The Impacts of Nanotechnology on Catalysis by Precious Metal Nanoparticles

Review

The impacts of nanotechnology on catalysis by precious metal nanoparticles

Rongchao Jin synthesis of ammonium in 1913, industrial catalysis has Department of Chemistry , Carnegie Mellon University, 4400 been practiced for nearly a century; its signifi cance for the Fifth Avenue, Pittsburgh, PA 15213 , USA , petrochemical industry was particularly realized when the e-mail: [email protected] oil crisis occurred in the 1970s. The importance of catalysis is also refl ected in environmental protection and public health; a well-known example is the catalytic converters for Abstract removing toxic emissions of automobiles, which were fi rst developed by General Motors Corporation and Ford Motor This review article focuses on the impacts of recent Company in 1974. advances in solution phase precious metal nanoparticles Catalysis is a complex and highly interdisciplinary sci- on heterogeneous catalysis. Conventional nanometal cata- ence; it integrates chemistry, materials science, and chemi- lysts suffer from size polydispersity. The advent of nano- cal reaction engineering. Apparently, catalysis constitutes a technology has signifi cantly advanced the techniques for central theme of chemical research as the primary activity of preparing uniform nanoparticles, especially in solution chemists is to perform reactions, which almost exclusively phase synthesis of precious metal nanoparticles with excel- involve catalysts, such as metals. Generally speaking, a cata- lent control over size, shape, composition and morphology, lyst is a special substance that can speed up chemical reac- which have opened up new opportunities for catalysis. This tions without itself being consumed in the reaction process; review summarizes some recent catalytic research by using note that in some (but rare) cases, a catalyst is used to slow well-defi ned nanoparticles, including shape-controlled down chemical reactions. The power of a catalyst lies in its nanoparticles, high index-faceted polyhedral nanocrys- capability in accelerating chemical reactions by reducing the tals, nanostructures of different morphology (e.g., core- energy barrier (i.e., activation energy) for the transition state shell, hollow, etc.), bi- and multi-metallic nanoparticles, as and in controlling reaction pathways for selective synthesis well as atomically precise nanoclusters. Such well-defi ned of desired product. With respect to materials with catalytic nanocatalysts provide many exciting opportunities, such as power, it is interesting that almost all types of substances identifying the types of active surface atoms (e.g., corner (e.g., acids, bases, metals, semiconductors, clays, carbon, and edge atoms) in catalysis, the effect of surface facets on organometallic complexes, nucleic acids, proteins, etc.) can catalytic performance, and obtaining insight into the effects serve as catalysts for certain chemical processes. The earliest of size-induced electron energy quantization in ultra-small observation of catalytic action dates back to the 19th century metal nanoparticles on catalysis. With well-defi ned metal when chemistry fi rst came into being. J ö ns Jacob Berzelius nanocatalysts, many fundamentally important issues are (1779 – 1848) is generally credited for being the fi rst person expected to be understood much deeper in future research, who carried out systematic, scientifi c study on catalysis. In such as the nature of the catalytic active sites, the metal- 1836, he integrated early observations of catalytic power support interactions, the effect of surface atom arrange- of special substances into a systematic body of knowledge ment, and the atomic origins of the structure-activity and and coined for the fi rst time the term “ catalysis ” . Currently, the structure-selectivity relationships. catalysis is recognized widely, not only in organic or inorganic chemistry (e.g., homogeneous and heterogeneous Keywords: heterogeneous catalysis; nanocatalysis; catalysis) but also in life and all living things (e.g., enzymatic nanoclusters; nanocrystals; nanoparticles; precious metals. catalysis). This review article focuses on heterogeneous catalysis by precious metal nanoparticles. Unlike homogeneous and 1. Introduction enzymatic catalysis, heterogeneous catalysis refers to a catalytic process that typically involves solid catalysts and reac- Catalysis is of tremendous importance for the chemical tants of different phases (e.g., gas or liquid). The solid-state industry. Approximately two-thirds of the chemical prod- catalysts are rather versatile, such as metals (e.g., Pt, Pd, Ag, ∼ ucts and 90 % of the chemical processes involve cataly- Au), semiconductors (e.g., TiO 2, CdS), zeolites, molecular sis (e.g., homogeneous, heterogeneous, or enzymatic sieves, and so on. Precious metal catalysis constitutes an type). Since the fi rst industrialized catalytic process – the important branch of heterogeneous catalysis in the chemical 32 R. Jin: Nanocatalysis industry; for example, the well-known catalytic converter 2. Conventional heterogeneous metal catalysts utilizes three metals (Rh, Pt, and Pd) of the platinum group. In retrospect, Paul Sabatier (1854 – 1941, awarded a Nobel Heterogeneous catalysts of metals are composed of two Prize in 1912) was the fi rst to carry out hydrogenation of major components: the active metal particles and the sup- organic compounds using metal (Ni) powder catalysts, and port. Typical supports are Al2 O 3 , SiO2 , MgO, Fe2 O3 , TiO 2 ,

Irving Langmuir (at General Electric) performed CO oxida- CeO2 , and many others. The most widely used conventional tion with Pd catalyst. In the Periodic Table, precious metals method for preparing metal catalysts is the wet impregnation (also called noble metals) refer to Ag, Au, Pd, Pt, Rh, Ir, Ru, method [1] . In this method, the support (usually in powders) and Os, with the latter six elements known as the platinum is soaked in a solution of metal salt (with metal loading, say, group. Rhenium (Re) may also be included in the group of of ∼ 0.1 – 10 wt % ), followed by drying, then by thermal treat- precious metals. ment (i.e., calcination) [2] . The metal salt decomposes under Heterogeneous catalysts of precious metals almost exclu- high temperatures and converts to nanoparticles, which are sively involve the small particle form, often on the nanoscale fi nally dispersed on the support surfaces. Hydrogen pretreat- (e.g., 1– 100 nm in diameter), although in some cases microscale ment is often applied to turn such pro-catalysts into active particles are also used (e.g., Ag particles of a few µ m in size ones (i.e., to reduce the surface oxidized metal nanoparticles are used in industrial epoxidation of ethylene). Physical “ see- into metallic state). ing ” of nanoparticles was apparently not feasible prior to the Apparently, the impregnation method has a poor invention of electron microscopy and, thus, early studies of control over the metal particle size and the size distribu- heterogeneous catalysis (e.g., during the fi rst half of the 20th tion. However, in practical work, researchers may not care century) were rather “ dark ” . Even today, catalysis is still some- about those aspects, as long as high activity and selectivity times called “ black art” owing to the nature of catalysts being of the catalyst can be attained, which is why catalysis was largely unknown. However, in the past two decades, signifi cant once called “ black art” . Nevertheless, in modern research, developments in nanotechnology have enabled researchers not the availability of many nanoscale characterization tools, only to see nanoparticles clearly at atomic resolution and create especially electron microscopy, has enabled researchers to well-defi ned nanoparticles but also to look into the very fun- obtain a great deal of invaluable information of catalysts damentals of catalytic processes. Such remarkable, signifi cant and to understand some mechanistic aspects of catalysis, progress in catalytic research repudiates the previous “ black hence no longer “ black art ” . The impregnation method, art ” notoriety of catalysis and brings catalysis into an exciting although very old, is still a very popular method for prepar- era of scientifi c research at the ultimate atomic level, which has ing heterogeneous catalysts, even today. Many variations long been dreamed by catalysis chemists. or new methods have been developed in the past decades, In this review article, I will fi rst briefl y discuss conven- such as coprecipitation, which involves simultaneous pre- tional nanocatalysts (Section 2) to usher new researchers into cipitation of the active metal and the support, deposition- the catalysis fi eld, then present the major advances in the cre- precipitation [3] , sol-gel method [4] , deposition of colloids ation of well-defi ned nanoparticles in solution phase (Chart [5] , and so on. 1 ) and their catalytic applications (Section 3), and fi nally pro- The metal nanocatalysts prepared by conventional, salt- vide my personal perspective on some future developments of based methods often consist of polydisperse nanoparticles, catalytic science, including challenging issues and prospects which preclude studies of particle size dependence or obscure of the catalysis fi eld (Section 4). As this review focuses on the size-dependent catalytic results. This situation is even the catalytic materials, theoretical aspects of catalysis and worse when it comes to very small nanoparticles, say below catalytic reaction mechanisms will not be covered. Also, this 5 nm (diameter), because it is even harder to control the uni- review article is not intended to be a comprehensive one, thus formity of such very small nanoparticles. Moreover, there many excellent works from my colleagues may not have been is no guarantee that the deposited metal salt has been com- included owing to my unawareness. pletely converted to nanoparticles during the calcination step.

Chart 1 Various types of nanoparticles synthesized in solution phase for application in catalysis [ligands are omitted for clarity, except

Au25 (SCH2 CH2 Ph)18 shown in the last panel in which the full structure is drawn]. R. Jin: Nanocatalysis 33

Some salt species can still possibly exist as ions on the sup- catalysis is essentially about the surface. Thus, achieving port surface, or some are present as few-atom clusters (< 1 control over the exposed facets of nanoparticles is critical nm) that are very hard to be detected by routine transmission for catalysis. The size dependence is well known in catalysis electron microscopy (TEM) analysis (e.g., bright fi eld imag- and has long been pursued [6] , but the shape dependence was ing). Thus, from the viewpoint of fundamental studies on much less explored in the past primarily owing to the unavail- catalysis, well-defi ned catalysts, at least well-defi ned metal ability of high-quality, shape-controlled nanocrystals. nanoparticles should be attained, although the support may With respect to the shape dependence, it is worth noting still be ill-defi ned; the support is, in most cases, less critical that previous studies on bulk single crystal surfaces explicitly compared with the metal component as the metal particles are demonstrated that the catalytic activity and/or selectivity are the active catalytic species. In previous research, it has been strongly dependent on the exposed crystallographic planes extensively demonstrated that the catalytic reactivity of nano- [7, 8] . For example, with regard to the Fe catalyst for ammo- particles is highly dependent on their size and composition, nia synthesis, the (111) surface was found to be most active, as well as on some other parameters. However, fundamental and the order of reaction rate follows (111) > > (110)> (100) understanding of real-world catalysts (as opposed to single [7, 8] . The availability of shape-controlled metal nanocrystals crystal model catalysts) was only met with limited success. in recent years has provided some exciting opportunities for Fortunately, the advent of nanotechnology has signifi cantly catalytic studies [9] . advanced the techniques for preparing well-defi ned nanopar- Among the precious metals, Ag, Au, Pd, Pt, Rh, and Ir are ticles, especially in solution phase synthesis of precious metal all of face-centered cubic (FCC) structure. The FCC unit cell nanoparticles with excellent control over particle size, shape, is shown in Figure 1 , together with the most common crystal and morphology, etc., which have opened up new, exciting planes of {100}, {110}, and {111}. “ Spherical ” nanoparticles opportunities for catalysis. are actually enclosed by many types of facets of different Miller indices, most commonly by {111} and {100} facets. The {111} surface exhibits hexagonal close-packing of atoms, 3. Impacts of nanotechnology on nanocatalysis and the {100} surface shows square-packed atoms, whereas by precious metals the {110} surface shows an open structure with the second layer of atoms also being accessible to reactants (Figure 1 ). Small metal nanoparticles have long been used in the fi eld The preferred exposure of these low-index facets is due to the of heterogeneous catalysis for many decades, although not fact that these surfaces have lower energy, and hence are more explicitly declared previously. Compared with bulk metals, stable. Nanoparticles of different shapes also possess differ- the nanoparticle form leads to a signifi cantly higher surface- ent fractions of atoms on the corners and edges. Such corner to-volume ratio (i.e., having a much higher fraction of surface and edge atoms are of particular interest in catalysis. atoms) as well as reduced cohesive energy; hence, the surface Many shapes of precious metal nanocrystals have been atoms (especially those at sharp corners and edges) become reported in the literature, each with specifi c crystallographic very reactive. The properties of surface atoms are inti- facets. With respect to chemical synthesis, several methods mately related to the underlying atoms which are not directly have been developed for preparing solution phase metal exposed; the latter are essentially in bulk-like coordination nanoparticles with excellent shape control. Some versatile environments, although size-induced electron energy quanti- approaches are, e.g., the kinetically controlled method, the zation may occur in small enough metal nanoparticles. photo-induced growth method, the seed-mediated method, When shrinking metal bulks into nanoparticles, the surface and the polyol method, which have enabled the synthesis of a atoms will experience distinct changes in atomic arrangement variety of nanostructures, such as nanotetrahedra [10] , nano- as well as in surface electronic properties. In addition, small prisms [11] , nanorods [12] , nanocubes [13, 14], and nanoparticles possess a higher degree of curvature, which could be wires [15] . benefi cial for those surface-sensitive reactions. The geomet- In 1996, El-Sayed and coworkers reported shape control of ric (e.g., size and shape), electronic, and other related effects Pt nanoparticles (including tetrahedral, cubic, and truncated of surface atomic sites are all of major importance for catal- octahedral shapes) using a hydrogen reduction method in the ysis. The signifi cant advances in nanotechnology have cre- presence of polymer capping agents (such as polyacrylate) ated some unique opportunities for catalytic research. With well-defi ned metal nanocatalysts, many important issues are expected to be understood much deeper, such as the nature of c the catalytic active sites, the metal-support interactions, the effect of surface atom arrangement, and the structure-activity and structure-selectivity relationships. Below we shall select some topics for a more detailed discussion. b 3.1. Shape effect of metal nanoparticles on catalysis a {100} {110} {111} One of the major advances in solution phase nanochemistry lies in the shape control of nanoparticles. As is known, Figure 1 FCC unit cell and low-index crystallographic planes. 34 R. Jin: Nanocatalysis

[10] . The Pt nanotetrahedra are enclosed by four {111} facets less than Pd nanoparticles (e.g., spherical ones); note that Pd and have sharp corners and edges, whereas the nanocubes are nanoparticles, regardless of particle shape, are highly active composed of {100} facets and have somewhat rounded cor- in catalyzing the Suzuki reaction. ners and edges, and the spherical nanoparticles are not ideally Tsung et al. [20] synthesized Pt nanocubes of various sizes spherical, instead, “ near spherical” , and possess many small from 5 to 9 nm by a polyol process (Figure 3 ) and loaded {111} and {100} facets with corners and edges located at such particles onto mesocellular foam (MCF)-17 mesopo- the boundaries of these facets (Figure 2 , left panels A, C and rous silica for evaluating the catalytic properties. The eth- E). These distinct differences in surface structures prompted ylene hydrogenation rate was found to be independent of a comparison of the catalytic efﬁ ciency of these different- both size and shape of the Pt nanocatalysts; this result is shaped nanoparticles. The Pt nanotetrahedra ( ∼ 4.8 nm) and indeed comparable to Pt single crystals. For the other reac- nanocubes (∼ 7.1 nm) were investigated for catalyzing the tion investigated, pyrrole hydrogenation, the Pt nanocubes electron-transfer reaction between hexacyanoferrate(III) were found to enhance the ring-opening ability and showed and thiosulfate ions in solution at room temperature and a higher selectivity to n-butylamine compared to the nano- pH 7 [16, 17] . The Pt nanotetrahedra [herein capped by polyhedra (Figure 3 , panels I and J). In related work, mono- poly(vinylpyrrolidone), PVP] showed the highest activity, disperse sub-10 nm Rh nanocubes were also synthesized by which was attributed to its highest fraction of surface atoms the polyol method [21] . The {100}-faceted Rh nanocubes on the corners and edges as Pt nanotetrahedra gave rise to the were capped by both tetradecyltrimethylammonium bromide lowest activation energy (Figure 2 , right panels B, D, and F) (TTAB) surfactant and PVP, but the capping agents did not [18] . The PVP-capped Pt nanotetrahedra were also utilized to prevent catalytic activity of the nanocube for pyrrole hydro- catalyze the Suzuki cross-coupling reaction between phenyl- genation and CO oxidation, respectively. Highly uniform Pd boronic acid and iodobenzene to form biphenyl at 100° C [19] , nanocubes were also reported, but their catalytic activities but such nanoparticles were found to be not very active, much have not been explored [14] .

