Recent Developments in Palladium-Based BimetallicCatalystsFenglin Liao, Tsz Woon Benedict Lo, and Shik Chi Edman Tsang*[a]

1. Introduction

Platinum-group metals (PGM) constitute the most studied cata-

lyst component for nearly 200 years.[1] In the 19th century, plat-inum was first used as a catalyst in industrial processes such as

the oxidation of sulfur dioxide to sulfuric acid, the oxidation ofethanol to acetic acid or acetaldehyde, and catalytic combus-

tion.[2] Thereafter, the development of new catalysts based onplatinum-group metals continued to bloom. The main reasons

for the selection of platinum-group elements include their in-

trinsic high catalytic ability under mild conditions, high selec-tivity towards a specific product, high stability under various

reaction conditions, and tolerance to poisons.[2] Although plati-num is the most widely used element in catalysis, palladium is

receiving increasing attention owing to its similar catalyticproperties to platinum and wider availability.

Palladium organometallic compounds (homogeneous cata-

lysts) or supported Pd-based nanoparticles (NPs; heterogene-ous catalysts) are both known to be active for a wide range ofreactions, which include hydrogenation, oxidation, dehydro-genation, hydrogenolysis, coupling reactions, carbonylation,

carbohydrate reforming, and hydrodesulphurization.[3–5] Al-though organometallic catalysts commonly exhibit impressive

activity and selectivity with clearly defined catalytic mecha-nisms, the applications in industry remain challenging becausethey are expensive, generally nonrecyclable, and difficult toseparate from the product mixture.[6–15] This minireview mainlyfocuses on the developments of some recently reported sup-

ported palladium bimetallic catalysts that may be more appli-

cable to industrial practice. Herein, the terms “alloy” or “bimet-

allic” are used interchangeably without much emphasis onwhether the added foreign atoms form a true solid solution, is-

lands, or core–shell morphology with Pd atoms.The intrinsic catalytic properties of Pd nanoparticles origi-

nate from the fundamental electronic configuration of its con-stituent element (4d105s0), which can be strongly influenced by

coordination environment.[16] By alloying with a secondary ele-

ment, anchoring to a support, or controlling particle size, thecoordination environment and the electronic band structure of

Pd NPs can be controlled and manipulated, which make it dis-tinct from that of monometallic Pd. Alloying is one of the

mostly investigated methods owing to a wide range of choicefor the alloying element and well-acquired skills for their syn-

theses. The alloying method, through inserting a secondary el-

ement into the Pd lattice, can affect the coordination environ-ment of Pd in two ways: i) formation of heteroatomic bonds

changes the electronic environment of the metal, giving rise tomodifications in the electronic configuration of Pd through the

so called “ligand effect”; ii) the geometry of the bimetallicstructure is typically different from its parent metal, for exam-

ple, the changes of lattice structure such as metal–metal bondlength and bond angle, which play important roles in funda-mental catalysis processes.[17] Meanwhile, the co-existence of

a foreign metallic atom may also provide a different adsorptionsite for some reactants or intermediates resulting a co-cata-

lyzed mechanism. Therefore, alloy nanoparticles consisting oftwo different elements have attracted great interest in the

field of science and technology.[18–20] Although the alteration of

catalytic performance by the alloying approach is well-known,further elucidation of structural change with respect to activity,

selectivity, or stability is not always given.[17, 21]

The purpose of this article is to provide a short and concise

review to cover some recent results on synthesis, characteriza-tion, and testing of Pd-based bimetallic nanoparticles for

As one of the alternative metals to platinum, which supportsa wide range of applications in chemistry and catalysis in in-

dustry, palladium increasingly receives attention because of its

similar physicochemical properties. However, Pd is generallyless expensive than Pt and has a richer natural reserve. Herein,

some recently developed techniques for the preparation andcharacterization of Pd-based bimetallic catalysts are reviewed.

The impact on catalytic reactions of interest, including hydro-

genation, dehydrogenation, hydrogenolysis, reforming, theoxygen reduction reaction, and hydrodesulfurization are also

discussed. It is shown that the catalytic performance of Pd-

based bimetallic catalysts is strongly dependent on the geo-metric and electronic states of Pd, which can be significantly

affected by blended foreign element(s). Rationalization of thestructure–activity relationship can provide useful guidelines to

the fine tuning of these important catalytic reactions.

[a] F. Liao, T. W. B. Lo, Prof. S. C. E. TsangWolfson Catalysis Centre, Department of ChemistryUniversity of OxfordOxford, OX1 3QR (UK)E-mail : edman.tsang@chem.ox.ac.uk

This publication is part of a Special Issue on Palladium Catalysis. To viewthe complete issue, visit :http://onlinelibrary.wiley.com/doi/10.1002/cctc.v7.14/issuetoc

ChemCatChem 2015, 7, 1998 – 2014 Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim1998

MinireviewsDOI: 10.1002/cctc.201500245

a number of important catalytic reactions. By comparison ofcatalytic performance and Pd structure, a more in-depth struc-

ture–activity relationship of the Pd catalysts can be estab-lished.

2. Preparation of Bimetallic NPs

Since the 1960s, when the importance of bimetallic NPs beganto be recognized, many methods have been developed andcan be classified into two categories—unsupported colloidalNPs and supported bimetallic NPs. Yonezawa and Toshimahave thoroughly reviewed the methods of preparing unsup-ported colloidal bimetallic NPs, in which a large amount of or-

ganic surfactants are used as stabilizers.[22] Consequently, thesurface of these bimetallic NPs is largely covered by a layer of

surfactant molecules, which reduces the number of exposedmetal sites for catalysis. Also, the small colloidal NPs without

support are more prone to aggregate under the reaction con-ditions after the total removal or partial removal of the surfac-

tant. Therefore, colloidal metallic NPs are mainly used as

model materials to investigate the alloy properties for compari-son with their monometallic NPs.[22] In contrast, supported

alloy or bimetallic NPs as catalysts are practically more usefulfor a wider range of industrial reactions.

2.1. Co-impregnation

Co-impregnation is one of the most compatible methods for

the preparation of supported bimetallic or alloy Pd-basedNPs.[16] Briefly, an inert support is immersed into a solution mix-

ture composed of Pd and the secondary element precursorswith a volume equal to the total pore volume of the support

material. With the evaporation of solvent, the complex precur-

sors of Pd and the secondary element are simultaneously de-posited on the surface of the support. After thermal reduction

at elevated temperature, Pd-based NPs can be synthesized ac-cordingly. This method for making PdCu bimetallic NPs was

successfully achieved by Haller and co-workers.[23] Pd(NO3)2 andCu(NO3)2 were used as the precursors with variable mole ratiosto control the ratio of Pd/Cu in their resulting alloy NPs. Thismethod is convenient and simple, but the precursors in solu-

tion generally have poor interaction with each other, hencethe formation of bimetallic species is not expected to takeplace in the impregnation or deposition step. Instead, the for-mation of bimetallic NPs precursor is thought to take place onthe surface of the support during the subsequent thermal or

reduction steps.[24] Thus, the composition of the bimetallic NPson the support surface could be unevenly distributed if the

latter processes (thermal or reduction steps) are improperly

carried out. For example, too rapid thermal heating or reduc-tion of deposited precursors on the support surface may result

in phase segregation. Also, compared to other methods de-scribed later in this review, the particle size prepared by this

method is relatively large because of the required high tem-perature for thermal reduction.

2.2. Sequential impregnation

In the co-impregnation method, the two elemental precursorsof bimetallic NPs are deposited on the surface of the support

simultaneously. In the process of sequential impregnation thetwo elements are introduced sequentially, as the name sug-

gests. The procedure involves several steps: A metallic com-plex containing one element (Pd or the modifier element) is

first deposited on an inert support. Then, a subsequent ther-

mal treatment at elevated temperature is required to partiallyremove the ligands of the central metal atom. Thereafter, the

secondary-element-containing complex is introduced to reactwith this unsaturated species with the formation of supported

bimetallic species. Following pre-reduction in a hydrogen at-mosphere, the bimetallic NPs can be produced. The synthesisof Pd-based alloys by this method is widely reported in the lit-

erature (PdPt by Auvray and Olsson,[25] PdRh by Araya andWeissmann,[26] PdCu by Yaakob and co-workers[27]) The compo-

sition of the alloy NPs prepared in this way can be very uni-form if the synthesis procedure is carefully controlled, also the

concentration of the produced bimetallic sites is generallyhigher than that produced by the conventional co-impregna-

tion method.[24] However, the whole procedure involves many

steps and usually requires exquisite thermal treatments, whichmay be difficult to handle on an industrial scale.[24]

2.3. Thermal decomposition of molecular bimetallic clusters

The thermal decomposition of molecular bimetallic clusters is

another promising method for the preparation of highly dis-persed supported Pd-based alloy or bimetallic catalysts. The

rapid developments in the synthesis of organometallic com-

plexes provides a wide range of heteroatomic metal clusters ofPd incorporated with other elements of tunable sizes.[28] Thiscatalyst synthesis method involves adsorption of a bimetalliccluster precursor on a support, followed by a thermal treat-ment in an inert atmosphere or hydrogen to remove the li-gands from the cluster frame. Braunstein et al. reported that

extremely small PdFe clusters can be prepared by the decom-position of [FePd2(CO)4(dppm)2] (dppm = 1,1-bis(diphenylphos-phino)methane) or [Fe2Pd2(CO)5(NO)2(dppm)2] on support

SiO2.[29] Typical transmission electron microscopy (TEM) imagesare shown in Figure 1; the average particle size can be as small

Figure 1. TEM images of a) PdFe bimetallic NPs on SiO2 produced by thethermal decomposition of PdFe organometallic compounds; b) PdFe parti-cles after catalysis. Taken from Ref. [29] .

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim1999

Minireviews

as 1–2 nm. The metallic NPs are normally well-dispersed andthe composition is uniform. The key factor for success relies on

strong bonding between Pd and the other element in the mo-lecular cluster. If the interaction

is not strong enough, comparedto the affinity for the individual

metal with the support surface,the heteroatomic cluster frame-

work is prone to segregate upon

adsorption, resulting in the for-mation of two separated mono-

metallic NPs on surface. Thus,this preparation method typical-

ly requires the precise synthesisof a stable organo–bimetalliccomplex. However, the draw-

backs of this method are thehigh cost for the complex syn-

thesis and the thermal instabilityof the produced bimetallic NPson the surface of the support.After the formation of bimetallic

NPs on the support (SiO2, Al2O3)

the weak interaction betweenthe support and NPs does not

offer much attenuation to the extent of their aggregation orsintering on the surface. As shown in Figure 1 b, the small par-

ticle size of PdFe bimetallic NPs prepared by Braunstein et al.was increased to 10 nm or more after their catalysis reaction.

