Applied Catalysis B: Environmental 66 (2006) 111118

www.elsevier.com/locate/apcatb

Single-reactor process for sequential aldol-condensation and

hydrogenation of biomass-derived compounds in water
C.J. Barrett, J.N. Chheda, G.W. Huber 1, J.A. Dumesic *
Chemical and Biological Engineering Department, University of Wisconsin, 1415 Engineering Drive,
Madison, WI 53706, USA
Received 11 January 2006; received in revised form 27 February 2006; accepted 1 March 2006
Available online 19 April 2006

Abstract
A bifunctional Pd/MgO-ZrO2 catalyst was developed for the single-reactor, aqueous phase aldol-condensation and hydrogenation of
carbohydrate-derived compounds, furfural and 5-hydroxymethylfurfural (HMF), leading to large water-soluble intermediates that can be
converted to liquid alkanes. The cross aldol-condensation of these compounds with acetone results in formation of water-insoluble monomer
(C8C9) and dimer (C13C15) product species, which are subsequently hydrogenated in the same batch reactor to form water-soluble products with
high overall carbon yields (>80%). After a cycle of aldol-condensation followed by hydrogenation, the Pd/MgO-ZrO2 catalyst undergoes a loss in
selectivity by 18% towards heavier product (dimer) during subsequent runs. However, the catalytic activity and dimer selectivity are completely
recovered when the catalyst is recycled with an intermediate calcination step at 873 K. The optimum temperatures for aldol-condensation of
furfural with acetone and for condensation of HMF with acetone are 353 and 326 K, respectively, representing a balance between dimer selectivity
and overall carbon yield for the process. The product selectivity can be controlled by the molar ratio of reactants. When the molar ratio of furfuralacetone increases from 1:9 to 1:1, the selectivity for the formation of dimer species increases by 31% and this selectivity increases further by 12%
when the ratio increases from 1:1 to 2:1. It is likely that this active, stable, and heterogeneous catalyst system can be applied to other base and/or
metal catalyzed reactions in the aqueous phase.
# 2006 Elsevier B.V. All rights reserved.
Keywords: Renewable energy; Heterogeneous catalysis; Biomass; Aldol-condensation; Hydrogenation

1. Introduction
Petroleum currently contributes more than 40% of the
energy for the world [1], and as this non-renewable resource
becomes less plentiful, it is becoming more important to
identify renewable alternatives [2,3]. One of the strategies for
replacement of petroleum fuels in the transportation sector is to
develop liquid, biomass-derived fuels that can use the existing
delivery infrastructure, while not needing extensive vehicle
engine modification. Recently, we developed a process that
produces liquid alkanes by aqueous phase processing of

* Corresponding author. Tel.: +1 608 262 1095/2; fax: +1 608 262 5434.
E-mail addresses: cjbarrett@wisc.edu (C.J. Barrett), jubenc@cae.wisc.edu
(J.N. Chheda), huber@ecs.umass.edu (G.W. Huber), dumesic@engr.wisc.edu
(J.A. Dumesic).
1
Present address: Department of Chemical Engineering University of Massachusetts, 159 Goessmann Lab 686 N. Pleasant, St. Amherst, MA 01003-9303,
United States.
0926-3373/$ see front matter # 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcatb.2006.03.001

biomass-derived carbohydrates. Alkane products from this

process have cetane numbers, a measure of the ignition quality
of diesel fuel, ranging from 63 to 97 for C7C15, whereas the
overall cetane number for diesel fuel is typically from 40 to 55.
The high cetane numbers for these biomass-derived alkanes
make them excellent blending agents to produce sulfur-free
diesel fuel for transportation applications [4].
Our process to convert renewable biomass-derived carbohydrates to alkanes with molecular weights in the range of
diesel fuel (C7C15) uses a series of consecutive reactions,
starting with acid-catalyzed dehydration of carbohydrates to
form carbonyl groups, followed by aldol-condensation over a
solid base catalyst (such as mixed MgAl-oxide) to form CC
bonds, and subsequent hydrogenation of C C and C O bonds
over a metal catalyst (such as Pd/Al2O3) to form large watersoluble organic compounds. These molecules are then
converted to liquid alkanes by aqueous-phase dehydration/
hydrogenation (APD/H) reactions over a bifunctional catalyst
containing acid and metal sites in a flow-reactor with two feed

