Comparison of kinetics and reaction pathways for hydrodeoxygenation of C3 alcohols on Pt/Al2O3

Article abstract of DOI:10.1016/j.cattod.2011.10.022

The catalytic hydrodeoxygenation of C₃ alcohols (1- and 2-propanol, 1,2- and 1,3-propanediol, and glycerol) on Pt/Al₂O ₃ has been mechanistically explored in the aqueous phase. Dehydrogenation on Pt and dehydration on alumina are the main elementary reaction pathways. In water, carbon-carbon bond cleavage for alcohols with terminal hydroxyl groups occurs via decarbonylation of aldehydes (generated by dehydrogenation of alcohols) and decarboxylation of acids, the latter being formed by disproportionation from aldehydes. The presence of water as solvent suppresses the dehydration for mono-alcohols mainly via blocking of Lewis acid sites by water. Dehydration is still the dominating primary reaction for 1,3-propanediol and glycerol, as the higher number of hydroxyl groups weakens the C-O bond strength. The overall reactivity of C₃ alcohols decreases in the order of 1,3-propanediol ≈ glycerol > 1,2-propanediol ≈ 1-propanol.

Full text of DOI:10.1016/j.cattod.2011.10.022

Catalysis Today 183 (2012) 3–9
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Comparison of kinetics and reaction pathways for hydrodeoxygenation of C3
alcohols on Pt/Al O3
2
1
1
Baoxiang Peng, Chen Zhao, Isidro Mejía-Centeno , Gustavo A. Fuentes , Andreas Jentys,
Johannes A. Lercher
∗
Department of Chemistry and Catalysis Research Center, Technische Universität München, Lichtenbergstraˇe 4, 85747 Garching, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 July 2011
Received in revised form 24 October 2011
Accepted 25 October 2011
Available online 21 November 2011
The catalytic hydrodeoxygenation of C3 alcohols (1- and 2-propanol, 1,2- and 1,3-propanediol, and glyc-
erol) on Pt/Al2O3 has been mechanistically explored in the aqueous phase. Dehydrogenation on Pt and
dehydration on alumina are the main elementary reaction pathways. In water, carbon–carbon bond cleav-
age for alcohols with terminal hydroxyl groups occurs via decarbonylation of aldehydes (generated by
dehydrogenation of alcohols) and decarboxylation of acids, the latter being formed by disproportionation
from aldehydes. The presence of water as solvent suppresses the dehydration for mono-alcohols mainly
via blocking of Lewis acid sites by water. Dehydration is still the dominating primary reaction for 1,3-
propanediol and glycerol, as the higher number of hydroxyl groups weakens the C–O bond strength. The
overall reactivity of C3 alcohols decreases in the order of 1,3-propanediol ≈ glycerol > 1,2-propanediol ≈ 1-
propanol.
Keywords:
Aqueous phase
Decarbonylation
Decarboxylation
Disproportionation
Hydrodeoxygenation of alcohols
Glycerol
© 2011 Published by Elsevier B.V.
1
. Introduction
Selective conversion of biomass resources such as polysaccha-
potential hydrodeoxygenation ability on alcohols and exhibited
exceeding 90% hydrogen selectivity from APR of alcohols [18,19],
was selected in this work.
rides [1,2], lignin [3,4], bio-ethanol, and glycerol [5–11] requires
highly efficient catalysts. Within these biomass feedstocks, glyc-
erol is a very attractive option due to its relatively low cost and
wide availability as a by-product of bio-diesel production. It is con-
sidered to be, therefore, one of the top 12 building block chemicals
of a biorefinery process [12].
Glycerol can be either converted to high value-added oxy-
genated chemicals such as propanediols via hydrodeoxygenation,
or can be used to produce hydrogen through aqueous phase
reforming process. The aqueous phase deoxygenation of glycerol
to propanediols requires selectively cleaving one of the C–O bonds,
but preserving the C–C bonds. It occurs catalytically on dual func-
tional catalysts via consecutive dehydration and hydrogenation in
the presence of H2 at moderate temperatures (453–543 K) and rel-
atively high pressures (20–150 bar) [13–17]. On the other hand,
the parallel route of aqueous phase reforming (APR) also occurs
at identical conditions used for hydrodeoxygenation of alcohols
The competing pathways of aqueous phase hydrodeoxygena-
tion and reforming have been extensively explored by Dumesic
[9–11], Davis [20,21], Tomishige [14,22], and ourselves [23]. The
key feature in determining the selectivity is the way and the extent
of C–C and C–O bond cleavage. The pathway for C–O bond cleavage
was suggested to occur either through dehydrogenation to form
surface adsorbed species followed by direct cleavage catalyzed by
metallic sites or via dehydration reactions catalyzed by acid sites
associated with catalyst support [9]. The mechanism for C–C bond
cleavage in aqueous phase reforming of alcohols producing smaller
alkanes, CO , and H2 has not been unequivocally explained. While,
2
the C–C bonds in alcohols were speculated to be cleaved by the
metal catalyzed direct hydrogenolysis [9,18], we showed first evi-
dence in a recent paper that direct hydrogenolytic cleavage of C–C
bonds does not occur [23].