2.4. Co-reduction of PdO/support oxide

Metal nanoparticles supported on high surface area oxides arethe most widely used catalysts in industrial reactions.[30] It has

been demonstrated that many catalytic reactions require a spe-

cific strong metal and support interaction (SMSI) to give a highcatalytic performance.[31] The underlying principle for the modi-

fication of metal nanoparticles by the support material is notwell-understood. However, it is becoming clear that one of theimportant causes for SMSI is the decoration of supportedmetal particles with reactive metallic elements derived from

the deep reduction of the support. This is more pronounced insupported noble-metal systems with NPs capable of activating

and transporting hydrogen for the deep reduction of the sup-port. Taking a traditional Cu/ZnO/Al2O3 catalyst for methanolproduction from synthesis gas as an example, recent research

using advanced and sensitive instruments has shown thata very small quantity of Zn atoms are reduced from ZnO at the

interface of Cu/ZnO by atomic H produced from the Cu site,generating catalytically active CuZn sites. CuZn is believed to

exhibit distinct electronic properties from those of monometal-

lic Cu, and gives enhanced catalytic performance.[32, 33] Similarly,Zhang and Chin et al. reported that PdZn can be formed

during the hydrogen pretreatment of PdO/ZnO. Pd is moreactive than Cu in dihydrogen activation to form active H

atoms, which promote more extensive reduction of ZnO.[34, 35]

The PdZn species were characterized by X-ray diffraction (XRD)

and transmission electron microscopy (TEM) and are shown inFigure 2.[34] Oyola-Rivera et al. reported the in situ formation of

the Pd2Ga phase from PdO/Ga2O3 during the CO2 hydrogena-

tion to methanol and dimethyl ether.[36] From these results, it isclear that co-reduction of PdO/support oxide can be used as

a synthetic tool for the preparation of Pd-based bimetallic oralloy NPs. Owing to the strong interaction with the support,

the supported Pd-containing NPs prepared by this method are

well-dispersed and also exhibit high stability in the catalytic re-action. However, the as-synthesized metallic NP is usually Pd

rich owing to the reduction tolerance of the support oxideunder mild conditions, leading to a limited production of the

modifier atom from the support reduction (it is noted thatmild reduction conditions are required to maintain the small

particle size of the metallic NPs and high-temperature treat-

ment would cause sintering of NPs). Consequently, the compo-sition is not easy to tune.

Recently, our group has disclosed a new method. By addingthe appropriate additive to the support material, the reductionbehavior of the supported oxide can be controlled rationallywith the assistance of Pd at a temperature of �100 8C. Thus,the composition of the formed Pd alloys with particle sizes as

small as 1–2 nm can be tunable.[37] The role of the additive isto establish a type II electronic heterojunction in the compo-

site semiconductor supports, which extend the lifetime of ther-mal- or photogenerated excitons owing to the spatial separa-

tion of electrons and holes of the support materials with differ-ent electronic energy levels, as shown in Scheme 1 a.[38–40] The

accumulated holes, in the form of electron-depleted [O] spe-

cies, in semiconductor 2 of lower energy react readily with theatomic H produced by Pd in the proximity to form water be-

cause their long dwell time allows the chemical reaction tooccur. Simultaneously, semiconductor 1 at higher energy facili-

tates the reduction of cations to the metallic state by their ex-cited electrons (as shown in Scheme 1 b). In contrast, for

Figure 2. Left) XRD patterns of formed PdZn alloys from the reduction of Pd/ZnO with a Pd loading amount ofa) 2 %, b) 4.8 %, and c) 9.1 %. Right) a, b, c) TEM images and d) area-indexed image of Fourier transforms of sup-ported PdZn alloy in the Pd/ZnO catalyst with a Pd loading of 9.1 %. Taken from Ref. [34] . Copyright 2014,American Chemical Society.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2000

Minireviews

a single-phase semiconductor oxide support, the majority of

their excited electrons and holes (excitons) will recombine rap-idly, blocking further reduction processes until high tempera-

ture is applied. Taking the synthesis of PdFe as an example,with the blending of limited amounts of ZnII into the PdO/Fe2O3 complex, a type II heterojunction of ZnFe2O4-Fe2O3 (Zn

reacts with Fe2O3 to form ZnFe2O4 in excess Fe2O3) is estab-lished with thermal excited electrons accumulated in Fe2O3

(lower energy) and holes residing in ZnFe2O4 (higher energy).Under hydrogen pretreatment, catalytic reduction of FeIII to Fe0

is strongly promoted at �100 8C by removing the active [O]species (holes). The characterization results [XPS and extended

X-ray absorption fine structure(EXAFS)] indicate that theamount of produced Fe0 is de-

termined by the ratio ofZnFe2O4/Fe2O3 in the support.

With the incorporation of Pd atthe interface, supported PdFe

NPs with high dispersion and

a tunable composition can beprepared successfully.

As a result, the Pd alloys pro-duced by this co-reduction

method are highly stable, well-dispersed, and have a tunable

composition. Thus, this method could be a superior and practi-cal method for the preparation of supported bimetallic cata-

lysts. However, because electron excitation, transfer, and reduc-tion processes occur at the interface between Pd and the sup-

port, and owing to the complexity and diversity of the materi-als interface, the composition of Pd NPs produced could some-

times be uneven.

3. Characterization Techniques for BimetallicNanoparticles

The common characterization techniques for metallic NPs in-

clude X-ray diffraction (XRD), transmission electron microscopy(TEM), Auger electron spectroscopy (AES), X-ray photoelectronspectroscopy (XPS), and extended X-ray absorption fine struc-ture (EXAFS) analysis. However, owing to the nanometer rangeof particle size and the existence of macroscopic support

phases, it is difficult to analyze the structure and compositionof the supported Pd-containing nanoparticles as accurately asthe bulk materials. Also, most of the characterization tech-niques are rather qualitative or only half-quantitative.

3.1. XRD

XRD is one of the most widely used characterization tech-niques for a range of materials and provides information onlattice structure, crystallite size, particle composition, and

atomic arrangement. By using Vegard’s law, XRD can be usedto estimate the composition of alloy NPs with particle sizes

larger than 3 nm. Taking typical PdAu bimetallic nanoparticlesas an example, it is well-understood that both Pd and Au

metal phases show a face centered cubic (fcc) structure. How-ever, owing to the difference in their lattice parameters, the

diffraction peaks of the two pure phases are positioned differ-

ently. As depicted by Moitra et al. , when PdAu alloy NPs areformed as a solid solution, the diffraction peak positions and

the lattice parameters vary smoothly as a function of the com-position (Vegard’s Law).[41]

The XRD diffractograms of a series of PdAu bimetallic nano-particles are shown in Figure 3 a. The plots of the estimated

ratios of Au/Pd demonstrate a good agreement with the nomi-

nal composition in the starting mixture (Figure 3 b). Apparently,XRD represents one of the simplest ways to estimate the com-

Scheme 1. a) The exciton separation of a type II heterojunction withZnFe2O4-Fe2O3 as an example. b) The proposed electron-flow cycle in the re-duction process of two semiconductor supports catalyzed by a noble metal(Pd). Taken from Ref. [37] .

Figure 3. a) X-ray diffraction patterns of PdAu nanoparticles prepared with different Au and Pd ratios; b) Correla-tion of nominal and estimated mol % of Au in an immobilized AuPd monolith. Taken from Ref. [41] . Copyright2014, Royal Society of Chemistry.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2001

Minireviews

position of bimetallic nanoparti-cles. However, this technique

can only be applied to particleswith long-range periodic atomic

arrangement. However, whenthe particle size is smaller than

3 nm, no clear diffraction peakcan be observed.

3.2. EXAFS

As previously stated, the coordi-nation environment of metal

atom in NPs is altered with the

formation of an alloy. EXAFS isan excellent technique for pro-

viding local structural informa-tion of the alloy NPs. When the

metal atom is excited by absorb-ing high-energetic X-rays, its core electrons can escape fromthe attraction of inner electron shells. The ejected electrons are

backscattered by the neighboring atom and then interferewith the forward ejected electrons. This interference, recorded

in EXAFS spectra, depends on the type of scattering atom(s)around the absorbing atom and the distance between them.Typically, Liu and Zhang synthesized a series of AuPd alloy NPswith variable composition.[42] The EXAFS results of Au demon-

strate that two scattering paths, including Au¢Au and Au¢Pdwith different atomic distances are observed. This confirms the

formation of a AuPd alloy. Also, the derived number of neigh-

boring Pd atoms around the Au absorbing atom is proportion-al to the concentration of Pd in the AuPd alloy, suggesting

that EXAFS can be used to estimate the composition of thealloy; these results agree well with the other characterization

techniques. The EXAFS data of PdFe alloys supported on Fe2O3

at Pd K-edge have been reported by our group recently.[43] The

particles with 1–2 nm size are too small to be detected by

XRD. Meanwhile, other techniques, including XPS and energy-dispersive X-ray analysis (EDX) cannot provide accurate infor-mation with the interference from the macroscopic support. Itis clearly shown that the coordination structure of Pd derived

from EXAFS can be used to confirm the existence of the neigh-boring Fe atoms around Pd as PdFe.

EXAFS is, therefore, a convenient and powerful technique inthe characterization of extremely small alloy particles withoutmuch long-range periodic order and the results do not appear

to be disturbed by the presence of the support. However, theresult represents only the average coordination environment

for the absorbing atom, which cannot be used to analyze anyfurther detailed variation of the composition of individual par-

ticles.

3.3. AES and XPS

AES and XPS are often used for the analysis of the elemental

composition of alloys by taking their sensitivity factors into ac-count.[44] Owing to poor photoelectron escaping depth, XPS

can only provide elemental information of a sample to a depth

of 3–4 nm under the surface. This means that this techniquemay not precisely reflect the composition of the active sites if

a complex interface is presented. By comparison with XPS, AES

is more sensitive to the surface composition. Li et al. have re-cently reported the XPS and AES characterization of a AuPd

alloy growing on a Pd substrate (Figure 4 a and 4 b).[45] The XPSresult is found to be strongly influenced by the Pd support

and has a lower sensitivity compared to AES.Besides chemical composition analysis, XPS can also be used

to detect the change of electronic environment of the noble

metal in the form of alloy NPs from the corresponding bind-ing-energy values. Figure 4 c shows typical Pd XPS data for

a series of PdFe bimetallic nanoparticles reported by Wuet al.[46] Clearly, the binding-energy value of Pd attenuates with

increasing content of Fe, indicating that the charge transferfrom Fe to Pd is due to the higher electron affinity of Pd thanthat of Fe (electronegativity of 1.8 and 2.2 for Fe and Pd, re-

spectively).The above techniques, and others, for the characterization

of Pd alloy nanoparticles have their own merits and drawbacks.A combination of techniques is commonly required to provide

a more accurate representation of supported Pd-containingNPs in catalyst characterization.

4. Catalytic Applications of Pd-Based Alloys

Pd-based catalysts have a wide range of applications in cataly-

sis. With the modification of a secondary element, supported

Pd alloy NPs are capable of catalyzing important reactions, forexample, hydrogenation, dehydrogenation, hydrogenolysis,

carbohydrate reforming, and the oxygen reduction reaction,with superior performance to that of many other reported cat-

alysts. Herein, we discuss some recent reports of catalysisusing Pd alloy NPs.

