Impact of solvent for individual steps of phenol hydrodeoxygenation with Pd/C and HZSM-5 as catalysts

Article abstract of DOI:10.1016/j.jcat.2013.09.009

Impacts of water, methanol, and hexadecane solvents on the individual steps of phenol hydrodeoxygenation are investigated over Pd/C and HZSM-5 catalyst components at 473 K in presence of H₂. Hydrodeoxygenation of phenol to cyclohexane includes four individual steps of phenol hydrogenation to cyclohexanone on Pd/C, cyclohexanone hydrogenation to cyclohexanol on Pd/C, cyclohexanol dehydration to cyclohexene on HZSM-5, and cyclohexene hydrogenation to cyclohexane on Pd/C. Individual phenol and cyclohexanone hydrogenation rates are much lower in methanol and hexadecane than in water, while rates of cyclohexanol dehydration and cyclohexene hydrogenation are similar in three solvents. The slow rate in methanol is due to the strong solvation of reactants and the adsorption of methanol on Pd, as well as to the reaction between methanol and the cyclohexanone intermediate. The low solubility of phenol and strong interaction of hexadecane with Pd lead to the slow rate in hexadecane. The apparent activation energies for hydrogenation follow the order E _{a phenol} > E_{a cyclohexanone} > E _{a cyclohexene}, and the sequences of individual reaction rates are reverse in three solvents. The dehydration rates 1.1-1.8×10 ³molmolBAS-1h^-1 and apparent activation energies (115-124 kJ mol^-1) are comparable in three solvents. In situ liquid-phase IR spectroscopy shows the rates consistent with kinetics derived from chromatographic evidence in the aqueous phase and verifies that hydrogenation of phenol and cyclohexanone follows reaction orders of 1.0 and 0.55 over Pd/C, respectively. Conversion of cyclohexanol with HZSM-5 shows first-order dependence in approaching the dehydration-hydration equilibrium in the aqueous phase. Published by Elsevier Inc.

Full text of DOI:10.1016/j.jcat.2013.09.009

Journal of Catalysis 309 (2014) 362–375
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Impact of solvent for individual steps of phenol hydrodeoxygenation
with Pd/C and HZSM-5 as catalysts
a,b,
Jiayue He a, Chen Zhao a, , Johannes A. Lercher
⇑
⇑
a Department of Chemistry and Catalysis Research Center, Technische Universität München, Lichtenbergstraße 4, 85747 Garching, Germany
b Institute for Integrated Catalysis, Pacific Northwest National Laboratory, 902 Battelle Boulevard, Richland, WA 99352, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Impacts of water, methanol, and hexadecane solvents on the individual steps of phenol hydrodeoxygen-
ation are investigated over Pd/C and HZSM-5 catalyst components at 473 K in presence of H2. Hydrodeox-
ygenation of phenol to cyclohexane includes four individual steps of phenol hydrogenation to
cyclohexanone on Pd/C, cyclohexanone hydrogenation to cyclohexanol on Pd/C, cyclohexanol dehydra-
tion to cyclohexene on HZSM-5, and cyclohexene hydrogenation to cyclohexane on Pd/C. Individual phe-
nol and cyclohexanone hydrogenation rates are much lower in methanol and hexadecane than in water,
while rates of cyclohexanol dehydration and cyclohexene hydrogenation are similar in three solvents.
The slow rate in methanol is due to the strong solvation of reactants and the adsorption of methanol
on Pd, as well as to the reaction between methanol and the cyclohexanone intermediate. The low solu-
bility of phenol and strong interaction of hexadecane with Pd lead to the slow rate in hexadecane. The
Received 5 May 2013
Revised 8 September 2013
Accepted 13 September 2013
Keywords:
Phenol hydrodeoxygenation
Individual steps
In situ liquid IR spectroscopy
Kinetics
apparent activation energies for hydrogenation follow the order Ea phenol > Ea cyclohexanone > Ea
,
Solvent effect
cyclohexene
and the sequences of individual reaction rates are reverse in three solvents. The dehydration rates
ꢀ
ꢁ
1:1—1:8 ꢀ 103 mol molBꢁA1S hꢁ1 and apparent activation energies (115–124 kJ molꢁ1) are comparable
in three solvents. In situ liquid-phase IR spectroscopy shows the rates consistent with kinetics derived
from chromatographic evidence in the aqueous phase and verifies that hydrogenation of phenol and
cyclohexanone follows reaction orders of 1.0 and 0.55 over Pd/C, respectively. Conversion of cyclohexanol
with HZSM-5 shows first-order dependence in approaching the dehydration–hydration equilibrium in
the aqueous phase.
Published by Elsevier Inc.
1. Introduction
selectivity of the catalysts, and the state of the reactant is limited
in the three-phase system (gas, liquid, and solid phases) during
Hydrodeoxygenation is considered to be an effective approach
to remove the oxygen from the products of low rank coal and oil
shale, and especially of biomass liquefaction [1]. Since the oxygen
content of biomass (up to 50 wt.%) markedly exceeds that of coal
(below 10 wt.%), advanced catalytic approaches are required for
economically feasible deoxygenation of biomass and its deriva-
tives. Solvents dramatically influence the rates and selectivity
in these catalytic processes, by changing the reactive state of
these complex bio-polymers and the products in the thermal
reactions. Therefore, the selection of a proper solvent is of great
importance for an efficient catalytic process. However, informa-
tion about the relation between solvent nature, the activity and
the hydrodeoxygenation.
Recently, we reported that the hydrodeoxygenation of bio-de-
rived phenolic components to cycloalkanes can be quantitatively
achieved using metallic (Pd, Pt, and Ni) and liquid (H3PO4) or solid
acid catalysts (Nafion, SO42ꢁ/ZrO2, HZSM-5, and HBEA,) at 473–
523 K in water [2–7]. The reaction sequence of phenol hydrodeox-
ygenation with bifunctional sites of metal and acid sites in the
aqueous phase is concluded to follow initial phenol hydrogenation
to cylcohexanol/cyclohexanone and subsequent cyclohexanone
hydrogenation to cyclohexanol catalyzed by metals, and then
acid-catalyzed dehydration of cyclohexanol to cyclohexene, and fi-
nal cyclohexene hydrogenation to cyclohexane [2–4]. It is, how-
ever, important to explore and compare the fundamental
chemistry of hydrodeoxygenation of phenolic compounds in sol-
vents of varying properties in order to understand, how the solvent
environment modifies the solvent–catalyst, solvent–reactant, and
reactant–catalyst interactions and how it might influence the
catalyst itself.
⇑
Corresponding authors at: Department of Chemistry and Catalysis Research
Center, Technische Universität München, Lichtenbergstrabe 4, 85747 Garching,
Germany (C. Zhao, J.A. Lercher). Tel.: +49 89 289 13540; fax: +49 89 289 13544.
E-mail addresses: chenzhao@mytum.de (C. Zhao), johannes.lercher@ch.tum.de
(J.A. Lercher).
0021-9517/$ - see front matter Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.jcat.2013.09.009

