Biomass to chemicals: Catalytic conversion of glycerol/water mixtures into acrolein, reaction network

Article abstract of DOI:10.1016/j.jcat.2008.04.016

Acrolein was obtained by reacting gas-phase glycerol/water mixtures with zeolite catalysts. Glycerol was converted through a series of reactions involving dehydration, cracking, and hydrogen transfer and catalyzed by the acid sites of the zeolite. Acrolein was the major product; short olefins, aromatics, acetaldehyde, hydroxyacetone, acids, and acetone also were formed through a complex reaction network.

Full text of DOI:10.1016/j.jcat.2008.04.016

Journal of Catalysis 257 (2008) 163–171
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Biomass to chemicals: Catalytic conversion of glycerol/water mixtures
into acrolein, reaction network
Avelino Corma a,∗, George W. Huber a,1, Laurent Sauvanaud a, Paul O’Connor a,b
a
Instituto de Tecnología Química, UPV-CSIC, Universidad Politécnica de Valencia, Avda. de los Naranjos, s/n, 46022 Valencia, Spain
BIOeCON BV, Hogebrinkerweg 15e, 3871KM Hoevelaken, The Netherlands
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Acrolein was obtained by reacting gas-phase glycerol/water mixtures with zeolite catalysts. Glycerol
was converted through a series of reactions involving dehydration, cracking, and hydrogen transfer and
catalyzed by the acid sites of the zeolite. Acrolein was the major product; short olefins, aromatics,
acetaldehyde, hydroxyacetone, acids, and acetone also were formed through a complex reaction network.
© 2008 Elsevier Inc. All rights reserved.
Received 3 January 2008
Revised 10 April 2008
Accepted 19 April 2008
Available online 23 May 2008
Keywords:
Catalytic cracking
Glycerol
Olefins
Acrolein
Biomass
Biochemicals
Zeolites
1. Introduction
feed [2,11]. Vegetable oils also can be converted into diesel fuel
range alkanes by hydrotreatment [12–14]. It is estimated that ap-
proximately 80% of the biofuels in Europe are biodiesel [11]. As
biodiesel production increases, the price of glycerol is expected
to drop significantly from the actual cost, which has already de-
creased by half in the last few years [15]. Thus, it has been
estimated that the cost of unrefined glycerol could decrease to
0.44 $/kg [16]. Importantly, the economics of biodiesel produc-
tion are such that a credit of 300 $/ton is given for the reselling
of crude glycerol, based on 2010 NREL estimates [17]. Such an
inexpensive feed makes the development of processes for the con-
version of glycerol into other chemicals desirable.
Today, there is renewed interest in the conversion of biomass
and biomass-derived products into fuels and chemicals [1–6]. Oxy-
genated petrochemicals are currently produced through the con-
trolled oxidation of hydrocarbons. The conversion of petroleum-
derived feedstocks into functionalized oxygenated chemicals in-
volves the high-temperature activation of thermally stable mole-
cules that have low functionality. In contrast, biomass-derived
feedstocks with high oxygen content are thermally unstable but
can be converted into oxygenated chemicals via deoxygenation
processes [7,8]. These oxygenated chemicals can be high-value
chemical feedstocks that possibly could replace petroleum-derived
products. The deoxygenation process typically involves dehydration
reactions to remove the oxygen as water. One example of this is
the selective dehydration of fructose and xylose in biphasic mix-
tures to produce 5-hydroxyfurfural and furfural, respectively, or
2,5-dimethylfuran from fructose [9,10].
One of the chemicals into which glycerol can be converted is
acrolein. The dehydration of glycerol into acrolein has been known
since the nineteenth century. Acrolein is used to produce acrylic
acid, acrylic acid esters, superabsorber polymers, and detergents.
As early as 1918, Sabatier and Gaudion [18] reported the synthe-
sis of acrolein (10% yield) from glycerol by gas-phase dehydration
◦
Glycerol is currently a valuable byproduct of biodiesel produc-
tion with a refined value close to 1 $/kg. Biodiesel is currently
produced by transesterification of vegetable oils (triglycerides) and
methanol, with 1 mol of glycerol produced per mol of triglyceride
on Al2O3 or UO2 catalysts at 360 and 350 C, respectively, with
ethanol, water, and allyl alcohol as other products. In 1928, Fre-
und reported that pure acrolein could be produced from glycerol
◦
at 180 C using diatomaceous silica [19]. Twenty years later, Hoyt
and Manninen reported a method to produce acrolein from glyc-
erol using solid phosphoric acid catalysts [20]. Heinemann et al.
studied the dehydration of organic compounds (including glycerol)
with activated bauxite catalysts and observed acrolein yields of up
Corresponding author.
E-mail address: acorma@itq.upv.es (A. Corma).
Present address: Chemical Engineering Department, University of Massachu-
*
1
◦
setts, 159 Goessmann Lab, 686 N. Pleasant St., Amherst, MA 01003-9303, USA.
to 42% from glycerol working in gas phase at 430 C [21]. In 1993,
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.04.016

164
A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
Table 1
2.2. Laboratory units
Physico-chemical characteristics of catalysts
2.2.1. Microactivity test (MAT)
Catalyst
ECat
ZSM-5 additive
Some of the experiments described herein were performed in
a fixed microactivity test (MAT) reactor. The reaction zone and
product recovery system were designed in accordance with ASTM
D-3907 [27]. Before each experiment, the MAT system was purged
with a 50 cc/min N2 flow for 30 min at the reaction tempera-
ture. All MAT reactions reported herein were conducted at 30 s
time on-stream. After the reaction, stripping of the catalyst was
done using a N2 flow of 40 cm3/min for 15 min. During the re-
action and stripping steps, the liquid products were collected in
the corresponding glass receivers located at the exit of the reactor,
kept at 278 K by means of a computer-controlled bath. Meanwhile,
the gaseous products were collected in a gas burette by water dis-
placement. After stripping, the catalyst was regenerated at 813 K
in a 100-cm3/min stream of air for 3 h.
