Fabrication of gold nanoparticles in confined spaces using solid-phase reduction: Significant enhancement of dispersion degree and catalytic activity

Author’s Accepted Manuscript Fabrication of Gold Nanoparticles in Confined Spaces using Solid-Phase Reduction: Significant Enhancement of Dispersion Degree and Catalytic Activity Zhi-Min Xing, Yu-Xia Gao, Li-Ying Shi, Xiao-Qin Liu, Yao Jiang, Lin-Bing Sun www.elsevier.com/locate/ces

PII: DOI: Reference:

S0009-2509(16)30559-0 http://dx.doi.org/10.1016/j.ces.2016.10.029 CES13200

To appear in: Chemical Engineering Science Received date: 1 August 2016 Revised date: 12 October 2016 Accepted date: 15 October 2016 Cite this article as: Zhi-Min Xing, Yu-Xia Gao, Li-Ying Shi, Xiao-Qin Liu, Yao Jiang and Lin-Bing Sun, Fabrication of Gold Nanoparticles in Confined Spaces using Solid-Phase Reduction: Significant Enhancement of Dispersion Degree and Catalytic Activity, Chemical Engineering Science, http://dx.doi.org/10.1016/j.ces.2016.10.029 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Fabrication of Gold Nanoparticles in Confined Spaces using Solid-Phase Reduction: Significant Enhancement of Dispersion Degree and Catalytic Activity Zhi-Min Xing, Yu-Xia Gao, Li-Ying Shi, Xiao-Qin Liu,* Yao Jiang, Lin-Bing Sun*

Jiangsu National Synergetic Innovation Center for Advanced Materials (SICAM), State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemistry and Chemical Engineering, Nanjing Tech University, 5 Xinmofan Road, Nanjing 210009, China

*Corresponding author. Tel: +86-25-83587177; Fax: +86-25-83587191 E-mail: [email protected]; [email protected]

1

Abstract: Au-containing catalysts are highly active in diverse reactions, and their activity strongly depends on the dispersion degree of Au. Here we report for the first time a solid-phase reduction strategy to promote Au dispersion in template-occluded SBA-15 (AS) by fully considering three crucial factors, namely (i) the interaction between Au and supports, (ii) the space where Au precursors locate during reduction, and (iii) the reduction method. First, both template and silica walls in AS offer interaction with Au species. Second, AS presents confined spaces between template and silica walls. Third, the reduction in solid phase avoids the competitive adsorption of solvent molecules. The results show Au-containing AS has a better dispersion of Au than its counterpart prepared from template-free SBA-15 (CS). Moreover, the obtained materials exhibit excellent catalytic activity in reduction reactions and that the organic template retained in mesopores promotes the reactions greatly.

Keywords: dispersion, gold nanoparticles, confined spaces, solid-phase reduction, catalytic reduction

2

1. Introduction Due to their unique physicochemical properties and excellent catalytic activity, noble metal nanoparticles have attracted much attention in recent years (Azubel et al., 2014; Ciracì et al., 2012; Savage et al., 2012; Valente et al., 2012; Wang et al., 2009). As one of the most concerned noble metal systems, gold nanoparticles (Au NPs), in contrast to the chemical inertness of pure Au films or bullion, are extensively studied (Boscoboinik et al., 2015; Yatabe et al., 2015). Au NPs have been demonstrated to be good catalysts for a variety of reactions such as oxidation (Wang et al., 2015), reduction (Mistry et al., 2014; Zhu et al., 2013), and dissociation (Lin et al., 2015). Because only those active species dispersed on the outermost layer are accessible to reactant molecules, the catalytic activity is strongly dependent on the dispersion degree of Au. An exponential increase in catalytic activity with the decrease of particle sizes has been observed (Lee et al., 2014; Lin et al., 2015). Unfortunately, noble metal nanoparticles are easily mobile on the surface of supports, often leading to severe aggregation and subsequent decrease/loss of catalytic activity after exposure to reaction conditions. As a kind of material with ordered pore structure, mesoporous silica shows high surface area and large pore volume, and is an ideal choice of support for Au dispersion. Although various methods including ion implantation (Guczi et al., 2003), sol-gel process (Fang et al., 2011), and sputtering (Raghuwanshi et al., 2014) have been reported to disperse Au on mesoporous silica, Au NPs still suffer from considerable aggregation between nanoparticles in close proximity owing to high specific surface energy. So far the development of a facile, efficient method to disperse and stabilize Au NPs has remained a great challenge. On the basis of previous reports and deep analysis, three factors are believed to correlate closely with the dispersion of Au, namely (i) the interaction between Au and supports, (ii) the space where Au precursors locate during reduction, and (iii) the reduction method for Au precursors. For the first factor, functional groups (e.g. ‒NH2 (Dhar et al., 2009; Yoon et al., 2013), ‒SH (Mayeda et al., 2012), and ‒ 3

SO3H (Rombi et al., 2012; Tschulik et al., 2015)) are usually grafted onto mesoporous silica to promote the interaction between Au and supports. A case in point is Wang’s work (Wang et al., 2010), where mesoporous silica SBA-15 was functionalized with amino groups prior to the introduction of Au. A monolayer of charged organic groups were attached on the surface of mesopores, which can adsorb and interact strongly with oppositely charged metal precursors to fabricate dispersed Au NPs. Albeit interesting, complicated surface functionalization is unavoidable by using this traditional method. For the second factor, Au precursors always locate in template-free mesopores during reduction. In other words, only after the complete removal of template, Au precursors are introduced and subsequently reduced. Nevertheless, there exists a special micro-environment between template and silica walls in template-occluded mesoporous silica. This micro-environment has been proven by Zhu and coworkers (Yue et al., 2008). They created a new kind of CO2 capturer by incorporating amino group into templateoccluded MCM­41, which is highly efficient in the adsorption of CO2. If Au precursors located in this special micro-environment during reduction, the confined effect may favor the dispersion of resultant Au NPs. For the third factor, a liquid-phase method is frequently used for the reduction of Au precursors to metallic Au, namely, the reduction process is performed in solutions (Yang et al., 2013). Taking account of the competitive adsorption of solvent molecules, the dispersion degree of resultant Au NPs by use of such a liquid-phase method is questionable (Wang et al., 2004). Moreover, the solubility makes precursors diffuse easily in both internal and external surfaces of supports, and the formation of aggregated nanoparticles becomes possible. In this regard, a new reduction method based on solid phase should benefit the formation of well dispersed Au NPs. Despite great efforts, the fabrication of Au NPs by fully considering the abovementioned three factors has never been reported up to now. Here, we report for the first time a solid-phase reduction (SPR) strategy to fabricate Au NPs in template-occluded mesoporous silica by fully considering the abovementioned three factors. The Au 4

