
measured and converted to a hardness value. This method is commonly suggested
to evaluate the hardness of a material on a microscopic scale [8]. In particular, the
Vickers Diamond Pyramid hardness number is calculated as the applied load (N)
divided by the surface area of the indentation (mm2).
 
F
HV 0:1891
d2
84 4 Mechanical Stability of Adsorbent Coatings

Fig. 4.4 Vickers indenter

where
F = Load in N
d = Arithmetic mean of the two diagonals, d1 and d2 in mm
HV = Vickers hardness
To avoid invalid measurements of the coating hardness it should be considered
that a minimum layer thickness is required, generally not less than 150 m. It has to
be guarantee, in fact, that the indenter does not touch the support during the test,
resulting a micro-hardness value influenced by the substrate characteristics. In the
micro-hardness test, the indenter imprint has to be measured by a microscope for
the small size of the diamond pyramid. In case of adsorption coatings, the layer
heterogeneity and its intrinsic brittleness cause the formation of irregular indenta-
tion marks difcult to observe and measure. To better visualize the indentation and
measure its size, a metallographic microscope could be used and the specimen
surface can be covered with graphite powder before indentation. Figure 4.5 shows
an example of indentation detected by the metallographic microscope.

Fig. 4.5 Micro-hardness test


on a zeolite coating
4.1 Static Mechanical Tests 85

Fig. 4.6 Scratch test scheme Load

Sample motion

After denition of the proper specimen conguration and the indentation mea-
surement method, several samples need to be prepared and properly characterized
to have an affordable measurement of the micro-hardness of the coating. The test is
a measure of the punctual mechanical resistance of the coating and can be used as
an indication of the layer brittleness. Usually micro-hardness values measured for
some zeolite adsorbent coatings are in the range of 26 HV.

4.1.4 Scratch Test

This test is used to evaluate the adhesion properties between the coating and the
substrate. An indenter, charged with a load, is moved across the coating at a
constant velocity while the normal force is steadily increased (Fig. 4.6).
The indenter causes stress at the coating/substrate interface resulting in delam-
ination of the coating. The normal force at which the rst failure of the coating can
be detected is termed the critical load Lc. Normally the rst failure is observed by
an optical microscope. In case of zeolite coatings, due to the intrinsic heterogeneity
of the lm, the Lc observed was always in the range of 110 N, but more often a
value close to 23 N have been recorded.

4.1.5 Pull-off Test

The tensile adhesion strength of the adsorbent coatings on metal substrate can be
measured by using a pull-off tester (in Fig. 4.7) [9]. The pull-off tester acquire the
strength required to detach the coating (glued with the dolly) from the substrate. In
a typical measurement for adsorption coatings an aluminum dolly (10 mm diam-
eter) was glued onto the coating surface using a ciano-acrylate adhesive which was
cured for 24 h at room temperature. A clamping xture was designed to avoid the
misalignment during the uniaxial tensile test.
The fracture surface of the sample can be evaluated by an optical stereomicro-
scope to highlight the rupture mechanism. Occasionally small cavities were seen on
the adhesive surface after pull-off, these cavities could have altered the stress dis-
tribution and effective area of contact and consequently were excluded.
86 4 Mechanical Stability of Adsorbent Coatings

Fig. 4.7 Example of pull-off


tester

For adsorption coatings the expected value of adhesion strength is around


0.81 MPa: lower values can be considered indication of scarce adhesion at the
support interphase or interlayer poor bonding between the adsorbent material and
the binder.

4.1.6 Peel Test

The peel test is widely used for the mechanical characterization of the adhesion
phenomenon. It is usually applied to evaluate the force to pull the free end of an
adhesive tape glued on the coating surface. In our case, this test is used to measure
the interlayer forces in the adsorbent coating causing a delamination among the
coating layers by the pulling action of the adhesive tape. By using a preliminary
approach the measure to compare the different coatings could be the quantity
(weight) of material detached by the pulling force during the test. A more reliable
approach is, according to standard ASTM D3359, to evaluate the adhesion quality
of the coating by tape peel test measuring the detached area of the coating after the
adhesive tape was pulled out. In the tape peel tests (according to ASTM D3359) a
square grid was realized with 20 mm long cuts at a distance of 2 mm each others.
A tape was placed over the grid and rubbed with the eraser end of a pencil to ensure
good contact. The tape was removed and then the samples were evaluated using a
polarized light microscope. The removed grid area was rated according the ASTM
specications, from 0B (low adhesion) to 5B (high adhesion) to evaluate quanti-
tatively the detachment of the coating from the substrate. The determination of the
detached area can be identied by visual inspection. A more sophisticate but more
reliable approach is to use a digital image analysis to calculate accurately the
detached area. In the following Fig. 4.8 is reported the digitalization procedure used
by our research group to identify the detached area [10].
4.2 Dynamic Tests 87