Si/Al
20
156
0.050
11
1.0
50
70
0.027
15
–
BET surface area (m2/g)
Micropore volume (cm3/g)
Zeolite content
Rare earth content, wt%
Particle size (μm)
40–120
40–120
Degussa reported a process to produce acrolein by dehydration
of glycerol in the liquid or gaseous phase with solid acidic cata-
lysts [22]. They claimed that aqueous solutions from 10 to 40 wt%
glycerol could be converted to acrolein at 180–340 C in the liquid
phase or 250–340 C in the gaseous phase. Gas-phase dehydration
◦
◦
◦
of glycerol at 300 C with phosphoric acid supported on alumina
gave acrolein yields of 70% and 65% as the concentration of glyc-
erol increased from 10 to 40 wt%; however, the yield to acrolein
decreased with increasing glycerol content. Consequently, a pro-
cess to produce acrolein using aqueous solutions from biodiesel
production (∼85 wt% glycerol in water) would be desirable.
Dehydration of glycerol has been conducted in liquid phase
with H-ZSM-5, Na-ZSM-5, mordenite, Li3PO4 and FePO4 as cata-
lysts. But acrolein yields were lower in liquid phase than in gas
phase; the highest liquid-phase yield was 36%. Acrolein also can
be prepared from glycerol using subcritical and supercritical wa-
ter [23–25] or with homogeneous catalysts [26]. The maximum
acrolein yield with subcritical and supercritical water was reported
2.2.2. Microdowner reactor
Hardware and detailed operation of the Microdowner unit has
been described previously [28]. The unit’s main features are a cata-
lyst preheater in which the catalyst is stored before testing, a once-
through reactor in which the feed and the preheated catalyst are
fed continuously during the test at a very short residence time
(0.3–2 s), and a separator that stores the catalyst used during the
test for regeneration and coke determination. The unit simulates
a steady-state regime during the test, which usually takes 1–2 min.
The catalyst separated from the reaction products is continuously
stripped during the reaction and for 60 s after the end of the reac-
tion. Liquids and gaseous products are recovered by known meth-
ods (cold traps and water displacement burette). As in many tests,
both an aqueous phase and a hydrocarbon phase appeared in the
traps. Acetone was used to dilute both phases, to allow extraction
of a single sample from each trap. The catalyst was regenerated
after the test with 500 cm3/min of air at 850 K for 3 h. Alterna-
tively, the coked catalyst could be withdrawn from the unit after
the stripping step, thus ignoring the regeneration step. A flow of
nitrogen was used for solid transportation and feed dispersion.
to be 37.5% at 360 C, 25 MPa, and 470 ppm (g g−1) of zinc sulfate
◦
as the catalyst [24]. Most earlier processes involved solid acid cata-
lysts in fixed-bed reactors, because catalyst deactivation with pure,
diluted glycerol feeds was considered sufficiently low to support
fixed-bed, continuous operation without regeneration for a sig-
nificant period [22]. Nevertheless, the presence of contaminants—
especially basic contaminants from the production of biodiesel and
the possible secondary reactions of glycerol cracking products—
may produce coke that causes deactivation of the catalyst, which
then must be regenerated.
2.2.3. Analysis of the reaction products
In this paper, we report that glycerol in water can be converted
into acrolein, olefins, and acetaldehyde by reaction with zeolites
in a continuous fluidized-bed reactor. This reaction system allows
better heat and mass transfer than fixed-bed reactors, along with
the possibility of performing continuous regeneration if needed.
We show that glycerol reacts through a complex reaction network
with numerous consecutive and parallel reactions that involve such
processes as dehydration, cracking, and hydrogen transfer, all of
which are catalyzed by acid sites.
The reaction gases were analyzed with a Varian 3800-GC gas
chromatograph equipped with three detectors, two thermal con-
ductivity detectors (TCDs) for analyzing H2 and CO/CO2/N2, sepa-
rated on a 2-m molecular sieve 5A column and a 2.5-m molec-
ular sieve 13X column, respectively; and a flame ionization de-
tector (FID) for C1 to C6 hydrocarbons separated on
a 50-m
Plot/Al2O3 column. Liquid samples were analyzed with a Varian
3900-GC gas chromatograph equipped with a Petrocol-100 fused
silica column connected to a FID following the PIONA proce-
dure. Gas chromatography–mass spectroscopy (Agilent Technolo-
gies 5973 and 6890N, with a 20-m HP1 column) was used to
identify the main oxygenates appearing in analyses of the gas and
liquid samples. The CO2 formed during the regeneration step was
monitored and quantified by means of an IR cell. The carbon yields
reported herein are defined as the mol of carbon in each prod-
uct divided by the carbon in the feed. The conversion presented is
the conversion of the feedstock. The conversion toward gas (carbon
oxides, hydrogen, and C1 to C4) plus hydrocarbons in liquids plus
coke (i.e., excluding the oxygenated products) is also presented in
addition to the traditional conversion.
2. Experimental
2.1. Materials and feeds
Table 1 summarizes the physical and chemical characteristics
of catalysts used in this study, including a commercial FCC cata-
lysts containing Y-zeolite in a silica–alumina matrix (Precision 1%
Rare Earth, supplied by BASF, tag FCC1) and a FCC additive based
on ZSM5 zeolite, mixed with a clay binder to ∼15% weight. These
equilibrium catalysts were subjected to hydrothermal deactivation
◦
(816 C, 4 h, 100% steam). The same catalyst sample was used to
3. Results and discussion
perform all of the tests with different feeds and operating condi-
tions with no noticeable changes in the catalyst parameters or in
the original activity or stability. Aqueous solutions of 20, 50, and
85 wt% of glycerol were prepared from 99.5 wt% glycerol (Aldrich
Chemicals).
◦
3.1. Conversion of glycerol at temperatures from 500 to 700 C
Table 2 gives the results of converting a 50% aqueous glycerol
solution at 500–650 C with a commercial USY-based equilibrium
◦

A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
165
Table 2
High temperature glycerol conversion in Microdowner (MD) and Microactivity test
(MAT) reactors (50 wt% glycerol aqueous solution)
Laboratory unit
MD
MD
MD
MD
MD
MAT
MAT
Test no.
