Aqueous-phase hydrodeoxygenation of sorbitol with Pt/SiO2-Al2O3: Identification of reaction intermediates

Article abstract of DOI:10.1016/j.jcat.2009.12.006

Aqueous-phase hydrodeoxygenation of sugar and sugar-derived molecules can be used to produce a range of alkanes and oxygenates. In this paper, we have identified the reaction intermediates and reaction chemistry for the aqueous-phase hydrodeoxygenation of sorbitol over a bifunctional catalyst (Pt/SiO₂-Al₂O₃) that contains both metal (Pt) and acid (SiO₂-Al₂O₃) sites. A wide variety of reactions occur in this process including C{single bond}C bond cleavage, C{single bond}O bond cleavage, and hydrogenation reactions. The key C{single bond}C bond cleavage reactions include: retro-aldol condensation and decarbonylation, which both occur on metal catalytic sites. Dehydration is the key C{single bond}O bond cleavage reaction and occurs on acid catalytic sites. Sorbitol initially undergoes dehydration and ring closure to produce cyclic C6 molecules or retro-aldol condensation reactions to produce primarily C3 polyols. Isosorbide is the major final product from sorbitol dehydration. Isosorbide then undergoes ring opening hydrogenation reactions and a dehydration/hydrogenation step to form 1,2,6-hexanetriol. The hexanetriol is then converted into hexanol and hexane by dehydration/hydrogenation. Smaller oxygenates are produced by C{single bond}C bond cleavage. These smaller oxygenates undergo dehydration/hydrogenation reactions to produce alkanes from C1-C5. The results from this paper suggest that hydrodeoxygenation chemistry can be tuned to make a wide variety of products from biomass-derived oxygenates.

Full text of DOI:10.1016/j.jcat.2009.12.006

Journal of Catalysis 270 (2010) 48–59
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Aqueous-phase hydrodeoxygenation of sorbitol with Pt/SiO2–Al2O3: Identification
of reaction intermediates
*
Ning Li, George W. Huber
Department of Chemical Engineering, University of Massachusetts, 159 Goessmann Lab., 686 North Pleasant Street, Amherst, MA 01003-9303, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Aqueous-phase hydrodeoxygenation of sugar and sugar-derived molecules can be used to produce a
range of alkanes and oxygenates. In this paper, we have identified the reaction intermediates and reaction
chemistry for the aqueous-phase hydrodeoxygenation of sorbitol over a bifunctional catalyst (Pt/SiO2–
Al2O3) that contains both metal (Pt) and acid (SiO2–Al2O3) sites. A wide variety of reactions occur in this
process including CAC bond cleavage, CAO bond cleavage, and hydrogenation reactions. The key CAC
bond cleavage reactions include: retro-aldol condensation and decarbonylation, which both occur on
metal catalytic sites. Dehydration is the key CAO bond cleavage reaction and occurs on acid catalytic
sites. Sorbitol initially undergoes dehydration and ring closure to produce cyclic C6 molecules or retro-
aldol condensation reactions to produce primarily C3 polyols. Isosorbide is the major final product from
sorbitol dehydration. Isosorbide then undergoes ring opening hydrogenation reactions and a dehydra-
tion/hydrogenation step to form 1,2,6-hexanetriol. The hexanetriol is then converted into hexanol and
hexane by dehydration/hydrogenation. Smaller oxygenates are produced by CAC bond cleavage. These
smaller oxygenates undergo dehydration/hydrogenation reactions to produce alkanes from C1–C5. The
results from this paper suggest that hydrodeoxygenation chemistry can be tuned to make a wide variety
of products from biomass-derived oxygenates.
Received 1 September 2009
Revised 4 December 2009
Accepted 5 December 2009
Available online 8 January 2010
Keywords:
Alkane production
Biofuels
Biomass conversion
Bifunctional catalysis
Pt
SiO2–Al2O3
Ó 2009 Elsevier Inc. All rights reserved.
1. Introduction
the oxygen-deficient molecules by hydrogenation on metal cata-
lytic sites. A wide variety of products can be generated by hydro-
Declining fossil fuel reserves combined with increasing fossil
fuel prices are making it imperative to develop new economical
and energy-efficient process for the production of fuel from renew-
able resources. Plant biomass is the only renewable source of car-
bon that can be used to make liquid fuels and chemicals [1].
Selectively removing oxygen from the biomass-derived species is
one of the key challenges with converting renewable biomass re-
sources into fuels and chemicals. Aqueous phase processing
(APP) [2,3] is a promising conversion option for the conversion of
aqueous biomass-derived feedstocks (including sugars, sugar alco-
hols and bio-oils) into hydrogen [4–10], light alkanes (even gaso-
line) [8,11,12], liquid alkanes [13–18], and oxygenates [3,19–24].
Aqueous-phase hydrodeoxygenation (APHDO), a key process in
APP, involves the conversion of oxygenated molecules into alkanes
by a series of dehydration and hydrogenation step [11,14,15]. In
APHDO process, the bifunctional catalysts containing metal (Pt
and Pd) and acid (such as SiO2–Al2O3 or mineral acid) sites were
used. The oxygen is removed from the biomass with APHDO by a
combination of CAC and CAO bond cleavage reaction. Once the
oxygen is removed from the biomass, hydrogen is then added to
deoxygenation reaction pathways. The products generated
depend on the relative rates of CAC vs. CAO bond cleavage. If lar-
ger alkanes or oxygenates (i.e. C5–C6) are desired, then CAC bond
cleavage should be inhibited and the rate of CAO bond cleavage
should increase. If smaller alkanes and oxygenates are desired,
then relative rate of CAC bond cleavage vs. CAO bond cleavage
should be increased. While Dumesic and co-workers [11] have pre-
viously shown that APHDO to convert sorbitol into alkanes ranging
from C1 to C6 alkanes, little is known about the reaction chemistry
that occurs in this process. The reaction intermediates produced in
this process are unknown. The exact nature of the CAC vs. CAO
bond cleavage is also unknown.
The purpose of this paper is to understand the reaction chemis-
try and identify the key reaction intermediates that are produced
in APHDO of sorbitol. In this paper, we have identified the key reac-
tion intermediates and discussed how these intermediates could
be selectively produced by adjusting the relative rates of CAC vs.
CAO bond cleavage. We showed that the product selectivity can
be adjusted by changing the relative rates of CAC vs. CAO bond
cleavage. As we will show in this paper, CAC bond cleavage occurs
over metal active sites by two major pathways retro-aldol conden-
sation and decarbonylation. CAO bond cleavage happens on acid
sites by dehydration reactions
* Corresponding author. Fax: +1 413 545 1647.
E-mail address: huber@ecs.umass.edu (G.W. Huber).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.12.006