In the present paper, the kinetics in the catalytic conversion of C3
alcohol molecules with different position of the hydroxyl group and
number of hydroxyl groups (mono-alcohols, i.e., 1-propanol and 2-
propanol; diols, i.e., 1,2-propanediol and 1,3-propanediol; and triol,
[
9–11]. The bifunctional catalyst Pt supported on Al O , which had
2 3
i.e., glycerol) were systematically studied over Pt/Al O3 in order to
2
elucidate the reaction pathways and fundamental chemistry that
lead to the C–C and C–O bond cleavage in the aqueous phase alcohol
transformation. In addition, the different activity performances of
2-propanol in the gas phase and in the aqueous phase over Pt/Al2O3
∗ Corresponding author. Tel.: +49 89 289 13540; fax: +49 89 289 13544.
E-mail address: johannes.lercher@ch.tum.de (J.A. Lercher).
1
Permanent address: Área de Ingeniería Química, Universidad A. Metropolitana-
Iztapalapa, México, D.F. 09340, Mexico.
0
920-5861/$ – see front matter © 2011 Published by Elsevier B.V.
doi:10.1016/j.cattod.2011.10.022

4
B. Peng et al. / Catalysis Today 183 (2012) 3–9
have also been compared to elucidate specific effects of the aqueous
environment.
considered for an accurate analysis. The solubility of CO2 in water
was, therefore, calculatedaccordingto thepublishedliterature[24].
At 473 K and a relatively low CO2 partial pressure, the solubility of
−
2
2
. Experimental
CO2 in water approximately equals to 10 ∗ pCO2 (mol/kg water,
pCO2 in unit of bar).
2.1. Catalysts preparation
3. Results and discussion
The 3 wt.% Pt/Al O3 catalyst was prepared by the incipient wet-
2
ness impregnation method, with platinum(II)-ammonium nitrate
3
.1. Catalysts characterization
(
(
[Pt(NH ) ](NO ) , Strem chemicals) as precursor and ␥-Al O
Aeroxide Alu C-Degussa, specific surface area: 105 m /g) as car-
3 4 3 2 2 3
2
The physicochemical properties of the 3 wt.% Pt/Al O3 are sum-
2
rier. After impregnating the carrier with the aqueous solution of Pt
precursor at ambient temperature, the catalyst was dried in air at
2
marized in Table 1. The BET specific surface area was 88 m /g, while
the dispersion of Pt was 90%. The acid and base site concentra-
tions of 3 wt.% Pt/Al O were 0.160 and 0.015 mmol/g, respectively.
It should be noted that the acid–base properties only represent
the starting catalyst, as the alumina support transforms during the
reaction under hydrothermal environment into aluminum hydrox-
ide (Böhmite).
3
93 K for 12 h and calcined in synthetic air for 2 h at 673 K. Prior to
2
3
the reaction and characterization, the catalyst was reduced in H2
at 573 K for 2 h.
2.2. Catalyst characterization
Atomic absorption spectroscopy (AAS) was used for analyz-
3
.2. Conversion of 2-propanol on Pt/Al O in the gas phase
ing the metal loading. The nitrogen adsorption–desorption was
adopted for measuring BET surface area and pore size distribu-
tion. The fraction of accessible Pt atoms was detected by hydrogen
chemisorption. The temperature programmed desorption (TPD) of
ammonia and carbon dioxide was used for acid and basic sites mea-
suring. The characterization methods have been described in detail
in a previous publication [23].
2 3
In order to compare the catalytic conversion of an alcohol at the
gas–solid interface with the conversion at the aqueous–solid inter-
face, we firstly explored the reactions of 2-propanol on Pt/Al O3
in the gas phase. It is known that the catalyst acidity is related to
its ability to dehydrate 2-propanol to propene, while its dehydro-
genation to acetone is catalyzed by a catalytic function that is able
to abstract a proton by a strong basic site and the hydride anion
by a Lewis acid site or a metal function [25,26]. The results of 2-
propanol conversion over 3 wt.% Pt/Al O3 at 1 bar in the gas phase
are shown in Fig. 1.
2
2
2
.3. Catalyzed reactions
.3.1. Gas phase reaction
The dehydrogenation/dehydration of 2-propanol was per-
formed in a continuous fixed bed flow reactor at atmospheric
pressure and temperature ranging from 393 to 523 K. The quartz
tubular reactor (4 mm diameter) was packed with 20 mg catalyst
2
It shows that the dehydrogenation and dehydration com-
peted over the whole conversion range. Dehydrogenation of
2
-propanol to acetone over 3 wt.% Pt/Al O3 was the dominat-
2
ing reaction at relatively low temperatures (below 473 K), while
propane formed via consecutive dehydration and hydrogenation
became the major product at higher temperatures. In addition,
diluted in 100 mg SiC. After the catalyst activation in H at 573 K for
2
2
h, the reaction was performed by introducing a H flow saturated
2
with 2-propanol at 286.2 K (the 2-propanol partial pressure was
5 mbar) into the reactor. The effluent products were analyzed by a
pure Al O3 selectively dehydrated of 2-propanol to propene at
2
2
identical conditions. Therefore, it can be concluded that dehy-
gas chromatography equipped with FID detector and Supelco-Wax
column.
