Gas-phase, catalytic hydrodeoxygenation of propanoic acid, over supported group VIII noble metals: Metal and support effects

Article abstract of DOI:10.1016/j.apcata.2013.10.025

The catalytic, gas-phase hydrodeoxygenation (HDO) of propanoic acid (PAc) over supported group VIII noble metals (M = Pd, Pt, Rh, Ru, Ni) was studied at 1 atm and 200-400 C. The activity and selectivity as a function of the reaction temperature was investigated. The reaction activity based on the TOF follows the order: Pd > Ru > Pt > Rh > Ni. The reaction over Pd, Pt and Rh catalysts proceeds mainly via decarbonylation (DCN) and decarboxylation (DCX) pathways at each reaction temperature. For Ru and Ni catalysts, while decarbonylation and decarboxylation pathways were predominant at lower temperatures (e.g., 200-250 C), at higher temperatures (>300 C) the formation of diethyl ketone was observed. Additionally, the kinetics of Pd over different supports (carbon, SiO₂ and TiO₂) were examined. The activity based on the TOF decreases in the following order: Pd/SiO₂ > Pd/TiO₂ > Pd/C. The reaction orders in acid and H₂ were found to be approximately 0.5 and zero, respectively, regardless of the support. The apparent activation energies studied in a temperature range of 200-240 C, were 16.7 ± 0.6, 19.3 ± 1.6 and 11.7 ± 0.7 kcal/mole for Pd/C, Pd/TiO₂ and Pd/SiO₂ catalysts, respectively. The selectivity for Pd/C and Pd/SiO₂ indicated mainly decarbonylation/decarboxylation and hydrogenation reaction pathways. In contrast, Pd/TiO₂ at low temperatures (200 C) could generate decarbonylation and esterification products.

Full text of DOI:10.1016/j.apcata.2013.10.025

Applied Catalysis A: General 469 (2014) 410–418
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Gas-phase, catalytic hydrodeoxygenation of propanoic acid, over
supported group VIII noble metals: Metal and support effects
Yuliana K. Lugo-José, John R. Monnier, Christopher T. Williams∗
Department of Chemical Engineering, University of South Carolina, Columbia 29208, SC, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 July 2013
Received in revised form 9 October 2013
Accepted 14 October 2013
Available online xxx
The catalytic, gas-phase hydrodeoxygenation (HDO) of propanoic acid (PAc) over supported group VIII
noble metals (M = Pd, Pt, Rh, Ru, Ni) was studied at 1 atm and 200-400 ◦C. The activity and selectiv-
ity as a function of the reaction temperature was investigated. The reaction activity based on the TOF
follows the order: Pd > Ru > Pt > Rh > Ni. The reaction over Pd, Pt and Rh catalysts proceeds mainly via
decarbonylation (DCN) and decarboxylation (DCX) pathways at each reaction temperature. For Ru and
Ni catalysts, while decarbonylation and decarboxylation pathways were predominant at lower tempera-
tures (e.g., 200–250 ◦C), at higher temperatures (>300 ◦C) the formation of diethyl ketone was observed.
Additionally, the kinetics of Pd over different supports (carbon, SiO2 and TiO2) were examined. The activ-
ity based on the TOF decreases in the following order: Pd/SiO2 > Pd/TiO2 > Pd/C. The reaction orders in
acid and H2 were found to be approximately 0.5 and zero, respectively, regardless of the support. The
apparent activation energies studied in a temperature range of 200–240 ◦C, were 16.7 0.6, 19.3 1.6
and 11.7 0.7 kcal/mole for Pd/C, Pd/TiO2 and Pd/SiO2 catalysts, respectively. The selectivity for Pd/C
and Pd/SiO2 indicated mainly decarbonylation/decarboxylation and hydrogenation reaction pathways.
In contrast, Pd/TiO2 at low temperatures (200 ◦C) could generate decarbonylation and esterification
products.
Keywords:
Biomass
Hydrodeoxygenation
Decarbonylation
Decarboxylation
Hydrogenation
Propanoic acid
Palladium catalyst
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
C:O ratio approaches unity. However, the fuel and chemical prod-
ucts derived from such feedstocks must have an oxygen content
The demand of energy is increasing constantly and the United
States is dependent on foreign sources for about half of its fuel
consumption [1]. The combustion of fossil fuels releases carbon
dioxide, a major greenhouse gas, into the atmosphere contribut-
ing to the problem of global warming. This factor, along with the
increasing demand for fossil fuels coupled with the decreasing
availability of those resources, has prompted a move towards alter-
native energy production and utilization [2,3]. New research efforts
promise to bring advances in a variety of areas, and biomass con-
version to fuels and chemicals will play a critical role in the future.
Industrial processes for converting biomass into biofuels include
gasification to syngas (leading to liquid fuels through FT synthe-
sis), pyrolysis to produce bio-oil, and hydroprocessing to remove
oxygen functionality [4–6]. During these processes, carbon diox-
ide is released as well; however, the emission gases are balanced
largely by the carbon dioxide that is captured during the growth of
the biomass, depending upon how much energy was used to grow,
harvest and process the fuel [7]. Feedstocks derived from lignocel-
lulosic substances contain great amounts of oxygen such that the
less than about 5.0% to be commercially viable [8,9]. Heteroge-
neous, catalytic deoxygenation (including hydrodeoxygenation)
is one of the most promising processes for generating hydro-
carbon fuels from biomass. During hydrodeoxygenation (HDO),
hydrogen is co-fed with the reactant to facilitate hydrogenolysis
(HYS) and hydrogenation reactions required for the removal of
oxygen.
During the 1980s, Maier et al. studied the gas-phase decarbox-
ylation of short chain carboxylic acids to hydrocarbons using Ni and
Pd catalysts [10]. However, most of this work focused on C–C bond
scission rather than removal or hydrogenation of the C=O bond.
