金催化作用的结构敏感性

黄伟新千 坤 邬宗芳 陈士龙

（中国科学技术大学化学物理系，微尺度物质科学国家实验室（筹），能量转换材料中国科学院重点实验室，合肥 230026）

金催化作用的结构敏感性

黄伟新*千 坤 邬宗芳 陈士龙

（中国科学技术大学化学物理系，微尺度物质科学国家实验室（筹），能量转换材料中国科学院重点实验室，合肥 230026）

金催化是纳米催化的代表性体系，金催化作用表现出复杂的结构敏感性。这篇综述总结了金催化作用研究的文献结果和我们利用从单晶到纳米晶的模型催化剂研究金催化作用的进展。展示了NO分解，CO氧化，丙烯在氢气和氧气气氛中环氧化等反应中金催化作用的结构敏感性和金催化剂的活性结构，讨论了金纳米粒子几何结构和电子结构、金纳米粒子–氧化物载体相互作用对金催化作用的影响和金表面低温高催化活性的来源，并展望了金催化作用结构敏感性的未来研究方向。

表面化学；多相催化；模型催化剂；CO氧化反应；NO分解反应；丙烯环氧化反应

1 Introduction

Au had been considered to be catalytically inert until Hutchings1,2and Haruta3,4et al. reported their pioneering works in 1980s. Since then Au catalysis has attracted extensive interests in heterogeneous catalysis. Supported Au nanoparticles（NPs） with the size below 10 nm have been demonstrated to actively catalyze a wide array of reactions including reactions involving CO oxidation reactions, hydrogenation reactions, and selective oxidation reactions. Au catalysis exhibits several unique characteristics5–9: （1） Supported Au catalysts are highly active and very selective at mild reaction temperatures. Au NPs with their sizes of ～5 nm supported on α-Fe2O3, Co3O4, andNiO catalyzed CO oxidation at temperatures as low as –70 °C2,4. Supported Au NPs and clusters catalyzed gas-phase propylene epoxidation with molecular O2to produce propylene oxide with selectivities higher than 90%10–12. （2） Catalytic performances of supported Au catalysts sensitively depend on the size of supported Au NPs. We herein would like to point out that the comparison of turn-over frequency （TOF） among supported Au catalysts with different Au NP sizes is meaningful only when the type of active site does not vary with the size of supported Au NPs. Fujitani and Nakamura13found that TOF of Au/TiO2（110）model catalysts calculated based on Au atoms exposed on Au NPs and based on Au atoms at the Au-TiO2perimeter interfaces varied in different trends as a function of the mean size of Au NPs. （3） Catalytic performances of supported Au catalysts depend on the support. The support not only affects the Au-support interaction and subsequently the structure and size of supported Au NPs, but also directly participates in catalytic reactions. Schubert et al.14grouped oxide supports into “active” supports （TiO2, α-Fe2O3, CoOx, and NiOx） and “inert” supports（SiO2, Al2O3, and MgO） with respect to CO oxidation. It was found that reactive oxygen species could facilely form on the“active” supports and supply for the catalytic CO oxidation but not on the “inert” supports. （4） Catalytic performances of supported Au catalysts in reactions involving molecular O2are affected by moisture （gas-phase reactions） and basicity （liquidphase reactions）. Daté et al.15reported that the moisture enhanced the catalytic activity of Au/TiO2, Au/Al2O3, and Au/SiO2in CO oxidation at 273 K for no less than two orders of magnitude and that the effect of moisture depended on the type of metal oxide support. Kung et al.16–18observed the promotion effect of H2O or H2in the reaction mixture on Au/Al2O3for low-temperature CO oxidation. Sanchez-Castillo et al.19reported that support-free Au nanotubes in polycarbonate membranes exhibited catalytic activity of CO oxidation at room temperature and that the activity is enhanced by liquid H2O and further promoted by increasing the pH value of the solution. Recently experimental evidence was reported for water-mediated O2activation for low temperature CO oxidation over Au/TiO2catalysts20.

The correlation among surface geometric and electronic structures, the geometric shape and electronic structure of a molecule, and the observed macroscopic reactivity represents a very important and long discussed, but not yet solved problem in heterogeneous catalysis. In the 1960s and early 1970s it was found that several hydrocarbon reactions like alkane hydrogenolysis21and isomerization22over platinum catalysts were strongly dependent on the Pt particle size while reactions like cyclopropane ring opening23and olefin hydrogenation24were independent of the Pt particle size. Boudart25classified the first type of reaction as structure-sensitive and the second as structure-insensitive in a review article. And then structure-sensitivity and structure-insensitivity have been widely used to distinguish heterogeneous catalytic reactions26–28. Structure-sensitivity reflects the complex nature of surface structure of catalyst particle and the resulting molecule–catalyst surface interaction and surface reactions. The catalyst particle surface provides various types of adsorption sites to chemisorb and activate reactants in different ways and subsequently catalyze the surface reactions. For each type of adsorption site and the corresponding surface reaction, the reaction rate depends on the chemisorption energy of involved surface adsorbates and the activation energy of involved elementary surface reactions. The measured catalytic activity is the sum of the rates of surface reactions catalyzed by various types of active sites while the measured catalytic selectivity of the catalyst is due to the different products produced by the surface reactions catalyzed over various types of active sites. The size of catalyst particle strongly affects the surface geometric structure （the relative concentrations of terraces, steps and kinks） and the surface electronic structure （the quantum size effect）, resulting in the variations of adsorbate bond strengths and the modification of reaction rates.

The reported Au-catalyzed reactions by far are typically structure-sensitive. Low-temperature CO oxidation is the most representative and extensively studied system. Haruta et al.29,30firstly observed that Au NPs supported on TiO2, Fe2O3and Co3O4below 5 nm exhibited sharply increased catalytic activity. Later Valden et al.31reported that 3.5 nm was the optimal size of Au NPs supported on TiO2（110） surface. Thereafter, studies exploring Au NPs supported on TiO2and TiO2-coated silica aerogel smaller than 2 nm also demonstrated the existence of an optimal size （2.0–2.5 nm）32,33. An operando study revealed an optimal size of （2.1 ± 0.3） nm for the catalytic activity in the Au/TiO2（110） model catalyst system34. However, the structure of supported Au NPs with the optimal size and the interpretation of structure-sensitivity of supported Au catalysts remain controversial. The decrease in the size of supported Au NPs could increase the Au-metal oxide perimeter interface length and thus enhance the catalytic activity of supported Au catalysts following the periphery reaction mechanism35,36. The morphology and geometric structure of supported Au NPs change with their size. Lattice contraction and structural changes of Au NPs supported on TiO2-coated silica aerogel became prominent at Au diameters less than 4–5 nm, in consistence with the enhanced catalytic activity33. On basis of density functional theory （DFT theoretical calculations, Nørskov et al.37–39considered low-coordinated Au atoms on supported Au NPs as the active sites and viewed the effect of support as structural and electronic promotion. The density of low-coordinated Au atoms at their corners, steps and edges generally increases with the size of Au NPs decreasing. The decrease of catalytic activity of Au NPs supported on TiO2with increasing calcination temperature was observed to be accompanied with the particle shape change from the shape containing a large proportion of low-coordinated sites to truncated octahedra with smooth facets32. The size of Au NPs also changes their electron-ic structure when it comes into the range for the quantum size effect, i.e., the metal-to-nonmetal transition. The critical thickness of supported Au NPs for the metal-to-nonmetal transition was reported to be two atomic layers31,40. Valden et al.31observed the size-dependent variations of electronic structure and catalytic activity of Au NPs supported on TiO2（110） and found that supported Au NPs of 2.5–3.5 nm with two layers of gold were most catalytically active. Later they also reported that two-layered Au film vacuum-deposited on a thin layer of TiO2on Mo（112） was the most active structure for low-temperature CO oxidation41. The high catalytic activity of coprecipitated Au/FeOOH catalysts dried at 393 K was also associated with the presence of Au clusters with a bi-layer structure42. However, Au/Fe2O3catalysts prepared with the colloidal deposition method containing supported Au NPs larger than 1 nm were also highly active, implying that bilayer structures and/or diameters of about 0.5 nm were not mandatory for supported Au NPs to achieve the high activity43. Au NPs supported on TiO2（110）with the optimal size （2.1 ± 0.3） nm） were demonstrated to be with a height of about six atomic layers by a recent simultaneous reactivity measurements and grazing incidence smallangle X-ray scattering characterizations34.