112

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

Nomenclature
APD/H aqueous-phase dehydration/hydrogenation reaction
BET
BrunauerEmmettTeller
HMF 5-hydroxymethylfurfural
MCM mesoporous crystalline material
MIBK methyl isobutyl ketone
PES
polyethersulphone
streams: an aqueous stream containing the organic reactant and
a hexadecane sweep stream. In this reactor, the large aqueous
organic reactants become more hydrophobic during APD/H
processing, and the hexadecane sweep stream removes these
hydrophobic species from the catalyst before they react further
to form coke [5]. Fig. 1 presents the essential features of the
bifunctional reaction pathways involved in the production of
large, water-soluble molecules from biomass-derived furfural
or 5-hydroxymethylfurfural, HMF (from xylose [6] or glucose
[7], respectively).
Aldol-condensation plays an important role in the aforementioned process. Many aldol-condensation reactions are
practiced industrially using homogenous base catalysts such as
sodium and calcium hydroxide. However, these processes
generate significant wastewater streams that must be neutralized, which leads to additional disposal cost. Developing a new
active and stable solid base catalyst can simplify the process,
promote clean manufacturing, and decrease the production
cost. Among the solid base catalysts that are active for aldolcondensation are alkali [811], alkaline earth oxides [12],
anion-exchange resins [1315], phosphates [16], MCM41
[17,18] and hydrotalcites [1922]. Most of the applications
with solid base catalysts for aldol-condensation have been
either in liquid phase using organic solvent or in vapor phase.
Magnesiazirconia and magnesiatitania have been tested as
base catalysts for acetone condensation in the vapor phase
[23,24]. Few studies have been carried out with solid base
catalysts for aldol-condensation in pure water as the solvent.

Aldol-condensation in an aqueous environment has been

reported for carbohydrate-derived molecules using heterogeneous catalysts such as weak anion exchange resins [13,14],
organocatalysts like cyclic secondary amines [25], and
homogeneous catalysts like NaOH [8,9]. Cross-condensation
of furfural with acetone, which is an important reaction for this
paper, has been conducted using heterogeneous aminofunctionalized mesoporous base [26] as well as chiral LProline [27] catalysts although in organic solvents. Aldolcondensation and subsequent hydrogenation using a bifunctional metal-base catalyst have been performed in an organic
environment for production of MIBK [28] and synthesis of aalkylated nitriles with carbonyl compounds [29] using Pd on
hydrotalcites.
We demonstrate in this paper that aldol-condensation and
hydrogenation steps in water can be combined in a singlereactor using a bifunctional Pd/MgO-ZrO2 catalyst that has
high activity and selectivity, as well as excellent recyclability
and hydrothermal stability. Stability studies show the catalyst
can be recycled without any loss in activity or selectivity. We
show how the performance of the catalyst is altered by changing
the reaction temperature and the molar feed ratio of the
reactants for the aldol-condensation step.
2. Experimental methods
2.1. Catalyst preparation
Magnesia-zirconia (MgO-ZrO2) catalyst was synthesized
using the solgel technique described by Aramenda et al.
[23,24], starting with magnesium nitrate {Mg(NO3)2
6H2O,
Aldrich} and zirconyl nitrate {ZrO(NO3)2, Aldrich}. The
catalyst was prepared by dissolving 50.9 g of magnesium
nitrate and 4.04 g of zirconyl nitrate in 1 l of deionized (DI)
water. The mixture was stirred at room temperature, and NaOH
(25 wt%) solution was added until the pH was equal to 10. The
gel was aged for 72 h and subsequently vacuum filtered. The
precipitate formed was washed with DI water until the Na ion
concentration in the filtrate was below 10 ppm, as measured by

Fig. 1. Reaction network for aldol-condensation of furfural (or HMF) and acetone, followed by hydrogenation of aldol-adducts.

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

113

A 5 wt% Pd/MgO-ZrO2 catalyst was prepared by incipient

wetness impregnation of Pd (using 5 wt% Pd in tetraaminepalladium(II) nitrate solution from Strem Chemicals) onto the
above-mentioned MgO-ZrO2 support. It was then calcined in
flowing O2 (120 cm3 (NTP) min1) with a 2 h ramp and a 2 h
hold at 723 K. The catalyst so obtained was used for all the
aldol-condensation and sequential hydrogenation runs, as
described below.
2.2. Reaction system and analysis method
Fig. 2. HMF disappearance versus time during aldol-condensation with acetone
over fresh and calcined, recycled mixed Mg-Al-oxide and MgO-ZrO2 catalysts.
Mixed Mg-Al-oxide: run 1 (black squares), run 2 (grey squares), run 3 (white
squares); MgO-ZrO2: run 1 (black circles), run 2 (grey circles), run 3 (white
circles). Mixed Mg-Al-oxide runs carried out at room temperature and pressure,
while MgO-ZrO2 runs carried out at 323 K and atmospheric pressure. HMF:acetone molar ratio equal to 1:10; 11.2 wt% organics in the aqueous phase;
organic/catalyst mass ratio equal to 6.

inductively coupled plasma (ICP) analysis [Perkin-Elmer

Plasma 400 ICP Emission Spectrometer]. It was then dried
in an oven at 393 K from 16 to 24 h. Calcination of the catalyst
was carried out in O2 (100 cm3 (NTP) min1) with a 3 h ramp
and a 3 h hold at 873 K. The catalyst thus obtained was used for
the initial activity runs (see Fig. 2) for aldol-condensation of
HMF with acetone, using a HMF:acetone molar ratio of 1:10.