N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
49
Both the injection port and the detector were held at 513 K. For
each analysis, 1 L liquid sample was injected. The column was
2. Experimental
l
held at 308 K for 5 min, ramped to 323 K at 5 K minꢀ1, then to
513 K at 20 K minꢀ1 and kept at 513 K for 7.5 min. A Shimadzu
HPLC with UV–Vis (SPD-20AV) and RID (RID-10A) detectors was
also used to analysis the liquid products. One microliter liquid
sample was injected for each sample. The separation of the prod-
ucts was realized by a BIO-RAD Aminex HPX-87H column (Catalog
no. 125-0140) maintained at 303 K with 0.005 M H2SO4 as mobile
phase flowing at a rate of 0.6 mL minꢀ1. Finally, a total organic car-
bon (TOC) analyzer (Shimadzu TOC-5000A) was used to check the
carbon balance. No major catalyst deactivation was observed dur-
ing the reactions performed in this study. We tested every reaction
in a continuous run of 24 h and did not observe catalyst deactiva-
tion. The data were collected after at least 6-h time on stream, and
multiple points were collected for each reaction condition.
2.1. Catalyst preparation
The Pt/SiO2–Al2O3 catalyst used in this work was prepared by
an incipient wetness impregnation of SiO2–Al2O3 (SIAL3125, Grace
Davison, SiO2/Al2O3 molar ratio about 4.0) with tetra-amine plati-
num nitrate (Strem Chemicals) aqueous solution according to liter-
ature [11,15]. The Pt content in the catalyst was 4 weight% (wt.%).
The mixture was then dried in an oven overnight at 373 K and cal-
cinated at 533 K for 3 h in air.
2.2. Catalyst characterization
XRD patterns of the 4 wt.% Pt/SiO2–Al2O3 as prepared were ob-
tained with a Philips XPERT powder diffractometer equipped with
an on-line computer. The Ni-filtered Cu Ka radiation was used. The
specific surface area of the catalyst was determined by nitrogen
adsorption at 77 K using a Quantchrome Autosorb Automated
Gas Sorption System. Before each experiment, the samples were
evacuated at 573 K for 24 h. The number of available Pt active sites
in the Pt/SiO2–Al2O3 catalyst was measured with same equipment
by the irreversible H2 adsorption at 303 K. The catalyst was re-
duced in situ at 723 K for 2 h with H2 flow before the chemisorption
measurement.
3. Results
3.1. Characterization of the Pt/SiO2–Al2O3 catalyst
The XRD pattern of the 4 wt.% Pt/SiO2–Al2O3 (omitted here) has
only a wide peak at 2h range from 15° to 30°, no peak of Pt or its
oxide can be observed, which means that the Pt species is well dis-
persed on the support (less than 5 nm). The catalysts had a hydro-
gen uptake of 74:25
Pt ratio of 0.724 (i.e. 72.4% dispersion). This further affirmed that
l
mol ꢁ H2 ꢁ gꢀca1talyst which corresponds to a H to
2.3. Catalytic activity
the Pt was highly dispersed. The specific BET surface area of the
Pt/SiO2–Al2O3 catalysts was 430 m2 gꢀ1
.
The hydrodeoxygenation reaction was carried out in a stainless
steel tubular flow reactor heated by a Lindberg (type 54032) fur-
nace. In order to ensure a uniform temperature profile along the
catalyst bed, an aluminum filler was inserted in the void between
the heating tube of the furnace and tubular reactor. The tempera-
ture difference between the middle and top (or end) of the tubular
reactor was less than 15 K. Before the reaction, the Pt/SiO2–Al2O3
(the particle size distribution about 100 mesh) was reduced in
the reactor with hydrogen flowing from the bottom at about
200 mL minꢀ1. For each activity test, 3.3 g Pt/SiO2–Al2O3 catalyst
was used. The temperature regime used for reduction of Pt/SiO2–
Al2O3 catalyst was: room temperature to 723 K at 50 K hꢀ1, then
hold at 723 K for 2 h. After cooling down the temperature to
493 K and slowly increasing the pressure to 2.93 MPa, the liquid
feed (5 wt.% sorbitol aqueous solution) was cofed with the hydro-
gen (45 mL minꢀ1) to the catalyst from reactor bottom with the
help of a JASCO PU980 HPLC pump. A gas–liquid separator was em-
ployed after the reactor tube. The gaseous products from the reac-
tor flow through a back pressure regulator, used to maintain the
pressure of the reaction system. The gaseous products were further
analyzed by two online gas chromatographs (HP 5890 series II).
CO2 in the gaseous product were analyzed by a Thermal Conductiv-
ity Detector (TCD). Alltech HAYESEP DB 100/120 packed column
(Part no. 2836PC) was used with the oven temperature held con-
stant at 348 K. The TCD and the injection port were held at 433 K
and 393 K respectively. The column flow rate was 41 mL minꢀ1
with helium carrier gas. Alkanes in the gaseous product were ana-
lyzed on a flam ion detector (FID) with Alltech AT-Q capillary col-
umn (Part no. 13950). Helium was used as the carrier gas with the
column flow rate of 1 mL minꢀ1. Both the injection port and the
detector were held at 473 K. Following GC oven temperature re-
gime was used: hold at 313 K for 6 min, ramp to 453 K at 5 K minꢀ1
and hold at 453 K for 25 min. Liquid product accumulated in the
gas–liquid separator was drained then. Liquid product was drained
periodically and analyzed by GC–MS and HPLC. The GC–MS is Shi-
madzu GC-2010 with an Rtx-VMS capillary column. Helium was
3.2. Catalytic activity
3.2.1. Catalytic performance of 4 wt.% Pt/SiO2–Al2O3 with sorbitol
We first investigated the APHDO of a 5 wt.% sorbitol aqueous
solution at 518 K, 2.93 MPa at weight hour space velocities (WHSV)
from 0.73 hꢀ1 to 11.64 hꢀ1. As shown in Tables 1–4 and Figs. 1–3, a
wide variety of products are formed in this reaction. The gas prod-
uct was mainly composed of CO2, C1–C6 straight chain alkanes,
unreacted H2 and trace amount of furans (furan, 2-methylfuran,
2,5-dimethylfuran) at high space velocities. No CO peak was ob-
served in the chromatogram which means that the CO concentra-
tion is very low. The liquid phase product includes alcohol, diol,
polyols, ketones, cyclic ether, isosorbide and unconverted sorbitol.
We have been able to identify over 60% of the carbon products that
are in the liquid phase and between 70% and 100% of the total car-
bon products in this study. We are unable to identify all of the car-
bon because of the following two reasons: (1) some of the products
are present as minor components with concentrations below
detection limits. (2) some non-volatile oligomers may be formed
during the reaction. These oligomers are not detected by our cur-
rent GC–MS and HPLC. Future work in analytical chemistry will
help in identifying more of the carbon in the liquid phase.
As shown in Table 1, the space velocity strongly influences the
amount of carbon presenting in the liquid and gaseous phase prod-
ucts. Fig. 1 shows the conversion of sorbitol and carbon distribu-
tion in the gas and liquid phase as a function of space velocity.
The sorbitol conversion increases almost linearly from 8.1% to
100% as the space velocity decreases from 11.64 hꢀ1 to 2.91 hꢀ1
.
At space velocities higher than 3 hꢀ1, over 70% of the carbon is still
in the liquid phase. As the space velocity decreases further, more
and more liquid products are converted into gas phase products.
Fig. 2 shows the distribution of products in the gas and liquid
phases. At high space velocities, over 80% of the gas is CO2. The
CO2 is most likely produced from the decarbonylation of aldehydic
species. Decarbonylation produces CO which is converted into CO2
used as the carrier gas with the column flow rate of 1.57 mL minꢀ1
.