Catalyst
1
ECat
2
ECat
3
4
5
6
7
ECat
ZSM5
ZSM5
ZSM5
ZSM5
Operating conditions
◦
Temperature ( C)
500
12
650
51
500
11
500
48
650
45
700
4
720
6
Catalyst to feed ratio
Residence time (s)
0.7
396
0.7
108
0.7
431
1.4
55
0.7
121
30
30
83
7
−1
WHSV (h
)
Glycerol conversion, wt%
100
100
100
100
100
100
100
Molar carbon selectivity (%)
Carbon monoxide
Carbon dioxide
C1–C4 alkanes
Ethylene
Propylene
Butenes
Acetaldehyde
Acrolein
Acetone
6.8
5.0
0.8
1.2
4.2
1.3
23.6
18.3
6.1
23.6
9.2
6.6
6.6
7.5
2.5
13.3
6.0
6.2
4.2
0.2
2.4
2.4
1.6
24.9
39.0
3.6
<0.1
<0.1
0.9
2.8
5.4
6.2
9.7
6.3
0.5
4.5
4.3
1.9
15.2
23.2
3.2
<0.1
<0.1
1.3
6.7
5.5
17.7
18.0
8.8
3.3
10.5
6.1
3.5
20.7
11.0
3.6
<0.1
<0.1
0.2
42.0
5.1
12.9
21.8
7.8
51.0
8.8
12.5
13.1
4.9
0.4
0.5
0.1
0.2
0.0
0.0
0.0
1.0
0.9
1.0
0.4
0.2
0.0
0.0
0.0
1.7
3.5
2-Propenol
Acetol
Acids
0.1
<0.1
<0.1
0.4
2.8
2.8
0.2
2.4
1.7
5.2
23.0
C5 + BTX
6.2
2.4
5.7
Others CHO
Coke
0.2
3.2
0.1
7.3
15.3
catalyst (ECat) and ZSM5-based additive in a Microdowner reactor.
The Microdowner unit was operated with catalyst residence times
close to 1 s and catalyst-to-feed (CTF) ratios of 12–51 to simulate
Fig. 1. Injection of glycerol into FCC type reactor.
◦
adiabatic mixing of the glycerol solution (100 C) and the catalyst
◦
◦
(close to 700 C from an industrial regenerator). A test at 500 C
with a CTF ratio of 48 using a ZSM5-based catalyst was also con-
ducted to study the effect of CTF ratio on product selectivity. These
reaction conditions were chosen to be representative of the opera-
tion conditions encountered at the several possible injection points
of glycerol in a FCC unit, as depicted in Fig. 1.
Table 3
◦
◦
Comparison of glycerol conversion at 700 C and naphtha steam cracking at 800
for the production of olefins, carbon yield basis
C
Glycerol cracking,
carbon wt% yield
Naphtha steam cracking,
carbon wt% yield
Carbon monoxide
Carbon dioxide
Methane
42
5.1
–
–
In all tests, the glycerol conversion was 100%. The main prod-
ucts of the reaction with both catalysts were CO, CO2, ethylene,
propylene, butenes, oxygenates, acrolein, acetaldehyde, acids (in-
10.5
1.8
21.8
0.6
7.8
15.7
3.3
30.8
0.6
14.0
3.8
6.5
–
Ethane
Ethylene
+
cluding acetic acid, propionic acid, and acrylic acid), C5 hydro-
Propane and butanes
Propylene
carbons, and coke. Acetone/propanal and 2-hydroxyacetone (ace-
◦
Butenes
C2–C4 alkynes
Acrolein
0.9
tol) were present at very low yields at 500 C and above. The
+
C5 hydrocarbons present included pentenes, benzene, toluene, and
<2
xylenes. A series of products grouped under the label “other CHO”
included hydrocarbons (hexanes and higher-boiling point compo-
nents) and oxygenated products, including methacrolein, methyl
vinyl ketone, butanone, cyclopentanone, and methylcyclopenenone.
No alcohols were found as products in these tests. Propane and bu-
tane yields were very low in all cases, whereas methane, ethane,
Other oxygenates
C5 hydrocarbons
Fuel oil
Coke
–
+
4.2
–
3.2
19.3
5.2
–
◦
◦
MAT testing at 700 C with long residence times (20–80 s). At
such a high temperature, the conversion of oxygenated products
was almost complete with both the ZSM5 and Y catalysts, whereas
ethylene yield increased to nearly 22% and propylene yield re-
mained around 8% (based on feed carbon content) with the ZSM5
catalyst. The CO yield was very high and coke yield was low with
both catalysts. We note that coke-on-catalyst was 10-fold higher
in the MAT reactor (fixed bed) than in the MD reactor (trans-
ported bed), due to the different residence times in the two re-
actors. The ethylene yields were similar to those obtained with
and butadiene yields were significant at 650 C due to thermal
cracking. Ethyne and propyne were not found as products, in con-
trast to the steam cracking operation, in which they are obtained
in significant amounts.
With increasing reaction temperature, the yields of CO, CO2,
+
light olefins, methane, and C5 hydrocarbons increased, whereas
the yields of the other major products, coke and oxygenates, de-
creased. Yields of acrolein and organic acids decreased faster than
those of acetaldehyde and acetone with increasing temperature.
This general pattern was seen regardless of the two catalysts used.
These findings indicate that the high-temperature operation is not
the ideal operating condition for oxygenate production as it is for
light-olefin production [29]. Conversion toward light olefins can
be further increased by increasing the temperature and decreasing
the space velocity. We simulated these conditions by performing
◦
steam cracking of naphtha at 800 C, as shown in Table 3, but with
lower yields of methane, dienes, and liquids in the former. High
levels of CO also were produced with glycerol conversion, which
could be used to produce additional hydrogen via the WGS reac-
tion.

166
A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
Table 4
temperature and the greater amount of glycerol and/or intermedi-
ates trapped on the catalyst. Surprisingly, the acrolein yield was
hardly affected by the temperature decrease, possibly indicating
a different formation pathway for acrolein and acetol, acids, and
Optimal conditions of temperature, glycerol dilution and space time for best acrolein
yield. Catalyst based on ZSM5 zeolite, operation in Microdowner unit
Laboratory unit
Test no.
MD
8
MD
9
MD
10
MD
11
MD
12
MD
13
◦
acetaldehyde. At a temperature of 290–350 C and space velocity of
300–400 h−1, the conversion of glycerol was still nearly complete,
Operating conditions
Temperature ( C)
Catalyst to feed ratio
Residence time (s)
WHSV (h
Feed dilution (wt% glycerol)
◦
◦
with a minimum conversion of 97% at 290 C (Table 4, test 8).