J. He et al. / Journal of Catalysis 309 (2014) 362–375
363
Water comprises typically 30 wt.% of raw bio-oil and is a suit-
able solvent for an upgrading process. Methanol has been success-
fully used for lignocellulose conversion especially for lignin
liquefaction [8], and nonpolar solvents such as hexadecane and tol-
uene have been employed for related hydrotreating of synthetic
crude coal and oil shale [9,10]. The solvent for hydrodeoxygenation
should, therefore, (i) be stable with respect to the process temper-
ature, pressure, and catalyst, (ii) hold reactants and intermediates
in solution; (iii) dissolve and transport H2, and (iv) dissipate the
heat of the highly exothermic reactions.
isotherm to zero H2 pressure, and this value was used to calculate
the dispersion of metal on supports.
2.3.3. Transmission electron microscopy (TEM)
TEM images were measured on a JEM-2010 Jeol transmission
electron microscope operating at 120 kV. Prior to the measure-
ments, catalyst was ground, suspended in ethanol, and dispersed
by ultrasonic treatment. The dispersion was dropped on a copper
grid-supported carbon film. The average cluster size was calculated
by counting 300 Pd particles.
In this contribution, we report the detailed kinetics of individual
steps for phenol hydrodeoxygenation with Pd/C or HZSM-5 in
water, methanol, and hexadecane. The turnover frequencies (TOFs)
and apparent activation energies for individual steps are explored.
The relationships between the chemical nature of the solvents and
the interactions at the gas–liquid–solid interfaces are investigated
in order to understand the fundamental differences between the
reactions occurring in media with differing properties. The sol-
vent–catalyst, solvent–reactant, and reactant–catalyst interactions
and their influence on the activity are discussed in this work as
well. In situ liquid-phase IR spectroscopy is employed to monitor
the concentrations of the reactants and products during individual
reaction steps in water to obtain kinetic information.
2.3.4. Temperature-programmed desorption (TPD)
HZSM-5 was activated in flowing He at 623 K for 1 h using a
heating rate of 5 K minꢁ1 from room temperature to 623 K. NH3
was adsorbed by adding 10 vol.% to the He carrier gas (total flow
30 mL minꢁ1) at 423 K. The sample was purged with He for 2 h
to remove physisorbed molecules. For TPD, the sample was heated
in He at a rate of 10 K minꢁ1 from 373 K to 1033 K for NH3 desorp-
tion. The species desorbing was monitored by mass spectrometry
(Balzers QME 200). For quantification, a reference material with
known acid site concentration (HZSM-5 with Si/Al = 45) was used.
2.3.5. IR spectra of adsorbed pyridine
IR spectroscopy with pyridine as probe molecule was used to
determine the acid site concentrations and distributions. The IR
spectra were measured with a PerkinElmer 2000 spectrometer
operated at a resolution of 4 cmꢁ1. The sample was activated at
723 K for 2 h in vacuum, and a background spectrum was recorded
after the temperature decreased to 423 K. The activated sample
was exposed to pyridine vapor (1.0 ꢀ 10ꢁ5 MPa) at 423 K for
0.5 h. After removing physisorbed pyridine by outgassing at
423 K for 0.5 h, the spectra were recorded.
2. Experimental section
2.1. Chemicals
All chemicals were provided from commercial suppliers: phenol
(Sigma–Aldrich, 99% GC assay), cyclohexene (Sigma–Aldrich, 99%
GC assay), cyclohexanol (Sigma–Aldrich, 99% GC assay), cyclohex-
anone (Sigma–Aldrich, P99.5%, GC assay), ethyl acetate (Sigma–
Aldrich, P99.5%, GC assay), methanol (Sigma–Aldrich, P99.93%,
GC assay), hexadecane (Merck, >99%, GC assay), toluene (Sigma–
Aldrich, P99.5%, GC assay), hydrogen (Westfalen AG,
99.999 vol.%), nitrogen (Westfalen AG, 99.999 vol.%), synthetic air
(Westfalen AG, 99.999 vol.%), and ultra water system (EASYpure
2.3.6. Atomic absorption spectroscopy (AAS)
A UNICAM 939 AA-Spectrometer was used to determine the
concentration of noble metals in the carbon supported catalyst. Be-
fore measurement, 20–40 mg of the sample was dissolved in a
mixture of 1.0 mL of 37 wt.% hydrochloric acid and 1.0 mL of
65 wt.% nitric acid at the boiling point of the mixture.
II, Resistivity: 18.2 MX cm).
2.2. Catalysts
2.4. Measurement of catalytic reactions
The Pd/C was obtained from Sigma–Aldrich (Sigma–Aldrich,
1 wt.% Pd), and HZSM-5 (Si/Al = 45) was supplied by Clariant AG.
They were used without further treatment.
2.4.1. Individual steps of phenol hydrodeoxygenation with Pd/C and
HZSM-5 catalysts
The hydrodeoxygenation of phenol to cyclohexane includes four
steps: Step 1: phenol hydrogenation to cyclohexanone on Pd/C, Step
2: cyclohexanone hydrogenation to cyclohexanol on Pd/C, Step 3:
cyclohexanol dehydration to cyclohexene on HZSM-5, and Step 4:
cyclohexene hydrogenation to cyclohexane on Pd/C. Exploration of
reaction rates was carried out at 473 K in three solvents (water,
methanol, and hexadecane). Apparent activation energies were
calculated from rates measured at 433, 453, 473, and 493 K.
The turnover frequencies (TOFs) defining initial reaction rates
were expressed as per mole of converted reactant on per mole of
surface Pd atom (or per mole of Brönsted acid site) per hour at con-
versions lower than 15% (except for the very fast cyclohexene
hydrogenation) at 473 K in the presence of 4 MPa H2. To control
conversions in this range, the ratios of reactant to catalyst have
been manipulated (Figs. 2, 4, 5, 7). H2 gas was introduced to the
reactor at the reaction temperature of 473 K in order to make the
conversion data start from zero at t = 0.
2.3. Catalyst characterization
2.3.1. Specific BET surface area and pore diameters
The surface areas and pore diameters were determined by N2
adsorption. The N2 adsorption was carried out at 77 K on a PMI
automated BET sorptometer. Prior to the measurements, the sam-
ples were outgassed at 523 K for 20 h. The specific surface areas
and micropore and mesopore distributions were calculated based
on the BET and BJH models.
2.3.2. H2 chemisorption
The Pd supported on activated carbon was first reduced at 733 K
in presence of 0.1 MPa H2 for 4 h before measurement. The reduced
catalyst was activated at 733 K for 1 h in vacuum and then cooled
to 313 K. The H2 chemisorption and physisorption were subse-
quently measured in a pressure range from 1 kPa to 40 kPa. Then,
the physisorbed H2 was removed by outgassing the sample at the
same temperature for 1 h. A second adsorption isotherm (physi-
sorption) was measured. The concentration of chemisorbed hydro-
gen on the metal was determined by extrapolating the differential
2.4.1.1. Kinetics of individual steps of phenol hydrodeoxygenation in
water. The reaction conditions for the TOF measurements are re-
ported as footnotes in the corresponding figures. In a typical exper-
iment exemplified by Step 1 of phenol hydrogenation, phenol

364
J. He et al. / Journal of Catalysis 309 (2014) 362–375
(12.5 g), Pd/C (0.020 g, 1 wt.%), and H2O (80 mL) were loaded in a
Parr reactor (Series 4848, 300 mL); then, the autoclave was
charged with 0.1 MPa N2. After the temperature was increased to
473 K, PH2 (partial pressure for hydrogen) was charged to
4.0 MPa (ambient temperature). Then, stirring was initiated and
the reaction time was recorded from that point. Reactions were
conducted at 473 K with different duration times, or at 433, 453,
473, and 493 K for 0.5 h, with a stirring speed of 700 rpm. After
reaction, the reactor was quenched by ice to ambient temperature,
and the organic products were extracted with ethyl acetate and
analyzed by GC/MS. Because two phases existed in the four indi-
vidual steps in water, the kinetics data were collected from sepa-
rate batches with varying reaction times. The number of surface
Pd atoms is determined by H2-chemisorption measurement. The
Brønsted acid measured by IR spectroscopy of adsorbed pyridine
is used to calculate the dehydration TOF. The measurements for
calculation of apparent activation energy were conducted at 433,
453, 473, and 493 K with the reaction times of 0.5 h.
3. Results and discussion
3.1. Catalyst characterizations
The physicochemical properties of 1 wt.% Pd/C and HZSM-5
catalysts are compiled in Table 1. Pd/C had a high BET surface area
of 1103 m2 gꢁ1 and porevolume of 0.58 cm3 gꢁ1 witha Pd dispersion
of 51% determined by H2 chemisorption. The average Pd
nanoparticle size was 2.6 nm measured by TEM (Fig. S1 in SI),
agreeing reasonably well with the H2 chemisorption result (2.2 nm).
For HZSM-5, the apparent BET specific surface area was 395 m2 gꢁ1
and the pore volume was 0.26 cm3 gꢁ1 determined by N2 sorption.
The total acid concentration of HZSM-5 was 0.359 mmol gꢁ1. The
Brønsted acid sites (BAS) and Lewis acid sites (LAS) were
0.278 mmol gꢁ1 and 0.048 mmol gꢁ1, respectively, measured by IR
spectra of adsorbed pyridine (Py-IR) (Table 1). The acid concentration
determinedbyTPD-NH3 was10%higherthanthoseanalyzedbyPy-IR,
because the smaller size of probe molecule NH3 allows it to reach
more sterically limited acid sites comparing to pyridine.
2.4.1.2. Kinetics of individual steps of phenol hydrodeoxygenation in
methanol. For the reactions carried out in methanol, the proce-
dures were similar to those described in water. But because reac-
tants, intermediates, and products were all soluble in methanol,
the in situ sampling was enabled and the solvent extractions are
unnecessary. The detailed reaction conditions are presented as
footnotes of the corresponding figures. For Steps 1, 2, 3, and 4,
the samples were collected in situ enabled by solubility in metha-
nol and then analyzed by GC/MS.
3.2. Hydrodeoxygenation of phenol in three solvents
We have previously evaluated combinations of mono-func-
tional catalysts with metal or acid sites for hydrodeoxygenation
of phenol in water [2–6]. We now investigate the kinetics of indi-
vidual steps of phenol hydrodeoxygenation with a specific combi-
nation of supported Pd and HZSM-5 in water, methanol, or
hexadecane. Water has a highest polarity index 9.0 (Table 2),
methanol is lower at 6.6, and hexadecane is nonpolar with the low-
est polarity index of 0.3 [11]. H2 solubility at 298 K and 0.1 MPa H2
decreases from hexadecane (4.7 ꢀ 10ꢁ6 mol cmꢁ3) via methanol
(3.5 ꢀ 10ꢁ6 mol cmꢁ3) to water (7.8 ꢀ 10ꢁ7 mol cmꢁ3) (Table 2)
[12,13].
We expect the solubility of reactants, intermediates, and prod-
ucts to influence the reaction rates and apparent energies of activa-
tion through variations in solvent–catalyst, solvent–reactant, and
reactant–catalyst interactions, as well as via potentially induced
diffusion limitations. Vigorous stirring of the liquid-phase ensures
the absence of external gas–liquid mass transfer diffusion limita-
tions in this four-phase reaction, verified in the previous kinetic
study in water that the individual reaction steps were not influ-
enced by stirring speeds varied from 700 to 1000 rpm or varied
reactant/water or reactant/catalyst ratios [4]. In addition, solubility
of phenol, its derived intermediates, and products is different in
various solvents (Table 3). Water is partially miscible with phenol,
cyclohexanone, and cyclohexanol at 298 K, but cyclohexene and
cyclohexane are nearly immiscible in water. Phenol, ketone,
alcohol, and alkene/ane are soluble in methanol at 298 K. In
2.4.1.3. Kinetics of individual steps of phenol hydrodeoxygenation in
hexadecane. When using hexadecane as solvent, the samples were
collected from separate batch modes at varying durations for Step
1 (phenol hydrogenation on Pd/C) because the phenol reactant is
insoluble in hexadecane. After reaction, the catalysts were col-
lected by filtration, and toluene was added to the reactor to dis-
solve all organics. The homogeneous solution was analyzed by
GC/MS. For Steps 2, 3, and 4, because reactants, intermediates,
and products were all soluble in hexadecane, the in situ sampling
was enabled and the solvent extractions are unnecessary. The de-
tailed reaction conditions are presented as footnotes of the corre-
sponding figures. The samples were collected in situ enabled by
solubility in hexadecane and then analyzed by GC/MS.
2.5. In situ liquid-phase IR spectroscopy for measuring individual steps
of phenol hydrodeoxygenation in water
The in situ liquid IR measurements were performed with a Re-
act IR 1000 spectrometer (Mettler Toledo) connected to an auto-
clave (100 mL, Parr). A diamond placed in the bottom of the
autoclave was used as a probe to collect the in situ liquid IR spectra
in the liquid phase. First, a background was collected with the
working system containing water solvent (50 mL) and catalyst
(1 wt.% Pd/C 0.10 g or HZSM-5 0.040 g) at 473 K in the presence
of 4.0 MPa H2. Second, after the reactor was cooled to ambient
temperature, the individual reactions for phenol and cyclohexa-
none hydrogenation (Steps 1 and 2) were carried out at 473 K.
When the temperature reached 473 K, a 4.0 MPa H2 was intro-
duced to the autoclave and the IR spectroscopy initiated. For dehy-
dration of cyclohexanol (Step 3), the autoclave was charged with
1.5 MPa H2 at ambient temperature and the IR spectrometer
started. The IR spectra were collected every 10 min for 240 min
during the overall reaction. Background spectra were subtracted
before analysis.
Table 1
Physicochemical properties of the selected catalysts.
Catalyst
Pd/C
1.0
HZSM-5 (Si/Al = 45)
–
Metal loading (wt.%)
BET surface area (m2 gꢁ1
)
1103 395
Mesopore surface area (m2 gꢁ1
)
)
107
996
0.58
0.14
0.44
51
173
222
0.26
0.18
0.08
–
–
–
Micropore surface area (m2 gꢁ1
Pore volume (cm3 gꢁ1
)
Mesopore volume (cm3 gꢁ1
)
)
Micropore volume (cm3 gꢁ1
Pd dispersion (%)
Pd particle diameter (nm) (H2 Chemisorption) 2.2
Metal particle diameter (nm) (TEM)
2.6
–
–
Acidity (mmol gꢁ1) (TPD)
0.359
BAS: 0.278
LAS: 0.048
Acidity (mmol gꢁ1) (Py-IR)