50
N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
Table 1
Conversion to gas and liquid products as a function of WHSV for sorbitol conversion over 4 wt.% Pt/SiO2–Al2O3 catalyst. (Reaction conditions: 5 wt.% sorbitol aqueous phase
solution, 518 K, 2.93 MPa and flow rate of H2 about 45 mL minꢀ1.)
WHSV of
CH2 at inlet
Csorbitol at inlet
Carbon in gas phase
detected by GC (%)
Carbon in liquid phase
detected by TOC analysis (%)
Carbon
balance (%)
% of liquid products and reactants
identified by GC–MS and HPLC (%)
a
b
c
liquid (hꢀ1
)
(mol Lꢀ1
)
(mmol Lꢀ1
)
0.73
1.45
2.91
5.82
11.64
0.668
0.660
0.643
0.617
0.562
3.99
7.46
14.3
24.7
47.2
82.1
61.0
23.9
9.2
18.2
36.1
75.1
92.5
100.0
100.3
97.1
99.0
101.7
101.8
74.2
90.3
67.1
58.0
104.7
1.8
a
Weight hourly space velocity (WHSV in units of hꢀ1) = (mass flow rate of feed solution in the unit of hꢀ1)/(the mass of catalyst).
b
CH2 at inlet (mol Lꢀ1) = (Mol of H2 fed into the reactor)/(sum of the volume of gas and liquid at the reaction condition). Assumes the liquid volume does not change with
temperature.
c
Csorbitol at inlet (mmol Lꢀ1) = (Mol of sorbitol fed into the reactor) ꢂ 1000/(sum of the volume of gas and liquid at the reaction condition). Assumes the liquid volume does
not change with temperature.
Table 2
Selectivity of C6 compounds in the liquid product as the function of WHSV. Reaction conditions: 5 wt.% sorbitol aqueous phase solution, 518 K, 2.93 MPa and flowrate of H2 about
45 mL minꢀ1 with inlet concentrations listed in Table 1.
WHSV Yield of
Sorbitol left in
the solution (%)
Selectivity (%)
(hꢀ1
)
C6 (%)a
Gas
phase
Liquid
phase
Hexanolb Hexanonec 2-Methyl-
2,5-Dimethyl-
tetra-hydro-
furan
1,2-
Tetra-hydro-
1,2,6-
Hexanetriol
Isosorbide
tetra-hydro-
pyran
Hexanediol pyran-2-
methanol
0.73
1.45
2.91
5.82
11.64
17.6
8.9
1.73
0.59
0.027
0.5
7.9
20.1
16.7
13.9
0
0
0
19.7
81.0
29.8
57.2
9.0
2.0
0
0
70.2
5.6
2.5
0.44
0
0
10.4
10.2
0
0
0
15.6
10.3
0
0
0
0
0
18.5
25.8
8.7
0
3.0
10.0
2.5
0
5.3
26.1
15.4
0
0.8
60.5
100
0
a
b
c
Yield of C6 (%) = (the sum of carbon in C6 compound in gas or liquid products)/(the carbon in the feed solution) ꢂ 100%.
The hexanol includes 1-hexanol, 2-hexanol, and 3-hexanol. The amount of 1-hexanol and 2-hexanol are comparable. Both of them are much higher than that of 3-hexanol.
The hexanone include 2-hexanone and 3-hexanone. The concentration of 2-hexanone and 3-hexanone are very similar.
Table 3
Selectivity of C1–C3 compounds in the liquid product as the function of WHSV. Reaction conditions: 5 wt.% sorbitol aqueous phase solution, 518 K, 2.93 MPa and flowrate of H2
about 45 mL minꢀ1 with inlet concentrations listed in Table 1.
WHSV (hꢀ1
)
Yield of C1–C3 (%)a
Selectivity (%)
Gas phase
Liquid phase
Methanol
Ethanol
Propanolb
Acetone
Ethylene glycol
1,2-Propanediol
2-Hydroxy-acetone
Glycerol
0.73
1.45
2.91
5.82
11.64
45.3
39.2
18.5
8.1
11.2
10.7
11.5
9.2
13.9
10.5
3.4
3.2
1.6
32.3
35.8
14.2
10.1
4.8
53.7
51.5
16.2
6.8
0
0
0
3.7
5.2
16.4
0
0
56.4
47.6
38.2
0
0
2.6
11.6
13.5
0
0
0
15.4
25.4
2.2
3.3
0
1.7
4.7
0
0
a
Yield of C1–C3 (%) = (the sum of carbon in C1–C3 compound in gas or liquid products)/(the carbon in the feed solution) ꢂ 100%.
b
The propanol include 1-propanol and 2-propanol. Then concentration of 2-propanol is more than five times of that of 1-propanol.
and H2 by the water–gas shift reaction. It has previously been
shown that the water–gas shift reaction is in near equilibrium un-
der APHDO conditions [25]. As the space velocity decreases, the gas
phase C2–C6 selectivity increases, while the C1 product selectivity
(which is primarily CO2) decreases. The liquid phase carbon selec-
tivities are shown in Fig. 2b. The amount of C6 products in the li-
quid phase increases with space velocity. The C6 product yield
reaches the maximum at space velocity of 2.91 hꢀ1 then decreases
as the space velocity decreases. The C6 products are shown in Table
2. The C6 products that are formed include isosorbide, 1,2,6-hex-
anetriol, tetrahydropyran-2-methanol, 1,2-hexanediol, 2,5-
dimethyltetrahydrofuran, 2-methyl-tetrahydropyran, hexanone,
and hexanol. We observed 100% isosorbide selectivity at a space
velocity of 11.64 hꢀ1. The isosorbide selectivity decreases from
100% to 60.5% as the space velocity decreases from 11.64 hꢀ1 to
5.82 hꢀ1. As the WHSV decreases further to 2.91 hꢀ1, the C6 selec-
tivity of isosorbide then decreases to 0.8%. 1,2,6-Hexanetriol is ob-
from 5.82 hꢀ1 to 2.91 hꢀ1. Then the selectivity of 1,2,6-hexanetriol
decreases as the WHSV decreases below 2.91 hꢀ1. At WHSV of
2.91 hꢀ1 and below, the major C6 liquid products include a combi-
nation of hexanediols, hexanols and some saturated cyclic ether
compounds (2-methyl-tetrahydropyran, 2,5-dimethyltetrahydro-
furan, tetrahydropyran-2-methanol).
As shown in Fig. 2b and Table 2, at a space velocity of 11.64 hꢀ1
,
over 95% of the liquid products are C6 compounds. The other 5% of
the products are primarily C3 compounds which are primarily 1,2-
propanediol (38% C1-3 selectivity), glycerol (25.4% C1–C3 selectiv-
ity), and hydroxyacetone (13.5% C1–C3 selectivity) as shown in Ta-
ble 3. These products are produced primarily from retro-aldol
condensation. As the space velocity decreased further, the amount
of liquid phase C1–C3 products increases. However, the distribu-
tion of C1–C3 products (reported in Table 3) is primarily the alco-
hols. At a higher space velocities, the majority of the products are
C2 and C3 polyols including 1,2-propanediol, glycerol, hydroxyac-
served as
a
major product. The 1,2,6-hexanetriol C6 liquid
etone, and ethylene glycol. As the WHSV decreases below 2.91 hꢀ1
the selectivity to these products decreases and the C1–C3 products
,
selectivity increases from 15.4% to 26.1% as the WHSV decreases

N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
51
Table 4
Selectivity of C4–C5 compounds in the liquid product as the function of WHSV. Reaction conditions: 5 wt.% sorbitol aqueous phase solution, 518 K, 2.93 MPa and flowrate of H2
about 45 mL minꢀ1 with inlet concentrations listed in Table 1.
WHSV
(hꢀ1
Yield to C4–C5
(%)a
Selectivity (%)
)
Gas
phase
Liquid
phase
Butanolb Pentanolc Butanone 2-
Tetra-
2-Methyl-
tetrahydro-furan
Butanediold 1,2-
Pentanediol furfural alcohol
Tetra-hydro-
Pentanone hydro-
pyran
0.73
1.45
2.91
5.82
11.64
19.2
12.9
3.7
0.54
0.075
1.7
14.0
19.0
8.8
37.8
14.9
3.6
0
26.2
32.1
7.3
4.1
0
0
30.8
32.3
0
0
0
3.9
4.2
0
36.0
5.9
2.4
2.7
0
0
0
2.4
20.5
36.2
74.4
0
2.9
18.4
46.9
25.6
0
8.9
8.7
0
2.1
2.9
5.9
0
0.4
0
0
0
a
b
c
Yield of C4–C5 (%) = (the sum of carbon in C4–C5 compound in gas or liquid products)/(the carbon in the feed solution) ꢂ 100%.
The butanol include 1-butanol and 2-butanol. The concentration of 1-butanol is much higher than that of 2-butanol, indicating that most of butanol is 1-butanol.
The pentanol include 1-pentanol and 2-pentanol. The amount of 1-pentanol is more than eight times of that for 2-pentanol.
The butanediol include 1,2-butanediol and 2,3-butanediol. More than 60% of butanediol detected are 1,2-butanediol.
d
30
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
(a)
25
20
15
10
5
0
0
2
4
6
8
10
12
0
2
4
6
8
10
12
-1
WHSV (h-1)
WHSV ( h )
Fig. 1. Sorbitol conversion (N), gas phase carbon (j) and liquid phase carbon (d) as
a function of weight hour space velocity (WHSV). Reaction conditions: 5 wt.%
sorbitol aqueous phase solution, 518 K, 2.93 MPa and flowrate of H2 about
45 mL minꢀ1 with inlet concentrations listed in Table 1.
(b)
30
20
10
0
are mainly alcohols including methanol, ethanol and propanol. The
yield of C1–C3 products increases with decreasing WHSV. This data
suggest that the C2 and C3 polyols are produced from the sorbitol
by retro-aldol condensation.
A wide variety of C4–C5 oxygenates are formed as shown in Ta-
ble 4. The yield of C4–C5 oxygenates increases from 0.4% to 38.7%
as the space velocity decreases from 11.64 hꢀ1 to 1.45 hꢀ1. The
yield of C4–C5 oxygenates decreases as the space velocity de-
creases from 1.45 hꢀ1 to 0.73 hꢀ1. At high space velocities (above
5.82 hꢀ1), the C4–C5 products are 1,2-petanediol, 1,2-butanediol
and 2,3-butanediol. Importantly, no C4–C5 oxygenates were iden-
tified that have more than two oxygenates. This indicated that C4
and C5 molecules are only formed after a large majority of the oxy-
gen is removed from the biomass feedstock. At space velocities
lower than 2.91 hꢀ1, the major C4 and C5 products are alcohols
(1-butanol, 2-butanol, 1-pentanol, 2-pentanol) and ketones (buta-
0
2
4
6
8
10
12
-1
WHSV (h
)
Fig. 2. Gas phase (a) and liquid phase (b) carbon selectivity of C1 (j), C2 (d), C3 (N),
C4 (.), C5 (4), C6 (5) and CO2 (h) as the function of WHSV. Reaction conditions:
5 wt.% sorbitol aqueous phase solution, 518 K, 2.93 MPa and flowrate of H2 about
45 mL minꢀ1 with inlet concentrations listed in Table 1.
none, 2-pentenone).
methyl-tetrahydrofuran, tetrahydrofurfural alcohol are also
formed in this process.
Fig. 3 shows the overall selectivity of C1–C6 carbon distribution
which was calculated by taking into consideration of carbon exists
in both the liquid and the gas phase. At high space velocity, the
majority of the carbon is present as C6 products. The C6 product
selectivity decreases from 94% to 36% as the space velocity
A
small amount of tetrahydropyran, 2-
decreases from 11.64 hꢀ1 to 5.82 hꢀ1. The C6 selectivity then
decreases from 36% to 22% as the space velocity decreases from
5.82 hꢀ1 to 2.91 hꢀ1. As the space velocity decreases further, the
C6 product selectivity does not change significantly. This indicates
that the majority of the CAC bond cleavage for the C6 products
occurs very early on. The overall product selectivity of C1 com-
pounds increases as the space velocity decreases from 11.64 hꢀ1
to 0.73 hꢀ1
.