290
12.6
1.0
282
50
350
5.6
0.5
1243
50
350
10.7
1.0
360
50
350
11.5
0.9
335
20
350
5.4
0.5
1315
85
350
9.9
0.9
388
85
−1
When the space velocity was increased to 1200 h
(by de-
creasing the residence time and CTF ratio), glycerol conversion
decreased to 89%, as shown in Table 4. This had no signifi-
cant effect on the selectivity to acrolein or acetaldehyde; how-
ever, coke yield decreased significantly due to the lower CTF
ratio, whereas the yields of acetol and the unidentified carbo-
hydrates/hydrocarbon increased. This demonstrates that acetol is
a primary but unstable product, and that acetone, 2-propenol, and
acids are secondary products. In addition, many of the noniden-
tified products had a boiling point in the range of that of glyc-
erol, suggesting that other minor intermediates besides acetol and
3-hydroxypropionaldehyde were formed from glycerol and were
rapidly converted as severity increased. We discuss the possible
formation of these intermediates in Section 3.4.
The results of tests 10, 11, and 13 in Table 4 demonstrate the
effect of the dilution of glycerol in water on the product dis-
tribution. This was little affected by the change in the glycerol
dilution, whereas the overall conversion was slightly lower with
the 85% glycerol solution (97% vs 100% at lower glycerol concen-
trations). The acrolein selectivity remained between 58 and 62%,
slightly higher with the highest dilution. The hydrocarbon yield
did not change with the dilution. The coke yield was lowest with
the 50% glycerol solution. This can be explained by the higher
product concentration of the 85% glycerol solution, which pro-
motes the hydrogen transfer reaction that produces coke; and the
higher catalyst-to-carbon ratio for the most diluted solution, which
proportionally increases the coke yield through the strong adsorp-
tion of feed or products by the strongest acid sites. Among the oxy-
genates, the larger molecules (i.e., acetol, other CHO) were favored
by higher glycerol concentrations, whereas the acetaldehyde yield
was the greatest at lower glycerol concentrations. Lower glycerol
concentrations may favor the monomolecular cracking of larger
molecules, whereas higher glycerol concentrations may favor bi-
molecular condensation reactions. A test carried out at high space
velocity and 85% glycerol concentration (test 12) showed very sim-
ilar results as were found with 50% glycerol solutions: decreased
glycerol conversion (86%) and coke yield and increased CHO yield.
These results demonstrate that it is possible to produce acrolein
as a bulk chemical from glycerol using commercial glycerol solu-
tion from biodiesel plants with a carbon selectivity >60%.
−1
)
Glycerol conversion, wt%
98
89
100
100
86
97
Molar carbon selectivity (%)
Carbon monoxide
Carbon dioxide
C1–C4 saturates
Ethylene
Propylene
Butenes
Acetaldehyde
Acrolein
Acetone
2.8
0.6
<0.1
0.3
0.8
0.2
4.6
58.8
2.4
1.2
0.9
2.0
0.6
4.2
20.6
1.8
0.1
2.6
0.8
<0.1
0.6
0.9
0.3
8.5
58.8
3.3
0.7
2.7
3.9
0.6
4.1
2.3
1.1
1.6
2.2
0.7
<0.1
0.5
1.0
0.3
5.9
58.6
2.1
0.9
1.8
2.0
0.9
6.8
16.2
0.5
<0.1
0.4
0.7
0.2
6.2
59.1
1.4
<0.1
0.4
0.6
0.2
8.9
58.9
2.0
0.2
4.0
2.4
0.5
12.3
7.8
<0.1
0.5
0.5
0.2
10.9
62.1
1.5
0.1
0.5
1.6
2-Propenol
Acetol
Acids
1.2
2.6
2.2
0.7
13.2
10.1
C5 + BTX
1.2
2.4
15.0
Others CHO
Coke
12.2
Comparing the behavior of the two catalysts under the same
operating conditions shows lower amounts of CO, CO2, and coke
and greater amounts of oxygenates and olefins with the ZSM5-
based catalyst. C5 hydrocarbon yield was higher with the ZSM5
additive.
When the CTF ratio was increased at constant temperature, the
yields of CO, CO2, coke, olefins, and C5 hydrocarbon increased,
whereas the yield of oxygenates decreased. Acrolein and acetalde-
hyde yields were decreased, probably because they were converted
mainly to olefins, CO, and coke with increasing CTF ratio. Clearly,
the yield to acrolein and other oxygenates was maximized at
higher space velocities and lower temperatures.
+
+
3.2. Low-temperature conversion of glycerol
As demonstrated by the previous results, high space velocities
and low temperatures are the best conditions for acrolein pro-
duction. Consequently, we performed several tests at 290–350 C
◦
−1
and space velocities close to 400 h
and 1200 h−1. To attain
these space velocities, we used catalyst residence times of 0.5–
1 s and CTF ratios of 5–12. The results, given in Table 4, show
that the acrolein yield increased with decreasing temperature.
