This is an open access article published under an ACS AuthorChoice License, which permits

copying and redistribution of the article or any adaptations for non-commercial purposes.

Review

pubs.acs.org/acscatalysis

Bifunctional Catalysts for Upgrading of Biomass-Derived

Oxygenates: A Review
Allison M. Robinson,† Jesse E. Hensley,‡ and J. Will Medlin*,†
†
Department of Chemical and Biological Engineering, University of ColoradoBoulder, UCB 596, Boulder, Colorado 80309, United
States
‡
National Bioenergy Center, National Renewable Energy Laboratory, 15013 Denver West Pikeway, Golden, Colorado 80401, United
States
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Downloaded via 88.202.186.59 on October 16, 2019 at 04:38:12 (UTC).

ABSTRACT: Deoxygenation is an important reaction in the conversion of biomass-derived oxygenates to fuels and chemicals. A
key route for biomass refining involves the production of pyrolysis oil through rapid heating of the raw biomass feedstock.
Pyrolysis oil as produced is highly oxygenated, so the feasibility of this approach depends in large part on the ability to selectively
deoxygenate pyrolysis oil components to create a stream of high-value finished products. Identification of catalytic materials that
are active and selective for deoxygenation of pyrolysis oil components has therefore represented a major research area. One
catalyst is rarely capable of performing the different types of elementary reaction steps required to deoxygenate biomass-derived
compounds. For this reason, considerable attention has been placed on bifunctional catalysts, where two different active materials
are used to provide catalytic sites for diverse reaction steps. Here, we review recent trends in the development of catalysts, with a
focus on catalysts for which a bifunctional effect has been proposed. We summarize recent studies of hydrodeoxygenation
(HDO) of pyrolysis oil and model compounds for a range of materials, including supported metal and bimetallic catalysts as well
as transition-metal oxides, sulfides, carbides, nitrides, and phosphides. Particular emphasis is placed on how catalyst structure can
be related to performance via molecular-level mechanisms. These studies demonstrate the importance of catalyst bifunctionality,
with each class of materials requiring hydrogenation and C−O scission sites to perform HDO at reasonable rates.
KEYWORDS: bifunctional, catalyst, hydrodeoxygenation, bimetallic, metal−metal oxide, tethered bifunctional catalysts

1. INTRODUCTION compounds, making upgrading through deoxygenation critical

Increasing environmental, economic, and political motivation
to develop renewable liquid fuels and chemicals has generated
significant interest in using biomass as a sustainable carbon
source. Upgrading of lignocellulosic biomass provides a unique
opportunity to utilize a renewable resource to create these fuels
and chemicals from waste materials and plants that do not
compete with food production. Through fast pyrolysis, the
biomass is rapidly heated in the absence of oxygen to produce a
condensable mixture of oxygenates along with char and light
gases.1−3 Pyrolysis oil composition varies depending on the
feedstock and process conditions, but it contains high
concentrations of phenolic compounds from lignin decom-
position and furanic compounds derived from cellulose, in
addition to organic acids and other low molecular weight
oxygenates.4 This complex mixture of oxygenates generally has
poor properties as a fuel.1,5,6 The undesirable fuel character-
istics stem from a high concentration of highly oxygenated
for improving heating value, stability, and physical properties
such as acidity and viscosity.
Conversion of the compounds making up pyrolysis oil
requires new types of catalysts that can selectively remove
oxygen from the oil. 9−11 One common strategy for
deoxygenating whole pyrolysis oil is hydrodeoxygenation
(HDO), which involves adding hydrogen to accomplish C−O
bond scission via removal of water. This is typically done
sequentially, with a hydrogenation-dehydration-hydrogenation
route in which dehydration often occurs over an acid site.
Alternatively, direct deoxygenation (DDO) does not require
intermediate hydrogenation steps; instead, the C−O bond is
directly cleaved from the reactant through hydrogenolysis. It
Received: March 31, 2016
Revised: June 18, 2016
Published: June 21, 2016

© 2016 American Chemical Society 5026 DOI: 10.1021/acscatal.6b00923

ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
should be noted that while these definitions of HDO and DDO
are used throughout this manuscript, some previous reports
have used different terminology for the same reactions. Other
reactions can also take place under these conditions, including
decarboxylation/decarbonylation, cracking, and hydrogenation
(Figure 1).7,8 It should be noted that while HDO reactions
Figure 1. Example reactions discussed in this review.
often produce saturated compounds and DDO may lead
directly to aromatic compounds, aromatic and saturated
products can also be interconverted subsequent to being
formed. For example, increasing the temperature shifts the
thermodynamic equilibrium toward aromatic products such as
toluene production over methylcyclohexane.
A wide variety of materials have been studied for pyrolysis oil
upgrading, ranging from carbides and nitrides to supported
transition metals, but despite their diversity, the most promising
materials are bifunctional: they contain disparate active sites
that catalyze complementary reaction steps. In many cases,
upgrading reactions demand a combination of the different
reaction types described in Figure 1; it is natural that each step
would require its own catalyst.
In this contribution, we review advances in catalysts currently
used for HDO with a focus on bifunctional catalysts. There
have been a wide variety of studies attempting to better
understand how HDO proceeds over different catalytic
materials using various probe molecules or whole oils.
Comparing these approaches may help elucidate which types
of materials are most promising for the end goal of making
pyrolysis oil a viable alternative to petroleum in fuel and
chemical production. Although pyrolysis oil is known to
contain many different compounds and types of functional
groups,9 the majority of the studies referenced in this report use
phenolic compounds as probe molecules. This will aid in
5027
Review
gaining an understanding of not only how the current state-of-
the-art deoxygenation catalysts perform but also how they
relate to one another.
2. BIFUNCTIONAL CATALYSIS
The general concept behind bifunctional heterogeneous
catalysis is that there are two distinct types of active sites that
function in tandem to perform a surface-catalyzed reaction.
Often, these two types of sites are expected to catalyze different
elementary steps within an overall reaction. In principle,
however, the two sites could participate in the same step, for
example through interaction of different parts of an adsorbed
reactant molecule with the catalyst as in a concerted reaction. In
this section, we highlight two major types of materials used as
bifunctional catalysts before considering the application of
similar types of materials in biomass refining reactions.
Proposed mechanisms for two major classes of catalysts
discussed in this review are shown schematically below in
Figure 2a,b. To point out the connections between these
examples for simple reactants of well-studied reactions and the
oxygenate upgrading reactions of primary interest for this
review, an example of bifunctional catalysts for deoxygenation
of phenolic oxygenates is shown in Figure 2c.
2.1. Examples of Bifunctional Catalysis. 2.1.1. Bimetallic
Catalysts. One of the most common methods for controlling
the catalytic performance of supported metal catalysts is via
control over the composition of bimetallic catalysts. In a large
number of cases, adding one metal to another is not intended
to introduce a new catalytic function, but instead, the intent is
to modify the activity of the original metal. A second metal may
influence the properties of the first via ligand effects, which is
often interpreted in terms of a shifting of the d-band density of
states of the first metal. A much-studied example of this
phenomenon is in the performance of Pt “skin” catalysts that
have been modified by 3d elements such as Ni, Co, Fe, and Ti
in the subsurface for improved performance in the electro-
catalytic oxygen reduction reaction.13−15 In many of the
systems dominated by this effect, the modifying metal is
assumed to not be present in the surface layer and thus has little
opportunity to directly create a bifunctional effect.16 Another
mechanism by which addition of a second metal can modify a
catalyst surface is via the so-called “ensemble” effect.17 For
example, addition of (relatively unreactive) Au to Pd olefin
hydrogenation catalysts has yielded an improved activity that
has been attributed to a decrease in the average number of
contiguous Pd atoms; contiguous surface Pd clusters have been
associated with accumulation of unreactive spectator species.18
In some cases, however, the presence of two metals at the
surface under reaction conditions has been associated with a
bifunctional effect. Bifunctional effects have been invoked, for
example, in some studies involving PdAu catalysts. Xu and co-
workers found that biphasic PdAu nanoparticleswhere
segregation of the two metals into relatively Pd-rich and Au-
richwere highly active for low-temperature CO oxidation.10
The authors found that Pd-rich sites were needed to carry out
the dissociative adsorption of oxygen, whereas Au-rich sites
were active for CO oxidation, as shown schematically in Figure
2a. The dispersion of the two metals in an active catalyst has
been seen experimentally through STM images of PdAu alloy
electrodes. In this example, PdAu thin films were electro-
deposited on Au(111), resulting in Pd monomers, dimers, and
trimers on the surface (Figure 3).19 Scanning tunneling
microscopy was used to obtain TM images of different PdAu
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
Figure 2. Schematic representation of proposed reaction mechanisms by (a) bimetallic catalysts,10 (b) metal−metal oxide catalysts,11 and (c)
bimetallic catalysts specifically for deoxygenation.12
Figure 3. Surface coverage of monomers, dimers, and trimers of Pd for
two surfaces along with experimental results for the coverage of H and
CO on these two electrodes. Reprinted with permission from ref 19.
Copyright 2001 American Association for the Advancement of
Science.
compositions, which were found to contain highly dispersed Pd
atoms. This dispersion is correlated with activity in Figure 3;
the coverage of hydrogen and CO on the surface was measured
along with the coverage of monomers, dimers, and trimers of
Pd.19 Because CO adsorbs on Pd sites, the coverage of CO
increases with Pd loading. Furthermore, IR results showed that
the CO adsorbed on a top site in these disperse samples, while
5028
Review
large Pd clusters (or a pure Pd surface) had a high
concentration of bridge/3-fold bound CO.19 This high
dispersion creates many of the neighboring Pd−Au sites
responsible for the bifunctional reaction shown schematically in
Figure 2a.
Similar effects have been proposed in olefin hydrogenation,
where Pd sites are necessary for dissociative hydrogen
adsorption while Au sites are active for hydrogenation.20
Bimetallic catalysts have also been employed such that different
types of sites are effective for reactions under different
conditions. For example, PtAu nanoparticles have been applied
as electrocatalysts for lithium−air batteries. The Au component
was found to be more active for the oxygen reduction reaction
(which occurs during battery discharging), while Pt sites were
associated with the reverse oxygen evolution reaction (which
occurs during charging).21 The utility of combining two active
sites in one material has also been shown for use in unitized
regenerative fuel cells (URFCs). These electrochemical cells
operate in one mode to electrolyze water, producing hydrogen
and oxygen, then operate in fuel cell mode to recombine them
and produce electricity.22,23 The challenge with this concept is
designing an electrocatalyst that is active for all of the reactions
involved; catalysts active for the oxygen reduction reaction
(ORR) perform poorly for the oxygen evolution reaction
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
(OER).22,23 Finding a catalyst that is both stable and active for
OER, which involves oxidizing water to molecular oxygen in
acidic media, has represented a major research focus in
electrocatalysis. This is due in part to the limited studies on
the relationship between catalyst active site structure and OER
activity.24 While Pt is nearly ideal for ORR, it has shown poor
OER activity. The lower rates have been attributed to reduced
surface area and formation of a poorly conductive PtOx layer
that resists current flow under the conditions of OER. This may
be overcome by using a bifunctional catalyst consisting of the
conventional Pt particles used for ORR supported on an oxide
such as IrO2 for the oxygen evolution reaction.22,23
2.1.2. Metal−Metal Oxide Catalysts. A frequently used type
of bifunctional catalyst combines metal sites that are active for
hydrogenation/dehydrogenation and acid sites that are active
for protonation/deprotonation. For example, catalytic metals
may be supported on a zeolite. In the isomerization of alkanes,
the saturated hydrocarbons are dehydrogenated to alkenes on
the metal sites; the alkenes can subsequently migrate to acid
sites, which protonate the alkenes to form carbenium ions.25
Structural rearrangement of the carbenium ion can lead to
formation of branched structures, which upon deprotonation
yield branched alkenes. Finally, the branched alkenes are
hydrogenated on metal sites. While these reactions can be
catalyzed by physical mixtures of metal and zeolite catalysts,
achieving intimate contact between the different types of
reactive sites significantly improves the kinetics. Producing
zeolite-supported metals with high dispersion is important for
improving the intimacy of contact and achieving rapid
isomerization kinetics.26
The intimacy of metal−metal oxide contact has also been
found to be important in other reactions, with considerable
focus devoted to Au nanoparticles dispersed on reducible oxide
supports.27 For example, Au supported on ceria28,29 or titania30
has been found to be active for the water−gas shift reaction at
relatively low temperatures, ranging from approximately 200−
350 °C. The metal oxide support has been reported to be active
for the required water dissociation step, whereas CO oxidation
has been associated with sites at the Au−metal oxide interface.
A related mechanism has been demonstrated on Pt/ZrO2
catalysts for dry reforming of methane (Figure 2b). It was
shown that the Pt metal serves to activate methane,
dehydrogenating it to form CHx, while sites at the Pt-ZrO2
interface are responsible for CO2 dissociation. Achieving good
catalyst performance was found to require optimization of the
interfacial contact area.11 Furthermore, bifunctional metal−
metal oxide catalysts in which the oxide serves as a base have
also been reported. For example, isobutanol may be produced
using syngas (CO and H2) on bifunctional catalysts consisting
of Cu supported on basic metal oxides such as MgO.31 The
syngas initially forms alcohols that may be coupled through
condensation reactions. In the proposed mechanism, the
alcohol dissociatively adsorbs on the basic oxide surface,
creating hydrogen and alkoxide species.31 Hydrogen is removed
by migration to Cu sites, where they recombine and desorb as
H2, while the alkoxide species may be further dehydrogenated
and eventually take part in condensation reactions to form
longer chain compounds.31 Chain growth rates are low due to
the steric hindrance that occurs as branches are added to the
initially formed linear alcohols, resulting in production of
isobutanol at high selectivity.31
2.2. Importance of Bifunctionality for Biomass
Reforming. A common theme across the classes of materials
5029
Review
studied for biomass deoxygenation is their bifunctionality. As
discussed later in this review, the importance of having sites
capable of activating hydrogen in close proximity to sites that
are effective for C−O scission has been proposed many times
throughout the literature for materials ranging from supported
transition metals to traditional hydrotreating catalysts such as
sulfides.32−44 Often these two reaction steps cannot be
effectively catalyzed by one type of active site. As a result, the
most commonly proposed mechanisms for deoxygenation of
these compounds involve one active site (such as a transition
metal) that readily activates hydrogen and another type of site
nearby, such as an acid site on an oxide support, for C−O bond
activation. Although this can be studied using whole pyrolysis
oil, the complex mixture of oxygenates present makes it very
difficult to elucidate reaction pathways to understand
mechanisms. For this reason, it can be useful to instead
consider more detailed studies using model compounds.
Phenolic compounds are useful probe molecules for pyrolysis
upgrading reactions; they make up a large fraction of the oil,
and the C−O bonds in these molecules are very difficult to
break.33,45−47 Other processes more representative of alter-
native conversion approaches, such as aqueous-phase process-
ing of polyols, are also described below.
3. BIFUNCTIONAL CATALYSTS FOR
DEOXYGENATION
3.1. Supported Metals. Many supported metal catalysts
have been investigated for deoxygenation reactions. While it is
difficult to directly compare the variety of different catalysts
tested under varying reaction conditions, it may be informative
to look at general trends in performance that has been observed
in the literature. Figure 4 shows the selectivity to deoxygenated
products when feeding aromatic oxygenates including cresol
and guaiacol.
Figure 4. Selectivity to deoxygenated products from feeding vapor-
phase aromatic oxygenates (cresol or guaiacol) over Pt (diamonds),
Ru (crosses), Fe (squares), Pd (triangles), and Ni (circles) catalysts.
Individual data points are taken from a number of prior literature
reports, with the citation number indication by the color for that data
point. Dashed lines are used to connect points for each metal, but note
that results from each metal come from differently prepared catalysts,
as reported in Table 1.
Across a range of conversions, Pt catalysts appear to have the
highest deoxygenation selectivity of these monometallic
samples. The data used in this figure can also be seen in
Table 1. However, the performance of these monometallic
catalysts can be dramatically changed through incorporation of
a second metal, as will be discussed below.
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis Review