Arguments in the structure-sensitivity of low-temperature CO oxidation catalyzed by supported Au catalysts majorly result from the inhomogeneity of supported Au NPs in their sizes. Although the surface structure, surface adsorbate structure, and catalytic performance of supported Au NPs can be comprehensively characterized with efforts, the inhomogeneity of their sizes makes the unambiguous correlation among the size, the catalyst surface structure and the catalytic performance very difficult. Another factor is that the catalyst preparation method strongly and sensitively affects the size of supported Au NPs44. The “same” preparation method adopted by different research groups could lead to supported Au catalysts with different sizes and catalytic activity45, which introduces additional complexity to unifying the structure-sensitivity of Au catalysis in low-temperature CO oxidation.

2 Model catalysts approach

With the respect to the complex and inhomogeneous sizes and surface structures of catalyst NPs in powder catalysts, an effective approach has been developed to employ uniform and well-defined surfaces as the model surfaces, traditionally single crystals-based model catalysts （single crystal/single crystal thin films and NPs supported on single crystal/single crystal thin films）, to explore the structure-sensitivity and establish surface structure–catalytic property relation46–48. Such an approach has been extensively used to investigate the structure-sensitivity of Au catalysis49–57. The chemical inertness of bulk Au makes activation of small molecules on Au single crystals difficult to be observed under ultra-high vacuum （UHV） conditions. CO did not chemisorb on closely-packed Au（111） at temperature as low as 78 K under UHV conditions but chemisorbed on more open Au single crystal surfaces such as Au（110） and Au（100）58–61. This demonstrates enhanced reactivity of low-coordinated Au atoms on Au single crystals with open surfaces. At elevated pressures CO could chemisorb on Au（111） even at 300 K62. However, no dissociative or molecular chemisorption of O2occurred on Au single crystal surfaces no matter how large the O2exposure is63. This makes investigations of oxidation reactions on Au single crystal surfaces employing O2unlikely. Alternatively, use of an atomic oxygen source64,65or ozone66,67, electron or photon irradiation of Au surfaces with physisorbed O2at 28 K68,69and electron bombardment of chemisorbed NO270were used to prepare atomic oxygen species including oxygen adatom and surface oxide. Reactivity of atomic oxygen species on Au single crystal surfaces toward CO51,54,65,71–74, H2O72,74–78, CO265,72,79, NOx80–82, NH383,84, propylene85, formic acid86, and alcohols72,75,87–94were studied. Atomic oxygen species reacted facilely with CO to produce CO2at low temperatures far below room temperature （RT）51,54,65,71–74and the CO2production was observed to follow the order of chemisorbed oxygen > oxygen in a surface oxide > oxygen in a bulk gold oxide51,54,73,74. These studies deepen the fundamental understanding of the reactivity of atomic oxygen species on Au surface, but fail to address the chemisorption and activation of molecular O2, the key step in the oxidation reactions catalyzed by supported Au catalysts.

Size-dependent structures and catalytic performance of supported Au catalysts have been approached employing Au NPs on oxide single crystals or single crystal thin films model catalysts. The critical size for the occurrence of the quantum size effect is a key issue for the size-dependent electronic structure of supported Au NPs. On rutile TiO2（110） Au NPs at a diameter of 3.2 nm and height of 1.0 nm exhibited a band gap of 0.2–0.6 eV and those finer than 2.0 nm exhibited a band gap of 1.4 eV31. On highly oriented pyrolitic graphite （HOPG） Au NPs at ～1 nm experienced the metal-to-nonmetal transition95. The effect of supported Au NPs size on the electronic structure was manifested by the generally-observed positive shift of the Au 4f core level binding energy of supported Au NPs as their size decreased96–102. This might indicate the occurrence of metal-tononmetal transition, but the contribution of final state effects to the core level binding energy shift is still unclear49. The morphology and geometric structure of supported Au NPs were also reported to vary with the size. A gold particle grown on a MgO（100） step was observed to rearrange continuously to maintain the lowest energy structure when held at room temperature103. Two-dimensional Au NPs on HOPG experienced smaller binding energy shifts as compared to more spherical NPs deposited on amorphous graphite with approximately the same total number of atoms154. Au 4f core level binding energy shifts of Au NPs supported on SiO2and TiO2were measured to be 1.6and 0.8 eV, respectively104. It should be noted that both the sizedependent structure of Au NPs and the size-dependent Au-support interaction can affect the structure of supported Au NPs. With this respect intrinsic size-dependent structures and properties of unsupported Au NPs were examined using small size-selective Au clusters52,106. Reactivity and catalytic performance of Au NPs supported on oxide single crystals or single crystal thinfilms were also explored and the size-dependence was observed. The size-dependent CO adsorption on Au NPs with various sizes supported on model surfaces of FeO, Fe3O4, Fe2O3, and Al2O3was observed and associated with increased percentages of highly uncoordinated Au atoms on small particles favoring CO adsorption, but supported Au NPs of a threshold diameter （3.0 nm） regardless of the substrate desorbed CO at the same maximum temperature107. The heats of CO adsorption on Au NPs supported on rutile TiO2（110） ranging in size from 1.8 to 3.1 nm increased sharply with decreasing particle size and the Au particle size corresponding to the largest heat of CO adsorption was very close to that exhibiting the maximum catalytic activity for CO oxidation108. However, similar to Au single crystals, supported Au NPs do not chemisorb molecular O2under UHV conditions. Supported Au NPs on rutile TiO2（110） exhibited catalytic activity in CO oxidation high pressure reaction cell with reactants' pressure up to Torr but the optimal size and the corresponding structure were debated. Valden et al.31firstly reported that most active supported Au NPs were 2.5–3.5 nm with two layers of gold, but a later simultaneous reactivity measurements and grazing incidence small-angle X-ray scattering characterizations revealed that most active supported Au NPs were （2.1 ± 0.3） nm with a height of about six atomic layers34.

Although model catalyst study employing Au single crystals and Au NPs supported on oxide single crystals or single crystal thin films have been extensively performed, the acquired fundamental understanding of structure-sensitivity of Au catalysis are somewhat diverse and even contradictory； moreover, several key issues in structure-sensitivity of Au catalysis, such as activation of molecular O2and reaction mechanisms, are not well understood. Au catalysis is very structure-sensitive and slightly different conditions in the preparation of supported Au model catalysts for the same system may result in quite different structures, chemisorption behaviors and catalytic properties； therefore, a comprehensive study of size, structure, chemisorption, and catalytic property of any given model system is preferred for unambiguous conclusions on Au catalysis, which, however, usually lacks in previous studies. The chemical inertness of bulk gold makes chemisorption and surface reactions occurring on the working supported Au catalysts difficult to be observed on single crystals-based Au model catalysts under UHV conditions. The contribution of “active” supports, for example the frequently-employed rutile TiO2（110） surface, whose extent remains uncertain adds additional complexity for fundamental understanding.

In order to overcome the so-called “pressure gap” and “materials gap” between surface chemistry and catalysis of single crystals-based model catalysts under UHV conditions and those of corresponding working catalysts, we have developed a novel model catalyst approach using both single crystals-based model catalysts and nanocrystals-based model catalysts for fundamental studies of the structure-property relation and reaction mechanism of working catalysts at the molecular level（Fig.1）109–113. Our model catalyst approach has been applied to investigate the structure-sensitivity of Au catalysis. Two groups of model catalysts for Au catalysis are adopted. One group aims to study the structure-sensitivity of intrinsic Au catalysis and includes Au single crystals, Au nanocrystals, and Au/SiO2catalysts. SiO2is an “inert” support and the observed catalytic activity of Au/SiO2catalysts can be associated with the structure of supported Au NPs. The other group aims to study the structure-sensitivity of supports-involved Au catalysis （Au-support interaction and support-participated catalysis） and includes Au NPs supported on oxide single crystals/single crystal thin films, oxide NPs supported on Au single crystals, Au NPs supported on oxide nanocrystals, oxide NPs supported on Au nanocrystals, and Au/oxide catalysts. We herein review the acquired progress so far and more works are still undergoing.