Acrolein selectivities of 58–62% were obtained. The acetaldehyde
and acetone yields were decreased at 290–350 C compared with
operation at 500 C, whereas the acid yield increased slightly. 2-
Propenol and acetol also were observed at these lower tempera-
tures. Interestingly, no saturated alcohols (1- and 2-propanol) were
detected. Methacrolein and methyl vinyl ketone were present in
3.3. Comparison of MD and MAT units for glycerol processing at low
temperature
◦
◦
Fig. 2 compares the processing of a glycerol solution (50% glyc-
◦
erol) at the same temperature (350 C). The MAT operating condi-
tions were adapted from the standard test conditions so that the
space velocity in these tests did not differ significantly from the
space velocity in the Microdowner tests (Table 5). In particular,
TOS and the CTF ratio were reduced so that space velocity in MAT
+
minor amounts. The production of CO, CO2, olefins, and C5 hydro-
carbons also decreased, whereas coke yield increased. This latter
finding may be due to product desorption problems in the same
way as Vacuum gasoil produces high coke yield when poorly va-
porized in the riser reactor or when the processing temperature
is too low [30]. This effect was particularly significant when the
reaction temperature approached the temperature of glycerol va-
porization (290 C), and a noticeable increase in the coke yield was
seen when the temperature was decreased from 350 to 290 C. The
yields of acetol, acid, acetaldehyde also decreased, possibly due to
the lower extent of secondary reactions resulting from the lower
−1
ranged from 360 to 90 h−1, compared with 20 to 60 h
in the
standard test. As a result, Fig. 2 shows that the conversion obtained
−1
at a space velocity of 90 h
was very close to the MD conversion
at 360 h−1. Moreover, the product distribution shown in Fig. 2 also
was very similar, whereas some significant differences were found
in the acetaldehyde, acetol, CHO, and coke yields for the MAT test
at 90 h−1. This difference is probably due to the amount of coke on
the catalyst (CoC), which is higher with MAT than with MD test-
ing (typically 1% vs 0.2%). It has been demonstrated [31–33] that
◦
◦

A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
167
Table 6
Product selectivity in the conversion of several main products obtained from the
glycerol processing. Tests in MAT and MD units with ZSM5-based catalyst, 50 wt%
aqueous solutions
Laboratory unit
Reactant
MAT
MAT
MAT
MAT
MD
Acetol
Acetol
Acetone
Acetone
2-Propenol
Operating conditions
◦
Temperature ( C)
350
2
20
90
50
500
2
20
90
50
350
2
20
90
50
500
2
20
90
50
500
11.9
0.9
391
50
Catalyst to feed ratio
residence time or TOS (s)
−1
WHSV (h
)
Feed dilution (wt% glycerol)
Feed conversion, wt%
23
86
14
35
99
Molar carbon selectivity (%)
Carbon monoxide
Carbon dioxide
C1–C4 alkanes
Ethylene
Propylene
Butenes
Acetaldehyde
Acrolein
Acetone
1.2
0.7
<0.1
0.1
0.3
0.2
1.4
0.1
4.0
0.0
–
11.6
3.1
0.5
6.1
4.8
2.7
8.0
0.2
17.1
0.0
–
0.2
0.7
0.3
0.4
1.7
27.9
0.6
0.0
–
0.0
2.3
20.2
6.6
7.3
31.9
1.2
7.3
0.7
1.5
7.6
47.2
1.0
0.0
–
0.0
0.0
11.1
4.8
5.0
12.4
1.4
1.0
0.6
6.7
22.2
4.5
0.7
5.0
11.6
–
0.0
1.3
8.4
16.1
20.5
2-Propenol
Acetol
Acids
9.4
2.9
52.0
27.7
10.3
4.3
24.8
6.5
C5 + BTX
Others CHO
Coke
Table 7
Product selectivity in the conversion of minor products observed in the glycerol
◦
conversion at 350 C, with ZSM5-based catalyst, and compared with acetol conver-
sion under similar conditions
Fig. 2. Comparison between MD and MAT units for glycerol processing, low temper-
Laboratory unit
Product
MAT
MAT
MAT
ature.
Acetol
1,2-Propanediol
Propionic acid
Table 5
Operating conditions
Temperature ( C)
◦
Standard MAT operating conditions and modifications for glycerol conversion
350
2
20
90
50
350
1
20
180
50
350
1.3
20
142
50
Catalyst to feed ratio
Operating conditions
Time on stream (s)
Standard, VGO
Glycerol cracking
Residence time (s)
30
2–6
20–60
500–550
20
−1
WHSV (h
)
Catalyst to feed ratio
0.5–2
90–360
350
Feed dilution (wt% glycerol)
−1
WHSV (h
)
◦
Temperature ( C)
Feed conversion, wt%
23
65
39
Molar carbon selectivity (%)
Carbon monoxide
Carbon dioxide
C1–C4 alkanes
Ethylene
Propylene
Butenes
Acetaldehyde
Acrolein
Acetone
1.2
0.3
0.3
0.2
0.9
5.1
2.4
1.3
0.0
60.8
0.4
0.3
8.4
1.5
9.4
8.7
0.3
2.0
0.0
0.6
0.6
0.4
0.9
0.0
2.3
0.0
0.0
0.7
1.8
6.2
84.2
catalyst activity is decreased by a factor of 2–4 for each 1% CoC
for typical hydrocarbon cracking reactions. This would explain the
difference between the two units in the space velocity required
to reach the same conversion. Notwithstanding the differences in
the two laboratory units, the MAT unit can give a sufficiently good
product selectivity pattern to perform catalyst screening or to ex-
plore the main conversion routes of the glycerol reagent and its
reaction products.
0.7
<0.1
0.1
0.3
0.2
1.4
0.1
4.0
0.0
–
2-Propenol
Acetol
Acids
9.4
2.9
52.0
27.7
C5 + BTX
3.4. Chemistry of glycerol conversion
Others CHO
Coke
3.4.1. Oxygenate distribution
3.4.2. Product conversion selectivities and reaction network
The main oxygenates produced by glycerol conversion, ex-
cluding acrolein, were acetaldehyde, acetone, 2-propenol, and
2-hydroxyacetone (acetol). Methacrolein and methyl vinyl ketone
were detected in trace amounts. No traces of propanol were de-
tected, and only trace amounts of methanol and ethanol, as well
as a number of unidentified compounds with 1 to 3 oxygen atoms
and 2 to 4 carbon atoms, were found. These oxygenated com-
pounds, plus acetol, were present at low levels (combined <5 wt%)
To track the primary, secondary, and stable or unstable charac-
ter of the different products detected in the conversion of glycerol,
we fed several of the main products in the MAT and MD units
under representative conditions of the previous MD tests (ZSM5-
◦
based catalyst, 350–500 C, space velocity of 100–400 h−1). The
aim was to help establish a reaction network. The operating con-
ditions and selectivity results are summarized in Tables 6 and 7.
Dehydration of glycerol has been shown to first produce ace-
tol and acrolein, as shown in Fig. 3 [34–36]. This mechanism has
◦
at 500 C and above, whereas at lower reaction temperatures, they
can represent >10% of the carbon balance.

168
A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
Fig. 4. Acetol conversion scheme.
Fig. 3. Glycerol primary products.
3.4.2.2. Acetone and 2-propenol conversion Acetone was fed at
been observed on a wide variety of catalysts, including base- and
acid-supported catalysts and zeolites [37–39]. Similarly, we found
the formation of significant amounts of acrolein during the MD ex-
periments with glycerol, as well as significant quantities of acetol
◦
◦
350 C and 500 C and a space velocity of 90 h−1, whereas 2-
◦
−1
propenol was fed at 500 C and a space velocity close to 90 h
.
(Propenol was not observed as a product in the conversion of glyc-
◦
◦
erol at 350 C.) The results, presented in Table 6, indicate that
at 350 C and below. But we also found large amounts of subprod-
2-propenol was nearly as reactive as glycerol, whereas acetone was
less reactive, with only 30% conversion at 500 C under the same
ucts, including acids, acetone, hydrocarbons, and coke, whose for-
mation pathways from glycerol and the primary products formed
from glycerol have not yet been described. In addition, it is known
that glycerol can form oligomers (polyglycerols), which may lead
to coking reactions [40]. (Water may decrease the amounts of
oligomers.) The same mechanism is applicable to acetol as well.