precursor (HAuCl4) was incorporated into the template-occluded SBA-15 (AS) by grinding, followed by the addition of reductant NaBH4 to grind for another 10 min during which the precursor can be converted to metallic Au completely (Scheme 1). First, our strategy avoids the complicated surface functionalization, while both template (with plenty of hydroxyl groups) and internal surface (containing more silanols than calcined SBA-15, CS) in the template-occluded support possess interaction with Au species. Second, the AS support provides a confined space between template and silica walls, which hinders the aggregation of Au during reduction and maintains the catalytic activity during reactions by limiting the movement of Au species. Third, the reduction in solid phase evades the competitive adsorption of solvent molecules and restricts the diffusion of Au species, which is beneficial to the dispersion of resultant Au NPs. The results show that our strategy is highly efficient in fabricating dispersed Au NPs, and the size of Au NPs is only 3.8 nm for the material containing 1.0 wt% of Au (1.0AuAS), which is much smaller than its counterpart prepared from template-free SBA-15 (1.0AuCS, possessing an Au size of 11.5 nm). We also demonstrate that the obtained materials exhibit excellent catalytic activity in the reduction of organics. Taking catalytic reduction of methylene blue (MB) as an example, 1.0AuAS can convert 100% of MB within 5 min, whereas only 35% of MB was converted even after 60 min over its counterpart 1.0AuCS. Similarly, the reduction of 4-aminophenol (4-NP) was finished quickly within 24 min over 1.0AuAS; however, only 22% of 4-NP was converted even after 60 min over 1.0AuCS. It is interesting to note that the organic template retained in mesopores is capable of enriching organic reactants during reactions, which favors the subsequent catalytic reactions greatly. The preservation of template is thus beneficial to not only the dispersion of Au NPs, but also the improvement of catalytic activity of resultant materials. 2. Experimental Section 2.1. Chemicals. 5

Pluronic 123 (P123), sodium borohydride (NaBH4), 4-nitrophenol (4-NP), and methylene blue (MB) were obtained from Aldrich. Tetraethylorthosilicate (TEOS) were obtained from Sinopharm. Hydrochloric acid (HCl) were obtained from Shanghai Ling Feng. Chloroauric acid (HAuCl43H2O) were obtained from Adamas. All reagents were used as received without further purification. 2.2. Materials Synthesis. Mesoporous silica SBA-15 was synthesized according to the reported method (Zhao et al., 1998b). In a typical synthesis, 2 g of Pluronic P123 was firstly dissolved in 75 g of aqueous HCl solution (1.6 M). Then 4.25 g of silica source tetraethylorthosilicate (TEOS) was added to the homogeneous solution and stirred at 40 oC for 24 h, followed by hydrothermal treatment at 100 oC for 24 h. The template-occluded mesoporous silica SBA-15 (denoted as AS) was recovered by filtration and dried under ambient conditions. After calcination in flowing air at 550 oC for 5 h, the template P123 was removed and template-free mesoporous silica SBA-15 (denoted as CS) was obtained. The Au precursor HAuCl4·3H2O was introduced to AS by grinding under ambient conditions, then the obtained solids were reduced by a SPR strategy. The SPR strategy means that Au precursor was reduced by grinding with the reductant under the free-solvent condition. The obtained samples were denoted as nAuAS (n represents the weight percent of Au per gram of the obtained samples). Typically, 0.0106 g of HAuCl4·3H2O was introduced to 0.5 g of AS by grinding under ambient conditions for 30 min. Then 0.0152 g of NaBH4 was added to the above mixtures and ground for another 10 min (the Au/NaBH4 molar ratio was 1:15). The color of mixtures turned red quickly. After reduction, the obtained solids were washed with deionized water sufficiently so that excess amounts of reagents including reductant NaBH4, and byproduct such as NaCl can be removed. The obtained sample was denoted as 1.0AuAS. By changing the amount of Au precursor, it can obtain other samples which have the different Au content. For comparison, HAuCl4·3H2O was also introduced to CS in a similar process as described above. The 6

obtained samples were denoted as nAuCS (n represents the weight percent of Au per gram of the obtained samples). In a typical solution phase syntheses, 0.0106 g of HAuCl4·3H2O, 100 mL of deionized water, and 0.5 g of support AS was mixed in a 500 mL breaker. Subsequently 0.0162 g of reductant NaBH4 was added under vigorous stirring. The solution was further stirred for 30 min, and then filtered, washed with a large amount of deionized water. The obtained sample was denoted as 1.0AuAS-r. For comparison, a given quantity of template P123 was introduced to the channels of 1.0AuCS by impregnation (the weight ratio of 1.0AuCS and P123 is 1:1 according to the content of P123 in 1.0AuAS). Typically, 0.1 g of P123 and 0.1g of 1.0AuCS were dissolved in 50 mL of deionized water and further stirred for 30 min. The mixture was centrifuged and dried. The obtained sample was denoted as [email protected]. 2.3. Materials Characterization. Powder X-ray diffraction (XRD) patterns of the samples were recorded using a Bruker D8 Advance diffractometer with Cu Kα radiation in the 2θ ranges from 0.6o to 4o and 5o to 80o at 40 kV and 40 mA. Metal content loading on the support was determined by Jarrell-Ash 1100 inductively coupling plasma (ICP) spectrometer. The N2 adsorption/desorption isotherms were carried out at 77 K on an ASAP 2020 instrument. Prior to the measurement, every sample was pretreated for 4 h at 343 K under vacuum (ca. 10−6 Torr). The measurements of high resolution transmission electron microscopy (HRTEM) and energy-dispersive X-ray spectroscopy (EDX) operated in scanning transmission electron microscopy (STEM) mode were performed on a Philips Analytical FEI Tecnai 30 electron microscope operated at an acceleration voltage of 300 kV. The X-ray photoelectron spectroscopy (XPS) analysis was conducted on a Physical Electronic PHI-550 spectrometer equipped with an Al Kα X-ray source (hν = 1486.6 eV) and was operated at 10 kV and 35 mA. Diffuse reflectance UV-vis spectroscopy was carried out on the 7