dration was catalyzed by the Lewis acid sites of Al O (reaction
2
3
−
1
−1
rate at 473 K: 69.0 mmol s molacid site ) and Pt was responsi-
ble for the dehydrogenation (reaction rate at 473 K: 96.0 mmol s
molPt-surf atom ). The apparent activation energies for dehydration
−
1
2.3.2. Aqueous phase reaction
−
1
Experiments with 1-propanol, 2-propanol, 1,2-propanediol, 1,3-
and dehydrogenation of 2-propanol over 3 wt.% Pt/Al O3 were 71
propanediol, and glycerol were conducted in a 300 ml batch
autoclave (Parr Instrument). The reactants and the catalyst loaded
in a closed glass vial were charged into the reactor, and then
2
and 44 kJ/mol, respectively.
the reactor was purged with N , which was also used as inter-
2
1
00
nal standard for vapor phase products analysis. When the required
temperature and pressure were reached, the reaction was started
after breaking the glass vial by stirring. The vapor phase was ana-
lyzed online by a gas chromatograph with TCD detector and two
capillary columns (MS-5A and HP-Plot Q). Liquid samples were
manually collected during the run and analyzed in a gas chromato-
graph equipped with an FID detector and a CP-Wax 57CB column.
Typical reactions were conducted under the following conditions:
8
0
60
4
2
0
0
0
4
73 K, 40 bar total pressure of H , 100 g of 10 wt.% reactant aque-
2
ous solution, 0.3 g of 3 wt.% Pt/Al O , and 600 rpm stirring speed.
2
3
All results were calculated and reported based on carbon basis.
373
423
473
523
2.4. Equilibrium and CO2 solubility calculation
Temperature (K
)
The reaction equilibrium compositions were calculated using
Fig. 1. Conversion of 2-propanol (ꢀ) and yield of acetone (ꢁ) and propane (ꢂ)
as a function of temperature on 3 wt.% Pt/Al2O3 in gas phase reaction. (Reaction
conditions: 3 wt.% Pt/Al2O3 20 mg, 2-propanol partial pressure 25 mbar, H2 flow
40 ml/min.)
the HSC software. Under reaction temperature and pressure, CO2
would be produced and partially dissolved in water. This part of
CO2 cannot be detected by gas chromatograph, but needs to be

B. Peng et al. / Catalysis Today 183 (2012) 3–9
5
Table 1
Physicochemical properties of Pt/Al2O3 catalyst.
Catalyst
Pt loading
wt.%)
Surface
area (m /g)
Dispersion
(H/Pt)
Acidity
(mmolNH3/g)
Basicity
(mmolCO2/g)
2
(
Pt/Al2O3
2.93
88
0.90
0.16
0.015
3
.3. Conversion of 2-propanol on Pt/Al O in the aqueous phase
3.4. Aqueous phase conversion of 1-propanol
2
3
The dehydrogenation to acetone was the dominating primary
The products of 1-propanol conversion versus reaction time
reaction of the conversion of 2-propanol on Pt/Al O3 at 473 K
over 3 wt.% Pt/Al O3 are plotted in Fig. 3. In contrast to 2-propanol
2
2
in the aqueous phase (see Fig. 2). The initial rate of dehydro-
conversion, the reaction of 1-propanol mainly led to CO and ethane
2
−1
−1
−1
−1
genation (48.0 mmol s molPt-surf atom ) was reduced to 50% of
(initial rate: 1.1 mmol s molPt-surf atom ). The ratio of ethane to
CO2 was approximately 2.0 based on carbon basis. The presence
of H2 (cannot be detected in H2 atmosphere) is inferred from the
reaction stoichiometry. Similar to the reaction of 2-propanol, only
small amounts of propane (not shown at here) were formed from
the C3 alcohol dehydration–hydrogenation on Al O . The dehy-
−
1
−1
the rate found in the gas phase (96.0 mmol s molPt-surf atom ).
As the hydrogen concentration (0.023 mol/L, calculated from
the solubility of hydrogen in water [27]) in aqueous phase
reaction is comparable with that in the gas phase reaction
(
0.026 mol/L), the activity coefficient of hydrogen is close to one.
2
3
−
1
−1
We speculate that the slower dehydrogenation rate in aqueous
phase is, therefore, mainly related to the competitive adsorp-
tion of water and 2-propanol on Pt active sites. Only a small
fraction was converted to propane, presumably via the slow
dration rate of 1-propanol (0.3 mmol s mol
) was slightly
acid site
−
1
−1
lower than the rate of 2-propanol (0.4 mmol s molacid site ) (see
Table 2), which is in agreement with the literature [30] that the sec-
ondary hydroxyl group exhibits higher dehydration activity than
the primary hydroxyl group.
dehydration of the alcohol on the Al O support (initial rate:
2
3
−1
−1
0
Pt.