Table 1. Data Used in Figure 4 To Compare Selectivity to Deoxygenation of m-Cresol or Guaiacol over Various Supported
Metals
catalyst temperature (K) P (atm) PH2 (atm) wt % metal reactant conversion (%) % deoxygenation ref
Pt/Al2O3 533 1 0.5 0.5 m-cresol 9.3 79.2 33
Pt/Al2O3 533 1 0.5 1 m-cresol 18 82.6 33
Pt/Al2O3 533 1 0.5 1.7 m-cresol 38.3 80.5 33
Pt/SiO2 533 1 0.5 1.6 m-cresol 54.9 82.1 33
Pt/Al2O3 533 1 0.5 1.7 m-cresol 30 76.1 32
Pt/TiO2 573 1 1 1 m-cresol 17 97 66
Pt/ZrO2 573 1 1 1 m-cresol 12 88.5 66
Ni/SiO2 523 1 1 5 m-cresol 55.8 7.2 68
Ni/C 623 39.5 37.6 5 guaiacol 30.7 17.5 60
Ni/SiO2 573 1 1 5 m-cresol 16.2 14.2 34
Pd/C 573 1 0.4 5 m-cresol 10 78 61
Pd/SiO2 523 1 1 1 m-cresol 54.9 24.1 68
Pd/C 623 39.5 37.6 1 guaiacol 15.5 23.3 60
Fe/SiO2 573 1 1 5 m-cresol 8.8 60.2 34
Fe/SiO2 623 1 0.9 not reported guaiacol 50 37 76
Ru/SiO2 573 1 1 9.4 m-cresol 4.5 38.5 67
Ru/C 623 39.5 37.6 1 guaiacol 34.2 39.6 60

3.1.1. Supported Noble Metals. One of the most widely
used types of catalysts for upgrading of biomass-derived
oxygenates are supported metals. To understand reactivity
trends for these systems, it is often useful to study reactivity on
model surfaces such as single crystals. Reactions of simple
organic oxygenates that can serve as probe molecules for
components in deconstructed biomass have been extensively
studied using surface spectroscopies and model surfaces. There
have been past reviews focusing on the decomposition of
simple oxygenates including alcohols, aldehydes, ketones, and
carboxylic acids over transition-metal surfaces.48−51 For
example, formic acid decomposition has been studied as a
probe for organic acids. A past review of surface science studies
over single crystals focused on formic acid adsorption and
decomposition processes.48 Formic acid tends to decompose
through a formate intermediate. The formation of a carboxylate
intermediate has been observed for larger carboxylic acids as
well, over a wide array of transition metals.51 Formate
decomposition can then occur either through dehydrogenation
to produce CO2 and H2, or C−O scission to form CO and
H2O.48 The pathway that a given transition metal will favor
depends on how strongly it binds oxygen: metals like Pt that
form relatively weak metal−oxygen bonds favor dehydrogen-
ation52 while metals further left on the periodic table such as Fe
favor C−O scission.53 On Pd(111) both pathways were
observed, although selectivity was greater toward dehydrogen-
ation.54 On Ni(111) the binding configuration was proposed to
be different, creating formic anhydride as an intermediate, but
the dominate decomposition pathway was C−O scission.55 A
major conclusion from these studies is that C−O bond scission
is often not favored on metals like Pt and Pd that do not
strongly bind oxygen, even when the C−O bond is relatively
weak. Similar observations have been made for reactions of
other oxygenates such as alcohols and aldehydes. On noble
metals such as Pt and Pd, decarbonylation via C−C scission is
often a dominant pathway, and C−O scission is generally not
observed.49 An exception is the reaction of unsaturated alcohols
and aldehydes such as prenal, furfural, and benzyl alcohol,
where the C−O bond is much weaker than in aliphatic or
phenolic reactants. However, noble metals are attractive for
their ability to activate hydrogen and to conduct hydrogenation
5030
reactions. This suggests that deoxygenation for reforming
pyrolysis oil may be best accomplished using a bifunctional
catalyst that combines hydrogenation sites with sites that form
strong metal−oxygen bonds.
This hypothesis is also supported by studies of technical
catalysts composed of combinations of metal and acid sites.
Late-transition-metal catalysts have several properties that make
them potentially attractive for pyrolysis oil upgrading; they are
typically very active, are effective at hydrogenation, and have
well-characterized structures. Additionally, they do not suffer
from some of the issues of traditional hydrotreating catalysts
discussed below, such as the need for cofeeding toxic H2S to
maintain activity as for sulfides.56 When an active support
material is used to impart bifunctionality, these catalysts can be
used for HDO of relatively recalcitrant compounds such as
phenolics. Before discussing the bifunctional materials, it is
useful to first discuss the individual components.
Noble metals including Ru, Rh, Pd, and Pt are a class of
promising upgrading catalysts; they activate hydrogen, are not
deactivated by water, and are already widely used in commercial
catalytic processes.57 Under typical upgrading conditions, the
differences in the chemistry catalyzed by the various noble
metals are subtle,8 and these materials have been highlighted in
a number of reviews discussing pyrolysis oil upgrading.1,7,8,39,58
As an illustration of their HDO activity as monofunctional
catalysts, Pd/C, Pt/C, and Ru/C were studied for vapor-phase
guaiacol (2-methoxyphenol) HDO.59 While the hydrogenation
activity was high, the paper concluded that carbon-supported
precious metals are poor HDO catalysts because they favor ring
saturation. At 250 °C, the C−O bond between the ring and
methoxy group was broken, but C−O scission of the strong
aromatic-hydroxyl bond did not readily occur, producing
phenol as the major product.59 Additionally, ring hydro-
genation to form cyclohexanone and cyclohexanol occurred,
and a small amount of benzene was formed as well.59 Even at
harsher reaction conditions (350 °C and 39.5 atm), noble-
metal catalysts show poor selectivity to complete deoxygena-
tion.60
Supports with acid/base activity have been widely used to
achieve better deoxygenation yields from phenolic reactants.
The aromatic-O bonds of phenolic compounds are weakened
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
following ring hydrogenation, so it has been proposed that the
deoxygenation mechanism over noble metals involves hydro-
genation followed by dehydration rather than direct C−O
scission.32−34 In this model, one site is proposed to activate C−
O bonds. This component is often proposed to be a Brønsted
acid site on the oxide support materials. In addition, sites are
needed to supply atomic hydrogen to the reaction; this role is
played by the noble metal.39 Proposed mechanisms from model
compound studies on noble-metal catalysts consistently invoke
hydrogenation followed by dehydration, after which further
hydrogenation steps can occur.8,32,33 This mechanism is shown
in Figure 5.
Figure 5. Schematic representation of metal-acid catalysis proposed
for aromatic deoxygenation on noble-metal-based catalysts.
The deoxygenated product in Figure 5 is shown converting
between the aliphatic and aromatic compound to suggest that,
depending on the catalyst and the reaction conditions, either
could be possible. Aromatic oxygenates may be fully saturated
over the metal particles, particularly if the reaction conditions
favor this (e.g., at lower temperatures).32,33 Bimetallic particles
may be used to eliminate the need for an acidic support by
providing oxophilic sites that catalyze dehydration. For
example, Fe can be used to promote Pd deoxygenation activity
either by depositing Pd on an Fe2O3 support or by creating
PdFe particles on a carbon support.59,61 In other systems, an
acidic support such as alumina may still be needed for
dehydration activity. For example, PtCo and PtNi bimetallic
particles were prepared to increase hydrogenation activity and
therefore increase the rate of deoxygenation through sequential
hydrogenation followed by dehydration on the support.32
One report studying the vapor-phase HDO of m-cresol (3-
methylphenol) over Pt/Al2O3 found that a bifunctional
mechanism was required, involving hydrogenation activity to
form alcohol intermediates as well as acidity for dehydration.32
It was proposed that ring hydrogenation first occurs to saturate
the phenyl ring and a partially or fully hydrogenated alcohol
intermediate is formed. On Pt/Al2O3 the partially hydrogenated
alcohols may then deoxygenate on acidic sites on the support to
5031
Review
form toluene.32 This hypothesis was further supported by
reacting the fully hydrogenated alcohol intermediate (3-
methylcyclohexanol) on a bare Al2O3 support, where it was
deoxygenated to form methylcyclohexene isomers.32 The rate
was much higher when the saturated intermediate 3-
methylcyclohexanol was fed compared to the rate when m-
cresol or an alkene was fed. This result suggested that at 533 K
ring hydrogenation over Pt sites to form the alcohol
intermediates was rate limiting. The proposed rate-limiting
step of ring hydrogenation was used to explain the higher
selectivity of Pt/Al2O3 toward toluene than methylcyclohexane,
because fewer steps are required to dehydrate the unsaturated
alcohol intermediates than to produce saturated alkanes. Note
that this mechanism requires intimate contact between the
hydrogenation and deoxygenation sites to prevent over-
hydrogenation of the relatively unstable intermediate. Also,
despite bare alumina easily catalyzing dehydration of the
saturated alcohol, no reaction was observed when feeding m-
cresol; at least partial hydrogenation was required to allow
cleavage of the C−O bond.33 To improve the HDO activity, Ni
and Co were added to form bimetallic catalysts. These 3d
metals have previously been shown to increase hydrogenation
activity of unsaturated compounds.62 By increasing the
hydrogenation activity through these bimetallic catalysts the
HDO activity increased, and the product distribution shifted to
produce more methylcyclohexane relative to the amount of
toluene formed. Total deoxygenation also increased over the
bimetallic catalysts, proposed to be a result of generating
aluminate species through interaction between the 3d metal
and the support, which are active for alcohol dehydration.32
Pt/SiO2 was also studied to compare the alumina-supported
material to a less acidic catalyst.33 Although both contain acid
sites, the Lewis and Brønsted acidity on Al2O363,64 is stronger
than that of SiO2.65 The less acidic Pt/SiO2 was found to be
ineffective at dehydrating 3-methylcyclohexanol to the alkane
product. A kinetic study of Pt/SiO2 also concluded that this
dehydration pathway is limited.66 The dominant deoxygenation
pathway was instead proposed to be dehydration of unsaturated
alcohol intermediates to form toluene. While SiO2 alone was
not active for dehydration of 3-methylcyclohexanol, in the
presence of Pt, its weakly acidic hydroxyl groups can dehydrate
partially hydrogenated alcohol intermediates present on the Pt
particles, as shown in Figure 5.33 Both the Pt and SiO2 sites are
necessary for this reaction. When the support was doped with
potassium to suppress the acidity, the product distribution was
substantially altered. Rather than primarily toluene, m-cresol
formed 3-methylcyclohexanol with 84% selectivity.33 This is in
agreement with density functional theory (DFT) calculations
over Pt(111), which show that m-cresol HDO is less
energetically favorable than ring hydrogenation.67 However,
increasing the reaction temperature may increase the
favorability of the HDO pathway.68
Another study of support effects on noble metals also
concluded that acidic supports facilitate phenolic HDO. Pt, Rh,
Pd, and Ru supported on Al2O3, SiO2−Al2O3, and nitric-acid-
treated carbon black (NAC) were studied for the liquid-phase
batch upgrading of guaiacol.35 The support was found to have a
greater influence on the product distribution than the noble
metal. Using ammonia TPD, the acidity of the catalysts was
found to decrease in the order SiO2-Al2O3 > Al2O3 > NAC.35
The same trend was found in production of cyclohexane, the
fully deoxygenated and hydrogenated product. The least acidic
catalysts, those supported on NAC, had the highest yields of 2-
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
methoxycyclohexanol, a result of ring hydrogenation without
any C−O scission and an indication of low HDO activity.35
Therefore, HDO was again proposed to need a bifunctional
catalyst where acid sites are required to produce a
deoxygenated product rather than stop at a ring-saturated
oxygenate.
In addition to HDO through hydrogenation reactions
followed by dehydration, direct deoxygenation (DDO) has
been observed over supported metal catalysts. As shown in
Figure 1, DDO does not require intermediate hydrogenation
steps; instead, the C−O bond is directly cleaved from the
aromatic reactant. This may be desirable for reducing the
hydrogen demand of the deoxygenation process. One example
of DDO is using noble metals supported on reducible oxides
such as TiO2. For instance, Ru/TiO2 has been shown to
catalyze DDO of phenol to benzene through a bifunctional
mechanism at the interface between the metal and oxide.69,70
Using both experimental and computational techniques, it was
demonstrated that at 573 K and 3.8 MPa, water that is present
on the TiO2 may either accept or donate protons across the
metal−metal oxide interface. This allows the phenolic OH
group to be directly removed through donation of a proton,
shown in Figure 6. Although this is typically difficult to achieve
Figure 6. Proposed mechanism of DDO at the bifunctional Ru/TiO2
interface. (a) H2 is activated on Ru particles and heterolytically
dissociates across the interface to protonate a bridging hydroxyl group
on TiO2, generating a Brønsted acid site. (b) This then assists in
phenol deoxygenation. (c) Side view of DFT calculation of water-
assisted phenol DDO on Ru/TiO2, including initial state (left panel),
transition state (center panel), and final state (right panel). Color
code: hydrogen, white; oxygen, red; carbon, dark gray; ruthenium, teal;
titanium, light gray. Reproduced with permission from ref 69.
Copyright 2015 American Chemical Society.
due to the strong aromatic−oxygen bond, the bifunctional sites
present at the interface containing both metal and acid
functionalities were calculated to have a lower energy barrier
than hydrogenation. On large Ru particles, where the interfacial
area is much smaller, ring hydrogenation was found to be most
favorable as expected from previous studies of noble metals.
Isotopic labeling experiments in combination with DFT
calculations indicated that the Ru particles bind the phenol
and activate hydrogen to catalyze the hydride attack (which
regenerates the water cocatalyst on the support), whereas
Brønsted acidic water on TiO2 catalyzes C−O scission.
5032
Review
This illustrates the importance of better understanding the
metal−support interface, especially for reducible metal oxides
like TiO2 that may act as either an acid or a base. Similar results
have been observed for Pt/TiO2. The combination of a metal
and a reducible oxide support were shown to enhance m-cresol
deoxygenation by introducing additional energetically favorable
reaction pathways.71 While Pt/C preferred ring hydrogenation,
Pt/TiO2 catalyzed tautomerization and DDO. This synergistic
effect was particularly important at elevated temperatures (623
K).71
There has also been some interest in using zeolites as the
solid acid component of these bifunctional systems. While their
low activity toward phenolic reforming and high rates of coke
generation make zeolites potentially less attractive than the
other upgrading catalysts highlighted here, these issues can be
addressed by adding metals to the zeolite. Using guaiacol as a
model compound, only transalkylation was found to occur over
HY zeolite, resulting in no deoxygenated products.72 Similar
results were observed for Pt/Al2O3 in the absence of hydrogen.
However, phenol could be hydrogenated and deoxygenated
over several zeolites that had been doped with Pt.73 Pt/HY has
also been used for HDO of phenolic compounds, with
deoxygenation as well as coupling reactions forming larger
molecules.74
In addition to these purely heterogeneous examples,
supported metal catalysts have also been combined with liquid
acids to achieve high deoxygenation selectivity. For example, an
aqueous-phase phenolic mixture of pyrolysis oil model
compounds was upgraded over carbon-supported noble-metal
catalysts and a mineral acid.38 In neutral solution, the noble-
metal catalysts favored hydrogenation of phenol to cyclo-
hexanol. However, in the presence of phosphoric acid, the
alcohol was easily dehydrated. The proposed pathway for this
process is initial metal-catalyzed hydrogenation of the ring
followed by acid-catalyzed dehydration to form cyclohexene
and subsequent metal-catalyzed hydrogenation to ultimately
produce cyclohexane.38 This combination of a carbon-
supported noble metal with a mineral acid was found to be
effective for selective deoxygenation of a variety of phenolic
compounds in the aqueous phase, performing multistep
reactions involving hydrogenation, hydrolysis, and dehydration
steps. For instance, in neutral solution, all catalysts produced no
deoxygenated products at 100% conversion; however, upon
addition of phosphoric acid to provide bifunctionality to the
catalytic mixture in the form of acid sites, the deoxygenated
alkane product was formed with about 90% selectivity over all
metals.38
3.1.2. Supported Base Metals. While supported noble-metal
catalysts are promising for biomass HDO, their high cost has
sparked interest in studying less-expensive base metals such as
Ni, Co, and Fe. As with the noble-metal catalysts, it is useful to
first investigate the monofunctional behavior of these metals.
Fe/SiO2 was used for the gas-phase upgrading of guaiacol as a
model compound to study iron’s ability to break hydroxyl and
methoxyl bonds on an aromatic ring.75 The catalyst was found
to be selective for production of aromatic hydrocarbons over a
range of temperatures.75 Kinetic studies of Fe/SiO2 for guaiacol
upgrading have also been performed.76 Results from Fe
catalysts were compared to the same reaction over a Co
catalyst. Cobalt was found to be more active than Fe but was
shown to catalyze decomposition of the desirable products.76 A
simplified pathway is shown in Figure 7. Transalkylation
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis Review