Perkin Elmer Lambda 950 spectrometer. Fourier transform infrared (IR) spectra were recorded on a Nicolet Nexus 470 spectrometer with a spectra resolution of 2 cm‒1 using transparent KBr pellets. Thermogravimetric (TG) analysis was performed in an air flow from room temperature to 800 oC on a thermobalance (STA-499C, NETZSCH). The UV-vis adsorption spectra were recorded on an UV-vis spectrometer (Jasco V-550) at 25 oC. 2.4. Catalytic Tests. Firstly, the reduction of 4-nitrophenol (4-NP) to 4-aminophenol (4-AP) was employed to evaluate the catalytic activity of Au catalysts. It was carried out in a quartz cuvette and monitored using UV-vis spectroscopy under ambient conditions. A total of 0.02 mL of aqueous 4-NP solution (0.005 M) was mixed with 0.2 mL of fresh NaBH4 (0.2 M) solution. Subsequently, 2.5 mL of aqueous dispersion of Au catalysts (2 mg) was added. The pH of 4-NP solution is not adjusted before adding catalysts and NaBH4. Mixed solution was quickly subjected to UV-vis measurements and the obtained data can be identified as the value of reaction time t = 0 min. Afterward, the solution was in situ measured every 4 min to obtain the successive information about the reaction. To study the reusability, 1.0 mL of aqueous 4-NP solution (0.005 M) was mixed with 10.0 mL of fresh NaBH4 (0.2 M) solution. Subsequently, 125.0 mL of aqueous dispersion of Au catalysts (100 mg) was added. The catalyst was separated from the reaction mixture by centrifugation after the entire reduction process was completed. The recycled catalyst was washed with deionized water for 3 times. The obtained catalyst was reused in the next run. After reaction for 24 min, the solution was monitored using UV-vis spectroscopy. The process was repeated 5 times. In order to investigate the function of template P123 in the catalytic reaction, 2 mg of [email protected] was also added to the reaction solution as described above. Meanwhile, about 2 mg of P123 was dissolved in the reaction solution before 2 mg of 1.0AuCS was dissolved (denoted as P123+1.0AuCS). The catalytic activity of Au catalysts was also explored by the catalytic reduction of MB to leuco 8

methylene blue (LMB). In a typical procedure, 1 mg of Au catalysts was dispersed into 2.5 mL of MB dye solutions (25 ppm), followed by rapid injection of 0.2 mL of NaBH4 solution (0.2 M). The reaction process was monitored as described above. In the recycling study, 50 mg of Au catalysts was dispersed into 125.0 mL of MB dye solutions (25 ppm), followed by rapid injection of 10.0 mL of NaBH4 solution (0.2 M) at the first time. The catalyst was separated from the reaction mixture by centrifugation when the whole reduction process was finished. The recycled catalyst was washed with deionized water for 3 times, and then reused in the next run. After reaction for 5 min, the solution was monitored using UV-vis spectroscopy. The process was repeated 5 times. 3. Results 3.1. Structural and Surface Properties. The low-angle XRD patterns show that all samples possess an intense diffraction peak and two weak ones, which can be respectively indexed to (100), (110), and (200) reflections and correspond to a twodimensional hexagonal pore symmetry (Figs. 1A and S1A) (Liu et al., 2013a; Yin et al., 2014; Zhao et al., 1998a). This reflects that the ordered mesostructure of all samples has no significant change after the introduction of Au NPs. However, in comparison with AuCS, the d-spacings of all AuAS samples shift toward higher values. Further calculation indicates that the unit cell constant values are 10.7 nm for AuCS and CS (Table S1). The unit cell constant values of AuAS samples range from 11.7 to 12.0 nm, which are higher than the unit cell constant of CS but close to AS (12.0 nm). These results indicate the contraction of silica frameworks in the process of template removal. The wide-angle XRD patterns show that the pristine AS and CS present a single broad diffraction peak centered at 23o (Fig. 1B), which can be attributed to amorphous silica walls (Zhao et al., 1998b). All AuAS and AuCS samples present four diffraction peaks at 2θ of 38o, 44o, 65o, and 78o corresponding to metallic Au (Figs. 1B and S1B), indicating the absence of other valence gold ions (such as Au3+ and 9

Au+) unreduced in the samples (López et al., 2014; Li et al., 2014). In comparison with AuCS, the samples AuAS exhibit broader diffraction peaks of metallic Au, suggesting better Au dispersion. The crystallite size of metallic Au was calculated by the Scherrer equation using the Au (111) diffraction peak. The AuNPs in 1.0AuAS shows a size of 3.8 nm, which is much smaller than that in its analogue 1.0AuCS (11.5 nm, Table S1). Moreover, Au NPs in all AuAS samples show a smaller size in contrast to AuCS samples. On the basis of these results, it is safe to say that the use of template-occluded mesoporous silica as the support promotes the dispersion of Au NPs obviously. TEM plays a vital role to characterize the structure of mesoporous materials and the dispersion degree of metal nanoparticles. The bright-field images illustrate that the ordered pore structure of both AuAS and AuCS is preserved (Fig. S2). The dark-field images provide clearer proofs of the presence of the Au NPs on the supports (Fig. 2). The analysis of the TEM image for 1.0AuAS reveals that the average size of Au NPs is 3.9 nm. However, the average size of Au NPs in its analogue 1.0AuCS is 12.2 nm. These results are in good agreement with the wide-angle XRD patterns. The Au NPs in 1.0AuAS are well dispersed instead of being aggregated in 1.0AuCS. The elemental mapping images show that silicon is equi-distributed, and Au in the sample 1.0AuAS has better dispersion than that in 1.0AuCS. In addition, the AuAS and AuCS samples exhibit different color (Figs. 2 insets and S3). The characteristic color of the sample 1.0AuAS is pink in contrast to the purple sample 1.0AuCS owing to the changes in the localized surface plasmon resonance (LSPR) absorption wavelength arising from size˗control (Mori et al., 2015). Other samples also have the same color contrast, which demonstrates that the AuAS samples have a higher dispersion degree of Au NPs than AuCS. To provide evidence showing that nanoparticles are formed within the pores, HRTEM and pore size distributions (PSDs) of samples were examined. The HRTEM images of 1.0AuAS either perpendicular or parallel to pore direction clearly demonstrate the location of Au nanoparticles within the pores (Fig. 10