AB Tetrahedral Pt NPs 100 -5.0 E=14.0±0.6 kJ mol-1 80 -5.1 ) 60 -1 -5.2 40 -5.3

20 In k (min -5.4 0 -5.5 RT DT S 3.10 3.15 3.20 3.25 3.30 3.35 3.40 10 nm Shape 1000/T (K-1)

CDCubic Pt NPs 100 -6.4 -1 80 E=26.4±1.3 kJ mol ) -6.6 60 -1 -6.8 40

In k (min -7.0 20

% of Nanoparticles % of Nanoparticles -7.2 0 RC DC TO 3.10 3.15 3.20 3.25 3.30 3.35 3.40 10 nm Shape 1000/T (K-1)

E F Spherical Pt NPs 100 -5.8

80 -6.0 E=22.6±1.2 kJ mol-1 ) -1 60 -6.2 40 -6.4 In k (min 20 -6.6 % of Nanoparticles 0 -6.8 S O 3.10 3.15 3.20 3.25 3.30 3.35 3.40 3.45 Shape 1000/T (K-1)

Figure 2 Shape-dependent catalytic activity of Pt nanotetrahedra (A), nanocubes (C), and nanospheres (E). Panels (B, D, and F) show shape distributions and activation energies from Arrhenius plots. From Ref. [16] . R. Jin: Nanocatalysis 35

A CEGI

BDFHJ

Figure 3 Pt nanocubes of (A, B) 9 nm, (C, D) 7 nm, (E, F) 6 nm, (G, H) 5 nm, and nanopolyhedra of 5 nm (I, J). The upper panels show regular TEM images (scale bar: 20 nm), and the lower panels for high resolution transmission electron microscopy (HR-TEM) images (scale bar: 1 nm). From Ref. [20] .

Somorjai and coworkers have carried out extensive inves- (∼ 160± 10 nm), and nanofl owers (300– 400 nm) via a photo- tigations on the shape-dependent catalysis. Pt nanoparticles chemical method [ 25 ]. All these Rh nanoparticles of different of cubic and cuboctahedral shapes capped by TTAB surfac- morphology are capped by CTAB surfactant. The Rh nanocubes tant were compared for the reaction of benzene hydrogena- exhibited the highest activity towards the catalytic reduction of tion [22] . The nanoparticle shape was found to strongly affect aromatic nitro compounds by NaBH4 in aqueous solution [ 25 ]. the catalytic selectivity: cuboctahedral Pt nanoparticles gave The different shape dependences (Au vs. Rh) are interesting rise to both cyclohexane and cyclohexene, whereas cubic Pt and further work is needed to gain insight into the differences. nanoparticles only produced cyclohexane. Interestingly, the The shape dependence of Pd nanocrystals (hexagonal vs. selectivities were similar to what were obtained on Pt (111) spherical) has also been investigated for selective alkyne and (100) single crystals, but the apparent activation ener- hydrogenation of 2-methyl-3-butyn-2-ol [26] . For Ag nano- gies (E a ) on nanocrystals were lower than on single crystal Pt catalysts, Xu et al. prepared three distinct shapes, nanoprisms = ± surfaces, e.g., E a 10.9 0.4 kcal/mol for cyclohexane produc- with predominant {111} facets, nanocubes with {100} facets, tion on cubic nanoparticles, 8.3± 0.2 and 12.2± 0.4 kcal/mol and quasi-spherical Ag nanoparticles (Figure 5 ) [27] . Catalytic for respective cyclohexane and cyclohexene production on evaluations for the styrene oxidation reaction showed that the cuboctahedral nanoparticles. Ag nanocubes was the most active, which was over 7 times Mostafa et al. investigated the infl uence of the shape of higher than that for the nanoplates and ∼ 3 times higher than ∼ γ Pt nanocatalysts ( 0.8 – 1 nm diameter, supported on -Al 2 O3 ) that for spherical particles. on the catalytic reactivity [23] . Such nanoparticles with dif- In electrocatalysis, Pt is generally the best catalyst for fuel ferent shapes were synthesized by inverse micelle encap- cell reactions, but its high cost and lower abundance on earth sulation, and the nanocatalysts were found to show distinct render Pt not feasible for world-wide application in fuel cell differences in the activity for 2-propanol oxidation to acetone cars, and thus searching for alternative catalysts is necessary. (below 100 °C). These authors found that the decreasing onset In this regard, signifi cant work has been carried out in recent reaction temperature for 2-propanol oxidation correlated well years, and nanotechnology apparently plays a vital role. The with the increasing number of missing bonds at the particle price of Pd is around one-third of the Pt price. With different surface, implying that undercoordinated Pt surface atoms shaped Pd nanoparticles (rods vs. spheres), Xiao et al. found a (those at corners and edges) are the most active reaction sites strong shape-dependence of the electrocatalytic activity in the [23] . oxygen reduction reaction (ORR) process [28] . The surface- Kundu et al. prepared Au nanospheres, nanorods, and nano- specifi c activity of Pd nanorods with exposed {110} facets prisms capped by cetyltrimethylammonium bromide (CTAB) was found to be ∼ 10-fold higher than that of Pd nanospheres, (Figure 4 ) and observed different catalytic activity for the becoming comparable to that of Pt at operating potentials of reduction of aromatic nitro compounds to amino derivatives at fuel cell cathodes for the ORR process. They ascribed the room temperature [24 ]. The reaction was found to be fastest superior activity of Pd nanorods to the much weaker interac- with nanospheres and slowest with nanorods, whereas nano- tion between the Pd {110} facet and O adatoms, which ben- prisms were intermediate; note that the numbers of particles in efi ts the ORR process. Seo et al. [29] prepared Au nanoplates - - the catalytic reaction was kept approximately the same. Kundu from Au(CN) 2 (as opposed to AuCl4 ) via electrodeposition. et al. also synthesized Rh nanospheres (60± 5 nm), nanocubes Such Au nanoplates were rich in Au {110} and {100} facets, 36 R. Jin: Nanocatalysis

A B

100 nm 50 nm

C D

100 nm 50 nm

Figure 4 TEM images of gold nanoparticles of different shapes: (A) nanospheres (45± 5 nm), (B) nanorods (aspect ratio 2.8± 0.2), (C) nanorods (aspect ratio 33± 0.5), and (D) nanoprisms (edge length ∼ 60 ± 5 nm, thickness measured by atomic force microscopy: 3.5 – 4 nm). From Ref. [ 24 ].

A BC

200 nm 50 nm 100 nm

D {100} {111} {111} ~15 nm ~50 nm ~50 nm {100}

~200 nm

Figure 5 TEM images of different shaped Ag nanoparticles: (A) truncated triangular nanoplates, (B) quasi-spherical nanoparticles, and (C) nanocubes; insets: scanning electron microscopy (SEM) images and selected area electron diffraction pattern (SAED) patterns, respectively. Panel (D) shows structural models for the different shapes. From Ref. [27] . and exhibited enhanced electrocatalytic activity for oxygen price for the time being. In a recent review, Koper discussed reduction and glucose oxidation, respectively [29] . It is worth the special role of two types of active sites on nanoparticle noting that the current gold price has become unexpectedly surfaces for electrocatalytic reactions, including (i) steps/ higher than that of platinum owing to the depressed economy; defects in the {111} terraces or facets and (ii) the {100} ter- thus, gold-based electrocatalysts do not have an incentive in races or facets [30] . R. Jin: Nanocatalysis 37

The shape dependence of metal nanoparticles is of major 3.2. High-index faceted nanoparticles interest and importance in catalysis [9] . The above examples only provide an introduction to this new exciting direction, Among the shape-controlled nanoparticles, there is a spe- and much work remains to be done in the future in terms cial type of polyhedral nanocrystals, which are exposed with of the catalytic properties of shaped-controlled nanoparti- high-index facets (Figure 6 ), such as {730}, {321}, {311}, cles. There are several disadvantages of currently available etc., rather than low-index ones such as {100} for nanocubes. shaped-controlled nanocrystals. First of all, the shape purity High-index (hkl ) facets should have at least one index being still needs to be improved; for example, in the studies of the larger than unity, such as (210). Pt shape dependence [16] , the tetrahedral nanoparticles were The relationship between different types of FCC metal only of ∼ 76 % purity and the cubic ones only of ∼ 61 % purity. polyhedrons bounded by low- and high-index crystallographic In this regard, single particle studies, which follows the facets is shown in Figure 7 . The three vertices of the triangu- catalysis of individual nanoparticles in real time (at single lar diagram show regular polyhedrons enclosed by low-index turnover resolution) using the single molecule fl uorescence facets, including octahedrons with {111} facets (bottom left microscopic technique, give rise to useful information [31, vertex), cubes with {100} facets (bottom right vertex), and 32] . Second, the currently available nanocrystals of distinct rhombic dodecahedrons with {110} facets (top vertex). The shapes are often rather large, typically of several tens to hun- high-index faceted polyhedrons are shown on the three sides dreds of nanometers, only in rare cases are shape-controlled of the triangle, as well as those within the triangle. There are nanocrystals of a few nm in size achieved [33] . The large four types of such polyhedrons, including trisoctahedron, tet- size gives rise to a low atomic effi ciency compared to smaller rahexahedron, hexoctahedron, and trapezohedron. nanoparticles. To achieve high activity and also reduce the In nanocrystals, high-index facets can be identifi ed in cost of precious metal catalysts, it is highly desirable to TEM imaging by the analysis of the projection angles and develop facile methods for making smaller (around < 10 – 20 the atomic arrangement of the edge-on facets. Such {hkl } sur- nm) nanocrystals of specifi c shapes. The development of faces are open structures, giving rise to rich atomic steps and facile synthetic methods of shape-controlled nanocrystals is kinks that are low coordinated. Such undercoordinated atoms also of particular importance to some catalysis researchers often serve as active sites and adsorb reactant molecules more who have less experience in nanoparticle synthesis and shape strongly, and hence can easily break the bonds and activate the control, as facile synthetic methods can be easily followed molecules. In previous electrochemical studies, high-indexed by them. single crystal planes of metals demonstrated high reactivity Other issues such as the stability of shape-controlled nano- and stability. Therefore, high index-faceted nanocrystals are crystals during the catalytic reaction processes and leaching of great interest to nanocatalysis. By choosing appropriate of metal atoms need to be more carefully studied in future ligands and exerting kinetic control, polyhedral nanopar- work. Non-spherical nanocrystals have generally been found ticles with certain types of high-index facets can indeed be to be more active than spherical counterparts, but inconsistent prepared. At fi rst sight, one would speculate that nanoparti- results certainly exist, and a unifi ed picture is still missing. cles should be enclosed by low energy facets (i.e., low-index Moreover, the nature of shape dependence is not fully under- facets) such as {111}, {100}, and {110}, for that they are stood. Comparison between nanocrystals with specifi c facets more favorable in energy than high-index facets. The surface and bulk single crystal surfaces also remains to be investi- energy of different crystallographic facets of a FCC metal gated in more detail. follows the increasing order of γ {111} < γ {100} < γ {110}

110

Zone [001] 551 320 _ 210 (331) 331 851 Zone221 [110] 310 (310) 731 510 332 421 (751) 721 732

111 533 211 _311 411 711 100 Zone [011]

(311)

Figure 6 Unit stereographic triangle of FCC metals and models of surface atomic arrangements. Adapted from Ref. [34] . 38 R. Jin: Nanocatalysis

Rhombic dodecahedron consideration the interfacial bonding. It should also be noted that ordinary nanoparticles that are thought to be enclosed by Tetraherxahedron { low-index facets could instead be enclosed by high-index fac- } (110) hhl ets; for example, Liz-Marz á n and coworkers recently identifi ed eight (250) lateral facets in Au nanorods [36] . (441) (430) Recent advances in nanoparticle synthesis have allowed hk0 Trisocathedron { (210) the preparation of some types of well-defi ned nanocrystals (221) (431) } hkl with high-index surfaces. In a recent article, Tian et al. [34] (553) Hexoctahedron { } (410) discussed the relationship between surface structure and (432) (421) catalytic functionality of model electrocatalysts and of the electrochemically synthesized, high index-faceted Pt and Pd (111) (100) (433) (211) (411) nanocrystals, including tetrahexahedral nanocrystals with 24 facets of {hk 0} type (Figure 7 ), trapezohedral nanocrystals with 24 facets of { hkk } type, concave hexoctahedral nano- Octahedron Trapezohedron {hkk} Cube crystals with 48 facets of {hkl } type, and multiple twinned Figure 7 Relationship of different types of FCC metal polyhe- nanorods with {hk 0} facets. Chemical reactions tend to pref- drons bounded by low and high index crystallographic facets. From erentially take place at the high-index facets or surface steps, Ref. [35] . as the latter sites are typically much more reactive due to the special atomic arrangement and the low coordination nature. In recent research, there has been a huge thrust on the high < γ { hkl }, where {hkl } represents high indices. However, one index-faceted nanocrystals, with a focus on the evaluation of should keep in mind that nanoparticles are largely due to such nanocrystals as electrocatalysts for small organic fuel kinetic stability, in contrast with the thermodynamic stability molecules, such as methanol, ethanol, formic acid, etc., as of bulk materials. Moreover, the total energy of a nanoparticle these are of great importance to fuel cells. is related to both the metal core and interfacial bonds with Tian et al. reported the fi rst synthesis of tetrahexahedral surface ligands. The interfacial bonding between metal sur- (THH) Pt nanocrystals (Figure 8 ) by an electrochemical treat- face atoms and ligands is of vital importance for the stability ment of Pt nanospheres supported on glassy carbon [37] . The of the entire nanoparticle, but is unfortunately often ignored THH nanocrystals were found to be enclosed by 24 high-in- in practical considerations and theoretical computations. The dex facets, such as (730), (210), and/or (520). All these sur- above order of surface energies apparently does not take into faces have a high density of atomic steps and dangling bonds

A B

∠α _ =133.6° (020)