.4 mmol s molacid site ) and the following hydrogenation on
Propanal was only observed in traces (concentration < 0.02%),
which suggests rapid decarbonylation to ethane and CO or dis-
proportionation of the formed propanal to propanol and propionic
It is important to note that the dehydration rate of 2-propanol
in the aqueous phase was two orders of magnitude slower than
that in the gas phase (see Table 2). This could be attributed either
to the decrease in Lewis acidity caused by the transformation of ␥-
Al O to aluminum hydroxide (Böhmite) [28,29], or to the fact that
acid followed by decarboxylation to ethane and CO . Propanal and
2
propionic acid were clearly detected as reaction intermediates in
relatively large amounts for 1-propanol conversion in the absence
of hydrogen [23]. The direct pathway of hydrogenolysis of 1-
propanol can be excluded, as the C–C bond strengths in 1-propanol
and 2-propanol are quite similar (see Table 3, i.e., 357 kJ/mol for
1-propanol and 368 kJ/mol for 2-propanol). This indicates that the
formation of hydrogen, ethane, and CO2 follows a reaction path-
way characterized by dehydrogenation, decarbonylation (–CO) or
disproportionation (Tishchenko or Cannizzaro type reactions), and
2
3
water essentially blocks the Lewis acid sites active in dehydrating
-propanol in the gas phase. As it has been established that the
transformation of ␥-Al O in water is a relatively slow process [29],
2
2
3
and the decrease of 2-propanol dehydration rate is substantial even
at initial time when alumina is still not measurably converted, we
conclude the drastic activity change of Pt/Al O in aqueous solution
2
3
is mainly caused by the competitive adsorption of water on Lewis
acid sites.
decarboxylation (–CO ).
2
At present, it is not possible to differentiate between these two
latter mechanisms (Tishchenko or Cannizzaro) for the dispropor-
tionation of aldehyde into an acid and an alcohol. The Tishchenko
reaction is catalyzed by acid catalysts, and an ester (propylpropi-
onate) is formed in the first step, which is subsequently rapidly
hydrolyzed at the current reaction conditions. On the other hand,
the occurrence of Cannizzaro reaction requires basic sites of the cat-
alysts, and it can be strongly enhanced by transition metals [32].
Experiments with pure ␥-Al O3 showed a significantly lower
2
conversion, which confirmed again that Pt is essential for 2-
propanol dehydrogenation. It is noted that the potential products
of C–C or C–O bond cleavage from 2-propanol were not observed,
suggesting that Pt/Al O3 is not able to cleave these bonds in 2-
2
propanol under the reaction conditions. It shows unequivocally
that hydrogenolysis reactions do not occur.
Both disproportionation pathways are possible with Pt/Al O , as
2
3
the Böhmite compound contains both acid and base sites.
0
0
0
0
0
0
.5
.4
.3
.2
.1
.0
6
4
3
1
0
.0
.5
.0
.5
.0
0.4
0.3
0.2
0.1
0
0
60
120
180
240
300
360
0
60
120
180
240
300
360
Time (min)
Time (min)
Fig. 2. Yield of acetone (ꢁ) and propane (ꢂ) on conversion of 2-propanol over
wt.% Pt/Al2O3 as a function of time in aqueous solution. (Experimental conditions:
T = 473 K, total pressure 40 bar H2, 2-propanol concentration 10 wt.%.)
Fig. 3. Yield of ethane (ꢁ), carbon dioxide (ꢂ) on conversion of 1-propanol over
3 wt.% Pt/Al2O3 as a function of time in aqueous solution. (Experimental conditions:
T = 473 K, total pressure 40 bar, 1-propanol concentration 10 wt.%.)
3

6
B. Peng et al. / Catalysis Today 183 (2012) 3–9
Table 2
Comparison of reaction rates for conversion of C3 alcohols in aqueous phase at 473 K.
Reaction pathways, products and reaction rates
Reactants
Dehydrogenation
Dehydration
(mmol s molacid site
Decarbonylation/decarboxylation
(mmol s molPt-surf atom
−1
−1
−1
−1
−1
−1
)
(mmol s molPt-surf atom
)
)
96.0b
Acetone
48.0b
Acetone
–
69.0
Propane
0.4
Propane
0.3
Propane
0.2
–
–
–
–
2
2
1
1
1
-Propanola
-Propanol
-Propanol
1.1*
Ethane, CO2
1.0
Ethanol, methane, CO2
1.0
Ethanol, ethane, CO2
1.4
Ethylene glycol, ethanol,
methanol, CO2
–
8.5
b
,2-Propanediol
,3-Propanediol
Hydroxyacetone
1-Propanol, 2-propanol
11.5
b
–
–
–
–
1-Propanol
b
Glycerol
4.6
Hydroxyacetone,
1,2-propanediol
a
Reaction performed under gas phase.
Primary reaction pathway and main products.
b
Table 3
C–C and C–O bond energies in C3 alcohols [31].