3). Moreover, N2 adsorption/desorption isotherms and corresponding PSDs of different samples are shown in Figs. 4 and S4. The isotherm shape of all samples is of type IV with an H1 hysteresis loop, which belongs to the structure of cylindrical mesopores (Yin et al., 2014). N2 adsorption at 77 K is a sensitive method to probe the change of pore structure. For noble metals, very small amount of loading can lead to some changes in isotherms and PSDs, which is different from some oxides (such as Al 2O3 and MgO) exhibiting strong interaction with silica surface. In our cases, the change in isotherms and PSDs can be observed for the sample with 0.5%-1.0% Au loading. Similar results were also reported in literature. For example, clear changes can be detected by N2 adsorption for the samples loading with 0.5%-1.0% (Li et al., 2009), 0.7% (Gutiérrez et al., 2012), or 1.0% of Au (Kumar et al., 2016). After the introduction of Au NPs in the pores of AS, the surface area and pore volume of AuAS, compared to the support AS, show an obvious decrease. Meanwhile, the PSDs exhibit an obvious tendency to decrease with the increase of Au loading. In line with AuAS, the surface area, pore volume, and pore size of AuCS have an obvious decrease with the increase of Au content. It is true that the size of Au particles is around 3.9 nm in 1.0AuAS, and the peak of PSD would shift about 3.9 nm if the pores are fully occupied by Au particles. However, there are only limited amount of Au particles in the pores, and only part of pore surface is occupied by Au particles (Fig. 3). Moreover, the results of N2 adsorption are average data of pore structure. As a result, the PSD peak shifted less 2 nm although the size of Au particles is 3.9 nm. Similar results were also reported in literature (Gao et al., 2016; Yang and Qi, 2013). Based on the results of HRTEM and PSD, it should be safe to say that Au nanoparticles are formed within the pores. Further information about the surface components of Au NPs is provided by XPS measurements. XPS spectra of samples AuAS exhibit doublet peaks of zero valence Au 4f7/2 and Au 4f5/2 centered at about 84.0 and 87.9 eV caused by metallic Au species (Fig. 5). No peaks ascribed to cationic Au species are observed, revealing the complete reduction of the precursor HAuCl4 to metallic Au by using the SPR 11

strategy (Dai et al., 2015; Liu et al., 2013b). It is interesting to note that 1.0AuCS exhibits Au 4f7/2 binding energy at about 84.2 eV, with an increase of 0.2 eV as compared with AuAS (84.0 eV), which is also observed in other AuCS samples with different Au contents (Fig. S5). This suggests that the AuAS samples have strong interaction between Au NPs and support (Wang et al., 2010; Wang et al., 2011). Diffuse reflectance spectroscopy in the UV-vis region is a sensitive method for characterizing the coordination environment of Au NPs. The pristine AS and CS have no absorption peak in the wavelength region >400 nm (Fig. 6). However, the samples AuAS and AuCS loaded with Au NPs reveal a surface plasmon resonance absorption peak due to the transition of valence electrons in the d band to the Fermi surface (Inouye et al., 1998), which indicates the formation of Au NPs in AuAS and AuCS samples (Fig. S6). The absence of other absorption peaks in the wavelength region <400 nm indicates the existence of only metallic Au, which is in accordance with the results of XRD and XPS. Interestingly, the characteristic absorption peak corresponding to Au NPs is observed at 522 nm for AuAS, but at 534 nm for AuCS. It is reported that the adsorption maximum moves progressively to longer wavelengths when the aggregation of Au NPs proceeds (Mei et al., 2013; Sacaliuc et al., 2007; Yu et al., 2011). The slightly more blue-shifted UV-vis absorption maxima of the samples 1.0AuAS compared to that of 1.0AuCS suggests that the size of Au NPs in 1.0AuAS was smaller than that in 1.0AuCS, which is in line with the results of XRD and TEM. On the basis of the results above, it is apparent that the SPR strategy can convert the precursor HAuCl4 to metallic Au efficiently regardless of AS and CS supports. By using the template-occluded mesoporous silica as support, well-dispersed Au NPs can be fabricated; however, severe aggregation of Au NPs occurs on the conventional support without template. 3.2. Catalytic Reduction of 4-NP. The catalytic reduction of 4-NP to 4-AP in the presence of NaBH4 (Fig. 7A) was first employed to 12

evaluate the catalytic activity of Au catalysts (Chen et al., 2015; Ma et al., 2010; Zhang and Su, 2012). 4-NP is one of the most common water pollutants with high toxicity and carcinogenic character (Yang et al., 2016), while 4-AP has been proposed as a dyeing agent (Zhang et al., 2016). In the UV-vis spectrum, 4-NP exhibits an absorption peak at 400 nm in NaBH4 solutions due to the formation of 4nitorphenolate, while 4-AP exhibits an absorption peak at 296 nm (Deng et al., 2010). Hence, the process of reaction can be monitored by recording the absorption spectra of the reaction solution as a function of the reaction time. Due to the linear relationship of concentration with absorbance, the concentration of 4-NP can be calculated from the absorbance value at 400 nm according to the calibration curve (Xiao et al., 2012). The plot of ct/c0 against the reaction time is shown in Figs. 7B and S7-S9, where ct and c0 respectively represents the concentration of 4-NP at the time of t and 0 min. Catalytic activity of the supports AS and CS was firstly examined. The decrease of 4-NP was negligible, which suggests that the reduction of 4-NP cannot proceed at all in the absence of Au. In the case of samples containing Au, the conversion of 4-NP can be observed. In contrast to the AuCS samples, the AuAS samples exhibit much higher catalytic activity under the same reaction conditions. Taking 1.0AuAS as an example, the reduction of 4-NP was finished quickly within 24 min; however, only 22% of 4-NP was converted even after 60 min over 1.0AuCS. The pseudo-first-order kinetics and rate constants of 1.0AuAS and 1.0AuCS from the rate equation ln(ct/c0) = kt are shown in Fig. S10. The reaction rate constant k is determined to be 0.175 min-1 for 1.0AuAS, while only 0.002 min-1 for 1.0AuCS. These results clearly show that AuAS are much more active than their AuCS counterparts in the catalytic reduction of 4-NP. The rate constant was also compared with those using other reported catalysts. For example, the rate constant is 0.019 min−1 for (PDDA/PSS)5/2Au, 0.069 min−1 for (PDDA/PSS)5/2Ag, and 0.156 min−1 for AuNP-2, which is lower than that of our material 1.0AuAS (0.175 min-1) (Wang et al., 2016; Zhang and Su, 2012). 13