J. He et al. / Journal of Catalysis 309 (2014) 362–375
365
Table 2
acetal product was able to react with a large excess of methanol.
The subsequent decrease in selectivity is mainly related to the fact
that the increasing concentrations of water formed from the
dehydration of methoxycyclohexanol (Compound A) shifted the
equilibrium of 1,10-dimethoxycyclohexane (Compound C) to meth-
oxycyclohexanol (Compound A) (Fig. 1b), and thus, decreased the
selectivity of 1,10-dimethoxycyclohexane (Compound C).
With methanol as solvent, cyclohexane was produced with
selectivity as low as 1.9% at 100% conversion after 170 min via
Route 1. The fast acetal reaction between ketone and methanol
(Route 2) dominated the phenol conversion in methanol. The ace-
tal reaction was much faster than cyclohexanone hydrogenation on
Pd/C in methanol under these reaction conditions (Fig. 1b). Thus,
alcohols or polyols are not suitable solvents for hydrodeoxygen-
ation of phenolic compounds to produce hydrocarbon biofuels,
although it has been reported that lignin is efficiently converted
in supercritical methanol [15].
Solvent polarity index and H2 solubility in three solvents (298 K and 0.1 MPa H2) [11–
13].
Solvent
Polarity
H2 solubility (mol cmꢁ3
)
Water
Methanol
Hexadecane
9.0
7.8 ꢀ 10ꢁ7
3.5 ꢀ 10ꢁ6
4.7 ꢀ 10ꢁ6
6.6
ca. 0.3a
a
The value is estimated from the polarity index of n-decane (0.3), n-octane (0.4),
and n-hexane (0).
hexadecane, phenol is completely insoluble, but the other interme-
diates and product ketone, alcohol, and alkene/ane are highly sol-
uble at 298 K (Table 3) [14].
3.2.1. Reaction sequence of phenol hydrodeoxygenation to cyclohexane
3.2.1.1. Reaction pathways of phenol hydrodeoxygenation in water
and in hexadecane. It has been established that with a combination
of two mono-functional catalysts, phenol hydrodeoxygenation to
cycloalkane proceeds in the aqueous phase through four steps
(Scheme 1), i.e., phenol hydrogenation to cyclohexanol/cyclohexa-
none on metal sites (Step 1), sequential cyclohexanone hydrogena-
tion to cyclohexanol on metal sites (Step 2), cyclohexanol
dehydration to cyclohexene on acid sites (Step 3), and finally
cyclohexene hydrogenation to cyclohexane on metal sites (Step
4). In hexadecane or water, the reaction pathways are identical
for phenol hydrodeoxygenation with Pd/C and HZSM-5; hydroge-
nation is followed by dehydration (Scheme 1), but in methanol,
the reaction pathway is more complex because of the acetal reac-
tions between the cyclohexanone intermediate and methanol.
3.2.2. Turnover frequencies (TOFs) of individual steps for phenol
hydrodeoxygenation in three solvents
3.2.2.1. TOFs for phenol hydrogenation over Pd/C. In the first phenol
hydrogenation step on Pd/C (Step 1, the first individual step), the
ꢀ
ꢁ
initial TOFs in methanol 1:3 ꢀ 103 mol molPꢁd1surf: hꢁ1 and hexa-
ꢀ
ꢁ
decane 7:8 ꢀ 102 mol molPꢁd1surf: hꢁ1 were significantly lower
ꢀ
ꢁ
than in water 4:2 ꢀ 103 mol molPꢁd1surf: hꢁ1 (Fig. 2, Table 4). Three
factors influence the initial hydrogenation rates in methanol. The
first is the solvation of phenol (solvent–reactant interaction). In
the case of acetophenone conversion, Bertero et al. [16] reported
that its solvation increased in the sequence 2-propanol < 1-propa-
nol < ethanol < methanol. The opposite trend was observed for ace-
tophenone hydrogenation rates with Ni/SiO2, explained by the fact
that stronger solvent–reactant interaction (solvation) reduced the
surface coverage of the reactant. The hydrogenation rate showed
0–1 reaction order with respect to the phenol concentrations,
e.g., a reaction order of 0.65 was obtained on Pd/Ca-Al2O3 at
503 K [17]. Thus, we hypothesize that the stronger solvation of
phenol in methanol than in water or hexadecane reduces its
adsorption on Pd and lowers the phenol hydrogenation rate.
The second consideration is related to the interaction of meth-
anol with the Pd surface (solvent–catalyst interaction). Solvent
adsorption on the catalyst surface may have strongly influenced
the Pd activity for phenol hydrogenation. For example, methanol
may compete for the catalyst surface and decrease surface cover-
age of phenol. Sexton et al. [18] reported that CH3OH adsorbed
more strongly than water on Pd and, thus, affects the surface
phenol coverage more strongly than water.
3.2.1.2. Hydrodeoxygenation pathway of phenol in methanol. The
product distribution from phenol conversion in methanol is plotted
in Fig. 1a as a function of time with Pd/C and HZSM-5 at 473 K. At
t = 0, the primary product was cyclohexanone [4]. The ketone was
rapidly consumed via the acetal reaction (Route 2 in Fig. 1b) but
not by the parallel hydrogenation route (Route 1 in Fig. 1b). The
selectivity to the primary acetal product (Compound A, methoxy-
cyclohexanol formed from cyclohexanone and methanol, Fig. 1b)
remained very low (<5 C%) during the reaction. Compound A was
either rapidly dehydrated by HZSM-5 (Route 3) and hydrogenated
over Pd to cyclohexyl methyl ether (Compound B), or this interme-
diate reacted further with methanol to form 1,10-dimethoxycyclo-
hexane (Compound C, Route 2, slower than Route 3) (Fig. 1b).
Thus, cyclohexyl methyl ether (Compound B) became the major
product with 81% selectivity at 170 min.
The selectivity to 1,10-dimethoxycyclohexane (Compound C)
first increased to a maximum of 29% at 10 min and then decreased
to 3.3% at 170 min. The initial selectivity increased, because the
The third impact results from the reaction between cyclohexa-
none and methanol (Fig. 1). The hydrogenated cyclohexanone
Table 3
Solubilities of phenol, cyclohexanone, cyclohexanol, cyclohexene, and cyclohexane at 298 K (gꢂ(100 mL)ꢁ1) [12].
Solvent
Phenol
Cyclohexanone
Cyclohexanol
Cyclohexene
Cyclohexane
Water
Methanol
Hexadecane
8.3
>62.5
Insoluble
5.0
>62.5
>62.5
3.6
>62.5
>62.5
0.025
>62.5
>62.5
Immiscible
30–33
>62.5
OH
OH
O
Pd/C,H2
Step 1
Hydrogenation
Pd/C, H2
HZSM-5
Pd/C, H2
Step 4
Hydrogenation
Step 2
Hydrogenation
Step 3
Dehydration
Scheme 1. Reaction sequence on hydrodeoxygenation of phenol to cyclohexane on metal and acid catalysts using water or hexadecane as solvents.