(200) ∠β=137.6°

50 nm {730} C D 1.0 nm (210)

(210) (730)

(310)

d = 0.20 nm (730)

Figure 8 THH Pt nanocrystals. (A) TEM, (B) SAED, (C) HR-TEM showing the THH Pt nanocrystal is bounded by (730) facets, and (D) atomic model of Pt(730) surfaces. From Ref. [37] . R. Jin: Nanocatalysis 39

and are found to be stable up to 800° C. Interestingly, the THH Zhang et al. recently made concave Au nanocubes enclosed nanocrystals exhibit signifi cantly enhanced (up to 400% ) by 24 surfaces of high-index (720) through a modifi ed seed- catalytic activity (normalized by surface area) for electrooxi- mediated approach involving CTAB surfactant [41] . A related dation of formic acid or ethanol. Very recent work by Sun nanostructure to the THH shape is the elongated tetrahexahe- and coworkers [38] demonstrated that, by decorating THH dral Au nanocrystals with (730) facets (Figure 10 ) [42] , which Pt nanocrystals (∼ 100 nm) with Bi adatoms, the catalytic were demonstrated to be more electrochemically active than activity for formic acid oxidation was signifi cantly enhanced octahedral Au nanocrystals enclosed by eight low-index (111) compared to undecorated nanocrystals, and the incompletely facets. oxidized CO product was inhibited; the latter avoids Pt poi- A high-yield synthetic method has been reported for fabri- soning caused by CO adsorption. They also found that the cating Aucore -Pd shell THH nanocrystals bounded by high-index catalytic activity of Bi-decorated THH Pt nanocrystals are (730) facets [43] . Using regular gold nanocubes ranging from approximately twice that of Bi-decorated Pt nanospheres, 30 to 70 nm, the size of THH Au-Pd core-shell nanocrystals demonstrating the additional shape effect apart from the dop- was tailored from 56 to 124 nm; their electrocatalytic activ- ing effect [38] . ity (e.g., oxidation peak current) for ethanol oxidation was ∼ In contrast to the THH nanocrystals that have protruded found to 2.1 higher than that of concave octahedral Aucore - ∼ (730) facets, Xia and coworkers prepared Pd concave nano- Pdshell nanocrystals, and 1.2 times higher than that of regular cubes with (730) facets (Figure 9 ) through preferential over- octahedral Aucore -Pdshell nanocrystals (Figure 11 ) [43] . Wang growth on Pd cubic seeds enclosed by (100) facets [39] . They et al. reported an epitaxial growth of Pd shells on high index- found that the preferential overgrowth at corners and edges faceted gold THH nanocrystals enclosed by 24 high-index of cubic Pd seeds could be triggered by lowering the concen- (730), and on trisoctahedral (TOH) nanocrystals enclosed by trations of Na 2 PdCl4 and KBr or by increasing the concentra- 24 high-index (221) [44] . The resultant Aucore -Pd shells nano- tion of ascorbic acid. The concave Pd nanocubes exhibited a structures preserve the original high-index facets (Figure 12 ). 1.9 times increase in the peak current in electrooxidation of In the Suzuki coupling reaction, the turnover numbers (per formic acid compared to regular nanocube catalysts. They surface atom) of these core-shell nanostructures were dem- also found that in the Suzuki coupling reaction, the Pd con- onstrated to be 3 – 7 times that of (100)-enclosed Pd or Aucore - cave nanocubes showed 99 % yield in the conversion of iodo- Pdshell nanocubes. benzene to biphenyl, whereas regular nanocubes only gave a A versatile method for delicately controlling the high- 38% yield, and the turnover frequency (TOF) of the Pd con- index facets of polyhedral nanocrystals has been reported cave nanocubes was 3.5 times that of the regular Pd nano- by Lee and coworkers [35] . Using concave TOH Au cubes. Herein, it should be noted that small Pd nanospheres are indeed very active in catalyzing this Suzuki reaction. In related work, Yu et al. investigated Pt concave nanocubes A with ample high-index facets, such as (510), (720), and (830), and demonstrated an enhanced electrocatalytic activity of such nanocrystals (per unit surface area), which was 3.5 times higher than that of the commercial Pt/C catalyst in the ORR process [40] .

Cube THH Elongated THH B A [110] [111] B

[100]

z y

C D

50 nm

Figure 9 Concave Pd nanocubes with (730) facets grown from regular nanocubes. From Ref. [39] . Figure 10 Elongated THH Au nanocrystals. From Ref. [42] . 40 R. Jin: Nanocatalysis

Tetrahexahedra Concave octahedra Octahedra

Figure 11 Core-shell Au-Pd nanocrystals of different morphologies (panels from left to right: THH, concave octahedral, and octahedral shapes). From Ref. [43] .

A B C

D E F

30 nm

2 nm

(310) Au (210) Au Pd + Pd 30 nm

Figure 12 THH Au nanocrystals (A: end view by SEM, B: side view by SEM, C: side view by TEM). Pd-coated THH Au nanocrystals [D: TEM image and electron diffraction (inset), E: HR-TEM and atomic model, F: HAADF-STEM image of a single core-shell nanocrystal and elemental maps]. From Ref. [44] . nanocrystals as seeds, heteroepitaxial growth of Pd lay- THH nanocrystals can be further ﬁ ne tailored, such as (210), ers on the seeds led to Aucore -Pdshell nanocrystals with three (520) or (720) faceted Au@Pd THH nanocrystals (Figure different classes of high-index facets, including concave 13 , right-most panel). The catalytic activities of these high TOH nanocrystals with ( hhl) facets, concave hexoctahe- index-faceted nanocrystals exhibited higher activity for the dron (HOH) nanocrystals with (hkl ) facets, and THH nanostructure-sensitive reaction, e.g., formic acid electrooxida- crystals with ( hk0) facets (Figure 13 ). The Miller indices of tion (Figure 14 ).

{hhl} Au@Pd TOH

{210} {hkl} Au@Pd HOH Au TOH {520}

{hk0} Au@Pd THH {720}

Figure 13 Seed-mediated synthesis of various Au@Pd polyhedral nanocrystals bounded by high-index facets. All geometrical models are oriented in the [110] direction. From Ref. [35] . R. Jin: Nanocatalysis 41

AB 20 20 ) ) 2 2 16 16

12 12

8 8

4 4 Current density (mA/cm Current density (mA/cm 0 0 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Potential (V vs. Ag/AgCl) Potential (V vs. Ag/AgCl)

C D

20 20 ) ) 2 2 16 16

12 12

8 8

4 4 Current density (mA/cm Current density (mA/cm 0 0 -0.20.0 0.2 0.4 0.6 0.8 1.0 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 Potential (V vs. Ag/AgCl) Potential (V vs. Ag/AgCl)

20 THH {210} THH {520} and {310}

) THH {720} and {410} 2 16

4 Current density (mA/cm

0 -0.20.0 0.2 0.4 0.6 0.8 1.0 Potential (V vs. Ag/AgCl)

Figure 14 Cyclic voltammograms of formic acid electrooxidation in 0.1 m HClO4 + 1 m HCOOH on Au@Pd nanocrystals (NCs) with different polyhedral shapes enclosed by different crystallographic facets. (A) Cubic Au@Pd NCs with (100) facets. (B) Octahedral Au@Pd NCs with (111) facets. (C) TOH Au@Pd NCs with (552) facets. (D) HOH Au@Pd NCs with (432) facets. (E) THH Au@Pd NCs with (hk 0) facets of different Miller indices. Scan rate: 10 mV/s. From Ref. [35] .

Liu and Pippel prepared quaternary PtCuCoNi nanotubes wet chemistry route (Figure 15 ). Excellent activity (per unit of low Pt content using a template-assisted one-step elec- surface area) in the electrocatalytic oxidation of both formic trodeposition method [46] . Interestingly, such nanostructures acid and ethanol was observed (i.e., 2.3 and 5.6 times greater exhibited markedly enhanced ORR activity. The authors than those of commercial Pt black and Pt/C, respectively). attributed the enhancement activity to several favorable fac- The enhanced activity of the Pt nanocrystals was attributed to tors, including the synergistic effect of multielements, the the high density of atomic steps present on high-index (411) strain and electronic effects associated with surface dealloy- exposed facets, although (100) facets also exist in the concave ing, and the hollow/porous geometry. Pt nanocrystals apart from high-index (411) facets. Zheng and coworkers [47] recently demonstrated amines A quantitative measurement of the ∼ 2 nm Pt nanoparticles as the surface controller and successfully prepared concave was attained by Shao-Horn and coworkers [50] . They found Pt octapod nanocrystals with (411) high-index facets via a that the intrinsic activity (normalized by Pt surface area) for 42 R. Jin: Nanocatalysis

ACB

-- - 53 022 002 53.1 - 220

40 nm [100] DFE

- 202 - 60 220 60.0

- 002

40 nm [111] GIH

- 002 109 45 -- 111 109.5 45.3 - 111

40 nm [110] JK

(411)

(111) (100) 0.20 nm 2 nm 0.23 nm

Figure 15 (A, D, G) TEM images, (B, E, H) SAED patterns, and (C, F, I) geometric models of individual concave Pt nanocrystals oriented along the (A, C) [100] , (D, F) [111] , and (G, I) [110] directions. (J) HR-TEM image of the region indicated by the box in (G). (K) Atomic model corresponding to the region indicated by the box in (J). From Ref. [47] .

CO and methanol electrooxidation processes increases lin- certain reactions but also promote mechanistic understand- early with the amounts of surface steps, e.g., n (111) × (111) ing of metal catalysis, especially the shape dependence. For type. Goodman and coworkers probed the terrace sites on Pt electrocatalysis, increasing surface steps on high-index nano- nanoparticles as well as step sites on such particles using CO particles is a promising strategy for fi nding highly active, as probe molecules, and temperature desorption spectroscopy new electrocatalysts for fuel cell applications. In terms of the (TDS) analysis revealed that the percentage of terrace sites shape- or facet-dependence in catalysis, it is meaningful to decreases by ∼ 50 % when the Pt particle size was reduced relate the results from nanocrystals to the conclusions from from 4.2 to 2.5 nm [51] . surface science single crystal studies [54, 55] . Future work on There are several other nanocrystals with high-index facets, the shape-controlled nanocatalysts (including both low and but their catalytic activities have not been evaluated, such as high index-faceted nanocrystals) will ultimately lead to use- high-index (210) faceted THH gold nanocrystals [52] , which ful working principles for new catalyst design. For the line of were synthesized by reducing HAuCl4 in the mixture solution of basis science research, future work will reveal deep insight N, N -dimethylformamide and PVP polymer at 80° C. Concave into the active sites, e.g., whether those atoms on the planes TOH gold nanocrystals (55– 120 nm) bounded by high-index (i.e., terraces) or at edges, kinks, and steps, which differ in facets, such as (221), (331), and/or (441), have also been syn- their coordination environments, are true active centers in cat- thesized by a seed-mediated growth method using cetyltrim- alytic reactions and, if so, how they are involved in activating ethylammonium chloride (CTAC) as the capping agent [53] . reactants and subsequent surface reactions. The authors suggested that a proper reduction rate of the metal ions in the growth solution and the preferential adsorption of 3.3. Core-shell, hollow, and multibranched + CTA on high-index facets led to the formation of the TOH nanoparticles for catalysis nanocrystals in the kinetically driven process [53]. Overall, well-defi ned nanocrystals with specifi c shapes In addition to polyhedral nanocrystals, other types of nano- will not only provide highly active and selective catalysts for structures that are of interest to catalysis include core-shell R. Jin: Nanocatalysis 43

nanoparticles, hollow nanostructures, nanoporous metals, and nanodendrites, etc. Such nanoparticles may be composed of a single, two or multiple metal elements. Core-shell nanopar- Pd Pt ticles with two or more metal components are typically made by sequential reduction of metal salt precursors in solution Au phase [56] . Hollow nanoparticles are often made by chemical etching methods; for example, using Ag nanoparticles as a sacriﬁ cial template in the etching reaction by Au(III) ions, Fe a galvanic replacement reaction (or transmetalation reaction) occurs and gives rise to porous gold nanostructures [57] .

02 468nm 3.3.1. Core-shell nanostructures The core-shell strategy is very useful in preparing catalysts, especially Pt-based electrocatalysts. Pt is the best catalyst for both cathode and anode reactions, but the high price of Pt has Au coating FePt coating long been a major concern in fuel cell applications. Thus, Pd fabricating Pt-coated nanoparticles of less expensive metals Pd/Au Pd/Au/FePt or Pt-based alloy nanoparticles constitutes a promising route to reduce the Pt content in electrocatalysts [58, 59] .

Sasaki et al. [60 ] prepared Pdcore Ptshell nanoparticle catalysts (denoted as Pd@Pt) at ultralow platinum content, in which Figure 16 HR-TEM image of a Pd@Au@FePt dual-shelled nanopar- Pd is much less expensive than Pt. Interestingly, the low Pt ticle with elemental mapping (shown by the proﬁ les). From Ref. [62] . content did not deteriorate much the ORR activity of the core-shell particles, and in fuel cell tests, no loss of Pt was thermal stability and thus were well suited to perform cata- observed even after 200,000 potential cycles. The improved lytic reactions at high temperatures (e.g., > 400 ° C) [64] . Pt stability was found to be imparted by the Pd core, which Related to the core-shell strategy, mesoporous multicom- was gradually lost to protect the Pt shell [ 60 ]. Shao et al. ponent nanocomposite colloidal spheres (Figure 17 ) were demonstrated that Au@Pt with one monolayer of Pt (∼ 3 reported by Li and coworkers [65], such as Ag(10 nm)-CeO2 , nm total diameter) can achieve a 1.6-fold increase in ORR Ag(10 nm)-TiO 2 -CeO2 , Au(5 nm)-CeO 2 , Au(3 nm)-CeO2 , activity compared to Pd@Pt nanoparticles of comparable Pd(3 nm)-CeO2 , Pt(6 nm)-CeO2 , Rh(3 nm)-CeO2 , Ru(3 nm)- size [61 ]. The activity of Au@Pt catalyst indeed surpasses CeO2 , and Pd(3 nm)-TiO2 . These caged nanoparticles were that of a state-of-the-art 2.8-nm Pt/C catalyst. Density demonstrated as high-temperature stable catalysts, such as in functional theory calculations imply that the signiﬁ cant strain CO oxidation and cyclohexene hydroconversion [65] . in the surface of the core nanoparticle plays a critical role Apart from inorganic shell materials, organic (e.g., polymer) in activity enhancement [61 ]. Sun and coworkers fabricated shells have also been utilized. Carregal-Romero et al. [66] core-shell nanoparticles composed of a 5-nm Pd core with encapsulated gold nanoparticles in a thermoresponsive micro- a controlled Au shell of 1, 1.5, and 2 nm, respectively [62] . gel (pNIPAM) and used such core-shell particles as catalysts In the catalytic evaluation for oxygen reduction in alkaline in the electron transfer reaction between hexacyanoferrate(III) fuel cell conditions (e.g., 0.5 m KOH), the Pd@Au activity and borohydride ions. The thermoresponsive polymer shell was found to be dependent on the Au shell thickness, with was demonstrated to act as a “ nanogate ” , which can be opened gold shells thinner than 1.5 nm being more active than those or closed to a certain extent (Figure 18 ) and thus control the with 2 nm shell and pure Au nanoparticles. Multi-metallic diffusion of reactants towards the catalytic Au core. Yuan Pd@Au@FePt core-shell nanoparticles (11 nm) have also et al. prepared Au@/polymer nanostructures and demonstrated been produced (Figure 16 ). Metal@oxide core-shell nanoparticles have also been reported. Yin et al. prepared Au@Fe2 O3 core-shell nanopar- A Au (3 nm)-CeO2 B ticle catalysts (deposited on SiO2 support) and found that the catalyst was highly active for low-temperature CO oxidation, even higher than that of Au/SiO2 or Au/Fe 2 O3 which were prepared by colloidal deposition [63] . Apart from the often observed synergistic effect in core-shell nanoparticles, another major advantage is to introduce an oxide shell to protect the metal particle from sintering under harsh reaction conditions, e.g., at high temperatures (for accelerating the reaction) and at high pressure (for increasing the yield). 50 nm 20 nm The durability of catalysts under harsh conditions is always a major issue. Metal nanoparticles in the form of metal core (Pt, Figure 17 TEM images of CeO 2 -caged Au nanoparticle catalysts. Co)@inorganic shell (SiO2 ) were found to exhibit exceptional From Ref. [65] . 44 R. Jin: Nanocatalysis