3 Structure-sensitivity of intrinsic Au catalysis

Low-coordinated Au atoms at the corners, steps, and edges generally increases with the size of Au NPs decreasing and exhibit enhanced reactivity and catalytic activity. Low-coordinated metal atoms can be modelled using stepped metal single crystals. In our model catalyst approach we choose stepped Au（997） and Au（110）-（1 × 2） single crystals to model low-coordinated Au atoms. As schematically illustrated and proved by low energy electron diffraction （LEED） patterns （Fig.2）, Au（997） vicinal surface is composed of close-packed （111） terraces and mono-atomic （111） steps respectively exhibiting ninefold-coordinated and sevenfold-coordinated Au atoms whereas Au（110）-（1 × 2） surface gives rise to on top of row atoms and side of row atoms with coordination numbers of 7 and 9, respectively. Therefore, Au（997） surface offers a model surface to compare the reactivity and catalytic activity of Au atoms with the same coordination environment but different coordination numbers, and Au（997） and Au（110）-（1 × 2） surfaces offer model surfaces to compare the reactivity and catalytic activity of Au atoms with the same coordination number but different coordination environments.

Fig.1 Model catalysts ranging from single crystals to uniform nanocrystals for fundamental studies of the structure-property relation and reaction mechanism of working catalysts at the molecular level

Sevenfold-coordinated Au atoms on the step of Au（997） ex-hibited stronger chemisorption ability toward CO, NO, and NO2than 9-coordinated Au atoms on the terrace of Au（997）, proving the enhanced reactivity of low-coordinated Au atoms114–116. Low-temperature NO decomposition was observed to occur on Au（997） and Au（110）-（1 × 2） and sensitively depend on the local structure of surface Au atoms （Fig.3）114. Following the saturating NO exposure 105 K, a similar NO desorption peak at 144 K appeared for both surfaces and was attributed to the desorption of species chemisorbed on sevenfold-coordinated Au atoms, and additional desorption peaks at 123 K with a shoulder at 116 K appeared for Au（110）-（1 × 2） surface. Desorption peaks of N2O and N2associated with the lower-temperature NO desorption peaks, the products of NO decomposition, were observed for Au（110）-（1 × 2） surface. Thus, NO adsorbed at surface Au sites other than sevenfold-coordinated Au atoms on the Au（110）-（1 × 2） surface, instead of more thermally stable NO adsorbed at sevenfold-coordinated Au atoms, is the active surface species to decompose. DFT calculation results suggested that chemisorbed （NO）2species and their decomposition dominated the surface chemistry of NO on the Au（997） and Au（110）-（1 × 2） surfaces. Different （NO）2species were identified, including （NO）2bonded to sevenfold-coordinated Au atoms via the N atom with the largest adsorption energy, （NO）2and （NO）2respectively bonded to the trench Au atoms of Au（110）-（1 × 2） surface via the N atom and the O and N atoms, and （NO）2bonded to both step and terrace Au atoms of Au（997） surface via the O and N atoms whose formation got suppressed due to its competition for the step sites with more stable （NO）2. The decomposition reactivity of various （NO）2species into N2O were calculated to follow the order opposite to their adsorption energy. The more stable （NO）2species bonded to sevenfold-coordinated Au atoms exhibits a much higher activation barrier for the decomposition reaction than other （NO）2species. Thus DFT calculation results agreed well with experimental results to reveal higher decomposition reactivity into N2O of （NO）2species with smaller adsorption energies and the structure-sensitivity of NO decomposition on Au surfaces. This sheds light on the origin of high activity of supported Au NPs at mild reaction temperatures. The coverage of a surface species on the catalyst surface is determined by the reactant pressure, the reaction temperature and the adsorption energy. The reactivity of all types of surface species contribute to the observed catalytic activity, and a catalyst on which surface species with small adsorption energies exhibit high reactivity will be active at low reaction temperatures under which they can be stable and get accumulated on the catalyst surface.

Fig.2 Schematic illustrations and LEED patterns of surface structures of Au（997） and Au（110）-（1 × 2） single crystals114

Fig.3 NO， N2O， and N2TDS spectra after saturating NO exposure on Au（997） and Au（110）-（1 × 2） at 105 K114

O2do not chemisorb on Au（997） under UHV conditions. We used the decomposition of amorphous N2O4multilayers to prepare site-selective atomic oxygen-covered Au（997） surfaces, including 0.02 ML oxygen adatoms （O（a） at the （111） steps, 0.12 ML O（a） at both （111） steps and terraces, and 0.26 ML O（a） at both （111） steps and terraces and oxygen islands at the （111）terraces, and studied their reactivity toward CO, NO, and H2O115–118. The thermal stability of different types of atomic oxygen species follows the order of O（a） on the （111） steps > O（a）adatoms on the （111） terraces > O（a） islands on the （111） terraces. All these atomic oxygen species facilely reacted with CO, NO, and H2O respectively to produce chemisorbed CO2（a）, NO2（a） and hydroxyl groups, and, similar to the case of NO decomposition, less thermally stable atomic oxygen species were also observed exhibiting enhanced reactivity. The desorption of CO2（a）/NO2（a） from the Au surface that occurred at higher temperatures than the surface reaction of atomic oxygen species with CO/NO was the rate-determining step respectively for the oxidation of CO and NO by atomic oxygen species. The CO2（a）desorption exhibited multi-channels that depended on the Au structure and was strongly affected by the co-adsorbates；moreover, the desorption of CO2（a） formed by O（a） + CO reaction to produce gaseous CO2at 105 K sensitively depended on the surface reaction extent （Fig.4）116. No gaseous CO2production was observed during the exposure of 0.05 L CO on 0.02 ML oxygen-covered Au（997） surface at 105 K whereas obvious gaseous CO2production occurred with the prolonging CO exposure under the same condition （0.5 L CO）. Accordingly, a CO2desorption peak emerged at 120 K in the subsequent CO2thermal desorption spectroscopy （TDS） spectrum following 0.05 L CO exposure while it was indistinguishable in the CO2TDS spectrum following 0.5 L CO exposure. Thus both thegaseous CO2produced upon CO exposure at 105 K and the CO2desorption peak at 120 K in the subsequent CO2TDS spectrum arose from CO2（a） formed at the bare （111） steps due to the locally complete consumption of pre-covered oxygen adatoms by CO at some （111） step sites. The CO exposure-dependent CO2production at 105 K, i.e., the surface reaction extent-dependent CO2production at 105 K, suggested that the increased reaction heat exerted by the prolonged CO + O（a） reaction upon the exposure of 0.5 L CO could locally drive the desorption of formed CO2（a） to produce gaseous CO2.

Fig.4 （A） CO2and CO QMS signals during 0.05 and 0.5 L CO exposures at a pressure of 5 × 10-7Pa on 0.02 ML oxygen-covered Au（997） surface at 105 K and （B） subsequent CO2TDS spectra following the exposures116

Fig.5 （A） Supported Au NPs size distributions， （B） Au 4f XPS spectra， （C） Au LIII-edge XAFS spectra， （D） DRIFTS spectra of CO adsorption at RT and （E） catalytic activity in CO oxidation of various Au/SiO2catalysts with the same 1.8%（w） Au loading123

Although our results on stepped Au single crystal model catalysts acquired so far have deepen the fundamental understanding of structure-sensitivity and elementary surface reactions of intrinsic Au catalysis, no insights were obtained on the activation of molecular oxygen and the relevant oxidation reactions on Au surfaces. We thus used SiO2-supported Au NPs as model catalysts to study structure-sensitivity of intrinsic Au catalysis in CO oxidation119–123. By calcining in air or reducing in H2the catalyst precursor at different temperatures, we acquired Au/SiO2catalysts with the same Au loading （1.8% （w）） but different Au NPs size distributions （Fig.5A）, including Au/SiO2-air-300 calcined in air at 300 °C with supported Au NPs mostly larger than 4.5 nm, Au/SiO2-H2-300 reduced in H2at 300 °C with ～23% supported Au NPs larger than 4.5 nm, ～70% supported Au NPs of 3–4.5 nm and ～7% supported Au NPs of 2–3 nm, Au/SiO2-H2-200 reduced in H2at 200 °C with ～24% supported Au NPs larger than 4.5 nm, ～54% supported Au NPs of 3–4.5 nm, and ～22% supported Au NPs of 2–3 nm, Au/SiO2-H2-120 reduced in H2at 120 °C with ～35.6% supported Au NPs larger than 4.5 nm, ～39% supported Au NPs of 3–4.5 nm,～23% supported Au NPs of 2–3 nm, and ～2.4% supported Au NPs finer than 2 nm123. Correlating with the Au 4f XPS and Au LIII-edge XANES results （Fig.5（B, C））, we found that Au NPs supported on SiO2larger than 3 nm exhibited the bulk Au-likeelectronic structure whereas those finer than 3 nm exhibited electronic structures deviating from that of bulk Au more or less. Thus Au NPs supported on SiO2exhibited the size-dependent electronic structures and the quantum size effect with～3 nm as the critical size. Similar results were also observed for Au colloids with different sizes121. CO adsorbed at low-coordinated Au atoms of supported Au NPs in Au/SiO2catalysts was observed at RT by in-situ diffuse reflectance infrared Fourier transform spectroscopy （DRIFTS spectra （Fig.5D） and their amounts followed the order of Au/SiO2-H2-300 > Au/SiO2-H2-200 > Au/SiO2-H2-120 >> Au/SiO2-Air-300. The catalytic activity of Au/SiO2catalysts in CO oxidation at RT followed the same order （Fig.5E） since CO must adsorb on Au NPs for catalytic CO oxidation. These observations indicated that Au NPs in Au/SiO2-Air-300 exposed few low-coordinated Au atoms due to their large size and likely smooth surfaces. The amount of adsorbed CO in Au/SiO2-H2increased with the reduction temperature, in consistence with the fractions of supported Au NPs of 3–4.5 nm but in contrast to the general assumption that finer Au NPs expose higher density of low-coordinated Au atoms because the population of fine Au NPs in various Au/SiO2-H2increased with the decrease of reduction temperature. This inferred that Au NPs finer than 3 nm with high density of low-coordinated Au atoms in Au/SiO2-H2should not actively chemisorb CO, which could be reasonably associated with their electronic structure deviating from bulk-Au electronic structure.