Taking account of its importance in practical applications, the recyclability of catalysts was investigated. Five successive cycles of the catalytic reduction by 1.0AuAS were carried out (Fig. 7C). The catalyst 1.0AuAS exhibited similar catalytic activity without detectable reduction in the conversion even after running for 5 cycles. The Au contents of liquid phase after reaction were checked, and no Au is detected by ICP. Thus, it is safe to say that it is really heterogeneous catalysis in the present study. Wide-angle XRD measurement was performed to examine if the sample 1.0AuAS underwent aggregation in the process of reactions. The results show that the dispersion degree of Au NPs is well retained without aggregation after five cycles (Fig. S11). TG technique was employed to analyze the content of template in the samples. After five cycles, the weight loss caused by template in 1.0AuAS is 44% (Fig. S12), which is comparable to that of the fresh catalyst (45%). This means that no loss of template takes place during reactions, which is beneficial to the maintenance of Au dispersion and catalytic activity. 3.3. Catalytic Reduction of MB. The catalytic reduction of MB to LMB was employed to further explore the catalytic activity of resultant materials (Fig. 8A) (Yao et al., 2014). The reactant MB is a heterocyclic aromatic dye, and the catalytic reduction of MB to LMB can reduce pollution in those colored waste waters (Wang et al., 2014). The UV-vis spectrum of initial MB solution with the color of dark blue shows two adsorption peaks with the maximum at 664 nm (Yonemoto et al., 2014). The pristine AS cannot catalyze the reduction of MB (Figs. 8B and S13). Interestingly, when CS was added to the reaction solution, a decrease of 40% in the concentration of MB was observed within 5 min; further increasing the time, the concentration of residual MB almost kept constant. This implies that the instant adsorption of MB by CS takes place and then the adsorption reaches equilibrium. These results indicate that the pristine AS and CS cannot catalyze the reduction of MB. For the samples containing Au, interesting catalytic 14

performance can be found, and all AuAS samples show much superior catalytic activity to their AuCS counterparts. The plot of ct/c0 against the reaction time was shown in Figs. 8B, S14, and S15. For instance, MB can be completely reduced by 1.0AuAS within 5 min; nevertheless, only 35% of MB was reduced even after 60 min over 1.0AuCS (some of them may be caused by adsorption). The catalytic performance of these samples on MB reduction is well consistent with that on 4-NP reduction as shown above. The samples with and without templates were all reduced under hydrogen gas at 200 oC, and the resultant catalysts were denoted as 1.0AuAS-g and 1.0AuCS-g, respectively. The catalytic activity of 1.0AuAS-g and 1.0AuCS-g was also explored by the catalytic reduction of MB to LMB. MB can be completely reduced by 1.0AuAS within 5 min; nevertheless, only 73% and 65% of MB was reduced even after 10 min over 1.0AuAS-g and 1.0AuCS-g, respectively (Fig. S16). Wide-angle XRD measurement was also performed to compare the dispersity (Fig. S17). The results show that the Au dispersion degree in 1.0AuAS was obviously better than that in 1.0AuAS-g and 1.0AuCS-g reduced under hydrogen gas. The AuAS samples present excellent recyclability. The sample 1.0AuAS can convert MB completely within 5 min even after recycling for 5 times (Fig. 8C). No Au leaching is detected by ICP, which demonstrates the really heterogeneous catalysis. Furthermore, no aggregation of Au NPs occurs after the fifth catalytic run from XRD patterns (Fig. S18). TG results show no loss of template in the process of reactions (Fig. S19). These results thus demonstrate the excellent stability of resultant AuAS samples. 4. Discussion 4.1 Mechanism for the Dispersion of Au NPs. Many attempts have been made to disperse Au NPs on mesoporous silica due to their diversity in catalytic reactions. It is known that the catalytic activity of Au-containing catalysts relies strongly on the dispersion degree of Au NPs. On the basis of the aforementioned results, it is obvious that Au can be 15

well dispersed on the support AS. The preservation of template in mesopores and the use of SPR strategy are believed to be responsible for the dispersion of Au. Three specific issues are addressed as follows. First, the confined space between template and silica walls in template-occluded mesoporous silica AS is considered. TG results show the decomposition temperature of template P123 in pristine AS is 170 o

C (Fig. 9), which is lower than the decomposition temperature of pure P123 (about 210 oC). This means

that silica walls can catalyze the decomposition of block copolymer (Yin et al., 2012a; Yin et al., 2012b). However, it is noticeable that the decomposition of template takes place at 243 oC in 1.0AuAS. Furthermore, the decomposition temperature increases gradually with the increase of Au contents in AuAS samples (Fig. S20). This gives evidence that Au NPs are successfully incorporated into the confined space between template and silica walls in mesopores, which separates the template P123 from the silica walls to some extent. The catalysis of silica walls on the decomposition of the template is somewhat hindered. Although no thin layer is formed, separate Au NPs can affect the decomposition temperature of organic compounds in mesopores, which was also reported in previous literature (Aslam et al., 2015; Selvakannan et al., 2012). Due to the energy barrier provided by the confined space, aggregation of Au NPs that usually occurs by using conventional methods can be restrained. As a result, the dispersion degree of Au NPs is improved for the samples prepared by use of confined spaces. Apparently, the confined space constructed between template and silica walls in silica nanopores provides an ideal platform for the formation and dispersion of active Au NPs. Of course, in contrast to AuCS derived from template-free mesoporous silica CS, all AuAS samples prepared from templateoccluded AS show better AuAS dispersion, since no confined space exists in the template-free support. Second, the interaction between Au NPs and the template-occluded mesoporous silica AS should be taken into account. XPS results have demonstrated the stronger interaction of Au NPs with the support 16

AS than that with CS. The Au NPs are located in confined spaces between template and silica walls, so they can interact with both of them. A series of obvious IR bands at 2850~3000 and 1350~1500 cm‒1 caused by C‒H stretching and bending vibrations of template are observed in AuAS samples (Fig. S21) (Yin et al., 2012a), which indicates the template P123 is well retained. TG results suggest the strong interaction between the template and Au NPs. Such strong interaction can increase the decomposition temperature of template, while obstruct the aggregation of Au NPs and subsequently improve the dispersion degree of Au NPs. In the meanwhile, IR results show that AuAS possess more silanol groups on the silica walls than AuCS, as indicated by the IR band at 960 cm‒1. The silanol groups on the silica walls can interact with metal nanoparticles, which has been demonstrated by density functional theory (DFT) calculations (Ewing et al., 2015). Higher hydroxyl density will lead to better resistance to aggregation and be conducive to the Au NPs dispersion when Au NPs are introduced to the support. By combining the results of IR and TG, it is clear that the template and the abundant silanols in the template-occluded samples enhance the interaction between Au NPs and the support. Third, the SPR method is utilized to reduce the Au precursor to metallic Au for the first time, which plays an important role in the dispersion of Au NPs. It is known that conventionally, most of these reduction processes were conducted in solutions, which not only requires extra energy to remove solvents but also produces wastes sometimes. More importantly, the surface of support undergoes the soakage-drying procedure, which changes not only the surface state of the host, but also the dispersion or distribution of the guest. Besides, competitive adsorption of solvent molecules will disturb the guesthost interaction, which may hinder the dispersion of guest species. By using the SPR strategy, the weaknesses for reduction in liquid phases can be overcome. The Au precursor is introduced to supports by grinding and the subsequent reduction is conducted in solid phases. The loading of Au on the support was determined by ICP. The results show that Au contents are rather close to the theoretical values 17