Figure 4. Results of AES and XPS on the analysis of the AuPd alloy grown on a Pd crystal, a) AES, b) XPS. Takenfrom Ref. [45] . Copyright 2007, Elsevier B.V. c) The XPS curves of Pd with increasing Fe content in the PdFe alloy(the labeled selectivity refers to corresponding methanol selectivity in ethylene glycol hydrogenolysis of the PdFecatalyst). Taken from Ref. [46] . Copyright 2012, Macmillan Publishers Limited.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2002

Minireviews

4.1. Hydrogenation reactions

Catalytic hydrogenation of unsaturated bonds is an importantchemical step in the chemical industry. Pd-based catalysts are

highly active in this type of reaction owing to the superiorityof Pd in dissociating H2 molecules into atomic H, attributed tothe appropriate strength of Pd¢H bond. The volcano relation-ship of H2-activation ability with M¢H (M = metal) bond energy

is shown in Figure 5. Clearly, the Pd metal locates in the top

area of the volcano plot, which implies excellent capability in

catalyzing hydrogenation reactions.[47] For example, catalytichydrogenation of benzene to cyclohexane is an important

chemical conversion on an industrial scale. The reaction is de-scribed by the following chemical equation: C6H6 + 3 H2!C6H12

[DH298K =¢206.2 kJ mol¢1] , which is the first step of the capro-lactam production process entitled CYCLOPOL. The typicalorder of catalytic activities of transition metals known for the

hydrogenation of benzene is, Co<Pd<Ni<Pt<Ru<Rh,which relates to the fundamental interactions of substrates(benzene and dihydrogen) with the supported metallic nano-particles.[48–52] As reported by Yoon et al. , Rh/CNT (CNT =

carbon nanotube) can convert benzene to cyclohexane favora-bly with a high conversion at room temperature, whereas Pd is

almost inactive under the same conditions because of pooreradsorption of benzene (although it shows excellent activity fordihydrogen activation).[53] It is noted that the alloy of Pd with

Rh gives an activity almost twice that of the monometallic Rhcatalyst. This enhancement has been ascribed to the distinctive

electronic structure of the alloy of Pd and Rh, which can acti-vate both benzene and hydrogen to cyclohexane simultane-

ously.

Selective hydrogenation towards specific products for sub-strates with multi-unsaturated bonds is even more difficult

owing to parallel pathways and multi-hydrogenation steps. Forexample, selective catalytic hydrogenation of acetylene to eth-

ylene represents an important industrial challenge owing tothe wide use of ethylene in polymer synthesis, in which sup-

ported Pd-based catalysts are commonly used.[54–57] Unfortu-nately, supported monometallic Pd catalysts give low ethylene

selectivity because of the formation of an ethylene hydridephase on large Pd ensembles, which can be readily trans-

formed into ethane.[58–65] It is reported by Choudhary et al. thatthe selectivity of Pd for the semi-hydrogenation of acetylene

(C2H2) to ethylene (C2H4) is strongly enhanced by the incorpo-ration with Au atoms through the formation of PdAu alloyNPs, as shown in Figure 6 a.[66] The ethylene selectivity was in-

creased to 80 % by Han et al. by using a Pd@Ag core–shellnanoparticle.[67] Schlçgl and co-workers investigated the Pd-based catalysts in this reaction systematically ; they found thatPdGa exhibits the highest selectivity towards ethylene

amongst PdAu, PdAg, and PdGa (Figure 6 b).[68] The reason forthe enhancement of selective semi-hydrogenation of acetylene

is the electron-charge transfer from the modifier (Au, Ag, Ga)

to Pd, as well as the isolation of Pd islands by the surface dec-oration of the modifier atoms. The strong interaction between

Pd and the modifier atom alters the electronic structure of Pdand creates a reduction of density states at its Fermi edge, as

Figure 5. A volcano relationship of metal¢hydrogen bond strength and ac-tivity of dissociating H2. Taken from Ref. [47] . Copyright 1972, Elsevier B.V.

Figure 6. a) Time-on-stream ethylene selectivity in the hydrogenation ofacetylene for different Pd-based catalysts (the ratio of Au/Pd for (I) and (II) is2.0 and 2.2, respectively) at 70 8C. GHSV (gas hourly space veloci-ty) = 90 000 cm3 g¢1 h¢1 (1 % C2H2 ; H2 :C2H2 = 2.5). Taken from Ref. [66] . Copy-right 2003, Springer. b) Activity and selectivity of ethylene (%) in acetylenehydrogenation for model catalysts and technical catalysts (dashed line isa guide). All data collected after 20 h time-on-stream at 200 8C in 0.5 % C2H2,5 % H2, and 50 % C2H4 in He with a total flow of 30 cm3 min¢1. Taken fromRef. [68] . Copyright 2010, American Chemical Society.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2003

Minireviews

shown in Figure 7 (the density of states for Pd and PdGa areshown in Figure 7), which decreases the d-band center posi-

tion of the alloy compared to Pd. This effect will reduce the

ability of Pd to dissociate hydrogen molecules (reduce the for-mation of metal hydride) and also restrict the availability of

active atomic hydrogen, leading to the partial hydrogenationof acetylene. Meanwhile, the isolation of Pd by the Au, Ag, Ga

atoms reduces the extensive formation of Pd ensembles forthe production of ethylene hydride, hence it reduces ethane

selectivity.

The selective catalytic hydrogenation of cinnamaldehyde(CAL) to cinnamyl alcohol (COL) through selective C=O hydro-

genation is also an important reaction because COL is an im-portant intermediate and precursor in many industrial chemical

syntheses. However, Pd is highly active in C=C bond hydroge-nation, but almost inert for the hydrogenation of carbonyl

groups owing to the weak adsorption of ¢C=O¢ at Pd.[70, 71]

Qiu et al. reported that with the incorporation of a smallamount of Ru, the activity of the PdRu alloy towards COL pro-duction from the selective hydrogenation of C=O, comparedwith that of C=C, is strongly enhanced. The performance of

this alloy is superior than both monometallic Pd and Ru, partic-ularly with the assistance of a carbon nanotube support.[72]

Such an enhancement in performance was ascribed to the syn-

ergism of Pd and Ru. As also reported, there is a “volcano” rela-tionship between the activity of the catalysts and their binding

energy with ¢C=O¢ species (the ¢C=O¢ species is representedin terms of formate), as shown in Figure 8; Pd locates in the as-

cending area, whereas Ru locates in the descending area ofthe volcano curve owing to the too-weak and too-strong bind-

ing energies with ¢C=O¢ species, respectively.[73, 74] The incor-

poration of Ru into Pd can strengthen the interaction of thealloy with the ¢C=O¢ group of CAL into a proper position,

which then gives a better catalytic performance than either ofthe monometallic species.

There are some recent developments in large-scale hydro-gen production from renewable sources (solar energy, hydro-

power, biomass, and excess chemical heat).[75, 76] With the ever

increasing CO2 emissions,[77, 78] catalytic hydrogenation of CO2

to value-added chemicals has become a hot issue in cataly-

sis.[79, 80] In particular, the hydrogenation of CO2 to methanollooks very attractive owing to the position of methanol as

a high-energy-density liquid fuel and a key chemical intermedi-ate [for the manufacture of formaldehyde, methyl tert-butyl

ether (MTBE), and acetic acid] .[81] The traditional catalyst for

methanol synthesis is Cu/ZnO.[82, 83] It is reported by Haller andco-workers and Fierro and co-workers that the addition of

a small amount of Pd into the Cu/ZnO catalyst can enhancemethanol production because of the synergy effect between

Pd and Cu atoms through the formation of a PdCu alloy.[84, 85]

The enhanced activity of the alloy is thought to be caused bythe hydrogen spillover from Pd to Cu, which accelerates the

hydrogenation rate of the adsorbed carbonaceous species tomethanol on the Cu site. A Pd-based catalyst is expected toprovide higher activity in CO2 hydrogenation than the Cu-based catalyst owing to the known superiority of Pd in the ac-

tivation of hydrogen molecules. However, the monometallic Pddispersed on an inert support (Pd/TiO2, Pd/Al2O3, Pd/SiO2)

shows very low activity and selectivity towards methanol syn-thesis because the weak adsorption of CO2 on pure Pd surfacesand the existence of the reverse water gas shift (RWGS) reac-

tion produce CO as the byproduct. The proposed pathways ofCO2 hydrogenation are shown in Scheme 2;[86] methanol syn-

thesis and CO production are parallel to each other. Bonivardiet al. reported that the Pd/SiO2 catalyst with incorporation of

Ga2O3 showed a 500 fold Pd turnover frequency (TOF) of mon-

ometallic Pd/SiO2 and the enhancement was ascribed to theformation of a Ga¢Pd interaction. The alloy surface is thought

to strengthen the stability of methanol synthesis intermediates,including the formate and methoxy species.[87] Iwasa et al. in-

vestigated a series of Pd alloys including PdZn, PdIn, andPdGa, which are formed in situ from the reduction of Pd/metal

Figure 7. Calculated electronic density of states (DOS) with measured XPSvalence-band spectra (dashed lines) for elemental Pd (top) and PdGa(bottom). Taken from Ref. [68, 69] . Copyright 2010, American Chemical Soci-ety. Pd 4d states of PdGa are shown in red. Note the pronounced changesnear the Fermi edge. The insets show the neighboring atoms. Figure 8. A plot of decomposition temperature at a normalized rate [an al-

ternative axis on natural logarithm of turnover frequency, Ln(TOF)] as a func-tion of heat of formation of transition metal formate. Taken from Ref. [74] .Copyright 2013, Royal Society of Chemistry.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2004

Minireviews

oxides (ZnO, In2O3, and Ga2O3) and exhibit methanol produc-

tion activities in the order of PdZn>PdIn>PdGa at 1 atmos-

phere (Table 1).[88] By combining the experimental results andtheoretical calculations, it was reported that there is an elec-

tron-charge transfer from the late transition metal (Ga, In, Zn)to Pd, which modifies the density of states of the valence

band allowing the alloys to resemble Cu. This feature stabilizesthe methanol synthesis intermediates, such as formate and

methoxy species.[86, 89, 90] It is noted that the heat formation of

formate on Cu is much higher than that on Pd, indicating itshigher ability to stabilize formate species in the CO2 hydroge-

nation reaction (Figure 8). Owing to the higher ability for theactivation of hydrogen molecules, Pd-based alloys show 20–

100 times higher activity (in terms of turnover frequency) thanCu, which makes them more promising for CO2 hydrogenation.

The detailed mechanism for the incorporation of a modifieratom to Pd to alter the properties of Pd is still unclear. From

electronic configurations, all of the late transition metals, in-cluding Zn, In, and Ga, with s-valence electrons can transfer

their charge (electrons) to Pd because of the empty Pd 5s orbi-tal. This transfer will increase the local electron density of thePd-based alloy surface, enhance the interaction(s) with inter-

mediates in high oxidation states, and catalyze further reduc-tion of CO2.