Fig.6 Schematic illustration of the size-structure-chemisorptioncatalytic activity relation and catalytically active structure of Au NPs supported on SiO2in low-temperature CO oxidation123

Our results of Au/SiO2provide novel insights into understanding the structure-sensitivity of intrinsic Au catalysis in low-temperature CO oxidation from the view of CO adsorption. As schematically illustrated in Fig.6, Au NPs of 3–4.5 nm with the bulk Au-like electronic structure supported on SiO2expose abundant low-coordinated Au atoms and effectively chemisorb CO and catalyze CO oxidation at RT； Larger supported Au NPs expose few low-coordinated Au atoms and are inactive in chemisorbing CO and catalyzing CO oxidation at RT； Finer supported Au NPs with electronic structures deviating from that of bulk Au are also not able to chemisorb CO and catalyze CO oxidation at RT although they expose more abundant low-coordinated Au atoms. Therefore, low-coordinated Au atoms on Au NPs with bulk Au-like electronic structure are active in adsorb CO and catalyze CO oxidation at RT. This is also supported by our results that nitrate anion could induce the surface restructuring of SiO2-supported Au NPs larger than 4.5 nm to create surface low-coordinated Au atoms active in catalyze CO oxidation at RT120. With the decrease of the size of Au NPs, the density of low-coordinated Au atoms generally increases while the electronic structure will eventually deviates from bulk Aulike electronic structure, therefore, the intrinsic catalytic activity of Au NPs in catalyzing low-temperature CO oxidation will the inevitably exhibit a volcano-shaped dependence on their size.

We also used Au/SiO2catalysts to prove hydroxyl group-induced activation of O2for CO oxidation at RT on inert Au surface122. NaOH additive substantially enhanced the catalytic activity of Au/SiO2with inactive large supported Au NPs in catalyzing CO oxidation at RT without changing the particle size distributions and electronic structure of supported Au NPs. The accompanying DFT calculations demonstrated two likely reaction mechanisms for CO oxidation on OH（a）-covered Au（111） （Fig.7） in which the key elementary step was the reaction of CO（a） with OH（a） to form COOH（a） with an activation energy of 0.44 eV. In mechanism （a）, the elementary step with the highest activation energy was the decomposition of COOH（a） into CO2and H（a） with an activation energy of 0.98 eV and a reaction enthalpy of –0.54 eV. In mechanism （b）, the elementary step with the highest activation energy was the reaction of COOH（a） with O2to form OOCOOH（a） with a hydrogen-bonded five-membered ring structure with an activation energy of 0.94 eV and a reaction enthalpy of –0.81 eV.

Fig.7 Two likely mechanisms for hydroxyl group-induced CO oxidation on Au（111） surface

4 Structure-sensitivity of supports-involved Au catalysis

“Active” supports can be involved in Au catalysis via strong Au-support interaction to affect the local structure of gold or direct participation into catalytic reactions. The structure of support strongly affects the structure-sensitivity of supports-involved Au catalysis. TiO2is a typical “active” support. We employed Au NPs physical vapor deposited on stoichiometric and reduced rutile-TiO2（110） surfaces as model catalysts to study the Au-TiO2interaction124. As shown in Fig.8A, resulted fromthe size effect, the 4f7/2XPS peak of supported Au clusters on stoichiometric rutile-TiO2（110） surface shifted monotonically toward lower binding energy with the increase of the Au coverage and reaches the binding energy value of the bulk Au at a deposited thickness of 1 nm； On reduced rutile-TiO2（110） surface, a shift of the Au 4f7/2peak toward higher binding energy was observed during its shifts toward lower binding energy with the increase of the Au coverage. This indicated the preferential nucleation of Au clusters on the oxygen vacancies of reduced rutile-TiO2（110） surface and the occurring of charge transfer from reduced rutile-TiO2（110） surface to supported Au clusters. As schematically illustrated in Fig.8B, the charge transfer from the oxygen vacancy sites to Au clusters decreases the Au 4f binding energy, which compensates the size effect on the Au 4f binding energy. Therefore, at the lowest thickness, Au clusters on reduced rutile-TiO2（110） surface exhibit a lower Au 4f binding energy than those on stoichiometric surface. The charge transfer effect depends on the size. A critical size should exist below which the charge transfer effect can effectively compensate the size effect but beyond which the charge transfer effect can not. When the Au cluster size is increased across the critical size, the Au 4f binding energy will increase rather than decrease. Our experimental observation of the upward binding energy shift of Au clusters on reduced rutile-TiO2（110）surface with the increase of their size provided the solid proof for such a model. The charge transfer from the oxygen vacancies of reduced rutile-TiO2（110） surface to Au clusters not only modified the electronic structure but also enhanced the thermal stability of Au clusters125. The thermal stability of Au clusters on reduced rutile-TiO2（110） surface was much stronger than that of Au clusters with the same thickness on stoichiometric rutile-TiO2（110） surface.

Fig.8 （A） Au 4f7/2binding energy of Au clusters deposited on rutile TiO2（110） surfaces annealed at 873 K （stoichiometric surface） and 1173 K （reduced surface） as a function of Au thickness. （B） Schematic illustration of the size effect and size-dependent charge transfer effect on the Au 4f7/2binding energy of Au clusters

Fig.9 In-situ DRIFTS spectra of CO adsorption on TiO2and Au/TiO2samples at 120 K with p = 250 Pa126

We also employed anatase TiO2nanocrystals predominantly enclosed the ｛001｝ facets, anatase TiO2nanocrystals predominantly enclosed the ｛100｝ facets, and P25 predominantly enclosed the ｛101｝ facets as the supports to study the effects of TiO2surface structure on the Au-TiO2interaction in powder samples126. Au NPs in Au/P25-｛101｝, Au/TiO2-｛100｝, Au/TiO2-｛001｝ exhibited size distributions of （3.1 ± 0.6） nm, （2.7 ± 0.6）nm, and （1.4 ± 0.6） nm & （4.6 ± 1.4） nm, respectively. XPS and Au LIII-edge XANES results demonstrated that all supported Au NPs showed the bulk Au-like electronic structure. This indicated the presence of charge transfer from TiO2support to supported Au NPs because Au NPs finer than ～3 nm supported on SiO2without charge transfter exhibited electronic structures different from the bulk Au. CO adsorption was used to probe the structures of Au/TiO2catalysts （Fig.9）. All supported Au NPs adsorbed CO but different charging states were observed. The Auδ–species corresponding to CO stretch vibrational bands of 2081–2089 cm–1was the largest in Au/TiO2-｛001｝ due to the creation of surface oxygen vacancies of TiO2-｛001｝ upon Auloading whereas the fraction of Auδ+species corresponding to CO stretch vibrational bands of 2099–2105 cm–1was the largest in Au/TiO2-｛100｝ due to the preserved surface stoichiometric of TiO2-｛100｝ upon Au loading. Auδ+species is most active than Auδ–species in catalyzing H2oxidation but Auδ–species is more selective than Auδ+species toward H2O2while is most selective toward H2O2. The ensemble consisting of intimatelycontacting Auδ–and Ti4+with weak adsorption ability is the active structure for C3H6epoxidation with O2and H2and Au/TiO2-｛001｝ catalyst containing the largest amount of active Auδ–-Ti4+ensemble is most active. Thus, TiO2facets strongly affect the Au-TiO2interaction, offering an effective strategy of TiO2morphology/facet-engineering to understand fundamental catalysis and optimize catalytic performances of TiO2-involved catalysts.

