hydrogenation of biomass-derived compounds in water.pdf

8/10/2019 hydrogenation of biomass-derived compounds in water.pdf

1/8

Single-reactor process for sequential aldol-condensation andhydrogenation 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-ZrO2catalyst 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 furfural-

acetone 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 vehicleengine modification. Recently, we developed a process that

produces liquid alkanes by aqueous phase processing of

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 carbo-

hydrates to alkanes with molecular weights in the range ofdiesel 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 water-

soluble 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

www.elsevier.com/locate/apcatbApplied Catalysis B: Environmental 66 (2006) 111118

* Corresponding author. Tel.: +1 608 262 1095/2; fax: +1 608 262 5434.

E-mail addresses:[email protected](C.J. Barrett),[email protected]

(J.N. Chheda),[email protected](G.W. Huber), [email protected]

(J.A. Dumesic).1 Present address: Department of Chemical Engineering University of Mas-

sachusetts, 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
mailto:[email protected]:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.apcatb.2006.03.001http://dx.doi.org/10.1016/j.apcatb.2006.03.001mailto:[email protected]:[email protected]:[email protected]:[email protected]


2/8

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. 1presents 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 afore-

mentioned 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 neutra-

lized, 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 aldol-

condensation 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 heteroge-

neous 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 amino-

functionalized mesoporous base [26] as well as chiral L-

Proline [27] catalysts although in organic solvents. Aldol-

condensation and subsequent hydrogenation using a bifunc-

tional metal-base catalyst have been performed in an organic

environment for production of MIBK[28]and synthesis ofa-

alkylated 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 single-

reactor 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. Weshow 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)26H2O,

Aldrich} and zirconyl nitrate {ZrO(NO3)2, Aldrich}. Thecatalyst 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

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118112

Nomenclature

APD/H aqueous-phase dehydration/hydrogenation reac-

tion

BET BrunauerEmmettTeller

HMF 5-hydroxymethylfurfural

MCM mesoporous crystalline materialMIBK methyl isobutyl ketone

PES polyethersulphone

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


3/8

inductively coupled plasma (ICP) analysis [Perkin-ElmerPlasma 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.

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

wetness impregnation of Pd (using 5 wt% Pd in tetraamine-

palladium(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

All reactions (seeTable 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 stoppedafter 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.

C.J. Barrett et al. / Applied Catalysis B: Environmental 66 (2006) 111118 113

Fig.2. HMF disappearanceversustime during aldol-condensation with acetone

over fresh and calcined, recycled mixed Mg-Al-oxide and MgO-ZrO2catalysts.

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 roomtemperature and pressure,

while MgO-ZrO2runs carried out at 323 K and atmospheric pressure. HMF:a-

cetone molar ratio equal to 1:10; 11.2 wt% organics in the aqueous phase;

organic/catalyst mass ratio equal to 6.

Table 1

Experimental data for aldol-condensation and hydrogenation batch reactions

Run # Feed Molar

ratio

Org/Cata Time

[h]bTemperature

[K]cVolume

[ml]

Disappearance [%] Selectivity Overall carbon

yield [%]Furfural C5 units [%]

C5 C8 C13

1 Fur:Ace [first] 1:1 6 26 326 250 79 23 34 43 91

2 Fur:Ace [second] 1:1 6 26 326 200 58 43 31 26 93

3 Fur:Ace [third] 1:1 6 26 326 125 58 45 31 24 90

4 Fur:Ace [fourth] 1:1 9 26 326 90 76 25 32 43 91

5 Fur:Ace 1:1 6 24 353 100 95 5 35 60 88

6 Fur:Ace 1:1 6 26 393 100 98 3 35 62 80

7 Fur:Ace 1:9 6 24 353 100 96 4 67 29 76

8 Fur:Ace 2:1 6 24 353 100 66 37 15 48 91

9 Fur:Ace 2:1 6 56 353 100 86 16 12 72 85

10 Fur:Ace 1:1 18 25 393 100 90 11 30 59 85

11 Fur:Ace 1:1 36 26 393 100 88 14 32 54 82

12 Fur:Ace0.5% Pdd 1:1 6 25 393 112 98 2 33 65 82

13 Fur:Acehexadece 1:1 6 24 353 100 100 0 15 85 71

HMF C6 units [%]

C6 C9 C15

14 HMF:Ace 1:1 6 26 298 70 51 42 20 38 100

15 HMF:Ace 1:1 6 26 326 100 79 21 18 61 94

16 HMF:Ace 1:1 6 26 353 100 88 14 21 65 84

17 HMF:Ace 1:1 6 26 393 100 93 11 38 51 67

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, andpressureof 55 bar (except run12 using 0.5 wt% Pd andhydrogenatedfor 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.