Compounds
1-Propanol
2-Propanol
1,2-Propanediol
Glycerol
Glycerol
OH
|
-CH-CH
OH OH
OH OH OH
OH OH OH
C
2
H
5
CH
2
—OH
|
|
|
|
|
|
|
|
Structure
CH
3
3
CH
3
-CH-CH
2
CH
2
-CH-CH
2
2 2
CH -CH-CH
C–O bond energy (kJ/mol)
Structure
392.0
397.9
–
335.6
333.0
OH
|
OH OH
OH OH OH
|
CH -CH—CH
2 2
C
2
H
5
—CH
2
OH
|
|
|
|
CH
3
—CH-CH
3
CH
3
-CH—CH
2
C–C bond energy (kJ/mol)
356.9
367.8
358.9
347.0
In the potential route of propanal decarbonylation to ethane
and CO, the produced CO is converted further to CO2 and hydro-
gen through the water gas shift reaction. It has been observed
that the concentration of CO in the products was slightly higher
or rapidly converted to 1,2-propanediol and 2-hydroxypropanoic
acid via Tishchenko/Cannizzaro type disproportionation followed
by decarboxylation of the acid, leading to ethanol and CO2.
The formation rate of ethanol decreased slightly, while the for-
mation rate of CO2 increased moderately with the reaction time,
leading to a ratio of ethanol to CO2 that is lower than the 2.0
on carbon basis expected for ideal stoichiometry. This indicates
that ethanol underwent further reaction to methane and CO2 via
analogous reaction pathways as described above. If the converted
ethanol was also taken into account, the initial ratio of ethanol to
CO (calculated by (ethanol + 2× methane)/(CO –methane) accord-
(
0.015 C%) than the expected equilibrium value (0.007 C%, cal-
culated from the observed CO2 concentration). Therefore, it can
be concluded that the parallel reaction of decarbonylation also
contributes to 1-propanol conversion. However, quantitative dif-
ferentiation between the reaction pathways of disproportionation
followed by decarboxylation of the acid and decarbonylation fol-
lowed by water gas shift is not attempted here.
2
2
ing to stoichiometry) was very close to 2.0 on carbon basis, clari-
fying that ethanol is involved to some extent in decarboxylation
3.5. Aqueous phase conversion of 1,2- and 1,3-propanediols
The reactions of 1,2-propanediol having a primary and a sec-
ondary hydroxyl group, and 1,3-propanediol with two primary
hydroxyl groups were studied at identical reaction conditions
0
0
0
.4
.3
.2
(
see Figs. 4 and 5). Hydroxyacetone, ethanol, and CO2 were the
main products of 1,2-propanediol conversion, and small amounts
of 1-propanol, 2-propanol, and methane were also formed. The
yield of hydroxyacetone formed through dehydrogenation of 1,2-
−1
−1
propanediol (initial rate: 8.5 mmol s molPt-surf atom ) rapidly
reached a constant value, which is determined by the high hydro-
gen pressure limiting further conversion.
0.1
Because the direct hydrogenolysis of C–C bonds of propanol does
not occur under the present reaction conditions (see above), and the
C–C bond strengths of 1,2-propanediol (359 kJ/mol) is quite close to
0
.0
0
60
120
180
240
300
360
that of propanol (357 kJ/mol) (see Table 3), ethanol and CO are con-
2
Time (min)
cluded to be formed via decarbonylation or decarboxylation (initial
−1
−1
rate: 1.0 mmol s molPt-surf atom ). For the reaction pathway, 1,2-
propanediol is dehydrogenated to 2-hydroxypropionaldehyde (not
detected due to its high reactivity), which is either instantly decar-
bonylated to ethanol and CO followed by water gas shift reaction,
Fig. 4. Yield of hydroxyacetone (ꢀ), ethanol (ꢁ), carbon dioxide (ꢂ), methane
ꢀ), 1-propanol (ꢀ), and 2-propanol (ꢃ) on conversion of 1,2-propanediol over
wt.% Pt/Al2O3 as a function of time in aqueous solution. (Experimental conditions:
T = 473 K, total pressure 40 bar, 1,2-propanediol concentration 10 wt.%.)
(
3

B. Peng et al. / Catalysis Today 183 (2012) 3–9
7
0
0
0
0
0
0
.25
.20
.15
.10
.05
.00
6
5
4
3
2
1
0
0
60
120
180
240
300
360
Time (min)
Fig. 5. Products distribution for 1,3-propanediol conversion over 3 wt.% Pt/Al2O3
in aqueous solution in the presence of H2; 1-propanol (᭹), ethanol (ꢀ), carbon
dioxide (ꢂ), ethane (ꢁ). (Experimental conditions: T = 473 K, total pressure 40 bar,
Fig. 6. Glycerol conversion over 3 wt.% Pt/Al2O3 in the aqueous solution in the pres-
ence of H2. 1,2-PD, 1,2-propanediol; acetol, hydroxyacetone; EG, ethylene glycol;
EtOH, ethanol; others = 1-propanol + 2-propanol + methanol. (Experimental condi-
tions: T = 473 K, total pressure 40 bar, glycerol concentration 10 wt.%.)