52
N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
3.2.2. Dehydration of sorbitol over pure SiO2–Al2O3 support
100
80
60
40
20
0
We also investigated the dehydration of 5 wt.% sorbitol solution
over the SiO2–Al2O3 support at the same temperature under inert
atmosphere (Helium) as shown in Fig. 4. The work was done with
a batch reactor. Only C6 compounds (1,4-sorbitan, isosorbide and
trace amount of 1,5-sorbitan) were detected, which means that
no CAC bond cleavage took place during the dehydration step. In
other words, the metal catalytic sites are necessary for CAC bond
cleavage. Fig. 5 illustrates the pathways for sorbitol dehydration.
We can see that the dehydration of sorbitol to isosorbide is a
two-step reaction: in the first step, the sorbitol is intermolecular
dehydrated to 1,4-sorbitan. Subsequently, the 1,4-sorbitan is fur-
ther dehydrated to isosorbide.
30
20
10
0
3.2.3. Hydrodeoxygenation of model compounds over Pt/SiO2–Al2O3
catalyst
0
2
4
6
8
10
12
WHSV (h-1)
We studied the APHDO of a series of model compounds includ-
ing isosorbide, 1,2,6-hexanetriol, 1,2-hexanediol, 2,3-butanediol,
1-butanol and 2-butanol. The concentrations of all model compounds
were fixed at 5 wt.% with a reaction temperature and pressure of
518 K and 2.93 MPa.
Fig. 3. Overall carbon selectivity of C1 (j), C2 (d), C3 (N), C4 (.), C5 (4), C6 (5) as
a function of WHSV. Reaction conditions: 5 wt.% sorbitol aqueous phase solution,
518 K, 2.93 MPa and flowrate of H2 about 45 mL minꢀ1 with inlet concentrations
listed in Table 1.
3.2.3.1. Hydrodeoxygenation of isosorbide and 1,2,6-hexanetriol. To
further understand the conversion of sorbitol after dehydration,
we also investigated the APHDO of isosorbide and 1,2,6-hexane-
triol, and the results are shown in Tables 5 and 6.
From Table 5, we observed that isosorbide has a significantly
higher gas phase alkane selectivity and hexane selectivity com-
100
80
60
40
20
0
100
80
60
40
20
0
pared to sorbitol or hexanetriol. For example, at
a
WHSV = 0.73 hꢀ1, a 90% alkane selectivity and 65% hexane selectiv-
ity can be obtained for isosorbide. In comparison, under the same
reaction conditions with sorbitol as the feed, the alkane and hex-
ane selectivity are about 70% and 30% respectively. In contrast to
sorbitol, the gas phase alkane selectivity for isosorbide is not a
strong function of conversion. With the isosorbide as feed stock,
when WHSV increased from 0.73 hꢀ1 to 2.91 hꢀ1, the carbon selec-
tivity to CO2 did not change, but the selectivity to hexane even in-
creased a little bit from 64.9% to 80.5%, which can be explained by
the retraining of the cyclic opening reaction over the acid sites by
decreasing the contacting time between isosorbide and catalyst.
From Table 6, we can see that all the intermediates identified from
the liquid product of APHDO of isosorbide are similar to the prod-
ucts we detected with sorbitol as feed stock, which indicates that
isosorbide is a key intermediate for the APHDO of sorbitol over
Pt/SiO2–Al2O3.
0
5
10
15
20
25
Reaction time (h)
Fig. 4. Conversion of sorbitol (d), the yield of 1,4-sorbitan (.) and isosorbide (N)
over pure SiO2–Al2O3 support as the function of reaction time. Reaction condition:
90 mL 5 wt.% sorbitol aqueous solution, 0.9 g SiO2–Al2O3 catalyst, 518 K, 2.93 MPa.
Fig. 5. Major reaction pathways for dehydration of 5 wt.% sorbitol aqueous solution over SiO2–Al2O3 catalyst at 518 K.