Fig.10 Representative TEM images of Au/CoOx/SiO2， Au/ZnO/SiO2， and Au/CeO2/SiO2catalysts127-129

Fig.11 Variations of the normalized vibrational peak intensities of major surface species formed on 0.18%-Au/CeO2of （1.7 ± 0.6） nm Au NPs， 0.96%-Au/CeO2of （2.6 ± 0.6） nm Au NPs， and 5.7%-Au/CeO2of （3.7 ± 0.9） nm Au NPs upon saturating CO adsorption at RT as a function of time during the initial purging processes in Ar and

Au/MOx/SiO2catalysts with both Au and “active” oxide supported on “inert” SiO2were also used as model catalysts to study Au-“active” oxide support interaction in powder catalysts, and their advantage was that the loadings of supported Au and MOxcould be varied to manifest the Au-“active” oxide support interaction45,127–129. In all investigated catalysts including Au/CoOx/SiO2, Au/CeO2/SiO2, and Au/ZnO/SiO2, Au NPs supported on MOxwere much finer and more active in catalyzing low-temperature CO oxidation than those supported on SiO2（Fig.10）, directly proving the much stronger Au-MOxinteraction than the Au-SiO2interaction. Several factors were identified to play important roles in the structure-sensitivity of supports-involved Au catalysis. Firstly, the types of hydroxyl groups on MOxaffected the nucleation and growth of suppor-ted Au NPs45. The Au precursor preferentially interacted with hydrogen-bonded hydroxyls in Co（OH）2supported on SiO2, forming clusters of gold precursors on the support surface and eventually large Au aggregates in the catalyst, and then interacted with isolated hydroxyls in Co（OH）2, forming isolated gold precursors on the support surface and eventually fine Au NPs in the catalyst. Secondly, Au（I） species and Au NPs competed for surface oxygen vacancies on CeO2, and highly dispersive CeO2on SiO2facilitated the formation of Au（I） species； meanwhile, Au species also facilitated the creation and stabilization of surface oxygen vacancies on CeO2. Au（I） species on CeO2alone was not active in CO oxidation and Au NPs in contact with CeO2in Au/CeO2/SiO2catalysts were active129. Thirdly, the deposition-precipitation （DP） agent affected the Au-MOxinteraction128. Au/ZnO/SiO2catalyst prepared employing Na2CO3aqueous solution as the DP agent exhibited stronger Au-ZnO interaction than that prepared employed ammonium hydroxide as the DP agent, resulting the larger Au（I） : Au（0） ratio, finer sizes of supported Au NPs, and higher catalytic activity in lowtemperature CO oxidation. Supported Au species was identified to interact with oxygen vacancies and grain boundary defects of ZnO.

Au/CeO2is a representative supported Au catalysts for lowtemperature CO oxidation and CeO2both strongly interacts with supported Au species and directly participates the catalytic reaction. Au/CeO2catalysts with Au NPs sizes ranging from（1.7 ± 0.6） to （3.7 ± 0.9） nm were prepared and comprehensively investigated130. Similar to Au/TiO2catalysts126, charge transfer from CeO2support to supported Au NPs occurred, modifying the electronic structures of supported Au NPs on CeO2to be independent of the size and similar to bulk Au. Employing time-resolved Operando-DRIFTS to monitor evolutions of various surface species during CO oxidation at RT, we revealed the size-dependent reaction pathways and their contributions to the catalytic activity. As shown in Fig.11, the intrinsic oxidation reactivity of CO（a） does not depend much on the Au particle size whereas the intrinsic oxygen-assisted decomposition reactivity of carbonate, bicarbonate and formate species strongly depend on the Au particle size and are facilitated over Au/CeO2catalysts with large Au particles. It is likely that larger Au particles are more capable of activating surface lattice oxygen on CeO2to participate in CO oxidation via surface intermediates of carbonate, bicarbonate, and formate. The size effects of supported Au NPs on CO（a） oxidation and oxygenassisted decomposition of carbonate, bicarbonate, and formate species are to affect the specific density of surface Au adsorption sites for CO（a） and open the decomposition reaction pathways at large supported Au NPs, respectively. These results fundamentally understand the structure-sensitivity of Au/CeO2in low-temperature CO oxidation.

5 Summary and outlook

Model catalyst approach is effective for the fundamental investigations of complex and structure-sensitive Au catalysis, but has to be comprehensive. Our strategy of model catalysts from single crystals to, nanocrystals and well-defined powder catalysts can achieve this goal, and so far we have provide solid experimental evidence for several important issues in the structure-sensitivity of Au catalysis: （1） The intrinsic active Au structure in adsorbing CO and catalyzing low-temperature CO oxidation without the involvement of supports is low-coordinated Au atoms of Au NPs with bulk Au-like electronic structure. The density of low-coordinated Au atoms generally increases with the decrease of the Au NPs size while the electronic structure eventually deviates from bulk Au-like electronic structure, therefore, the intrinsic catalytic activity of Au NPs will inevitably exhibit the volcano-shaped dependence on their size. （2） Au-“active” oxide interaction sensitively depends on the microstructure of “active” oxide during the catalyst preparation. Charge transfer occurs from “active” oxides to supported Au NPs, stabilizing fine Au NPs and modifying their electronic structures to be independent of the size and similar to bulk Au. This conveys low-coordinated Au atoms of Au NPs supported on “active” oxides always active in adsorbing CO and catalyzing low-temperature CO oxidation. （3） Multi surface reaction pathways contribute to the catalytic activity of Au/CeO2in lowtemperature CO oxidation and exhibit different structure-sensitivities. The intrinsic oxidation reactivity of CO（a） does not depend much on the Au particle size and the size effect of supported Au NPs is to affect the specific density of surface Au adsorption sites for CO（a） while the intrinsic oxygen-assisted decomposition reactivity of carbonate, bicarbonate and formate species strongly depend on the Au particle size and the size effect of supported Au NPs is to open the decomposition reaction pathways at large supported Au NPs. （4） The reactivity of surface Au atoms toward reactants enhances with the decrease of the coordination number. Weakly-adsorbed but reactive surface species form on Au surface. This makes Au surfaces catalytically active at low temperatures under which weakly-adsorbed but reactive species can effectively accumulate on Au surface and contribute to the catalytic activity. （5） Hydroxyls on inert Au surface can act as co-catalyst to catalyze low-temperature CO oxidation in which COOH（a） species formed by CO + OH reaction can activate molecular O2.

We are still working on model catalyst approach to fundamentally understand the structure-sensitivity of Au catalysis and the efforts are devoted to: （1） Fabrications of model catalysts for supported Au catalysts with precise structural control of both Au NPs and supports. （2） In-situ and Operando spectroscopic and microscopic studies of Au catalysis under working condition to identify the active structure and the active surface species, particularly active oxygen species on supported Au catalysts for oxidation reactions. （3） Quantification of contributions of different reaction pathways to the catalytic activity of Au catalysts. We hope that, via our comprehensive studies, Au catalysis will become a novel and typical system to demon-strate both the classic concept of structure-sensitivity in heterogeneous catalysis and the novel concept of model catalysts approach using both single crystals-based model catalysts and nanocrystals-based model catalysts for fundamental studies of heterogeneous catalysis at the molecular level.

（1）Hutchings, G. J. J. Catal. 1985， 96, 292. doi: 10.1016/0021-9517（85）90383-5

（2）Nkosi, B.； Coville, N. J.； Hutchings, G. J. Appl. Catal. 1988， 43, 33. doi: 10.1016/S0166-9834（00）80898-2

（3）Haruta, M.； Kobayashi, T.； Sano, H.； Yamada, N. Chem. Lett. 1987， 16, 405.

（4）Haruta, M.； Yamada, N.； Kobayashi, T.； Iijima, S. J. Catal. 1989， 115, 301. doi: 10.1016/0021-9517（89）90034-1

（5）Bond, G. C.； Thompson, D. T. Cat. Rev. Sci. Eng. 1999， 41, 319. doi: 10.1081/CR-100101171

（6）Haruta, M.； Daté, M. Appl. Catal. A: Gen. 2001， 222, 427. doi: 10.1016/S0926-860X（01）00847-X

（7）Hashmi, A. S. K.； Hutchings, G. J. Angew. Chem. Int. Edit. 2006， 45, 7896.