1
,3-propanediol concentration 10 wt.%.)
or decarbonylation reactions. 1-Propanol and 2-propanol were
produced through sequential dehydration–hydrogenation of 1,2-
confirms the proposed reaction route. Therefore, the dehydration
to hydroxyacetone is the dominating primary reaction step for glyc-
erol conversion (initial rate: 4.6 mmol s molacid site ).
−
1
−1
−1
−1
propanediol over Al O3 (initial rate: 0.2 mmol s molacid site ),
2
and the produced amount of 1-propanol was higher than that of
It is important to emphasize that all the other products
appear to be formed along the same pathways as for other alco-
hols. Glyceraldehyde generated by dehydrogenation of glycerol
2
-propanol, which is attributed to the higher dehydration activ-
ity of secondary hydroxyl group than that of the primary hydroxyl
group.
is rapidly converted to ethylene glycol and CO , either through
2
The results of the conversion of 1,3-propanediol at 473 K
and 40 bar hydrogen are compiled in Fig. 5. 1-Propanol was the
sequential decarbonylation and water gas shift reaction or via
Tishchenko/Cannizzaro type disproportionation reactions over
main product at initial conversion. Ethane, ethanol, CO , and
acid–base sites of Pt/Al O3 followed by subsequent decarboxy-
2
2
trace amounts of propane, propanal, propionic acid, methane, and
methanol were also observed.
lation of glyceric acid. These intermediates were not detected,
because their concentrations were very low due to their high reac-
tivity and the equilibrium limitations caused by the high hydrogen
pressure. Ethylene glycol can be further converted, either it is dehy-
drated to acetaldehyde, which in turn is hydrogenated to ethanol,
or it is decomposed to CO2 via dehydrogenation, disproportiona-
tion, and decarboxylation reactions (or decarbonylation reaction).
The slight difference (5%) of C–C bond strengths in glycerol and
propanols allows us to exclude direct hydrogenolysis for C–C bond
cleavage pathway (see Table 3). 1-Propanol and 2-propanol were
formed by subsequent conversion of propanediols along the path-
ways discussed above.
It is still not clear, if the acid sites active in the dehydration in
water are of Brønsted or Lewis acid character. The strong absorp-
tion of water onto Lewis acid sites allows the conversion of Lewis
acid sites to surface hydroxyl groups via rehydroxylation (or rehy-
dration). This two-step rehydroxylation transformation involves
1
-Propanol is highly selectively produced through the dehydra-
−
1
−1
tion of 1,3-propanediol (initial rate: 11.5 mmol s molacid site )
and the subsequent hydrogenation, which then further under-
goes the same reaction sequences of dehydrogenation, followed
by either decarbonylation with a subsequent water gas shift
reaction or disproportionation with a subsequent decarboxy-
lation, leading to CO2 and ethane. The observation of trace
amounts of propanal and propionic acid provides strong evidence
for this proposed decarbonylation and decarboxylation reaction
pathway. CO2 also can be produced directly from the starting
reactant 1,3-propanediol via the parallel reaction pathway of dehy-
drogenation, decarbonylation and decarboxylation (initial rate:
− −1
.0 mmol s molPt-surf atom ). The ratio of C compounds (ethanol
2
1
1
and ethane) to CO2 is approximately 2.0 based on carbon basis,
suggesting that CO2 is produced through both reaction pathways.
Propane is concluded to be formed via the dehydration of 1-
propanol. We speculate at present that methane is formed by the
dehydrogenation, decarboxylation/decarbonylation of hydrogena-
tion of CO2.
nondissociative adsorption of H O on the Lewis acid sites followed
2
by the subsequent dissociative chemisorption of H O and modifi-
2
cation of the alumina surface [33,34]. Although the transformation
of bulk ␥-Al O to hydroxide is slow, the surface rehydroxylation is
2
3
fairly rapid and thorough and so the concentration of exposed Lewis
acid sites is limited. We speculate at present that weak Brønsted
acid sites resulting e.g., from some isolated surface hydroxyl groups
and identified by the shift of ꢁOH on CO infrared adsorption [35], act
as active sites for dehydration.
3
.6. Aqueous phase conversion of glycerol
The results of glycerol conversion in the aqueous phase at 473 K
in the presence of H2 (40 bar) are compiled in Fig. 6. Hydroxyace-
tone, 1,2-propanediol, ethylene glycol, ethanol, and CO2 were the
main products, and 1-propanol, 2-propanol, and methanol were
formed in small concentrations.