4.2. Dehydrogenation reactions

As the reverse process of hydrogenation, most dehydrogena-tion reactions also involve the interaction of H atoms with thecatalyst surface. Thus, dehydrogenation reactions can be cata-lyzed by Pd-based catalysts and the situation is quite similar to

that of hydrogenation. There is one type of dehydrogenationreaction widely demanded in the field of fuel cells, that is, the

in situ production of H2 in polymer electrolyte membrane(PEM) fuel cells to supply electronic devices. For example, thedecomposition of formic acid (HCOOH) to CO2 and H2 is one ofthe typical representatives. Formic acid is regarded as a conven-ient H2-storage chemical for the supply of hydrogen to fuel

cells owing to its high energy content and low decompositiontemperature.[91–93] It has been widely reported that formic acid

can be dehydrogenated under 100 8C catalyzed by Pd.[93–95]

However, accompanying the dehydrogenation process,HCOOH can also be dehydrated to produce CO and H2O.[96] For

the subsequent conversion of hydrogen into electrical energy,the latter pathway that produces CO impurity, which is toxic to

fuel-cell catalysts, should be avoided.[93] Gu et al. reported thata supported monometallic Pd catalyst initially showed high ac-

tivity in the decomposition of formic acid, but deactivates

quickly owing to CO poisoning.[97] After screening a large varie-ty of mono- or bimetallic catalysts, the supported AuPd alloy

was found to exhibit the highest activity with an excellent sta-bility for the dehydrogenation of HCOOH to H2. The enhanced

activity should be attributed to the high CO tolerance of thealloy by the incorporation of Au. With the same principle,

Zhang et al. successfully prepared a PdAg alloy with controlla-

ble composition and they showed that Ag42Pd58 NPs is thebest catalyst for the reaction. An initial TOF of 382 h¢1 and an

apparent activation energy of 22 kJ mol¢1 were obtained.[98]

Kim and Barteau and Hoshi et al. proposed that CO2 and CO

are generated from the decomposition of formate species indifferent distinguished adsorption states, as shown in Fig-ure 9 a and 9 b: the adsorbed bidentate leads to the produc-

tion of CO2 and the adsorbed monodentate is the precursorfor CO.[99, 100] The enhancement of electronic back donation

from the Pd-based surface to the formate species couldstrengthen the adsorption of bidentate and favors the dehy-

drogenation pathway. To tune the interaction of Pd with theadsorbed formate species, a series of core–shell Pd(shell)-

M(core) alloys were prepared in our group. A negative linearcorrelation between catalytic activity and the work function ofthe metal core is observed (Figure 9 c).[101] This work clearly in-

dicates that the enhancement of catalytic performance iscaused by the electron transfer from the core metal to the Pd

shell, which increases the adsorption strength of the bidentateformate species. As a result, the Pd(shell)Ag(core) alloy is

found to display the highest selectivity of CO2 owing to the

lowest work function of Ag.

4.3. Hydrogenolysis reactions

It is generally accepted that the present nonrenewable naturalresources (oil, gas, and coal) cannot continue to dominate the

Scheme 2. The proposed pathways for CO2 hydrogenation to methanol andCO.

Table 1. The catalytic performances of Pd-based catalysts in CO2 hydro-genation at 1 atmosphere, 463 K, CO2/H2 = 1:9.[88]

Catalyst Rate of formation Methanol Methanol[mmol min¢1 g¢1] selectivity [%] TOF Õ 103 [s¢1]

CH3OH CO CH4

Pd/ZnO 1.28 0.682 0.01 65.1 1.1Pd/Ga2O3 0.169 0.607 0.01 23.3 0.12Pd/In2O3 0.246 15.4 0.01 1.72 0.21Pd/SiO2 0 0.522 0.09 0 0Pd/MgO 0 2.05 0.07 0 0Pd/ZrO2 0.222 13.2 0.54 1.65 0.06Pd/CeO2 0.11 9.90 0.16 1.08 0.05Pd/black 0 0.431 0.02 0 0Cu/ZnO 1.74 3.96 0 30.4 0.075

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2005

Minireviews

energy and chemical markets for long into the future owing totheir finite reserves and rapid consumption.[102–105] As one ofthe primary renewable energy resources in nature, biomass, es-

pecially carbohydrates with repeating ¢CHOH¢CHOH¢ units,becomes a favorable alternative.[106] Hydrogenolysis is widely

used as a first step in the utilization of biomass molecules, usu-ally involving the cleavage of C¢C or C¢O bonds by produced

atomic hydrogen on the catalyst surface. A monometallic Pdcatalyst was found to be highly active in C¢O hydro-

genolysis. Mitra et al. reported that in the hydroge-nolysis of 5-hydroxymethylfurfural (HMF; a glucosederivative) catalyzed by Pd/C, 2,5-dimethylfuran

(DMF), produced from the hydrogenolysis of the C¢Obond, and further hydrogenation products are ob-

served.[107] The reaction pathways are displayed inScheme 3. With the addition of appropriate additives,

the DMF selectivity can be further improved to over

80 % (Table 2), which reflects the high activity ofmonometallic Pd in C¢O hydrogenolysis. Pd catalysts

are also widely applied in the production of alkanesin the hydrogenolysis of CO moieties of biomass mol-

ecules. As Huber et al. have reported, the alkane se-lectivity in sorbitol hydrogenolysis reaches 98 %, im-

plying a near complete elimina-tion of the C¢O bonds.[108] How-

ever, owing to the high reactiontemperature employed, some of

the C¢C bonds in the ¢CHOH¢CHOH¢ units are also cleaved,

increasing the selectivity towardslighter alkanes at the expense of

higher alkanes. Our group first

reported methanol synthesisfrom the selective C¢C hydroge-

nolysis of ethylene glycol overPd/Fe2O3 (the simplest biomass-

derived molecule), whichopened up a new renewable

non-syngas route for methanol

production.[46] The in situ-formedPdFe alloy nanoparticles sup-

ported on Fe2O3 are believed toprovide active sites for the hy-

drogenolysis reaction. The coor-dination of Fe atoms around Pd atom shown by EXAFS

(Table 3) confirms the formation of PdFe. It was also found that

Figure 9. Summary of the interactions of a) bridging formate (bidentate) on a flat terrace of metal M sites andb) linear formate (monodentate) on isolated or low-coordinated M sites. Taken from Refs. [99, 100]. Copyright1990, American Chemical Society and 2007, Elsevier B.V. , respectively). c) Correlation between the M-core workfunction and the rate of dehydrogenation of HCOOH with M@Pd as the catalysts. Taken from Ref. [101] . Copyright2011, Macmillan Publishers Limited.

Scheme 3. Hydrogenolysis of 5-hydroxymethylfurfural under H2 (30 psi) overPd/C catalyst. Taken from Ref. [107]. Copyright 2015, Royal Society ofChemistry.

Table 2. The effect of an additive on the hydrogenolysis of 5-hydroxy-methylfurfural over Pd/C catalyst.[a][107]

Entry Additive Yield [%] Conv.[c]

DMF BHMTHF DMDHF DMTHF [%]

1 none 0 24 11 64 >952 HCO2H (5 mmol) 85 0 8 4 >953 CH3CO2H (5 mmol) 42 0 10 42 >954 CO2 (30 psi)/H2O[b] 37 0 50 12 >955 CO2 (30 psi) 0 57 0 42 >956 [CH3OC(O)]2O

(5 mmol)/H2O[b]

52 0 40 7 >95

7 [CH3OC(O)]2O(5 mmol)

0 trace 0 >95 >95

[a] Reaction conditions: HMF (0.5 mmol), substrate/Pd = 10 molar ratio, H2

(30 psi), dioxane (5 mL), 120 8C, 15 h. [b] 0.5 mL H2O and 4.5 mL dioxane.[c] Conversion. BHMTHF, DMDHF, and DMTHF are abbrievations for 2,5-bis(hydroxymethyl)tetrahydrofuran, 2,5-dimethyldihydrofuran, 2,5-dime-thyltetrahydrofuran.

Table 3. The EXAFS results of the Pd/Fe2O3 catalyst for the hydrogenolysis of ethyleneglycol.[46]

Sample Scattering CN[a] Bond length R-factor DE DW factor[b]

pair [æ] [%] [eV] [s2 Õ æ2]

Pd/C Pd¢Pd[d] 8.0 (3) 2.73 (1) 2.6 ¢7.6 0.006

Pd/Fe2O3

(pre-reduced)

Pd¢Fe[c] 1.0 (2) 2.59 (1)1.6 ¢7.4

0.003Pd¢Pd[c] 1.9 (2) 2.66 (1) 0.004Pd¢Pd[d] 2.4 (2) 2.75 (1) 0.003

[a] Coordination number, values in parenthesis indicate the number of neighbor atomaround each Pd absorbing atom. [b] Debye–Waller factor; DE, the absorbing energydifference between the experiment value and the calculated value in the fitting.[c,d] indicate the scattering paths derived from PdFe(1:1) face center cubic structure(fcc) and Pd fcc structure, respectively

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2006

Minireviews

the methanol selectivity rises with increasing reaction time,this finding is attributed to the increase in Fe coordination to

Pd from the progressive reduction of the support (Figure 10).Further investigations by our group also found that the Fe-rich

supported PdFe nanoparticles are more selective in the hydro-

genolysis of C¢C bonds compared with C¢O bonds for a seriesof biomass-related molecules that produce small alcohol mole-

cules (methanol, ethanol, and propanol) as the degradationproducts.[43] According to density function theory (DFT) calcula-

tions on the interaction of ethylene glycol (the C¢O functionalgroup) with the Pd surface (Figure 11), the specificity of PdFe

in the hydrogenolysis of C¢C bonds compared with C¢O

bonds is caused by the significant decrease in both the d-bandfilling and d-band center position of the alloy at increasing Fe

content. This could then strengthen the s-donation from thebiomass substrate to Pd and simultaneously weaken the C¢C

bond. Thus, the new alloy catalyst can provide an exciting wayto break down biomass compounds to valuable small alcohol

molecules with a narrow product distribution. In principle, the

catalytic ability of Pd in the biomass hydrogenolysis can bewell-tuned from selective C¢O cleavage to selective C¢C cleav-

age, depending on the d-band filling and d-band center posi-

tion of the alloy nanoparticle of variable composition. The se-lectivity of C¢O bond breakage in the biomass molecules

should increase with the up-shift of the d-band center positionand the increasing d-band filling, both of which strengthen

the electronic p-back donation from Pd to the substrate mole-

cule and eventually facilitates the breaking of the C¢O bond.

4.4. Aqueous-phase reforming (APR) reactions

Besides hydrogenolysis, the aqueous-phase reforming (APR) re-action has recently attracted a wide interest in biomass utiliza-

tion to produce hydrogen gas. Hydrogen gas is considered to

be the cleanest energy carrier. However, its large scale produc-tion is presently derived from steam reforming of natural gas,

which is obviously unsustainable. Thus, hydrogen productionfrom the aqueous-phase reforming (APR) reaction of biomass

molecules is receiving increasing attention and is regarded asan alternative route to obtain renewable hydrogen.[109–111] Thedesired reaction pathways involve cleavage of C¢C bonds as

well as C¢H and O¢H bonds to form adsorbed CO species onthe catalyst surface, which is then removed by the water–gasshift reaction (WGS) to form CO2 and H2 (Scheme 4).[112] Howev-er, the hydrogenation of the adsorbed CO species by thein situ-formed H2 may also produce alkanes as byproducts,which reduces the overall H2 yield. Pt metal is found to be the

most active monometallic catalyst for the APR followed by

Ni>Ru>Rh�Pd> Ir, but Pd is known to be the most selectivemetal in H2 production.[111] Thus, the modification of Pd to en-

hance the activity is being intensively investigated owing tothe relatively lower cost of Pd compared with Pt and the

Figure 10. Catalytic performances for hydrogenolysis of ethylene glycol. a) 5 % Pd with different oxides (prepared by co-precipitation) ; b) time–fraction selec-tivity over 5 % Pd/Fe2O3 at different periods. Taken from Ref. [46] . Copyright 2012, Macmillan Publishers Limited.

Figure 11. Electronic interaction of Pd and the C¢O functional group of eth-ylene glycol. Taken from Ref. [43] . Copyright 2015, Royal Society of Chemis-try.