All reactions (see Table 1) were carried out in a Parr batch

reactor (Model # 4566) with an external temperature and
stirring controller (Model # 4836). The reactor was initially
loaded with the reaction mixture and air was purged out by
adding helium up to 55 bar for three times before starting the
condensation reaction. The reactor was then pressurized to
8 bar with He, heated to the reaction temperature, and stirred at
1000 rpm. After reaching the reaction temperature, the reactor
was pressurized to 10 bar. Aldol-condensation was stopped
after 2426 h of reaction time, and the reactor was then cooled
to room temperature. The hydrogenation reaction was started
by a similar purging procedure with H2 and pressurizing the
reactor to 44 bar before heating. The stirring speed was
maintained at 1000 rpm and the reactor was heated to 393 K at
which time H2 was added to reach a pressure of 55 bar.

Table 1
Experimental data for aldol-condensation and hydrogenation batch reactions
Run #

1
2
3
4
5
6
7
8
9
10
11
12
13

14
15
16
17

Feed

Fur:Ace [first]
Fur:Ace [second]
Fur:Ace [third]
Fur:Ace [fourth]
Fur:Ace
Fur:Ace
Fur:Ace
Fur:Ace
Fur:Ace
Fur:Ace
Fur:Ace
Fur:Ace0.5% Pdd
Fur:Acehexadece

HMF:Ace
HMF:Ace
HMF:Ace
HMF:Ace

Molar
ratio

1:1
1:1
1:1
1:1
1:1
1:1
1:9
2:1
2:1
1:1
1:1
1:1
1:1

1:1
1:1
1:1
1:1

Org/Cata

6
6
6
9
6
6
6
6
6
18
36
6
6

6
6
6
6

Time
[h]b

26
26
26
26
24
26
24
24
56
25
26
25
24

26
26
26
26

Temperature
[K] c

326
326
326
326
353
393
353
353
353
393
393
393
353

298
326
353
393

Volume
[ml]

250
200
125
90
100
100
100
100
100
100
100
112
100

70
100
100
100

Disappearance [%]
Furfural

Selectivity
C5 units [%]

Overall carbon
yield [%]

C5

C8

C13

79
58
58
76
95
98
96
66
86
90
88
98
100

23
43
45
25
5
3
4
37
16
11
14
2
0

34
31
31
32
35
35
67
15
12
30
32
33
15

43
26
24
43
60
62
29
48
72
59
54
65
85

91
93
90
91
88
80
76
91
85
85
82
82
71

HMF

C6 units [%]
C9
C6
42
20
21
18
14
21
11
38

C15
38
61
65
51

100
94
84
67

51
79
88
93

All runs were carried out in a Parr batch reactor over 5 wt% Pd/MgO-ZrO2, 5 wt% organics in the aqueous solution, condensation pressure of 10 bar, hydrogenation
time of 24 h, temperature of 393 K, and pressure of 55 bar (except run 12 using 0.5 wt% Pd and hydrogenated for 40 h). Recycle runs (run 14) were carried out using
the same catalyst, first run with fresh catalyst, second and third runs with recycled catalyst without calcinations, and fourth run with calcination.
a
Organic to catalyst ratio by mass.
b
Time for aldol-condensation.
c
Reaction temperature for aldol-condensation.
d
Reaction carried out over 0.5 wt% Pd/MgO-ZrO2.
e
Aldol-condensation carried out in water and hydrogenation conducted in hexadecane solvent at 393 K, 55 bar.