(Table S1), which indicates that all of the Au precursors have been reduced to Au nanoparticles. As a result, the dispersion of Au is promoted in comparison with conventional reduction in liquid phases. The typical solution phase synthesis was employed and a reference sample was obtained. The results of XRD show that the diffraction lines of Au nanoparticles in AuAS are obviously weaker than those in the reference sample despite the same Au content (Fig. S22), indicating that the better dispersion of Au by use of our method. The catalytic activity of AuAS was also compared with that of the reference sample. The reduction of 4-NP was finished within 48 min over 1.0AuAS-r (Fig. S23), which is far slower than that over 1.0AuAS (within 24 min). Also, the reduction of MB was completed within 16 min over 1.0AuAS-r, while only 5 min was required for the same reaction over 1.0AuAS (Fig. S24). These results demonstrate that our materials present better dispersion state and higher catalytic activity in contrast to commonly used gold nanoparticles. Furthermore, we examined the stability of our materials by XRD (Fig. S25). The results show that after six months, the intensity of diffraction lines in 1.0AuAS is comparable to that of the fresh sample. This gives evidence of the high stability of our materials. According to the discussion above, it is conclusive that the high dispersion of Au NPs in AuAS can be attributed to the confined space between template and silica walls, the strong interaction between Au NPs and supports, as well as the use of SPR method. 4.2 Clarification of Factors Affecting the Catalytic Activity. Different samples show quite different performance in catalytic reduction reactions. All AuAS samples exhibit much superior activity to their AuCS counterparts. The dispersion degree of Au NPs is considered one of the main factors influencing the catalytic activity. A series of results including XRD, TEM, and UV-vis indicate that all AuAS samples possess higher dispersion degree of Au NPs than AuCS. Moreover, the difference in size increases progressively with increasing Au contents for AuAS 18

and AuCS. A large number of studies have demonstrated that the catalytic activity of Au catalysts strongly depends on the dispersion of Au NPs. Better dispersion of Au NPs results in higher catalytic activity in spite of the same Au contents (Kuzyk et al., 2012). Therefore, all AuAS samples exhibit higher activity in the catalytic reduction of both 4-NP and MB. It is interesting to note that the existence of template P123 in the channels has a significant effect on the catalytic activity. In order to explore the function of template in the catalytic reactions, two control experiments were designed. First, the template P123 was dissolved in the reaction solution followed by the addition of the catalyst 1.0AuCS (denoted as P123+1.0AuCS). The time-dependent absorption spectra of the reaction solution in the presence of P123+1.0AuCS for the catalytic reduction of 4-NP to 4-AP was shown in Figs. S26 and S27. The conversion of 4-NP is 23% in 60 min under the catalysis of P123+1.0AuCS, and such activity is analogous to that of 1.0AuCS (22% at 60 min). Second, the template P123 was introduced to the channels of 1.0AuCS by impregnation (denoted as [email protected]) and then used to catalyze the same reaction. Interestingly, the catalytic activity was improved obviously. The conversion of 4-NP over [email protected] reaches 43% at 24 min and 100% at 60 min, which is lower than that over 1.0AuAS (100% at 24 min) but obviously higher than that over 1.0AuCS (22% at 60 min). Similar results were also observed in the catalytic reduction of MB to LMB (Figs. S28 and S29). Only 34% of MB was reduced by P123+1.0AuCS in 60 min, while 86% of MB was reduced by [email protected] under the same reaction conditions. These results show that, in both reactions, P123+1.0AuCS exhibits comparable catalytic activity to 1.0AuCS, while [email protected] has better catalytic activity than 1.0AuCS but still poorer than 1.0AuAS. Apparently, the addition of template outside the channels of 1.0AuAS (P123+1.0AuCS) have no effect on catalytic activity, while the existence of template in the channels of 1.0AuCS ([email protected]) can enhance the catalytic activity. This may be explained by the enhanced hydrophobicity caused by the presence of organic 19

template in the mesoporous channels, which enriches the organic substrates and favors their interaction with Au NPs. On the contrary, the hydrophilic nature of template-free mesoporous silica CS is not beneficial to the enrichment of organic substrates in mesopores and subsequent conversion on Au NPs. The capacity of catalysts on enriching organic reactants was also reported in previous literature (Liu et al., 2012). It is possible that the separation of products becomes harder. However, one should be very happy to work on the separation issue provided that the catalytic performance is greatly improved. On the basis of the analysis, it is safe to say that the high catalytic activity of AuAS samples can be ascribed to the high dispersion of Au NPs and the existence of template in the channels of supports. 5. Conclusions A facile, efficient SPR strategy was developed for the formation and dispersion of Au NPs in confined spaces, for the first time. High Au dispersion was achieved in the template- occluded mesoporous silica AS, whereas severe Au aggregation takes place for the samples derived from the conventional templatefree support CS. The confined space between template and silica walls, the strong interaction of Au NPs with supports, as well as the utilization of SPR strategy is responsible for the high Au dispersion in AuAS. In contrast to aggregated Au in AuCS, well dispersed Au in AuAS makes the materials highly active in the catalytic reduction of both 4-NP and MB. It is interesting to find that the template P123 retained in the channels of AuAS plays a significant role in enhancing the catalytic activity by enrichment of organic substrates. The excellent catalytic activity of AuAS is attributed to high Au dispersion and the existence of template P123 in mesoporous channels. The present strategy should enable us to incorporate various metal, metal oxides, and even composites into confined spaces, leading to the fabrication of functional materials with well-dispersed and aggregation-resistant active sites that are highly promising for catalytic and adsorptive applications.