Scheme 4. The steps of the APR reaction, converting carbohydrate toCO2 + H2.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2007

Minireviews

demand for higher H2 selectivity over alkanes. Huber et al. re-ported that the metallic iron atom has an extremely strong

promotional effect on the activity of Pd in the APR of ethyleneglycol ; the turnover frequency (TOF) of Pd for H2 production

with the incorporation of Fe is 500 times and 150 times higherthan those of monometallic Pd and Pt supported on Al2O3, re-spectively.[112] After screening over 130 platinum-group cata-lysts, they reported that PdFe is the most active catalyst withobtained H2 selectivity of over 95 %. This enhancement in cata-lytic performance is assigned to the superior activity of PdFefor the water–gas shift of CO to CO2/H2, which is considered asthe rate-determining step of the APR reaction. However,a huge amount of Fe atoms is needed to modify Pd, and this

also blocks the surface Pd active sites and reduces the overallH2 yield. Tanksale et al. reported that the incorporation of

a small fraction of Pd into Ni metal also shows a strong syner-

gy effect between the two, giving higher activity and H2 selec-tivity than the monometallic Ni, Pd, and even Pt, in the sorbitol

APR reaction.[113] Their characterization results indicate that theaddition of Pd reduces the CO adsorption strength on the

alloy surface, hence enhancing the production of H2 and CO2

through the WGS reaction. In summary, the key steps of the

APR reaction for biomass molecules appears to lie in the WGS

reaction of pre-adsorbed CO. Owing to its relatively higherability to catalyze the WGS reaction compared with other plati-

num-group elements, Pd exhibits an excellent H2 selectivity inthe APR reaction. The addition of promoter(s) to reduce CO ad-

sorption can further facilitate the activity and selectivity of thePd-containing alloys.

4.5. Steam-reforming reactions

As previously stated, hydrogen is a clean and efficient energycarrier. However, as the lightest gas, its high-pressure storage

and transportation are considered to be inefficient, especially

when it is used to supply fuel cells for providing power to elec-tronic appliances. An instant hydrogen production from steam-

reforming reactions of liquidifed hydrocarbons could be a feasi-ble solution.[114] The difference between steam reforming and

aqueous-phase reforming (APR) reactions lies in the reactionphase: the APR reaction is catalyzed at the aqueous phase

inside a batch reactor for a period of time and H2 gas can beproduced on a large-scale at high pressure. For the steam-re-

forming reaction, the hydrocarbon molecules react with H2O ina flowing gas stream over a solid catalyst and the produced H2

is transferred for consumption immediately. Owing to the

short residence times of the reactants on the catalyst surface,both the reactivity of a specific hydrocarbon and the activity of

the catalyst are required to be high for instant H2 production.The most widely used hydrocarbons are small alcohol mole-

cules, particularly methanol, owing to its high ratio of H/C

(4:1), its industrial availability, relatively low reforming tempera-ture, and low sulfur content.[34, 115–117] Previous studies of metha-

nol steam reforming (MSR) to CO2 and H2 were focused on theuse of Cu-based catalysts.[118–120] However, the Cu-based cata-

lysts show a poor thermal stability and are pyrophoric, there-fore, are not suitable for general fuel-cell applications.[35] Re-

cently, investigations have shifted towards Pd-based catalysts.Iwasa et al. reported that monometallic Pd shows very low ac-tivity and selectivity for the MSR reaction owing to the decom-position of methanol to CO, which not only reduces the H2

yield, but also causes deactivation of the catalyst (through COpoisoning).[121] However, Pd/ZnO was found to be an excellentcatalyst for the MSR reaction owing to the modification of Pdby reactive Zn atoms from the reduction of the ZnO sup-port.[35, 122] PdZn is then regarded as the active phase. In addi-

tion, it was reported that CO2 selectivity of the MSR reaction isrelated to the ratio of Pd/Zn in the alloy. The Zn-rich PdZnalloy catalysts show higher activity and higher CO2 selectivitythan that of Pd-rich catalysts.[123–125] Zhang et al. recently re-

ported the Zn-rich PdZn b-phase (Pd/Zn = 1:1) was prone toform on polar (0 0 0 1) facets of ZnO rather than on nonpolar

(1 0 1 0) facets.[34] Thus, the catalytic ability can be further en-

hanced by the morphology control of the support. Besides,PdZn, PdGa, and PdIn formed during in situ reduction of Pd/

metal oxide (ZnO, Ga2O3, In2O3) are reported to show enhancedperformances compared with monometallic Pd, as illustrated

in Table 4.[126] The possible reaction pathways of the MSR reac-tion are shown in Scheme 5. Iwasa and Takezawa proposed

that the role of the additive atoms is to stabilize the adsorbed

formaldehyde and to facilitate the attack of H2O to produceformic acid, which further decomposes to CO2 and H2.[126]

Apart from the MSR reaction, hydrogen production from thesteam reforming of ethanol is also being intensively investigat-

Scheme 5. The parallel pathways of CO and CO2 production in the MSR reac-tion.

Table 4. The catalytic performance of a series of Pd-based catalysts inthe MSR.[126]

Catalyst Conversion Selectivity [%] Dispersion H2 TOF[%] CO2 CO DME

Pd/ZnO 54.2 99.2 0.8 0 2.2 0.497Pd/In2O3 28.3 95.5 4.5 0 2.1 0.291Pd/Ga2O3 21.2 94.6 5.4 0 2.6 0.177Pd/SiO2 15.7 0 100 0 9.0 0.021Pd/Al2O3 58.9 0 69.6 30.4 14.0 0.036Pd/MgO 41.0 6.6 93.4 0 10.4 0.053Pd/ZrO2 64.3 18.4 81.6 0 6.6 0.139Pd/CeO2 62.4 22.7 77.3 0 36.2 0.025Pd/Ta2O3 6.0 0 100 0 N.A. N.A.Pd/HfO2 13.6 0 100 0 N.A N.APd/C black 10.9 0 100 0 2.1 0.0071Cu/ZnO 19.3 100 0 0 1.0 0.300

[a] Reaction conditions: 493 K, inlet partial pressure of methanol =10.1 kPa, Pd loading = 10 wt %. N.A = not applicable.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2008

Minireviews

ed. Goula et al. reported that Pd/g-Al2O3 exhibits good activityand high selectivity towards CO2 and H2 from ethanol reform-

ing, but this catalyzed reaction requires a relatively high tem-perature of �500 8C to give hydrogen-rich gas.[127]

To summarize, Pd-based catalysts are clearly active and se-lective in steam-reforming reactions of small alcohol molecules

and their catalytic performance is strongly enhanced by the in-corporation of late transition metals.

4.6. Oxygen reduction reaction

The oxygen reduction reaction (ORR) shown in the two chemi-cal equations below consitutes the cathodic half reaction ofa fuel cell.

2 Mþ O2 ! 2 MO ð1Þ

2 MOþ 4 Hþ þ 4 e! 2 Mþ 2 H2O ð2Þ

Thus, understanding the catalysis actions of the ORR is sig-

nificantly important in the development of fuel-cell technology.The activity of the electrocatalyst for ORR on cathodes is

mainly related to the interaction of the O atom with the metalsurface, as shown in Figure 12.[128] Until now, Pt is still regarded

as the best catalyst for the ORR reaction (at the top area of the

“volcano” curve) because of its appropriate binding energywith the O atom [weak M¢O bond cannot activate O2 mole-

cule efficiently whereas strong M¢O bond leads to negative re-duction potential of Equation (2)] . However, because of the

limited availability of Pt, recent efforts in electrocatalysis havebeen focused on decreasing the Pt content in fuel-cell electro-

catalysts or replacing it with less expensive materials. As

shown in Figure 12, Pd is the second most active metal, and isgenerally considered as a promising alternative of Pt. Mazumd-

er et al. reported that the core–shell structure of the Pd(core)-Pt(shell) alloy with the decoration of a small amount of Fe

atoms not only shows excellent catalytic performance in ORR,but also exhibits a higher durability in acidic solution than the

monometallic Pt catalyst.[129]

Even in the absence of Pt, Pd alloys are also active in catalyz-ing ORR reaction. It is widely reported that the PdCo alloy

shows a comparable activity to Pt in the ORR owing to themodification of Co on the electronic structure of Pd.[130–136] Xu

et al. reported that a nanoporous PdCo alloy with uniformstructure size of �5 nm exhibits much higher specific and

mass activities relative to nanoporous Pd and commercial Pt/Ccatalysts as shown in Figure 13.[137] A long-term stability test

demonstrated that nanoporous PdCo has comparable catalyticdurability to that of Pt/C, but has a smaller loss of ORR activityand electrochemical surface area.

4.7. Hydrodesulfurization (HDS) reactions

Hydrodesulfurization (HDS) is probably the most important hy-

drotreatment process in the petroleum refining industry toproduce clean fuels.[138] The recent trend of environmental reg-

ulation is becoming increasingly stringent with regard to sulfurcontent in fuels.[139–144] The conventional Ni¢Mo sulfide catalyst

requires a high temperature for activation and it also deacti-vates rapidly owing to problems of particle sintering and

carbon deposition.[145, 146] Thus, it is difficult for the convention-

al catalysts to manage a deep HDS. With the expandingdemand for more efficient catalysts, modification or replace-

ment of the Ni¢Mo catalysts is receiving considerable atten-tion. Owing to the higher hydrogenation ability, Pd-based cata-

lysts can work under a much lower temperature regime com-pared with the conventional Ni¢Mo sulfide catalysts. Niquille-

Figure 12. Trends in oxygen reduction activity plotted as a function ofoxygen binding energy. Taken from Ref. [128] . Copyright 2004, AmericanChemical Society.

Figure 13. a) Specific kinetic current densities for all catalysts at different po-tentials ; b) specific kinetic and mass kinetic current densities at room tem-perature for all catalysts at 0.9 V.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2009

Minireviews

Rçthlisberger and Prins reported that PdPt bimetallic catalystsshows a much higher activity in HDS of 4,6-dimethyldibenzo-

thiophene and dibenzothiophene than the correspondingmonometallic Pd or Pt catalysts owing to the chemical syner-

gism of Pd and Pt.[147] Venezia et al. reported that the incorpo-ration of Au with Pd can further promote the HDS rates of

thiophene and dibenzothiophene, which are considered asmodels for sulfur-containing compounds in petroleum.[145] Theenhancement was attributed to the size reduction of Pd en-

sembles by the introduction of Au atoms and the charge trans-fer from Au to Pd. Both factors strengthen the interaction ofthe substrate with Pd. Moreover, AuPd alloy particles alsoshow better H2S tolerance than Pd monometallic nanoparticles

and hence exhibit excellent stability.

5. Modification Mechanism of SecondaryAtoms on Pd-Based Catalysis

By reviewing Pd-based bimetallic catalysis, it is clearly demon-strated that the catalytic ability of Pd for a specific reaction

can be strongly promoted by the appropriate selection of al-

loying element. To rationally design novel catalysts with excel-lent catalytic properties, it is essential to understand how the

insertion of a secondary atom into a Pd lattice influences thecatalytic performance. At this stage, the factors affecting the

catalytic properties of metals are briefly classified into threecategories: i) the geometric factor, ii) electronic factor, and iii)

synergy effect.