（8）Takei, T.； Akita, T.； Nakamura, I.； Fujitani, T.； Okumura, M.；Okazaki, K.； Huang, J.； Ishida, T.； Haruta, M. Adv. Catal. 2012，55, 1.

（9）Ide, M. S.； Davis, R. J. Accounts Chem. Res. 2014， 47, 825. doi: 10.1021/ar4001907

（10）Ojeda, M.； Iglesia, E. Chem. Commun. 2009， No. 3, 352.

（11）Lee, S.； Molina, L. M.； López, M.； Alonso, J. A.； Hammer, B.；Lee, B.； Seifert, S.； Winans, R. E.； Elam, J. W.； Pellin, M. J.；Vajda, S. Angew. Chem. Int. Edit. 2009， 48, 1467. doi: 10.1002/anie.v48:8

（12）Huang, J.； Akita, T.； Faye, J.； Fujitani, T.； Takei, T.； Haruta, M. Angew. Chem. Int. Edit. 2009， 48, 7862. doi: 10.1002/anie.v48:42

（13）Fujitani, T.； Nakamura, I. Angew. Chem. Int. Edit. 2011， 50, 10144. doi: 10.1002/anie.201104694

（14）Schubert, M. M.； Hackenberg, S.； van Veen, A. C.； Muhler, M.；Plzak, V.； Behm, R. J. J. Catal. 2001， 197, 113.

（15）Daté, M.； Okumura, M.； Tsubota, S.； Haruta, M. Angew. Chem. Int. Edit. 2004， 43, 2129.

（16）Costello, C. K.； Kung, M. C.； Oh, H. S.； Wang, Y.； Kung, H. H. Appl. Catal. A: Gen. 2002， 232, 159. doi: 10.1016/S0926-860X（02）00092-3

（17）Costello, C. K.； Yang, J. H.； Law, H. Y.； Wang, Y.； Lin, J. N.；Marks, L. D.； Kung, M. C.； Kung, H. H. Appl. Catal. A 2003，243, 15. doi: 10.1016/S0926-860X（02）00533-1

（18）Kung, H. H.； Kung, M. C.； Costello, C. K. J. Catal. 2003， 216, 425. doi: 10.1016/S0021-9517（02）00111-2

（19）Sanchez-Castillo, M. A.； Couto, C.； Kim, W. B.； Dumesic, J. A. Angew. Chem. Int. Edit. 2004， 43, 1140.

（20）Saavedra, J.； Doan, H. A.； Pursell, C. J.； Grabow, L. C.；Chandler, B. D. Science 2014， 345, 1599. doi: 10.1126/science.1256018

（21）Anderson, J. R.； Shlmoyama, Y. In Proc. 5th Int. Congr. Catal.；Hightower, J. W. Ed.； Elsevier B. V.: North-Holland, Amsterdam, 1973； p 965.

（22）Boudart, M.； Aldag, A. W.； Ptak, L. D.； Benson, J. E. J. Catal. 1968， 17, 35.

（23）Boudart, M.； Aldag, A. W.； Benson, J. E.； Dougharty, N. A.；Harklns, G. C. J. Catal. 1966， 6, 92. doi: 10.1016/0021-9517（66）90113-8

（24）Dorllng, T. A.； Moss, R. L. J. Catal. 1966， 5, 111. doi: 10.1016/S0021-9517（66）80130-6

（25）Boudart. M. Adv. Catal. 1969， 20, 153.

（26）Volta. J. C.； Portefaix, J. L. Appl. Catal. 1985， 18, 1. doi: 10.1016/S0166-9834（00）80296-1

（27）Carberry, J. J. J. Catal. 1987， 107, 248. doi: 10.1016/0021-9517（87）90291-0

（28）Carberry, J. J. J. Catal. 1988， 114, 277. doi: 10.1016/0021-9517（88）90031-0

（29）Haruta, M.； Tsubota, S.； Kobayashi, T.； Kageyama, H.； Genet, M. J.； Delmon, B. J. Catal. 1993， 144, 175. doi: 10.1006/jcat.1993.1322

（30）Bamwenda, G. R.； Tsubota, S.； Nakamura, T.； Haruta, M. Catal. Lett. 1997， 44, 83. doi: 10.1023/A:1018925008633

（31）Valden, M.； Lai, X.； Goodman, D. W. Science 1998， 281, 1647. doi: 10.1126/science.281.5383.1647

（32）Zanella, R.； Giorgio, S.； Shin, C. H.； Henry, C. R.； Louis, C. J. Catal. 2004， 222, 357. doi: 10.1016/j.jcat.2003.11.005

（33）Tai, Y.； Yamaguchi, W.； Tajiri, K.； Kageyama, H. Appl. Catal. A: Gen. 2009， 364, 143. doi: 10.1016/j.apcata.2009.05.041

（34）Laoufi, I.； Saint-Lager, M. C.； Lazzari, R.； Jupille, J.； Robach, O.； Garaudée, S.； Cabailh, G.； Dolle, P.； Cruguel, H.； Bailly, A. J. Phys. Chem. C 2011， 115, 4673. doi: 10.1021/jp1110554

（35）Haruta, M. Chem. Rec. 2003， 3, 75.

（36）Green, I. X.； Tang, W.； Neurock, M.； Yates, J. T., Jr. Science 2011， 333, 736. doi: 10.1126/science.1207272

（37）Lopez, N.； Janssens, T. V. W.； Clausen, B. S.； Xu, Y.；Mavrikakis, M.； Bligaard, T.； Nørskov, J. K. J. Catal. 2004，223, 232. doi: 10.1016/j.jcat.2004.01.001

（38）Janssens, T. V. W.； Clausen, B. S.； Hvolbæk, B.； Falsig, H.；Christensen, C. H.； Bligaard, T.； Nørskov, J. K. Topic Catal. 2007， 44, 15. doi: 10.1007/s11244-007-0335-3

（39）Falsig, H.； Hvolbæk, B.； Kristensen, I. S.； Jiang, T.； Bligaard, T.； Christensen, C. H.； Nørskov, J. K. Angew. Chem. Int. Edit. 2008， 47, 4835.

（40）Maeda, Y.； Okumura, M.； Tsubota, S.； Kohyama, M.； Haruta, M. Appl. Surf. Sci. 2004， 222, 409. doi: 10.1016/j.apsusc.2003.09.007

（41）Chen, M. S.； Goodman, D. W. Science 2004， 306, 252.

（42）Herzing, A. A.； Kiely, C. J.； Carley, A. F.； Landon, P.；Hutchings, G. J. Science 2008， 321, 1331. doi: 10.1126/science.1159639

（43）Liu, Y.； Jia, C. J.； Yamasaki, J.； Terasaki, O.； Schüth, F. Angew.Chem. Int. Edit. 2010， 49, 5771. doi: 10.1002/anie.v49:33

（44）Haruta, M. Catal. Today 1997， 36, 153. doi: 10.1016/S0920-5861（96）00208-8

（45）Qian, K.； Fang, J.； Huang, W. X.； He, B.； Jiang, Z. Q.； Ma, Y. S.； Wei, S. Q. J. Mol. Catal. A: Chem. 2010， 320, 97. doi: 10.1016/j.molcata.2010.01.010

（46）Somorjai, G. A.； Carrazza, J. Ind. Eng. Chem. Fundam. 1986，25, 63. doi: 10.1021/i100021a009

（47）Parker, I. B.； Waugh, K. C.； Bowker, M. J. Catal. 1988， 114, 457. doi: 10.1016/0021-9517（88）90049-8

（48）Bennett, C. O.； Che, M. J. Catal. 1989， 120, 293. doi: 10.1016/0021-9517（89）90270-4

（49）Meyer, R.； Lemire, C.； Shaikhutdinov, S. K.； Freund, H. J. Gold Bull. 2004， 37, 72. doi: 10.1007/BF03215519

（50）Chen, M. S.； Goodman, D. W. Accounts Chem. Res. 2006， 39, 739. doi: 10.1021/ar040309d

（51）Min, B. K.； Friend, C. M. Chem. Rev. 2007， 107, 2709. doi: 10.1021/cr050954d

（52）Risse, T.； Shaikhutdinov, S.； Nilius, N.； Sterrer, M.； Freund, H. J. Accounts Chem. Res. 2008， 41, 949. doi: 10.1021/ar800078m