Since then, increasing amounts of research have been conducted
in the area of biomass-derived platform molecules such as organic
acids and esters (e.g., palmitic, oleic and stearic acid, ethyl stearate
and tristearine) via decarboxylation and hydrogenation reactions
on monometallic and bimetallic catalysts [11–28]. For example,
Murzin and co-workers have focused on biomass- derived, liquid-
phase hydrogenation on supported palladium catalysts [29], while
Resasco and Lobban reported the direct conversion of triglycerides
to olefins and paraffins over Pt/Al2O3, Pt/SiO2, and PtSnK/SiO2 cata-
lysts [30]. Additionally, Kozhevnikov et al. recently reported studies
on heteropoly acid and bifunctional metal-loaded heteropoly acid
catalysts for the deoxygenation of PAc [31].
∗
Corresponding author. Tel.: +1 803 777 0143; fax: +1 803 777 8265.
E-mail address: willia84@cec.sc.edu (C.T. Williams).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.10.025

Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
411
Table 1
Physical and chemical properties of M/SiO2.
Metal loading (wt%)
Catalyst
Mx/SiO2
Calcination
temp. (◦C)
Reduction temp. (◦C)
100% H2
Dispersion (%)
O2–H2 titr.
Particle
diameter (nm)
1.1
4.0
2.1
2.2
1.6
1.8
Pd
Pd*
Pt
Rh
Ru
Ni
300
–
–
–
300
450
350
350
300
300
300
450
3.9
55.0
11.7
21.0
11.7
4.8
31.4
2.0
9.7
5.3
11.4
22.6
*
Pd/SiO2 synthesized by SEA.
Eqs. (1)–(6) illustrate the hydrogenation, decarbonylation, and
decarboxylation pathways proposed for the HDO of organic acids
and esters. In addition, the CO2 and CO products from the decar-
boxylation (DCX) and decarbonylation (DCN) routes can undergo
further reaction through water–gas shift and methanation, con-
suming large amount of H2.
precursor and the –OH group on the support took place, allowing
a controlled impregnation with highly dispersed metal precursors
[34–36]. The catalysts prepared by SEA or incipient wetness were
dried overnight at 70 ◦C, followed by calcination in air and reduction
with 100% H2 at their respective temperatures for 2 h (see Table 1).
The exception was the 4 wt% Pd/SiO2, which was not subjected to
a calcination step due to the fact that the 1.1 wt% Pd/SiO2 catalyst
exhibited very large particle size. The commercial catalysts were
already prepared, and therefore it was decided that no calcination
step was required.
Hydrogenation : R1 − COO − R2 + 4H2 ↔ R1 − CH3
+ R2 − H + 2H2O
(1)
(2)
R − COOH + 3H2 ↔ R − CH3 + 2H2O
2.2. Catalyst synthesis (Pd/X)
Decarbonylation : R1 − COO − R2 + 2H2 ↔ R1 − H
+ R2 − H + CO + H2O
Three different palladium catalysts were investigated for the
detailed kinetics of the HDO of PAc: 4.0 wt% Pd/SiO2 (Aerosil 300,
SBET = 330 m2/g, Evonik), 5.0 wt% Pd/C (CP-97, SBET = 615 m2/g, com-
mercial BASF catalyst) and 2.3 wt% Pd/TiO2 (TiO2, SBET = 46.1 m2/g,
Evonik). The Pd/SiO2 and Pd/TiO2 catalysts were synthesized by
SEA, while the commercial Pd/C was supplied by BASF and followed
the same procedure described in the previous section.
(3)
(4)
R − COOH + H2 ↔ R − H + CO + H2O
Decarboxylation : R1 − COO − R2 + H2 ↔ R1 − H
+ R2 − H + CO2
(5)
(6)
R − COOH ↔ R − H + CO2
3. Pulsed H2 chemisorption
The present work aims to explore the catalytic chemistry of
the HDO of PAc (a model compound for aliphatic organic acids)
over supported group VIII noble metal catalysts. The monometallic
catalysts have been selected based on their potential to catalyze
different bond breaking events relevant to the above reactions. The
catalysts include Pd, Pt, Ru, Rh and Ni supported on SiO2. Addi-
tionally, Pd was screened on different supports (i.e. carbon, SiO2,
and TiO2) to investigate the influence of the support material on
activity, selectivity and activation energy. These studies provide
insight on the different reaction pathways to form paraffins/olefins
or oxygenated products from biomass derived organic acids.
The metal dispersions of the catalysts were determined by
pulsed hydrogen titration of oxygen pre-covered sites using a
Micromeritics 2920 AutoChem II Analyzer. Prior to the analysis,
the catalysts (0.1–0.2 g) were reduced in flowing 100% H2 for 2 h,
followed by purging with Ar for 2 h to remove any physisorbed
hydrogen. After cooling to 40 ◦C in flowing Ar, the catalysts were
exposed to 10% O2/He for 30 min followed by purging with Ar
(30 min). Titration with pulses of 10% H2/Ar was then employed
until no further H2 uptake occurred. This method was applied for
Pd, Pt, and Rh based catalysts. For Ru and Ni, a modified hydrogen
chemisorption method was implemented. The sample experienced
the same pretreatment described previously; however, no H2-
chemisorption at 40 ◦C was done. After being exposed to argon,
the sample was heated up to 100 ◦C by temperature programmed
oxidation, following by flushing with argon for 30 min. Then, the
sample temperature was raised up to 225 ◦C and exposed to 10%
H2/Ar. This titration temperature was confirmed based on the
results of temperature programmed oxidation/reduction described
in the Supplemental Information.
2. Experimental
2.1. Catalyst synthesis (M/SiO2)
For the screening reaction on M/SiO2 (M = Pd, Ru, Ni), the cat-
alysts were synthesized by incipient wetness utilizing aqueous
solutions of the following metal salts: palladium nitrate hydrate
(Pd(NO3)2·xH2O, 99.9 Sigma-Aldrich), ruthenium nitrosyl nitrate
(Ru(III)(NO)(NO3)3, 31.3% min Alfa Aesar) and nickel perchlorate
hexahydrate (Ni(II)(ClO4)2·6H2O, 99.99% Alfa Aesar). The SiO2 sup-
port (Silica Star, SBET = 100 m2/g) and the commercial catalysts of
2.14 wt% Pt and 2.06 wt% Rh over SiO2, were obtained from BASF.