N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
53
Table 5
Activity and selectivity of 4 wt.% Pt/SiO2–Al2O3 catalyst for the aqueous-phase hydrodeoxygenation (APHDO) of isosorbide, 1,2,6-hexanetriol and sorbitol. Reaction conditions:
5 wt.% aqueous phase solution, 518 K, 2.93 MPa and flow rate of H2 about 45 mL minꢀ1
.
Feedstock
WHSV (hꢀ1
)
Carbon conversion to gas (%) from GCa Alkane selectivity (%)b CO2 selectivity (%) Specific alkane selectivity (%)c
C1
C2
3.4
C3
4.2
3.0
C4
7.0 19.3 64.9
4.2 9.0 80.5
C5
C6
Isosorbide
Isosorbide
1,2,6-hexanetriol 0.73
1,2,6-hexanetriol 1.45
1,2,6-hexanetriol 2.91
Sorbitol
Sorbitol
Sorbitol
0.73
2.91
68.9
12.5
66.2
44.8
16.7
82.1
61.0
23.9
89.1
89.2
70.2
66.3
51.4
68.8
59.7
37.7
10.9
10.8
29.8
33.7
48.6
31.2
40.3
62.3
1.2
0.7
2.6
0.20
0.12
0.18
5.5
8.7
9.0
0.04
0.03
0.04
14.9
17.2
16.2
3.02 38.2 44.2 14.3
3.08 36.8 46.0 14.1
3.61 40.4 40.1 15.8
0.73
1.45
2.91
12.6
16.9 19.5 30.4
17.5 18.0 24.5
21.4 19.2 19.1
14.1
15.0
a
b
c
Carbon conversion to gas (%) from GC = (sum of the carbon detected by GC)/(carbon in the feed solution) ꢂ 100%.
Alkane selectivity (%) = (sum of carbon in alkanes detected by GC)/(sum of carbon detected in gas phase) ꢂ 100%.
Specific alkane selectivity (%) = (carbon in specific alkanes)/(sum of carbon in alkanes detected by GC) ꢂ 100%.
Table 6
Liquid product selectivity (%)a for the APHDO of isosorbide and 1,2,6-hexanetriol over 4 wt.% Pt/SiO2–Al2O3 catalyst at different WHSV. Reaction conditions: 5 wt.% aqueous phase
solution, 518 K, 2.93 MPa and flow rate of H2 about 45 mL minꢀ1
.
Carbon number
Component
Isosorbide
0.73 hꢀ1
1,2,6-Hexanetriol
0.73 hꢀ1
2.91 hꢀ1
1.45 hꢀ1
2.91 hꢀ1
C1
C2
Methanol
Ethanol
0
4.3
0.1
0.5
0
0
0
0
0
0
C3
Propanolb
1,2-Propanediol
17.6
0
1.0
0.9
0
0
0
0
0
0
C4
Butanolc
Tetrahydrofuran
2,3-Butanediol
6.8
0.7
0
0.5
0.1
1.2
4.9
0
0
3.8
0
0
3.1
0
0
C5
Pentanold
Pentanal
2-Pentanone
Tetrahydropyran
2-Methyl-tetrahydrofuran
Tetra-hydrofurfural alcohol
1,2-Pentanediol
7.8
0
0
2.0
2.1
4.9
0
0.8
0
39.1
0
0
31.6
0
0
38.8
0
0
12.4
0
0
32.7
0.8
0
6.5
0
0
8.8
0
0.6
0.2
1.3
1.5
0
0
0
0
0
1-Hydroxy-2-pentanone
30.3
8.5
C6
Hexanole
20.0
0
2.5
2.3
0
1.5
0
1.5
0
4.9
5.2
0
0.6
0
0.4
4.1
0.6
67.1
9.1
0
5.9
0
4.6
4.9
0
15.7
0
4.5
0
3.4
20.4
0
17.6
0.7
3.3
0
1.7
23.9
0
Hexanonef
2-Methyl-tetrahydropyran
2,5-Dimethyltetrahydrofuran
Oxepane
Tetrahydropyran-2-methanol
1,2-Hexanediol
1,2,6-Hexanetriol
Isosorbide
0
0
0
0
0
0
a
b
c
d
e
f
The value in the table is the carbon selectivity (%) of specific compound in the liquid phase.
The propanol includes 1-propanol and 2-propanol. The concentration of 1-propanol is more than 40 times of 2-propanol.
The butanol includes 1-butanol and 2-butanol. For 1,2,6-hexanediol, the butanol detected is totally 1-butanol. While for isosorbide, the butanol is 2-butanol.
The pentanol includes 1-pentanol and 2-pentanol. The amount of 1-pentanol is more than 30 times of 2-pentanol.
The hexanol includes 1-hexanol, 2-hexanol and 3-hexanol. The concentration of 1-hexanol is more than twice of 2-hexanol and 3-hexanol.
The hexanone includes 2-hexanone and 3-hexanone. For 1,2,6-hexanetriol, only 2-hexanone was detected. For isosorbide, the concentration of 3-hexanone is more than
twice of 2-hexanone.
As shown in Table 5, when 1,2,6-hexanetriol was the feedstock,
propane, butane, pentane, hexane and CO2 are the main products.
The selectivity to methane and ethane are very low (less than 1%).
This result indicates the CAC bond cleavage for oxygenates mainly
occurs at the C with CAO bond.
The alkane selectivity for 1,2,6-hexanetriol is only slightly high-
er than sorbitol, the hexane selectivity is also lower for hexanetriol
than sorbitol. However, the butane and pentane selectivity is much
higher than either sorbitol or isosorbide. The gas phase carbon con-
version decreases with increasing space velocity. The CO2 selectiv-
ity increases slightly with increasing space velocity and the alkane
selectivity does not change significantly with space velocity. After
the analysis of liquid products (illustrated in Table 6), some cyclic
ether species (such as tetrahydropyran, tetrahydrofurfural alcohol,
tetrahydropyran-2-methanol, 2-methyl-tetrahydropyran, oxe-
pane) were detected. We also noticed from Table 6 that the selec-
tivities of 6-ring cyclic ethers (tetrahydropyran, 2-methyl-
tetrahydropyran, tetrahydropyran-2-methanol) are higher that of
7-ring cyclic ether (oxepane). The pathway for the APHDO of
1,2,6-hexanetriol is shown in Fig. 6.
3.2.3.2. Hydrodeoxygenation of 1,2-hexanediol and 2,3-butane-
diol. Subsequently, we also investigated the APHDO of diols
including 1,2-hexanediol and 2,3-butanediol. These compounds
were selected because their hydroxyl groups are in different posi-
tions: The 1,2-hexanediol represents the diols with OH locates
respectively at the edge and in the middle of molecule. The 2,3-

54
N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
Fig. 6. Major reaction pathways for hydrodeoxygenation of 1,2,6-hexanetriol at 518 K.
butanediol stands for the diols with both OH groups in the middle
of molecule.
relative stable cyclic ether can (partially) eliminate the number
of OH at the edge of 1,2,6-hexanetriol. As a direct result, the CAC
bond cleavage by the decarbonylation of aldehyde was inhibited.
From the liquid phase analysis (cf. Table 8), 1-pentanol, 1-hex-
anol, 2-hexanol, 2-hexanone were identified as intermediates.
Reaction routes were shown in Fig. 7. We can see that the APHDO
of 1,2-hexanediol produces 1-pentanol. This is then converted into
butane by dehydrogenation and decarbonylation. The main gas
phase products for 2,3-butanediol are butane, propane and CO2.
With the increasing of space velocity, the selectivity does not
change. After analysis the liquid phase products, 2-butanol,
As shown in Table 7, 1,2-hexanediol produced propane, butane,
hexane and CO2. The selectivity of propane, butane and CO2 in-
creased with the space velocity, which is due to the preferable
CAC bond cleavage at higher space velocity. It seems that the effect
of space velocity on the CAC cleavage of 1,2-hexanediol is more
significant than 1,2,6-hexanetriol even though the latter has more
OH groups at the end of the molecules, this phenomena can be
rationalized by the generation of cyclic ether (tetra-hydropyran-
2-methanol and 2-oxepanol). As we know, the formation of
Table 7
Activity and selectivity for the APHDO of 1,2-hexanediol and 2,3-butanediol over 4 wt.% Pt/SiO2–Al2O3 catalyst. Reaction conditions: 5 wt.% 1,2-hexanediol and 2,3-butanediol
aqueous phase solution, 518 K, 2.93 MPa and flow rate of H2 about 45 mL minꢀ1
.
Compound
WHSV (hꢀ1
)
Carbon conversion to gas (%) from GC
Alkane selectivity (%)
CO2 selectivity (%)
Specific alkane selectivity (%)
C1
C2
C3
0.0
0.0
9.6
C4
9.6
20.9
90.2
87.6
C5
C6
1,2-Hexanediol
1,2-Hexanediol
2,3-Butanediol
2,3-Butanediol
0.73
1.45
0.73
1.45
76.0
40.0
100.0
69.1
85.6
77.0
96.5
95.5
14.4
23.0
3.5
0.1
0.1
0.0
0.0
0.0
0.0
0.1
0.2
48.4
53.0
0
41.6
26.0
0
4.5
12.2
0
0
Table 8
Liquid product selectivity (%)a for the APHDO of 1,2-hexanediol and 2,3-butanediol over 4 wt.% Pt/SiO2–Al2O3 catalyst at different WHSV. Reaction conditions: 5 wt.% aqueous
phase solution, 518 K, 2.93 MPa and flow rate of H2 about 45 mL minꢀ1
.
Carbon number
Component
1,2-Hexanediol
0.73 hꢀ1
2,3-Butanediol
0.73 hꢀ1
1.45 hꢀ1
1.45 hꢀ1
C3
C4
2-Propanol
Acetone
0
0
0
0
0
0
0.7
0.2
2-Butanol
0
0
0
0
0
0
26.4
73.6
0
18.7
79.2
1.1
0
2-Butanone
2,3-Butanediol
1-Pentanol
C5
C6
98.4
65.1
0
1-Hexanol
2-Hexanol
2-Hexanone
1,2-Hexanediol
1.6
0
0
0.7
34.1
0.06
0.03
0
0
0
0
0
0
0
0
0
a
The value in the table is the carbon selectivity (%) of specific compound in the liquid phase.