（53）Chen, M. S.； Goodman, D. W. Chem. Soc. Rev. 2008， 37, 1860. doi: 10.1039/b707318f

（54）Gong, J. L.； Mullins, C. B. Accounts Chem. Res. 2009， 42, 1063. doi: 10.1021/ar8002706

（55）Carabineiro, S. A. C.； Nieuwenhuys, B. E. Gold Bull. 2010， 43, 252. doi: 10.1007/BF03214995

（56）Gong, J. L. Chem. Rev. 2012， 112, 2987. doi: 10.1021/cr200041p

（57）Stowers, K. J.； Madix, R. J.； Friend, C. M. J. Catal. 2013， 308, 131. doi: 10.1016/j.jcat.2013.05.033

（58）Outka, D. A.； Madix, R. J. Surf. Sci. 1987， 179, 351. doi: 10.1016/0039-6028（87）90062-8

（59）Gottfried, J. M.； Schmidt, K. J.； Schroeder, S. L. M.；Christmann, K. Surf. Sci. 2003， 536, 206. doi: 10.1016/S0039-6028（03）00595-8

（60）Meier, D. C.； Bukhtiyarov, V.； Goodman, D.W. J. Phys. Chem. B 2003， 107, 12668. doi: 10.1021/jp030499u

（61）McElhiney, G.； Pritchard, J. Surf. Sci. 1976， 60, 397. doi: 10.1016/0039-6028（76）90324-1

（62）Peters, K. F.； Steadman, P.； Isern, H.； Alvarez, J.； Ferrer, S. Surf. Sci. 2000， 467, 10. doi: 10.1016/S0039-6028（00）00777-9

（63）Olsen, C. W.； Masel, R. I. Surf. Sci. 1988， 201, 444. doi: 10.1016/0039-6028（88）90496-7

（64）Canning, N. D. S.； Outka, D.； Madix, R. J. Surf. Sci. 1984， 141, 240. doi: 10.1016/0039-6028（84）90209-7

（65）Outka, D. A.； Madix, R. J. Surf. Sci. 1987， 179, 351. doi: 10.1016/0039-6028（87）90062-8

（66）Saliba, N.； Parker, D. H.； Koel, B. E. Surf. Sci. 1998， 410, 270. doi: 10.1016/S0039-6028（98）00309-4

（67）Parker, D. H.； Koel, B. E. J. Vac. Sci. Tech. A 1990， 8, 2585. doi: 10.1116/1.576675

（68）Gottfried, J. M.； Schmidt, K. J.； Schroeder, S. L. M.； Christmann, K. Surf. Sci. 2003， 525, 184. doi: 10.1016/S0039-6028（02）02560-8

（69）Gottfried, J. M.； Schmidt, K. J.； Schroeder, S. L. M.；Christmann, K. Surf. Sci. 2002， 511, 65. doi: 10.1016/S0039-6028（02）01555-8

（70）Deng, X.； Min, B. K.； Guloy, A.； Friend, C. M. J. Am. Chem. Soc. 2005， 127, 9267. doi: 10.1021/ja050144j

（71）Gottfried, J. M.； Christmann, K. Surf. Sci. 2004， 566–568, 1112. doi: 10.1016/j.susc.2004.06.067

（72）Lazaga, M. A.； Wickham, D. T.； Parker, D. H.； Kastanas, G. N.；Koel, B. E. ACS Symp. Ser. 1993， 523, 90. doi: 10.1021/symposium

（73）Min, B. K.； Alemozafar, A. R.； Pinnaduwage, D.； Deng, X.；Friend, C. M. J. Phys. Chem. B 2006， 110, 19833. doi: 10.1021/jp0616213

（74）Gong, J. L.； Ojifinni, R. A.； Kim, T. S.； Stiehl, J. D.； McClure, S. M.； White, J. M.； Mullins, C. B. Topic Catal. 2007， 44, 57. doi: 10.1007/s11244-007-0278-8

（75）Outka, D. A.； Madix, R. J. J. Am. Chem. Soc. 1987， 109, 1708. doi: 10.1021/ja00240a018

（76）Kim, T. S.； Gong, J.； Ojifinni, R. A.； White, J. M.； Mullins, C. B. J. Am. Chem. Soc. 2006， 128, 6282. doi: 10.1021/ja058263m

（77）Ojifinni, R. A.； Froemming, N. S.； Gong, J.； Pan, M.； Kim, T. S.； White, J. M.； Henkelman, G.； Mullins, C. B. J. Am. Chem. Soc. 2008， 130, 6801. doi: 10.1021/ja800351j

（78）Quiller, R. G.； Baker, T. A.； Deng, X.； Colling, M. E.； Min, B. K.； Friend, C. M. J. Chem. Phys. 2008， 129, 064702.

（79）Ojifinni, R. A.； Gong, J. L.； Froemming, N. S.； Flaherty, D. W.；Pan, M.； Henkelman, G.； Mullins, C. B. J. Am. Chem. Soc. 2008， 130, 11250. doi: 10.1021/ja803482h

（80）Bartram, M. E.； Koel, B. E. Surf. Sci. 1989， 213, 137. doi: 10.1016/0039-6028（89）90255-0

（81）Wang, J.； Koel, B. E. Surf. Sci. 1999， 436, 15. doi: 10.1016/S0039-6028（99）00457-4

（82）McClure, S. M.； Kim, T. S.； Stiehl, J. D.； Tanaka, P. L.；Mullins, C. B. J. Phys. Chem. B 2004， 108, 17952. doi: 10.1021/jp047335d

（83）Gong, J. L.； Ojifinni, R. A.； Kim, T. S.； White, J. M.； Mullins, C. B. J. Am. Chem. Soc. 2006， 128, 9012. doi: 10.1021/ja062624w

（84）Lopez, N.； Garcia-Mota, M.； Gomez-Diaz, J. J. Phys. Chem. C 2008， 112, 247. doi: 10.1021/jp077205f

（85）Davis, K. A.； Goodman, D. W. J. Phys. Chem. B 2000， 104, 8557. doi: 10.1021/jp001699y

（86）Outka, D. A.； Madix, R. J. Surf. Sci. 1987， 179, 361. doi: 10.1016/0039-6028（87）90063-X

（87）Gong, J. L.； Flaherty, D. W.； Ojifinni, R. A.； White, J. M.；Mullins, C. B. J. Phys. Chem. C 2008， 112, 5501. doi: 10.1021/jp0763735

（88）Gong, J. L.； Mullins, C. B. J. Am. Chem. Soc. 2008， 130, 16458. doi: 10.1021/ja805303s

（89）Gong, J. L.； Flaherty, D. W.； Yan, T.； Mullins, C. B.ChemPhysChem 2008， 9, 2461. doi: 10.1002/cphc.v9:17

（90）Liu, X.； Xu, B.； Haubrich, J.； Madix, R. J.； Friend, C. M. J. Am. Chem. Soc. 2009， 131, 5757. doi: 10.1021/ja900822r

（91）Xu, B.； Liu, X.； Haubrich, J.； Madix, R. J.； Friend, C. M. Angew. Chem. Int. Edit. 2009， 48, 4206. doi: 10.1002/anie.v48:23

（92）Xu, B.； Liu, X.； Haubrich, J.； Friend, C. M. Nat. Chem. 2009， 2, 61.

（93）Xu, B.； Madix, R. J.； Friend, C. M. J. Am. Chem. Soc. 2010，132, 16571. doi: 10.1021/ja106706v

（94）Xu, B.； Haubrich, J.； Baker, T. A.； Kaxiras, E.； Friend, C. M. J. Phys. Chem. C 2011， 115, 3703. doi: 10.1021/jp110835w

（95）Vinod, C. P.； Kulkarni, G. U.； Rao, C. N. R. Chem. Phys. Lett. 1998， 289, 329. doi: 10.1016/S0009-2614（98）00447-3

（96）DiCenzo, S. B.； Berry, S. D.； Hartford, E. H. Phys. Rev. B 1988，38, 8465. doi: 10.1103/PhysRevB.38.8465

（97）Boyen, H. G.； Herzog, T.； Kastle, G.； Weigl, F.； Ziemann, P.；Spatz, J. P.； Möller, M.； Wahrenberg, R.； Garnier, M. G.；Oelhafen, P. Phys. Rev. B 2002， 65, 075412. doi: 10.1103/PhysRevB.65.075412

（98）Mason, M. G. Phys. Rev. B 1983， 27, 748. doi: 10.1103/PhysRevB.27.748

（99）Liang, K. S.； Salaneck, W. R.； Aksay, I. A. Solid State Commun. 1976， 19, 329. doi: 10.1016/0038-1098（76）91344-2

（100）Cordes, O.； Harsdorff, M. Appl. Surf. Sci. 1988， 33, 152.