In addition, a 4.0 wt% Pd catalyst with a different SiO2 support
(Aerosil 300, SBET = 330 m2/g, Evonik) was synthesized by strong
electrostatic adsorption (SEA), where the pH of the metal salt com-
plex solution (200 ppm of [Pd(NH3)4]2+Cl2, 99.9% Sigma Aldrich)
was controlled based on the PZC of the SiO2 [32,33]. Once the
pH of the solution is acquired, the support was impregnated and
shaken for 1 h. After the final pH was obtained, the difference (ꢀpH)
indicated that a strong electrostatic interaction between the metal
3.1. Catalyst evaluation
The activity and selectivity of each catalyst was measured for
the HDO of PAc. The experiments were carried out in a single pass,
packed bed, plug flow reactor system. The catalysts were reduced
in-situ at temperatures ranging from 200 to 400 ◦C at a total flow
of 50 sccm H2 for 2 h at 1 atm pressure. The reduction tempera-
ture was chosen depending on the different reduction treatments
required for each metal (Table 1). The feed stream for the screening
experiments consisted of 1.0% PAc (Alfa Aesar, 99%) and 20% H2,

412
Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
Fig. 1. Schematic of the vapor phase continuous plug-flow reactor system.
balances He. For the kinetic experiments, PAc and H2 concentra-
tions ranged from XPAc ∼ 0.0020–0.017 and XH2 ∼ 0.20–0.99. Fig. 1
shows a schematic of the plug flow reactor system including the
Brooks mass flow controllers (MFC), vapor–liquid equilibrium sat-
urator (VLE), switching valve box, and the GC inlet. PAc (Alfa Aesar,
97%) was used as received. The gas feed and product lines (going
through the valve and reactor) were heated >100 ◦C, to prevent acid
and product condensation. Reaction products were analyzed by an
on-line Hewlett-Packard 5890 gas chromatograph equipped with
two Poraplot Q capillary columns; one was linked to a flame ioniza-
tion detector (FID) for analysis of organic products and unreacted
PAc, while the second column was connected to a thermal con-
ductivity detector for analysis of permanent gases such as CO and
CO2. Carbon monoxide hydrogenation experiments were also con-
ducted for two M/SiO2 (M = Ru, Ni) catalysts to explore their degree
of methanation activity. The concentration of carbon monoxide was
maintained at 1.0% to simulate what might be encountered dur-
ing the HDO reaction based on the amount of PAc that was being
fed. The conversion and selectivity for a given catalyst was defined
according to the following equations:
where PAcin and PAcout denote the concentrations of PAc before
and after the reaction, and ꢁi, Pi are the stoichiometric factor and
product concentration, respectively.
4. Results and discussion
4.1. Hydrodeoxygenation of PAc over M/SiO2: Effect of reaction
temperature
The effect of reaction temperature on the hydrodeoxygenation
activity of PAc was investigated over the range 200–400 ◦C. The
reactions are classified, based on the formation of main products,
as decarbonylation/decarboxylation for hydrocarbon products and
hydrogenation and ketonization (for oxygenated products). Prior
to examining the HDO of PAc on various SiO2-supported group VIII
noble metals, the activity and selectivity over two different SiO2
supports were explored. Fig. 2a and 2b shows the conversion and
selectivity as a function of temperature for Silica Star and Aerosil
300. These supports were capable of catalyzing the ketonization
reaction due to the acid properties of the silica surface [37]. Reac-
tion over Silica Star resulted in the formation of diethyl ketone
(DEK) at temperatures above 300 ◦C around 92% selectivity. The
remaining 8% of products consisted of small amounts of ethane
and propionaldehyde. The reaction over SiO2 Aerosil 300, resulted
primarily in the formation of DEK (73% selective), but also ethyl-
ene (27% selective) at temperatures above 350 ◦C, suggesting the
Aerosil 300 has lower surface acidity. Pestman and co-workers
have reported reactions of carboxylic acids over metal oxides (i.e.,
Al2O3, Fe2O3, SiO2) [38,39]. According to their work, a bimolecular
ketonization reaction was observed that produced DEK along with
ꢀ
ꢁ
PAcin − PAcout
Conversion =
× 100
PAcin
⎛
⎞
⎜
⎜
⎜
⎜
⎝
⎟
⎟
Pi/ꢁi
⎟
Selectivity =
× 100
n
⎟
⎠
ꢂ
Pi/ꢁi
i

Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
413
Fig. 2. Conversion and selectivity for (a) Silica Star and (b) Aerosil 300 SiO2 supports; (c) 1.1 wt% Pd/Silica Star and (d) 4.0 wt% Pd/Aerosil 300. C2 denotes the sum of ethylene
and ethane. EtCHO and DEK refer to propionaldehyde and diethyl ketone, respectively. Mass of catalyst (100–300 mg); feed composition 1% PAc, 20% H2, balance He. Total
flow rates for 1.1 wt% and 4 wt% Pd/SiO2 were 50 and 200 sccm, respectively.
CO2 and water. According to their findings, for metal oxide surfaces
with high metal-oxygen bond strength (i.e. SiO2, TiO2), the reactant
required at least one reactive ␣-hydrogen atom to form a ketene-
like intermediate which subsequently reacted with an adsorbed
carboxylate species (from a second molecule of adsorbed carboxylic
acid) to form the ketone [40].
2 nm particles present for the 4 wt% catalyst have a significantly
larger fraction of step and kink sites than the 31 nm particle for the
1.1 wt% catalyst (Table 1), which contain primarily low index (1 1 1)
and (1 0 0) crystal faces. Another explanation may be the different
degrees of acidity for the two SiO2 supports, as indicated by the data
in Fig. 2a and b. Such possible effects are currently under further
investigation.