Fig. 4.8 Digital imaging of a detached area after peel test

4.2 Dynamic Tests

The tests discussed up to now are static or quasi-static mechanical tests where the
application velocity of the stress not influences the nal results. To complete the
characterization of the adsorbent coatings, an evaluation of the dynamic properties
of the composites is important in view of their stability and resistance to injuries.
Two simplied experimental approaches are customized impact and vibrational
tests (a scheme is reported in Fig. 4.9).

4.2.1 Impact Test

This test is carried out in order to simulate sudden crush and injuries events to the
coated element. The drop weight impact test is the preferred experimental set-up. In
this test a weight (usually hemispherical or spherical indenter) is dropped down on

Fig. 4.9 Scheme of the


dynamic tests
88 4 Mechanical Stability of Adsorbent Coatings

Fig. 4.10 Impact test set-up

the coated sample. If the coating is the side impacted by the indenter the test is
called direct impact test. If the coating is on the back side of the impacted area the
test is called reverse impact test. The last set-up is not preferred because is heavy
influenced by the substrate thickness. If the thickness of the metal support is rel-
evant it will not be deformed by the impact and consequently the coating located on
the back side will not be cracked.
The weight is dropped at increasing heights until the coating of the sample
cracks. Knowing the impact height and indenter weight is possible to calculate the
impact energy by the simplied equation:

E mgh 4:1

where E is the impact energy of the drop-weight (measured in J), related with the
mass m (in kg) and drop height h (in m) of the indenter, g is the acceleration of
gravity (9.81 m/s2). A problem in this type of test is to ensure that the indenter once
impacted the surface does not continue to bounce on the damaged specimen,
leading to further impact conditions that make difcult to interpret the results. In
this regard, a variant of the test has been proposed in Fig. 4.10, showing that the
sample under test was placed on a proper support with a tilt angle of 45 to avoid
the rebound of the falling sphere over the sample [5].

4.2.2 Vibrational Test

The objective of the characterization of adsorbent coatings by vibration tests is to


evaluate the durability of deposited coatings when subjected to recurring stresses.
Crush and injuries events will be better simulated by the impact test. The experi-
mental replication of a real vibrational condition is quite difcult particularly
because not many information on the effective dimensions and mass of the
4.2 Dynamic Tests 89

Fig. 4.11 Scheme of a low frequency vibrational test

adsorbing machine are available. For this reason two different experimental set-up
have been developed [5], in order to investigate two different ranges of vibration
modes:
1. a low-frequency eld by a vibrating system able to reach 30 Hz
2. a high-frequency eld obtained by an ultrasonic applicator at 20 kHz.
In the low-frequency test a coated sample is xed on the vibrating system and
then put in vibration, while in the high-frequency test, the coated sample is
immersed in a not-wetting oil medium, which acts as transmitter, and exposed to
ultrasonic waves. In both cases, the measure is the time (s) needed to causes the
coating detachment from the support. A measure of the energy involved in the
coating cracking could be also possible and can be used as term of comparison for
the mechanical evaluation.
In Fig. 4.11 is reported an example of low-frequency test, where the coated
sample has one extremity xed to the shaker while at the free-end a weight is
clamped on it. Such a system allows increasing the flexural response of the alu-
minum substrate so to transmit stresses to the coating.
The specimen dimension are not restrictive, however, a reference sample could
be a rectangular foil having dimensions: 50 80 1.5 mm, whose surface is
partially coated with an adsorbent layer of size 50 70 0.5 mm. The testing
frequency can be increased from 0 to 30 Hz and the weight can be varied between
10 and 60 g. The time needed to reach the coating rst-failure will be used as
quantitative parameter. Figure 4.12 shows two coated samples after testing. Failure
was observed after 30 s at 15 Hz (weight = 55 g).
The arrangement of the high-frequency test is shown in Fig. 4.13. The ultrasonic
probe is positioned few millimeters over the coated sample. The sample is plunged
in the oil with the uncoated face up, so that the waves action is transmitted to the
90 4 Mechanical Stability of Adsorbent Coatings