5.1. Geometric factor

Geometric factors apply to the elements that are related to the

geometry of substance, the atomic packing, lattice parameter,bond length, and site occupancy for instance, ignoring the cor-

responding electronic influences. With the progressive substi-

tution or decoration of Pd by a secondary element, a series ofgeometric variations may occur leading to the alteration of the

local chemical environment of Pd: i) the bond length of Pd¢Pdmay vary owing to a compression or expansion, or even phase

transition of the parent lattice; ii) the size of the Pd ensembledecreases with the insertion of the modifier atom; iii) the num-

bers of corner, edge, low-coordinated sites may be reducedbecause these active sites are favored by the decoration of

a secondary element.[17] These variations play important rolesin the adsorption of reactants or intermediates, particularly forbi- or multiadsorbate(s). Shao et al. reported that PdFe shows

an excellent catalytic property in the oxygen reduction reac-tion in fuel cells, superior to that of monometallic Pd and even

the active Pt catalyst. This enhancement is caused by the influ-ence of the Fe atom on Pd¢Pd distance, as shown in Fig-

ure 14 (a).[148] The activity increases with the decrease in Pd¢Pd

bond length and reaches a peak at a specific ratio of Pd/Fewith the smallest Pd¢Pd distance. Jones et al. reported that Bi

atoms dwell specifically on low-coordinated sites (corners/edges) in the Pd crystal model.[149] In the decomposition of

formic acid, this site blockage suppresses the dehydration reac-tion (HCOOH!CO + H2O) by decreasing the adsorption site of

monodentate formate (the precursor of CO as shown inFigure 9) on the corners/edges. This reduces the deactivationof Pd, resulting in a higher rate of formic acid dehydrogena-tion. Further surface coverage of Bi beyond the corner/edge

sites rapidly attenuates the catalytic activity, giving a typical“volcano” response as shown in Figure 14 (b).

5.2. Electronic factor

One of the most fundamental properties of a metal surface is

its ability to form bonds with the surrounding atoms and mol-

ecules. This bonding ability determines the catalytic propertyof the metal surface because the catalyst surface forms chemi-

cal bonds with the reactants and it helps to break intramolecu-lar bonds and form new ones during catalysis.[150] Meanwhile,

the bonding ability is governed by the electronic structure ofthe metal because the bonding process involves a redistribu-

tion of electron density between the metal and the surround-

ing atoms or molecules through the overlapping of electronorbitals. All of the elements that influence the electronic struc-

ture are so-called “electronic factors”, which play key roles incatalysis processes. For transition metals, the electronic struc-

ture is mainly described by its d-band structure, in terms of itsd-band center position and d-band filling.

Figure 14. a) Kinetic current density (electrochemical area) and Pd¢Pd bondlength calculated from XRD data against the concentration of Fe in thePdFe/C electrocatalysts treated at 500 8C. Taken from Ref. [149] . Copyright2006, American Chemical Society. b) Total gas production (2 h) of Bi decorat-ed Pd-PVP with increasing Bi content in the decomposition of formic acid.Taken from Ref. [150] . Copyright 2015, Royal Society of Chemistry.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2010

Minireviews

It is widely reported that the variation in the d-band struc-ture of Pd by the modification of a secondary element eventu-

ally influences the catalytic performance. To avoid the influ-ence of surface active site blockage by the modifier atom and

to simply investigate the electronic effects, our group demon-strated the synthesis of a non-surface decorated PdB catalyst

by introducing boron atoms into the interstitial octahedral siteof the Pd lattice.[151] This novel catalyst was found to give ultra-selectivity in a series of selected partial hydrogenation reac-

tions, outperforming the Lindlar catalysts in the liquid phaseowing to the strong electronic perturbation by the interstitialB atoms to the parent Pd atoms.[151] The DFT calculation showsa broadening and downshifting d-band of Pd upon the incor-

poration of B atoms, which dramatically attenuate the adsorp-tivity of the metal surface to achieve ultra-selectivity in partial

hydrogenation. Also, the occupancy of the interstitial boron

atoms can displace H atoms (Pd b-hydride) that dwell at thesame sites to reduce undesirable over-hydrogenation by the

subsurface hydrogen.[152]

5.3. Co-catalyzing mechanism

It is well-understood that a typical catalytic reaction usually in-

volves several elementary steps. The adsorption of each inter-mediate or reactant of the reaction on the catalyst surface can

influence the catalytic performance.[84, 85, 153, 154] In the catalysisprocess of monometallic catalysts, all of the intermediates or

reactants are adsorbed on the same metallic site with competi-

tion. With the addition of a secondary element into Pd lattice,the modification atoms may provide a stronger adsorption site

for some of the intermediates or reactants and enhance theirinteraction with the catalyst. This may eventually strengthen

the catalytic ability of bimetallic NPs. Li and co-workers report-ed the synthesis of a series of PdAg alloys with controllable

compositions for the hydrodechlorination of 4-chlorophe-

nol.[153] The catalytic results show that the addition of Ag signif-icantly promotes the activity of PdAg NPs and the activity is

superior to that of monometallic NPs, despite the fact thatmonometallic Ag is inert for this reaction. They found that the

4-chlorophenol molecules adsorbed onto the Ag site and H2

dissociated on Pd site. Thus, this co-catalyzing mechanism ofAg and Pd atoms in a bimetallic phase strengthens the catalyt-ic ability, as shown in Figure 15. The Li group also reported the

catalysis of PdNi NPs in the Miyaura–Suzuki reaction. Theadded Ni site, with a lower electronegativity value (1.9) thanthe Pd site (2.2), accelerates the reaction rate of the first oxida-tive addition of aryl halide to the Pd¢Ni nanoparticle surfacebecause Ni is a good nucleophile.[154] As a result, PdNi shows

better activities than monometallic Pd NPs because of the syn-ergism with Ni atoms.

Although the electronic factor, geometric factor, and co-cat-

alysis mechanism may be separately assessed through someextreme methods, they are commonly entangled with each

other in most of the practical catalysts. For example, the inser-tion of a foreign atom causes a geometric change in the Pd

parent lattice; this geometric change must influence the over-lap of metal d-orbitals, which also affects the d-band structure

of the formed bimetallic phase. The variation in electronic

structure, particularly the d-band filling, may also result in thechange of metal¢metal bond length, or even the transition of

phase structure, owing to the altered electron repulsion be-tween two neighboring atoms.[155] Additionally, the co-catalyz-

ing phenomenon occurs simultaneously with electronic and

geometric factors if the secondary atoms show significant in-teraction with the intermediates or reactants of the catalysis

reaction. The convolution of the three factors makes the modi-fication mechanism of foreign atoms on Pd catalytic properties

complicated. This complication could sometimes limit the sys-tematic investigation of geometry and electronic structures of

a catalyst, and ultimately restrict the rational design of a bimet-

allic catalyst.

Conclusions

Supported Pd-based NPs act as excellent catalysts in a large

variety of industrially important reactions, which include hydro-genation, dehydrogenation, hydrogenolysis, reforming, and hy-

drodesulphurization, to name a few. By incorporating a secon-dary element as a modifier, the catalytic properties of Pd can

be tuned. The approach could also enhance both the activityand selectivity towards a specific product. This minireview has

presented the preparation, characterization, and catalytic test-ing of some recent Pd-based alloy catalysts in a series of im-

portant chemical reactions. It is believed that the resulting cat-alytic properties of Pd are the convolution of all of the modifi-cation effects including geometric factor, electronic factor, and

co-catalyzing. In general, the alloying effects of the transitionmetals with Pd can be briefly described as followed: i) early

transition metal(s) result in a decrease in the d-band filling,whereas the incorporation of late transition metals increases

the electron filling of the resulting d-band; ii) 3rd row ele-ment(s) usually cause a lattice compression for its alloy, result-ing in a better overlap of 4th row Pd d-orbitals and hence de-

creasing the d-band center position, whereas 5th row ele-ment(s) of larger atom size than Pd would cause the opposite

effect; iii) elements of relatively low work function (Ag, Ga, Zn,etc) would raise the Fermi level of Pd through a simple charge

Figure 15. a) TOFs in the first hour for the HDC reaction of 4-chlorophenol.Conditions: 4-chlorophenol (1 mmol, 128.6 mg), base (1.5 mmol, 60.0 mg),catalyst loading (0.6 mol % based on initial 4-chlorophenol), THF (3 mL), n-tri-decane (0.3 mmol), H2 balloon; b) proposed reaction mechanism of hydro-dechlorination of 4-chlorophenol catalyzed by AgPdx. Taken from Ref. [154].Copyright 2013, American Chemical Society.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2011

Minireviews

transfer, whereas the other PGM show little or even oppositeeffects. Thereby, a rational selection of the modifier atom for

Pd may result in the formation of alloys with predictable andtailorable electronic properties. The catalytic performance is

mainly dictated by the adsorption energy of the catalyst sur-face with the substrate. Alloys with a lower d-band filling and

d-band center position favor s-donation from the substratemolecule to the Pd-based alloy surface, whereas a higher d-

band filling and d-band center strengthen the p-back donation

from the alloy surface to the substrate with the p system. Al-though establishing the catalytic mechanism for a specific re-

action is always more complex than using the simple adsorp-tion enthalpy of a substrate, and there are many factors to

consider, these simple rules may still be useful in providinggeneral guidelines for the selection of appropriate modifieratoms in Pd-based alloys for targeted chemical reactions.

Acknowledgements

The authors wish to thank the EPSRC and University of Oxford

for funding some of the research mentioned in this review.

Keywords: alloys · geometric effects · ligand effects ·nanoparticles · palladium

[1] L. S. Benner, T. Suzuki, K. Meguro, Precious Metals Science and Technolo-gy (Ed. : S. Tanaka), IPMI Press, Allentown, 1991.

[2] P. B. Kettler, Org. Process Res. Dev. 2003, 7, 342 – 354.[3] G. W. Parshall ; S. D. Ittel, Homogeneous Catalysis, 2nd ed. , Wiley, New

York, 1992.[4] Catalytic Asymmetric Synthesis, 2nd ed. (Ed. : I. Ojima), Wiley-VCH, New

York, 2000.[5] Applied Homogeneous Catalysis with Organometallic Compounds (Eds. :

B. Cornils, W. A. Herrmann), Wiley-VCH, Weinheim, 2002.[6] N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457 – 2483.[7] L. Yin, J. Liebscher, Chem. Rev. 2007, 107, 133 – 173.[8] A. F. Littke, G. C. Fu, Angew. Chem. Int. Ed. 2002, 41, 4176 – 4211;

Angew. Chem. 2002, 114, 4350 – 4386.[9] R. Martin, S. L. Buchwald, Acc. Chem. Res. 2008, 41, 1461 – 1473.

[10] A. K. Diallo, C. Ornelas, L. Salmon, J. R. Aranzaes, D. Astruc, Angew.Chem. Int. Ed. 2007, 46, 8644 – 8648; Angew. Chem. 2007, 119, 8798 –8802.

[11] G. Altenhoff, R. Goddard, C. W. Lehmann, F. Glorius, Angew. Chem. Int.Ed. 2003, 42, 3690 – 3693; Angew. Chem. 2003, 115, 3818 – 3821.

[12] W. A. Herrmann, K. Fele, S. K. Schneider, E. Herdtweck, S. D. Hoffmann,Angew. Chem. Int. Ed. 2006, 45, 3859 – 3862; Angew. Chem. 2006, 118,3943 – 3947.