366
J. He et al. / Journal of Catalysis 309 (2014) 362–375
(a)
(b)
100
100
80
60
40
20
0
80
Route 1
Methoxyl cyclohexanol
Cyclohexyl methyl ether
Dimethoxyl cyclohexane
Cyclohexanol
Cyclohexanone
Cyclohexane
Covnersion
60
Route 2
40
A
C
B
20
Route 3
0
0
20 40 60 80 100 120 140 160 180
Time (min)
Fig. 1. (a) The product distributions during phenol hydrodeoxygenation in methanol at 473 K and 4 MPa H2. Reaction condition: Phenol (10.0 g), CH3OH (80 mL), Pd/C (1 wt.%,
0.20 g), HZSM-5 (Si/Al = 45, 0.02 g), stirring at 700 rpm. (b) Reaction pathways for phenol hydrodeoxygenation over Pd/C and HZSM-5 catalysts in methanol at 473 K.
10.0
9.0
8.0
7.0
6.0
5.0
4.0
20
16
12
8
(a)
(b)
Water
TOF = 7.8×102 h-1
R/C = 5.3× 103
mol•mol-1
Ea = 70kJ•mol-1
Methanol
Hexadecane
TOF = 1.3×103 h-1
R/C = 1.1× 104
mol•mol-1
TOF = 4.2×103 h-1
R/C = 6.8× 104
mol•mol-1
Ea = 50 kJ•mol-1
Ea = 79kJ•mol-1
Water
Methanol
4
Hexadecane
0
0
20
40
60
80
100
120
2.0
2.1
2.2
2.3
1000 × (1/T) (K-1)
Time (min)
Fig. 2. (a) TOFs on phenol hydrogenation over Pd/C in three solvents at 473 K and 4 MPa H2 at a stirring speed of 700 rpm. Note: R/C is mol (reactant) to mol (catalyst) ratio.
Reaction conditions: phenol (12.5 g), H2O (80 mL), and Pd/C (1 wt.%, 0.020 g); phenol (15.0 g), CH3OH (80 mL), and Pd/C (1 wt.%, 0.15 g); phenol (5.0 g), hexadecane (80 mL),
and Pd/C (1 wt.%, 0.10 g). (b) The TOF values for Arrhenius plots were recorded after 30 min at 433, 453, 473, and 493 K.
reacted quickly with methanol to form a series of products via the
acetal reaction (Fig. S2 in SI). Consequently, the acetal products
cyclohexyl methyl ether and dimethoxycyclohexane dominated
with selectivity exceeding 85%. However, if 1,10-dimethoxycyclo-
hexane were preferably adsorbed on Pd sites, phenol hydrogena-
tion would have been inhibited.
water (8.3 gꢂ100 mLꢁ1) at ambient temperature, but the phenol
solubility increased to completely miscible in water at 338 K. Thus,
the solubility of phenol in water at 473 K would not limit the reac-
tion rate. In the second consideration, the H2 solubility in water is
lower (7.8 ꢀ 10ꢁ7 mol cmꢁ3) than in hexadecane (4.7 ꢀ 10ꢁ6
-
mol cmꢁ3) or methanol (3.5 ꢀ 10ꢁ6 mol cmꢁ3), and the soluble H2
concentration in water was 0.0232 mol cmꢁ3 at 473 K in the pres-
ence of 4 MPa H2 [21]. The result shows that the hydrogenation
rates are more affected by the solvent properties, but not (as to
be expected) by the H2 solubility at conditions selected, which is
in accordance with the previous report for hydrogenation of aro-
matics [22].
In hexadecane, the rate of phenol hydrogenation was lowered
by two major factors. One is the low solubility of phenol in hexa-
decane (Table 3), limiting the phenol concentrations (0.26 mg Lꢁ1
at 298 K, detected by GC) close to the accessible Pd/C sites and
inducing a diffusion limitation. The second is related to the sol-
vent–Pd interaction, as the phenol hydrogenation rate decreased
with increased strength of solvent adsorption on the Pd surface.
It has been reported that the adsorption energies of hexadecane
on Pd(111) and Pt(111) were ca. 118 kJ molꢁ1 [19], while in a
comparative study, the adsorption energy of water on Pd(111)
was 50 kJ molꢁ1 [20]. This indicates again that the hexadecane–
Pd interaction is much stronger than the water–Pd interaction,
which would reduce the surface concentration of phenol on Pd.
In line with this argument, the hydrogenation rate in water is five
times that in hexadecane.
3.2.2.2. TOFs of cyclohexanone hydrogenation over Pd/C. Cyclohexa-
none reacted with methanol to form methoxycyclohexanol (Com-
pound A) as initial product through the acetal reaction (Fig. 3a
and b) and further to form 1,1-dimethoxycyclohexane (Compound
C) via Route 1 (Fig. 3b). The methoxycyclohexanol (Compound A)
intermediate can also dehydrate to cyclohexenylmethyl ether cat-
alyzed by protons in hot water (Route 2 in Fig. 3b) and hydroge-
nate to cyclohexylmethyl ether (Compound B, major product at
selectivity of 50% at 110 min). In a separate experiment on
cyclohexanol conversion in methanol using Pd/C, the ether yield
In comparison with methanol and hexadecane, water gave the
highest phenol hydrogenation rate. Phenol had low solubility in

J. He et al. / Journal of Catalysis 309 (2014) 362–375
367
Table 4
TOFs (hꢁ1) of hydrodeoxygenation of phenol in three solvents (473 K, 4 MPa H2).
Solvent
Phenol hydrogenation
Cyclohexanone hydrogenation
over Pd/C
Cyclohexanol dehydration
over HZSM-5
Cyclohexene hydrogenation
over Pd/C
over Pd/C
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
ꢀ
ꢁ
ꢁ1
ꢁ1
ꢁ1
ꢁ1
TOF1 mol mol
hꢁ1
TOF2 mol mol
hꢁ1
TOF3 mol mol
hꢁ1
TOF4 mol mol
hꢁ1
Pd surf:
Pd surf:
BAS
Pd surf:
Water
Methanol
Hexadecane
4.2 ꢀ 103
1.3 ꢀ 103
7.8 ꢀ 102
2.1 ꢀ 104
3.5 ꢀ 103
5.1 ꢀ 103
1.6 ꢀ 103
1.8 ꢀ 103
1.1 ꢀ 103
6.0 ꢀ 106
5.6 ꢀ 106
7.7 ꢀ 106
was only 0.4%, and, thus, another possible route (Route 3) via
intermolecular dehydration forming cyclohexyl methyl ether is ex-
cluded. The kinetics of cyclohexanone conversion in methanol with
Pd/C showed that conversion already attained 15% at t = 0 after the
pre-heating process (Fig. 3a), and it is concluded that this conver-
sion is due to the non-catalytic acetal reaction between cyclohex-
anol and methanol at 473 K. This is supported by the separate
blank test that the cyclohexanone already reached 13% conversion
after 15 min. In the first 10 min, this acetal reaction between cyclo-
hexanone and methanol consumed 15% of the cyclohexanone. This
would slow the rate for Step 2 over Pd/C as the cyclohexanone
hydrogenation rate had nearly first-order (n = 0.8) dependence on
the reactant concentration [23]. It is estimated that with the
lowered cyclohexanone concentration, the rate in methanol would
monomer–oligomer equilibrium is rapidly shifted toward mono-
mers to accelerate the dehydration rate [26].
The initial activity for alcohol dehydration of HZSM-5 dropped
by 60% in methanol and water after 40 min (Fig. 5a), indicating
deactivation appears in the liquid phase of high concentrations of
cyclohexanol (38.5–63.6 wt.%, see conditions in Fig. 5a). We also
found that the color of HZSM-5 changed from white to brown,
which may be caused by carbon deposition in the autoclave after
reaction in water or methanol, while the color of catalyst in hexa-
decane was unchanged. Noller and Kladnig [27] reported that the
gas-phase alcohol dehydration was accompanied by carbon depo-
sition from the oligomerization of the formed olefins in the zeolite.
Xu and Haw [28] also observed with 13C MAS NMR oligomerized
cyclohexene from cyclohexanol on HZSM-5. In fact, more carbon
was deposited from the concentrated cyclohexanol liquid phase
(38.5–63.6 wt.%) than from the gas phase after considering the dif-
ferent densities of the two phases. The decreased catalytic activi-
ties in water and methanol are mainly due to the carbon
deposition formation on the catalytic material.
The catalyst, however, did not deactivate in hexadecane, and
this may be attributed to the different stabilities of carbenium ions
(from cyclohexanol) in the three solvents. The counter ion and the
carbenium ion may have a range of associations, from a covalent
bond, tight ion pair (un-separated), to solvent-separated ion pair
(partially separated) and free ions (completely dissociated)
(Scheme 2) [29,30]. The covalent bond association is strongest
and the association of free ions is weakest [30]. In cationic poly-
merization, the ions are in equilibrium between an associated ion
pair and free ions [29]. Ion pair separation is favored by more polar
solvents. Since free ions are more reactive than ion pairs, the rate of
propagation is faster in more polar solvents. Therefore, polar water
(polarity index: 9.0) and methanol (polarity index: 6.6) yield high-
er polymerization selectivity than hexadecane (polarity index: 0.3).
The intermolecular dehydration of cyclohexanol in water
formed bicyclohexyl ether with 10% selectivity at 110 min. Cyclo-
hexanol reacted with methanol via intermolecular reaction as well.
Shown from the kinetics of cyclohexanol dehydration (Fig. 6a), the
parallel routes of intra-molecular dehydration producing cyclohex-
ene and of intermolecular dehydration forming cyclohexyl methyl
ether maintained individual selectivities of 75% and 25% during the
conversion (Fig. 6b), respectively.
be increased to 4:1 ꢀ 103 mol molPꢁd1surf: hꢁ1 over Pd/C, which is
ꢀ
slightly higher than the experimental rate
3:5 ꢀ 103 mol
ꢁ1
mol
hꢁ1Þ. Strong solvation of cyclohexanone and the Pd
Pd surf:
surface by methanol [18] further decreased the cyclohexanone
ꢀ
hydrogenation rate
experimental result : 3:5 ꢀ 103 mol molPd
surf:ꢁ1 hꢁ1; expected : 4:1 ꢀ 103 mol molPꢁd1surf: hꢁ1Þ.
In the cyclohexanone hydrogenation rate in water over Pd/C
ꢀ
ꢁ
TOF ¼ 2:1 ꢀ 104 mol molPꢁd1surf: hꢁ1 was six times that in metha-
ꢀ
ꢁ
nol TOF ¼ 3:5 ꢀ 103 mol molPꢁd1surf: hꢁ1 and four times that in
ꢀ
ꢁ
hexadecane TOF ¼ 5:3 ꢀ 103 mol molPꢁd1surf: hꢁ1 (Fig. 4a, Table 4).
The rates of cyclohexanone hydrogenation were three to seven
times those of phenol hydrogenation (Table 4), because Pd metal
sites have a stronger interaction with a ketone functional group
than with the aromatic ring. Since cyclohexanone did not populate
Pd active sites, a lower rate of cyclohexanone hydrogenation was
observed in hexadecane. This is because hexadecane has much
stronger adsorption strength than water to Pd as noted previously
[19,20].
3.2.2.3. TOFs for cyclohexanol dehydration over HZSM-5. The TOFs of
cyclohexanol dehydration (Step 3, the third individual step) were
similar at 1:1 ꢀ 103—1:8 ꢀ 103 mol molBꢁA1S hꢁ1 in water, methanol,
and hexadecane, respectively (Fig. 5a and Table 4). HZSM-5-cata-
ꢀ
ꢁ
lyzed dehydration rates in water 1:6 ꢀ 103 mol molBꢁA1S hꢁ1 were
3.2.2.4. TOFs for cyclohexene hydrogenation over Pd/C. Cyclohexene
hydrogenation rates (Step 4, the fourth individual step) were similar
in these three solvents at attaining TOFs of 5:6 ꢀ 106—7:7
ꢀ106 mol molꢁPd1surf: hꢁ1 (Fig. 7, Table 4). This is in line with results
obtained by Gonzo and Boudart [31], who also reported the absence
of a solvent effect for hydrogenation of cyclohexene on Pd/SiO2 after
studying the influence of diverse solvents of methanol, n-heptane,
ethyl acetate, and cyclohexane at 308 K.
Combining all the data obtained from kinetics of the four indi-
vidual reactions distinguishes the rate-determining step for the
phenol hydrodeoxygenation in the three solvents (Table 4). In
water, cyclohexanol dehydration (Step 3) was the slowest step.
The rate-determining step in methanol and hexadecane was
two
orders
of
magnitude
higher
than
H3PO4
did
ꢀ
ꢁ
15 mol molꢁ1 hꢁ1 [4]. The water solvent shifted the dehydra-
½Hþꢃ
tion–hydration equilibrium and inhibited dehydration. It has been
reported that Lewis acid sites are ineffective for dehydration in
aqueous phase, because water blocks Lewis active sites [24]. Con-
sequently, the Brønsted acid sites in the pores of zeolite are con-
cluded to be the active catalytic sites for dehydration. We
attribute the high dehydration rates on HZSM-5 to the unique
properties of zeolite stabilizing a favorable transition state. High
alcohol adsorption capacity even in aqueous solution [25] as well
as micropore size of HZSM-5 only allows the smaller alcohol
monomers to approach the zeolite BAS sites. Thus, the alcohol