N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
55
Fig. 7. Major reaction pathways for hydrodeoxygenation of 1,2-hexanediol and 2,3-butanediol over the 4 wt.% Pt/SiO2–Al2O3 catalyst at 518 K.
butanone, isopropanol and acetone were identified as intermedi-
ates. The reaction pathways as shown in Fig. 7 were drawn.
bond cleavage is more preferable under such condition. In contrast,
for 2-butanol, butane is the only product (no CAC bond cleavage
occurs). The liquid product only contains unreacted 1-butanol or
2-butanol indicating that the intermediates are not stable. 2-buta-
nol is more reactive than 1-butanol. This is because a second car-
bocation is more stable than primary carbocation.
We propose the reaction pathways for 1-butanol and 2-butanol
as shown in Fig. 8. Both 1-butanol and 2-butanol can be dehy-
drated to butylenes then hydrogenated to butane. The difference
3.2.3.3. Hydrodeoxygenation of 1-butanol and 2-butanol. Table 9 lists
the result from the analysis of gas phase products from the APHDO
of 5 wt.% 1-butanol and 2-butanol in aqueous solutions at 518 K
and 2.93 MPa. For 1-butanol, butane, propane and CO2 were iden-
tified as the final products. Higher CO2 and propane selectivity
were observed at higher space velocity, indicating that the CAC
Table 9
Activity and selectivity for the APHDO of 1-butanol and 2-butanol over 4 wt.% Pt/SiO2–Al2O3 catalyst. Reaction conditions: reaction conditions: 5 wt.% butanol aqueous phase
solution, 518 K, 2.93 MPa and flow rate of H2 about 45 mL minꢀ1
.
Compound
WHSV (hꢀ1
)
Carbon conversion to gas (%) from GC
Alkane selectivity (%)
CO2 selectivity (%)
Specific alkane selectivity (%)
C1
C2
C3
C4
42.4
30.1
100.0
100.0
1-Butanol
1-Butanol
2-Butanol
2-Butanol
0.73
1.45
0.73
1.45
51.7
40.6
98.9
96.0
82.9
81.7
100
17.1
18.3
0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
57.6
69.8
0.0
100
0
0.0
Fig. 8. Major reaction pathways for the hydrodeoxygenation of 1-butanol and 2-butanol over 4 wt.% Pt/SiO2–Al2O3 catalyst at 518 K.

56
N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
is that 1-butanol can go through a continuous dehydrogenation
and decarbonylation reaction to produce propane and CO2. 2-Buta-
nol cannot undergo the CAC bond cleavage by decarbonylation be-
cause its dehydrogenation product is a ketone and not an aldehyde.
present on the catalyst surface. Any CO produced is rapidly con-
verted into CO2 and H2 by the water–gas shift reaction [10]. The
intermediates between isosorbide and hexanetriol probably also
undergo CAC bond cleavage. Most of the C4 and C5 oxygenates de-
tected were either diols or alcohols. This indicates that the major-
ity of the C4 and C5 oxygenates are formed from decarbonylation
reactions. The C4 and C5 oxygenates can either undergo dehydra-
tion/hydrogenation reactions to form their respective alkanes or
decarbonylation reactions to form lighter oxygenates, alkanes
and CO. The C2–C3 oxygenates were produced by two different
pathways. The C2–C3 oxygenates that were diols or triols were
produced primarily by retro-aldol condensation of sorbitol. These
products are then converted into alcohols and alkanes by hydro-
deoxygenation reactions. The lighter oxygenates can also be
formed through decarbonylation reactions. Methanol and methane
may be generated by the retro-adol reaction from glycerol as pro-
posed by Davis et al. [26] or decarbonlyation of glycol or ethanol.
As shown in Fig. 9, this reaction scheme hydrodeoxygenation of
sorbitol involves a number of different products all of which are
formed by four key reactions. These four key reactions can be ad-
justed to make targeted products.
4. Discussion
4.1. Reaction pathways for APHDO of sorbitol
The overall reaction pathway for the conversion of sorbitol to
light alkanes is shown in Fig. 9. This reaction pathway involves
at least four key reactions: dehydration, hydrogenation, decarb-
onylation and retro-aldol condensation. The first step in this reac-
tion pathway involves either dehydration or retro-aldol
condensation. The sorbitol undergoes intermolecular dehydration
to 1,4-sorbitol followed by further dehydration to isosorbide. Iso-
sorbide then undergoes hydrogenation (or hydrogenolysis) of the
CAOAC bond followed by a further dehydration/hydrogenation
step to form 1,2,6-hexanetriol. We are unable to detect any inter-
mediates between the isosorbide and the hexanetriol indicating
that these intermediates are most likely rapidly converted into
other products. The hexanetriol then undergoes further dehydra-
tion/hydrogenation to hexanol or decarbonylation reactions to pro-
duce CO and lighter alcohols or polyols. The hexanols can also
undergo decarbonylation to form CO and pentane or dehydra-
tion/hydrogenation to form hexane. The polyols are most likely
converted into their aldehydes before they undergo decarbonyla-
tion. We only detected small amounts of aldehydes in our products
most likely because the aldehydes were rapidly hydrogenated and
decomposed on the catalyst surface. These aldehydes may only be
4.2. Carbon–carbon bond cleavage reactions
A key step in this chemistry is CAC bond cleavage. If lighter oxy-
genates and alcohols are desired then the rate of this reaction must
be increased. In contrast, when larger products are expected then
this reaction should be inhibited. It seems that the majority of
the CAC bond cleavage occurs on Pt sites by two pathways:
decarbonylation and retro-aldol condensation.
Fig. 9. Major reaction pathways and chemical reactions for the hydrodeoxygenation of sorbitol at 518 K.

N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
57
Decarbonylation occurs by CAC bond cleavage of aldehydes on
the metal surface forming a CO. The aldehydes are generated by
the dehydrogenation of primary alcohol or polyol over metal sur-
faces. In our reaction, we do not detect large amounts of aldehydes
because these species are very reactive and will either undergo
hydrogenation or decarbonylation. Dumesic et al. [27] have stud-
ied how CAC bonds are broken for ethanol-derived surface species
over a Pt surface. These results indicate what may happen in our
reaction chemistry for the surface chemistry for decarbonylation
of an alcohol or polyol. Ethanol initially adsorbs on the surface
most likely as a hydroxyethyl species. The adsorbed hydroxyethyl
species then undergoes dehydrogenation reactions to form ad-
sorbed carbonyl species and hydrogen. These adsorbed carbonyl
species include acetyl (CH3CO), ketene (CH2CO) and ketenyl
(CHCO). The species that has the lowest CAC bond cleavage is
the ketenyl species. They also show that the rate of CAC bond
cleavage is significantly lower than the rate of CAO bond cleavage
on Pt surfaces. This suggests that CAC bond cleavage by decarony-
lation occurs by three major steps: (1) adsorption of the alcohol on
the surface; (2) dehydrogenation of the adsorbed species to a car-
bonyl compounds and (3) CAC bond cleavage.
Retro-adol condensation reactions occur primarily only for sor-
bitol on the metal catalytic site. This reaction produces primarily
C2 and C3 oxygenates. It has previously been shown that sorbitol
can be converted into glycerol, 1,2-propanediol and ethylene glycol
over a series of noble metal-based [28–34] catalysts by retro-aldol
reactions. These are the same products that we observe for hydro-
deoxygenation of sorbitol over Pt/SiO2–Al2O3. A hydroxyl group at
the third C atom or a C@C between the second and third carbon
atoms is also necessary for this reaction. In our case, the retro-adol
condensation is most likely to happen at the very beginning of the
reaction when there are more oxygen atoms on the feed molecule.
A pathway for retro-aldol condensation of sorbitol is shown in
Fig. 10. The first step in this pathway is dehydrogenation of sorbitol
which most likely occurs on metal catalytic sites. The ketone can
form at either the two or three site of the sorbitol. CAC bond
cleavage can then occur at the adjacent carbon atom by retro-adol
condensation. According to literature, this step can happen at acid
[35–37], basic [29,34,38,39] or even neutral conditions [29]. The
glyceraldehyde and dihydroxyacetone are produced when sobitol
is dehydrogenated at two site. These products can be hydrogenated
to glycerol and then further converted to 1,2-propanediol by dehy-
dration/hydrogenation. In contrast, for sorbitol dehydrogenated at
three sites, hydroxyl-acetaldehyde and erythrose are generated as
the products of retro-adol condensation. The erythrose can be
hydrogenated to threitol or can undergo further retro-adol conden-
sation to two hydroxyl-acetaldehyde. Hydroxyl-acetaldehyde can
be hydrogenated to form glycol. We have been unable to identify
the aldehydes that form in the reaction pathway in Fig. 10 probably
primarily due to the high concentration of hydrogen in our reactor.
4.3. Carbon–oxygen bond cleavage reactions
Carbon oxygen bond cleavage occurs by dehydration mainly
over Brønsted acid catalyst sites. For APHDO of sorbitol, a number
of different functionalities are hydroxyl including polyols and alco-
hols. Dehydration of polyols is done through three different routes:
CAOAC bond formation, C@O bond formation and C@C bond for-
mation. If large amounts of oxygen are present on the molecules
then CAOAC bonds will be formed as is the case for dehydration
of sorbitol and 1,4-sorbitan. This type of dehydration results in
the formation of 5- and 6-member rings. For CAOAC bonds to
form, there must be hydroxyl groups in the correct positions on
the feedstocks.
Dehydration of diols with adjacent hydroxyl groups (e.g. 1,2- or
2,3- polyols) results in a C@O group and water as shown in Fig. 7.
This is in agreement with dehydration experiments performed by
Olah and co-workers on dehydration of diols [40]. From the carbon
Fig. 10. Reaction pathways for dehydrogenation and retro-adol condensation of sorbitol.