Coke is also formed from acetaldehyde [41], and probably acrolein;
coke can be formed from many sources.
◦
operating conditions. These two reactants showed high selectivity
toward olefins, but 2-propanol was more selective to propene and
ethylene, whereas acetone was more selective to butenes, as re-
ported previously for acetone [45]. The conversion of acetone also
produced a significant yield of CO2. Acetic acid also was a major
◦
◦
product at 350 C but not at 500 C. Some bimolecular mechanism
may explain the butene and acid selectivity at low temperature,
with the overall stoichiometry
3.4.2.1. Acetol conversion To detect the secondary products rising
◦
from acetol, we fed acetol into the MAT unit at 350 and 500 C
2 Acetone → butene + acetic acid.
(2)
and a WHSV close to 90 h−1. The results, presented in Table 6,
show that acetol was very reactive; with 86 wt% conversion at
Interestingly, coke selectivity, which usually increases with a bi-
molecular mechanism, was low in this case. At higher tempera-
◦
500 C, but still was less reactive than glycerol. As was the case
◦
tures (>450 C), acetic acid was converted into CO2 and hydro-
for glycerol conversion, oxygenate and coke yields decreased and
gas and hydrocarbon yields increased with increasing temperature.
Nevertheless, the detailed product distribution of acetol conver-
sion was quite different from that of glycerol conversion; whereas
coke selectivity was similar or slightly increased at low tempera-
ture, acetol yielded more acids and acetone and had unidentified
high-molecular-weight oxygenates/hydrocarbons as major prod-
ucts. Meanwhile, acetaldehyde yield was lower, and acrolein was
nearly missing as a reaction product. This confirms that glycerol
dehydrates through two distinct and independent pathways, one
leading to acrolein through 3-hydroxypropenal (a very unstable
product) and the other forming acetol, as shown in Fig. 3. The first
pathway implies the removal of the central alcohol function in the
glycerol molecule, whereas the second implies the removal of one
of the two terminal alcohol groups. Statistically, the glycerol would
then split into 66% acetol (and subproducts) and 33% acrolein.
carbons, as was shown by Gayubo et al. [45]. There was a higher
coke selectivity with 2-propanol, as well as a higher selectivity
toward high-molecular-weight oxygenated components. Production
of pentenes and BTX production also was greater. The 2-propenol
also could undergo hydrogenation to form propanol, which in turn
could undergo dehydration to yield propene. No propanol was
found as a product; another donor may exist that leads to propene,
similar to the following reaction:
propen-2-ol + H-donor → propene + H-receptor + H2O.
(3)
As in the case of acetol, such a donor can be coke, an olefin,
or another oxygenated oligomer obtained in the conversion of
2-propenol. Aromatics may be formed from small olefins gener-
ated during the cracking of acrolein. The aromatics thus formed
can then condense with small olefins to form coke, as is well
known in the cracking of hydrocarbons [46]. Although acetone and
2-propenol are similar in structure, they have very different con-
version networks, as shown in Fig. 5.
◦
A selectivity to acrolein of close to 60% was observed at 350 C,
indicating preferential removal of the central alcohol group with
ZSM5.
Acetone can be obtained from the hydrodeoxygenation of ace-
tol, as shown in Fig. 4. Whether the reaction proceeds through an
intermediate, 1,2-propane diol, which is not observed as a product,
or occurs directly in a combined reaction with H2 or a hydrogen
donor in a similar way as in catalytic cracking [42–44] as set in
Eq. (1) is not clear.
3.4.2.3. Propionic acid and propanediol conversion We also fed pro-
pionic acid and propanediol into the MAT reactor at 350 C to
◦
check for minor conversion pathways. The findings are compared
with those of acetol under similar operating conditions in Table 7.
Propanediol exhibited a reactivity similar to that of acetol along
with a high selectivity to acetone, obtained by direct dehydration.
The propionic acid reactivity was much lower, with a very high
selectivity to coke at low temperature and low conversion. The
main conversion pathways of these components complete the ace-
tol conversion scheme as shown in Fig. 4.
Acetol + H-donor → acetone + H-receptor + H2O.
(1)
Coke, olefins, and their acetol species can be H-donor species. The
formation of acid can be explained by the isomerization of acetol,
forming propionic acid.

A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
169
3.4.2.4. Reaction network Fig. 6 summarizes all of the foregoing
findings in a general glycerol conversion scheme. Along with the
information included in Figs. 3–5, Fig. 6 presents an additional
reaction pathway for the formation of acetaldehyde. Note that ac-
etaldehyde was found as a main product in glycerol conversion
and, to a lesser extent, in acetol conversion. No product demon-
strated acetaldehyde yields as high as glycerol yields, indicating
the formation of acetaldehyde by a different route than that de-
picted in Fig. 4. Chai et al. [37] and Tsukuba et al. [39], based
on a previous computational study [36], proposed that the de-
composition of 3-hydroxylpropenal should yield acetaldehyde and
formaldehyde by retro aldol condensation. Formaldehyde then may
be decomposed into CO + H2 or hydrogenated into methanol. In
hydrogen-transfer mechanism, with acetaldehyde formed from a
dehydrogenated dione intermediate, as shown in Fig. 6. The unsta-
ble dione reaction intermediates rapidly decomposed into CO and
acetaldehyde, and thus diones were not observed in reaction prod-
ucts.
Fig. 6 depicts the main conversion pathways for glycerol and
◦
◦
subproducts at 350–500 C. At higher temperatures (above 600 C),
acrolein and other oxygenates may undergo decarbonylation, yield-
ing ethylene and other small olefins as well as CO [47]. These
reactions account for the high olefin yield at high temperature
in a more direct way than at low temperature, where more oxy-
genated intermediates are preserved. Finally, steam cracking of
glycerol and all of the dehydrated species may produce CO and H2
at high temperature, and the WGS reaction also can produce more
H2 as well as CO2. The additional reactions to Fig. 6 that directly
produce CO, CO2, and H2 are then
◦
our study, at 350 C, formaldehyde was not detected as a prod-
uct and methanol was present in only trace amounts. Moreover,
◦
our hydrogen/CO molar ratio was much lower than 1 at 350 C,
whereas a ratio of 1 should result from formaldehyde decompo-
sition. Acetaldehyde also was found as a reaction product in the
conversion of acetol. Consequently, we propose that the mecha-
nism of acetaldehyde formation was coupled with some type of
CxHyOz + (x − z)H2O → xCO + (y/2 + x − z)H2
(4)
and
CO + H2O → CO2 + H2.