support and thus one may investigate the intrinsic catalytic properties of Au core. Ma and Dai recently summarized research progress in the design of novel structured nanogold catalysts such as core-shell and yolk-shell nanocatalysts ∆ T [71] . Au Au The ordered porous metal materials are also promising for catalysis. Lu and Eychm ü ller [72] discussed the colloidal crystal template technique, which is rather general for fabricating ordered porous metal structures. Ordered bi-metallic nanostructures with hierarchical porosity can be fabricated and utilized for catalysis. Such hollow nanostructures have some unique advantages, such as the uniform macropores of a Figure 18 Au@pNIPAM core-shell nanoparticle consisting of a few hundred nanometers (i.e., high mass transport efﬁ ciency) gold core and thermoresponsive pNIPAM polymer shell. The shell and a very high surface area due to hierarchical macropo- can reversibly swell or collapse below or above its lower critical solu- rous and mesoporous structures, and thus may exhibit some tion temperature (LCST) when dispersed in water. From Ref. [66] . unusual properties.

3.3.3. Dendritic nanostructures for catalysis Another selective catalytic hydrogenation of nitrobenzene (as opposed interesting type of nanostructure is the nanodendrites, also to nitrophenol) [67] . A key to that is the porous polymer shell called multipod, multiarm, multibranch, or nanofl ower was made to be hydrophobic, which selected hydrophobic structures. Compared to those shaped-controlled nanocrystals, molecules (e.g., nitrobenzene) to pass through freely, but bar- nanodendrites are not well defi ned in terms of the exposed ricades hydrophilic molecules (e.g., nitrophenol). Upon reach- surfaces as the exposed facets tend to be a mixture of various ing the Au core, the nitro-compound can be hydrogenated. surfaces. Xia and coworkers synthesized Pd-Pt nanodendrites by 3.3.2. Hollow and porous nanostructures Apart from reducing K2 PtCl4 with l-ascorbic acid in the presence of trun- solid nanoparticles of metals, hollow or porous nanoparticles cated octahedral Pd seeds in aqueous solution (Figure 19 A) have also been the focus of considerable interest for catalytic [73] . Such heterogeneous structures were found to be very applications. Kim et al. demonstrated that hollow Pd active in the ORR process under acidic conditions, which nanoparticles showed excellent catalytic activity in Suzuki provides a promising approach for solving the otherwise reactions, and that such nanocatalysts could also be recycled slow kinetics of the ORR process at the cathode. Yang and without the loss of the catalytic activity [68] . In catalytic coworkers prepared Pt-on-Pd heterogeneous nanostructures photodegradation of methyl orange, hollow Au nanocages were [75] , which showed high activity and durability superior to found to be more effi cient than semiconductor nanomaterials the commercial Pt/C catalysts. In a recent review on nano-

(e.g., TiO2 and ZnO) [69] . The Au nanocages were made by dendrites [76] , Lim and Xia summarized several strategies Au(III) etching of Ag nanocubes and thus had remaining Ag for the synthesis of branched metal nanocrystals, such as the on their interior walls which could be oxidized to Ag2 O and kinetically controlled overgrowth, aggregation-based growth, acted as a photocatalyst [69] . A yolk-shell type of catalyst, heterogeneous seeded growth, selective etching, and tem- e.g., Au nanoparticles located inside a porous shell of SiO 2 , plate-directed methods. was prepared by Wang et al. [70] . A unique aspect of the yolk- A universal method for the synthesis of nanodendrites of shell catalyst lies in that Au has a minimum effect from the precious metals (Pd, Pt, and Au) has been reported (Figure

ABC

10 nm

20 nm

50 nm 20 nm

Figure 19 Nanodendrites of (A) Pd@Pt, (B) Pd, and (C) Pt. Image (A) from Ref. [73] ; images (B) and (C) from Ref. [74] . R. Jin: Nanocatalysis 45

19 B, C) [74] , which utilized an amino acid-based surfactant, much higher electrocatalytic activity towards methanol oxi- sodium N -(4-n-dodecyloxybenzoyl)-l -isoleucinate (SDBIL). dation reaction than the conventional catalysts of platinum Such nanostructures showed abundances of (220) and (311) black and commercial E-TEK Pt/C. Yeo et al. [82 ] prepared facets compared to typical nanoparticles. The Pt nanofl ow- another type of Pt nanodendrites attained by Au seed-medi- ers were demonstrated to be capable of catalyzing the Suzuki ated growth inside hollow silica nanoshells and demonstrated coupling reaction of phenyl boronic acid and iodobenzene in that such Pt nanocatalysts had substantially higher activity water with very high yield (> 99 % ) and excellent recyclabil- than commercial Pt black in the ORR process. There are also ity; this is indeed the fi rst demonstration of complete Suzuki many other reports on the synthesis of highly branched nano- coupling with Pt nanocatalysts; note that the Pt spherical structures. Lee and Park [83 ] devised a facile one-pot method counterparts only gave 7 % yield for this reaction. The Pd for synthesizing multibranch gold nanoparticles by reducing nanofl owers were prepared and utilized to catalyze the Heck Au(III) with hydroxyphenol in a biphasic, kinetically con- coupling reaction of styrene and iodobenzene in water at trolled process. The as-prepared, highly branched gold nano- 95 % yield, much higher than the yield (25 % ) for spherical Pd particles possess high-index facets. Different hydroxyphenol nanoparticle catalysts under comparable experimental condi- derivatives were found to produce various nanostructures, but tions [74] . no catalytic evaluation has been done. A multi-armed star-like Pt nanowire catalyst was synthesized by reduction of a Pt precursor with formic acid [77] . 3.3.4. Plasmon enhancement in composite The ORR activity (on a mass basis) was demonstrated to be nanocatalysts Apart from the intrinsic catalytic role 1.5-times that of the commercial Pt/C (E-TEK) catalyst. In of precious metal nanoparticles (such as Au and Ag), the terms of surface area-based specifi c activity, the new catalyst plasmonic properties of metal nanoparticles can also be utilized exhibited an activity nearly 3 times that for E-TEK Pt/C cata- for catalysis. In a recent study, Mori et al. anchored a Ru(II) lyst. An accelerated durability test of the catalysts by potential complex catalyst onto SiO2 -coated plasmonic Ag nanoparticles cycling between 0.6 and 1.2 V (vs. RHE) in O 2-purged, 0.5 m [84] . An enhancement in the photoinduced oxidation activity

H2 SO4 solution at room temperature, showed much improved of the Ru(II) catalyst for selective oxidation of styrene has stability: after 4000 cycles the star-like Pt/C catalyst lost 40% been observed, which is attributed to the Ag surface plasmon of the initial activity, compared to the loss of 67.5% for the properties. Metal nanoparticles are also capable of promoting the E-TEK Pt/C catalyst and a larger decrease in the half-wave catalytic properties of semiconductors. For example, in a recent potential for the latter catalyst. An interesting fi nding was that study, Yin and coworkers designed a TiO2 -based, sandwich- unsupported Pt nanostars showed even higher durability (only structured photocatalysts decorated with Au nanoparticles and 13 % loss of activity after 4000 cycles) [77] . found that Au played an important role in obtaining the high Citrate-capped, dendritic Pt nanoparticles of ∼ 20 nm catalytic effi ciency in decomposition of organic compounds size have also been prepared and loaded onto carbon nano- under light [85] . The Au nanoparticles were rationalized to tubes (CNTs) via electrostatic interactions [78] . Their cata- enhance light harvesting and charge separation as well as lytic activity for the reduction of hexacyanoferrate(III) by boosting the role of implanted nitrogen dopants in TiO2 [85] . sodium borohydride was studied, and the dendritic Pt/CNT In summary, the various types of nanostructures discussed heterostructures gave rise to a much lower activation energy in Section 3.3 remain to be further explored in future research. ∼ ( E a 10.3 kJ/mol vs. 27.0 kJ/mol for spherical Pt nanopar- It is worth developing some versatile synthetic methods to ticles). The exceptional catalytic behavior was attributed to prepare core-shell nanostructures with delicately controlled the presence of many corners and edges in the particles with distribution of metal elements in the core and/or the shell at a dendritic morphology. Such hybrid heterostructures also controlled ratios. The synergistic effect in metal-semiconduc- have potential application in proton exchange membrane fuel tor hybrids [86] deserves to be systematically studied. Overall, cells [79] . the various nanostructures of core-shell, hollow, multipod, and Using Au nanocubes, nanorods, and nanooctahedra as porous morphology are expected to open up some opportu- seeds, Han and coworkers deposited dendritic Pt onto those nities in the future development of multicomponent nanopar- seeds via reduction of K2 PtCl 4 by ascorbic acid in the pres- ticles with improved activities for catalytic applications. ence of CTAB surfactant in solution [80] . Such Au-Pt heterostructures exhibited higher mass and area-specifi c activities 3.4. Bi- and multi-metallic nanoparticles for catalysis for ORR than do monometal Pt nanodendrites. Interestingly, the core geometry was found to signifi cantly affect the ORR Tuning the composition of metal catalysts is one of the effec- activity: the Au-Pt nanostructure with octahedral Au core tive ways of tuning the catalytic activity of metal nanopar- showed the highest enhancement of the catalytic activity. This ticles. The miscibility of metals at the nanoscale can be much is attributed to the reduced oxygen binding energy on Au-Pt wider than in bulk alloys. Bi-metallic nanoparticles have long compared to that on the bulk Pt(111) surface [80] . Wang and been recognized for the synergistic effects on catalytic perfor- coworker prepared Pt-on-Pd bi-metallic nanodendrites sup- mance [87] . In conventional preparation of bi-metallic cata- ported on graphene nanosheets [81] , in which small Pt nano- lysts, wet impregnation of supports with a mixed solution of branches were grown on Pd nanocrystals, forming porous different metal salts works well. Recent development in solu- structures. This type of metal/graphene heterostructure pos- tion phase nanochemistry has permitted precise control of the sessed a high electrochemical active area and thus exhibited composition and distribution of the two metallic components 46 R. Jin: Nanocatalysis

in the nanoparticle, such as alloyed, core-shell, and dumbbell surface enrichment and, hence, higher mass activity, whereas nanostructures. the higher temperature-treated catalyst showed a reduced sur- Kang and Murray [88] prepared cubic Pt-Mn alloy nano- face concentration of Pt (i.e., enhanced Pt-alloying surface crystals and found that the ORR activity (mass-speciﬁ c) in sites), and hence higher speciﬁ c activity for electrocatalysis ∼ H2 SO4 was much higher ( 3 times) than that of the commer- [93] . Yu and coworkers reported PdAuCu ternary nanotube cial Pt catalysts (e.g., commercial Pt/C and Pt black catalysts). catalysts for ORR and H2 O 2 reduction reactions [94] . They In addition, the Pt-Mn nanocubes also showed a higher ORR found that the incorporation of Au into a PdCu system can activity than Pt-Mn nanospheres. By incorporating Co into lower the oxidation potential of Cu and favor the formation of Pt nanocubes, Yang et al. demonstrated enhanced catalytic active Cu(II) species, giving rise to a synergistic effect in the activity of Pt3 Co nanocubes compared to plain Pt nanocubes catalytic activity of the nanotubes [94] . in electrocatalytic methanol oxidation [89] . The enhancement Dumbbell-like nanoparticles of Pt-Fe3 O4 (Figure 21 ) were was attributed to the slower and weaker adsorption of adverse also synthesized by Wang et al. through epitaxial growth of

CO onto Pt 3Co (vs. plain Pt). Lee et al. reported a simple one- Fe onto Pt nanoparticles, followed by Fe oxidation [95] . The pot aqueous synthetic method for preparing rhombic dodeca- nanoparticle size in the structure is tunable from 2 to 8 nm for hedral Au-Pd alloy nanocrystals enclosed exclusively by 12 Pt and 6 to 20 nm for Fe3 O 4. Pt nanoparticles in the Pt-Fe3 O4 (110) facets (Figure 20 ), which exhibited higher electrocata- structure show a 20-fold increase in mass activity towards lytic activities than (111)-faceted nanoparticles [90] . Wong oxygen reduction reaction compared with the single compo- and coworkers reported high activity of Pd-on-Au bi-metallic nent Pt nanoparticles and the commercial 3-nm Pt particles. nanocatalysts (with only submonolayer Pd coverages) com- This work proves that it is possible to maximize catalytic pared to monometallic Pd catalysts for hydrodechlorination of activity of the Pt nanoparticle catalyst through the control not trichloroethene [91] , with catalytic activity being two orders only of Pt size and shape (as discussed in the above sections) of magnitude greater than Pd black (on a Pd atom basis). In but also of its interaction with Fe3 O 4 nanoparticles. this system, apart from the synergistic effect, gold was also Recently, Zhong and coworkers summarized the phase found to improve oxidation of the catalyst and deactivation properties of multi-metallic nanoparticle catalysts. Their sys- resistance to chloride and sulﬁ de species, a common problem tematic work has demonstrated that the bi-metallic nanopar- for Pd-based catalysis [91] . ticles may possess phases ranging from alloy, partial alloy, Multi-metallic nanoparticles can also be made. Zhong and to phase segregation (Figure 22 ) depending on the prepara- coworkers [92] investigated PtNiCo catalysts for oxygen tion, the bi-metallic composition, and the supporting materi- reduction reaction to understand the effects of lattice strain als [87] . Overall, for bi- or multi-metallic catalysts, detailed and surface properties of the carbon-supported tri-metallic analyses of atomic distributions, their oxidation states, inter- catalyst on its activity and stability. The increase in the ther- atomic distances, the segregation of metal component(s), and mal treatment temperature was found to shrink the lattice con- the formation of surface metal oxide or active oxygen species stants of the alloys and increased faceting of the catalysts. are of particular importance. Such information will provide Such effects were correlated to the enhancement of electro- deep insight into the synergistic effect between the metal catalytic activity [92] . Recent work on PtNiFe/C catalysts components and, ultimately, one can attain precise correla- revealed that catalysts treated at lower temperatures led to Pt tion of metal doping with the catalytic properties.