20

Acknowledgements We acknowledge the financial support of this work by the National Natural Science Foundation of China (21576137 and 21676138), the Distinguished Youth Foundation of Jiangsu Province (BK20130045), the Fok Ying-Tong Education Foundation (141069), the National Basic Research Program of China (973 Program, 2013CB733504), and the Project of Priority Academic Program Development of Jiangsu Higher Education Institutions. Appendix A. Supplementary material Supplementary information associated with this article, including additional characterization and catalytic data. References Aslam, S., Subhan, F., Yan, Z., Xing, W., Zeng, J., Liu, Y., Ikram, M., Rehman, S., Ullah, R., 2015. Rapid functionalization of as-synthesized KIT-6 with nickel species occluded with template for adsorptive desulfurization. Microporous Mesoporous Mater. 214, 54–63. Azubel, M., Koivisto, J., Malola, S., Bushnell, D., Hura, G.L., Koh, A.L., Tsunoyama, H., Tsukuda, T., Pettersson, M., Häkkinen, H., Kornberg, R.D., 2014. Electron microscopy of gold nanoparticles at atomic resolution. Science 345, 909-912. Boscoboinik, A.M., Manzi, S.J., Tysoe, W.T., Pereyra, V.D., Boscoboinik, J.A., 2015. Directed nanoscale self-assembly of molecular wires interconnecting nodal points using monte carlo simulations. Chem. Mater. 27, 6642-6649. Chen, P.-C., Liu, G., Zhou, Y., Brown, K.A., Chernyak, N., Hedrick, J.L., He, S., Xie, Z., Lin, Q.-Y., Dravid, V.P., O’Neill-Slawecki, S.A., Mirkin, C.A., 2015. Tip-directed synthesis of multimetallic nanoparticles. J. Am. Chem. Soc. 137, 9167-9173. Ciracì, C., Hill, R.T., Mock, J.J., Urzhumov, Y., Fernández-Domínguez, A.I., Maier, S.A., Pendry, J.B., 21

Chilkoti, A., Smith, D.R., 2012. Probing the ultimate limits of plasmonic enhancement. Science 337, 1072-1074. Dai, B., Wang, Q., Yu, F., Zhu, M., 2015. Effect of Au nano-particle aggregation on the deactivation of the AuCl3/AC catalyst for acetylene hydrochlorination. Sci. Rep. 5, 10553-10562. Deng, Y., Cai, Y., Sun, Z., Liu, J., Liu, C., Wei, J., Li, W., Liu, C., Wang, Y., Zhao, D., 2010. Multifunctional mesoporous composite microspheres with well-designed nanostructure: A highly integrated catalyst system. J. Am. Chem. Soc. 132, 8466-8473. Dhar, S., Daniel, W.L., Giljohann, D.A., Mirkin, C.A., Lippard, S.J., 2009. Polyvalent oligonucleotide gold nanoparticle conjugates as delivery vehicles for platinum(IV) warheads. J. Am. Chem. Soc. 131, 14652-14653. Ewing, C.S., Veser, G., Mccarthy, J.J., Johnson, J.K., Lambrecht, D.S., 2015. Effect of support preparation and nanoparticle size on catalyst–support interactions between Pt and amorphous silica. J. Phys. Chem. C 119, 19934-19940. Fang, Q., Xuan, S., Jiang, W., Gong, X., 2011. Yolk-like micro/nanoparticles with superparamagnetic iron oxide cores and hierarchical nickel silicate shells. Adv. Funct. Mater. 21, 1902-1909. Gao, D., Zhang, X., Dai, X., Qin, Y., Duan, A., Yu, Y., Zhuo, H., Zhao, H., Zhang, P., Jiang, Y., Li, J., Zhao, Z., 2016. Morphology-selective synthesis of active and durable gold catalysts with high catalytic performance in the reduction of 4-nitrophenol. Nano Research 9, 3099-3115. Guczi, L., Petö, G., Beck, A., Frey, K., Geszti, O., Molnár, G., Daróczi, C., 2003. Gold nanoparticles deposited on SiO2/Si(100):  Correlation between size, electron structure, and activity in co oxidation. J. Am. Chem. Soc. 125, 4332-4337. Gutiérrez, L.-F., Hamoudi, S., Belkacemi, K., 2012. Effective gold catalyst supported on mesoporous silica decorated by ceria for the synthesis of high value lactobionic acid. Appl. Catal. A: General 425426, 213-223. 22

Inouye, H., Tanaka, K., Tanahashi, I., Hirao, K., 1998. Ultrafast dynamics of nonequilibrium electrons in a gold nanoparticle system. Phys. Rev. B 57, 11334-11340. Kumar, A., Kumar, V.P., Srikanth, A., Vishwanathan, V., Chary, K.V.R., 2016. Vapor phase oxidation of benzyl alcohol over nano Au/SBA-15 catalysts: Effect of preparation methods. Catal. Lett. 146, 35-46. Kuzyk, A., Schreiber, R., Fan, Z., Pardatscher, G., Roller, E.M., Högele, A., Simmel, F.C., Govorov, A.O., Liedl, T., 2012. DNA-based self-assembly of chiral plasmonic nanostructures with tailored optical response. Nature 483, 311-314. López, J.M., Arenal, R., Puértolas, B., Mayoral, Á., Taylor, S.H., Solsona, B., García, T., 2014. Au deposited on CeO2 prepared by a nanocasting route: A high activity catalyst for CO oxidation. J. Catal. 317, 167-175. Lee, W.-S., Cem Akatay, M., Stach, E.A., Ribeiro, F.H., Nicholas Delgass, W., 2014. Gas-phase epoxidation of propylene in the presence of H2 and O2 over small gold ensembles in uncalcined TS-1. J. Catal. 313, 104-112. Li, L., Jin, C., Wang, X., Ji, W., Pan, Y., van der Knaap, T., van der Stoel, R., Au, C.T., 2009. Cyclohexane oxidation over size-uniform Au nanoparticles (SBA-15 hosted) in a continuously stirred tank reactor under mild conditions. Catal. Lett. 129, 303-311. Li, X., Zhu, X.-H., Fang, Y., Yang, H.-L., Zhou, X., Chen, W., Jiao, L., Huo, H., Li, R., 2014. Programmed synthesis of magnetic mesoporous silica nanotubes with tiny Au nanoparticles: a highly novel catalyst system. J. Mater. Chem. A 2, 10485-10491. Lin, J., Abroshan, H., Liu, C., Zhu, M., Li, G., Haruta, M., 2015. Sonogashira cross-coupling on the Au(111) and Au(100) facets of gold nanorod catalysts: Experimental and computational investigation. J. Catal. 330, 354-361. Liu, F., Wang, L., Sun, Q., Zhu, L., Meng, X., Xiao, F.S., 2012. Transesterification catalyzed by ionic 23