reactions were proposed to occur over the acid sites, explaining

the formation of toluene from guaiacol.

Figure 7. Simplified pathway of guaiacol HDO on iron and cobalt

catalysts. Both iron and cobalt catalysts follow black arrows; only
cobalt catalyzes additional reactions, as shown in blue. Adapted from
ref 76.
Figure 8. Proposed reaction network for guaiacol HDO on precious
metals (Pd, Pt, Ru, pathways shown with black arrows) and base
A study comparing carbon-supported Fe and Cu to noble metals (Cu, Fe, pathways shown with red, dashed arrows). Adapted
metals obtained similar results for guaiacol HDO.59 Guaiacol from ref 59.
HDO was found to proceed through a phenol intermediate
over all catalysts, but unlike the noble metals, these base metals
did not lead to ring hydrogenation and C−C scission products.
For example, Fe/C was able to convert some of the phenol to
aromatic deoxygenated products such as benzene and
toluene.59 However, the activity of the base-metal catalysts
was lower than the noble metals. On Pd, Pt, and Ru the
intermediate phenol was primarily hydrogenated to form
cyclohexanone and cyclohexanol as expected in the absence
of an acid site for deoxygenation. In addition, ring opening to
form light gases also occurred. Meanwhile, Fe and Cu were not
active for ring saturation and ring opening reactions,
predominately forming benzene.59 This proposed network
can be seen in Figure 8.
These base metals show even more promising HDO activity
when used as part of a bifunctional metal-acid catalytic system.
For example, Raney Ni and Nafion on SiO2 were used for the
aqueous-phase HDO of a series of phenolic compounds.36 The
Ni was found to act as a hydrogenation catalyst, and the
Nafion/SiO2 provided acid sites for hydrolysis and dehydration,
resulting in high selectivity to hydrocarbon production.36 The Figure 9. Proposed reaction pathways of phenolic compounds over
proposed reaction pathway for phenolics over this catalyst Raney Ni and Nafion/SiO2. Adapted from ref 36.
system is depicted in Figure 9.
With Nafion/SiO2 still present in the aqueous mixture to hydrogen at the metal−support interface. The TPD trace for
provide acid sites, Ni/SiO2 has also shown activity for the smallest Ni loading had a small shoulder at this temperature
deoxygenation of phenol with the properly selected reaction along with a larger peak and additional high-temperature
temperature.77 Temperatures below 523 K did not catalyze shoulders at 720 and 760 K.77 The peak at 720 K was attributed
hydrogenolysis to form benzene; this was attributed to to hydrogen that is associated with the surface Ni sites, which
thermodynamic limitations. Instead, cyclohexanone and cyclo- has also been proposed to be responsible for hydrogenation
hexanol were produced. Higher Ni loading was shown to selectivity, while the peak at 760 K was attributed to hydrogen
increase activity as well as selectivity to benzene production, that had spilled over onto the support. Because the interfacial
while lower loadings of Ni produced more cyclohexanol and area was much higher for the catalysts with high Ni loading, it
cyclohexanone.77 This was proposed to be due to a greater was suggested that increases in the Ni loading increased the
population of hydrogen on the surface that can be used for availability of weakly bound hydrogen accessible for hydro-
hydrogenolysis.77 It was further proposed through H2 TPD genolysis to benzene.
results that on the larger particles the majority of the hydrogen A different bifunctional approach to deoxygenation using Ni
available is at metal−support interfacial sites, which are most is to utilize MgO as a basic support. Liquid-phase guaiacol
active for hydrogenolysis. Hydrogen TPD showed a sharp peak upgrading was studied over Ni/MgO at 433 K and 29.6 atm
at 663 K on large particles, which was proposed to come from H2.78 At 98% conversion, there was 100% selectivity to
5033 DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
demethoxylation to produce cyclohexanol. This high selectivity
was attributed to the acid−base interaction between the
phenolic OH group of guaiacol and the MgO support, with
Ni particles providing hydrogenation sites.78 Guaiacol deme-
thoxylation has also been observed over Ru/MgO79 and Ru-
MnOx/C, where in the latter case, the MnOx provides
basicity.80 Selective demethoxylation and hydrogenation may
be desirable as a renewable method of producing cyclohexanol
as a feedstock for polymer production or as an intermediate for
fuel production.
Although the previous literature convincingly demonstrates
that a combination of acid and metal sites leads to more
complete deoxygenation, it is important to point out that
metals themselves have some activity toward HDO reactions.
For reactants with relatively weak C−O bonds such an allyl
alcohol, furfuryl alcohol, and benzyl alcohol, deoxygenation can
even be observed on single crystals in ultrahigh vacuum.81−83
Deoxygenation of these compounds on supported metals has
often not been interpreted in terms of the support having a
role. Studies of such reactants on supported Pd catalysts have
shown that selective poisoning of different types of Pd sites can
be used to favor selectivity for deoxygenation.84,85 However, as
the studies investigating support effects show, the interaction
between the metal particle and the support can significantly
impact the surface chemistry. Going forward, a major challenge
will be to develop better models for the structure of metal−
support interface and to learn how to control interfacial active
sites. This is discussed in greater detail in the final section.
3.2. Bimetallics. The previous examples of transition-metal-
catalyzed HDO of phenolic compounds show that these
catalysts may be promising materials for pyrolysis oil upgrading
if acid sites are present. This mechanism (as shown in Figure 5)
may also take place over bimetallic catalysts. Oxide supports
such as alumina are often used as the source of acid sites, but
they also can strongly bind oxygenates like phenols, leading to
intermediates that block sites and can be precursors to coke
formation.33,86,87 Phenols and polyphenols have been shown to
be coke precursors over a wide range of catalysts, including
supported metals, oxides, and zeolites.88 This has prompted
studies of the effects of bimetallic catalysts containing a
hydrogenating metal and an oxophilic metal (an element that
forms strong metal−oxygen bonds) to provide intimate contact
between diverse sites for pyrolysis oil upgrading.89 Many of
these have been noble-metal-based alloys. Figure 10 shows one
proposed mechanism for the interaction of these bimetallic
catalysts with an organic oxygenate. The metals used as a
modifier site may be present in their metallic state or as an
oxide; regardless, their role is to introduce an additional active
site to produce a bifunctional nanoparticle. Table 2 shows
Figure 10. Hypothesized mechanism for interaction of oxygenates
with bimetallic surfaces containing a metal capable of dissociating
hydrogen and a metal that forms strong M−O bonds.
5034
Review
metal−oxygen bond strengths for some commonly used
oxophilic modifiers.
Table 2. Metal−Oxygen Bond Strengths for Oxophilic
Metals Used in Bimetallic Catalystsa
metal−oxygen bond bond strength (kJ mol−1) ref
oxophilic metals
Co−O 397 90
Cr−O 461 90
Fe−O 407 90
W−O 653 91
Mo−O 607 91
Re−Ob 644 92
hydrogenating metals
Ni−O 366 90
Pd−O 234 91
Pt−O 347 91
Rh−O 377 91
a
Values represent the bond dissociation energies measured at 298 K.
Both modifier sites used as the oxophilic site and hydrogenating metals
are included for comparison. bSpecifically for the Re−O bond in ReO3
The hydrogenating sites provide H while the oxophilic metal
more strongly binds the reactant to the surface through its
oxygen functionality, allowing easier C−O scission and
potentially creating new reaction pathways unavailable on a
monometallic catalyst. The majority of bimetallic catalysts
investigated have involved 3d metals as a modifier to noble
metals. These systems are known to create subsurface alloys
with the surface noble-metal atoms significantly electronically
modified.93,94 Subsurface transition-metal alloys have been
shown to increase HDO activity as well as selectivity toward
saturated hydrocarbons.32,95−97 DFT results show that the Pt-
terminated alloys bind adsorbates less strongly than pure Pt,
resulting in a more optimized adsorption strength which
increases hydrogenation.32 An investigation of the hydro-
genation of the unsaturated compounds benzene and 1,3-
butadiene over Pt and PtNi catalysts also observed that
hydrogenation rates increased with the Ni loading.97 DFT
calculations have suggested this could be due to shifting the d-
band center of surface Pt atoms by subsurface Ni, resulting in
weaker binding of the unsaturated compounds.98 The hydro-
genation activity was shown to have a volcano-curve-type
relationship with the binding energy of the reactant.98
Therefore, tuning of this binding strength allows maximization
of the hydrogenation activity. In addition to influencing
hydrogenation rates, it was proposed that the bimetallic
catalysts contained additional sites for dehydration. Metals
such as Co and Ni can create additional acid sites through
interaction with the alumina support to increase the rates of
dehydration reactions.32 This is proposed to be due to
formation of an aluminate species, which previous work has
shown increases the acidity on alumina.99 Because acid-
catalyzed dehydration steps are critical for C−O scission in
phenolic compounds, this increases HDO rates.
Many previous studies with bimetallic catalysts have been
performed using probe molecules such as glycerol. While the
precise role of the modifier is unknown, many studies have
shown increased activity and selectivity for C−O bond scission
reactions for a variety of probe molecules and metal
combinations.37,100−104 PtRe/C catalysts were shown to be
more active than monometallic Pt for converting glycerol to
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
syngas.37 Here, the role of the oxophilic metal may not be
directly related to its affinity for oxygen. Rather, CO-TPD
showed that CO binding energies were reduced on PtRe
compared to either monometallic sample, suggesting that
reduced site blocking could be at least partially responsible for
the increased activity.37 Another glycerol reforming study in
which PtRe/C was used for hydrogenolysis to form propane-
diols reported both increased rate and selectivity toward
deoxygenation of the secondary alcohol.37 Similar results were
obtained for RhRe/SiO2 catalysts.100 In trying to understand
the nature of these modifier sites, two main hypotheses have
