Glycerol as a source for fuels and chemicals by low-temperature catalytic processing

Article abstract of DOI:10.1002/anie.200600212

(Figure Presented) Lean and green: Glycerol can be catalytically converted into synthesis gas at relatively low temperatures. The combination of this process with Fischer-Tropsch and methanol-synthesis steps offers a new strategy for the production of liquid fuels and chemicals from renewable resources.

Full text of DOI:10.1002/anie.200600212

Communications
Biofuels
from biomass gasification, gas cleaning, and synthesis-gas
processing. The method we present herein may allow for
economic operation of a small-scale Fischer–Tropsch reactor
by producing an undiluted H2/CO gas mixture. Indeed, our
method reduces the capital cost of the Fischer–Tropsch plant
by eliminating the O2 plant or biomass gasifier and subse-
quent gas-cleaning steps.
DOI: 10.1002/anie.200600212
Glycerol as a Source for Fuels and Chemicals by
Low-Temperature Catalytic Processing**
The conversion of glycerol into CO and H2 takes place by
Equation (1). The endothermic enthalpy change of this
Ricardo R. Soares, Dante A. Simonetti, and
James A. Dumesic*
C3O3H8 ! 3 CO þ 4H2
ð1Þ
In this era of diminishing petroleum reserves, it is imperative
that industrialized society should develop ways to utilize more
effectively the abundant and renewable biomass resources
available to provide new sources of energy and chemical
intermediates.[1] To this end, various processes have been
developed to convert biomass and biomass-derived molecules
into specialty chemicals (methanol), light alkanes (C1–C6),
liquid fuels (ethanol and C7–C15 alkanes), and synthesis gas.[2]
As a new direction, we show herein that glycerol can be
converted over platinum-based catalysts into gas mixtures of
H2 and CO (synthesis gas)at temperatures from 498 to 620 K.
These temperatures are lower than those for conventional
gasification of biomass (e.g. 800–1000 K).[3,4] Synthesis gas can
be used to produce fuels and chemicals; therefore, the
endothermic conversion of glycerol into synthesis gas can be
combined with exothermic Fischer–Tropsch and methanol
syntheses to provide low-temperature and energy-efficient
routes for the production of these compounds. The glycerol
used can be sourced from waste glycerol streams that are
currently generated as by-products from the production of
biodiesel. This heat-integrated catalytic process could also be
a less energy-intensive alternative to current methods of
converting carbohydrates into fuel-grade ethanol.
reaction (350 kJmolꢀ1)corresponds to about 24% of the
heating value of the glycerol (ꢀ1480 kJmolꢀ1). The heat
generated by Fischer–Tropsch conversion of the CO and H2 to
liquid alkanes such as octane (ꢀ412 kJmolꢀ1)corresponds to
about 28% of the heating value of the glycerol. Thus,
combining these two reactions results in the following
exothermic process, with an enthalpy change (ꢀ63 kJmolꢀ1)
that is about 4% of the heating value of the glycerol:
7
C3O3H8 ! C8H18
25
19
25
37
CO2 þ H2O
25
ð2Þ
þ
One source of glycerol for our catalytic process is the low-
value waste stream from the transesterification of plant
oils[1,7] and animal fat[1] in the production of biodiesel. This
waste contains glycerol in water at high concentrations (e.g.
80%).[8] The current production levels of biodiesel in the
United States and Europe are 1 ” 108 and 2 ” 109 Lyearꢀ1,
respectively, and are expected to double in the near future.[8,9]
Selling waste glycerol solution can lower the production costs
of biodiesel by 6%.[10] The production of methanol from this
glycerol is an attractive alternative that could help alleviate
the burden of low-value waste glycerol and decrease the
production costs of biodiesel by lowering the dependence on
natural-gas-derived methanol.
Glycerol can also be produced by fermentation of sugars
such as glucose, either directly[11] or as a by-product of the
industrial conversion of lignocellulose into ethanol.[12] This
form of ethanol production in the U.S. is expected to increase
to 30 ” 109 Lyearꢀ1 by 2020 (three times the expected increase
in ethanol production from corn over the same period);[13] this
growing industry may be a major biofuel provider in the
future. Technical and economic analyses show that the co-
production of glycerol with ethanol lowers production costs
and increases profit margins.[11,14–16] The process pathways that
generate liquid fuels and chemicals from glycerol are outlined
in the Supporting Information.