[13] T. Fujihara, S. Yoshida, H. Ohta, Y. Tsuji, Angew. Chem. Int. Ed. 2008, 47,8310 – 8314; Angew. Chem. 2008, 120, 8434 – 8438.

[14] D. J. M. Snelders, G. van Koten, R. J. M. Klein Gebbink, J. Am. Chem. Soc.2009, 131, 11407 – 11416.

[15] E. Bustelo, C. Gurot, A. Hercouet, B. Carboni, L. Toupet, P. H. Dixneuf, J.Am. Chem. Soc. 2005, 127, 11582 – 11583.

[16] W. Yu, M. D. Porosoff, J. G. Chen, Chem. Rev. 2012, 112, 5780 – 5817.[17] J. R. Kitchin, J. K. Norskov, M. A. Barteau, J. G. Chen, Phys. Rev. Lett.

2004, 93, 156801 – 156804.[18] M. K. Sharma, R. D. Buchner, W. J. Scharmach, V. Papavassiliou, M. T.

Swihart, Aerosol Sci. Technol. 2013, 47, 858 – 866.[19] P. Kumar Maharana, S. Bharadwaj, R. Jha, J. Appl. Phys. 2013, 114,

014304.[20] U. Banin, Nat. Mater. 2007, 6, 625 – 626.[21] L. E. Murillo, A. M. Goda, J. G. Chen, J. Am. Chem. Soc. 2007, 129,

7101 – 7105.[22] N. Toshima, T. Yonezawab, New J. Chem. 1998, 22, 1179 – 1201.

[23] M. Fern�ndez-Garc�a, J. A. Anderson, G. L. Haller, J. Phys. Chem. 1996,100, 16247 – 16254.

[24] O. S. Alexeev, B. C. Gates, Ind. Eng. Chem. Res. 2003, 42, 1571 – 1587.[25] X. Auvray, L. Olsson, Appl. Catal. B 2015, 168, 342 – 352.[26] P. Araya, C. Weissmann, Catal. Lett. 2000, 68, 33 – 39.[27] Z. Yaakob, M. R. Yosfiah, L. J. Minggu, M. A.Ibrahim, W. R. W.Daud, Pro-

ceedings of the 1st International Conference on Natural Resources Engi-neering & Technology, Putrajaya, Malaysia, 2006, (24 – 25th July), 290 –296.

[28] Metal Clusters in Chemistry (Eds. : P. Braunstein, L. A. Oro, P. R. Raithby),Wiley-VCH, Weinheim, 1999.

[29] P. Braunstein, R. Devenish, P. Gallezot, B. T. Heaton, C. J. Hurnphreys, J.Kervennal, S. Muliey, M. Ries, Angew. Chem. Int. Ed. Engl. 1988, 27,927 – 929; Angew. Chem. 1988, 100, 972 – 973.

[30] B. C. Gates, Chem. Rev. 1995, 95, 511 – 522.[31] S. J. Tauster, S. C. Fung, J. Catal. 1978, 55, 29 – 35.[32] S. Derrouiche, H. Pernot, C. Louis, Chem. Mater. 2012, 24, 2282 – 2291.[33] I. Kasatkin, P. Kurr, B. Kniep, A. Trunschke, R. Schlçgl, Angew. Chem. Int.

Ed. 2007, 46, 7324 – 7327; Angew. Chem. 2007, 119, 7465 – 7468.[34] H. Zhang, J. Sun, V. L. Dagle, B. Halevi, A. K. Datye, Y. Wang, ACS Catal.

2014, 4, 2379 – 2386.[35] Y. Chin, R. Dagle, J. Hu, A. C. Dohnalkova, Y. Wang, Catal. Today 2002,

77, 79 – 88.[36] O. Oyola-Rivera, M. A. Baltan�s, N. Cardona-Mart�nez, J. CO2 Util. 2015,

9, 8 – 15.[37] a) F. Liao, B. T. Lo, D. Sexton, J. Qu, C. Ma, R. C.-T. Chan, Q. Lu, R. Che,

W.-M. Kwok, H. He, S. Fairclough, S. C. E. Tsang, ChemCatChem 2015, 7,230 – 235.

[38] C. Chen, J. Yu, S. Pu, Y. Cheng, P. Chou, J. Phys. Chem. B 2004, 108,10687 – 10691.

[39] P. Chou, C. Cheng, S. Pu, K. Wu, Y. Cheng, C. Lai, Y. Chou, H. Chiu,ChemPhysChem 2006, 7, 222 – 228.

[40] S. Kaniyankandy, S. Rawalekar, S. Verma, H. N. Ghosh, J. Phys. Chem. C2011, 115, 1428 – 1435.

[41] N. Moitra, K. Kanamori, Y. H. Ikuhara, X. Gao, Y. Zhu, G. Hasegawa, K.Takeda, T. Shimadad, K. Nakanishi, J. Mater. Chem. A 2014, 2, 12535 –12544.

[42] F. Liu, P. Zhang, Appl. Phys. Lett. 2010, 96, 043105.[43] F. Liao, T. Woon, B. Lo, D. Sexton, J. Qu, C. Wu, S. Chi, E. Tsang, Catal.

Sci. Technol. 2015, 5, 887 – 896.[44] J. G. Chen, C. A. Menning, M. B. Zellner, Surf. Sci. Rep. 2008, 63, 201 –

254.[45] Z. Li, F. Gao, Y. Wang, F. Calaza, L. Burkholder, W. T. Tysoe, Surf. Sci.

2007, 601, 1898 – 1908.[46] C. Wu, K. M. K. Yu, F. Liao, N. Young, P. Nellist, A. Dent, A. Kroner,

S. C. E. Tsang, Nat. Commun. 2012, 3, 1050.[47] S. Trasatti, J. Electroanal. Chem. 1972, 39, 163.[48] J. A. Widegren, R. G. Finke, J. Mol. Catal. A 2003, 191, 187 – 207.[49] Heterogeneous Catalysis for the Synthetic Chemist, (Ed. : R. L. Augustine),

Marcel Dekker, New York, 1996.[50] Industrial Catalysis : A Practical Approach, 2nd ed. (Ed. : J. Hagen), Wiley-

VCH, Weinheim, 2006.[51] Catalytic Hydrogenation oVer Platinum Metals (Ed. : P. N. Rylander), Aca-

demic Press, New York, 1967.[52] H. Greenfield, N. Y. Ann. Acad. Sci. 1973, 214, 233 – 242.[53] B. Yoon, H. Bin Pan, C. M. Wai, J. Phys. Chem. C 2009, 113, 1520 – 1525.[54] D. Duca, F. Frusteri, A. Parmaliana, G. Deganello, Appl. Catal. A 1996,

146, 269 – 284.[55] H. R. Aduriz, P. Bodnariuk, B. Coq, F. Figueras, J. Catal. 1989, 119, 97 –

107.[56] R. G. Nuzzo, L. H. Dubois, N. E. Bowles, M. A. Trecoske, J. Catal. 1984,

85, 267 – 271.[57] F. M. Mckenna, R. Wells, J. A. Anderson, Chem. Commun. 2011, 47,

2351 – 2353.[58] A. A. Lamberov, S. R. Egorova, I. R. Il’yasov, Kh. Kh. Gil’manov, S. V. Trifo-

nov, V. M. Shatilov, A. Sh. Ziyatdinov, Kinet. Catal. 2007, 48, 136 – 142.[59] A. Moln�r, A. Sarkany, M. Varga, J. Mol. Catal. A 2001, 173, 185 – 221.[60] B. Coq, F. Figueras, J. Mol. Catal. A 2001, 173, 117 – 134.[61] G. Rupprechter, M. Morkel, H. J. Freund, R. Hirschl, Surf. Sci. 2004, 554,

43 – 59.[62] A. Borodzinski, A. Janko, React. Kinet. Catal. Lett. 1977, 7, 163 – 170.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2012

Minireviews

[63] D. Mei, M. Neurock, C. M. Smith, J. Catal. 2009, 268, 181 – 195.[64] Y. Jin, A. K. Datye, E. Rightor, R. Gulotty, W. Waterman, M. Smith, M.

Holbrook, J. Maj, J. Blackson, J. Catal. 2001, 203, 292 – 306.[65] N. A. Khan, S. Shaikhutdinov, H. J. Freund, Catal. Lett. 2006, 108, 159 –

164.[66] T. V. Choudhary, C. Sivadinarayana, A. K. Datye, D. Kumar, D. W. Good-

man, Catal. Lett. 2003, 86, 1 – 8.[67] Y. Han, D. Peng, Z. Xu, H. Wan, S. Zheng, D. Zhu, Chem. Commun.

2013, 49, 8350 – 8352.[68] M. Armbrìster, K. Kovnir, M. Behrens, D. Teschner, Y. Grin, R. Schlçgl, J.

Am. Chem. Soc. 2010, 132, 14745 – 14747.[69] K. Kovnir, D. Teschner, M. Armbrìster, P. Schnçrch, M. H�vecker, A.

Knop-Gericke, Yu. Grin, R. Schlçgl, BESSY Highlights 2007, 22.[70] C. Sungbom, K. Tanaka, Bull. Chem. Soc. Jpn. 1962, 65, 2275 – 2276.[71] V. Ponec, Appl. Catal. A 1997, 149, 27 – 48.[72] J. Qiu, H. Zhang, X. Wang, H. Han, C. Liang, C. Li, React. Kinet. Catal.

Lett. 2006, 88, 269 – 276.[73] H. Knozinger, K. Kochloefl, Heterogeneous Catalysis and Solid Catalysts,

Ullman’s Encyclopedia of Industrial Chemistry, Wiley-VCH, Weinheim,2003.

[74] M. Gao, Y. Xu, J. Jiang, S. Yu, Chem. Soc. Rev. 2013, 43, 2986 – 3017.[75] G. A. Olah, Angew. Chem. Int. Ed. 2005, 44, 2636 – 2639; Angew. Chem.

2005, 117, 2692 – 2696.[76] J. Turner, G. Sverdrup, M. K. Mann, P. C. Maness, B. Kroposki, M. Ghirar-

di, R. J. Evans, D. Blake, Int. J. Energy Res. 2008, 32, 379 – 407.[77] K. M K. Yu, I. Curcic, J. Gabriel, S. C. E. Tsang, ChemSusChem 2008, 1,

893 – 899.[78] J. Hansen, L. Nazarenko, R. Ruedy, M. Sato, J. Willis, A. Del Genio, D.

Koch, A. Lacis, K. Lo, S. Menon, T. Novakov, J. Perlwitz, G. Russell, G. A.Schmidt, N. Tausnev, Science 2005, 308, 1431 – 1451.

[79] Handbook of Homogeneous Hydrogenation (Eds. : P. G. Jessop, H. de V-ries, K. Elsevier), 2007, Wiley-VCH, Weinheim.

[80] T. Sakakura, J.-C. Choi, H. Yasuda, Chem. Rev. 2007, 107, 2365 – 2387.[81] F. Liao, Z. Zeng, C. Eley, Q. Lu, X. Hong, S. C. E. Tsang, Angew. Chem. Int.