been proposed. First, it was suggested that hydroxylated Re−
OH groups may be responsible for C−O activation by acting as
Brønsted acid sites, with the strong metal−oxygen bond
weakening the hydroxyl’s oxygen−hydrogen bond, thus
allowing the hydrogen to be more easily donated to a
neighboring oxygenate.37 NH3 TPD has shown increased
acidity on PtRe catalysts in the presence of water relative to
what is observed over the monometallic Pt sample.105 The
second hypothesis is that these oxophilic sites are strongly
anchoring the reactant to the catalyst surface through its oxygen
functionality, allowing surrounding sites to donate hydrogen for
deoxygenation.100 This strong interaction of the oxophilic site
and the oxygenated reactant can result in altered reaction
pathways. For example, m-cresol was shown to preferentially
follow the ring hydrogenation route to deoxygenation over pure
group VIII metals such as Ni, Pd, and Pt (shown by the
saturated compounds in Figure 5).34,71 However, when an
oxophilic modifier was added to the catalyst a tautomerization
pathway became accessible, creating a keto intermediate which
may then be dehydrated to form toluene. By making this
additional deoxygenation pathway energetically favorable the
deoxygenation activity is increased. This has been shown for m-
cresol HDO over PtMo/Al2O3, as illustrated in Figure 2c.12
Although more work is needed to fully understand the role of
oxophilic modifiers, model compound studies show their
promise for biomass deoxygenation. In addition to improving
deoxygenation selectivity, some studies using bimetallic
catalysts also reported DDO mechanisms. This is advantageous
if the goal is to remove oxygen with minimal hydrogen
consumption. Increased selectivity to C−O scission products
has been observed over noble-metal catalysts modified with
oxophilic metals such as Fe, Re, Mo, and W; the modifier sites
could potentially bind oxygenates more strongly to the surface
or create acid sites when they are hydroxylated by water present
under typical reaction conditions.37,58,101,102,104,106,107
In one study, PdFe alloys were used for the vapor-phase
upgrading of guaiacol.59 PdFe/C showed better selectivity
toward production of deoxygenated aromatics than either of the
monometallic catalysts, and no ring-saturation or ring-opening
products were observed. This study led to the hypothesis that
phenolics prefer to adsorb on the Pd-modified Fe sites.
Compared to the monometallic Fe surface, these sites
hypothetically have altered electronic properties and are in a
more reduced state due to increased hydrogen activation by Pd.
This results in a catalyst that can activate phenolic C−O bonds
without first needing to saturate the ring. This mechanism was
explored through DFT calculations by substituting Pd atoms
into a Fe(110) surface. Phenol was adsorbed with varying
distances from the Pd atoms. The results indicated that
adsorption and C−O cleavage are both more favorable on the
Fe sites.59 TPR results on the PdFe/C catalysts showed that Pd
assists in the reduction of FeOx.59 This evidence, along with
5035
Review
EXAFS and STEM/EDS results, was used to propose that Fe is
the active site for HDO and the role of Pd is facilitating
reduction of FeOx as well as modifying the Fe sites through
electronic effects due to alloy formation.59 STEM images of the
morphologies of these materials can be seen in Figure 11.
Figure 11. Representative aberration-corrected high-angle annular
dark field scanning transmission electron microscopy (AC-HAADF-
STEM) images of reduced and (left) and spent (right) Pd/Fe2O3
studied for m-cresol HDO. Brighter spots show Pd atoms, and darker
spots show Fe atoms. Reproduced with permission from ref 59.
Copyright 2013 Elsevier.
Bright regions with a diameter of 1−2 nm are Pd clusters on
the Fe surface; the darker spots underneath these are Fe
atoms.59 Both reduced and spent catalysts were imaged; the
comparison between these was used to suggest that there is
negligible change in morphology during reaction conditions,
and therefore, the Pd−Fe structure is robust.59
Similar results were obtained from a Pd/Fe2O3 catalyst for
HDO of m-cresol: Pd was proposed to activate hydrogen and
allow reduction of the Fe2O3 as well as prevent oxidation by
water formed during the HDO reaction.61 Addition of Pd was
also found to increase the activity compared to monometallic
Fe while maintaining a similar selectivity toward the production
of aromatic deoxygenated products.61
Cyclic oxygenates have also been studied for selective
hydrogenolysis on bimetallic catalysts. A common conclusion
among many of these reports is that the combination of a noble
metal and an oxophilic metal such as Re, Mo, or W can
dramatically increase selectivity to cleaving more sterically
hindered C−O bonds. For example, Rh catalysts were modified
with Re to improve selectivity to 1,6-hexanediol production
from tetrahydropyran-2-methanol (THPM).108,109 While Rh/C
had only 15% selectivity to this desired product, Rh-ReOx/C
had 96% selectivity as well as increasing the reaction rate from 2
to 180 mmol h−1 g catalyst−1. This improved activity was
proposed to be due to strong Re−OH interactions that allowed
an alkoxide species to form, strongly tethering the THPM to
the surface. Hydrogenolysis of the C−O bond adjacent to the
alkoxide group is favored and therefore results in high
selectivity to 1,6-hexanediol.108 A similar mechanism was
proposed using Rh-ReOx/SiO2 for tetrahydrofurfuryl alcohol110
and glycerol100,111 hydrogenolysis. Another study using Rh-
ReOx/C for hydrogenolysis of cyclic ethers proposed that this
increase in selectivity to C−O bond scission at the more
sterically hindered carbon may be due to the ability of OH
groups on ReOx sites to stabilize the intermediate carbenium
ions.112 To illustrate this mechanism, Figure 12 shows the
proposed reactant and intermediate adsorption states for
tetrahydrofurfuryl alcohol hydrogenolysis on Rh(111) modified
with an Re−OH site.
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis Review

been studied for m-cresol HDO.34 Over monometallic Ni,

Figure 12. DFT-calculated reactant and transition-state adsorption
structures for tetrahydrofurfuryl alcohol hydrogenolysis. The mecha-
nism is proposed to be a concerted protonation, hydride transfer, and
ring opening over Re−OH modified Rh(111). Reproduced with
permission from ref 112. Copyright 2011 American Chemical Society.
Increased hydrogenolysis selectivity has been observed for
acyclic oxygenates as well.112 Ir-ReOx/SiO2 showed higher
selectivity to 1,4-butanediol production from erythritol hydro-
genolysis than unmodified Ir.113 This increase in selectivity
from 0% to 33% was again proposed to be due to formation of
an alkoxide species on ReOx sites followed by hydrogenolysis
by a hydride species on a neighboring Ir site. Similar results
were found for glycerol hydrogenolysis on Ir-ReOx/SiO2, with
increased C−O scission attributed to formation of an
alkoxide.114−116 Characterization work on Ir-ReOx surfaces
has supported the hypothesis that interaction between the two
sites is important in the enhanced hydrogenolysis; a
combination of XAS, TPR, XPS, and CO chemisorption
suggested that there is significant interaction between the Ir
sites and the low-valent ReOx.117 A similar investigation for Rh-
ReOx also found that the two sites were well mixed and
interacting.118
Although bimetallics containing noble metals have been the
most extensively studied, another promising area is in the use of
bimetallics formed from base metals. NiFe/SiO2 catalysts have
methylcyclohexanone was found to be the major product, while
on the bimetallic catalyst, toluene was dominant. A physical
mixture of Ni/SiO2 and Fe/SiO2 did not show this change in
the reaction pathway, indicating that interactions between the
two metals are important. It was proposed that alloying with
oxophilic metals such as Fe weakened the interaction between
the aromatic ring and the surface, favoring strong binding with
the carbonyl group and leading to production of unsaturated
compounds.34 Studies like these have recently generated a great
deal of interest in better understanding bimetallic catalysts
containing less-expensive base metals for deoxygenation
applications.
A summary of these studies and the improvement of
deoxygenation over the monometallic catalyst can be seen in
Table 3. For both noble-metal and base-metal-containing
bimetallic catalysts, one of the main challenges is understanding
and controlling the active sites at the interface of the two
metals. Continuing to work toward a better, atomic-scale
picture of how the two metals interact under reaction
conditions will be a key challenge to overcome going forward.
3.3. Oxides. The discussion above mainly considers metal
oxides that serve as a carrier for supported metal nanoparticles,
and which require the presence of the metal to effectively
catalyze deoxygenation. However, some metal oxide surfaces
contain diverse types of active sites so that they themselves may
be considered bifunctional without the addition of other metals.
In particular, reducible metal oxides have shown deoxygenation
activity. For example, benzoic acid deoxygenation was studied
over ZnO and ZrO2.119 Hydrogen was used to create oxygen
vacancies in the oxide surface which could then be reoxidized
by the benzoic acid reactant to form benzaldehyde.119
However, these vacancies were susceptible to poisoning by
compounds like CO2 and water. In addition to model
compound studies, oxides have been studied for catalytic
upgrading of whole pyrolysis vapors. A ZrO2 and TiO2 mixed

Table 3. Summary of Bimetallic Catalysts Studied for Deoxygenationa

PH2 conversion % increase over
catalyst, wt % metals temperature (K), phase (atm) reactant (%) % deoxygenation monometallic ref
Pt−Co/Al2O3 533 0.5 m-cresol 57 97.1 21 32
1.3% Pt 4.3% Co vapor (1 atm pressure)
Pt−Ni/Al2O3 533 0.5 m-cresol 63 94.2 18.1 32
1.5% Pt 4.9% Ni vapor (1 atm pressure)
Pt−Re/C 443 39.5 glycerol 20 26b 26 37
5.7% Pt 4.6% Re aqueous
Rh-ReOx/SiO2 393 39.5 glycerol 10 22b 17 100
4% Rh, 0.5 molar ratio Rh/ReOx aqueous
Pt−Mo/Al2O3 523 4.6 m-cresol 95 96 76 12
5.1% Rh 1.8% ReOx vapor (5 atm pressure)
Pd−Fe/C 623 0.4 guaiacol 90 29 26 59
2% Pd 10% Fe vapor (1 atm pressure)
Pd−Fe/C 723 0.4 guaiacol 95 85 80 59
2% Pd 10% Fe vapor (1 atm pressure)
Rh-ReOx/C 373 39.5 THPM 36 97 58 108
4% Rh, 0.5 molar ratio Rh/ReOx aqueous
Rh-ReOx/SiO2 393 39.5 THFA 57 94c 76 110
4% Rh, 0.5 molar ratio Rh/ReOx aqueous

a
Reactants include m-cresol, glycerol, guaiacol, tetrahydropyran-2- methanol (THPM), and tetrahydrofurfuryl alcohol (THFA). bSelectivity to 1,3-
propanediol. cSelectivity to 1,5-pentanediol.