（101）Zafeiratos, S.； Kennou, S. Surf. Sci. 1999， 443, 238. doi: 10.1016/S0039-6028（99）01014-6

（102）Lim, D. C.； Lopez-Salido, I.； Dietsche, R.； Bubek, M.； Kim, Y. D. Angew. Chem. Int. Edit. 2006， 45, 2413.

（103）Kizuka, T.； Tanaka, N. Phys. Rev. B 1997， 56, 10079. doi: 10.1103/PhysRevB.56.R10079

（104）Costanzo, E.； Faraci, G.； Pennisi, A. R.； Ravesi, S.； Terrasi, A.；Margaritondo, G. Solid State Commun. 1992， 81, 155.

（105）Chusuei, C.； Lai, X.； Luo, K.； Goodman, D. W. Top. Catal. 2001， 14, 71.

（106）Häkkinen, H. Chem. Soc. Rev. 2008， 37, 1847. doi: 10.1039/b717686b

（107）Shaikhutdinov, S.； Meyer, R.； Naschitzki, M.； Bäumer, M.；Freund, H. J. Catal. Lett. 2003， 86, 211. doi: 10.1023/A:1022616102162

（108）Meier, D. C.； Goodman, D. W. J. Am. Chem. Soc. 2004， 126, 1892. doi: 10.1021/ja030359y

（109）Huang, W. X. Science China Chemistry 2012， 42, 469. ［黄伟新.中国科学: 化学, 2012， 42, 469.］

（110）Huang, W. X. Top. Catal. 2013， 56, 1363. doi: 10.1007/s11244-013-0139-6

（111）Huang, W. X.； Gao, Y. X. Catal. Sci. Technol. 2014， 4, 3772. doi: 10.1039/C4CY00679H

（112）Huang, W. X. In RSC Catalysis Series-Heterogeneous Gold Catalysts and Catalysis； Ma, Z., Dai, S. Eds., Royal Society of Chemistry: Cambridge, UK, 2014； pp 533–574.

（113）Huang, W. X.； Cao, T. In Catalysis by Materials with Well-Defined Structure； Wu, Z. L.,Overbury, S. H. Eds., Elsevier B.V.: Amsterdam, Netherlands, 2015； pp 1–29.

（114）Wu, Z. F.； Xu, L. S.； Zhang, W. H.； Ma, Y. S.； Yuan, Q.； Jin, Y. K.； Yang, J. L.； Huang, W. X. J. Catal. 2013， 304, 112. doi: 10.1016/j.jcat.2013.04.013

（115）Wu, Z. F.； Ma, Y. S.； Zhang, Y. L.； Xu, L. S.； Chen, B. H.；Yuan, Q.； Huang, W. X. J. Phys. Chem. C 2012， 116, 3608.

（116）Wu, Z. F.； Jiang, Z. Q.； Jin, Y. K.； Xiong, F.； Sun, G. H.；Huang, W. X. Science China Chemistry 2016， in press. doi: 10.1007/s11426-015-5510-y.

（117）Wu, Z. F.； Jin, Y. K.； Xu, L. S.； Yuan, Q.； Xiong, F.； Jiang, Z. Q.； Huang, W. X. J. Phys. Chem. C 2014， 118, 8397.

（118）Wu, Z. F.； Jiang, Z. Q.； Jin, Y. K.； Xiong, F.； Huang, W. X. J. Phys. Chem. C 2014， 118, 26258. doi: 10.1021/jp509551d

（119）Qian, K.； Jiang, Z. Q.； Huang, W. X. J. Mol. Catal. A: Chem. 2007， 264, 26. doi: 10.1016/j.molcata.2006.08.074

（120）Qian, K.； Sun, H. X.； Huang, W. X.； Fang, J.； Lv, S. S.； He, B.；Jiang, Z. Q.； Wei, S. Q. Chem. Eur. J. 2008， 14, 10595. doi: 10.1002/chem.200801199

（121）Wang, S.； Qian, K.； Bi, X. Z.； Huang, W. X. J. Phys. Chem. C 2009， 113, 6505.

（122）Qian, K.； Zhang, W. H.； Sun, H. X.； Fang, J.； He, B.； Ma, Y. S.；Jiang, Z. Q.； Wei. S. Q.； Yang, J. L.； Huang, W. X. J. Catal. 2011， 277, 95. doi: 10.1016/j.jcat.2010.10.016

（123）Qian, K.； Luo, L. F.； Bao, H. Z.； Hua, Q.； Jiang, Z. Q.； Huang, W. X. Catal. Sci. Technol. 2013， 3, 679. doi: 10.1039/C2CY20481A

（124）Jiang, Z. Q.； Zhang, W. H.； Jin, L.； Yang, X.； Xu, F. Q.； Zhu, J. F.； Huang, W. X. J. Phys. Chem. C 2007， 111, 12434.

（125）Yu, Y.； Xu, L. S.； Zhang, W. H.； Jiang, Z. Q.； Zhu, J. F.； Huang, W. X. Chin. J. Chem. Phys. 2009， 22, 339. doi: 10.1088/1674-0068/22/04/339-345

（126）Chen, S. L.； Zhang, B. S.； Su, D. S.； Huang, W. X. ChemCatChem 2015， 7, 3290. doi: 10.1002/cctc.201500599

（127）Qian, K.； Huang, W. X.； Jiang, Z. Q.； Sun, H. X. J. Catal. 2007，248, 137. doi: 10.1016/j.jcat.2007.02.010

（128）Qian, K.； Huang, W. X.； Fang, J.； Lv, S. S.； He, B.； Jiang, Z. Q.；Wei, S. Q. J. Catal. 2008， 255, 269. doi: 10.1016/j.jcat.2008.02.012

（129）Qian, K.； Lv, S. S.； Xiao, X. Y.； Sun, H. X.； Lu, J. Q.； Luo, M. F.； Huang, W. X. J. Mol. Catal. A: Chem. 2009， 306, 40. doi: 10.1016/j.molcata.2009.02.014

（130）Chen, S. L.； Luo, L. F.； Jiang, Z. Q.； Huang, W. X. ACS Catal. 2015， 5, 1653. doi: 10.1021/cs502067x

Structure-Sensitivity of Au Catalysis

HUANG Wei-Xin*QIAN Kun WU Zong-Fang CHEN Shi-Long
（CAS Key Laboratory of Materials for Energy Conversion, Hefei National Laboratory for Physical Sciences at the Microscale, Department of Chemical Physics, University of Science and Technology of China, Hefei 230026, P. R. China）

Au catalysis is representative of nanocatalysis. Au catalysis has been demonstrated to be very complex and structure-sensitive. In this short review we summarize the literature reports on Au catalysis and our recent progress on the fundamental understanding of Au catalysis using model catalysts from single crystals to nanocrystals. We demonstrate the structure-sensitivity of Au catalysis used for NO decomposition, CO oxidation, and propylene epoxidation with H2and O2and the corresponding active Au structures. We discuss the effects of the geometric and electronic structures and the Au–oxide support interactions on Au catalysis, and the origin of high catalytic activity of the Au surface at low temperatures. Finally, we provide an outlook for future research directions of structure-sensitive Au catalysis.

Surface chemistry; Heterogeneous catalysis; Model catalyst; CO oxidation reaction; NO decomposition reaction; Propylene epoxidation reaction

O643

10.3866/PKU.WHXB201511092

Received: October 12, 2015； Revised: November 9, 2015； Published on Web: November 9, 2015.

*Corresponding author. Email: huangwx@ustc.edu.cn； Tel: +86-551-63600435.

The project was supported by the National Key Basic Research Program of China （973） （2013CB933104）, National Natural Science Foundation of China （21525313, 20973161, 21373192）, MOE Fundamental Research Funds for the Central Universities, China （WK2060030017）, MPG-CAS Partner Group Program and Collaborative Innovation Center of Suzhou Nano Science and Technology.

国家重点基础研究发展规划项目（973） （2013CB933104）,

国家自然科学基金（21525313, 20973161, 21373192）, 教育部中央高校基本科研业务费（WK2060030017）, 德国马普协会-中国科学院伙伴小组计划和教育部苏州纳米科学与技术协同创新中心资助

©Editorial office of Acta Physico-Chimica Sinica