Synthesis gas production methods for conventional
Fischer–Tropsch plants require either an O2 plant or a large
Fischer–Tropsch reactor to process the synthesis-gas stream
diluted with N2 from air, thereby increasing the capital costs
of such facilities.[5] Furthermore, Hamelinck et al. showed that
nearly 50% of the cost of producing Fischer–Tropsch liquids
from biomass is related to capital cost,[6] of which 50% stems
[*] D. A. Simonetti, Prof. J. A. Dumesic
Department of Chemical and Biological Engineering
University of Wisconsin
Madison, WI 53706 (USA)
Fax: (+1)608-262-5434
E-mail: dumesic@engr.wisc.edu
R. R. Soares
Universidade Federal de Uberlândia
Faculdade de Engenharia Química
Uberlândia, MG 38408-100 (Brazil)
Although the fermentation of glucose produces ethanol at
concentrations of about 5 wt% in water, it can be carried out
to produce glycerol at concentrations of at least 25 wt%.[11]
This higher concentration of glycerol decreases the energy
costs required to remove water from the oxygenated hydro-
carbon fuel. Indeed, distillation is one of the most energy-
intensive steps involved in the production of fuel-grade
ethanol from glucose.[17,18] For example, the energy required
to vaporize a 5-wt% aqueous solution of ethanol (as an
estimate of the energy of distillation)is 1680 kJmol ꢀ1, which
[**] This work was supported by the Office of Basic Energy Sciences,
U.S. Department of Energy (DOE), and we acknowledge a post-
doctoral grant from CAPES–Brazil for R.R.S. We thank Edward L.
Kunkes for help with catalyst preparation and characterization. We
thank Randy D. Cortright and Manos Mavrikakis for discussions
throughout this project.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3982
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3982 –3985

Angewandte
Chemie
corresponds to 1.4 times the lower heating value of the
ethanol product (ꢀ1241 kJmolꢀ1). In contrast, the energy
required to vaporize a 25-wt% aqueous solution of glycerol
(531 kJmolꢀ1)corresponds to 0.33 times the lower heating
value of the liquid-alkane products (ꢀ1418 kJmolꢀ1, assum-
ing octane as the product molecule and utilizing the enthalpy
change in Equation (2)). Detailed energy calculations are
presented in the Supporting Information.
Figure 1 shows results for the gas-phase conversion of
glycerol into synthesis gas over supported Pt catalysts at
623 K and atmospheric pressure with a feed solution of
glycerol in water (30 wt%). This concentration is similar to
that reported by Gong et al.[11] Moreover, the energy required
to separate the glycerol from water is similar to that of
vaporizing the solution and condensing the water after the
glycerol has reacted. Catalysts consisting of Pt supported on
Al2O3, ZrO2, CeO2/ZrO2, and MgO/ZrO2 exhibited deactiva-
tion during time-on-stream, whereas the Pt/C catalyst showed
stable conversion of glycerol into synthesis gas for at least
30 h (Figure 1 a). The catalyst with the most acidic support,
Pt/Al2O3, showed a period of apparently stable catalytic
activity, followed by a period of rapid catalyst deactivation.
Because the reactor initially operates at 100% conversion,
glycerol is present only in the upstream portion of the catalyst
bed in the tubular reactor, and a deactivation front moves
from the reactor inlet to the outlet as olefinic species are
formed from glycerol on acid sites associated with alumina,
followed by deposition of coke from these species on the Pt
surface sites. The most basic catalyst support, MgO/ZrO2,
showed rapid deactivation for all times-on-
Figure 1. Performance of supported Pt catalysts. Variation with time-
on-stream of a) percentage glycerol conversion to gas-phase products,
b) H2-turnover frequency, c) CO/CO2 molar ratio, and d) C2-TOF/H2-
~
&
TOF ratio for Pt catalysts supported on Al2O3 ( ), CeO2/ZrO2 ( ), C
!
^
*
(
), ZrO2 ( ), and MgO/ZrO2 ( ). Percentage conversion to gas
phase is calculated as [(Catoms in gas-phase product stream)/(total C
atoms into reactor as feed)]”100. Reactions were carried out at 1 bar
and 623 K with aqueous glycerol feed solution (30 wt%) over oxide-
supported Pt catalysts (1.0 g) or Pt/Ccatalyst (0.060 g) and a feed rate
of 0.32 cm3 minꢀ1. See Supporting Information for further experimental
details.
stream. The most stable oxide-supported
Table 1: Experimental data for catalytic processing of glycerol into synthesis gas under various
catalyst appears to be Pt on CeO2/ZrO2;
however, the performance of this catalyst is
inferior to that of Pt supported on carbon.