Fig. 4.12 Samples of coated


aluminum after 30 s at 15 Hz

Fig. 4.13 High-frequency


vibrational test set-up

coating by the elastic deformation of the metal base. No particular restrictions are
evidenced for the sample dimensions. Reference parameters could be:
Thickness of the metal support: 1 mm
Optimal coating thickness: 0.5 mm
4.2 Dynamic Tests 91

Other important parameters are:


Oil volume: 4 ml
Specimen-probe distance: 5 mm
Output Power range: 530 W.
Tests carried out on coated aluminum demonstrated that the coating fails with a
measured power of 20 W, in 10 s at a frequency of 20 kHz.

4.3 Methods Comparison

In the following Table 4.2, all the mechanical methods proposed for the charac-
terization of the adsorbent coatings are summarized for an easier comparison.
Table 4.3 presents the results of mechanical characterizations carried out on the
two reference adsorbent coatings reported below:
Laboratory scale coating prepared according to Freni et al. [5]
Pilot scale coating prepared according to Okamoto et al. [11]

Table 4.2 Comparative scheme of mechanical methods proposed for zeolite coatings
characterization
Test Reliability Result typology Repeatability Handiness
Flexural test High Quantitative Medium Medium
Shear test High Quantitative High Medium
Micro-hardness Medium Quantitative Medium Good
Scratch test Medium Semi-quantitative Medium Medium
Pull off test Good Quantitative Good Good
Peel test Good Semi-quantitative Medium Good
Impact test Good Semi-quantitative Medium Good
Vibration test Medium Semi-quantitative Good Medium

Table 4.3 Results of


Test Freni et al. [5] Okamoto et al. [11]
mechanical characterizations
of two reference adsorbent Micro-hardness 0.5 HV 810 HV
coatings Shear test 75 N/cm2 85 N/cm2
Peel test 40 wt% 4020 wt%
Bending test 7 mm 5.5 mm
Impact test 130150 mJ 200220 mJ
Low Freq. V. Test 30 s @15 Hz 80 s @15 Hz
High Freq. V. Test 10 s @ 20 kHz 90 s @ 20 kHz
92 4 Mechanical Stability of Adsorbent Coatings

The coating from Okamoto et al. [11], evidenced more stable mechanical
properties compared with coating from Freni et al. [5]. Indeed, the coating from
Freni et al. have a quite brittle behavior that influences its mechanical stability
especially under dynamic loading. Instead the coating from Okamoto et al., char-
acterized by a high hardness, showed good adhesion with the substrates and very
good performance in static and dynamic testing conditions. Some tests were carried
out also in a third coating prepared in according to Sauer et al. [12]. This coating
evidenced good hardness and better adhesion properties compared with coating
from Okamoto et al.

4.4 Experimental Case History: Adhesion Quality Control


on a Coated Adsorber

20 sequenced Lamella samples from a dip coated adsorber according to Okamoto


et al. [11] have been investigated.
Furthermore each side of each lamella (front and rear) was discretized in six
positions; three on the left side (1L, 4L and 7L) and three on the right side (1R, 4R
and 7R).
At rst, we realized thickness measurements of the coatings. Details of thickness
distribution are reported in Fig. 4.14. The red dotted line is referred to a reference
value of 500 m. The thickness is quite constant for all laminas. The average value
is 449.7 m with a standard deviation of 39.6 m. Furthermore, to evidence
graphically the distribution of coating thickness on the heat exchanger laminas we
reported the data in a scheme as reported in Fig. 4.15.
Afterward we realized 3 pull-off test on each lamina. The measurements were
carried out on conguration 1L, 4L, 7L, respectively in front and rear conguration
for odd and even lamella position. The average pull-off strength is 0.91 Mpa, and
the observed fracture is mainly cohesive. With the purpose to better evaluate the

Fig. 4.14 Scatter plot of


coating thickness
4.4 Experimental Case History: Adhesion Quality Control on a Coated Adsorber 93