114

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

Hydrogenation of the furfural:acetone 1:1 system was complete

in 46 h at 393 K, and this temperature was employed for all
hydrogenation runs with no further optimization. Hydrogenation was stopped after a constant carbon yield in the aqueous
phase was reached, which was ensured for all runs by allowing
the reaction to proceed for 24 h with monitoring. For Table 1,
run 13, after condensation was complete, the aqueous layer was
evaporated off, leaving catalyst and precipated monomer and
dimer in the reactor. At which point hexadecane was added in
equal volume to that of the evaporated aqueous layer and the
subsequent hydrogenation reaction was conducted.
Samples were withdrawn from the sampling port during the
condensation and hydrogenation reaction. Samples were
filtered (using 0.2 mm PES syringe membrane filter) before
being analyzed by GC (Shimadzu GC-2010 with an FID
detector and a DB-5 column from Alltech). For catalyst recycle
experiments without calcination (Table 1, runs 2 and 3), the
reaction mixture was filtered after the hydrogenation run and
the catalyst was dried in an oven at 393 K for 1216 h before
reuse. Additionally for the recycle run with catalyst calcination
(Table 1, run 4), the catalyst was calcined after use as described
above for Pd/MgO-ZrO2. The ICP analysis for Na, Mg and Pd
in the final reaction mixture showed negligible leaching of the
catalyst components. Total organic carbon (TOC) analysis
(Shimadzu TOC-6001 with autosampler) was performed on
final reaction mixtures to quantify the total carbon present and
to calibrate the GC for reaction products. Furfural:acetone
dehydrated monomer (4-(2-furyl)-3-buten-2-one) was hydrogenated for calibration purposes. The self-condensation
product of acetone was not identified in HPLC during the
condensation runs as confirmed by running standards of
diacetone alcohol (4-hydroxy-4-methyl-2-pentanone). Both
chemicals were purchased from Aldrich.
The initial studies (see Fig. 2) of aldol-condensation with
HMF:acetone (molar ratio of 1:10) were carried out in 50 ml
glass reactor vessels using an oil bath to control the reaction
temperature. Regular samples were withdrawn, filtered (using
0.2 mm PES syringe membrane filter) and the HMF disappearance was monitored using HPLC (Waters 2695 system with
a Zorbax SB-C18 5 mm column from Agilent and PDA 960 and
RI 410 detectors).
Overall carbon yield and selectivity were calculated based
on C5 (for furfural) or C6 (for HMF) units (see Fig. 1). For
furfural:acetone reactions:
Overall carbon yield%

3moles C3 5  moles C5 8moles C8 13moles C13

3moles C3 fed 5  moles C5 fed
 100

C5 selectivity%

moles C5
 100
moles C5 moles C8 2  moles C13

An analogous definition applies for HMF:acetone reactions

on a C6 basis.

2.3. Catalyst characterization

2.3.1. CO and CO2 chemisorption and BET surface area
The irreversible uptakes of CO and CO2 on catalysts at
300 K were measured using a standard gas adsorption apparatus
described elsewhere [30]. Prior to CO or CO2 adsorption
measurements, the catalyst was reduced in flowing H2, with an
8 h ramp and 2 h hold at 393 K. After reduction, the
temperature was ramped to 573 K for 30 min and held for
30 min, while evacuating the cell. The cell was cooled to room
temperature, and the adsorbant was then dosed onto the catalyst
in 1015 doses until the equilibrium pressure was approximately 5 Torr. Gas in the cell was then evacuated for 30 min at
room temperature to a pressure of 106 Torr, and the adsorbant
was again dosed on the sample to determine the amount of
reversibly adsorbed CO or CO2. Irreversible uptake was
determined by subtracting the second isotherm from the first.
BET surface areas were measured by N2 adsorption at 77 K on
this same system.
2.3.2. X-Ray diffraction (XRD)
X-ray diffraction data were collected with a Cu Ka source
using a Scintag PADV diffractometer operating at 40.0 mA and
35.0 kV. Diffraction patterns were collected in continuous scan
mode with steps of 0.028 s1. The Scherrer equation was used
to estimate crystal size.
2.3.3. Thermo-gravimetric analysis (TGA)
A thermo-gravimetric analyzer from Netzsch Thermal
Analysis (model TG 209 with a TASC 414/3 temperature
controller) was used to analyze the amount of coke formed on
the catalyst surface. Approximately 4.5 mg of spent catalyst
was weighed and heated to 423 K in 13 min in presence of
flowing O2. The temperature was held at that point for
additional 30 min and ramped to 723 K at a rate of 10 K min1.
The amount of carbon on the catalyst was obtained by
comparing TGA data for fresh versus spent catalyst samples.
3. Results and discussion
3.1. Overview of reaction chemistry
Fig. 1 outlines the bifunctional reaction pathways involved
in the production of large, water-soluble molecules from
xylose-derived furfural or glucose-derived HMF. In this
example, the base catalyst abstracts an a-H atom from acetone
to form an intermediate carbanion (enolate ion) species, which
then attacks the carbon atom of the carbonyl group of furfural to
form a CC bond. Molecules such as furfural or HMF cannot
form carbanion intermediates for aldol-condensation because
they do not have a-H atoms. This reaction between furfural and
acetone results in formation of a C8-species that undergoes
dehydration to form an ab unsaturated aldehyde (monomer)
species, which has limited solubility in water. The C8-monomer
can further condense with a second furfural molecule to form a
dehydrated C13 aldol-adduct (dimer), which has very low
solubility in water because of its non-polar structure. Following