{110} 200 5 nm Pt

5-12 nm Pt-Fe3O4 5-17 nm Pt-Fe O 0 3 4

-200

-400 Mass activity (mA/mg Pt) 20 nm -600

200 nm -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 Potential (V vs. Hg/HgO)

Figure 20 Rhombic dodecahedral Au-Pd alloy nanocrystals Figure 21 Dumbbell Pt-Fe3 O4 nanoparticles and enhanced ORR enclosed by 12 (110) facets. From Ref. [90] . activity. From Ref. [95] . R. Jin: Nanocatalysis 47

Support Support Support Support Support Support

Alloyed Partially alloyed or phased segregated Completely phased segregated

Figure 22 Several scenarios of atomic distribution in bi-metallic nanoparticles. From Ref. [87] .

3.5. Size control of metal nanoparticles and effects obtained in solution phase. For example, uniform Au nano- on catalysis particles with size ranging from ∼ 1 nm to ∼ 100 nm [100–103] can now be routinely made; for example, Eah and coworkers The high surface-to-volume ratio of nanocatalysts is one of reported uniform Au nanoparticles ranging from ∼ 3 to ∼ 8 nm the primary factors in catalysis, which largely enhances atom (Figure 23 ) [101, 102] . The available uniform nanoparticles effi ciency and reduces the cost of precious metal catalysts. will allow one to carry out more precise measurements of size The size dependence has long been an outstanding issue in dependence in catalytic tests in future work. catalysis [96 – 98] . However, in previous studies the unavail- Joo et al. [104] synthesized uniform Ru nanoparticles rang- ability of well-defi ned nanocatalysts with suffi ciently narrow ing from 2.1 to 6.0 nm by a polyol reduction of Ru(acac)3 size distribution makes it diffi cult to carry out in-depth studies precursor in the presence of a PVP stabilizer and further inves- on the particle size dependence, and some ambiguity appar- tigated their catalytic activity for CO oxidation. The CO oxida- ently exists. In surface science studies, although narrow size tion activity was interestingly found to increase with particle distributed metal clusters (even nanoparticles of a few nm in size (as opposed to the common observation – the smaller the size) from gas phase beams can be mass selected and deposited better – in many reactions), with the 6.0-nm Ru nanoparticle onto supports, but the landing of clusters may cause aggrega- catalyst showing an 8-fold higher activity than the 2.1-nm cata- tion of such clusters as they are bare, resulting in a broadened lysts. In related work by Grass et al. [105] , Rh nanoparticles size distribution. Similarly, the solvated metal atom disper- ranging from 1.9 to 11.3 nm were supported on mesoporous sion method [99] may also lead to a size distribution dur- SBA-15 in hydrothermal synthesis and such Rh/SBA-15 cata- ing the trapping of vaporized metal atoms in frozen organic lysts gave a result for CO oxidation opposite to the Ru nano- matrices and subsequent loading onto supports. In contrast, particles. With increasing size from 1.9 to 11.3 nm, the turnover ligand-protected nanoparticles of extremely narrow size dis- frequency decreases from 1.7 to 0.4 s. For Pt nanoparticle- tribution (e.g., 5 % standard deviation) can now be readily catalyzed pyrrole hydrogenation, Kuhn et al. [106] compared

ABC

20 nm 20 nm 20 nm DEF

20 nm 20 nm 20 nm

Figure 23 SEM images of 3 – 8 nm (nearly) monodisperse gold nanoparticles protected by 1-dodecanethiol (DDT). Particle diameters: (A): 3.2, (B): 4.0, (C): 5.2, (D): 6.3, (E): 7.4, and (F): 8.2 nm. (A – C): Size control in water before ligand coating [ 101 ], and (D – F): size control in boiling DDT after ligand coating [ 102 ]. Images are from Refs. [ 101 ] and [ 102 ], courtesy of Prof. Sang-Kee Eah. 48 R. Jin: Nanocatalysis

the size dependence of dendrimer/polymer capped Pt nanopar- 3.6. Atomically precise nanoclusters for catalysis ticles between 0.8 and 5.0 nm diameter, which were supported onto SBA-15 silica, and found that the ring opening reaction Another remarkable advance in solution phase synthesis of occurred more easily over larger particles than smaller ones. nanoparticles lies in the successful preparation of atomically These examples illustrate the intriguing size dependence in precise nanoclusters. Current efforts focus on gold [111, 112], various reactions. Further efforts in understanding the nature of but it is feasible to extend to other precious metals. These such dependences are expected in future work. well-deﬁ ned nanoparticles provide new opportunities for Finke and coworkers [107] have carried out a series of work achieving fundamental understanding of metal nanocatalysis. on the synthesis of uniform, ultra-small Ir(0) nanoparticles Recent research has produced several atomically precise γ thiolate-protected Au (SR) nanoclusters of extraordinary sta- supported on -Al 2 O3 by using oxide supported-organome- n m tallic compound (as precatalysts). The as-prepared catalysts, bility, such as Au25 (SR)18 , Au38 (SR)24 , Au144 (SR)60 , and other sizes (Figure 25 ) [113] . These protected nanoclusters are well e.g., Ir(0)∼ 900/Al2 O3 , were found to be highly active (2.2- to deﬁ ned on the ultimate atomic scale, as each of them is com- 4.8-fold higher than two literature Ir(0) n /Al2 O3 catalysts) and have good lifetime (≥ 220,000 total turnovers) in cyclohexene posed of a precise number of metal atoms (n ) and of ligands hydrogenation. (m ), with n ranging from approximately tens to hundreds of Dendrimers have been exploited for preparing ultra-small atoms (equivalent diameter ranging from subnanometer to ∼ 2 metal nanoparticles [108, 109]. Dendrimers serve as a support nm). Such nanoclusters are vastly different from their larger (or matrix) to host ultra-small nanoparticles (e.g., < 2 nm diam- counterparts – metallic or plasmonic Au nanocrystals (typi- eter) and prevent the nanoparticles from aggregation (Figure cally >2 nm) in terms of atomic packing structure and physi-

24 ). By contrast, the hyperbranched structure of dendrimers cochemical properties [111, 112]. Relatively small Au n (SR) m allows reactant molecules to readily access the nanoparticles nanoclusters (e.g., n< 100) exhibit strong quantum confi ne- [108] . Dendrimer-supported Pt and Rh nanoparticles have ment effects, manifested by the discrete electron energy states been utilized for catalytic cyclization and hydroformylation and signifi cant band gaps, e.g., 1.3 eV for Au 25 (SR)18 [114] reactions. Nakamula et al. utilized the fourth-generation (G4) and 0.9 eV for Au 38 (SR)24 [115] nanoclusters, as opposed to phenylazomethine dendrimers (abbreviated as TPP-DPA G4) the quasicontinuous conduction band in plasmonic nanocrys- < < for synthesizing ultra-small Rh and bi-metallic RhFe nano- tals. Larger Au n (SR) m nanoclusters with 100 n 200 exhibit particles with a very narrow size distribution [48] . They found intermediate properties between molecular-like semiconduct- that, compared to Rh nanoparticle catalysts in dendrimer ing behavior and metallic properties [116] . cages, the bi-metallic Rh/Fe nanoparticles showed signifi - The non-metallic behavior of Au n (SR) m nanoclusters is of cantly higher catalytic ability in the respective hydrogenation interest to catalysis [45, 49, 117, 118]. These atomically pre- reaction of olefi ns and of nitroarenes under mild conditions. cise Au n (SR) m nanoclusters are expected to become a promis- The TOF (calculated by the conversion in 30 min of reac- ing class of model catalysts. Unlike conventional nanogold tion) for the bi-metallic catalyst was up to 8.5-times higher catalysts [119] having a major issue of polydispersity, atomi- than that of the Rh nanoparticle catalyst. The high activity of cally monodisperse nanoclusters provide a well-defi ned sys- Rh/Fe catalyst was attributed to the Rh-Fe synergistic elec- tem; thus, the observed catalytic properties refl ect the intrinsic tronic effect [48] . properties of the nanocluster catalysts, which is particularly For ultra-small metal nanocatalysts, some subnanometer important for obtaining insight into the size dependence of cluster species have been reported to be particularly active nanogold catalysis. Furthermore, correlation of the crystal in certain reactions, but the presence of mixed sizes some- structures of Au n (SR) m nanoclusters and their catalytic prop- times obscures the conclusion. Liu et al. [110] prepared erties permits one to achieve a fundamental understanding of

FeOx -supported gold catalysts through a colloidal deposition the catalytic mechanism and will ultimately contribute to the method and such catalysts were found to be highly active for new design of highly selective catalysts for speciﬁ c chemical CO oxidation. Their high-angle annular dark-ﬁ eld scanning processes. transmission electron microscopy (HAADF-STEM) analyses Herein, we illustrate the catalytic properties of protected showed that gold nanocluster size for this catalyst was larger Au n nanoclusters with spherical and rod-like Au25 for the than 1 nm (diameter). They pointed out that bilayer structures selective hydrogenation process [117, 120] . Bulk solution ∼ and/or diameters of 0.5 nm are not mandatory in order to phase syntheses of ligand-protected Au25 nanospheres [114] achieve the high activity. and Au25 nanorods have been achieved recently [121, 122] .

FeCl3 RhCl3 Reduction

TPP-DPA G4 Rh/Fe Bimetallic nanoparticle

Figure 24 Preparation of Rh-Fe nanoclusters in the templates of TPP-DPA G4. From Ref. [48] . R. Jin: Nanocatalysis 49

A B

- [Au25(SC2H4Ph)18]

4000 5000 6000 7000 8000 9000 10,000 400 500 600 700 800 9001000 1100

+ [Au38(SC2H4Ph)24-Cs]

6000 8000 10,000 12,000 14,000 400 600 800 1000 1200 1400 1600

E F

2+ [Au144(SC2H4Ph)60-Csx]

X=0,1,2,3,4

14,000 16,000 18,000 20,000 22,000 24,000 400 600 800 1000 1200 1400 1600 Mass (m/z) Wavelength (nm)

Figure 25 Electrospray mass spectrometry (ESI-MS) and UV-vis absorption spectroscopy characterization of Au25 (SR)18 , Au38 (SR) 24 , and = Au 144 (SR)60 nanoclusters. R CH2 CH2 Ph. From Ref. [112] .

α β The true monodispersity of these Au n nanoclusters allows for (or aldehydes), and a 100 % selectivity for , -unsaturated the determination of their atomic structure by singe crystal alcohol products was obtained at ∼ 20 % conversion [117] . X-ray crystallography. Interestingly, these structures do not The nanoclusters were conﬁ rmed to be intact (i.e., no resemble the FCC structure of their larger counterparts – Au fragmentation) after the catalytic reaction. With oxide- nanocrystals. In the case of Au25 nanospheres, X-ray crystal- supported nanoclusters, the activity of Au25 was largely lographic analysis shows that the Au25 sphere features a cen- enhanced (approx. twice that of unsupported particles) tered icosahedral Au13 core (Figure 26 A, in magenta), which is further capped by a second shell comprising the 12 remaining A B Au atoms (in green); the latter form six pairs and are situated S P ± ± ± S around the x -, y -, and z -axes (C 2 axes of the icosahedron), Au respectively [114] . One can ﬁ nd that the 12 exterior Au atoms Au Cl do not form a closed shell on the Au13 icosahedron, for that S an icosahedron has 20 triangular faces (Au3 ), but only 12 of them are face capped, that is, leaving eight Au3 triangular faces uncapped. These holes or volcano-like sites may act as catalytic active centers. In contrast, the Au25 rod was found to adopt a vertex-sharing bi-icosahedral structure (Figure 26 B). Figure 26 Atom packing structures of 25 gold atom nanoclus- This structure is a closed one and no volcano sites were found = ters: (A) icosahedral, two shelled Au25 (SR) 18 (R CH2 CH2 Ph, omit- [121] . ted for clarity); (B) bi-icosahedral Au (PPh ) (SR) Cl (the Ph and Both the spherical and rod-shaped Au nanoclusters 25 3 10 5 2 25 CH2 CH2 Ph groups are omitted for clarity). The core atoms of gold were found to be capable of catalyzing chemoselective are shown in magenta. Surface gold atoms and ligand atoms are as hydrogenation of the C =O bond in α ,β -unsaturated ketones labeled. From Ref. [120] . 50 R. Jin: Nanocatalysis

[117] . Interestingly, Au25 spheres were found to have much electronic properties of metal nanoclusters and correlation higher activity than the Au25 rod (34 % vs. 9 % yield of unsat- with their catalytic performance. Third, such nanoclusters urated alcohol, on CeO2 support) [120] . The high activity possess a unique core-shell structure – an electron-rich core of the Au25 nanosphere was ascribed to its non-closed sur- with delocalized valence electrons and an electron-deﬁ cient face structure and unique core-shell electronic structure. shell, such as the case of Au25 nanoclusters [114] . The pres-

Compared to Au25 nanorods, the unique volcano-like sites ence of both electron-rich and electron-defi cient Au atoms at = in Au25 nanospheres favor adsorption of the C O group of the cluster surface creates a unique environment for simul- α ,β -unsaturated ketone (Figure 27 ), and electron transfer taneous activation of multiple types of reactant molecules. between the electron-rich Au13 core and the O atom should Last but not least, the structural determination of Au n (SR) m activate the C= O bond. Subsequently, the weakly nucleo- nanoclusters permitted atomic level insight into the nature of philic hydrogen could attack the activated C= O group, lead- nanocatalysis and the structure-property relationship. On the ing to the unsaturated alcohol product. These two types of basis of the structure of Au n nanoclusters, future theoretical Au 25 nanoparticles clearly demonstrate the importance of work will reveal more information about the molecular acti- surface atom arrangement on catalysis. vation and reaction mechanisms. The ligand-protected nanoclusters also provide an oppor- tunity for exerting infl uence by surface ligands on catalytic reaction. Diastereoselective catalytic capability of Au25 (SR)18 4. Future opportunities of catalysis by metal nanocluster catalyst has been demonstrated [123] . The -R nanoparticles group (phenylethyl vs. long-chain alkyl) was found to infl u- ence the selectivity of products. This scenario is somewhat Section 3 has briefl y discussed some remarkable progress like organometallic catalysis and should be further exploited in nanocatalysis by precious metal nanoparticles. Those in future research. well-defi ned colloidal nanoparticles, such as polyhedral q = nanocrystals with low- or high-index facets, the bi- or multi- The charge states of [Au25 (SR)18 ] ( q -1, 0) do not affect the atomic structure of the Au25 spherical clusters [124] , which is metallic nanoparticles, the atomically precise nanoclusters, important for understanding the catalytic mechanism of Au25 as well as other types of nanostructures, have created many in selective oxidation reactions. The Au 25 nanospheres were exciting new opportunities for catalytic research. In future again found to be more active than the Au25 rods [120] . In research, the well-defi ned nanocatalysts are expected to = other work, Liu et al. investigated thiolate-capped Au n ( n 10, mediate the knowledge gap between single crystal model 18, 25, 39) supported on hydroxyapatite for aerobic oxidation catalysts and real-world conventional nanocatalysts. of cyclohexane to cyclohexanol and cyclohexanone [45] . A Surface science has made major impacts on catalysis in monotonic increase in the turnover frequency was found with the past decades, and the knowledge learned provides n up to 39, and then followed by a decrease with a further invaluable information towards rational catalyst design, but increase in n to ∼ 85. the apparent gaps of the material and pressure for model

Overall, the Au n (SR) m nanoclusters, as a new class of nano- single crystal surfaces have long been recognized, which catalyst, hold great promise in future research. By carefully raises questions whether the conclusions from single crys- controlling the experimental conditions in the synthesis, a tal surface studies are transferable to the case of nanoscale series of well-deﬁ ned Au n (SR) m nanoclusters as well as other catalysts. With more in-depth research work to be carried noble metals should be achievable in future research. The out in the future, we anticipate that catalytic science will advantages of Au n (SR) m nanoparticles lie in several aspects. achieve remarkable advances and contribute to the chemi-

First of all, Au n (SR) m nanoclusters provide the access to cal, energy, and pharmaceutical industries. Steady advances the ultra-small size regime (< 2 nm) that was not accessible. in understanding the catalytic mechanism will eventually

Second, the ultra-small size of Au n (SR) m nanoclusters ren- lead to new design of high-performance catalysts for chem- ders them non-metallic electronic properties, manifested in ical processes of environmental and energy importance and the quantized electron energy levels [112] . Future research to develop new catalytic technologies that eliminate, or at may reveal more fundamental aspects of the semiconducting least minimize, the use and release of hazardous materials.