The different deactivation profiles dis-
played in Figure 1 for the various catalysts
suggest that the support plays an important
role in the deactivation process. Figure 1d
shows the rate of formation of C2-hydro-
carbons (ethane and ethylene)normalized
to the rate of H2 production for the various
supported Pt catalysts. Negligible amounts
of C2-hydrocarbons were formed on the Pt/
C catalyst. In contrast, catalysts consisting
of Pt supported on the various oxides
formed measurable amounts of C2-hydro-
carbons, and the C2-TOF/H2-TOF ratio
(TOF = turnover frequency)increased
with time-on-stream. This behavior suggests
that one of the modes of catalyst deactiva-
tion is caused by dehydration on the oxide
catalyst supports, which leads to the forma-
tion of unsaturated hydrocarbon species
conditions.
Conditions
Conversion into
gas phase [%]
H2-TOF
H2/CO
CO/CO2
CH4/H2
[minꢀ1
]
Feed flow rate
[cm3 minꢀ1 [a]
0.08
0.16
0.32
0.64
68
71
64
39
111
241
373
449
1.6
1.4
1.3
1.3
5.7
8.8
12
0.038
0.036
0.045
0.038
]
17
Glycerol
20
64
50
26
265
285
267
1.4
1.3
1.2
8.7
14
37
0.025
0.032
0.050
concentration
30[c]
50
[wt%][b]
T [K][d]
573
623
17
54
100
72
61
43
104
335
600
450
419
300
1.31
1.31
1.33
1.38
1.68
1.83
90
17
11
–
4.6
–
0.037
0.027
0.027
–
0.019
–
673
673[e]
723
723[e]
For these studies of reaction kinetics, 0.060 g of 5 wt% Pt/Cwas used. [a] Glycerol feed 30 wt%, 623 K,
1 bar. [b] Feed flow rate 0.32 cm3 minꢀ1, 623 K, 1 bar. [c] Point taken after 2 h time-on-stream. [d] Glycerol
feed 30 wt%, 0.32 cm3 minꢀ1, 1 bar. [e] Point taken after 3 h time-on-stream.
that form carbonaceous deposits on the Pt surface, thereby
decreasing the rate of H2 production and increasing the C2-
TOF/H2-TOF ratio.
The H2/CO ratio for the product stream from the Pt/C
catalyst is approximately 1.3:1 (Table 1), which is in agree-
ment with the stoichiometry of Equation (1). In contrast, the
H2/CO ratios obtained over the other catalysts were higher
than 1.5:1, which indicates some contribution from water–gas
shift (WGS). This behavior is demonstrated more clearly by
the CO/CO2 ratio (Figure 1c). The initial CO/CO2 ratio for
Pt/C is 12:1, whereas for the other catalysts it is less than 3:1.
Thus, it appears that the WGS reaction is facilitated by the
Angew. Chem. Int. Ed. 2006, 45, 3982 –3985ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3983

Communications
presence of the oxide support, as reported in other studies of
glycerol feed, and at different glycerol feed concentrations
(30 and 50 wt%)at a pressure of 1 bar. The catalyst showed
good stability for at least 48 h time-on-stream for both the
higher glycerol feed concentration (50 wt%)and the higher
reaction pressure (20 bar). It also showed excellent stability
for converting a 30-wt% glycerol feed into synthesis gas at
20 bar, with a H2/CO ratio ( ꢁ 2:1)that is appropriate for
Fischer–Tropsch[22] or methanol synthesis.[23]
To allow the efficient combination of the endothermic
conversion of glycerol into synthesis gas with the exothermic
Fischer–Tropsch or methanol-synthesis steps, the former
should be performed at the lowest-possible temperature. In
doing so, flexibility in the Fischer–Tropsch materials used (e.g,
Co-based versus Fe-based catalysts)becomes possible. There-
fore, we carried out studies of glycerol conversion into
synthesis gas at 498–573 K. The catalytic conversion of
WGS over supported metal catalysts.[19–21]
Reaction conditions (Table 1)that result in lower con-
versions of glycerol (i.e. higher flow rates of the glycerol feed
(30 wt%)and higher glycerol concentrations at constant feed
flow rate)lead to higher CO/CO
ratios. This behavior
2
suggests that the primary reaction in glycerol conversion is the
formation of CO and H2, and the production of CO2 by WGS
is a secondary reaction. Therefore, water acts primarily as a
diluent in the production of synthesis gas from glycerol, thus
leading to higher conversions (Table 1). The important
consequence of this behavior is that it is possible to adjust
the space velocity of the catalytic process to ensure that the
conversion of glycerol is complete, while simultaneously
controlling the H2/CO ratio for the Fischer–Tropsch synthesis
step. We note that the rate of methane formation remained
low for all of the conditions in Table 1.