The physical and chemical properties (i.e. dispersion and parti-
cle size) for M/SiO2 catalysts are summarized in Table 1. Catalyst
evaluation data for 1.1 wt% Pd/SiO2 (incipient wetness) and 4.0 wt%
Pd/SiO2 (SEA) catalysts are summarized in Fig. 2c and d. For
1.1 wt% Pd/SiO2, complete conversion is reached at 350 ◦C with
98% selectivity to ethane. Conversely, at lower temperatures of
200 ◦C, the product distribution is mainly ethane (83%) and some
propionaldehyde (17%), a deoxygenation product formed by C–OH
cleavage. Thus, even at 200 ◦C, the reaction proceeds mainly
through decarbonylation and hydrogenation pathways supported
by CO formation. The reaction result is in agreement with previ-
ous work conducted by Lamb et al., where the liquid-phase HDO of
carboxylic acids over Pd/SiO2 catalyst occurred mainly via decar-
bonylation at 300 ◦C [41]. The reaction over 4 wt% Pd/SiO2 shows
94% conversion at 350 ◦C with 97% and 3% selectivities toward
ethane and methane, respectively. At lower temperatures (200 ◦C),
the product distribution is mainly ethane (94%), with lower produc-
tion of propionaldehyde (6% selective). The different selectivities
with respect to propionaldehyde formation at 200 ◦C could be
attributed to a combination of particle and support effects. The
Given that DEK was formed on both the Silica Star and (to a
much lesser extent) Aerosil 300 supports at temperatures of 300 ◦C
and above, it was important to consider whether this species could
undergo reaction to form ethane and/or propionaldehyde at these
temperatures on the supported Pd catalyst. Thus, the activity of
DEK on 1.1 wt% Pd/SiO2 (incipient wetness) catalyst was explored
to address this issue. Hydrodeoxygenation of DEK was conducted
between 200 and 400 ◦C using 1.0% DEK, 20% H2, and the balance
He. The results confirmed negligible activity (<3% conversion) only
by 400 ◦C, indicating that the observed product formation over the
Pd catalyst during HDO of PAc was not proceeding through this
pathway on the support surface.
Catalyst performances for 2.1 wt% Pt and 2.0 wt% Rh on SiO2
show selectivity only towards hydrocarbons. The results indicate
mainly decarbonylation based on the observation of CO production
as opposed to CO2. Fig. 3(a) and (b) shows the conversion and cor-
responding selectivities to C2 and C3 hydrocarbons for the Pt and
Rh catalysts, respectively, as a function of temperature. Measur-
able PAc conversions were detected at temperatures above 200 ◦C

414
Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
Fig. 4. Conversion and selectivity for (a) Ru and (b) Ni over SiO2 Star. C2 denotes
as the sum of ethylene and ethane C3 denote as propane. EtCHO and DEK denote
propionaldehyde and diethyl-ketone respectively. Mass of catalyst (200–300 mg);
feed composition 1% PAc, 20% H2, balance He; total flow rate 50 sccm.
Fig. 3. Conversion and selectivity for (a) Pt and (b) Rh supported on SiO2 Star. C2
and C3 denote ethane and propane, respectively. Mass of catalyst (200–300 mg);
feed composition 1% PAc, 20% H2, balance He; total flow rate 50 sccm.
and the conversion was complete at temperatures above 300 ◦C for
Rh and above 350 ◦C for Pt. At complete conversion, Pt/SiO2 shows
100% selectivity to ethane, while Rh/SiO2 mainly shows selectiv-
ity to ethane (89%) and some propane (10%), revealing that some
hydrodeoxygenation/hydrogenation is occurring for the formation
of propane.
For the reactions with Ru and Ni supported on SiO2, at high
conversion levels, both oxygenated compounds and cracking prod-
ucts are observed. Fig. 4(a) shows that the selectivity for Ru/SiO2 at
200 ◦C, is mainly towards methane (72%), ethane (22%) and small
traces of propane and butane (6%). As the temperature approaches
to 350 ◦C, the selectivity changes toward DEK (45%), suggesting
PAc interaction at the surface the support. At this temperature,
considerable amounts of methane were formed (38% selectivity).
Therefore, methanation experiments were conducted to verify
whether the catalyst can produce methane by CO hydrogenation
at similar temperatures. Based on these methanation experiments
(see Supporting information Table S1), no substantial amount of
methane is produced over the temperature range 200–400 ◦C.
Therefore, methane is considered to be produced via the cracking
of PAc and/or ethane.
propane (12%), indicative of decarbonylation and hydrodeoxygena-
tion/hydrogenation. At 350 ◦C, the catalyst shows methane (9%) and
DEK (6%) formation. This methane is formed by CO hydrogenation,
as a similar yield of methane is obtained during direct methanation
of CO at similar temperature (see Supporting information, Table S2).
In addition to selectivity pattern studies at variable tempera-
tures, experiments were conducted at low conversion at 200 ◦C
and 1 atm for all the M/SiO2 catalysts. Table 2 summarizes the
reaction rates in ␮mol product formed/min gcat and turnover fre-
quencies (TOF, molecules site−1min−1). The non-rate performance
parameters such as conversion, selectivities towards oxygenated
and nonoxygenated hydrocarbon products, and the carbon bal-
ance, which is the ratio of carbon atoms in the reactor outlet
to carbon atoms in the reactor inlet. The reaction rates were
restricted to differential conversion (<10%), and followed the order:
Pd > Rh ∼ Pt > Ru > Ni. According to the rates per g catalyst, 4.0 wt%
Pd/SiO2 was approximately 25 times higher than for 1.1 wt%
Pd/SiO2. However, when activities are expressed as turn over fre-
quencies (activity per surface site of Pd), the orders are reversed,
with the lower dispersion 1.1 wt% Pd catalyst being more active
by a factor of 2. While there may be some support acidity effects,
the most likely explanation is that the nature and distribution of
the exposed Pd sites are different in these catalysts due to differ-
ent particle sizes. Overall, the TOF values followed a similar trend
as reaction rates with the order: Pd > Ru > Pt > Rh > Ni. The work
For Ni/SiO2 (Fig. 4(b)), very low activity (<1% conversion) is
observed at 200 ◦C, although at 250 ◦C, conversion increases to
11% to form mainly methane (32%) and ethane (59%). However,
as the temperature is raised to 350 ◦C, the conversion reaches
100% and the product distribution is mainly ethane (71%) and

Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
415
Table 2
Kinetic data for the screened catalysts in the HDO of PAc at 200 ◦C and 1 atm. ∼1.0% PAc, 20% H2/He, catalyst mass 100–300 mg, total flow 50 sccm. Pd* was measured at a
total flow 200 sccm.
Metal loading (wt%)
Catalyst
Mx/SiO2
Rate
(␮mol/min gcat
TOF
Conv.