Fig. 4.15 Coating thickness


distribution

homogeneity of zeolite coating performance on the dip coated adsorber we


graphically referred each pull-off strength to each lamina (Figs. 4.16 and 4.17).
The pull-off strength is well distributed, and although the coating surface is quite
rough, the experimental results are homogeneous. It is only possible to evidence a
higher dispersion of data for the former lamellas (lamella 16). In fact, the standard
deviation varies signicantly at increasing the lamella position (Fig. 4.18). Higher
standard deviation was observed for the former lamella and progressively it is
reduced for lamellas located in the middle of the adsorber. At this moment, the data
population is not statistically signicant to identify a valid interpretation of the
observed trend. E.g. it could be related with the coating deposition or curing phases.
In particular, during the drying procedure a thermal/flux gradient between lamellas
94 4 Mechanical Stability of Adsorbent Coatings

Fig. 4.16 Coating pull-off


strength distribution

Fig. 4.17 Coating pull-off


strength
4.4 Experimental Case History: Adhesion Quality Control on a Coated Adsorber 95

Fig. 4.18 Standard deviation


of pull-off strength

could induces a different heating treatment. The experimental data evidenced a


slight fluctuation on the mechanical performances of the coating. Eventually, fur-
ther tests could be useful to clarify these considerations.
To conclude the experimental characterization, the fracture surface induced by
pull-off tests showed a clear cohesive fracture.

References

1. Chalker PR, Bull SJ, Rickerby DS (1991) A review of the methods for the evaluation of coating-
substrateadhesion.MaterSciEngA140:583592.doi:10.1016/0921-5093(91)90482-3
2. Forster A, Neudeck T, Blatcher S (2008) Approaches to the physico-chemical and mechanical
characterisation of functional coatings, part I. Med Device Technol 19:2528, 3031
3. Forster A, Neudeck T, Blatcher S (n.d.) Approaches to the physico-chemical and mechanical
characterisation of functional coatings, part II. Med Device Technol 19:424, 467. http://
www.ncbi.nlm.nih.gov/pubmed/19133628 (accessed 12 Jan 2015)
4. Dvok M (2012) New methods testing of adhesion of the coating to sheet metal by bending.
J Surf Eng Mater Adv Technol 02:6164. doi:10.4236/jsemat.2012.22011
5. Freni A, Frazzica A, Dawoud B, Chmielewski S, Calabrese L, Bonaccorsi L (2013) Adsorbent
coatings for heat pumping applications: verication of hydrothermal and mechanical
stabilities. Appl Therm Eng 16581663. doi:10.1016/j.applthermaleng.2011.07.010
6. DeVries P, Borgmeier KL (2000) Adhesion testing. In: ASM handbook, vol 8
7. Marot G, Lesage J, Dmarcaux P, Hadad M, Siegmann S, Staia MH (2006) Interfacial
indentation and shear tests to determine the adhesion of thermal spray coatings. Surf Coat
Technol 201:20802085. doi:10.1016/j.surfcoat.2006.04.046
8. Brotzen FR (2000) Evaluation of mechanical properties of thin lms. In: ASM handbook, vol 6
9. Calabrese L, Bonaccorsi L, Capr A, Proverbio E (2014) Adhesion aspects of hydrophobic
silane zeolite coatings for corrosion protection of aluminium substrate. Prog Org Coat
77:13411350. doi:10.1016/j.porgcoat.2014.04.025
10. Calabrese L, Bonaccorsi L, Proverbio E (2012) Corrosion protection of aluminum 6061 in
NaCl solution by silanezeolite composite coatings. J Coat Technol Res 9:597607. doi:10.
1007/s11998-011-9391-5
96 4 Mechanical Stability of Adsorbent Coatings

11. Okamoto K, Teduka M, Nakano T, Kubokawa S, Kakiuchi H (2010) The development of


AQSOA water vapor adsorbent and AQSOA coated heat exchanger. In: Proceedings of
International Symposium on Innovative Materials for Processes in Energy Systems
(IMPRES2010), Singapore (ISBN-10981-08-7614-9)
12. Sauer J, Dawoud B, van Heyden H, Klaschinsky H, Lohmoller S (2008) Adsorberelement und
Verfahren zur Herstellung eines Adsorberelements, Patent DE102008050926 A1