58
N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
conversion to gas phase of 1,2-hexanediol and 2,3-butanediol (see
Table 7) at the same WHSV, the APHDO of 2,3-butanediol is rela-
tively faster than the APHDO of 1,2-hexanediol. Moreover, as we
noticed in the APHDO of sorbitol, the concentration of 1,2-butane-
diol in the liquid product is also higher than that of 2,3-butanediol.
Dehydration of alcohols results in the formation of C@C bonds
and water. The rate of dehydration is a function of the location of
the alcohol on the molecules. As shown in Table 9, the rate of dehy-
dration of 2-butanol is significantly higher than the rate of dehy-
dration of 1-butanol. The rate of dehydration of alcohols is
known to decrease for tertiary alcohols > secondary alco-
hols > primary alcohols [41–43]. Dehydration occurs through for-
mation of carbocations with the stability of the carbocation
decreasing as tertiary carbocation > secondary carbocat-
ion > primary carbocation. It should be noted that for APHDO of
sorbitol, we observed more primary alcohols than secondary alco-
hols. No tertiary alcohols are formed in our reaction chemistry.
These dehydration reactions are most likely to occur over
Brønsted acid sites. The presence of water also has a critical role
in determining the rate of the dehydration reaction. In recent work
of Iglesia et al. [44,45], it was found that the turnover rate for dehy-
dration of 2-butanol increased with Brønsted acid strength; there-
fore, they suggested that the dehydration of 2-butanol can be used
as a rigorous method to estimate the deprotonation energy (DPE)
and acid strength for solid Brønsted acids. Dumesic et al. [46] also
studied the catalytic performance of a series of solid acid catalysts
for the dehydration 2-butanol with high water concentration.
SiO2–Al2O3, niobium phosphate and niobic acid were found to be
stable and active for the dehydration of butanol. Their activities in-
creased in the presence of water due to the increase in the concen-
tration of Brønsted acid sites. Zeolites catalyst (Beta, USY, H-ZSM-
5) and zirconia-based superacid catalysts (WOx/ZrO2 and MoOx/
ZrO2) were ineffective due to deactivation or low catalytic activity.
The lower activity of WOx/ZrO2 and MoOx/ZrO2 may be explained
by their lower acidity in the presence of large amount of water.
Brønsted acid sites, high acid strength, tolerance to poisoning) to-
gether with proper metals (or alloys) with higher activity for the
hydrogenation reaction, we believe it will be possible to more
selectively produce larger alkanes from sorbitol or other bio-
mass-derived oxygenates.
4.5. Potential of hydrodeoxygenation routes to make targeted products
Biomass conversion involves oxygen removal [1]. Hydrodeoxy-
genation chemistry is one of the major critical routes to success-
fully remove oxygen from biomass-derived species and make
targeted products liquid or oxygenated products. We have shown
the chemistry of hydrodeoxygenation with sorbitol in this paper.
This same chemistry can be applied to conversion of other bio-
mass-derived species including bio-oils [47], C5 sugars, C6 sugars
and even potentially lignin [16]. As we have shown in this paper,
hydrodeoxygenation routes can produce a large variety of products
including polyols, alcohols, ketones, cyclic ethers, and alkanes. Sev-
eral other researchers have used hydrodeoxygenation chemistry to
make targeted products including dimethylfuran [24], polyols [48],
and tetrahydrofuran [49].
The products from hydrodeoxygenation pathways can be used
for a variety of applications. Straight chain C5 and C6 alkanes could
be isomerized into gasoline. Lighter alkanes can be used for natural
gas or liquefied petroleum gas. Oxygenates and alkanes could also
be used as gasoline or as chemicals. Targeted oxygenates can also
undergo CAC bond formation reactions to produce diesel and jet
fuel [13,14]. Short chain alcohols could be converted into olefins
for polymer industry. Cyclic ethers can be separated and used as
solvents, additives, or blending agents for transportation fuels
[17]. Upgraded products can also be passed over zeolites catalysts
to produce aromatics [17]. In this respect, APHDO is a critical reac-
tion for conversion of biomass-derived species into targeted fuels
and chemicals. However, it is necessary to understand and control
the key fundamental reactions to use hydrodeoxygenation chemis-
try to make these targeted products. We project that with future
understanding based on mechanistic insights into spectroscopic
studies, theoretical study of catalyst surfaces and kinetic study im-
proved APHDO processes will be developed which will allow us to
more selectively make targeted products by hydrodeoxygenation
chemistry.
4.4. Hydrogenation reactions
Hydrogenation occurs on the surface of active metal. Three dif-
ferent functionalities are hydrogenated in APHDO of sorbitol
including: C@O bonds, C@C bonds, and CAOAC bonds. Only trace
amounts of C@C bonds were detected which indicated that under
our reactions conditions this reaction was fast. The only C@C bonds
that were detected were in the furan form, which is more stable
than straight chain C@C bonds. Only small amounts of C@O bonds
were detected also indicating that this reaction is fast under our
reaction conditions. However, large amounts of CAOAC bonds
were detected. This indicates that the rate of hydrogenation in-
creases for the different type of functionalities as
CAOAC < C@O < C@C bonds. The C@O bonds may be present in
higher concentrations on the metal surface. As discussed earlier,
the C@O bonds are necessary for CAC bond cleavage. Therefore,
it is desirable to find a metal catalyst that has a higher activity
for both C@O bond and C@C hydrogenation if one wants to sup-
press CAC bond cleavage.
From the result mentioned earlier, we can see that both
Brønsted acid site and metal sites are needed for production of al-
kanes. When a Pt catalyst is added to a support without Brønsted
acid sites (such as carbon, alumina) [10,11], only a small amount
of C5 and C6 alkanes are formed with CO2, H2 and light alkanes
being the major product. In contrast, isosorbide is the final product
when SiO2–Al2O3 is the catalyst. At the same time, the cooperation
of Pt and Brønsted acid sites (dehydration followed by hydrogena-
tion) is very important for the high activity and selectivity of the
catalyst towards larger alkanes. In the future, with the utilization
of some new catalyst materials with strong Brønsted acidity (more
5. Conclusions
In this work, we investigated the reaction pathways for the
aqueous-phase hydrodeoxygenation (APHDO) of sorbitol over the
4 wt.% Pt/SiO2–Al2O3 catalyst. A wide variety of reaction intermedi-
ates were identified by GC–MS and HPLC. From the analysis of the
gas phase and liquid phase products from the APHDO of sorbitol
and a series of model compounds, we can see that APHDO process
is mainly composed of three classes of reactions: CAC bond cleav-
age, CAO bond cleavage, and hydrogenation reactions. The key
CAC bond cleavage reactions include: retro-aldol condensation
and decarbonylation which most likely occur on the surface of Pt
metal. Dehydration is the most important CAO bond cleavage reac-
tion and most likely occurs on Br£nstead acid sites. Sorbitol ini-
tially undergoes either dehydration, to form 1,4-sorbitan and
then isosorbide, or CAC retro-aldol condensation reactions to pro-
duce primarily C3 polyols. The isosorbide generated is converted to
1,2,6-hexanetriol by a ring opening hydrogenation step followed
by dehydration/hydrogenation. The 1,2,6-hexanetriol is further
dehydrated and hydrogenated to form hexane as the final product.
Lighter alkanes and alcohols are produced by CAC bond cleavage
decarbonylation reactions to form C1–C5 oxygenates and alkanes.
It is likely that future advances in understanding the surface