(%)
Select. to organic
products (%)
Select. to
hydrocarbons (%)
C atoms outlet/
C atoms inlet
)
(min−1
)
1.1
4.0
2.1
2.1
1.6
1.8
Pd
Pd*
Pt
Rh
Ru
Ni
2.57
69.1
1.60
1.89
4.55
0.13
0.64
0.32
0.12
0.05
0.24
0.009
7.0
11.0
5.0
4.0
6.0
17
83
94
100
100
100
100
1.0
1.0
1.0
0.9
1.0
–
6
0
0
0
0
1.0
conducted by Kozhevnikov et al. on the deoxygenation of PAc, sug-
gested comparable activity trends when Pd and Pt are compared
(i.e., Pd > Pt) [31]. The Ni/SiO2 catalyst was the least active at 200 ◦C,
consistent with previous investigations that gave similarly poor
activities for nickel-based catalysts for the HDO of carboxylic acids
[42,43].
the carbon-supported Pd catalyst. These issues are currently under
further investigation in our laboratory.
A series of plausible reaction pathways (Scheme 1) have been
developed to better understand the performance of these catalysts.
For Pd/C at 200–250 ◦C, the reaction pathway is more consistent
with decarbonylation, based on the simultaneous formation of
ethane and CO. At higher temperatures (>300 ◦C), CO2 is produced,
which could be due to decarboxylation, but also the water–gas
shift (WGS) reaction. Therefore, the WGS equilibrium constants
of Keq ∼ 38.1, 19.9 and 11.5 at 300, 350 and 400 ◦C [47], respec-
tively, were used to calculate the maximum amount of CO2 that
could be produced by WGS reactions between 350-400 ◦C. Results
showed that WGS could account for only < 5% of the CO2 actu-
ally observed. Consequently, at 350–400 ◦C the reaction proceeds
mainly via decarbonylation and decarboxylation, with negligible
4.2. Catalytic evaluation of palladium on various supports
In addition to examining the catalytic performance of vari-
ous group VIII noble metals over SiO2, the influence of different
supports on the activity and selectivity of Pd for the hydrodeoxy-
genation of PAc was explored. Three different supported Pd
catalysts were evaluated for detailed kinetics: 5.0 wt% Pd/C, 4.0 wt%
Pd/SiO2 and 2.3 wt% Pd/TiO2. Prior to these studies, the perform-
ances of the carbon and TiO2 supports alone were examined,
as shown in Fig. 5(a–b). For carbon, no significant activity was
observed until 350 ◦C, where at 10% conversion, small amounts of
ethane (78% selectivity) were formed; the remaining products con-
sist of C2 and C3 hydrocarbons and propionaldehyde. In contrast,
for TiO2 at 300 ◦C conversion was 20% and DEK was produced at
100% selectivity. This indicates a strong interaction between the
acid and the support that lead to ketonization catalyzed by acidic
sites on TiO2.
Fig. 6(a) for 5.0 wt% Pd/C shows very high selectivity (100%) at
200 ◦C and 91% selective at >350 ◦C, towards ethane with balance
CH4 over the entire temperature range of 200–350 ◦C, producing
mainly CO. The results are in agreement with work of Murzin and
co-workers on the continuous liquid-phase deoxygenation of lau-
ric acid over 5 wt% Pd/C [44]. They reported an activity of 44%
conversion and 98% selectivity to undecane at 300 ◦C. Both stud-
ies indicate that decarbonylation and/or decarboxylation is/are the
main reaction pathways. In contrast, for 2.3 wt% Pd/TiO2 (Fig. 6(b))
the reaction temperature has a significant effect on product distri-
bution. At 200 ◦C, the selectivity is primarily towards oxygenated
compounds (15% propanol, 59% n-propyl-propionate) and some
ethane (23%). As the temperature increases, the selectivity towards
ethane is increased with no evidence of n-propyl-propionate for-
mation above 300 ◦C, which indicates that decarbonylation is
favored above this temperature, due to the formation of CO, verified
by gaseous products analyses generated by the TCD at the different
temperatures studied. At complete conversion (300 ◦C), the reac-
tion was selective toward ethane (90%) with balance propanol (10%)
due to decarbonylation and hydrodeoxygenation respectively.
Table 3 summarizes the physical properties (i.e., metal disper-
sion, particle diameter) and catalyst evaluation results for these
three catalysts. The TOF values decrease in the following order:
Pd/SiO2 > Pd/TiO2 > Pd/C. The Pd/TiO2 and Pd/SiO2 catalysts exhibit
TOF values 20 and 30 times higher, respectively, than Pd/C, which
can be attributed to a combination of particle size [45,46] and/or
support effects. It is difficult to assign the higher activities to sup-
port effects alone, since TiO2 and SiO2 have markedly different
physiochemical properties, yet both are much more active than
Fig. 5. Conversion and selectivity for (a) carbon and (b) TiO2. C2 denotes as
the sum of ethylene and ethane. DEK denote as diethyl-ketone. Mass of catalyst
(200–300 mg); feed composition 1% PAc, 20% H2, balance He; total flow rate 50 sccm.

416
Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
Table 3
Kinetic data for the HDO of PAc over supported Pd catalyst. Reaction conditions: 200 ◦C and 1 atm, XPAc = 0.01, 20% H2/balance He, catalyst weight: Pd/C—200 mg, Pd/SiO2 and
Pd/TiO2—100 mg.
Metal loading (wt%)
Catalyst
Disp. (%)
Particle
size dnm
Conv. (%)
Sel. oxygenate
(%)
Sel. non-
oxygenate (%)
C atoms outlet/
C atoms inlet
TOF (min−1
)
5.0
4.0
2.3
Pd/C
Pd/SiO2
Pd/TiO2
16.9
55.2
27.3
6.8
2.0
4.1
2.0
4.0
3.3
0
6
42
100
94
58
1.0
1.0
1.4
0.01
0.32
0.20
WGS. However, at 300 ◦C, the amount of CO2 that could be formed
based on WGS equilibrium would be equivalent to ∼50% of the CO2
that was actually observed. Therefore, it is possible that some of the
CO2 formed at 300 ◦C was from WGS, with the remaining amount
produced from decarboxylation.
pre-exponential factor, Ea is the activation energy, Pi is the partial
pressure of species i, and xi is the reaction order of species i. The
apparent reaction order is obtained from the slope of ln reaction
rate vs. ln mole fractions of PAc and H2 (XPAc and XH2, respectively).