368
J. He et al. / Journal of Catalysis 309 (2014) 362–375
60
50
40
30
20
10
0
30
(a) Dimethoxy cyclohexane
(b)
Route 3
25
Conversion
20
15
10
5
A
B
Cyclohexyl methyl ether
Route 2
Cyclohexanol
20 40
Methoxy cyclohexanol
Route 1
C
0
0
60
80
100
120
Time (min)
Fig. 3. (a) The conversion/selectivity of cyclohexanone hydrogenation in methanol at 473 K and 4 MPa H2 at a stirring speed of 700 rpm. Reaction condition: cyclohexanone
(20.0 g), CH3OH (80 mL), and Pd/C (1 wt.%, 0.010 g). (b) Reaction pathways for cyclohexanone hydrogenation over Pd/C in methanol.
Eoabs ¼ Eatrue þ ð1 ꢁ hAÞ
where Eaobs; Eatrue hA, hB,
tion energy, the true activation energy, the coverage by reactant, the
coverage by product, the reactant heat of adsorption, and the prod-
uct heat of adsorption, respectively.
The phenol hydrogenation rates were recorded in three solvents
at 433, 453, 473, and 493 K. Based on the Arrhenius law, the acti-
vation energies (Ea) were calculated to be 70, 50, and 79 kJ molꢁ1
in water, methanol, and hexadecane, respectively. Vishwanathan
and Mahata [33] obtained an apparent activation energy of
65 kJ molꢁ1 for phenol hydrogenation over Pd/MgO in the gas
phase in fair agreement of our result (Table 5). By comparison, Ea
for the phenol hydrogenation in hexadecane was similar to that
in water at 70–79 kJ molꢁ1. Ea in methanol was significantly lower
at 50 kJ molꢁ1 (Table 5). We attribute this to the strong solvation of
both phenol [16] and surface Pd [18] by methanol. Solvation would
lower the hA and hB in Eq. (1). Consequently, Eoabs in methanol
would be lower than in water or hexadecane. Moreover, the reac-
tion with methanol of the cyclohexanone generated from phenol
hydrogenation would also decrease hA and hB because of competing
by-product adsorption, and thus lower Eoabs in methanol.
D
HA ꢁ hB
D
HB
ð1Þ
phenol hydrogenation (Step 1). Based on these results, it is ex-
pected that in the use of dual-functional catalysts of Pd/C and
HZSM-5 for phenol hydrodeoxygenation in water, if not consider-
ing the competitive adsorption of the reactants in the overall kinet-
ics, the most abundant product is cyclohexanol as dehydration
(Step 3) is relatively slow compared to hydrogenation. When phe-
nol starts to be hydrogenated on Pd/C (Step 1) in hexadecane,
cyclohexane is expected to appear as the major product in a short
time, because the rates of cyclohexanone hydrogenation (Step 2),
via cyclohexanol dehydration (Step 3), to cyclohexene hydrogena-
tion (Step 4) are accelerated step by step. Rates of cyclohexanol
dehydration and cyclohexene hydrogenation were similar in three
solvents, but the rates of hydrogenation of phenol and cyclohexa-
none were much slower in methanol and hexadecane compared
to the rates in water.
D
HA, and
DHB represent the observed activa-
3.2.3. Ea of the individual steps in phenol hydrodeoxygenation
The individual steps during phenol hydrodeoxygenation can be
grouped into (i) hydrogenations of aromatic ring, ketone, and al-
kene on Pd and (ii) dehydration of cyclohexanol on HZSM-5. For
the three hydrogenation steps, it was found that the apparent acti-
vation energies for hydrogenation were similar in hexadecane and
water, but significantly lower in methanol (Table 5).
The observed activation energy equation can be expressed as
follows[32]:
The Ea of cyclohexanone hydrogenation followed the same
trend as phenol hydrogenation; hexadecane (48 kJ molꢁ1) > water
(37 kJ molꢁ1) > methanol (28 kJ molꢁ1). Ea for the hydrogenation
11.0
7
(b)
(a)
TOF = 2.1×104 h-1
R/C = 3.2 105 mol•mol-1
6
5
4
3
2
1
0
Water
10.5
Ea = 37 kJ•mol-1
TOF = 3.5×103 h-1
R/C = 2.1 105
mol•mol-1
Methanol
Hexadecane
10.0
9.5
Water
Ea = 28 kJ•mol-1
9.0
8.5
8.0
Methanol
Hexadecane
TOF = 5.1×103 h-1
Ea = 48 kJ•mol -1
R/C = 2.1 105 mol•mol-1
0
20
40
60
80
100
120
2.0
2.1
2.2
2.3
2.4
1000×(1/T) (K-1)
Time (min)
Fig. 4. (a) TOFs of cyclohexanone hydrogenation over Pd/C in three solvents at 473 K and 4 MPa H2 at a stirring speed of 700 rpm. Note: R/C stands for mol (reactant) to mol
(catalyst). Reaction conditions: cyclohexanone (60.0 g), H2O (80 mL), and Pd/C (1 wt.%, 0.020 g); cyclohexanone (20.0 g), CH3OH (80 mL), and Pd/C (1 wt.%, 0.010 g);
cyclohexanone (20.0 g), hexadecane (80 mL), and Pd/C (1 wt.%, 0.010 g). (b) The TOF values for Arrhenius plots are recorded after 30 min at 433, 453, 473, and 493 K.

J. He et al. / Journal of Catalysis 309 (2014) 362–375
369
7
6
5
4
3
2
1
0
12.0
(a)
(b)
TOF=1.1×103 h-1
R/C=4.5 × 104
mol•mol-1
10.0
8.0
Water
Methanol
Hexadecane
Ea=115 kJ •mol-1
6.0
Water
4.0
2.0
0.0
TOF=1.8×103 h-1
Ea=124 kJ•mol-1
Ea=124 kJ •mol-1
R/C=9 × 104 mol•mol-1
Methanol
Hexadecane
TOF=1.6×103 h-1
R/C=1 × 105 mol•mol-1
0
20
40
60
80
100
120
2.0
2.1
2.2
2.3
2.4
1000×(1/T) (K-1)
Time (min)
Fig. 5. (a) TOFs of cyclohexanol dehydration in three solvents at 473 K in presence of 4 MPa H2 at a stirring speed of 700 rpm. R/C stands for mol (reactant) to mol (catalyst).
Reaction conditions: cyclohexanol (140.0 g), H2O (80 mL), HZSM-5 (0.050 g); cyclohexanol (50.0 g), CH3OH (80 mL), HZSM-5 (0.020 g); cyclohexanol (50.0 g), hexadecane
(80 mL), HZSM-5 (0.040 g). (b) The TOF values for Arrhenius plots are recorded after 3 min at 433, 453, 473, and 493 K.
In the cyclohexene hydrogenation, Ea in water (25 kJ molꢁ1) and
in hexadecane (24 kJ molꢁ1) were almost the same. Ea in methanol
was only 12 kJ molꢁ1. The Ea of cyclohexene hydrogenation was
32 kJ molꢁ1 in the liquid phase over Pd/SiO2 reported by Gonzo
Scheme 2. Range of associations between the carbenium ion (R+) and anion (Xꢁ).
and Boudart [31], which was a bit higher than our result. The ad-
duct formation from methanol and cyclohexene [37] may lead to
a much lower apparent Ea4 in methanol.
reaction in methanol was significantly lower than that in water or
hexadecane (Table 5). Due to the similar hydrogenation mecha-
Summarizing the three hydrogenation reactions in the
sequence at 473 K, the activation energies decreased in the order
nisms, the strong solvation of cyclohexanone by methanol [16] (Ta-
Ea phenol hydrogenation > Ea cyclohexanone hydrogenation > Ea cyclohexene hydroge-
ble 3) and the stronger adsorption of methanol on Pd [18]
decreased the hA and hB, and resultantly lowered the apparent acti-
vation energy.
nation, while the reaction rates rose in the reverse sequence TOFphe-
hydrogenation < TOFcyclohexanone
hydrogenation < TOFcyclohexene
nol
hydrogenation. It also shows that the three hydrogenation reaction
apparent activation energies were much lower in methanol than
in water or hexadecane, apparently due to the solvent effect
including methanol–phenol and methanol–Pd interactions. The
interactions of methanol with phenol, cyclohexanone, or cyclohex-
ene altered the occupancy of Pd sites by decreasing the reactant
coverage and thus also lower the apparent activation energies
eventually. The rates and activation energies for dehydration cata-
lyzed by HZSM-5 were similar in the three solvents, indicating that
they led to similar interactions with reactant–HZSM-5 and prod-
uct–HZSM-5.
The Ea values for cyclohexanol dehydration were 115, 124, and
124 kJ molꢁ1 with HZSM-5 in water, methanol, and hexadecane,
respectively. Knözinger et al. [34] measured Ea for cyclohexanol
dehydration (110 kJ molꢁ1) over
c-Al2O3 in a continuous flow reac-
tor at low conversion (<10%). In our preliminary work, Ea for
H3PO4-catalyzed cyclohexanol dehydration was calculated to be
120 kJ molꢁ1. In the present study (Table 5), Ea for HZSM-5-cata-
lyzed cyclohexanol dehydration in water (115 kJ molꢁ1) was simi-
lar to that in hexadecane (124 kJ molꢁ1
) and in methanol
(124 kJ molꢁ1) and to Ni/ZSM-5-catalyzed cyclohexanol dehydra-
tion (Ea: 112 kJ molꢁ1) in our works [35,36].
90
6
5
4
3
2
1
0
(a)
Cyclohexene
(b)
80
70
60
50
Conv.
40
30
20
Cyclohexylmethyl ether
10
0
0
20
40
60
80
100
120
Time (min)
Fig. 6. (a) The product distribution from cyclohexanol dehydration in methanol at 473 K in presence of 4 MPa H2. Reaction condition: cyclohexanol (50.0 g), CH3OH (80 mL),
HZSM-5 (Si/Al = 45, 0.020 g). (b) Reaction route for cyclohexanol dehydration over HZSM-5 in methanol.