5036 DOI: 10.1021/acscatal.6b00923

ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
oxide catalyst increased the detected percent of hydrocarbons
to 13.1%, compared to 0.1% in the raw vapor.120
One of the most promising uses for oxide catalysts in
biomass upgrading is for ketonization. The carboxylic acids
present in pyrolysis oil are especially important to remove due
to their corrosivity and reactivity. One approach to eliminating
these functional groups while also increasing carbon chain
length is ketonization, or ketonic decarboxylation, which
converts two carboxylic acid molecules into a ketone with
water and CO2 formed as byproducts.121,122
R1COOH + R 2COOH → R1COR 2 + CO2 + H 2O
(1)
This is also beneficial because ketones can undergo aldol
condensation reactions to form larger molecules in the gasoline
to diesel range. The majority of studies for ketonization have
been done using oxide catalysts. Reducible metal oxides such as
CeO2, ZrO2, and TiO2 have been found to be more active and
selective than purely acidic or basic oxides.121 It has been
proposed that reducible oxides allow the carboxylic acids to
dissociatively adsorb to form carboxylates and hydroxyls on the
surface. These oxides also have available cation sites that
promote the adsorbed acids to interact and form the required
carboxylate intermediates.121 These two distinct types of active
sites were proposed to create a bifunctional surface capable of
improving activity for ketonization. An example of this
mechanism is shown in Figure 13.
Figure 13. Ketonization mechanism proposed over reducible metal
oxides. M = metal site. Adapted from ref 121. Reduced M+ sites act as
Lewis acids to bind the carboxylic acid and oxygen anions provide
Lewis basicity to deprotonate the adsorbed acid.
A combination of surface science and reactor studies has led
to proposing that the active site for these reactions contains
coordinatively unsaturated metal cations that act as Lewis acid
sites for binding of intermediates and oxygen anions, acting as
Lewis bases for deprotonation of the acid. The availability of
these sites as well as the redox properties of the material can be
influenced by adding dopants such as transition metals and
metal oxides.121,122 This can improve catalyst activity because
both types of active sites are important for ketonization to
5037
Review
occur; metal oxides that are not reducible and therefore lack the
metal cations for binding the oxygen functional group do not
show the same reactivity. These bifunctional surfaces are
effective because they provide a surface containing Lewis acid
sites for binding the intermediate in close proximity to Lewis
basic sites for deprotonation. More information on ketonization
catalysts and mechanisms can be found in past reviews.51,123
3.4. Sulfides. The most commonly used industrial hydro-
treating catalysts are sulfided molybdenum or tungsten
promoted by cobalt or nickel on an alumina support; these
materials are widely used in petroleum refining for hydro-
desulurization (HDS) and hydrodenitrogation (HDN).8 Due to
this large-scale production, these catalysts can already be made
economically and are well-characterized and understood in the
context of HDS and HDN. The wide variety of past studies has
shown that the edge area of these materials is important for
HDS activity, leading to the conclusion that edges are the active
sites for this reaction.124−127 However, the nature of the active
sites for HDO is still under debate. It is typically proposed that
the materials are bifunctional, with one site for activation of
heteroatom-C bonds and another for hydrogenation.39,40
Oxygen can adsorb to coordinatively unsaturated sites (sulfur
anion vacancies) at the catalyst edges.39,40 An example of this is
shown schematically in Figure 14.
Figure 14. Proposed mechanism of deoxygenation over transition-
metal sulfides.
Modifiers including Ni and Co donate electrons to the metal,
weakening the metal−S bond and allowing more vacancies to
be created.39,128 Hydrogen is thought to come from S−H
groups on the surface (formed either through activation of
molecular hydrogen or from decomposition of molecules such
as water or alcohols).8 This hydrogen can be used for HDO by
breaking the C−O bond or for decarbonylation/decarbox-
ylation by using acid sites to remove end groups from
aldehydes or carboxylic acids, respectively128 (see Figure 1).
The selectivity to a given pathway depends on reaction
conditions as well as the transition metal being used. For the
HDO pathway, unsaturated carbon bonds as well as carbonyl
groups are often first hydrogenated to form alcohols, which are
then dehydrated. An example of this can be seen for phenolics
and ketones in Figure 15.
Recent work has focused on materials that more clearly fit
the definition of bifunctional (i.e., incorporating additional
metals to create sites that assist in HDO of compounds relevant
to pyrolysis oil reforming).129−133 Previous investigations have
shown that Co and Ni can promote the HDO pathway over
hydrogenation or decarboxylation/decarbonylation for Mo2S
catalysts.130,131,134 Experimental and DFT work has suggested
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
Figure 15. Example mechanism proposed for the HDO pathway over
sulfides.
that incorporation of Ni−Mo sites lowers the activation barrier
for breaking aldehyde CO bonds and alcohol C−OH bonds
by altering the preferred adsorption orientation, enhancing
HDO over decarbonylation/decarboxylation or hydrogenation.
Instead of binding with just the oxygen group on top of the Mo
site, it was proposed that the bimetallic surface allows bidentate
binding where the carbon atom of the functional group (for
example, in a carbonyl) is in close proximity to an Ni−H site.
This lowers the energy of the transition state for HDO.
However, if too much Ni is incorporated, Ni3S2 is formed,
which increases selectivity to decarbonylation over HDO.131,134
These studies have also expanded the composition of
transition-metal sulfides being investigated, incorporating
multiple metals and modifiers, as well as studying support
effects. For example, on a W−Mo−S catalyst promoted by Ni,
tuning the W/Mo ratio was found to change the morphology of
the catalyst, influencing the number of coordinatively
unsaturated sites that have been previously proposed as the
active site for activating C−O bonds.129 This resulted in
increased activity as well as selectivity to deoxygenation.
3.5. Carbides, Nitrides, and Phosphides. Transition-
metal carbides, nitrides, and phosphides have also been
explored for HDO, showing promising hydrotreating abilities
because they combine the stability of ceramics with the
electronic properties of a metal.135,136 These materials are well-
characterized HDS and HDN catalysts due to their use in
petroleum refining, but less attention has been given to their
applications for HDO until recently.
A current hypothesis for the mechanism of deoxygenation
over transition-metal phosphides relies on the presence of both
metallic and acidic groups.41−43 For example, it has been
proposed that for Ni2P, P−OH groups could contribute
Brønsted acid sites while Ni sites add metallic behavior.41−43
Nickel phosphide has received the most attention as an HDO
catalyst due to its activity and selectivity.136 For example,
anisole HDO over supported Ni2P has been investigated; the
proposed reaction scheme involves demethylation to form
phenol followed by hydrogenolysis and hydrogenation.42 The
active sites on this catalyst are proposed to be Ni, modified by
its P neighbors.8 Ni sites with 4 neighboring P atoms, or Ni(1)
sites, have a tetrahedral geometry and are active for
decarbonylation; Ni(2) sites have 5 P neighbors in a square
pyramidal geometry and catalyze HDO.8,136,137 Bulk Ni2P
contains equal amounts of Ni(1) and Ni(2), but the surface is
enriched with Ni(2), and this enrichment increases as the
particle size decreases. This has been confirmed experimentally
using a range of particle sizes and measuring both coordination
numbers as well as Ni/P ratios.136 Increasing the number of
Ni(2) sites correlated with an increase in hydrotreating activity
and a decrease in decarbonylation.136
A series of phosphides including Ni2P, Co2P, Fe2P, WP, and
MoP supported on SiO2 were studied for guaiacol HDO and
5038
Review
compared to a Pd/Al2O3 catalyst at the same conditions.138 A
proposed reaction scheme for guaiacol HDO on these catalysts
can be seen in Figure 16. The Pd catalyst had a much higher
Figure 16. Proposed guaiacol HDO reaction pathways on transition-
metal phosphides. Proposed intermediates shown in parentheses.
Adapted from ref 138.
activity than the phosphides under the experimental conditions
but formed primarily catechol, an undesired product that also is
a precursor to coke formation.138 Ni2P was the most active of
the phosphides tested and avoided this pathway, favoring
production of phenol, benzene, and methoxybenzene.138 The
decrease in catechol formation may be attributed to the P−OH
groups on the phosphide that are proposed to add acidity to the
surface.41−43 The suppression of catechol production on
phosphides is promising; however, additional studies are
needed to improve selectivity to fully deoxygenated products.
The selectivity toward phenol and methoxybenzene seen in this
study are not ideal for a final upgrading catalyst.
Carbides have been compared to platinum group metals,
showing similar catalytic properties; reactions that would not
occur on W or Mo alone are observed over the carbide surface,
because binding of the metal to carbon atoms shifts the energy
of the transition-metal d electrons to lower energies.137,139−141
The ability of tungsten and molybdenum carbides to remove
heteroatoms suggests they could be promising catalysts for
biomass deoxygenation. For example, through DFT calculations
as well as surface science and gas-phase reactor studies using
propanol and propanal as reactants, WC was shown to activate
C−O bonds but not C−C; the dominant product formed was
propene.139 The catalysts have again been suggested to be
bifunctional in nature. The presence of oxygen on the WC
surface was proposed to introduce acidic sites, on which
dehydration reactions are faster than when oxidized W or WC
is not present.139 A similar effect of surface oxygen has been
proposed for other carbide surfaces as well. For the HDO of
model compounds over NiMo carbide, it was proposed that
metallic sites as well as metal−OH sites are generated under
reaction conditions, leading to enhanced dehydration activity.44
Mo2C has also been studied for HDO and has shown
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
promising selectivity similar to that of WC.142−144 In addition
to these C3 oxygenates, studies have also been performed using
larger, aromatic oxygenates. For instance, a study of group 4, 5,
and 6 metal carbides and nitrides investigated the HDO of
benzofuran at 643 K and 3.1 MPa.145 The carbides tested
(Mo2C, WC, NbC, and VC) showed poor selectivity to HDO
(about 9, 10, 5, and 3% respectively). As a comparison, sulfided
Ni−Mo/Al2O3, which is a common industrial hydrotreating
catalyst, showed over 60% selectivity to HDO.
Nitrides have been studied extensively, particularly for
hydrogenation/dehydrogenation reactions. As discussed for
carbides, the electronic effects induced by alloying can create a
material that has some catalytic properties similar to noble
metals.146 They have been explored as hydrotreating catalysts,
primarily for HDS and HDN, but there has been increasing
interest in using these materials for HDO.146,147 Nitrides are
also thought to be bifunctional, with active sites being both
coordinatively unsaturated metal atoms as well as 4-fold
nitrogen vacancies.148,149 Through studying guaiacol HDO, it
was proposed that demethoxylation and dehydroxylation occur
over the unsaturated Mo sites and demethylation is promoted
by nitrogen vacancies.148 When Mo2N was studied for guaiacol
HDO, similar results were obtained for that of the phosphides
reported above, avoiding the pathway to produce catechol but
producing mainly phenol rather than achieving complete
deoxygenation.150 Mo2N also showed low selectivity to HDO
in the study by Ramanathant and Oyama using benzofuran.145
All nitrides tested had low selectivity (<10%) to deoxygenation
with the exception of VN, which had about 50% HDO
selectivity. Although this is lower than the control case of