C C CH CCC CH C C C CH CH3 3 3 H H H OH H O H O H H

α β Figure 27 Proposed mechanism for Au25 (SR) 18 nanocluster-catalyzed chemoselective hydrogenation of , -unsaturated ketone to unsaturated alcohol. For clarity, the -SR ligands are omitted in the drawing. From Ref. [117] . R. Jin: Nanocatalysis 51

Apart from the research themes discussed in Section 3, 4.3. Metal-support interactions some additional opportunities (especially on fundamental catalysis) are also highly attractive in the future research of The interaction of nanoparticles and supports is one of the catalysis. These are perhaps more challenging but very excit- major topics in nanocatalysis. In the early days, the support ing to pursue. was introduced to merely disperse metal nanoparticles and was thought to be inert in the catalytic process, but in later 4.1. Chiral catalysis research the support has been demonstrated in many cases for its signifi cant role in catalytic reactions [119, 131, 132] . Chirality is an intriguing phenomenon, which is of particu- Recent work by Sonströ m et al. demonstrated that ligand- lar importance in biology and drug molecules; for example, capped colloidal Pt nanoparticles, after deposition onto nano- the right-handed Dopa (a drug) is not active in the treatment structured iron oxide, can catalyze CO oxidation at unusually of Parkinson’ s disease. Ordinary syntheses often produce a low temperatures [133] , similar to the Au-catalyzed CO oxi- mixture of enantiomers in equal amounts, i.e., the enantio- dation by Haruta and coworkers [134]. Sonströ m et al. [133] meric excess (ee) is zero, but with certain catalysts one may discussed that the ligands could effi ciently modify the metal- selectively obtain one of the enantiomers. Although chiral support interaction, and the benefi cial interaction remediated catalysis has long been practiced in homogeneous catalysis CO poison of the catalyst at low temperatures and, hence, using chiral ligand-modifi ed oragnometallic complexes, it much higher activity. It is worth noting that some well-defi ned has not been achieved to the same level in heterogeneous support materials have become available, such as the different catalysis. Thus, chiral nanocatalysis is another major theme shapes of TiO2 nanocrystals [135, 136], highly ordered mes- γ to pursue and is anticipated to attain signifi cant progress in oporous supports of -Al 2 O3 [137] , SiO2 [138] , aluminosili- future research. With regard to chiral nanoparticles, there are cate such as MCM-41 [139] , MCM-48 [132] , ZSM-5 [139] , generally two types. Type I nanoparticles pertain to chiral SBA-15 [140] , TS-1 [141] , mesoporous carbon [142] , and ligand-protected nanoparticles, which mimic chiral ligand- so forth. Combining these well-defi ned supports with well- modifi ed complexes [125] . For example, Li and coworkers defi ned metal nanoparticles is anticipated to provide new recently attained enantioselective hydrogenation (96% ee) of opportunities for in-depth catalytic research. In addition to α-ketoesters using chirally modifi ed Pt nanoparticles encap- the metal nanoparticle-support interactions, the infl uence of sulated inside CNTs, in which the high activity and enanti- surface ligands on particles also deserves to be carefully stud- oselectivity were attributed to the unique 1D nanochannels of ied [143] . In previous work, some reports have indicated very the CNTs as the nanochannels readily enrich both the chiral little infl uence of ligands on catalytic performance [21] , but modifi er and the reactants [126] . Another example is that Pd some indicate large effects. The discrepancy in results may be nanoparticles modifi ed by a chiral N -heterocyclic carbene due to the specifi c types of ligands (and the specifi c reactions were demonstrated to catalyze asymmetric α -arylations at as well). A consistent picture has yet to be obtained in future up to 85% ee [127] . In contrast with the type I chiral nano- research. particles, recent research has obtained a new type of chiral nanoparticles (referred to type II chiral nanoparticles), which 4.4. Catalytic active sites exhibits intrinsic chirality in the metal core, such as the chiral arrangement of gold atoms in the 38-atom (Au 38 ) nanopar- Apart from the size, shape, and composition dependences in ticles capped by thiolate ligands [128] . These nanoparticles nanocatalysis, which are all of practical importance, identi- hold promise in chiral metal catalysis. fying the active sites and their role in the catalytic mechanism is of fundamental signifi cance [144] . The nature or 4.2. Atomic level control of catalysts structure of active sites in catalysts has intrigued researchers for many decades. However, without well-defi ned nanocata- Correlation of catalytic properties of nanoparticles with lysts, it would not be possible to pinpoint the catalytic active their atomic level structure is the ultimate goal in catalytic sites. Some interesting research has started to tackle this task research. Such knowledge will provide important guidelines [123, 145] . Future research efforts should devise strategies for the future design of new catalysts for specifi c chemi- of enabling observation of the catalytic reaction steps at the cal processes. In this regard, well-defi ned nanocatalysts are catalytic site at atomic resolution. Revealing the nature and expected to play signifi cant roles in unraveling fundamental the structure of active sites of catalysts is certainly a grand principles of catalysis. In terms of the shape effect in polyhe- challenge. dral nanocrystals, correlation of the structure of various (hkl ) facets with catalytic properties is highly desirable. For bi- or multi-metallic catalysts, detailed information on the atomic 5. Concluding remarks distributions and the oxidation states of metal atoms [93, 129], as well as the dynamics of metal segregation, are of particular Future progress in catalysis science is expected to bring about importance. The atomically precise nanoclusters hold prom- exciting progress not only in the design of new nanostruc- ise in investigating the fundamental aspects of nanocatalysis tured catalysts with high selectivity, activity, and durability, as the total structure can be solved by X-ray crystallography but also in the fundamental understanding of the elementary and the nanoclusters are particularly robust [130] . reaction mechanisms at the atomic level [146, 147]. Catalysts 52 R. Jin: Nanocatalysis

with high selectivity [148] and of low energy consumption with controllable sizes. Crystal Growth Design 2008, 8, (i.e., in mild processes) should be of particular importance in 4440 – 4444. future catalytic research. As discussed by Somorjai et al. in [15] Huo Z, Tsung C-K, Huang W, Zhang X, Yang P. Sub-two their recent work [149] ‘ the challenge of chemistry in the 21st nanometer single crystal Au nanowires. Nano Lett. 2008, 8, century is to achieve 100 % selectivity of the desired prod- 2041 – 2044. [16] Narayanan R, El-Sayed MA. Catalysis with transition metal uct molecule in multipath reactions ( “ green chemistry ” ) and nanoparticles in colloidal solution: nanoparticle shape depen- develop renewable energy based processes’ . Looking into the dence and stability. J. Phys. Chem. B 2005, 109, 12663 – 12676. future of catalytic science, there are many huge challenges [17] Narayanan R, El-Sayed MA. Effect of nanocatalysis in colloi- but they offer very exciting opportunities. Tailoring catalysts dal solution on the tetrahedral and cubic nanoparticle shape: at the atomic level will constitute the ultimate goal of cataly- electron-transfer reaction catalyzed by platinum nanoparticles. sis research. J. Phys. Chem. B 2004, 108, 5726 – 5733. [18] Narayanan R, El-Sayed MA. Shape-dependent catalytic activity of platinum nanoparticles in colloidal solution. Nano Lett. 2004, Acknowledgments 4, 1343 – 1348. [19] Narayanan R, El-Sayed MA. Effect of colloidal nanocatalysis on Rongchao Jin acknowledges support of research by Carnegie Mellon the metallic nanoparticle shape: the suzuki reaction. Langmuir University, the Air Force Offi ce of Scientifi c Research (AFOSR) 2005, 21, 2027 – 2033. under Award No. FA9550-11-1-9999 (FA9550-11-1-0147), and the [20] Tsung C-K, Kuhn JN, Huang W, Aliaga C, Hung L-I, Somorjai Camille Dreyfus Teacher-Scholar Awards Program for fi nancial sup- GA, Yang P. Sub-10 nm platinum nanocrystals with size and port. Rongchao Jin is a Camille Dreyfus Teacher-Scholar. shape control: catalytic study for ethylene and pyrrole hydrogenation. J. Am. Chem. Soc. 2009, 131, 5816 – 5822. [21] Zhang Y, Grass ME, Kuhn JN, Tao F, Habas SE, Huang W, References Yang P, Somorjai GA. Highly selective synthesis of catalyti- cally active monodisperse rhodium nanocubes. J. Am. Chem. [1] Bond GC. Heterogeneous Catalysis: Principles and Soc. 2008, 130, 5868 – 5869. Applications . Clarendon Press, 1974. [22] Bratlie KM, Lee H, Komvopoulos K, Yang P, Somorjai GA. [2] Jin R, Chen Y, Li W, Cui W, Ji Y, Yu C, Jiang Y. Mechanism for Platinum nanoparticle shape effects on benzene hydrogenation selectivity. Nano Lett. 2007, 7, 3097 – 3101. catalytic partial oxidation of methane to syngas over a Ni/Al2 O3 catalyst. Appl. Catal. A 2000, 201, 71 – 80. [23] Mostafa S, Behafarid F, Croy JR, Ono LK, Li L, Yang JC, [3] Cunningham DAH, Vogel W, Kageyama H, Tsubota S, Haruta Frenkel AI, Cuenya BR. Shape-dependent catalytic properties of M. The relationship between the structure and activity of nano- Pt nanoparticles. J. Am. Chem. Soc. 2010, 132, 15714 – 15719. [24] Kundu S, Lau S, Liang H. Shape-controlled catalysis by cetylt- meter size gold when supported on Mg(OH) 2 . J. Catal. 1998, 177, 1 – 10. rimethylammonium bromide terminated gold nanospheres, nano- [4] Chen Y, Qiu J, Wang X, Xiu J. Preparation and application rods, and nanoprisms. J. Phys. Chem. C 2009, 113, 5150 − 5156. of highly dispersed gold nanoparticles supported on silica for [25] Kundu S, Wang K, Liang H. Photochemical generation of cata- catalytic hydrogenation of aromatic nitro compounds. J. Catal. lytically active shape selective rhodium nanocubes. J. Phys. 2006, 242, 227 – 230. Chem. C 2009, 113, 18570 – 18577. [5] Bradley JS, Hill E, Leonowicz ME, Witzke H. Clusters, colloids [26] Semagina N, Kiwi-Minsker L. Palladium nanohexagons and and catalysis. J. Mol. Catal. 1987, 41, 59 – 74. nanospheres in selective alkyne hydrogenation. Catal. Lett. [6] Davis SC, Klabunde KJ. Unsupported small metal particles: 2009, 127, 334 – 338. preparation, reactivity, and characterization. Chem. Rev. 1982, [27] Xu R, Wang D, Zhang J, Li Y. Shape-dependent catalytic activ- 82, 153 – 208. ity of silver nanoparticles for the oxidation of styrene. Chem. [7] Spencer ND, Schoonmaker RC, Somorjai GA. Iron single crys- Asian J. 2006, 1, 888 – 893. tals as ammonia synthesis catalysts: effect of surface structure [28] Xiao L, Zhuang L, Liu Y, Lu J, Abruna HD. Activating Pd by on catalyst activity. J. Catal. 1982, 74, 129 – 135. morphology tailoring for oxygen reduction. J. Am. Chem. Soc. [8] Bozso F, Ertl G, Grunze M, Weiss M. Interaction of nitrogen with 2008, 131, 602 – 608. iron surfaces: I. Fe(100) and Fe(111). J. Catal. 1977, 49, 18 – 41. [29] Seo B, Choi S, Kim J. Simple electrochemical deposition of Au [9] Li Y, Liu Q, Shen W. Morphology-dependent nanocatalysis: nanoplates from Au(I) cyanide complexes and their electrocata- metal particles. Dalton Trans. 2011, 40, 5811 – 5826. lytic activities. ACS Appl. Mater. Interfaces 2011, 3, 441 – 446. [10] Ahmadi TS, Wang ZL, Green TC, Henglein A, El-Sayed MA. [30] Koper MTM. Structure sensitivity and nanoscale effects in elec- Shape-controlled synthesis of colloidal platinum nanoparticles. trocatalysis. Nanoscale 2011, 3, 2054 – 2073. Science 1996, 272, 1924 – 1925. [31] Xu W, Kong JS, Yeh Y-TE, Chen P. Single-molecule nanoca- [11] Jin R, Cao Y, Mirkin CA, Kelly KL, Schatz GC, Zheng JG. talysis reveals heterogeneous reaction pathways and catalytic Photoinduced conversion of silver nanospheres to nanoprisms. dynamics. Nat. Mater. 2008, 7, 992 – 996. Science 2001, 294, 1901 – 1903. [32] Chen P, Xu W, Zhou X, Panda D, Kalininski A. Single- [12] Murphy CJ, Thompson LB, Alkilany AM, Sisco PN, Boulos SP, nanoparticle catalysis at single-turnover resolution. Chem. Sivapalan ST, Yang JA, Chernak DJ, Huang J. The many faces Phys. Lett. 2009, 470, 151 – 157. of gold nanorods. J. Phys. Chem. Lett. 2010, 1, 2867 – 2875. [33] Liu Y, Walker ARH. Monodisperse gold-copper bimetallic [13] Sun YG, Xia YN. Shape-controlled synthesis of gold and silver nanocubes: facile one-step synthesis with controllable size and nanoparticles. Science 2002, 298, 2176 – 2179. composition. Angew. Chem. Int. Ed. 2010, 49, 6781 – 6785. [14] Niu W, Li Z-Y, Shi L, Liu X, Li H, Han S, Chen J, Xu G. Seed- [34] Tian N, Zhou Z-Y, Sun S-G. Platinum metal catalysts of high- mediated growth of nearly monodisperse palladium nanocubes index surfaces: from single-crystal planes to electrochemically R. Jin: Nanocatalysis 53