3.7. Summary of reaction pathways and reaction rates for
aqueous phase conversion of C3 alcohols over Pt/Al O
2
3
1
,2-Propanediol was selectively produced from glycerol (75%
selectivity). It can be seen from Fig. 6 that the increase of 1,2-
propanediol nearly equals the decrease of hydroxyacetone in
selectivity as a function of time, suggesting that 1,2-propanediol
is produced from the hydrogenation of hydroxyacetone. A con-
trol experiment with hydroxyacetone as reactant shows 98.5%
selectivity to 1,2-propanediol under identical conditions, which
Summarizing the aqueous phase reactions with C3 alcohols
over Pt/Al O , dehydrogenation of alcohols to ketones on Pt is the
main route for conversion of 2-propanol and 1,2-propanediol, and
dehydration on alumina is the main route for conversion of 1,3-
propanediol and glycerol. It has been demonstrated that the direct
hydrogenolysis of C–C and C–O bonds of the alcohols does not
2
3

8
B. Peng et al. / Catalysis Today 183 (2012) 3–9
R
decarboxylation
O
R '
R
R'
R
H
R
2
OH
CO
2
R '
OH
R'
O
disproportionation*
R'
dehydrogenation
R
OH
WGS
R, R’ = H or OH
CO
decarbonylation
Fig. 7. Proposed main reaction pathways for C–C bond cleavage in aqueous phase conversion of glycerol derived alcohols over Pt/Al2O3 (* either via Tishchenko or Cannizzaro
type reactions).
take place over Pt/Al O3 under the used reaction conditions. In
The C–O bonds of C3 alcohols are cleaved by dehydration,
while the C–C bonds of C3 alcohols with terminal hydroxyl
groups are cleaved by sequential dehydrogenation to aldehyde,
followed by either disproportionation (Tishchenko or Cannizzaro
type reactions) with a subsequent decarboxylation reaction, or
decarbonylation with a subsequent water gas shift reaction. The
presence of terminal hydroxyl group of alcohols is proved to be
critical for C–C bond cleavage in this reaction sequence, as it
allows forming the essential aldehyde intermediate, which opens
the reaction pathway to decarbonylation and decarboxylation. The
overall reaction rates decrease in the sequence of 1,3-propanediol
≈ glycerol > 1,2-propanediol ≈ 1-propanol, which depends on the
number of hydroxyl groups in the molecule, as well as the num-
ber of primary hydroxyl groups. The higher concentration of the
hydroxyl groups in one molecule weakens the C–O bond strengths,
leading to higher dehydration rates.
2
comparison, the C–O bonds of C3 alcohols are cleaved by dehy-
dration reaction in the present work, while the C–C bonds of C3
alcohols with terminal hydroxyl groups are cleaved by sequential
dehydrogenation to aldehyde, followed by either disproportiona-
tion (Tishchenko or Cannizzaro type reactions) with a subsequent
decarboxylation reaction, or decarbonylation with a subsequent
water gas shift reaction (see Fig. 7).
The reaction rates of individual steps for C3 alcohols conversion
in aqueous phase are compiled in Table 2. For 2-propanol and 1,2-
propanediol conversion, dehydrogenation is the major step (initial
−
1
−1
rates: 48.0 and 8.5 mmol s molPt-surf atom , respectively), while
dehydration is the minor reaction for mono-alcohols conversion
−
1
−1
with slow dehydration rates (i.e., 0.4 and 0.3 mmol s molacid site
for 2-propanol and 1-propanol, respectively), which is mainly
attributed to the blocking of Lewis acid sites by water.
Dehydration, however, is the dominating primary reaction for
−
−1
1
−1
both 1,3-propanediol (initial rate: 11.5 mmol s molacid site ) and
Acknowledgements
−1
glycerol (initial rate: 4.6 mmol s molacid site ), with the dehy-
dration rates being 10–40 times faster than that of mono-alcohols.
We attribute this to the distinctly weaker C–O bond strengths
with the increase of hydroxyl group number in alcohol molecule,
e.g., 398 kJ/mol for 2-propanol and 333 kJ/mol for glycerol (see
Table 3), which supposedly should lower the activation energies
for 1,3-propanediol and glycerol dehydration compared with
mono-alcohols. The C–C bond cleavage through decarbonylation
and decarboxylation occurs at the C3 alcohols with terminal
This work was partly supported by the European Union in
the framework of the Integrated Project TOPCOMBI (NMP2-CT-
2
005-515792-2). C.Z. acknowledges support from the Technische
Universität München in the framework of the European Gradu-
ate School for Sustainable Energy. GAF acknowledges CONACYT
(Mexico) and the Alexander von Humboldt Stiftung (Germany) for
support and Prof. Dr. J.A. Lercher for the hospitality during a sabbati-
cal stay in Munich. IMC acknowledges the Alexander von Humboldt
Stiftung for a Study Fellowship for Junior Scientists.
hydroxyl groups and produces smaller alkanes and CO , attaining
the comparable rates of ca. 1.0 mmol s molPt-surf atom . The
overall reactivity decrease in the sequence of 1,3-propanediol
2
−1
−1
−
6
−1
−1
−6
−1
References
(
g
2.3 × 10 mol s gcatalyst ) ≈ glycerol
(1.2 × 10 mol s
−
1
−7
−1
−1
) > 1,2-propanediol
(5.1 × 10 mol s
g
) ≈ 1-
catalyst
catalyst
[
1] G.W. Huber, J.N. Chheda, C.J. Barrett, J.A. Dumesic, Science 308 (2005) 1446.
−
7
−1
−1
propanol (2.2 × 10 mol s
. Conclusions
The catalytic conversion of 2-propanol at the gas–solid and
g
).
catalyst
[2] J.P. Lange, R. Price, P.M. Ayoub, J. Louis, L. Petrus, L. Clarke, H. Gosselink, Angew.