sulfided Ni−Mo/Al2O3 (which had over 60% selectivity to
HDO), VN was relatively selective to the aromatic product
ethylbenzene, whereas the traditional hydrotreating catalyst
favored ethylcyclohexane. It was proposed that the much more
promising activity of VN than any other nitride or carbide
tested was due to its optimized binding strength of the reactant;
a volcano curve plotting HDO activity as a function of
benzofuran binding energy peaked near VN.145
While carbides, nitrides, and phosphides are effective
hydrotreating catalysts for HDS and HDN, overall they do
not seem to perform as well as other materials for HDO, such
as transition-metal-based catalysts. Most of the studies
summarized here report low activity and/or selectivity to
deoxygenation. However, they have not been thoroughly
investigated, and their low cost compared to noble-metal-
based catalysts makes them attractive to study further. While
hypotheses exist regarding the nature of active sites for HDO
on these materials, a more fundamental understanding of their
surface, particularly how the surface is influenced by the
presence of oxygen, would aid in designing better catalysts. A
good comparison of a variety of carbide, nitride, and phosphide
catalysts for deoxygenation of various oxygenates can be found
in a recent review.8 One promising area to explore in greater
depth is supported nanoparticles of these materials, which may
be expected to have different surface structures and facets
exposed than bulk phosphides, carbides, or nitrides. For
example, nanoparticle Rh2P/SiO2 and Ni2P/SiO2 were studied
for guaiacol deoxygenation and showed similar levels of
deoxygenation selectivity to transition-metal catalysts.151
4. CONCLUSIONS AND OUTLOOK
Bifunctional materials have been widely studied for use as
deoxygenation catalysts. The key reason that bifunctional
5039
Review
materials are important for deoxygenation is the necessity of
both hydrogenation and acid/base sites. While many bifunc-
tional combinations have demonstrated promising deoxygena-
tion activity, key questions about the mechanism remain. In
particular, there is a critical need to understand the role of the
interface between the different active sites in the reaction
mechanism. In most studies to date, the density of interfacial
sites has not been explicitly controlled, so it is difficult to
determine the role proximity between distinct types of sites
plays in the reaction mechanism. It is important to resolve these
questions because design of improved catalysts would ideally
involve tailoring the interface between the two types of catalyst
functionalities.
Certain materials lend themselves to studies in which the
interface structure can be more systematically varied. Bimetallic
catalysts in which different metals at the surface play different
catalytic roles (e.g., PdFe catalysts in which Pd serves as a
hydrogenation center and Fe serves as a deoxygenation center)
hypothetically allow the density of one metal to be varied with
respect to another and for parameters such as average ensemble
size to be varied in a controlled manner. Surface-level
investigations with model surfaces have previously been
employed to understand how specific arrangements of metal
atoms lead to improved performance for other types of
reactions.152−155 Such combinations of metals are also relatively
straightforward to investigate using computational methods. In
all these investigations, care must be taken to ensure that the
surface composition used in modeling the reaction is
representative of that present under reaction conditions.
Thus, use of in situ and operando spectroscopies that are
surface-sensitive, such as near-ambient pressure X-ray photo-
electron spectroscopy, will provide important information for
tracing catalyst performance back to the true surface
composition.156−158
Metal−metal oxide interfaces are significantly more difficult
to model but are clearly of interest for combining metal and
acid/base activity. There have been many recent advances in
the computational modeling of such interfaces that show the
importance of interfacial sites in facilitating catalysis.69,159,160
Employing materials’ synthesis techniques that enable better
control over interfacial sites is one likely method for
establishing structure−property relations in deoxygenation
catalysis. For example, techniques such as atomic layer
deposition could be employed to generate well-defined
coverages of different metal and metal oxide layers on both
model surfaces and high-surface area catalysts.161−167 Another
opportunity for generating control over the interface (and
being able to employ reliable computational modeling
approaches) is to employ metal nanoparticles that are very
well-defined in their structure,168−174 including the use of single
metal atoms as catalysts.175−178
Finally, work using tethered organocatalysts suggests an
interesting path for designing bifunctional catalysts for biomass
conversions. In addition to inorganic catalysts, tethered organic
acids and bases are a promising route for acid−base chemistry.
A significant research effort has been devoted to studies of
these bifunctional (or “cooperative”) catalysts; one of the key
concepts in this area is controlling the spatial arrangement of
the two functionsfor example, an acid and a baseto allow
simultaneous interaction of the two functions with different
positions on a reactant molecule or a transition state. Part of
the motivation for this approach is to mimic the behavior of
enzymes, where spatial arrangement of multiple catalytic
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
domains can be used to enable specific catalysis.179 This
approach can be used to combine catalyst functional groups
that would be incompatible in solution. For example, acidic
sulfonic acid groups have been cotethered with basic amine
groups on a silica support, shown in Figure 17.180 The
Figure 17. Example of the catalytic cycle proposed for a tethered
bifunctional catalyst in the aldol condensation of 4-nitrobenzaldehyde
and acetone at 323 K. Adapted from ref 191.
bifunctional catalyst was found to be more active for a model
aldol condensation reaction than the monofunctional controls.
Interestingly, it was also more active than a physical mixture of
acid- and base-functionalized materials, indicating that prox-
imity of the acid and base sites is important in maximizing
catalyst rate. Whereas in this example the proximity of the
functional groups was not controlled in a direct way, other
examples of tethered bifunctional catalysts have sought to tailor
the proximity of groups with high precision, for example
through molecular imprinting approaches.181 Other work has
exploited the bifunctional effect of a tethered organic catalyst
combined with acidic groups on the support.182 A related
example is the use of molecular catalysts for homogeneously
catalyzing chiral syntheses. Using transition-metal-based
complexes that contain both a Lewis acid and Lewis base,
C−H, C−C, C−O, and C−N bonds may be formed.183,184
Through proper tuning of the bifunctional complex issues with
acid−base neutralization can be mitigated and high degrees of
enantioselective reduction and C−C bond formation may be
achieved.185
As an example of the promise of these catalysts, aldol
condensation reactions may be considered. These reactions of
aldehydes such as furfurals with short ketones followed by
hydrogenation can yield gasoline/diesel range hydrocarbons
(C8−C13).186−188 Coupling reactions like aldol condensation
may be improved by utilizing a bifunctional catalyst rather than
a single acidic or basic material.189 The two components of this
bifunctional catalyst can each activate one of the substrates to
be reacted, resulting in a lower energy pathway.190 These
catalysts are often composed of an acid tuned to activate one of
the substrates for electrophilic addition and a base that is
selected to increase the nucleophilicity of the other substrate.
5040
Review
For example, Figure 17 shows the aldol condensation of 4-
nitrobenzaldehyde and acetone using a bifunctional tethered
catalyst consisting of a sulfonic acid and an amine immobilized
on a support.191 In this example, the acetone is deprotonated to
create a nucleophilic enolate which may then attack the
electron-deficient aldehyde to form the aldol product.191 Proper
design and synthesis prevents the acid and base from
neutralizing each other and creates a unique bifunctional
catalyst.190 A similar approach is taken by many enzymes, which
operate at high selectivity.189 By being able to stabilize
incompatible functional groups in a way that allows them to
maintain their individual functionalities, enzymes such as
adolases are able to catalyze multistep reactions that would
not take place in solution.189 Mimicking this, the field of
heterogeneous catalysis has adopted a similar approach for
these acid−base-catalyzed reactions.
There is the potential to integrate the use of well-defined
tethered organocatalysts with reactive surfaces such as metal
oxides. Bifunctional surfaces have already been employed for
such reactions. For example, this has been demonstrated using
solid base catalysts in single-phase, mixed oxides containing
both acidic and basic sites, as well as on novel catalysts at the
water/oil interface of an emulsion.186−188,192 The high water
content of pyrolysis oil can cause phase separation; taking
advantage of differing partition coefficients of these reactant
molecules could lead to improved selectivites to targeted
products. Smaller aldehydes have also been used as probe
molecules to better understand condensation reaction mecha-
nisms over novel materials. For example, the mixed oxide
Ce0.5Zr0.5O2 has been studied for propanal conversion to C6−
C9 products.192
Overall, the approach of using bifunctional (or multifunc-
tional) catalysts is an important direction for catalysis in general
and particularly for reactions that require significantly different
types of elementary steps as part of an overall reaction. For this
reason, efforts to employ rational design of interfaces will
become important for supplementing the promising initial work
in demonstrating bifunctional catalysis for biomass conversion.
■ AUTHOR INFORMATION
Corresponding Author
*E-mail: will.medlin@colorado.edu.
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
This work was supported by the Department of Energy
Bioenergy Technologies Office (BETO) under Contract no.
DE-AC36-08-GO28308. A.M.R. acknowledges support from
the National Science Foundation (Award CHE-1464979) and
partial support from the Department of Education Graduate
Assistantships in Areas of National Need (GAANN).
■
REFERENCES
(1) Bridgwater, A. V. Biomass Bioenergy 2012, 38, 68−94.
(2) Di Blasi, C. Prog. Energy Combust. Sci. 2009, 35, 121−140.
(3) Bridgwater, A.; Peacocke, G. Renewable Sustainable Energy Rev.
2000, 4, 1−73.
(4) Pham, T. N.; Shi, D.; Resasco, D. E. Appl. Catal., B 2014, 145,
10−23.
(5) Mettler, M. S.; Vlachos, D. G.; Dauenhauer, P. J. Energy Environ.
Sci. 2012, 5, 7797−7809.
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
(6) Lu, Q.; Li, W.-Z.; Zhu, X.-F. Energy Convers. Manage. 2009, 50,
1376−1383.
(7) Zacher, A. H.; Olarte, M. V.; Santosa, D. M.; Elliott, D. C.; Jones,
S. B. Green Chem. 2014, 16, 491−515.
(8) Ruddy, D. A.; Schaidle, J. A.; Ferrell, J. R., III; Wang, J.; Moens,
L.; Hensley, J. E. Green Chem. 2014, 16, 454−490.
(9) Ingram, L.; Mohan, D.; Bricka, M.; Steele, P.; Strobel, D.;
Crocker, D.; Mitchell, B.; Mohammad, J.; Cantrell, K.; Pittman, C. U.,
Jr Energy Fuels 2008, 22, 614−625.
(10) Xu, J.; White, T.; Li, P.; He, C.; Yu, J.; Yuan, W.; Han, Y.-F. J.
Am. Chem. Soc. 2010, 132, 10398−10406.
(11) Bitter, J.; Seshan, K.; Lercher, J. J. Catal. 1998, 176, 93−101.
(12) Robinson, A. M.; Ferguson, G. A.; Gallagher, J. R.; Cheah, S.;
Beckham, G. T.; Schaidle, J. A.; Hensley, J. E.; Medlin, J. W. ACS
Catal. 2016, 6, 4356−4368.
(13) Kitchin, J.; Nørskov, J. K.; Barteau, M.; Chen, J. J. Chem. Phys.
2004, 120, 10240−10246.
(14) Stamenkovic, V. R.; Mun, B. S.; Mayrhofer, K. J.; Ross, P. N.;
Markovic, N. M. J. Am. Chem. Soc. 2006, 128, 8813−8819.
(15) Stamenkovic, V. R.; Mun, B. S.; Arenz, M.; Mayrhofer, K. J.;
Lucas, C. A.; Wang, G.; Ross, P. N.; Markovic, N. M. Nat. Mater. 2007,
6, 241−247.
(16) Chen, J. G.; Menning, C. A.; Zellner, M. B. Surf. Sci. Rep. 2008,
63, 201−254.
(17) Liu, P.; Nørskov, J. K. Phys. Chem. Chem. Phys. 2001, 3, 3814−
3818.