shape-controlled nanoparticles. J. Phys. Chem. C 2008, 112, [51] Lundwall MJ, McClure SM, Goodman DW. Probing terrace and 19801 – 19817. step sites on Pt nanoparticles using CO and ethylene. J. Phys. [35] Yu Y, Zhang Q, Liu B, Lee JY. Synthesis of nanocrystals with Chem. C 2010, 114, 7904 – 7912. variable high-index Pd facets through the controlled heteroepi- [52] Kim DY, Im SH, Park OO. Synthesis of tetrahexahedral gold taxial growth of trisoctahedral Au templates. J. Am. Chem. Soc. nanocrystals with high-index facets. Crystal Growth Design 2010, 132, 18258 – 18265. 2010, 10, 3321 – 3323. [36] Carb ó -Argibay E, Rodr í guez-Gonz á lez B, G ó mez-Gra ñ a S, [53] Yu Y, Zhang Q, Lu X, Lee JY. Seed-mediated synthesis of Guerrero-Mart í nez A, Pastoriza-Santos I, P é rez-Juste J, Liz- monodisperse concave trisoctahedral gold nanocrystals with Marz á n LM. The crystalline structure of gold nanorods revis- controllable sizes. J. Phys. Chem. C 2010, 114, 11119 – 11126. ited: evidence for higher-index lateral facets. Angew. Chem. Int. [54] Campbell CT. Studies of model catalysts with well-defi ned Ed. 2010, 49, 9397 – 9400. surfaces combining ultrahigh vacuum surface characterization [37] Tian N, Zhou Z-Y, Sun S-G, Ding Y, Wang ZL. Synthesis of with medium- and high-pressure kinetics. Adv. Catal. 1989, 36, tetrahexahedral platinum nanocrystals with high-index fac- 1 – 54. ets and high electro-oxidation activity. Science 2007, 316, [55] Somorjai GA. Modern surface science and surface technolo- 732 – 735. gies: an introduction. Chem. Rev. 1996, 96, 1223 – 1236. [38] Chen Q-S, Zhou Z-Y, Vidal-Iglesias FJ, Solla-Gullon J, Feliu [56] Cao Y, Jin R, Mirkin CA. DNA-modifi ed core-shell Ag/Au JM, Sun S-G. Signifi cantly enhancing catalytic activity of tet- nanoparticles. J. Am. Chem. Soc. 2001, 123, 7961 – 7962. rahexahedral Pt nanocrystals by Bi adatom decoration. J. Am. [57] Sun Y, Xia Y. Alloying and dealloying processes involved in the Chem. Soc. 2011, 133, 12930 – 12933. preparation of metal nanoshells through a galvanic replacement [39] Jin M, Zhang H, Xie Z, Xia Y. Palladium concave nanocubes reaction. Nano Lett. 2003, 3, 1569 – 1572. with high-index facets and their enhanced catalytic properties. [58] Wagner FT, Lakshmanan B, Mathias MF. Electrochemistry Angew. Chem. Int. Ed. 2011, 50, 7850 – 7854. and the future of the automobile. J. Phys. Chem. Lett. 2010, 1, [40] Yu T, Kim DY, Zhang H, Xia Y. Platinum concave nano- 2204 – 2219. cubes with high-index facets and their enhanced activity for [59] Peng Z, Wu J, Yang H. Synthesis and oxygen reduction elec- oxygen reduction reaction. Angew. Chem. Int. Ed. 2011, 50, trocatalytic property of platinum hollow and platinum-on-silver 2773 – 2777. nanoparticles. Chem. Mater. 2009, 22, 1098 – 1106. [41] Zhang J, Langille MR, Personick ML, Zhang K, Li S, Mirkin [60] Sasaki K, Naohara H, Cai Y, Choi YM, Liu P, Vukmirovic MB, CA. Concave cubic gold nanocrystals with high-index facets. J. Wang JX, Adzic RR. Core-protected platinum monolayer shell Am. Chem. Soc. 2010, 132, 14012 – 14014. high-stability electrocatalysts for fuel-cell cathodes. Angew. [42] Ming T, Feng W, Tang Q, Wang F, Sun L, Wang J, Yan C. Chem. Int. Ed. 2010, 49, 8602 – 8607. Growth of tetrahexahedral gold nanocrystals with high-index [61] Shao M, Peles A, Shoemaker K, Gummalla M, Njoki PN, Luo facets. J. Am. Chem. Soc. 2009, 131, 16350 – 16351. J, Zhong C-J. Enhanced oxygen reduction activity of platinum [43] Lu C-L, Prasad KS, Wu H-L, Ho J-AA, Huang MH. Au nano- monolayer on gold nanoparticles. J. Phys. Chem. Lett. 2011, 2, cube-directed fabrication of Au-Pd core-shell nanocrystals with 67 – 72. tetrahexahedral, concave octahedral, and octahedral structures [62] Mazumder V, Chi M, More KL, Sun S. Synthesis and character- and their electrocatalytic activity. J. Am. Chem. Soc. 2010, 132, ization of multimetallic Pd/Au and Pd/Au/FePt core/shell nano- 14546 – 14553. particles. Angew. Chem. Int. Ed. 2010, 49, 9368 – 9372. [44] Wang F, Li C, Sun L-D, Wu H, Ming T, Wang J, Yu JC, Yan [63] Yin H, Ma Z, Chi M, Dai S. Heterostructured catalysts prepared

C-H. Heteroepitaxial growth of high-index-faceted palladium by dispersing Au@Fe2 O3 core-shell structures on supports and nanoshells and their catalytic performance. J. Am. Chem. Soc. their performance in CO oxidation. Catal. Today 2011, 160, 2011, 133, 1106 – 1111. 87 – 95. [45] Liu Y, Tsunoyama H, Akita T, Xie S, Tsukuda T. Aerobic oxida- [64] Dai Y, Lim B, Yang Y, Cobley CM, Li W, Cho EC, Grayson B, tion of cyclohexane catalyzed by size-controlled Au clusters on Fanson PT, Campbell CT, Sun Y, Xia Y. A sinter-resistant cata-

hydroxyapatite: size effect in the sub-2 nm regime. ACS Catal. lytic system based on platinum nanoparticles supported on TiO2 2011, 1, 2 – 6. nanofi bers and covered by porous silica. Angew. Chem. Int. Ed. [46] Liu L, Pippel E. Low-platinum-content quaternary PtCuCoNi 2010, 49, 8165 – 8168. nanotubes with markedly enhanced oxygen reduction activity. [65] Chen C, Nan C, Wang D, Su Q, Duan H, Liu X, Zhang L, Angew. Chem. Int. Ed. 2011, 50, 2729 – 2733. Chu D, Song W, Peng Q, Li Y. Mesoporous multicomponent [47] Huang X, Zhao Z, Fan J, Tan Y, Zheng N. Amine-assisted nanocomposite colloidal spheres: ideal high-temperature stable synthesis of concave polyhedral platinum nanocrystals hav- model catalysts. Angew. Chem. Int. Ed. 2011, 50, 3725 – 3729. ing {411} high-index facets. J. Am. Chem. Soc. 2011, 133, [66] Carregal-Romero S, Buurma NJ, Perez-Juste J, Liz-Marzan LM, 4718 – 4721. Herves P. Catalysis by Au@pNIPAM nanocomposites: effect of [48] Nakamula I, Yamanoi Y, Imaoka T, Yamamoto K, Nishihara H. the cross-linking density. Chem. Mater. 2010, 22, 3051 – 3059. A uniform bimetallic rhodium/iron nanoparticle catalyst for the [67] Yuan C, Luo W, Zhong L, Deng H, Liu J, Xu Y, Dai L. Gold@ hydrogenation of olefi ns and nitroarenes. Angew. Chem. Int. Ed. polymer nanostructures with tunable permeability shells for 2011, 50, 5830 – 5833. selective catalysis. Angew. Chem. Int. Ed. 2011, 50, 3515 – 3519. [49] Liu Y, Tsunoyama H, Akita T, Tsukuda T. Effi cient and selective [68] Kim S-W, Kim M, Lee WY, Hyeon T. Fabrication of hollow

epoxidation of styrene with TBHP catalyzed by Au 25 clusters on palladium spheres and their successful application to the recy- hydroxyapatite. Chem. Commun. 2010, 46, 550 – 552. clable heterogeneous catalyst for Suzuki coupling reactions. [50] Lee SW, Chen S, Sheng W, Yabuuchi N, Kim Y-T, Mitani T, J. Am. Chem. Soc. 2002, 124, 7642 – 7643. Vescovo E, Shao-Horn Y. Roles of surface steps on Pt nanopar- [69] Yen CW, Mahmoud MA, El-Sayed MA. Photocatalysis in ticles in electro-oxidation of carbon monoxide and methanol. gold nanocage nanoreactors. J. Phys. Chem. A 2009, 113, J. Am. Chem. Soc. 2009, 131, 15669 – 15677. 4340 – 4345. 54 R. Jin: Nanocatalysis

[70] Wang S, Zhang M, Zhang W. Yolk-shell catalyst of single Au [88] Kang Y, Murray CB. Synthesis and electrocatalytic properties nanoparticle encapsulated within hollow mesoporous silica of cubic Mn-Pt nanocrystals. J. Am. Chem. Soc. 2010, 132, microspheres. ACS Catal. 2011, 1, 207 – 211. 7568 – 7569. [71] Ma Z, Dai S. Design of novel structured gold nanocatalysts. [89] Yang H, Zhang J, Sun K, Zou S, Fang J. Enhancing by weaken-

ACS Catal. 2011, 1, 805 – 818. ing: electrooxidation of methanol on Pt3 Co and Pt nanocubes. [72] Lu L, Eychm ü ller A. Ordered macroporous bimetallic nano- Angew. Chem. Int. Ed. 2010, 49, 6848 – 6851. structures: design, characterization, and applications. Acc. [90] Lee YW, Kim M, Kang SW, Han SW. Polyhedral bimetal- Chem. Res. 2008, 41, 244 – 253. lic alloy nanocrystals exclusively bound by {110} facets: [73] Lim B, Jiang M, Camargo PHC, Cho EC, Tao J, Lu X, Au-Pd rhombic dodecahedra. Angew. Chem. Int. Ed. 2011, 50, Zhu Y, Xia Y. Pd-Pt bimetallic nanodendrites with high activity 3466 – 3470. for oxygen reduction. Science 2009, 324, 1302 – 1305. [91] Nutt MO, Heck KN, Alvarez P, Wong MS. Improved Pd-on-Au [74] Mohanty A, Garg N, Jin R. A universal approach to the synthe- bimetallic nanoparticle catalysts for aqueous-phase trichlo- sis of noble metal nanodendrites and their catalytic properties. roethene hydrodechlorination. Appl. Catal. B 2006, 69, Angew. Chem. Int. Ed. 2010, 49, 4962 – 4966. 115 – 125. [75] Peng Z, Yang H. Synthesis and oxygen reduction electrocata- [92] Wanjala BN, Loukrakpam R, Luo J, Njoki PN, Mott D, Zhong lytic property of Pt-on-Pd bimetallic heteronanostructures. C-J, Shao M, Protsailo L, Kawamura T. Thermal treatment of J. Am. Chem. Soc. 2009, 131, 7542 – 7543. PtNiCo electrocatalysts: effects of nanoscale strain and struc- [76] Lim B, Xia Y. Metal nanocrystals with highly branched mor- ture on the activity and stability for the oxygen reduction reac- phologies. Angew. Chem. Int. Ed. 2010, 50, 76 – 85. tion. J. Phys. Chem. C 2010, 114, 17580 – 17590. [77] Sun S, Zhang G, Geng D, Chen Y, Li R, Cai M, Sun X. A highly [93] Wanjala BN, Fang B, Luo J, Chen Y, Yin J, Engelhard MH, durable platinum nanocatalyst for proton exchange membrane Loukrakpam R, Zhong C-J. Correlation between atomic fuel cells: multiarmed starlike nanowire single crystal. Angew. coordination structure and enhanced electrocatalytic activity Chem. Int. Ed. 2010, 50, 422 – 426. for trimetallic alloy catalysts. J. Am. Chem. Soc. 2011, 133, [78] Sanles-Sobrido M, Correa-Duarte MA, Carregal-Romero S, 12714 – 12727. Rodriguez-Gonzalez B, Alvarez-Puebla RA, Herves P, Liz- [94] Cui C-H, Li H-H, Yu J-W, Gao M-R, Yu S-H. Ternary hetero- Marzan LM. Highly catalytic single-crystal dendritic Pt nanostructured nanoparticle tubes: a dual catalyst and its synergistic

structures supported on carbon nanotubes. Chem. Mater. 2009, enhancement effects for O2 /H2 O2 reduction. Angew. Chem. Int. 21, 1531 – 1535. Ed. 2010, 49, 9149 – 9152.