Chem. Int. Ed. 49 (2010) 4479.
[
3] C. Zhao, Y. Kou, A.A. Lemonidou, X. Li, J.A. Lercher, Angew. Chem. Int. Ed. 48
(2009) 3987.
4
[4] C. Zhao, J. He, A.A. Lemonidou, X. Li, J.A. Lercher, J. Catal. 280 (2011) 8.
[5] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044.
[6] S. Czernik, R. French, C. Feik, E. Chornet, Ind. Eng. Chem. Res. 41 (2002) 4209.
aqueous–solid interfaces with Pt/Al O3 shows that the rate of 2-
[7] S. Adhikari, S. Fernando, S.D. Filio, M. Bricka, P.H. Steele, A. Haryanto, Energy
Fuels 22 (2008) 1220.
2
propanol dehydration in the aqueous phase was two orders of
magnitude slower than that in the gas phase. Because the cat-
alytic dehydration in the gas phase is generally accepted to be
[
[
8] G.W. Huber, A. Corma, Angew. Chem. Int. Ed. 46 (2007) 7184.
9] R.D. Cortright, R.R. Davda, J.A. Dumesic, Nature 418 (2002) 964.
[10] G.W. Huber, J.W. Shabaker, J.A. Dumesic, Science 300 (2003) 2075.
[11] G.W. Huber, J.A. Dumesic, Catal. Today 111 (2006) 119.
12] T. Werpy, G. Petersen, Top Value Added Chemicals From Biomass, US DOE
Report, 2004.
−
initiated by OH abstraction from the alcohol, we conclude that
[
this step is drastically retarded when the reaction is conducted
in water. As ␥-Al O3 slowly transforms into aluminum hydrox-
[13] S. Wang, H. Liu, Catal. Lett. 117 (2007) 62.
2
ide (Böhmite) under hydrothermal conditions, the weaker Lewis
acidity due to rehydroxylation is possibly responsible for the rate
decrease to some extent. However, the substantial decrease of 2-
propanol dehydration rate even at initial conversion, when alumina
is still not measurably converted, suggests that blocking of Lewis
acid sites by the abundant water is the main reason leading to the
drastic decrease of dehydration rate.
[14] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Appl. Catal. A 329 (2007) 30.
[
15] J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion, C. Pinel, C. Rosier, Green
Chem. 6 (2004) 359.
[
16] D.G. Lahr, B.H. Shanks, Ind. Eng. Chem. Res. 42 (2003) 5467.
[17] D.G. Lahr, B.H. Shanks, J. Catal. 232 (2005) 386.
[
18] R.R. Davda, J.W. Shabaker, G.W. Huber, R.D. Cortright, J.A. Dumesic, Appl. Catal.
B: Environ. 43 (2003) 13.
[
19] J.W. Shabaker, G.W. Huber, R.R. Davda, R.D. Cortright, J.A. Dumesic, Catal. Lett.
88 (2003) 1.

B. Peng et al. / Catalysis Today 183 (2012) 3–9
9
[
[
[
20] E.P. Maris, R.J. Davis, J. Catal. 249 (2007) 328.
21] E.P. Maris, W.C. Ketchie, M. Murayama, R.J. Davis, J. Catal. 251 (2007) 281.
22] T. Miyazawa, Y. Kusunoki, K. Kunimori, K. Tomishige, J. Catal. 240 (2006)
[28] P. Raybaud, M. Digne, R. Iftimie, W. Wellens, P. Euzen, H. Toulhoat, J. Catal. 201
(2001) 236.
[29] E. Laurent, B. Delmon, J. Catal. 146 (1994) 281.
2
13.
23] A. Wawrzetz, B. Peng, A. Hrabar, A. Jentys, A.A. Lemonidou, J.A. Lercher, J. Catal.
69 (2010) 411.
[30] H. Noller, W. Kladnig, Catal. Rev. Sci. Eng. 13 (1976) 149.
[31] Y.R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, Boca
Raton, FL, 2007.
[
2
[
[
[
24] Z. Duan, R. Sun, Chem. Geol. 193 (2003) 257.
25] A. Gervasini, A. Aroux, J. Catal. 131 (1991) 190.
26] M.A. Aramendia, V. Borau, C. Jimenez, J.M. Marinas, A. Porras, F.J. Urbano, React.
Kinet. Catal. Lett. 53 (1994) 397.
[32] J. Cook, P.M. Mailis, J. Chem. Soc. Chem. Commun. 17 (1981) 924.
[33] D. Coster, J.J. Fripiat, M. Muscas, A. Auroux, Langmuir 11 (1995) 2615.
[34] M. Trueba, S.P. Trasatti, Eur. J. Inorg. Chem. 17 (2005) 3393.
[35] W. Daniell, U. Schubert, R. Glöckler, A. Meyer, K. Noweck, H. Knözinger, Appl.
Catal. A 196 (2000) 247.
[27] H.A. Pray, C.E. Schweickert, B.H. Minnich, Ind. Eng. Chem. 44 (1952) 1146.