(18) Rebelli, J.; Detwiler, M.; Ma, S.; Williams, C. T.; Monnier, J. R. J.
Catal. 2010, 270, 224−233.
(19) Maroun, F.; Ozanam, F.; Magnussen, O.; Behm, R. Science 2001,
293, 1811−1814.
(20) Neurock, M.; Mei, D. Top. Catal. 2002, 20, 5−23.
(21) Lu, Y.-C.; Xu, Z.; Gasteiger, H. A.; Chen, S.; Hamad-Schifferli,
K.; Shao-Horn, Y. J. Am. Chem. Soc. 2010, 132, 12170−12171.
(22) Zhang, Y.; Zhang, H.; Ma, Y.; Cheng, J.; Zhong, H.; Song, S.;
Ma, H. J. Power Sources 2010, 195, 142−145.
(23) Kong, F.-D.; Zhang, S.; Yin, G.-P.; Zhang, N.; Wang, Z.-B.; Du,
C.-Y. J. Power Sources 2012, 210, 321−326.
(24) Dau, H.; Limberg, C.; Reier, T.; Risch, M.; Roggan, S.; Strasser,
P. ChemCatChem 2010, 2, 724−761.
(25) Chiusoli, G. P.; Maitlis, P. M. Metal-catalysis in Industrial Organic
Processes, 3rd ed.; Chiusoli, G. P. M., Peter, M, Eds.; Royal Society of
Chemistry: Cambridge, U.K., 2008.
(26) Ono, Y. Catal. Today 2003, 81, 3−16.
(27) Gong, J. Chem. Rev. 2012, 112, 2987−3054.
(28) Fu, Q.; Saltsburg, H.; Flytzani-Stephanopoulos, M. Science 2003,
301, 935−938.
(29) Fu, Q.; Deng, W.; Saltsburg, H.; Flytzani-Stephanopoulos, M.
Appl. Catal., B 2005, 56, 57−68.
(30) Sakurai, H.; Ueda, A.; Kobayashi, T.; Haruta, M. Chem.
Commun. 1997, 271−272.
(31) Xu, M.; Gines, M. J.; Hilmen, A.-M.; Stephens, B. L.; Iglesia, E. J.
Catal. 1997, 171, 130−147.
(32) Do, P. T.; Foster, A. J.; Chen, J.; Lobo, R. F. Green Chem. 2012,
14, 1388−1397.
(33) Foster, A. J.; Do, P. T.; Lobo, R. F. Top. Catal. 2012, 55, 118−
128.
(34) Nie, L.; de Souza, P. M.; Noronha, F. B.; An, W.; Sooknoi, T.;
Resasco, D. E. J. Mol. Catal. A: Chem. 2014, 388-389, 47−55.
(35) Lee, C. R.; Yoon, J. S.; Suh, Y.-W.; Choi, J.-W.; Ha, J.-M.; Suh,
D. J.; Park, Y.-K. Catal. Commun. 2012, 17, 54−58.
(36) Zhao, C.; Kou, Y.; Lemonidou, A. A.; Li, X.; Lercher, J. A. Chem.
Commun. 2010, 46, 412−414.
(37) Daniel, O. M.; DeLaRiva, A.; Kunkes, E. L.; Datye, A. K.;
Dumesic, J. A.; Davis, R. J. ChemCatChem 2010, 2, 1107−1114.
(38) Zhao, C.; Kou, Y.; Lemonidou, A. A.; Li, X.; Lercher, J. A.
Angew. Chem. 2009, 121, 4047−4050.
(39) Mortensen, P. M.; Grunwaldt, J.-D.; Jensen, P. A.; Knudsen, K.;
Jensen, A. D. Appl. Catal., A 2011, 407, 1−19.
(40) Furimsky, E. Appl. Catal., A 2000, 199, 147−190.
5041
Review
(41) Chen, J.; Shi, H.; Li, L.; Li, K. Appl. Catal., B 2014, 144, 870−
884.
(42) Li, K.; Wang, R.; Chen, J. Energy Fuels 2011, 25, 854−863.
(43) Lee, Y.-K.; Oyama, S. T. J. Catal. 2006, 239, 376−389.
(44) Zhang, W.; Zhang, Y.; Zhao, L.; Wei, W. Energy Fuels 2010, 24,
2052−2059.
(45) Talukdar, A. K.; Bhattacharyya, K. G.; Sivasanker, S. Appl. Catal.,
A 1993, 96, 229−239.
(46) Park, C.; Keane, M. A. J. Colloid Interface Sci. 2003, 266, 183−
194.
(47) Zhao, C.; He, J.; Lemonidou, A. A.; Li, X.; Lercher, J. A. J. Catal.
2011, 280, 8−16.
(48) Columbia, M.; Thiel, P. J. Electroanal. Chem. 1994, 369, 1−14.
(49) Mavrikakis, M.; Barteau, M. A. J. Mol. Catal. A: Chem. 1998,
131, 135−147.
(50) Ferrin, P.; Simonetti, D.; Kandoi, S.; Kunkes, E.; Dumesic, J. A.;
Nørskov, J. K.; Mavrikakis, M. J. Am. Chem. Soc. 2009, 131, 5809−
5815.
(51) Rajadurai, S. Catal. Rev.: Sci. Eng. 1994, 36, 385−403.
(52) Columbia, M.; Crabtree, A.; Thiel, P. J. Am. Chem. Soc. 1992,
114, 1231−1237.
(53) Hung, W.-H.; Bernasek, S. L. Surf. Sci. 1996, 346, 165−188.
(54) Davis, J.; Barteau, M. Surf. Sci. 1991, 256, 50−66.
(55) Erley, W.; Sander, D. J. Vac. Sci. Technol., A 1989, 7, 2238−2244.
(56) Wildschut, J.; Mahfud, F. H.; Venderbosch, R. H.; Heeres, H. J.
Ind. Eng. Chem. Res. 2009, 48, 10324−10334.
(57) Sheu, Y.-H. E.; Anthony, R. G.; Soltes, E. J. Fuel Process. Technol.
1988, 19, 31−50.
(58) Bu, Q.; Lei, H.; Zacher, A. H.; Wang, L.; Ren, S.; Liang, J.; Wei,
Y.; Liu, Y.; Tang, J.; Zhang, Q.; Ruan, R. Bioresour. Technol. 2012, 124,
470−477.
(59) Sun, J.; Karim, A. M.; Zhang, H.; Kovarik, L.; Li, X. S.; Hensley,
A. J.; McEwen, J.-S.; Wang, Y. J. Catal. 2013, 306, 47−57.
(60) Chang, J.; Danuthai, T.; Dewiyanti, S.; Wang, C.; Borgna, A.
ChemCatChem 2013, 5, 3041−3049.
(61) Hong, Y.; Zhang, H.; Sun, J.; Ayman, K. M.; Hensley, A. J.; Gu,
M.; Engelhard, M. H.; McEwen, J.-S.; Wang, Y. ACS Catal. 2014, 4,
3335−3345.
(62) Lu, S.; Lonergan, W. W.; Bosco, J. P.; Wang, S.; Zhu, Y.; Xie, Y.;
Chen, J. G. J. Catal. 2008, 259, 260−268.
(63) Chapman, K. W.; Sava, D. F.; Halder, G. J.; Chupas, P. J.;
Nenoff, T. M. J. Am. Chem. Soc. 2011, 133, 18583−18585.
(64) Medema, J.; Van Bokhoven, J.; Kuiper, A. J. Catal. 1972, 25,
238−244.
(65) Auroux, A. Top. Catal. 1997, 4, 71−89.
(66) Nie, L.; Resasco, D. E. J. Catal. 2014, 317, 22−29.
(67) Tan, Q.; Wang, G.; Nie, L.; Dinse, A.; Buda, C.; Shabaker, J.;
Resasco, D. E. ACS Catal. 2015, 5, 6271−6283.
(68) Chen, C.; Chen, G.; Yang, F.; Wang, H.; Han, J.; Ge, Q.; Zhu, X.
Chem. Eng. Sci. 2015, 135, 145−154.
(69) Nelson, R. C.; Baek, B.; Ruiz, P.; Goundie, B.; Brooks, A.;
Wheeler, M. C.; Frederick, B. G.; Grabow, L. C.; Austin, R. N. ACS
Catal. 2015, 5, 6509−6523.
(70) Newman, C.; Zhou, X.; Goundie, B.; Ghampson, I. T.; Pollock,
R. A.; Ross, Z.; Wheeler, M. C.; Meulenberg, R. W.; Austin, R. N.;
Frederick, B. G. Appl. Catal., A 2014, 477, 64−74.
(71) Griffin, M. B.; Ferguson, G. A.; Ruddy, D. A.; Biddy, M. J.;
Beckham, G. T.; Schaidle, J. A. ACS Catal. 2016, 6, 2715−2727.
(72) Nimmanwudipong, T.; Runnebaum, R. C.; Block, D. E.; Gates,
B. C. Energy Fuels 2011, 25, 3417−3427.
(73) Horácě k, J.; Št’ávová, G.; Kelbichová, V.; Kubička, D. Catal.
Today 2013, 204, 38−45.
(74) Hong, D.-Y.; Miller, S. J.; Agrawal, P. K.; Jones, C. W. Chem.
Commun. 2010, 46, 1038−1040.
(75) Olcese, R. N.; François, J.; Bettahar, M.; Petitjean, D.; Dufour,
A. Energy Fuels 2013, 27, 975−984.
(76) Olcese, R.; Bettahar, M.; Petitjean, D.; Malaman, B.; Giovanella,
F.; Dufour, A. Appl. Catal., B 2012, 115-116, 63−73.
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis
(77) Shin, E.-J.; Keane, M. A. Ind. Eng. Chem. Res. 2000, 39, 883−
892.
(78) Long, J.; Shu, S.; Wu, Q.; Yuan, Z.; Wang, T.; Xu, Y.; Zhang, X.;
Zhang, Q.; Ma, L. Energy Convers. Manage. 2015, 105, 570−577.
(79) Nakagawa, Y.; Ishikawa, M.; Tamura, M.; Tomishige, K. Green
Chem. 2014, 16, 2197−2203.
(80) Ishikawa, M.; Tamura, M.; Nakagawa, Y.; Tomishige, K. Appl.
Catal., B 2016, 182, 193−203.
(81) Pang, S. H.; Medlin, J. W. ACS Catal. 2011, 1, 1272−1283.
(82) Pang, S. H.; Román, A. M.; Medlin, J. W. J. Phys. Chem. C 2012,
116, 13654−13660.
(83) Davis, J.; Barteau, M. J. Mol. Catal. 1992, 77, 109−124.
(84) Pang, S. H.; Schoenbaum, C. A.; Schwartz, D. K.; Medlin, J. W.
Nat. Commun. 2013, 4, 1−6.
(85) Lien, C.-H.; Medlin, J. W. J. Phys. Chem. C 2014, 118, 23783−
23789.
(86) Huber, G. W.; Iborra, S.; Corma, A. Chem. Rev. 2006, 106,
4044−4098.
(87) Popov, A.; Kondratieva, E.; Goupil, J. M.; Mariey, L.; Bazin, P.;
Gilson, J.-P.; Travert, A.; Maugé, F. J. Phys. Chem. C 2010, 114,
15661−15670.
(88) Meloni, D.; Monaci, R.; Solinas, V.; Berlier, G.; Bordiga, S.;
Rossetti, I.; Oliva, C.; Forni, L. J. Catal. 2003, 214, 169−178.
(89) Alonso, D. M.; Wettstein, S. G.; Dumesic, J. A. Chem. Soc. Rev.
2012, 41, 8075−8098.
(90) Haynes, W. M. CRC Handbook of Chemistry and Physics; CRC
Press: New York, 2014.
(91) Dean, J. A. Lange’s Handbook of Chemistry; McGraw-Hill: New
York, 1985.
(92) Gisdakis, P.; Antonczak, S.; Rösch, N. Organometallics 1999, 18,
5044−5056.
(93) Skoplyak, O.; Barteau, M. A.; Chen, J. G. G. Catal. Today 2009,
147, 150−157.
(94) Skoplyak, O.; Barteau, M. A.; Chen, J. G. G. J. Phys. Chem. B
2006, 110, 1686−1694.
(95) Lu, S.; Lonergan, W. W.; Bosco, J. P.; Wang, S.; Zhu, Y.; Xie, Y.;
Chen, J. G. J. Catal. 2008, 259, 260−268.
(96) Lu, S.; Menning, C. A.; Zhu, Y.; Chen, J. G. ChemPhysChem
2009, 10, 1763−1765.
(97) Lonergan, W. W.; Vlachos, D. G.; Chen, J. G. J. Catal. 2010,
271, 239−250.
(98) Humbert, M. P.; Chen, J. G. J. Catal. 2008, 257, 297−306.
(99) Stanislaus, A.; Absi-Halabi, M.; Al-Dolama, K. Appl. Catal. 1989,
50, 237−245.
(100) Shinmi, Y.; Koso, S.; Kubota, T.; Nakagawa, Y.; Tomishige, K.
Appl. Catal., B 2010, 94, 318−326.
(101) Kunkes, E. L.; Simonetti, D. A.; Dumesic, J. A.; Pyrz, W. D.;
Murillo, L. E.; Chen, J. G.; Buttrey, D. J. J. Catal. 2008, 260, 164−177.
(102) Simonetti, D.; Kunkes, E.; Dumesic, J. J. Catal. 2007, 247,
298−306.
(103) Lee, J.; Kim, Y. T.; Huber, G. W. Green Chem. 2014, 16, 708−
718.
(104) Chia, M.; Pagán-Torres, Y. J.; Hibbitts, D.; Tan, Q.; Pham, H.
N.; Datye, A. K.; Neurock, M.; Davis, R. J.; Dumesic, J. A. J. Am. Chem.
Soc. 2011, 133, 12675−12689.
(105) Zhang, L.; Karim, A. M.; Engelhard, M. H.; Wei, Z.; King, D.
L.; Wang, Y. J. Catal. 2012, 287, 37−43.
(106) Pallassana, V.; Neurock, M. J. Catal. 2002, 209, 289−305.
(107) Ma, L.; He, D. Top. Catal. 2009, 52, 834−844.
(108) Chen, K.; Koso, S.; Kubota, T.; Nakagawa, Y.; Tomishige, K.
ChemCatChem 2010, 2, 547−555.
(109) Koso, S.; Nakagawa, Y.; Tomishige, K. J. Catal. 2011, 280,
221−229.
(110) Koso, S.; Furikado, I.; Shimao, A.; Miyazawa, T.; Kunimori, K.;
Tomishige, K. Chem. Commun. 2009, 2035−2037.
(111) Amada, Y.; Koso, S.; Nakagawa, Y.; Tomishige, K.
ChemSusChem 2010, 3, 728−736.
5042
Review
(112) Chia, M.; Pagan-Torres, Y. J.; Hibbitts, D.; Tan, Q.; Pham, H.
N.; Datye, A. K.; Neurock, M.; Davis, R. J.; Dumesic, J. A. J. Am. Chem.
Soc. 2011, 133, 12675−12689.
(113) Amada, Y.; Watanabe, H.; Hirai, Y.; Kajikawa, Y.; Nakagawa,
Y.; Tomishige, K. ChemSusChem 2012, 5, 1991−1999.
(114) Tomishige, K.; Tamura, M.; Nakagawa, Y. Chem. Rec. 2014, 14,
1041−1054.
(115) Nakagawa, Y.; Ning, X.; Amada, Y.; Tomishige, K. Appl. Catal.,
A 2012, 433-434, 128−134.
(116) Amada, Y.; Shinmi, Y.; Koso, S.; Kubota, T.; Nakagawa, Y.;
Tomishige, K. Appl. Catal., B 2011, 105, 117−127.
(117) Amada, Y.; Watanabe, H.; Tamura, M.; Nakagawa, Y.;
Okumura, K.; Tomishige, K. J. Phys. Chem. C 2012, 116, 23503−
23514.
(118) Koso, S.; Watanabe, H.; Okumura, K.; Nakagawa, Y.;
Tomishige, K. J. Phys. Chem. C 2012, 116, 3079−3090.
(119) De Lange, M.; Van Ommen, J.; Lefferts, L. Appl. Catal., A
2001, 220, 41−49.
(120) Lu, Q.; Zhang, Y.; Tang, Z.; Li, W.-z.; Zhu, X.-f. Fuel 2010, 89,
2096−2103.
(121) Pham, T. N.; Sooknoi, T.; Crossley, S. P.; Resasco, D. E. ACS
Catal. 2013, 3, 2456−2473.
(122) Gaertner, C. A.; Serrano-Ruiz, J. C.; Braden, D. J.; Dumesic, J.
A. J. Catal. 2009, 266, 71−78.
(123) Renz, M. Eur. J. Org. Chem. 2005, 2005, 979−988.
(124) Iwata, Y.; Araki, Y.; Honna, K.; Miki, Y.; Sato, K.; Shimada, H.
Catal. Today 2001, 65, 335−341.
(125) Chianelli, R. R.; Siadati, M. H.; De la Rosa, M. P.; Berhault, G.;
Wilcoxon, J. P.; Bearden, R., Jr; Abrams, B. L. Catal. Rev.: Sci. Eng.
2006, 48, 1−41.
(126) Tauster, S.; Pecoraro, T.; Chianelli, R. J. Catal. 1980, 63, 515−
519.
(127) Bahl, O.; Evans, E.; Thomas, J. Proc. R. Soc. London, Ser. A
1968, 306, 53−65.
(128) Romero, Y.; Richard, F.; Brunet, S. Appl. Catal., B 2010, 98,
213−223.
(129) Wang, W.; Zhang, K.; Qiao, Z.; Li, L.; Liu, P.; Yang, Y. Catal.
Commun. 2014, 56, 17−22.
(130) Bui, V. N.; Laurenti, D.; Afanasiev, P.; Geantet, C. Appl. Catal.,
B 2011, 101, 239−245.
(131) Ruinart de Brimont, M.; Dupont, C.; Daudin, A.; Geantet, C.;
Raybaud, P. J. Catal. 2012, 286, 153−164.
(132) Yoosuk, B.; Tumnantong, D.; Prasassarakich, P. Fuel 2012, 91,
246−252.
(133) Popov, A.; Kondratieva, E.; Mariey, L.; Goupil, J. M.; El Fallah,
J.; Gilson, J.-P.; Travert, A.; Maugé, F. J. Catal. 2013, 297, 176−186.
(134) Dupont, C.; Lemeur, R.; Daudin, A.; Raybaud, P. J. Catal.
2011, 279, 276−286.
(135) Oyama, S. Catal. Today 1992, 15, 179−200.
(136) Oyama, S. T.; Gott, T.; Zhao, H.; Lee, Y.-K. Catal. Today 2009,
143, 94−107.
(137) Hwu, H. H.; Chen, J. G. Chem. Rev. 2005, 105, 185−212.
(138) Zhao, H.; Li, D.; Bui, P.; Oyama, S. Appl. Catal., A 2011, 391,
305−310.
(139) Ren, H.; Chen, Y.; Huang, Y.; Deng, W.; Vlachos, D. G.; Chen,
J. G. Green Chem. 2014, 16, 761−769.
(140) Weigert, E. C.; Stottlemyer, A. L.; Zellner, M. B.; Chen, J. G. J.
Phys. Chem. C 2007, 111, 14617−14620.
(141) Levy, R.; Boudart, M. Science 1973, 181, 547−549.
(142) Ren, H.; Yu, W.; Salciccioli, M.; Chen, Y.; Huang, Y.; Xiong,
K.; Vlachos, D. G.; Chen, J. G. ChemSusChem 2013, 6, 798−801.
(143) Han, J.; Duan, J.; Chen, P.; Lou, H.; Zheng, X.; Hong, H. Green
Chem. 2011, 13, 2561−2568.
(144) Jongerius, A. L.; Gosselink, R. W.; Dijkstra, J.; Bitter, J. H.;
Bruijnincx, P. C.; Weckhuysen, B. M. ChemCatChem 2013, 5, 2964−
2972.
(145) Ramanathan, S.; Oyama, S. J. Phys. Chem. 1995, 99, 16365−
16372.
(146) Hargreaves, J. Coord. Chem. Rev. 2013, 257, 2015−2031.
DOI: 10.1021/acscatal.6b00923
ACS Catal. 2016, 6, 5026−5043
ACS Catalysis Review