Ed. 2012, 51, 5832 – 5836; Angew. Chem. 2012, 124, 5934 – 5938.[82] J. G. Wu, M. Saito, M. Takeuchi, T. Watanabe, Appl. Catal. A 2001, 218,

235 – 240.[83] G. C. Chinchen, P. J. Denny, J. R. Jennings, M. S. Spencer, K. C. Waugh,

Appl. Catal. 1988, 36, 1 – 65.[84] M. Fern�ndez-Garc�a, J. Anderson, G. L. Haller, J. Phys. Chem. 1996, 100,

16247 – 16254.[85] I. Meli�n-Cabrera, M. L. Granados, J. L. G. Fierro, J. Catal. 2002, 210,

285 – 294.[86] Y.-F. Zhao, Y. Yang, C. Mims, C. H. F. Peden, J. Li, D. Mei, J. Catal. 2011,

281, 199 – 211.[87] A. L. Bonivardia, D. L. Chiavassaa, C. A. Querinib, M. A. Baltan�sa, Stud.

Surf. Sci. Catal. 2000, 130, 3747 – 3752.[88] N. Iwasa, H. Suzuki, M. Arai, N. Takezawa, Catal. Lett. 2004, 96, 75 – 78.[89] K. M. Neyman, R. Sahnoun, C. Inntam, S. Hengrasmee, J. Phys. Chem. B

2004, 108, 5424 – 5430.[90] Z. X. Chen, K. M. Neyman, K. H. Lim, N. Rçsch, Langmuir 2004, 20,

8068 – 8077.[91] T. C. Johnson, D. J. Morris, M. Wills, Chem. Soc. Rev. 2010, 39, 81 – 88.[92] F. Joû, ChemSusChem 2008, 1, 805 – 808.[93] B. Loges, A. Boddien, F. Gartner, H. Junge, M. Beller, Top. Catal. 2010,

53, 902 – 914.[94] X. Zhou, Y. Huang, W. Xing, C. Liu, J. Liao, T. Lu, Chem. Commun. 2008,

3540 – 3542.[95] O. Metin, X. Sun and S. Sun, Nanoscale, 2013, 5, 910 – 912 .[96] N. Akiya, P. E. Savage, AIChE J. 1998, 44, 405 – 415.[97] X. Gu, Z. H. Lu, H. L. Jiang, T. Akita, Q. Xu, J. Am. Chem. Soc. 2011, 133,

11822 – 11825.[98] S. Zhang, ©. Metin, D. Su, S. Sun, Angew. Chem. Int. Ed. 2013, 52,

3681 – 3684; Angew. Chem. 2013, 125, 3769 – 3772.[99] K. S. Kim, M. A. Barteau, Langmuir 1990, 6, 1485 – 1488.

[100] N. Hoshi, M. Nakamura, K. Kida, Electrochem. Commun. 2007, 9, 279 –282.

[101] K. Tedsree, T. Li, S. Jones, C. W. A. Chan, K. M. K. Yu, P. J. Bagot, E. A.Marquis, G. D. W. Smith, S. C. E. Tsang, Nat. Nanotechnol. 2011, 6, 302 –307.

[102] D. E. Bloom, Science 2011, 333, 562 – 569.

[103] R. D. Cortright, R. R. Davda, J. A. Dumesic, Nature 2002, 418, 964 – 967.[104] A. M. Omer, Renew. Sust. Energ. Rev. 2008, 12, 1789 – 1821 .[105] J. Tollefson, Nature 2011, 473, 134 – 134.[106] G. Bolye, Renewable Energy – Power for a Sustainable Future, Oxford

University Press, Oxford, 1996.[107] J. Mitra, X. Zhou, T. Rauchfuss, Green Chem. 2015, 17, 307 – 313.[108] G. W. Huber, R. D. Cortright, J. A. Dumesic, Angew. Chem. Int. Ed. 2004,

43, 1549 – 1551; Angew. Chem. 2004, 116, 1575 – 1577.[109] J. R. Rostrup-Nielsen, Steam Reforming Catalysts, Danish Technical

Press, Copenhagen, 1975.[110] J. Rostrup-Nielsen, Phys. Chem. Chem. Phys. 2001, 3, 283 – 288.[111] R. R. Davda, J. W. Shabaker, G. W. Huber, R. D. Cortright, J. a. Dumesic,

Appl. Catal. B 2005, 56, 171 – 186.[112] G. W. Huber, J. W. Shabaker, S. T. Evans, J. A. Dumesic, Appl. Catal. B

2006, 62, 226 – 235.[113] A. Tanksale, J. N. Beltramini, J. A. Dumesic, G. Q. Lu, J. Catal. 2008, 258,

366 – 377.[114] Z. Wie, J. Sun, Y. Li, A. Datye, Y. Wang, Chem. Soc. Rev. 2012, 41, 7994 –

8008.[115] S. S�, H. Silva, L. Brandao, J. M. Sousa, A. Mendes, Appl. Catal. B 2010,

99, 43 – 57.[116] J. C. Amphlett, K. A. M. Creber, J. M. Davis, R. F. Mann, B. A. Peppley,

D. M. Stokes, Int. J. Hydrogen Energy 1994, 19, 131 – 137.[117] S. D. Davidson, H. Zhang, J. Sun, Y. Wang, Dalton Trans. 2014, 43,

15038 – 15047.[118] J. P. Breen, J. R. H. Ross, Catal. Today 1999, 51, 521 – 533.[119] C. J. Jiang, D. L. Trimm, M. S. Wainwright, N. W. Cant, Appl. Catal. A

1993, 93, 245 – 255.[120] M. Shimokawabe, H. Asakawa, N. Takezawa, Appl. Catal. 1990, 59, 45.[121] N. Iwasa, S. Kudo, H. Takahashi, S. Masuda, N. Takezawa, Catal. Lett.

1993, 19, 211 – 216.[122] C. Cao, G. Xia, J. Holladay, E. Jones, Y. Wang, Appl. Catal. A 2004, 262,

19 – 29.[123] B. Halevi, E. J. Peterson, A. De La Riva, E. Jeroro, V. M. Lebarbier, Y.

Wang, J. M. Vohs, B. Kiefer, E. Kunkes, M. Havecker, M. Behrens, R.Schlçgl, A. K. Datye, J. Phys. Chem. C 2010, 114, 17181 – 17190.

[124] E. J. Peterson, B. Halevi, B. Kiefer, M. N. Spilde, A. K. Datye, J. Peterson,L. Daemen, A. Llobet, H. Nakotte, J. Alloys Compd. 2011, 509, 1463 –1470.

[125] B. Halevi, E. J. Peterson, A. Roy, A. DeLariva, E. Jeroro, F. Gao, Y. Wang,J. M. Vohs, B. Kiefer, E. Kunkes, M. Havecker, M. Behrens, R. Schlçgl,A. K. Datye, J. Catal. 2012, 291, 44 – 54.

[126] N. Iwasa, N. Takezawa, Top. Catal. 2003, 22, 215 – 224.[127] M. Goula, S. K. Kontou, P. E. Tsiakaras, Appl. Catal. B 2004, 49, 135 – 144.[128] J. K. Nørskov, J. Rossmeisl, a. Logadottir, L. Lindqvist, J. R. Kitchin, T. Bli-

gaard, H. Jûnsson, J. Phys. Chem. B 2004, 108, 17886 – 17892.[129] V. Mazumder, M. Chi, K. L. More, S. Sun, J. Am. Chem. Soc. 2010, 132,

7848 – 7849.[130] J. L. Fern�ndez, D. A. Walsh, A. J. Bard, J. Am. Chem. Soc. 2005, 127,

357 – 365.[131] J. L. Fern�ndez, V. Raghuveer, A. Manthiram, A. J. Bard, J. Am. Chem.

Soc. 2005, 127, 13100 – 13101.[132] V. Raghuveer, A. Manthiram, A. J. Bard, J. Phys. Chem. B 2005, 109,

22909 – 22912.[133] V. Raghuveer, P. J. Ferreira, A. Manthiram, Electrochem. Commun. 2006,

8, 807 – 814.[134] W. E. Mustain, K. Kepler, J. Prakash, Electrochem. Commun. 2006, 8,

406 – 410.[135] M. R. Tarasevich, A. E. Chalykh, V. A. Bogdanovskaya, L. N. Kuznetsova,

N. A. Kapustina, B. N. Efremov, M. R. Ehrenburg, L. A. Reznikova, Electro-chim. Acta 2006, 51, 4455 – 4462.

[136] T. Huang, P. Liu, J. Zhang, R. R. Adzic, The 207th Electrochemistry SocietyMeeting, 2006, 501, 1588.

[137] C. Xu, Y. Liu, H. Zhang, H. Geng, Chem. Asian J. 2013, 8, 2721 – 2728.[138] M. M. Hossain, M. A. Al-Saleh, M. A. Shalabi, T. Kimura, T. Inui, Appl.

Catal. A 2004, 274, 43 – 48.[139] R. T. Yang, A. J. Hernandez-Maldonado, F. H. Yang, Science 2003, 301,

79 – 81.[140] C. Song, Catal. Today 2003, 86, 211 – 263.[141] J. Torres-Nieto, W. W. Brennessel, W. D. Jones, J. J. Garc�a, J. Am. Chem.

Soc. 2009, 131, 4120 – 4126.

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2013

Minireviews

[142] D. Buccella, K. E. Janak, G. Parkin, J. Am. Chem. Soc. 2008, 130, 16187 –16189.

[143] A. Sattler, G. Parkin, Nature 2010, 463, 523 – 526.[144] A. Sattler, G. Parkin, J. Am. Chem. Soc. 2011, 133, 3748 – 3751.[145] A. M. Venezia, V. La Parola, G. Deganello, B. Pawelec, J. L. G. Fierro, J.

Catal. 2003, 215, 317 – 325.[146] W. Fu, L. Zhang, T. Tang, Q. Ke, S. Wang, J. Hu, G. Fang, J. Li, F. S. Xiao,

J. Am. Chem. Soc. 2011, 133, 15346 – 15349.[147] A. Niquille-Rçthlisberger, R. Prins, J. Catal. 2006, 242, 207 – 216.[148] M. H. Shao, K. Sasaki, R. R. Adzic, J. Am. Chem. Soc. 2006, 128, 3526 –

3527.[149] S. Jones, S. M. Fairclough, M. Gordon-Brown, W. Zheng, A. Kolpin, B.

Pang, W. C. H. Kuo, J. M. Smith, S. C. E. Tsang, Chem. Commun. 2015,51, 46 – 49.

[150] A. Nilsson, L. G. M. Pettersson, B. Hammer, T. Bligaard, C. H. Christen-sen, J. K. Nørskov, Catal. Lett. 2005, 100, 111 – 114.

[151] C. Wong, A. Chan, A. H. Mahadi, M. M. Li, E. C. Corbos, C. Tang, G.Jones, W. Chun, H. Kuo, J. Cookson, C. M. Brown, P. T. Bishop, S. C. E.Tsang, Nat. Commun. 2014, 5, 5787.

[152] R. Burch, F. A. Lewis, Trans. Faraday Soc. 1970, 66, 727 – 735.[153] H. Rong, S. Cai, Z. Niu, Y. Li, ACS Catal. 2013, 3, 1560 – 1563.[154] Y. Wu, D. Wang, P. Zhao, Z. Niu, Q. Peng, Y. Li, Inorg. Chem. 2011, 50,

2046 – 2048.[155] S. Lee, R. Hoffmann, J. Am. Chem. Soc. 2002, 124, 4811 – 4823.

Received: March 7, 2015Revised: May 9, 2015Published online on June 26, 2015

ChemCatChem 2015, 7, 1998 – 2014 www.chemcatchem.org Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim2014

Minireviews