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

these aldol-condensation reactions, hydrogen is introduced into

the reactor, leading to hydrogenation of the furan rings in the
aldol-adducts over the metallic sites on the catalyst, thereby
increasing the solubility of the products in water to levels
higher than 35 wt%. This two-step process results in formation
of large water-soluble organic compounds that can subsequently undergo dehydration and hydrogenation reactions in
presence of hydrogen during APD/H processing to form
alkanes (C8 and C13 alkanes in the present example). An
analogous series of reactions can be depicted for HMF, leading
to the formation of C9 and C15 molecules. The hydrogen
required for the process can be obtained from biomass through
aqueous phase reforming (APR), as demonstrated in our
previous work [31,32].
3.2. Reaction experiment
In our previous work, we carried out the above aldolcondensation reactions using a mixed Mg-Al-oxide catalyst
derived from hydrotalcite synthesis, and we then added a Pd/
Al2O3 catalyst to carry out the subsequent hydrogenation step
in a batch reactor. However, as seen from Fig. 2, the mixed MgAl-oxide catalyst lost approximately 79% and subsequently
96% of its initial activity after the second and third uses of this
catalyst for the aldol-condensation of HMF with acetone
(HMF:acetone ratio of 1:10). This loss of activity persisted after
calcination of the catalyst between recycle runs. In addition, the
mixed Mg-Al-oxide catalyst when physically mixed with Pd/
Al2O3 showed a 15% loss of selectivity to formation of dimer
and a 10% loss in overall carbon yield, thereby indicating a
negative effect of the Pd/Al2O3 catalyst on the performance of
the mixed Mg-Al-oxide catalyst for aldol-condensation. The
deactivation upon recycling of mixed MgAl-oxide could be
caused by irreversible structural changes resulting from
interactions with water in the aqueous environment, because
these catalysts are known to undergo structural and subsequent
activity changes in presence of water. As seen from Fig. 2, an
MgO-ZrO2 catalyst exhibits excellent recyclability upon
repeated use for aldol-condensation reactions in the aqueous
phase. It will be shown below that the selectivity of a Pd/MgOZrO2 catalyst for the formation of C15-dimer species also
remains constant with repeated use for sequential aldolcondensation and hydrogenation reactions.
Fig. 3 shows the aqueous phase concentration of carbon
(normalized to the initial concentration of carbon in the batch
reactor) versus time during aldol-condensation over a bifunctional Pd/MgO-ZrO2 catalyst at various temperatures, followed
by sequential hydrogenation in the same batch reactor at 393 K.
As aldol-condensation proceeds, monomer and dimer species
form and precipitate out of the aqueous solution, and the
amount of carbon in the aqueous phase decreases accordingly.
It is important to note that during this reaction the Pd on the
catalyst is inert, because the performance of the Pd/MgO-ZrO2
catalyst is identical to the performance of MgO-ZrO2 during
aldol-condensation. Approximately 80% of the furfural has
disappeared after a period of 24 h under these reaction
conditions. The reactor is then pressurized to about 55 bar

115

Fig. 3. Overall carbon yield in the aqueous phase versus time for aldolcondensation at various temperatures of HMF with acetone (HMF:acetone
molar ratio equal to 1:1), followed by hydrogenation at 393 K. Black circles,
298 K; dark grey squares, 326 K; light grey triangles, 353 K; white diamonds,
393 K.

with hydrogen to initiate subsequent hydrogenation of the furan

rings and thereby increase the solubility of monomer and dimer
species in the aqueous phase. As seen in Fig. 3, this
hydrogenation step leads to an increase in the concentration
of carbon in the liquid phase. For example, while the carbon
concentration in the aqueous phase after aldol-condensation at
326 K decreases to about 44% of the initial carbon concentration, this value increases to about 94% after the hydrogenation
step. This figure thus illustrates the ability of the bifunctional
Pd/MgO-ZrO2 catalyst to facilitate a single-reactor, aqueous
phase process that combines aldol-condensation with sequential hydrogenation, in which the aqueous phase carbon lost
during the aldol-condensation step is returned to the aqueous
phase during the hydrogenation step. Table 1 shows the details
of various runs conducted. TGA identified 48, 21 and 95% of
the carbon missing from the carbon balance to be located on the
catalyst for runs 1, 7 and 15 in Table 1, respectively. For run 7,
furfural:acetone 1:9, roughly 63% of the missing carbon is
caused by the initial purging of gas from the reactor (because of
the high concentration and volatility of acetone), leading to an
overall carbon yield equal to 96%.
3.3. Catalyst stability studies
Experiments were conducted to study the stability and
recyclability of the bifunctional 5 wt% Pd/MgO-ZrO2 for aldolcondensation of acetone with furfural (molar ratio 1:1) at
326 K, followed by hydrogenation at 393 K. The catalyst was
recycled for use in runs 2 and 3 without any intermediate
regeneration, whereas the catalyst was subjected to a
calcination treatment prior to run 4. Fig. 4 (Table 1, runs 1
4) shows that selectivity for the formation of the dimer adduct
decreases by about 18% for recycle runs 2 and 3, while still
maintaining good overall carbon yield (>90%), and returns to
original levels for run 4. This result shows that the catalyst
retains most of its activity and selectivity for at least three runs
without requiring regeneration and can be completely
regenerated through calcination. As shown in Table 2, metal
sites (50  2 mmol/g), surface area (300  30 m2/g), and
average particle size (11  2 nm) of the catalyst for before