370
J. He et al. / Journal of Catalysis 309 (2014) 362–375
70
60
50
40
30
20
10
0
16.8
TOF = 7.7×106 h-1
(a)
(b)
Water
R/C=6.4 × 105 mol•mol-1
16.4
16.0
15.6
15.2
14.8
Ea = 24 kJ •mol-1
Methanol
Hexadecane
Ea= 12 kJ•mol-1
Ea = 25 kJ •mol-1
TOF = 5.6×106 h-1
Water
R/C=6.4× 105 mol•mol-1
Methanol
TOF = 6.0×106 h-1
R/C=6.4× 105 mol•mol-1
Hexadecane
2.0
2.1
2.2
2.3
2.4
0
1
2
3
4
5
6
1000×(1/T) (K-1)
Time (min)
Fig. 7. (a) TOFs of cyclohexene hydrogenation over Pd/C in water, methanol and hexadecane at 473 K in 4 MPa H2 at a stirring speed of 700 rpm. Note: R/C stands for mol
(reactant) to mol (catalyst). Reaction conditions: cyclohexene (50.0 g), H2O (80 mL), Pd/C (1 wt.%, 0.010 g); cyclohexene (50.0 g), CH3OH (80 mL), Pd/C (1 wt.%, 0.010 g);
cyclohexene (50.0 g), hexadecane (80 mL), Pd/C (1 wt.%, 0.010 g). (b) The TOF values for Arrhenius plots were recorded after 3 min at 433, 453, 473, and 493 K.
Table 5
Ea (kJ molꢁ1) on hydrodeoxygenation of phenol in three solvents (473 K, 4 MPa H2).
Solvent
Phenol hydrogenation over Pd/ Cyclohexanone hydrogenation over
Cyclohexanol dehydration over
Cyclohexene hydrogenation over
C
Pd/C
HZSM-5
Pd/C
Ea1
Ea2
Ea3
Ea4
Water
Methanol
Hexadecane 79
70
50
37
28
48
115
124
124
25
12
24
OH
OH
O
OH
(a)
(b)
Abs
0.20
0.15
0.10
0.05
0.0
OH
O
OH
Pd/C
Pd/C
3500
3000
2500
2000
1500
1000
Time
Wavenumber (cm-1)
0.14
0.12
0.1
100%
(c)
1228 cm-1
(d)
90%
1058 cm-1
Phenol
80%
70%
60%
50%
40%
30%
20%
10%
0%
1698 cm-1
0.08
0.06
0.04
0.02
0
Cyclohexanone
Cyclohexanol
Experimental
Data
Fitted Line
0
40
80
120
Time (min)
160
200
240
0
40
80
120
Time (min)
160
200
240
Fig. 8. (a) In situ liquid IR spectra for aqueous-phase phenol hydrogenation. Reaction conditions: phenol (10.0 g), H2O (50.0 mL), 1 wt.% Pd/C (0.10 g), 4 MPa H2, 473 K, stirred
at 700 rpm. The spectra were collected every 10 min for 240 min. (b) Reaction pathway for phenol conversion on Pd/C in the aqueous phase. (c) IR absorbance as a function of
reaction time. (d) Fitted line and experimental conversion data as a function of time based on the assumption of first-order dependence to phenol concentrations.

J. He et al. / Journal of Catalysis 309 (2014) 362–375
371
Table 6
Kinetic data for individual steps of phenol hydrodeoxygenation in the aqueous phase determined by in situ liquid IR and GC measurements (Normalized to the conditions in the
in situ liquid IR experiments).
Kinetic data
Reactions
Phenol hydrogenation
Cyclohexanone hydrogenation
Cyclohexanol dehydration
Cyclohexene hydrogenation
–
5:0 ꢀ 103 mol molPd surfꢁ1 hꢁ1
1:8 ꢀ 104 mol molPd surfꢁ1 hꢁ1
2:2 ꢀ 103 mol molBꢁA1S hꢁ1
n(Cyclohexanol): 1
In situ liquid IR TOF
:
:
Reaction order
Rate constant
GC TOF
n(Phenol): 1
n(Cyclohexanone): 0.55
–
–
9.1 ꢀ 10ꢁ3 minꢁ1
4.6 ꢀ 10ꢁ2 mol0.45 Lꢁ0.45 minꢁ1
1.1 ꢀ 10ꢁ2 minꢁ1
a
4:2 ꢀ 103 mol molPd surfꢁ1 hꢁ1
2:1 ꢀ 104 mol molPd surfꢁ1 hꢁ1
1:6 ꢀ 103 molBꢁA1S hꢁ1
n(Cyclohexanol): 1
6:0 ꢀ 106 mol molPd surfꢁ1 hꢁ1
:
:
:
Reaction order
Rate constant
n(Phenol): 1
n(Cyclohexanone): 0.55
n(Cyclohexene): 0
3.3 ꢀ 10ꢁ3 minꢁ1
4.5 ꢀ 10ꢁ2 mol0.45 Lꢁ0.45 minꢁ1
7.5 ꢀ 10ꢁ3 minꢁ1
5.0 ꢀ 10ꢁ2 mol Lꢁ1 minꢁ1
b
a
The value represents the sum of kdehydration and khydration in the aqueous phase.
The value represents the kdehydration in the aqueous and organic phases.
b
(a)
(b)
Abs
0.20
0.15
0.10
0.05
0.0
Time
3500
3000
2500
2000
1500
1000
Wavenumber (cm-1)
0.2
0.18
0.16
0.14
0.12
0.1
100%
(c)
(d)
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
1058 cm-1
Cyclohexanone
Cyclohexanol
Experimental Data
Fitted line
0.08
0.06
0.04
0.02
0
1698 cm-1
0
40
80
120
160
200
240
0
10
20
30
40
50
60
Time (min)
Time (min)
Fig. 9. (a) In situ liquid IR spectra during aqueous-phase cyclohexanone hydrogenation. Reaction conditions: cyclohexanone (10.0 g), H2O (50.0 mL), 1 wt.% Pd/C (0.10 g),
4 MPa H2, 473 K, stirred at 700 rpm. The spectra were collected every 10 min for 240 min. (b) Reaction pathway for cyclohexanone conversion on Pd/C in the aqueous phase.
(c) IR absorbance as a function of reaction time. (d) Fitted line and experimental conversion data as a function of time based on the assumption that the order dependence to
cyclohexanone concentrations is 0.55.
3.3. In situ liquid-phase IR spectroscopy for measuring the individual
steps in aqueous-phase phenol hydrodeoxygenation
and cyclohexane are immiscible in water (Table 3). With the IR dia-
mond probe on the reactor floor, three individual reactions (Steps
1, 2, and 3) were carried out in the aqueous phase at 473 K and
4.0 MPa H2.
During phenol hydrogenation over Pd/C, the intensity of the
characteristic IR peak of phenol at 1228 cmꢁ1 steadily decreased
throughout the run shown in Fig. 8a. The intensity of the peak
characteristic of cyclohexanone (partial hydrogenated from phe-
nol) at 1698 cmꢁ1 initially increased from 0 to 0.085 at 100 min
and then decreased from 0.085 to 0 at extended time up to
240 min. The intensity of the characteristic peak of cyclohexanol
at 1058 cmꢁ1 increased very slowly in the initial 100 min, because
cyclohexanol was formed by direct hydrogenation from phenol in
In order to detect the intermediates and products formed in situ
during phenol conversion, and to estimate and compare the rates
of individual steps, variations in concentrations of compounds
are monitored by attenuated total reflectance (ATR) IR spectros-
copy using diamond as a probe window for phenol hydrogenation,
cyclohexanone hydrogenation, and cyclohexanol dehydration reac-
tions in the aqueous phase. The standard IR reference spectra of
water solvent with reactant, intermediates, and products are dis-
played in Fig. S3. The IR signals of cyclohexene hydrogenation
are not observable in the aqueous phase, because cyclohexene