The results in Table 1 show that the rate of H2 production
passes through a maximum with respect to reaction temper-
ature at constant-feed conditions. The rate increases with
increasing temperature from 573 to 623 K (with an activation
energy of about 70 kJmolꢀ1). However, although the rate of
hydrogen production increases further when the temperature
is initially increased to 673 K, the Pt/C catalyst eventually
undergoes deactivation at this higher temperature. We
suggest that at high temperatures, dehydration is fast relative
to H2 formation, thus leading to catalyst deactivation.
Figure 2a and b show the performance of the Pt/C catalyst
under different pressures (1 and 20 bar)for the 30-wt%
polyols to H2, CO2, and CO involves the preferential cleavage
[24,25]
ꢀ
ꢀ
of C C bonds as opposed to C O bonds,
and Pt-based
catalysts are particularly active and selective for this process.
Under these reaction conditions, the surface is covered
primarily by adsorbed CO species;[26] therefore, a strategy
for a catalyst that converts polyols into synthesis gas and is
active at low temperatures is to facilitate the desorption of
CO, thereby suppressing the subsequent WGS step and
improving the turnover of the catalytic cycle by regenerating
vacant surface sites. Accordingly, we require materials that
possess the catalytic properties of Pt with respect to selective
ꢀ
ꢀ
cleavage of C C versus C O bonds, but that have less
exothermic enthalpy changes for CO adsorption; Pt–Ru and
Pt–Re alloy catalysts fit this description.[27,28]
Figure 2c and d show that the percentage conversion of
glycerol and the product-gas ratios remained constant for at
least 72 h time-on-stream at these low temperatures over Pt–
Ru and Pt–Re (atomic ratio 1:1)bimetallic catalysts. The
overall carbon balances for these runs came to within 10%.
The main condensable organic compound in the effluent
streams of the catalysts was unconverted glycerol, with
smaller amounts of ethylene glycol, methanol, hydroxypro-
panone, and ethanol. These results demonstrate that the
conversion of glycerol to synthesis gas can, in fact, be
accomplished at temperatures well within the ranges
employed for Fischer–Tropsch and methanol syntheses,[22,23,29]
thus allowing for the efficient combination of these processes.
Moreover, the H2/CO ratio can be adjusted by adding a WGS
functionality (see Supporting Information).
We have presented a low-temperature catalytic route for
converting glycerol into H2/CO gas mixtures that are suitable
for combination with Fischer–Tropsch and methanol synthe-
ses. This new route opens new opportunities for use with
processes that produce polyols, such as glycerol through
fermentation of carbohydrate feedstocks, at high concentra-
tions in water.
Figure 2. Performance of carbon-supported Pt and Pt-bimetallic cata-
lysts under various process conditions. Variation with time-on-stream
of a) percentage glycerol conversion to gas-phase products and
&
b) molar ratios for conversion of 30 wt% glycerol at 1 bar ( ), 50 wt%
~
*
at 1 bar ( ), and 30 wt% at 20 bar ( ) over 0.090 g Pt/Cat 623 K.
Variation with time-on-stream of c) percentage glycerol conversion to
gas-phase products and d) CO/CO2 (filled symbols) and H2/CO
(empty symbols) molar ratios for conversion of 30 wt% glycerol at
1 bar over Pt–Ru/Cat 548 K (triangles; 0.435 g catalyst) and 573 K
(squares; 0.513 g catalyst) and over Pt–Re/Cat 498 K (inverted
triangles; 0.535 g catalyst) and 523 K (circles; 0.535 g catalyst). A feed
rate of 0.16 cm3 minꢀ1 for a) and b) and 0.08 cm3 minꢀ1 for c) and d)
was used.
Received: January 18, 2006
Published online: May 9, 2006
Keywords: biomass · gas-phase reactions ·
.
heterogeneous catalysis · liquid fuels · synthesis gas
3984 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3982 –3985

Angewandte
Chemie
[1] D. L. Klass, Biomass for Renewable Energy, Fuels, and Chem-
icals, Academic Press, San Diego, 1998.