116

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

and after reaction did not change appreciably, while the phases
found (MgO (2 0 0, 2 2 0), ZrO2 (1 1 1, 2 2 0)) remained
constant. Hence XRD, CO chemisorption and BET measurements show that catalyst has excellent recycling ability and
hydrothermal stability. Aldol-condensation does not take place
homogeneously in the aqueous phase by dissolved basic
species, because the rate of aldol-condensation was negligible
after the MgO-ZrO2 catalyst was removed from the aqueous
solution, which shows further that the catalyst is stable.
3.4. Effects of process variables

Fig. 4. Selectivity based on C5 (furfural:acetone) units in the aqueous phase

after aldol-condensation followed by hydrogenation over 5 wt% Pd/MgO-ZrO2
catalyst. Selectivity and overall carbon yield defined in Section 2.2. Furfural:acetone (molar ratio of 1:1) over fresh and recycled catalyst. 5 wt% organics
in aqueous phase; 326 K and 10 bar He for condensation and 393 K and 55 bar
H2 for hydrogenation; organic/catalyst mass ratio of 6, except run 4 which had a
ratio of 9. (Bar 1, black) fresh catalyst with calcinations; (bar 2, dark grey) first
recycle without calcinations; (bar 3, light grey) second recycle without calcinations; (bar 4, white) third recycle with calcination.

Fig. 5A (Table 1, runs 1, 5, 6) shows experimental results

obtained at reaction temperatures from 298 to 393 K for aldolcondensations of furfural with acetone at a molar ratio of 1:1.
The rate of reaction increases with temperature; however, the
overall carbon yield in the aqueous solution after aldolcondensation (followed by hydrogenation) decreases at
temperatures above 353 K, probably caused by the formation
of coke on the catalyst during aldol-condensation. As the
temperature is increased from 326 to 353 K, the selectivity for
dimer increases by 17% with no significant change in the

Table 2
Characterization of 5 wt% Pd/MgO-ZrO2
Chemisorption and BET

Catalyst Before Reaction

After Fur:Ace Reaction

After HMF:Ace Reaction

Metal sites (mmol/g)

Base sites (mmol/g)
Surface area (m2/g)

49.0
103
292

51.7

329

48.8

299

XRD, identified phase, Miller indice, 2u

Average particle size (nm)

Before reaction

After Fur:Ace (run 1)

After HMF:Ace (run 15)

MgO (2 0 0), 2u = 30.658

MgO (2 2 0), 2u = 42.798
ZrO2 (1 1 1), 2u = 51.088
ZrO2 (2 2 0), 2u = 62.058

9
10
10
10

10
12
13
10

9
11
12
13

All catalysts were calcined and reduced before chemisorption, BET and XRD analysis. Mean diameter by XRD was estimated by line broadening of powder XRD
peaks using the Scherrer equation (1 nm).

Fig. 5. Selectivity based on C5 (furfural:acetone) or C6 (HMF:acetone) units in the aqueous phase after aldol-condensation followed by hydrogenation over 5 wt%
Pd/MgO-ZrO2 catalyst. Selectivity and overall carbon yield were in Section 2.2. (A) Furfural:acetone (molar ratio of 1:1) over fresh catalyst at various condensation
temperatures followed by hydrogenation at 393 K. (B) HMF:acetone (molar ratio of 1:1) over fresh catalyst at various condensation temperatures followed by
hydrogenation at 393 K.