(5)
3.5. Energy balance for acrolein production
The FCC unit is attractive because the heat necessary for the
reaction is generated readily within the process. In state-of-the-art
units, 3–5% of the feed is converted to coke, which is burned in
the regenerator, generating sufficient heat to vaporize the feed and
support the endothermic reaction. Compared with VGO cracking,
the processing of glycerol solutions requires that more water to be
vaporized per unit of feed, thus increasing the heat requirement
(and hence the coke yield) per kg of feed processed.
To calculate the amount of coke that must be burned to main-
tain the process autothermal, we need the following parameters:
• The heat capacity and the heat of vaporization of the glycerol
Fig. 5. Acetone and 2-propenol conversion scheme.
solution, as well as the temperature of the feed and of the
Fig. 6. Proposed general glycerol conversion scheme. Reactant and main reaction products have been highlighted.

170
A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
Table 8
Increasing the acrolein selectivity requires minimizing any sec-
ondary reaction that leads to hydrogenated products or coke—that
is, limit the amount of hydrogen transfer.
Coke yield compared with heat balance needs
Glycerol dilution
% coke yield on feed
% coke on carbon in feed
20 wt%
50 wt%
85 wt%
7.9
6.3
3.7
101
32
11
4. Conclusion
In this paper, we have demonstrated how dehydration on a
zeolite-based catalyst in a moving-bed reactor can be used to
produce acrolein and other oxygenated chemicals from glycerol.
These oxygenated compounds are typically produced from the ox-
idation of nonrenewable petroleum derived feedstocks. The pro-
cess is based on standard FCC technology. We used a moving-bed
reactor (Microdowner reactor) and a fixed-bed reactor (Microac-
tivity test reactor) to test an equilibrated FCC catalyst (ECat) and
process, to estimate the heat requirement to vaporize and heat
up the glycerol solution;
• The heat of combustion of coke;
• The enthalpy of the reaction and the conversion.
◦
We assumed a feed temperature of 70 C and a reaction temper-
◦
ature of 350 C. The heat of combustion (complete combustion) is
◦
a ZSM5-based catalysts at temperatures of 290–650 C and catalyst
31,500 kJ/kg for carbon and 115,000 kJ/kg for hydrogen, and we
assumed 8 wt% hydrogen in coke. The reaction enthalpy was es-
timated to be slightly endothermic (10 kJ/mol). In any case, an
enthalpy value of 100 kJ/mol would not change the results sig-
nificantly, because this value is very small compared with the va-
porization and heating needs. Finally, the glycerol conversion was
considered complete. Calculations based on the foregoing assump-
tions led to the results presented in Table 8.
residence times of 0.5–30 s. The microdowner reactor simulates
the industrial fluid catalytic cracking process. The highest yield of
acrolein (55–61% molar carbon yield) was obtained at 350 C with
◦
the ZSM5 zeolite-based catalyst, at low catalyst-to-oil ratios (6–11)
and contact times of 0.5–2 s, which correspond to weight hourly
space velocities of 300–1300 h−1. Water did not significantly in-
fluence the yield of acrolein, and we were able to obtain high
yields of acrolein (55% and 62%) with aqueous glycerol solutions
ranging from 20 to 85 wt% glycerol. Increasing the temperature
We calculated the theoretical coke yield (referred to the amount
of feed) needed to maintain the autothermal process for several
glycerol concentrations and compared this value to the amount
of carbon (39.1% carbon in glycerol) present in the feed. Table 8
shows that processing diluted solutions of glycerol would require
burning an amount of coke similar to the amount of carbon
present in the feed; thus, to keep the process autothermal, all of
the carbon would be converted to coke, and hardly any valuable
chemicals would be formed. Consequently, an external heat source
is needed in this case. In contrast, with concentrated solutions,
such as those produced directly during the transesterification of
oils, the autothermal process would require conversion of only 10%
of the carbon present in the glycerol to coke, a yield similar to
◦
◦
from 350 C and 500 C caused a partial shift toward the produc-
tion of acetaldehyde. The advantage of using a moving-bed reactor
is that the catalysts can be continuously separated and regenerated
while producing the energy to keep the reaction. Future improve-
ments to catalyst and process improvements promise to achieve
higher yields of oxygenates.
Acknowledgments
Financial support was provided by CICYT (project MAT2006-
14274-C02-01) and BIOeCON BV.
◦
the experimental yield obtained in previous tests at 350 C with
References
a ZSM5-based catalyst. Thus, the processing of concentrated glyc-
erol solutions obtained from the synthesis of FAME could be car-
ried out in an autothermal process, without the need for external
heat.
[1] A.J. Ragauskas, C.K. Williams, B.H. Davison, G. Britovsek, J. Cairney, A. Eck-
ert, W.J. Frederick Jr., J.P. Hallet, D.J. Leak, C.H. Liotta, J.R. Mielenz, R. Murphy,
R. Templer, T. Tschaplinski, Science 311 (2006) 484.
[2] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044.
[3] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411.
[4] S.E. Koonin, Science 311 (2006) 435.
3.6. Costs, challenges, and future directions
[5] L.R. Lynd, C.E. Wyman, T.U. Gerngross, Biotechnol. Prog. 15 (1999) 777.
[6] G.W. Huber, A. Corma, Angew. Chem. Int. Ed. 46 (2007) 7184.
[7] J.N. Chheda, G.W. Huber, J.A. Dumesic, Angew. Chem. Int. Ed. 46 (2007) 7164.
[8] J.P. Lange, Biofuels Bioprod. Biorefin. 1 (2007) 39.
[9] J.N. Chheda, Y. Roman-Leshkov, J.A. Dumesic, Green Chem. 9 (2007) 342.
[10] Y. Roman-Leshkov, C.J. Barrett, Z.H. Liu, J.A. Dumesic, Nature 447 (2007) 982.
[11] G. Knothe, J. Krahl, J.V. Gerpen, The Biodiesel Handbook, AOCS Press, Cham-
paign, IL, 2005.