372
J. He et al. / Journal of Catalysis 309 (2014) 362–375
0.16
(b)
(a)
0.14
0.12
0.10
0.08
0.06
0.04
Abs
0.20
Cyclohexanol
1058 cm-1
0.15
0.10
0.05
0.0
OH
OH
0.02
0.00
0
40
80
120
160
200
240
3500 3000
2500
2000
1500
1000
Time (min)
Time
Wavenumber (cm-1)
Fig. 10. (a) In situ liquid IR spectra during aqueous-phase cyclohexanol dehydration. Reaction conditions: cyclohexanol (10.0 g), H2O (50.0 mL), HZSM-5 (0.040 g), 4 MPa H2,
473 K, stirred at 700 rpm. The spectra were collected every 10 min for 240 min. (b) IR absorbance versus reaction time.
0
(Fig. 8b). From the plotted data of IR absorbance versus reaction
Experimental data
Fitted line
time (Fig. 8c), it can be calculated that the phenol hydrogenation
rate was 5:0 ꢀ 103 mol molꢁPd1surf: hꢁ1, which is similar to that ob-
tained value from the above kinetic measurement detected by GC
-2
-4
-6
ꢀ
ꢁ
4:2 ꢀ 103 mol molPꢁd1surf: hꢁ1 . In addition, if assuming that it was
first-order to phenol concentration, the fitted curve can be per-
fectly matched with the experimental data with a rate constant
of 9.1 ꢀ 10-3 minꢁ1 (Fig. 8d and Table 6).
In the conversion of cyclohexanone hydrogenation over Pd/C
(Fig. 9a), the intensity of the characteristic IR absorbance of cyclo-
hexanone reactant at 1698 cmꢁ1 decreased rapidly and vanished at
60 min, while the intensity of characteristic IR absorbance of cyclo-
hexanol at 1058 cmꢁ1 increased quickly accordingly. The fast loss
of cyclohexanone is agreement with the high cyclohexanone
hydrogenation TOF. The rate of cyclohexanone hydrogenation
was shown to be much faster than phenol hydrogenation through
the comparison of two in situ liquid IR spectra (Figs. 8a and 9a).
y = -0.0111x -2.8731
0
40
80
120
160
200
240
Time (min)
Fig. 11. Estimation the validity of a first-order reaction rate hypothesis for the
cyclohexanol dehydration via ln([Cyclohexanol]–[Cyclohexanol]eq) versus time.
The cyclohexanone hydrogenation rate was 1:8 ꢀ 104
100
90
80
70
60
50
40
30
20
10
0
100
90
80
70
60
50
40
30
20
10
0
ꢁ1
mol mol
hꢁ1 calculated from the plots of IR absorbance versus
Pd surf:
Cyclohexane
Cyclohexanone
Cyclohexanol
Conversion
reaction time (Fig. 9c), consistent with the value 2:1ꢀ
104 mol molꢁ1 hꢁ1 monitored by GC. As shown in Fig. 9d, the
Pd surf:
cyclohexanone hydrogenation rate showed the order dependence
to cyclohexanone concentrations of 0.55, attaining the rate
constant of 4.6 ꢀ 10-2 mol0.45 Lꢁ0.45 minꢁ1
.
In a sequence of IR spectra taken during cyclohexanol dehydra-
tion in water over HZSM-5 (Fig. 10a), only the characteristic IR
band of cyclohexanol varied, because the cyclohexene floated on
the aqueous phase and was not detected by the IR detector in
the bottom of the autoclave. The intensity of the cyclohexanol
absorbance at 1058 cmꢁ1 decreased more slowly than the former
two individual reactions, in accordance with the kinetic measure-
ment that dehydration is the rate-determining step in aqueous-
phase phenol hydrodeoxygenation. Determined from the IR absor-
bance versus reaction time (Fig.ꢁ11 0b), the cyclohexanol dehydra-
tion rate was 2:2 ꢀ 103 mol molBAS hꢁ1, agreeing with the rate of
1:6 ꢀ 103mol molBꢁA1Shꢁ1 derived from chromatographic evidence.
The dehydration–hydration reaction of cyclohexanol is equili-
brated over HZSM-5 in the aqueous phase as Eq. (2):
0
20 40 60 80 100 120 140 160 180 200
Time (min)
Fig. 12. The product distributions during phenol hydrodeoxygenation in the
aqueous phase in the presence of Pd/C and HZSM-5. Reaction conditions: Phenol
(1.0 g), 1 wt.% PdC (0.040 g), HZSM-5 (0.20 g), H2O (80 mL), 473 K, 4 MPa H2, stirring
at 700 rpm.
the beginning. Then, the intensity rose more steeply from 100 to
240 min, as both phenol and cyclohexanone were converted to
cyclohexanol at this time interval. This interesting kinetic perfor-
mance is consistent with step-wise hydrogenation of phenol to
cyclohexanone and subsequent hydrogenation to cyclohexanol
ð2Þ
For the first-order reversible reaction with excess water:

J. He et al. / Journal of Catalysis 309 (2014) 362–375
373
d½Aꢃ
dt
phenol and cyclohexanone hydrogenation were comparable.
Phenol hydrogenation attained a rate constant of 9.1 ꢀ 10ꢁ3 and
3.3 ꢀ 10ꢁ3 minꢁ1 as measured by in situ liquid IR spectroscopy
and GC, respectively, while cyclohexanone hydrogenation showed
a rate constant of 4.6 ꢀ 10ꢁ2 and 4.5 ꢀ 10ꢁ2 mol0.45 Lꢁ0.45 minꢁ1
from these two techniques. It is notable that the dehydration rate
constants were quite different from two approaches, because
in situ liquid-phase IR spectroscopy estimates the sum of kdehydra-
ꢁ
¼ k1½Aꢃ ꢁ k ½Bꢃ
ð3Þ
ð4Þ
ꢁ1
½Bꢃ ¼ ½Bꢃ0 þ ð½Aꢃ0 ꢁ ½AꢃÞ
d½Aꢃ
So ꢁ
¼ k1½Aꢃ ꢁ kꢁ1ð½Bꢃ0 þ ð½Aꢃ0 ꢁ ½AꢃÞÞ
ð5Þ
ð6Þ
dt
and khydration (1.1 ꢀ 10ꢁ2 minꢁ1) in the aqueous phase, while
d½Aꢃ
dt
k
ꢁ1
k1 þk
tion
At equilibrium;ꢁ
¼ 0 and ½Aꢃeq
¼
ð½Bꢃ0 þ½Aꢃ0Þ
ꢁ1
the GC analysis determines kdehydration (7.5 ꢀ 10ꢁ3 minꢁ1) in the
sum of aqueous and organic phases. The rate constant for cyclo-
hexene hydrogenation was calculated to be 5.0 ꢀ 10ꢁ2 mol Lꢁ1
-
dð½Aꢃ ꢁ ½Aꢃeq
dt
Þ
d½Aꢃ
minꢁ1 from GC measurement, as the reaction order toward
cyclohexene concentration is assumed to be 0 [31].
ꢁ
¼ ꢁ
¼ ðk1 þ kꢁ1Þð½Aꢃ ꢁ ½AꢃeqÞ
ð7Þ
ð8Þ
dt
In a comparative study, we also performed the overall phenol
hydrodeoxygenation with dual Pd/C and HZSM-5 components in
the aqueous phase in order to demonstrate the cooperative effect
of acid and metal sites (Fig. 12). Phenol was quantitatively con-
verted to cyclohexane at 473 K and 4 MPa H2 along the reaction
routes discussed above. Due to the fast hydrogenation of cyclo-
hexene intermediate on Pd/C, it was not observed in the presence
of the catalyst mixture. This result agrees well with our previous
studies on hydrodeoxygenation of phenol monomers and dimers
as well as bio-oil with bifunctional catalysts consisting of Pd and
acids [4–6]. The elementary steps in the overall phenol
hydrodeoxygenation include phenol hydrogenation to produce
cyclohexanone (k1), subsequently cyclohexanone hydrogenation
to afford cyclohexanol (k2) which is in an equilibrium with
cyclohexanol dehydrogenation to cyclohexanone (kꢁ2), in turn
cyclohexanol is dehydrated and hydrogenated to cyclohexane
(k3) (Scheme 3).
For hydrogenation of phenol and cyclohexanone over Pd/C
with varying hydrogen pressures of 3, 4, and 5 MPa, the reaction
orders of hydrogen concentration for phenol and cyclohexanone
hydrogenation were determined to be 1.5 and 0.6, respectively
(see Figs. 13 and 14). It was observed that the hydrogen pressure
before and after reaction kept unchanged; thus, the hydrogen
pressures were considered as constant values in the elementary
rate equations.
½Aꢃ ꢁ ½Aꢃeq ¼ ð½Aꢃ0 ꢁ ½AꢃeqÞeꢁðk
1þkꢁ1Þt
lnð½Aꢃ ꢁ ½Aꢃeq
ꢁðk1 þ kꢁ1Þ
Þ
lnð½Aꢃ0 ꢁ ½Aꢃeq
ðk1 þ kꢁ1Þ
Þ
t ¼
þ
ð9Þ
As shown in Fig. 11, cyclohexanol dehydration was determined to
be a first-order reaction with HZSM-5 at 473 K in the aqueous
phase, and k1 + k equaled to 1.1 ꢀ 10-2 minꢁ1
.
ꢁ1
The rate constants, reaction orders, and TOFs of three individual
steps in the aqueous phase, as determined by GC and in situ liquid
IR detectors, are compiled in Table 6. The reaction orders in the
aqueous solution have been identified by in situ liquid IR spectros-
copy, that is, phenol and cyclohexanone hydrogenation over Pd/C
followed reaction orders of 1.0 and 0.55, respectively. Conversion
of cyclohexanol with HZSM-5 shows first-order dependence in
the dehydration–hydration equilibrium. The TOFs of three aque-
ous-phase individual steps determined from GC measurement
were very close to the values from in situ liquid IR, with minor
differences of 10–20%. As shown in Table 6, the rate constants for
OH
O
OH
k2
k1
k3
k-2
4
It is assumed first-order reaction to phenol, cyclohexanone, and
cyclohexanol concentrations and zero-order reaction to cyclohex-
ene concentration. Therefore, the overall phenol hydrodeoxygen-
ation includes the following elementary rate equations:
3
2
1
Scheme 3. Proposed elementary reaction steps in the overall phenol
hydrodeoxygenation.
9.0
10
(a) Conv. Versus Time
(b) Ln (TOF) Versus Ln (P)
9
8.8
8
5 Mpa Hydrogen
MPa
y = 1.5469x + 6.2296
8.6
8.4
8.2
8.0
7.8
7.6
7
6
5
4
3
2
1
0
4 M Hydrogen
MPa
pa
MPa
3 Mpa hydrogen
1
1.2
1.4
1.6
1.8
0
30
60
Time (min)
90
120
Ln (P (MPa))
Fig. 13. Hydrogenation of phenol over Pd/C with varying hydrogen pressures. Reaction conditions: phenol (12.5 g), H2O (80 mL), 1 wt.% Pd/C (0.02 g), 473 K, stirring at
700 rpm.