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

overall carbon yield. In contrast, as the temperature is increased

further from 353 to 393 K, the dimer selectivity remains the
same but the overall carbon yield decreases by 8%. Thus, it
appears that the optimum temperature for aldol-condensation of
furfural is near 353 K, with this temperature providing a
compromise between the selectivity for heavier product and
overall carbon yield. In the case of aldol-condensation between
HMF and acetone (Fig. 5B; Table 1, runs 1417), the overall
yield of carbon is only 67% at 393 K. At lower temperatures,
similar trends were seen for furfural, with a marked increase in
selectivity for dimer by 23% as temperature increases from 298
to 326 K, and a 10% decrease in overall carbon yield as
temperature is increased further from 326 to 353 K. Accordingly, the optimum temperature for aldol-condensation of HMF
with acetone is near 326 K. This study shows that the aldolcondensation temperature has a significant effect on the
selectivity of the reaction and the overall yield of the process,
with the optimum temperature for condensation with acetone
being higher for furfural compared to HMF. At these optimum
temperatures, the furfural:acetone reaction achieves a higher
final conversion (by 16%) but a lower dimer to monomer ratio
(1.8 versus 3.4) as compared to the HMF:acetone reaction.
Results in Fig. 6 (Table 1, runs 5, 79) show that the molar
ratio of reactants for aldol-condensation plays a significant role in
controlling the reaction selectivity. The presence of excess
acetone (furfural:acetone molar ratio of 1:9) leads primarily to
the formation of monomer, because it is more probable that a
furfural molecule will react with an acetone molecule in contrast
to reacting with a monomer species. When the molar ratio of
furfural:acetone is increased from 1:9 to 1:1, the selectivity for
the formation of dimer species increases by 31%, and this
selectivity increases further by 12% when the furfural:acetone
ratio increases from 1:1 to 2:1. As the furfural:acetone ratio is
increased the condensation step requires additional time as
shown by an increase in dimer selectivity by 24% when the
condensation step is carried out for 56 h instead of 24 h.

117

Experiments were carried out to study the effects of varying

the organic/catalyst ratio, the palladium loading, and of
performing the hydrogenation step in hexadecane instead of
water. Increasing the organic/catalyst mass ratio from 6 to 36
(Table 1, runs 6, 10, 11) does not have an effect on the
selectivity and the overall carbon yield of the process.
Decreasing the amount of Pd on the MgO-ZrO2 catalyst from
5 to 0.5 wt% (Table 1, runs 6, 12) increased by about an order of
magnitude the time required to reach high overall yields of
carbon in the aqueous phase at 393 K (i.e., from 5 to nearly
40 h). In Table 1, run 13, the aqueous solution was removed at
the end of the aldol-condensation step, leaving the insoluble
monomer and dimer species on the catalyst surface; and the
reactor was then filled with hexadecane, followed by
hydrogenation at 393 K. This treatment led to the formation
of hydrogenated monomer and dimer species in the hexadecane
solvent, with an overall carbon yield of around 71%, indicating
that the hydrogenated form of monomer and dimer can be
blended with diesel fuel without the need to convert these
species into alkanes, thereby eliminating the need for the
further APD/H processing step.
4. Conclusions
We have shown that Pd/MgO-ZrO2 is an active, selective,
and hydrothermally stable catalyst for aldol-condensation over
basic sites (MgO-ZrO2) followed by sequential hydrogenation
over metal sites (Pd). This bifunctional catalyst thus allows
carbohydrate-derived compounds, like furfural and HMF, to be
converted in a single reactor to large water-soluble intermediates for further aqueous phase processing to produce liquid
alkanes. The selectivity and overall yield of the process can be
controlled by the reaction temperature and the molar ratio of the
aldol-condensation reactants. This aqueous phase process is
still in its infancy, and it is likely that this process can be applied
to other biomass-derived compounds containing carbonyl
compounds, such as replacing acetone with glyceraldehyde
or dihydroxyacetone. The development of this new metal-base
catalyst will be of particular interest for other base-catalyzed
reactions in aqueous phase solutions to develop environmentally benign processes.
Acknowledgements
This work was supported by the U.S. Department of Energy
Office of Basic Energy Sciences, and the National Science
Foundation Chemical and Transport Systems Division of the
Directorate for Engineering. We thank R. Gounder, J. Philipsek,
R. Schleich, and C. Zimmerman for technical assistance.

Fig. 6. Selectivity based on C5 (furfural:acetone) units in the aqueous phase

after aldol-condensation followed by hydrogenation over 5 wt% Pd/MgO-ZrO2
catalyst. Selectivity and overall carbon yield defined in Section 2.2. Various
furfural:acetone molar ratios were carried out over fresh catalyst at 353 K for
condensation followed by hydrogenation at 393 K. First three bars (black, dark
grey and white) show condensation for 24 h, while bar 4 (light grey) shows
condensation for 56 h.