[2] J. O. Metzger, Angew. Chem. 2006, 118, 710 – 713; Angew. Chem.
Int. Ed. 2006, 45, 696 – 698.
[3] M. Asadullah, S. Ito, K. Kunimori, M. Yamada, K. Tomishige,
Environ. Sci. Technol. 2002, 36, 4476 – 4481.
[4] R. L. Bain, D. C. Dayton, D. L. Carpenter, S. R. Czernik, C. J.
Feik, R. J. French, K. A. Magrini-Bair, S. D. Phillips, Ind. Eng.
Chem. Res. 2005, 44, 7945 – 7956.
[5] C. H. Bartholomew, R. J. Farrauto, Fundamentals of Industrial
Catalytic Processes, Wiley, Hoboken, 2006.
[6] C. N. Hamelinck, A. P. C. Faaij, H. den Uil, H. Boerrigter,
Energy 2004, 29, 1743 – 1771.
[7] J. M. Encinar, J. F. Gonzµlez, A. Rodríguez-Reinares, Ind. Eng.
Chem. Res. 2005, 44, 5491 – 5499.
[8] J. Van Gerpen, Fuel Process. Technol. 2005, 86, 1097 – 1107.
[9] K. Bozbas, Renew. Sust. Ener. Rev., in press.
[10] M. J. Haas, A. J. McAloon, W. C. Yee, T. A. Foglia, Bioresour.
Technol. 2006, 97, 671 – 678.
[11] C. S. Gong, J. X. Du, N. J. Cao, G. T. Tsao, Appl. Biochem.
Biotechnol. 2000, 84, 543 – 560.
[12] “Carbohydrate Economy Bulletin”, to be found under http://
www.carbohydrateeconomy.org/library/admin/uploadedfiles/
Carbohydrate_Economy_Bulletin_Volume_1_Numb_3.htm,
2000.
[13] P. L. Rogers, Y. J. Jeon, C. J. Svenson, Process Saf. Environ. Prot.
2005, 83, 499 – 503.
[14] C. E. Wyman, Biotechnol. Prog. 2003, 19, 254 – 262.
[15] J. J. Bozell, R. Landucci (Eds.), Alternative Feedstocks Program
Technical and Economic Assessment: Thermal/Chemical and
Bioprocessing Components, U.S. Department of Energy, Wash-
ington D.C., 1993, pp. 1 – 196.
[16] R. Landucci, B. J. Goodman, C. E. Wyman, Appl. Biochem.
Biotechnol. 1994, 45–46, 677 – 696.
[17] H. Shapouri, J. A. Duffield, M. Wang, (report no. 814, Office of
the Chief Economist, U. S. Department of Agriculture; available
at http://www.usda.gov/oce/reports/energy/aer-814.pdf) 2002.
[18] D. Pimentel, T. W. Patzek, Nat. Resour. Res. 2005, 14, 65 – 76.
[19] A. Basinska, L. Kepinski, F. Domka, Appl. Catal. A 1999, 183,
143 – 153.
[20] R. J. Gorte, S. Zhao, Catal. Today 2005, 104, 18 – 24.
[21] F. Esch, S. Fabris, L. Zhou, T. Montini, C. Africh, P. Fornasiero,
G. Comelli, R. Rosei, Science 2005, 309, 752 – 755.
[22] E. Iglesia, S. C. Reyes, R. J. Madon, S. L. Soled, Adv. Catal. 1993,
39, 221 – 302.
[23] K. A. Pokrovski, M. D. Rhodes, A. T. Bell, J. Catal. 2005, 235,
368 – 377.
[24] R. D. Cortright, R. R. Davda, J. A. Dumesic, Nature 2002, 418,
964 – 967.
[25] R. Alcala, M. Mavrikakis, J. A. Dumesic, J. Catal. 2003, 218,
178 – 190.
[26] R. He, R. R. Davda, J. A. Dumesic, J. Phys. Chem. B 2005, 109,
2810 – 2820.
[27] E. Christoffersen, P. Liu, A. Ruban, H. L. Skriver, J. K. Norskov,
J. Catal. 2001, 199, 123 – 131.
[28] J. Greely, M. Mavrikakis, Nat. Mater. 2004, 3, 810 – 815.
[29] A. P. Steynberg, R. L. Espinoza, B. Jager, A. C. Vosloo, Appl.
Catal. A 1999, 186, 41 – 54.
Angew. Chem. Int. Ed. 2006, 45, 3982 –3985ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3985