Fig. 7 shows the plots for 5.0 wt% Pd/C, 4.0 wt% Pd/SiO2 and 2.3 wt%
Pd/TiO2. For all catalysts, the reaction orders in PAc and H2 are
approximately 0.5 and 0, respectively, over the range of concentra-
tions examined. It can be seen based on the formation of CO, that
decarbonylation is the dominant deoxygenation pathway under
these conditions, which is consistent with the computational stud-
ies of Heyden [47] for the DCX and DCN of PAc on Pd(1 1 1) model
surfaces. Heyden’s computations indicated the most kinetically-
favorable pathway was decarbonylation and the rate controlling
steps included dehydroxylation of PAc (C–OH bond cleavage) to
produce propanoyl (CH3CH2CO), which was followed by ␣ carbon
dehydrogenation to form C2 hydrocarbons. Thus, the rate determin-
ing step is dependent on the capacity to remove the hydrocarbon
pool (CH3CHx) from the surface, which could explain why the reac-
tion order for PAc was less than first order. For H2, the reaction
order of zero implies a high coverage of hydrogen on the Pd surface,
relative to PAc reaction intermediates. Interestingly, even though
Pd/SiO2 and Pd/TiO2 have different reaction pathways that lead to
hydrogenation and esterification products such as aldehydes, alco-
hols and esters, the reaction orders are not affected by the supports
even though the selectivity trends are significantly different.
The Arrhenius plots for each catalyst are shown in Fig. 8. Activa-
tion energy values of 16.7 0.6, 19.3 1.6 and 11.7 0.7 kcal/mole
were obtained for the Pd/C, Pd/TiO2 and Pd/SiO2 catalysts, respec-
tively. The apparent activation energies were calculated based on
the total consumption rates of PAc, and therefore represent an
For Pd/SiO2, similar results are observed. The reaction pathways
consist mainly of decarbonylation and hydrogenation steps. PAc is
primarily decarbonylated to produce ethane and CO, although the
formation of small amounts of propionaldehyde indicates hydro-
genation of PAc also occurs, especially at lower temperatures
(200–250 ◦C). In contrast, Pd/TiO2 exhibits different reaction path-
ways. At low temperatures (∼200 ◦C), the formation of propanol
and n-propyl propionate is observed. At temperatures >225 ◦C,
ethane, CO and CO2 are observed, indicating decarbonylation and
decarboxylation. The results are consistent with previous studies
for the hydrogenation of carboxylic acids over Pt/TiO2, where it
was claimed that oxygen vacancies were generated on the oxide
support from the hydrogen spillover of Pd during reduction [50].
Hence, a plausible reaction pathway over Pd/TiO2 could be the gen-
eration of oxygen vacancies at the interface of the catalyst, which
in turn weakens the C=O bond of PAc to favor the formation of
propanol and water. Esterification of propanol with PAc can then
occur to form n-propyl propionate. At higher temperatures, how-
ever, decarbonylation of PAc to form ethane and CO becomes the
predominant reaction pathway.
4.3. Kinetics of palladium on various supports
The reaction rates of the three catalysts can be expressed as
ꢃ
ꢄ
ꢅ
n
E
RT
a
power rate law equations, r = A exp
−
i=1Pxi, where A is the
i
Scheme 1. Plausible reaction mechanism for the HDO of PAc over Pd over SiO2, C and TiO2, between 200 and 400 ◦C. Mx/SiO2 (Pd, Pt, Rh), My/SiO2 (Ru, Ni).

Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
417
Fig. 8. Arrhenius apparent activation energies for (ꢀ) 4.0 wt% Pd/Aerosil 300, (ꢁ)
2.3 wt% Pd/TiO2 and (ꢂ) 5 wt% Pd/C.
aggregate of multiple pathways. Pd/C exhibit mainly decarbonyla-
tion/decarboxylation reactions, where Pd/SiO2 and Pd/TiO2 exhibit
hydrogenation/decarbonylation reactions. For Pd/C, the results are
in general agreement with previous work [48,49], where activa-
tion energy values for the vapor-phase hydrogenation of carboxylic
acids over platinum group metals ranged from 13–16 kcal/mol. In
the case of Pd/SiO2, a recent investigation [31] has reported an Ea
value of 8.4 kcal/mol over the temperature range 250–300 ◦C, which
is lower than the value reported here (11.7 kcal/mol). However,
in addition to the temperature range being higher, the selectivity
in the previous study was towards ethylene as opposed to ethane
observed here, suggesting different mechanistic behavior.
While the lower Ea value of 11.7 kcal/mol for Pd/SiO2 might
suggest the presence of some external mass transfer limitations
[50], the experiments here were conducted in the kinetic regime.
This was confirmed by varying the flow rates over a range of
50–250 sccm for each catalyst to determine the minimum flow rate
where the reaction rate remained constant (and thus free of exter-
nal mass transfer limitations). The minimum flow rates where no
external mass transport limitations were observed were found to
be 50, 200 and 150 sccm for Pd/C, Pd/SiO2 and Pd/TiO2, respectively,
and these were the flow rates used in the kinetic studies.
Fig. 6. Conversion and selectivity for (a) Pd/C and (b) Pd/TiO2. C2 denotes as the
sum of ethylene and ethane. Ester denotes as propyl-propionate. Mass of catalyst
(100–200 mg); feed composition 1% PAc, 20% H2, balance He; total flow rate 50 sccm.
In addition to confirming the absence of external mass transfer
limitations, intraparticle diffusion limitations could be dismissed
since application of the Weisz–Prater criterion showed that the sil-
ica pore diameters were much lower than the mean free path (ꢁ) of
PAc, indicating that diffusivity was governed by Knudsen diffusion
and that internal mass transport limitations were absent [51–53].