374
J. He et al. / Journal of Catalysis 309 (2014) 362–375
9
8
7
6
5
4
3
2
1
0
10.2
(b) Ln (TOF) Versus Ln (P)
(a) Conv. Versus Time
5 Mpa hydrogen
MPa
4 M
MPa
10.1
10.0
9.9
M
p
P
a
a
hydrogen
y = 0.6201x + 9.0796
3 Mpa hydrogen
9.8
9.7
0
20
40
60
80
100
120
1
1.2
1.4
1.6
1.8
Time (min)
Ln (P (MPa))
Fig. 14. Hydrogenation of cyclohexanone over Pd/C with varying hydrogen pressures. Reaction conditions: cyclohexanone (60.0 g), H2O (80 mL), 1 wt.% Pd/C (0.02 g), 473 K,
stirring at 700 rpm.
Table 7
Comparison of fitted rate constants in the overall phenol hydrodeoxygenation and in the individual steps analyzed by GC measurements (conditions: 1.0 g phenol, 0.04 g Pd/C,
0.20 g HZSM-5, 473 K, 4 MPa H2).
Reaction rate constants
k1 (minꢁ1
)
k2 (minꢁ1
)
k
ꢁ2
(minꢁ1
)
k3 (minꢁ1
)
Fitness parameter
Fitted elementary steps in the overall HDO
Data from individual steps analyzed by GCa
9.0 ꢀ 10ꢁ2
3.0 ꢀ 10ꢁ1
5.4
n.d.
4.0 ꢀ 10ꢁ1
0.46
–
1.4 ꢀ 10ꢁ2
1.8 ꢀ 10ꢁ1
3.8 ꢀ 10ꢁ1
a
Normalized to the conditions in the overall phenol hydrodeoxygenation experiment.
Fitting the pseudo-rate constants of the individual steps of
phenol hydrodeoxygenation (Table 7 and Fig. 15), the fitted phe-
nol hydrogenation rate constant (9.0 ꢀ 10ꢁ2 minꢁ1) was much
higher than the value (1.4 ꢀ 10ꢁ2 minꢁ1) deduced as the first step
in individual measurements, probably due to the added acid
effect for accelerating hydrogenation rates in the overall phe-
nol hydrodeoxygenation. But fitted apparent rate constants
(3.0 ꢀ 10ꢁ1 minꢁ1) in the subsequent cyclohexanone hydrogena-
tion elementary steps varied from that determined in individual
steps (1.8 ꢀ 10ꢁ1 minꢁ1) (Table 7) due to strong competitive
adsorption of phenol and cyclohexanone onto active sites during
the overall phenol hydrodeoxygenation. The fitted cyclohexanol
dehydration rate constant (4.0 ꢀ 10ꢁ1 minꢁ1) was slightly higher
than the value (3.8 ꢀ 10-1 min-1) analyzed by the individual
measurement, because the dehydration–hydration equilibrium
is shifted by the subsequent cyclohexene hydrogenation over
combined Pd/C and HZSM-5 catalysts.
1.0
0.8
Phenol
Cyclohexanone
Cyclohexanol
0.6
Cyclohexane
0.4
0.2
0.0
0
40
80
120
160
200
Time (min)
Fig. 15. Fitting the experimental data of phenol hydrodeoxygenation with Pd/C
and HZSM-5 catalysts in presence of hydrogen, solid points: experiment data,
line: fitted curve. Conditions: phenol (1.0 g), 1 wt.% Pd/C (0.04 g), HZSM-5 (0.20 g),
473 K, 4 MPa H2.
4. Conclusion
Solvents influence the rates of catalytic phenol hydrodeoxygen-
ation by altering the H2 solubility and the solvent–reactant interac-
tions, as well as by competitive solvent/reactant adsorption on the
catalyst surface. Impacts of water, methanol, and hexadecane sol-
vents on the individual steps of phenol hydrodeoxygenation are
investigated over Pd/C or HZSM-5 catalyst at 473 K in the presence
of hydrogen. Hydrodeoxygenation of phenol to cyclohexane in-
cludes four individual steps of phenol hydrogenation to cyclohex-
anone on Pd/C, cyclohexanone hydrogenation to cyclohexanol on
Pd/C, cyclohexanol dehydration to cyclohexene on HZSM-5, and
cyclohexene hydrogenation to cyclohexane on Pd/C. Methanol pro-
vides an additional acetal reaction route of cyclohexanone with
methanol during phenol hydrodeoxygenation, compared to water
and hexadecane. The hydrogenation reactions, especially of phenol
dc1
¼ ꢁk1 ꢀ c1
ð10Þ
ð11Þ
ð12Þ
ð13Þ
dt
dc2
dt
¼ k1 ꢀ c1 ꢁ k2 ꢀ c2 þ kꢁ2 ꢀ c2
¼ k2 ꢀ c2 ꢁ kꢁ2 ꢀ c2 ꢁ k3 ꢀ c3 ꢀ cðHþÞ
¼ k3 ꢀ c3 ꢀ cðHþÞ
dc3
dt
dc4
dt

J. He et al. / Journal of Catalysis 309 (2014) 362–375
375
and cyclohexanone, show lower rates in methanol than in water or
References
hexadecane. Methanol not only strongly solvates phenol and cyclo-
hexanone but also adsorbs strongly on Pd. This strong adsorption
either prevents phenol/cyclohexanone from from interacting with
Pd active sites or reduces the phenol/cyclohexanone coverage cov-
erage on the Pd/C surface, affording slower phenol/cyclohexanone
hydrogenation in methanol. The low solubility of phenol and
strong interaction of hexadecane–Pd surface in hexadecane lead
to the slow hydrogenation rate.
[1] E. Furimsky, Appl. Catal. A 199 (2000) 147.
[2] C. Zhao, Y. Kou, A.A. Lemonidou, X. Li, J.A. Lercher, Angew. Chem. Int. Ed. 48
(2009) 3987.
[3] C. Zhao, J.A. Lercher, Angew. Chem. Int. Ed. 51 (2012) 5935.
[4] C. Zhao, J. He, A.A. Lemonidou, X. Li, J.A. Lercher, J. Catal. 280 (2011) 8.
[5] C. Zhao, D.M. Camaioni, J.A. Lercher, J. Catal. 288 (2012) 92.
[6] C. Zhao, J.A. Lercher, ChemCatChem 4 (2012) 64.
[7] J. He, C. Zhao, J.A. Lercher, J. Am. Chem. Soc. 134 (2012) 20768.
[8] J.S. Shabtai, W.W. Zmierczak, E. Chornet, US Patent, The University of Utah
Research Foundation, 6,172,272 B1, 2001.
[9] C.-L. Li, Z.-R. Xu, Z.-A. Cao, B.C. Gates, AIChE J. 31 (1985) 170.
[10] Y.Q. Yang, C.T. Tye, K.J. Smith, Catal. Commun. 9 (2008) 1364.
[11] L.R. Snyder, J. Chromatogr. 92 (1974) 223.
[12] U.K. Singh, M.A. Vannice, Appl. Catal. A 213 (2001) 1.
[13] M. Boudart, D.J. Sajkowski, Faraday Discuss. 92 (1991) 57.
[14] Y. Xiang, L. Ma, C. Lu, Q. Zhang, X. Li, Green Chem. 10 (2008) 939.
[15] J.S. Shabtai, W.W. Zmierczak, E. Chornet, US Patent, The University of Utah
Research Foundation, 5,959,167, 1999.
The apparent activation energies for hydrogenation show the
order Ea phenol hydrogenation > Ea cyclohexanone hydrogenation > Ea cyclohexene
hydrogenation, and the reaction rates follow TOFcyclohexene
hydrogena-
tion > TOFcyclohexanone hydrogenation > TOFphenol
at 473 K.
hydrogenation
The in situ liquid IR spectroscopy results confirm that the rates
are consistent with kinetics derived from chromatographic evi-
dence in the aqueous phase and verify that hydrogenation of phe-
nol and cyclohexanone follows reaction orders of 1.0 and 0.55 over
Pd/C, respectively. Conversion of cyclohexanol with HZSM-5 shows
first-order dependence in the dehydration–hydration equilibrium.
The dehydration rates and apparent activation energies are compa-
rable in three solvents. The control of the balance between metal
and acid sites is crucial to optimize the cascade reactions consist-
ing of the integrated hydrogenation and dehydration steps, which
allows developing active and stable catalysts.
[16] N.M. Bertero, A.F. Trasarti, C.R. Apesteguia, A.J. Marchi, Appl. Catal. A 394
(2011) 228.
[17] N. Mahata, K.V. Raghavan, V. Vishwanathan, Catal. Today 49 (1999) 65.
[18] B.A. Sexton, K.D. Rendulic, A.E. Hughes, Surf. Sci. 121 (1982) 181.
[19] Y. Morikawa, H. Ishii, K. Seki, Phys. Rev. B 69 (2004) 041403.
[20] S. Andersson, C. Nyberg, C.G. Tengstål, Chem. Phys. Lett. 104 (1984) 305.
[21] H.A. Pray, C.E. Schweickert, B.H. Minnich, Ind. Eng. Chem. 44 (1952) 1146.
[22] A. Drelinkiewicza, A. Waksmundzka, W. Makowski, J.W. Sobczak, A. Król, A.
Zieba, Catal. Lett. 94 (2004) 143.
[23] K.A. Manbeck, N.E. Musselwhite, L.M. Carl, C.A. Kauffman, O.D. Lyons, J.K.
Navin, A.L. Marsh, Appl. Catal. A 384 (2010) 58.
[24] B. Peng, C. Zhao, I. Mejía-Centenoa, G.A. Fuentesa, A. Jentys, J.A. Lercher, Catal.
Today 183 (2012) 3.
[25] R. Krishna, J.M. van Baten, Langmuir 26 (2010) 10854.
[26] J. Macht, M.J. Janik, M. Neurock, E. Iglesia, Angew. Chem. Int. Ed. 46 (2007)
7864.
Acknowledgments
J.H. gratefully acknowledges the support of the TUM graduate
school’s faculty graduate center of chemistry (FGCh) at the Techni-
sche Universität München, the Elitenetzwerk Bayern (graduate
school NanoCat). (graduate school NanoCat). M.Sc. Stanislav Kasa-
kov is acknowledged for fitting the experimetal data of overall phe-
nol hydrodeoxygenation. C.Z. thanks the support from European
Graduate School for Sustainable Energy. J.A.L. acknowledges the
support from the US Department of Energy, Office of Basic Energy
Sciences. Pacific Northwest National Laboratory (PNNL) is a multi-
program national laboratory operated for DOE by Battelle.
[27] H. Noller, W. Kladnig, Catal. Rev. 13 (1976) 149.
[28] T. Xu, J.F. Haw, J. Am. Chem. Soc. 116 (1994) 7753.
[29] G. Odian, Principles of Polymerization, fourth ed., Wiley-Interscience,
Hoboken, NJ, 2004.
[30] J.M.G. Cowie, V. Arrighi, Polymers Chemistry and Physics of Modern Materials,
third ed., Taylor & Francis, Boca Raton, 2008.
[31] E.E. Gonzo, M. Boudart, J. Catal. 52 (1978) 462.
[32] F. Kapteijn, J.A. Moulijn, R.A. van Santen, R. Wever, Stud. Surf. Sci. Catal. 123
(1999) 81.
[33] N. Mahata, V. Vishwanathan, J. Mol. Catal. A 120 (1997) 267.
[34] H. Knözinger, K. Kochloef, H. Bühl, J. Catal. 24 (1972) 57.
[35] C. Zhao, Y. Yu, A. Jentys, J.A. Lercher, Appl. Catal. B 132–133 (2013) 282.
[36] C. Zhao, S. Kasakov, J. He, J.A. Lercher, J. Catal. 296 (2012) 12.
[37] R. West, J. Am. Chem. Soc. 81 (1959) 1614.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.09.009.