[12] G.W. Huber, P. O’Connor, A. Corma, Appl. Catal. A: Gen. 329 (2007) 120.
[13] M. Stumborg, A. Wong, E. Hogan, Bioresour. Technol. 56 (1996) 13.
[14] M. Snåre, I. Kubicˇková, P. Mäki-Arvela, D. Chichova, K. Eränen, D.Y. Murzin, Fuel
87 (2008) 933.
We obtained a 58% carbon acrolein yield, or 28 wt% acrolein
yield, with an 85 wt% glycerol feed. Typical yields for acrolein
production from propylene oxidation are around 85% mol; that is,
ca. 900 kg of propylene (combined with oxygen from air) is needed
to obtain 1 ton of acrolein. With feedstock costs of 1000 $/ton for
propene and 300 $/ton for glycerol solution, the feed cost for the
glycerol-based process (1120 $/ton of acrolein) is close to that for
the propene-based process (900 $/ton of acrolein). These calcu-
lations show that acrolein production from crude glycerine from
a biodiesel plant is already nearly competitive with acrolein pro-
duced from propylene oxidation based on feedstock costs, with
a safer and less energy-intensive process. The price of acrolein
from glycerol can be further improved through further process im-
provements to achieve higher yields of acrolein (with a theorical
maximum of 52 wt% from the raw glycerol solution). Any car-
bon emission credit obtained from the use of a renewable stock
(e.g., glycerol from biomass) can further improve the economics of
the process. In addition, significant amounts of acetaldehyde also
can be recovered, adding more commercial value. Moreover, the
acrolein-to-acetaldehyde ratio can be readily changed by shifting
the process temperature.
[15] M. McCoy, Chem. Eng. News 83 (8) (2005) 19.
[16] T. Werpy, G. Peterson, Top Value Added Chemicals from Biomass, Pacific North-
west National Laboratory, National Renewable Energy Laboratory, 2004.
[17] M. Bender, Bioresour. Technol. 70 (1999) 81.
[18] P. Sabatier, G. Gaudion, Compt. Rend. 166 (1918) 1033.
[19] E. Freund, US Patent 1 672 378 (1928), to Chemische Fabrik Auf Actien.
[20] H.E. Hoyt, T.H. Manninen, US Patent 2 558 520 (1951), to U.S. Industrial Chem-
icals.
[21] H. Heinemann, R.W. Wert, Ind. Eng. Chem. 41 (1949) 2928.
[22] A. Neher, T. Haas, US Patent 5 387 720 (1993), to Degussa Aktiengesellschaft.
[23] W. Buhler, E. Dinjus, J. Supercrit. Fluids 22 (2002) 37.
[24] L. Ott, M. Bicker, Green Chem. 8 (2006) 214.
[25] S. Ramayya, A. Brittain, Fuel 66 (1987) 1364.
[26] E. Drent, W.W. Jager, US Patent 6 080 898 (2000), to Shell Oil Company.
[27] A. Corma, C. Corell, F. Llopis, A. Martínez, J. Pérez-Pariente, Appl. Catal. A: Gen.
115 (1994) 121.
As shown in Fig. 6, glycerol readily yields acrolein through de-
hydration; thus, the theorical maximum carbon selectivity is 100%.
[28] A. Corma, F.V. Melo, L. Sauvanaud, Appl. Catal. A: Gen. 232 (2002) 247.

A. Corma et al. / Journal of Catalysis 257 (2008) 163–171
171
[29] A. Corma, F.V. Melo, L. Sauvanaud, F. Ortega, Catal. Today 107–108 (2005) 699.
[30] R. Sadeghbeigi, Fluid Catalytic Cracking Handbook, Eds. Gulf Publishing Com-
pany, Houston, TX, 1995.
[31] P. Hagelberg, I. Eilos, J. Hiltunen, K. Lipiänen, V.M. Niemi, J. Aittamaa, A.O.I.
Krause, Appl. Catal. A: Gen. 223 (2002) 73.
[32] H.C. Beinaert, R. Vermeulen, G.F. Froment, in: B. Delmon, G.F. Froment (Eds.),
Catalytic Deactivation, in: Stud. Surf. Sci. Cat., vol. 88, Elsevier, Amsterdam,
1994, p. 97.
[37] S. Chai, H. Wang, Y. Liang, B. Xu, Green Chem. 9 (2007) 1130.
[38] S. Chai, H. Wang, Y. Liang, B. Xu, J. Catal. 350 (2007) 342.
[39] E. Tsukuda, S. Sato, R. Takahashi, T. Sodewasa, Catal. Commun. 8 (2007) 1349.
[40] J. Barrault, J.M. Clacens, Y. Pouilloux, Top. Catal. 27 (2004) 137.
[41] A.G. Gayubo, A.T. Aguayo, A. Atutxa, R. Aguado, J. Bilbao, Ind. Eng. Chem. Res.
43 (2004) 2610.
[42] A. Corma, B.W. Wojciechowski, Catal. Rev. 27 (1985) 29.
[43] A. Corma, P.J. Miguel, A.V. Orchillés, J. Catal. 145 (1994) 171.
[44] A. Corma, V. Fornés, E. Ortega, J. Catal. 92 (1985) 284.
[45] A.G. Gayubo, A.T. Aguayo, A. Atutxa, R. Aguado, M. Olazar, J. Bilbao, Ind. Eng.
Chem. Res. 43 (2004) 2619.
[33] A. Corma, F.V. Melo, L. Sauvanaud, Appl. Catal. A: Gen. 287 (2005) 34.
[34] W. Buhler, E. Dinjus, H.J. Ederer, A. Kruse, C. Mas, J. Supercrit. Fluids 22 (2002)
37.
[35] M.J. Antal, W.S.L. Mok, J.C. Roy, A.T. Raissi, J. Anal. Appl. Pyrolysis 8 (1985) 291.
[36] M.R. Nimlos, S.J. Blanksby, X. Qian, M.E. Himmel, D.K. Johnson, J. Phys. Chem. A
110 (2006) 6145.
[46] M. Guisnet, P. Magnoux, in: B. Delmon, G.F. Froment (Eds.), Catalytic Deactiva-
tion, in: Stud. Surf. Sci. Cat., vol. 88, Elsevier, Amsterdam, 1994, p. 53.
[47] A. Corma, G.W. Huber, L. Sauvanaud, P. O’Connor, J. Catal. 247 (2007) 307.