References
[1] P. Davies, BP Statistical Review of World Energy 2005, can be found
under http://www.bp.com/downloads.do?categoryId=9003093&contentId=
7005944, 2005.
[2] J. Chow, R. Kopp, P. Portney, Science 302 (2003) 1528.
[3] R. Kerr, R. Service, Science 309 (2005) 101.

118

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118

[4] D. Moulton, N. Sefer, Diesel Fuel Quality and Effects of Fuel Additives,
Transportation Research Board, Washington, 1984, pp. 1015.
[5] G. Huber, J. Chheda, C. Barrett, J. Dumesic, Science 308 (2005) 1446.
[6] C. Moreau, R. Durand, D. Peyron, J. Duhamet, P. Rivalier, Ind. Crop. Prod.
7 (1998) 95.
[7] K. Lourvanij, G. Rorrer, Ind. Eng. Chem. Res. 32 (1993) 11.
[8] C. Gutsche, D. Redmore, R. Buriks, K. Nowotny, H. Grassner, C.
Armbruster, J. Am. Chem. Soc. 89 (1967) 1235.
[9] Y. Shigemasa, K. Yokoyama, H. Sashiwa, H. Saimoto, Tetrahedron Lett.
35 (1994) 1263.
[10] G.-W. Wang, Z. Zhang, Y.-W. Dong, Org. Process Res. Dev. 8 (2004) 18.
[11] Z. Zhang, Y.-W. Dong, G.-W. Wang, Chem. Lett. 32 (2003) 966.
[12] J.I. Di Cosimo, V.K. Diez, C.R. Apesteguia, Appl. Catal. A 137 (1996) 149.
[13] V. Serra-Holm, T. Salmi, P. Maki-Arvela, E. Paatero, L. Lindfors, Org.
Process Res. Dev. 5 (2001) 368.
[14] V. Serra-Holm, T. Salmi, J. Multamaki, J. Reinik, P. Maki-Arvela, R.
Sjoholm, L.P. Lindfors, Appl. Catal. A 198 (2000) 207.
[15] T. Salmi, V. Serra-Holm, T.K. Rantakyla, P. Maki-Arvela, L.P. Lindfors,
Green Chem. 1 (1999) 283.
[16] R. Zeng, X. Fu, C. Gong, Y. Sui, X. Ma, X. Yang, J. Mol. Catal. A: Chem.
229 (2005) 1.
[17] B.M. Choudary, M.L. Kantam, P. Sreekanth, T. Bandopadhyay, F. Figueras, A. Tuel, J. Mol. Catal. A: Chem. 142 (1999) 361.

[18] J. Yu, S.Y. Shiau, A. Ko, Catal. Lett. 77 (2001) 165.

[19] M. Campanati, S. Franceschini, O. Piccolo, A. Vaccari, A. Zicmanis,
Catal. Commun. 5 (2004) 145.
[20] J.C.A.A. Roelofs, D.J. Lensveld, A.J. van Dillen, K.P. de Jong, J. Catal.
203 (2001) 184.
[21] J.C.A.A. Roelofs, A.J. van Dillen, K.P. de Jong, Catal. Lett. 74 (2001)
91.
[22] M.J. Climent, A. Corma, S. Iborra, A. Velty, J. Catal. 221 (2004) 474.
[23] M. Aramenda, V. Borau, C. Jimenez, A. Marinas, J. Marinas, J. Ruiz, F.
Urbano, J. Mol. Catal. A: Chem. 218 (2004) 81.
[24] M. Aramenda, V. Borau, C. Jimenez, A. Marinas, J. Marinas, J. Navo, J.
Ruiz, F. Urbano, Colloids Surf. A 234 (2004) 17.
[25] A. Cordova, W. Notz, C. Barbas, Chem. Commun. (2002) 3024.
[26] B. Choudary, M. Kantam, P. Sreekanth, T. Bandopadhyay, F. Figueras, A.
Tuel, J. Mol. Catal. A: Chem. 142 (1999) 361.
[27] H. Liu, L. Peng, T. Zhang, Y. Li, New J. Chem. 27 (2003) 1159.
[28] A.A. Nikolopoulos, B.W.-L. Jang, J.J. Spivey, Appl. Catal. A 296 (2005)
128.
[29] K. Motokura, N. Fujita, K. Mori, T. Mizugaki, K. Ebitani, K. Kaneda,
Tetrahedron Lett. 46 (2005) 5507.
[30] B.E. Spiewak, J. Shen, J.A. Dumesic, J. Phys. Chem. 99 (1995) 17640.
[31] R.D. Cortright, R.R. Davda, J.A. Dumesic, Nature 418 (2002) 964.
[32] G.W. Huber, J.W. Shabaker, J.A. Dumesic, Science 300 (2003) 2075.