The criterion represents a useful technique, which includes param-
eters that can be calculated to determine if the reaction process
is operating in a diffusion or reaction-limited regime. In order to
determine the absence of pore diffusion limitations the following
equation has to be satisfied:
ꢀPAcꢂcRp2
CsDeff
NW−P
=
≤ 0.3
(9)
where ꢀPAc (moles/s gcat) denotes the rate of reaction of PAc, ꢂc
(gcat/cm3) is the catalyst pellet density, Rp (cm) is the radius of the
particle, Cs (moles/cm3) is the concentration of PAc at the surface
of the catalyst (assuming that it is the same as the bulk concen-
tration) and Deff is the effective diffusion coefficient. If the support
pore diameter is much less than the mean free path (ꢁ), the diffu-
sivity is governed by Knudsen diffusion. The criterion values for Pd
over carbon, TiO2 and SiO2 were 4.5E − 04, 2.4E − 05 and 1.8E − 02
Fig. 7. Power rate law for the kinetic dependencies of PAc and H2 for PAc conver-
sion at 200 ◦C and 1 atm pressure. (ꢀ) 4.0 wt% Pd/Aerosil 300, (ꢁ) 2.3 wt% Pd/TiO2,
(ꢂ) 5 wt% Pd/C, reactor conditions: 20% H2/balance, total flow: Pd/C—50 sccm,
Pd/SiO2—200 sccm, Pd/TiO2—150 sccm. For hydrogen kinetics ∼ 1.0% PAc balances
He.

418
Y.K. Lugo-José et al. / Applied Catalysis A: General 469 (2014) 410–418
respectively, confirming the absence of internal mass transport lim-
itations. Recently, Lopez-Ruiz and Davis used the same criterion
for confirming the absence of internal mass transport during the
decarbonylation of heptanoic acid over Pt/C [54]. Therefore, we
are confident that the lower Ea value of 11.7 kcal/mol reflects the
true kinetics of the reaction. The observations for the lower Ea for
Pd/SiO2 compared to Pd/TiO2 and Pd/C might be due again to a
combination of support and particle size effect, and this is currently
under investigation.
[9] A.V. Bridgwater, D. Meier, D. Radlein, Organic Geochemistry 30 (1999)
1479–1493.
[10] W.F. Maier, W. Roth, I. Thies, P.V. Schleyer, Chemische Berichte/Recueil 115
(1982) 808–812.
[11] J.C. Serrano-Ruiz, A. Pineda, A.M. Balu, R. Luque, J.M. Campelo, A.A. Romero, J.M.
Ramos-Fernández, Catalysis Today 195 (2012) 162–168.
[12] B.J. O’Neill, E.I. Gürbüz, J.A. Dumesic, Journal of Catalysis 290 (2012) 193–201.
[13] C.A. Henao, D. Braden, C.T. Maravelias, J.A. Dumesic, A novel catalytic strategy
for the production of liquid Fuels from Ligno-cellulosic biomass, in: M.C.G.E.N.
Pistikopoulos, A.C. Kokossis (Eds.), Computer Aided Chemical Engineering, Else-
vier, 2011, pp. 1723–1727.
[14] R.M. West, E.L. Kunkes, D.A. Simonetti, J.A. Dumesic, Catalysis Today 147 (2009)
115–125.
[15] J.Q. Bond, D. Wang, D.M. Alonso, J.A. Dumesic, Journal of Catalysis 281 (2011)
290–299.
5. Conclusion
[16] S.G. Wettstein, J.Q. Bond, D.M. Alonso, H.N. Pham, A.K. Datye, J.A. Dumesic,
Applied Catalysis B: Environmental 117–118 (2012) 321–329.
[17] E.W. Ping, J. Pierson, R. Wallace, J.T. Miller, T.F. Fuller, C.W. Jones, Applied Catal-
ysis A: General 396 (2011) 85–90.
[18] S.J. Tauster, S.C. Fung, R.L. Garten, Journal of the American Chemical Society 100
(1978) 170–175.
[19] T.C. Chang, J.J. Chen, C.T. Yeh, Journal of Catalysis 96 (1985) 51–57.
[20] S.J. Tauster, Accounts of Chemical Research 20 (1987) 389–394.
[21] I. Simakova, O. Simakova, P. Mäki-Arvela, A. Simakov, M. Estrada, D.Y. Murzin,
Applied Catalysis A: General 355 (2009) 100–108.
[22] H. Bernas, K. Eränen, I. Simakova, A.-R. Leino, K. Kordás, J. Myllyoja, P. Mäki-
Arvela, T. Salmi, D.Y. Murzin, Fuel 89 (2010) 2033–2039.
[23] D.Y. Murzin, M. Snare, I. Kubickova, P. Maki-Arvela, K. Eranen, J. Warna, The
Chemical Engineering Journal 134 (2007) 29–34.
[24] S. Lestari, I. Simakova, A. Tokarev, P. Mäki-Arvela, K. Eränen, D. Murzin, Catalysis
Letters 122 (2008) 247–251.
[25] S. Lestari, P. Mäki-Arvela, I. Simakova, J. Beltramini, G. Lu, D. Murzin, Catalysis
Letters 130 (2009) 48–51.
[26] I. Kubicková, M. Snåre, K. Eränen, P. Mäki-Arvela, D.Y. Murzin, Catalysis Today
106 (2005) 197–200.
[27] J.N. Chheda, J.A. Dumesic, Catalysis Today 123 (2007) 59–70.
[28] M. Chia, J.A. Dumesic, Chemical Communications 47 (2011) 12233–12235.
[29] S. Lestari, P.i. Ma¨ki-Arvela, H. Bernas, O. Simakova, R. Sjo¨holm, J. Beltramini,
G.Q.M. Lu, J. Myllyoja, I. Simakova, D.Y. Murzin, Energy & Fuels 23 (2009)
3842–3845.
[30] M. Chiappero, P.T.M. Do, S. Crossley, L.L. Lobban, D.E. Resasco, Fuel 90 (2011)
1155–1165.
[31] M.A. Alotaibi, E.F. Kozhevnikova, I.V. Kozhevnikov, Applied Catalysis A: General
447–448 (2012) 32–40.
[32] T.E. Feltes, L. Espinosa-Alonso, E.d. Smit, L. D’Souza, R.J. Meyer, B.M. Weckhuy-
sen, J.R. Regalbuto, Journal of Catalysis 270 (2010) 95–102.
[33] R. John, Strong electrostatic adsorption of metals onto catalyst supports, in:
Catalyst Preparation, CRC Press, Boca Raton, FL, 2006, pp. 297–318.