N. Li, G.W. Huber / Journal of Catalysis 270 (2010) 48–59
59
[17] E.L. Kunkes, D.A. Simonetti, R.M. West, J.C. Serrano-Ruiz, C.A. Gartner, J.A.
Dumesic, Science 322 (2008) 417.
[18] J.N. Chheda, J.A. Dumesic, Catal. Today 123 (2007) 59.
[19] N. Ji, T. Zhang, M.Y. Zheng, A.Q. Wang, H. Wang, X.D. Wang, J.G.G. Chen, Angew.
Chem. Int. Edit. 47 (2008) 8510.
chemistry of catalysts combined with improvements in design of
catalysts and theoretical studies will allow us to use hydrodeoxy-
genation chemistry to make targeted products from biomass-de-
rived species.
[20] A. Corma, G.W. Huber, L. Sauvanauda, P. O’Connor, J. Catal. 257 (2008) 163.
[21] J.N. Chheda, Y. Roman-Leshkov, J.A. Dumesic, Green Chem. 9 (2007) 342.
[22] Y. Roman-Leshkov, J.A. Dumesic, Top. Catal. 52 (2009) 297.
[23] Y. Roman-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933.
[24] Y. Roman-Leshkov, C.J. Barrett, Z.Y. Liu, J.A. Dumesic, Nature 447 (2007) 982.
[25] R.R. Davda, J.A. Dumesic, Angew. Chem. Int. Edit. 42 (2003) 4068.
[26] E.P. Maris, R.J. Davis, J. Catal. 249 (2007) 328.
Acknowledgments
This work was supported by a NSF-MRI grant and by the De-
fense Advanced Research Project Agency through the Defense Sci-
ence Office Cooperative Agreement W911NF-09-2-0010 (Surf-Cat:
Catalysts for production of JP-8 range molecules from lignocellu-
losic Biomass. Approved for Public Release, Distribution Unlim-
ited). We would like to thank T.P. Vispute, Dr. V.S. Ayyagari and
H. Olcay for technical assistance.
[27] R. Alcala, M. Mavrikakis, J.A. Dumesic, J. Catal. 218 (2003) 178.
[28] E. Tronconi, N. Ferlazzo, P. Forzatti, I. Pasquon, B. Casale, L. Marini,
A
Mathematical-model for the Catalytic Hydrogenolysis of Carbohydrates,
Pergamon-Elsevier Science Ltd., June 28–July 01, Turin, Italy, 1992, p. 2451.
[29] D.K. Sohounloue, C. Montassier, J. Barbier, React. Kinet. Catal. Lett. 22 (1983)
391.
[30] L. Zhao, J.H. Zhou, Z.J. Sui, X.G. Zhou, Chem. Eng. Sci. (2009), doi:10.1016/
j.ces.2009.03.026.
[31] B.W. Hoffer, R. Prochazka, WO 2008071642, BASF SE, Germany, 2008.
[32] B.W. Hoffer, R. Prochazka, WO 2008071641, BASF SE, Germany, 2008.
[33] B.W. Hoffer, R. Prochazka, WO 2008071616, BASF SE, Germany, 2008.
[34] C. Cao, J.F. White, Y. Wang, J.G. Frye, Chem. Ind. 115 (2007) 289.
[35] O. Munoz-Muniz, M. Quintanar-Audelo, E. Juaristi, J. Org. Chem. 68 (2003)
1622.
References
[1] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044.
[2] R.R. Davda, J.W. Shabaker, G.W. Huber, R.D. Cortright, J.A. Dumesic, Appl. Catal.
B – Environ. 56 (2005) 171.
[3] J.N. Chheda, G.W. Huber, J.A. Dumesic, Angew. Chem. Int. Edit. 46 (2007) 7164.
[4] J.W. Shabaker, G.W. Huber, J.A. Dumesic, J. Catal. 222 (2004) 180.
[5] J.W. Shabaker, R.R. Davda, G.W. Huber, R.D. Cortright, J.A. Dumesic, J. Catal. 215
(2003) 344.
[6] G.W. Huber, J.W. Shabaker, S.T. Evans, J.A. Dumesic, Appl. Catal. B: Environ. 62
(2006) 226.
[7] G.W. Huber, J.W. Shabaker, J.A. Dumesic, Science 300 (2003) 2075.
[8] R.R. Davda, J.W. Shabaker, G.W. Huber, R.D. Cortright, J.A. Dumesic, Appl. Catal.
B: Environ. 43 (2003) 13.
[9] R.R. Davda, J.A. Dumesic, Chem. Commun. (2004) 36.
[10] R.D. Cortright, R.R. Davda, J.A. Dumesic, Nature 418 (2002) 964.
[11] G.W. Huber, R.D. Cortright, J.A. Dumesic, Angew. Chem. Int. Edit. 43 (2004)
1549.
[12] Y.C. Lin, G.W. Huber, Energy Environ. Sci. 2 (2009) 68.
[13] R.M. West, Z.Y. Liu, M. Peter, C.A. Gartner, J.A. Dumesic, J. Mol. Catal. A – Chem.
296 (2008) 18.
[14] R.M. West, Z.Y. Liu, M. Peter, J.A. Dumesic, ChemSusChem 1 (2008) 417.
[15] G.W. Huber, J.N. Chheda, C.J. Barrett, J.A. Dumesic, Science 308 (2005) 1446.
[16] N. Yan, C. Zhao, P.J. Dyson, C. Wang, L.T. Liu, Y. Kou, ChemSusChem 1 (2008)
626.
[36] J. Binapfl, DE 875512, BAYER AG, 1953.
[37] J.E. Dolfini, J. Glinka, S.C. Bosch, US Patent 4927805, 1990.
[38] C.S. Cao, J.F. White, Y. Wang, J.G. Frye, in: S.R. Schmidt (Ed.), Selective
Hydrogenolysis of Sugar Alcohols Over Structured Catalysts, CRC Press-Taylor
& Francis Group, Orlando, FL, 2006, p. 289 (2–6 April).
[39] I.T. Clark, Ind. Eng. Chem. 50 (1958) 1125.
[40] I. Bucsi, Á. Molnár, M. Bartók, G.A. Olah, Tetrahedron 51 (1995) 3319.
[41] M. Clugston, R. Flemming, Advanced Chemistry, Oxford University Press, 2000.
[42] V.K. Ahluwalia, M. Goyal, A Textbook of Organic Chemistry, Alpha Science Int’l
Ltd., 2000.
[43] T. Nishiguchi, C. Kamio, J. Chem. Soc. – Perkin Trans. 1 (1989) 707.
[44] J. Macht, R.T. Carr, E. Iglesia, J. Am. Chem. Soc. 131 (2009) 6554.
[45] J. Macht, R.T. Carr, E. Iglesia, J. Catal. 264 (2009) 54.
[46] R.M. West, D.J. Braden, J.A. Dumesic, J. Catal. 262 (2009) 134.
[47] T.P. Vispute, G.W. Huber, Green Chem. (2009), doi:10.1039/b912522c.
[48] S. Koso, I. Furikado, A. Shimao, T. Miyazawa, K. Kunimori, K. Tomishige, Chem.
Commun. (2009) 2035.
[49] R.M. West, E.L. Kunkes, D.A. Simonetti, J.A. Dumesic, Catal. Today (2009),
doi:10.1016/j.cattod.2009.02.004.