4/8

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. Hydrogena-

tion 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. ForTable 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 0.2mm 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 hydro-

genated 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). Bothchemicals 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

0.2mm PES syringe membrane filter) and the HMF disap-

pearance was monitored using HPLC (Waters 2695 system with

a Zorbax SB-C18 5mm 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 C35moles C58moles C813moles C13

3moles C3fed 5 moles C5fed

100

C5selectivity%

moles C5

moles C5 moles C8 2 moles C13 100

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 approxi-

mately 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 N2adsorption at 77 K on

this same system.

2.3.2. X-Ray diffraction (XRD)

X-ray diffraction data were collected with a Cu Kasource

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 0.028s1. 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 catalystwas 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. 1outlines the bifunctional reaction pathways involved

in the production of large, water-soluble molecules from

xylose-derived furfural or glucose-derived HMF. In thisexample, 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



5/8

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 subse-

quently 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 aldol-

condensation 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 fromFig. 2, the mixed Mg-

Al-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/Al2O3catalyst on the performance of

the mixed Mg-Al-oxide catalyst for aldol-condensation. Thedeactivation 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 fromFig. 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/MgO-

ZrO2 catalyst for the formation of C15-dimer species also

remains constant with repeated use for sequential aldol-

condensation and hydrogenation reactions.

Fig. 3 shows the aqueous phase concentration of carbon

(normalized to the initial concentration of carbon in the batchreactor) versus time during aldol-condensation over a bifunc-

tional Pd/MgO-ZrO2catalyst 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-ZrO2catalyst 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

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, thishydrogenation 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 concentra-

tion, 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 sequen-

tial hydrogenation, in which the aqueous phase carbon lost

during the aldol-condensation step is returned to the aqueous

phase during the hydrogenation step.Table 1shows the details

of various runs conducted. TGA identified 48, 21 and 95% ofthe carbon missing from the carbon balance to be located on the

catalyst for runs 1, 7 and 15 inTable 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 aldol-

condensation of acetone with furfural (molar ratio 1:1) at

326 K, followed by hydrogenation at 393 K. The catalyst wasrecycled 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 2mmol/g), surface area (300 30 m2/g), and

average particle size (11 2 nm) of the catalyst for before


Fig. 3. Overall carbon yield in the aqueous phase versus time for aldol-

condensation 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.


6/8

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 measure-

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

obtained at reaction temperatures from 298 to 393 K for aldol-

condensations 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 aldol-

condensation (followed by hydrogenation) decreases at

temperatures above 353 K, probably caused by the formation

of coke on the catalyst during aldol-condensation. As thetemperature 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) 49.0 51.7 48.8

Base sites (mmol/g) 103

Surface area (m2/g) 292 329 299

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 9 10 9

MgO (2 2 0), 2u= 42.798 10 12 11

ZrO2 (1 1 1), 2u= 51.088 10 13 12

ZrO2 (2 2 0), 2u= 62.058 10 10 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. 4. Selectivity based on C5 (furfural:acetone) units in the aqueous phase

after aldol-condensation followed by hydrogenation over 5 wt% Pd/MgO-ZrO2catalyst. Selectivity and overall carbon yield defined in Section2.2. Furfur-

al: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

H2for 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 calcina-tions; (bar 4, white) third recycle with calcination.

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-ZrO2catalyst. Selectivity and overall carbon yield were in Section2.2. (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.


7/8

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. Accord-

ingly, the optimum temperature for aldol-condensation of HMF

with acetone is near 326 K. This study shows that the aldol-

condensation 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 higherfinal conversion (by 16%) but a lower dimer to monomer ratio

(1.8 versus 3.4) as compared to the HMF:acetone reaction.

Results inFig. 6(Table 1, runs 5, 79) show that the molar

ratioof reactants for aldol-condensationplays a significant rolein

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:acetoneratio 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.

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 to0.5 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). InTable 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 thefurther 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 inter-

mediates for further aqueous phase processing to produce liquid

alkanes. The selectivity and overall yield of the process can becontrolled 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 environmen-

tally 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.

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.


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

after aldol-condensation followed by hydrogenation over 5 wt% Pd/MgO-ZrO2catalyst. 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.
http://www.bp.com/downloads.do%3FcategoryId=9003093%26contentId=7005944http://www.bp.com/downloads.do%3FcategoryId=9003093%26contentId=7005944http://www.bp.com/downloads.do%3FcategoryId=9003093%26contentId=7005944http://www.bp.com/downloads.do%3FcategoryId=9003093%26contentId=7005944


8/8

[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. Fig-

ueras, 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.


Documents

hydrogenation of biomass-derived compounds in water.pdf