[34] S. Lambert, N. Job, L. D’Souza, M.F.R. Pereira, R. Pirard, B. Heinrichs, J.L.
Figueiredo, J.-P. Pirard, J.R. Regalbuto, Journal of Catalysis 261 (2009) 23–33.
[35] L. Jiao, J.R. Regalbuto, Journal of Catalysis 260 (2008) 329–341.
[36] J.T. Miller, M. Schreier, A.J. Kropf, J.R. Regalbuto, Journal of Catalysis 225 (2004)
203–212.
[37] R.L. Augustine, Catalyst support, in: M. Dekker (Ed.), Heterogeneous catalysis
for the synthetic chemist, Marcel Dekker Inc, New York, 1996, pp. 153–178.
[38] R. Pestman, R.M. Koster, A. vanDuijne, J.A.Z. Pieterse, V. Ponec, Journal of Catal-
ysis 168 (1997) 265–272.
In conclusion, gas-phase HDO of PAc over various supported
metal catalysts has been carried out. For M/SiO2 catalysts, the
choice of group VIII noble metal has a significant effect on the
selectivity, which is also significantly influenced by the temper-
ature. At 200 ◦C the main reaction pathways for the HDO of PAc
under the explored conditions are: (1) hydrogenation, decarbony-
lation and decarboxylation over Pd, Pt and Rh based catalysts, and
(2) decarbonylation and methanation, followed by ketonization at
high temperatures over Ru and Ni respectively. The TOF for the PAc
conversion is highest over Pd-based catalysts, suggesting that this
metal is the most effective for this class of reactions of the ones
tested. Kinetic studies over Pd on different supports indicate that
the activity (TOF) changes depending on the support, thus indi-
cating a combination of particle size and support effects. For all
Pd catalysts, the apparent reaction order with respect of PAc and
hydrogen is half and zero order, respectively. These reaction orders
are consistent with rate determining steps involving C–OH bond
cleavage and the removal of hydrocarbon pool species (e.g., CH3Cx)
from the surface of the catalyst. These initial findings have formed
the basis for more focused studies of the effect of Pd particle size
on activity and the influence of adding second metal to Pd (i.e.,
bimetallic effects), and will be the subject of future reports.
Acknowledgments
The authors wish to thank the financial support from National
Science Foundation (NSF) grant number CHE-1153012. Yuliana
Lugo-José thanks the Sloan Foundation and the NSF Southeastern
Alliance for Graduate Education and the Professoriate (SEAGEP) for
financial support. The authors thank Professors Andreas Heyden
and John R. Regalbuto from the University of South Carolina for
their helpful and insightful conversations.
[39] R. Pestman, A. Vanduijne, J.A.Z. Pieterse, V. Ponec, Journal of Molecular Catalysis
A: Chemical 103 (1995) 175–180.
[40] M. Renz, European Journal of Organic Chemistry (2005) 979–988.
[41] J.P. Ford, J.G. Immer, H.H. Lamb, Topics in Catalysis 55 (2012) 175–184.
[42] A.T. Madsen, E.H. Ahmed, C.H. Christensen, R. Fehrmann, A. Riisager, Fuel 90
(2011) 3433–3438.
[43] W.F. Maier, W. Roth, I. Thies, P.V.R. Schleyer, Chemische Berichte 115 (1982)
808–812.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2013.10.025.
[44] P. Mäki-Arvela, M. Snåre, K. Eränen, J. Myllyoja, D.Y. Murzin, Fuel 87 (2008)
3543–3549.
[45] R. Van Hardeveld, F. Hartog, Surface Science 15 (1969) 189–230.
[46] G.C. Bond, Surface Science 156 (1985) 966–981, Part 2.
[47] J. Lu, S. Behtash, A. Heyden, The Journal of Physical Chemistry C 116 (2012)
14328–14341.
[48] D.A. Bulushev, J.R.H. Ross, Catalysis Today 163 (2011) 42–46.
[49] W. Rachmady, M.A. Vannice, Journal of Catalysis 192 (2000) 322–334.
[50] G.W. Roberts, Chemical Reactions and Chemical Reactors, John Wiley & Sons,
Hoboken, NJ, 2008.
[51] P.G. Smirniotis, Angewandte Chemie International Edition 46 (2007), 1007-
1007.
[52] P.B. Weisz, C.D. Prater, Interpretation of measurements in experimental catal-
ysis, in: V.I.K.W.G. Frankenburg, E.K. Rideal (Eds.), Advances in Catalysis,
Academic Press, New York, 1954, pp. 143–196.
References
[1] B. ASSOC, Annu. Energy Rev. 2010 (1949–2010) 384.
[2] S. National Academy of, E. National Academy of, M. Institute of, P. National
Academies, What You Need to Know about Energy, National Academies Press,
Washington, DC, 2008.
[3] D.A. Simonetti, J.A. Dumesic, Catalysis Review 51 (2009) 441–484.
[4] D.G. Lima, V.C.D. Soares, E.B. Ribeiro, D.A. Carvalho, É.C.V. Cardoso, F.C. Rassi,
K.C. Mundim, J.C. Rubim, P.A.Z. Suarez, Journal of Analytical and Applied Pyrol-
ysis 71 (2004) 987–996.
[5] A.V. Bridgwater, Chemical Engineering Journal 91 (2003) 87–102.
[6] J. Michael Robinson, C.E. Burgess, M.A. Bently, C.D. Brasher, B.O. Horne, D.M.
Lillard, J.M. Macias, H.D. Mandal, S.C. Mills, K.D. O’Hara, J.T. Pon, A.F. Raigoza,
E.H. Sanchez, J.S. Villarreal, Biomass and Bioenergy 26 (2004) 473–483.
[7] C. Bessou, F. Ferchaud, B. Gabrielle, B. Mary, Agronomy for sustainable devel-
opment, EDP Sciences 31 (2011) 1–79.
[53] M.E. Davis, R.J. Davis, Fundamentals of chemical reaction engineering, 1st ed.
ed, McGraw-Hill, Dubuque, IA, 2002.
[54] J.A. Lopez-Ruiz, R.J. Davis, Green Chemistry (2013), Advance Article.
[8] T.V. Choudhary, C.B. Phillips, Applied Catalysis A: General 397 (2011) 1–12.