(147) Dolce, G.; Savage, P.; Thompson, L. Energy Fuels 1997, 11, (176) Boucher, M. B.; Zugic, B.; Cladaras, G.; Kammert, J.;
668−675. Marcinkowski, M. D.; Lawton, T. J.; Sykes, E. C. H.; Flytzani-
(148) Ghampson, I. T.; Sepúlveda, C.; Garcia, R.; Radovic, L. R.; Stephanopoulos, M. Phys. Chem. Chem. Phys. 2013, 15, 12187−12196.
Fierro, J.; DeSisto, W. J.; Escalona, N. Appl. Catal., A 2012, 439-440, (177) Flytzani-Stephanopoulos, M.; Gates, B. C. Annu. Rev. Chem.
111−124. Biomol. Eng. 2012, 3, 545−574.
(149) Frapper, G.; Pélissier, M.; Hafner, J. J. Phys. Chem. B 2000, 104, (178) Zhai, Y.; Pierre, D.; Si, R.; Deng, W.; Ferrin, P.; Nilekar, A. U.;
11972−11976. Peng, G.; Herron, J. A.; Bell, D. C.; Saltsburg, H.; Mavrikakis, M.;
(150) Ghampson, I.; Sepulveda, C.; Garcia, R.; Frederick, B.; Flytzani-Stephanopoulos, M. Science 2010, 329, 1633−1636.
Wheeler, M.; Escalona, N.; DeSisto, W. Appl. Catal., A 2012, 413-414, (179) Margelefsky, E. L.; Zeidan, R. K.; Davis, M. E. Chem. Soc. Rev.
78−84. 2008, 37, 1118−1126.
(151) Griffin, M. B.; Baddour, F. G.; Habas, S. E.; Ruddy, D. A.; (180) Zeidan, R. K.; Hwang, S. J.; Davis, M. E. Angew. Chem., Int. Ed.
2006, 45, 6332−6335.
Schaidle, J. A. Top. Catal. 2016, 59, 124−137.
(181) Katz, A.; Davis, M. E. Nature 2000, 403, 286−289.
(152) Ruff, M.; Takehiro, N.; Liu, P.; Nørskov, J. K.; Behm, R. J.
(182) Notestein, J. M.; Katz, A. Chem. - Eur. J. 2006, 12, 3954−3965.
ChemPhysChem 2007, 8, 2068−2071.
(183) Sawamura, M.; Ito, Y. Chem. Rev. 1992, 92, 857−871.
(153) Sachtler, J.; Somorjai, G. J. Catal. 1983, 81, 77−94.
(184) Takamura, M.; Funabashi, K.; Kanai, M.; Shibasaki, M. J. Am.
(154) Chen, M.; Kumar, D.; Yi, C.-W.; Goodman, D. W. Science
Chem. Soc. 2000, 122, 6327−6328.
2005, 310, 291−293. (185) Ikariya, T. In Bifunctional Molecular Catalysis; Springer: Tokyo,
(155) Scott, R. W.; Sivadinarayana, C.; Wilson, O. M.; Yan, Z.; 2011.
Goodman, D. W.; Crooks, R. M. J. Am. Chem. Soc. 2005, 127, 1380− (186) Chheda, J. N.; Dumesic, J. A. Catal. Today 2007, 123, 59−70.
1381. (187) Zapata, P. A.; Faria, J.; Ruiz, M. P.; Resasco, D. E. Top. Catal.
(156) Divins, N. J.; Angurell, I.; Escudero, C.; Pérez-Dieste, V.; 2012, 55, 38−52.
Llorca, J. Science 2014, 346, 620−623. (188) Crossley, S.; Faria, J.; Shen, M.; Resasco, D. E. Science 2010,
(157) Tao, F.; Grass, M. E.; Zhang, Y.; Butcher, D. R.; Aksoy, F.; 327, 68−72.
Aloni, S.; Altoe, V.; Alayoglu, S.; Renzas, J. R.; Tsung, C.-K.; Zhu, Z.; (189) Brunelli, N. A.; Jones, C. W. J. Catal. 2013, 308, 60−72.
Liu, Z.; Salmeron, M.; Somorjai, G. A. J. Am. Chem. Soc. 2010, 132, (190) Brunelli, N. A.; Venkatasubbaiah, K.; Jones, C. W. Chem. Mater.
8697−8703. 2012, 24, 2433−2442.
(158) Tao, F.; Dag, S.; Wang, L.-W.; Liu, Z.; Butcher, D. R.; Bluhm, (191) Zeidan, R. K.; Davis, M. E. J. Catal. 2007, 247, 379−382.
H.; Salmeron, M.; Somorjai, G. A. Science 2010, 327, 850−853. (192) Gangadharan, A.; Shen, M.; Sooknoi, T.; Resasco, D. E.;
(159) Saavedra, J.; Doan, H. A.; Pursell, C. J.; Grabow, L. C.; Mallinson, R. G. Appl. Catal., A 2010, 385, 80−91.
Chandler, B. D. Science 2014, 345, 1599−1602.
(160) Green, I. X.; Tang, W.; Neurock, M.; Yates, J. T. Science 2011,
333, 736−739.
(161) Stair, P.; Marshall, C.; Xiong, G.; Feng, H.; Pellin, M.; Elam, J.;
Curtiss, L.; Iton, L.; Kung, H.; Kung, M.; Wang, H.-H. Top. Catal.
2006, 39, 181−186.
(162) Lu, J.; Stair, P. C. Angew. Chem., Int. Ed. 2010, 49, 2547−2551.
(163) Lu, J.; Fu, B.; Kung, M. C.; Xiao, G.; Elam, J. W.; Kung, H. H.;
Stair, P. C. Science 2012, 335, 1205−1208.
(164) Pellin, M.; Stair, P.; Xiong, G.; Elam, J.; Birrell, J.; Curtiss, L.;
George, S.; Han, C.; Iton, L.; Kung, H.; Kung, M.; Wang, H.-H. Catal.
Lett. 2005, 102, 127−130.
(165) Wank, J. R.; George, S. M.; Weimer, A. W. J. Am. Ceram. Soc.
2004, 87, 762−765.
(166) Li, J.; Liang, X.; King, D. M.; Jiang, Y.-B.; Weimer, A. W. Appl.
Catal., B 2010, 97, 220−226.
(167) Gould, T. D.; Lubers, A. M.; Neltner, B. T.; Carrier, J. V.;
Weimer, A. W.; Falconer, J. L.; Medlin, J. W. J. Catal. 2013, 303, 9−15.
(168) Lei, Y.; Mehmood, F.; Lee, S.; Greeley, J.; Lee, B.; Seifert, S.;
Winans, R. E.; Elam, J. W.; Meyer, R. J.; Redfern, P. C.; Teschner, D.;
Schlogl, R.; Pellin, M. J.; Curtiss, L. A.; Vajda, S. Science 2010, 328,
224−228.
(169) Vajda, S.; Pellin, M. J.; Greeley, J. P.; Marshall, C. L.; Curtiss, L.
A.; Ballentine, G. A.; Elam, J. W.; Catillon-Mucherie, S.; Redfern, P. C.;
Mehmood, F.; Zapol, P. Nat. Mater. 2009, 8, 213−216.
(170) Yin, Y.; Lu, Y.; Gates, B.; Xia, Y. J. Am. Chem. Soc. 2001, 123,
8718−8729.
(171) Xu, Z.; Xiao, F.-S.; Purnell, S.; Alexeev, O.; Kawi, S.; Deutsch,
S.; Gates, B. Nature 1994, 372, 346−348.
(172) Gates, B. C. J. Mol. Catal. A: Chem. 2000, 163, 55−65.
(173) Bratlie, K. M.; Lee, H.; Komvopoulos, K.; Yang, P.; Somorjai,
G. A. Nano Lett. 2007, 7, 3097−3101.
(174) Kwon, G.; Ferguson, G. A.; Heard, C. J.; Tyo, E. C.; Yin, C.;
DeBartolo, J.; Seifert, S. n.; Winans, R. E.; Kropf, A. J.; Greeley, J.;
Johnston, R. L.; Curtiss, L. A.; Pellin, M. J.; Vajda, S. ACS Nano 2013,
7, 5808−5817.
(175) Matsubu, J. C.; Yang, V. N.; Christopher, P. J. Am. Chem. Soc.
2015, 137, 3076−3084.

5043 DOI: 10.1021/acscatal.6b00923

ACS Catal. 2016, 6, 5026−5043