[79] Michel M, Taylor A, Sekol R, Podsiadlo P, Ho P, Kotov N, [95] Wang C, Daimon H, Sun S. Dumbbell-like Pt-Fe 3 O4 nanopar- Thompson L. High-performance nanostructured membrane ticles and their enhanced catalysis for oxygen reduction reac- electrode assemblies for fuel cells made by layer-by-layer tion. Nano Lett. 2009, 9, 1493 – 1496. assembly of carbon nanocolloids. Adv. Mater. 2007, 19, [96] Qi C, Huang J, Bao S, Su H, Akita T, Haruta M. Switching of 3859 – 3864. reactions between hydrogenation and epoxidation of propene

[80] Kim Y, Hong JW, Lee YW, Kim M, Kim D, Yun WS, Han SW. over Au/Ti-based oxides in the presence of H2 and O2 . J. Catal. Synthesis of AuPt heteronanostructures with enhanced electro- 2011, 281, 12 – 20. catalytic activity toward oxygen reduction. Angew. Chem. Int. [97] Sakurai H, Tsunoyama H, Tsukuda T. Oxidative homo-cou- Ed. 2010, 49, 10197 – 10201. pling of potassium aryltriﬂ uoroborates catalyzed by gold nan- [81] Guo S, Dong S, Wang E. Three-dimensional Pt-on-Pd bimetal- ocluster under aerobic conditions. J. Organomet. Chem. 2007, lic nanodendrites supported on graphene nanosheet: facile syn- 692, 368 – 374. thesis and used as an advanced nanoelectrocatalyst for methanol [98] Bethke GK, Kung HH. Selective CO oxidation in a hydrogen- γ oxidation. ACS Nano 2009, 4, 547 – 555. rich stream over Au/ -Al 2 O3 catalysts. Appl. Catal. A 2000, [82] Yeo KM, Choi S, Anisur RM, Kim J, Lee IS. Surfactant-free 194 – 195, 43 – 53. platinum-on-gold nanodendrites with enhanced catalytic perfor- [99] Klabunde KJ. Nanoscale Materials in Chemistry . Wiley- mance for oxygen reduction. Angew. Chem. Int. Ed. 2011, 50, Interscience, 2001. 745 – 748. [100] Whetten RL, Shaﬁ gullin MN, Khoury JT, Schaaff TG, [83] Lee Y, Park TG. Facile fabrication of branched gold nanopar- Vezmar I, Alvarez MM, Wilkinson A. Crystal structures of ticles by reductive hydroxyphenol derivatives. Langmuir 2011, molecular gold nanocrystal arrays. Acc. Chem. Res. 1999, 32, 27, 2965 – 2971. 397 – 406. [84] Mori K, Kawashima M, Che M, Yamashita H. Enhancement [101] Martin MN, Basham JI, Chando P, Eah S-K. Charged gold of the photoinduced oxidation activity of a ruthenium(II) com- nanoparticles in non-polar solvents: 10-min synthesis and 2D plex anchored on silica-coated silver nanoparticles by localized self-assembly. Langmuir 2010, 26, 7410 – 7417. surface plasmon resonance. Angew. Chem. Int. Ed. 2010, 49, [102] Eah S-K. A very large two-dimensional superlattice domain of 8598 – 8601. monodisperse gold nanoparticles by self-assembly. J. Mater. [85] Zhang Q, Lima DQ, Lee I, Zaera F, Chi M, Yin Y. A highly Chem. 2011, 21, 16866 – 16868. active titanium dioxide based visible-light photocatalyst with [103] Njoki PN, Lim IIS, Mott D, Park H-Y, Khan B, Mishra S, nonmetal doping and plasmonic metal decoration. Angew. Sujakumar R, Luo J, Zhong C-J. Size correlation of optical Chem. Int. Ed. 2011, 50, 7088 – 7092. and spectroscopic properties for gold nanoparticles. J. Phys. [86] Wang C, Thompson RL, Baltrus J, Matranga C. Visible light Chem. C 2007, 111, 14664 – 14669.

photoreduction of CO 2 using CdSe/Pt/TiO 2 heterostructured [104] Joo SH, Park JY, Renzas JR, Butcher DR, Huang W, Somorjai catalysts. J. Phys. Chem. Lett. 2010, 1, 48 – 53. GA. Size effect of ruthenium nanoparticles in catalytic carbon [87] Wanjala BN, Luo J, Fang B, Mott D, Zhong C-J. Gold-platinum monoxide oxidation. Nano Lett. 2010, 10, 2709 – 2713. nanoparticles: alloying and phase segregation. J. Mater. Chem. [105] Grass ME, Joo SH, Zhang Y, Somorjai GA. Colloidally syn- 2011, 21, 4012 – 4020. thesized monodisperse Rh nanoparticles supported on SBA-15 R. Jin: Nanocatalysis 55

for size- and pretreatment-dependent studies of CO oxidation. [123] Zhu Y, Wu Z, Gayathri C, Qian H, Gil RR, Jin R. Exploring

J. Phys. Chem. C 2009, 113, 8616 – 8623. stereoselectivity of Au25 nanoparticle catalyst for hydrogena- [106] Kuhn JN, Huang W, Tsung C-K, Zhang Y, Somorjai GA. tion of cyclic ketone. J. Catal. 2010, 271, 155 – 160. Structure sensitivity of carbon-nitrogen ring opening: impact [124] Zhu M, Eckenhoff WT, Pintauer T, Jin R. Conversion of - of platinum particle size from below 1 to 5 nm upon pyrrole anionic [Au25 (SCH2 CH 2 Ph)18 ] cluster to charge neutral cluster hydrogenation product selectivity over monodisperse platinum via air oxidation. J. Phys. Chem. C 2008, 112, 14221 – 14224. nanoparticles loaded onto mesoporous silica. J. Am. Chem. [125] Shukla N, Bartel MA, Gellman AJ. Enantioselective separa- Soc. 2008, 130, 14026 – 14027. tion on chiral Au nanoparticles. J. Am. Chem. Soc. 2010, 132, [107] Mondloch JE, Wang Q, Frenkel AI, Finke RG. Development 8575 – 8580. plus kinetic and mechanistic studies of a prototype supported- [126] Chen Z, Guan Z, Li M, Yang Q, Li C. Enhancement of the per- nanoparticle heterogeneous catalyst formation system in con- formance of a platinum nanocatalyst conﬁ ned within carbon

tact with solution: Ir(1,5-COD)Cl/Al2 O3 and its reduction by nanotubes for asymmetric hydrogenation. Angew. Chem. Int.

H 2 to Ir(0) n /Al2 O3 . J. Am. Chem. Soc. 2010, 132, 9701 – 9714. Ed. 2011, 50, 4913 – 4917. [108] Wilson OM, Scott RWJ, Garcia-Martinez JC, Crooks RM. [127] Ranganath KVS, Kloesges J, Sch ä fer AH, Glorius F. Synthesis, characterization, and structure-selective extraction Asymmetric nanocatalysis: n-heterocyclic carbenes as chiral

of 1– 3-nm diameter AuAg dendrimer-encapsulated bimetallic modiﬁ ers of Fe3 O4 /Pd nanoparticles. Angew. Chem. Int. Ed. nanoparticles. J. Am. Chem. Soc. 2004, 127, 1015 – 1024. 2010, 49, 7786 – 7789. [109] Boisselier E, Diallo AK, Salmon L, Ornelas C, Ruiz J, Astruc [128] Qian H, Eckenhoff WT, Zhu Y, Pintauer T, Jin R. Total struc-

D. Encapsulation and stabilization of gold nanoparticles with ture determination of thiolate-protected Au38 nanoparticles. “ click ” polyethyleneglycol dendrimers. J. Am. Chem. Soc. J. Am. Chem. Soc. 2010, 132, 8280 – 8281. 2010, 132, 2729 – 2742. [129] Frondelius P, H ä kkinen H, Honkala K. Formation of gold(I) [110] Liu Y, Jia C-J, Yamasaki J, Terasaki O, Sch ü th F. Highly active edge oxide at ﬂ at gold nanoclusters on an ultrathin MgO ﬁ lm iron oxide supported gold catalysts for CO oxidation: how under ambient conditions. Angew. Chem. Int. Ed. 2010, 49, small must the gold nanoparticles be ? Angew. Chem. Int. Ed. 7913 – 7916. 2011, 49, 5771 – 5775. [130] Zhu Y, Qian H, Jin R. Catalysis opportunities of atomically pre- [111] Jin R, Zhu Y, Qian H. Quantum-sized gold nanoclusters: bridg- cise gold nanoclusters. J. Mater. Chem. 2011, 21, 6793 – 6799. ing the gap between organometallics and nanocrystals. Chem. [131] Baker LR, Hervier A, Seo H, Kennedy G, Komvopoulos K, Eur. J. 2011, 17, 6584 – 6593. Somorjai GA. Highly n -type titanium oxide as an electroni- [112] Jin R, Qian H, Zhu Y, Das A. Atomically precise nanoparticles: cally active support for platinum in the catalytic oxidation of a new frontier in nanoscience. J. Nanosci. Lett. 2011, 1, 72 – 78. carbon monoxide. J. Phys. Chem. C 2011, 115, 16006 – 16011. [113] Jin R, Qian H, Wu Z, Zhu Y, Zhu M, Mohanty A, Garg N. Size [132] Uphade BS, Akita T, Nakamura T, Haruta M. Vapor-phase

focusing: a methodology for synthesizing atomically precise epoxidation of propene using H 2 and O2 over Au/Ti-MCM-48. gold nanoclusters. J. Phys. Chem. Lett. 2010, 1, 2903 – 2910. J. Catal. 2002, 209, 331 – 340. [114] Zhu M, Aikens CM, Hollander FJ, Schatz GC, Jin R. Correlating [133] Sonstr ö m P, Arndt D, Wang X, Zielasek V, B ä umer M. Ligand

the crystal structure of a thiol-protected Au25 cluster and opti- capping of colloidally synthesized nanoparticles – a way to cal properties. J. Am. Chem. Soc. 2008, 130, 5883 – 5885. tune metal-support interactions in heterogeneous gas-phase [115] Qian H, Zhu Y, Jin R. Size-focusing synthesis, optical and catalysis. Angew. Chem. Int. Ed. 2011, 50, 3888 – 3891.

electrochemical properties of monodisperse Au38 (SC2 H4 Ph) 24 [134] Haruta M, Yamada N, Kobayashi T, Iijima S. Gold catalysts nanoclusters. ACS Nano 2009, 3, 3795 – 3803. prepared by coprecipitation for low-temperature oxidation [116] Qian H, Jin R. Controlling nanoparticles with atomic preci- of hydrogen and of carbon monoxide. J. Catal. 1989, 115,

sion: the case of Au144 (SCH2 CH 2 Ph)60 . Nano Lett. 2009, 9, 301–309. 4083 – 4087. [135] Zhao X, Jin W, Cai J, Ye J, Li Z, Ma Y, Xie J, Qi L. Shape- and

[117] Zhu Y, Qian H, Drake BA, Jin R. Atomically precise Au25 (SR)18 size-controlled synthesis of uniform anatase TiO2 nanocuboids nanoparticles as catalysts for the selective hydrogenation of α , enclosed by active {100} and {001} facets. Adv. Funct. Mater. β-unsaturated ketones and aldehydes. Angew. Chem. Int. Ed. 2011, 21, 3554 – 3563. 2010, 49, 1295 – 1298. [136] Wu B, Guo C, Zheng N, Xie Z, Stucky GD. Nonaqueous [118] Zhu Y, Qian H, Jin R. An atomic-level strategy for unraveling production of nanostructured anatase with high-energy facets.

gold nanocatalysis from the perspective of Au n (SR) m nanoclus- J. Am. Chem. Soc. 2008, 130, 17563 – 17567. ters. Chem. Eur. J. 2010, 16, 11455 – 11462. [137] Yuan Q, Yin A-X, Luo C, Sun L-D, Zhang Y-W, Duan W-T, [119] Bond GC, Louis C, Thompson DT. Catalysis by Gold . Imperial Liu H-C, Yan C-H. Facile synthesis for ordered mesoporous College Press, 2006. aluminas with high thermal stability. J. Am. Chem. Soc. 2008, [120] Zhu Y, Qian H, Das A, Jin R. Comparison of the catalytic 130, 3465 – 3472. properties of 25-atom gold nanospheres and nanorods. Chin. [138] Cheng K, Landry CC. Diffusion-based deprotection in mes- J. Catal. 2011, 32, 1149 – 1155. oporous materials: a strategy for differential functionaliza- [121] Shichibu Y, Negishi Y, Watanabe T, Chaki NK, tion of porous silica particles. J. Am. Chem. Soc. 2007, 129, Kawaguchi H, Tsukuda T. Biicosahedral gold clusters 9674 – 9685. 2 + = [Au25 (PPh 3 )10 (SC n H2 n + 1 ) 5 Cl2 ] (n 2 – 18): a stepping stone [139] Gaomeng L, Ji D, Qian G, Qi Y, Wang X, Suo J. Gold nano- to cluster-assembled materials. J. Phys. Chem. C 2007, 111, particles in mesoporous materials showing catalytic selective 7845 – 7847. oxidation cyclohexane using oxygen. Appl. Catal. A 2005, [122] Qian H, Zhu M, Lanni E, Zhu Y, Bier ME, Jin R. Conversion 280, 175 – 180.

of polydisperse Au nanoparticles into monodisperse Au25 [140] Zhu K, Hu J, Richards R. Aerobic oxidation of cyclohexane nanorods and nanospheres. J. Phys. Chem. C 2009, 113, by gold nanoparticles immobilized upon mesoporous silica. 17599 – 17603. Catal. Lett. 2005, 100, 195 – 199. 56 R. Jin: Nanocatalysis

[141] Zhan G, Du M, Huang J, Li Q. Green synthesis of Au/TS-1 most active and selective ? J. Am. Chem. Soc. 2011, 133, catalysts via two novel modes and their surprising performance 12787 – 12794. for propylene epoxidation. Catal. Commun. 2011, 12, 830 – 833. [146] Bell AT. The impact of nanoscience on heterogeneous cataly- [142] Liang C, Li Z, Dai S. Mesoporous carbon materials: synthesis sis. Science 2003, 299, 1688 – 1691. and modiﬁ cation. Angew. Chem. Int. Ed. 2008, 47, 3696 – 3717. [147] Freyschlag CG, Madix RJ. Precious metal magic: catalytic [143] Lopez-Sanchez JA, Dimitratos N, Hammond C, Brett GL, wizardry. Mater . Today 2011, 14, 134 – 142. Kesavan L, White S, Miedziak P, Tiruvalam R, Jenkins RL, [148] Zaera F. Regio-, stereo-, and enantioselectivity in hydrocar- Carley AF, Knight D, Kiely CJ, Hutchings GJ. Facile removal bon conversion on metal surfaces. Acc. Chem. Res. 2009, 42, of stabilizer-ligands from supported gold nanoparticles. Nat. 1152 – 1160. Chem. 2011, 3, 551 – 556. [149] Somorjai GA, Frei H, Park JY. Advancing the frontiers in [144] Vinod CP. Surface science as a tool for probing nanocatalysis nanocatalysis, biointerfaces, and renewable energy conversion phenomena. Catal. Today 2010, 154, 113 – 117. by innovations of surface techniques. J. Am. Chem. Soc. 2009, [145] Crespo-Quesada M, Yarulin A, Jin M, Xia Y, Kiwi-Minsker 131, 16589 – 16605 . L. Structure sensitivity of alkynol hydrogenation on shape- and size-controlled palladium nanocrystals: which sites are Received September 4, 2011; accepted October 9, 2011

Rongchao Jin is an Assis- tant Professor of Chemistry at Carnegie Mellon Univer- sity. He received his BS in Chemical Physics from the University of Science and Technology of China (Hefei, China) in 1995, his MS in Physical Chemistry/Cataly- sis from Dalian Institute of Chemical Physics (Dalian, China) in 1998, and his PhD in Chemistry from Northwest- ern University (Illinois, USA) in 2003. After 3 years of post- doctoral research at the University of Chicago (IL, USA), he joined the chemistry faculty of Carnegie Mellon University (PA, USA) in 2006. His current research interests focus on atomically precise noble metal nanoparticles and their applications in catalysis, optics, and sensing.