liquids on superhydrophobic mesoporous polymers: heterogeneous catalysts that are faster than homogeneous catalysts. J. Am. Chem. Soc. 134, 16948-16950. Liu, X.-Y., Sun, L.-B., Lu, F., Liu, X.-D., Liu, X.-Q., 2013a. Low-temperature generation of strong basicity via an unprecedented guest-host redox interaction. Chem. Commun. 49, 8087-8089. Liu, Y., Liu, B., Wang, Q., Liu, Y., Li, C., Hu, W., Jing, P., Zhao, W., Zhang, J., 2013b. Three dimensionally ordered macroporous Au/CeO2 catalysts synthesized via different methods for enhanced CO preferential oxidation in H2-rich gases. RSC Adv. 4, 5975-5985. Ma, C.Y., Mu, Z., Li, J.J., Jin, Y.G., Cheng, J., Lu, G.Q., Hao, Z.P., Qiao, S.Z., 2010. Mesoporous Co3O4 and Au/Co3O4 catalysts for low-temperature oxidation of trace ethylene. J. Am. Chem. Soc. 132, 2608-2613. Mayeda, M.K., Kuan, W.-F., Young, W.-S., Lauterbach, J.A., Epps, T.H., 2012. Controlling particle location with mixed surface functionalities in block copolymer thin films. Chem. Mater. 24, 26272634. Mei, B., Pougin, A., Strunk, J., 2013. Influence of photodeposited gold nanoparticles on the photocatalytic activity of titanate species in the reduction of CO2 to hydrocarbons. J. Catal. 306, 184189. Mistry, H., Reske, R., Zeng, Z., Zhao, Z.J., Greeley, J., Strasser, P., Cuenya, B.R., 2014. Exceptional size-dependent activity enhancement in the electroreduction of CO2 over Au nanoparticles. J. Am. Chem. Soc. 136, 16473-16476. Mori, K., Verma, P., Hayashi, R., Fuku, K., Yamashita, H., 2015. Color-controlled ag nanoparticles and nanorods within confined mesopores: Microwave-assisted rapid synthesis and application in plasmonic catalysis under visible-light irradiation. Chem. Eur. J. 21, 11885-11893. Raghuwanshi, V.S., Ochmann, M., Polzer, F., Hoell, A., Rademann, K., 2014. Self-assembly of gold nanoparticles on deep eutectic solvent (DES) surfaces. Chem. Commun. 50, 8693-8696. 24

Rombi, E., Cutrufello, M.G., Cannas, C., Occhiuzzi, M., Onida, B., Ferino, I., 2012. Gold-assisted E[prime or minute] centres formation on the silica surface of Au/SBA-15 catalysts for low temperature CO oxidation. Phys. Chem. Chem. Phys. 14, 6889-6897. Sacaliuc, E., Beale, A.M., Weckhuysen, B.M., Nijhuis, T.A., 2007. Propene epoxidation over Au/TiSBA-15 catalysts. J. Catal. 248, 235-248. Savage, K.J., Hawkeye, M.M., Esteban, R., Borisov, A.G., Aizpurua, J., Baumberg, J.J., 2012. Revealing the quantum regime in tunnelling plasmonics. Nature 491, 574-577. Selvakannan, P., Mantri, K., Tardio, J., Bhargava, S.K., 2012. High surface area Au-SBA-15 and AuMCM-41 materials synthesis: Tryptophan amino acid mediated confinement of gold nanostructures within the mesoporous silica pore walls. J. Colloid Interface Sci. 394, 475-484. Tschulik, K., Ngamchuea, K., Ziegler, C., Beier, M.G., Damm, C., Eychmueller, A., Compton, R.G., 2015. Core–shell nanoparticles: Characterizing multifunctional materials beyond imagingdistinguishing and quantifying perfect and broken shells. Adv. Funct. Mater. 25, 5149-5158. Valente, C., Çalimsiz, S., Hoi, K.H., Mallik, D., Sayah, M., Organ, M.G., 2012. The development of bulky palladium NHC complexes for the most-challenging cross-coupling reactions. Angew. Chem., Int. Ed. 51, 3314-3332. Wang, C., Salmon, L., Li, Q., Igartua, M.E., Moya, S., Ciganda, R., Ruiz, J., Astruc, D., 2016. From mono to tris-1,2,3-triazole-stabilized gold nanoparticles and their compared catalytic efficiency in 4nitrophenol reduction. Inorg. Chem. 55, 6776-6780. Wang, L.-N., Li, Z.-Y., Liu, Q.-Y., Meng, J.-H., He, S.-G., Ma, T.-M., 2015. CO oxidation promoted by the gold dimer in Au2VO3− and Au2VO4− clusters. Angew. Chem., Int. Ed. 54, 11720-11724. Wang, L., Meng, X., Wang, B., Chi, W., Xiao, F.-S., 2010. Pyrrolidone-modified SBA-15 supported Au nanoparticles with superior catalytic properties in aerobic oxidation of alcohols. Chem. Commun. 46, 5003-5005. 25

Wang, L., Wang, H., Hapala, P., Zhu, L., Ren, L., Meng, X., Lewis, J.P., Xiao, F.-S., 2011. Superior catalytic properties in aerobic oxidation of olefins over Au nanoparticles on pyrrolidone-modified SBA-15. J. Catal. 281, 30-39. Wang, M., Anghel, A.M., Marsan, B., Cevey Ha, N.-L., Pootrakulchote, N., Zakeeruddin, S.M., Grätzel, M., 2009. CoS supersedes Pt as efficient electrocatalyst for triiodide reduction in dye-sensitized solar cells. J. Am. Chem. Soc. 131, 15976-15977. Wang, W., Ye, M., He, L., Yin, Y., 2014. Nanocrystalline TiO2-catalyzed photoreversible color switching. Nano Lett 14, 1681-1686. Wang, Y.M., Ying Wu, Z., Zhu, J.H., 2004. Surface functionalization of SBA-15 by the solvent-free method. J. Solid State Chem. 177, 3815-3823. Xiao, Y., Jaskula-Sztul, R., Javadi, A., Xu, W., Eide, J., Dammalapati, A., Kunnimalaiyaan, M., Chen, H., Gong, S., 2012. Co-delivery of doxorubicin and SiRNA using octreotide-conjugated gold nanorods for targeted neuroendocrine cancer therapy. Nanoscale 4, 7185-7193. Yang, J., Shen, D., Zhou, L., Li, W., Li, X., Yao, C., Wang, R., El-Toni, A.M., Zhang, F., Zhao, D., 2013. Spatially confined fabrication of core-shell gold nanocages@mesoporous silica for nearinfrared controlled photothermal drug release. Chem. Mater. 25, 3030-3037. Yang, L., Qi, M., 2013. Rapid fabrication of confined Au nanoparticles with tunable sizes and morphologies by a simple glucose-assisted vacuum impregnation method. Mater. Lett. 98, 74-77. Yang, Y., Mao, Y., Wang, B., Meng, X., Han, J., Wang, C., Yang, H., 2016. Facile synthesis of cubical Co3O4 supported Au nanocomposites with high activity for the reduction of 4-nitrophenol to 4aminophenol. RSC Adv. 6, 32430-32433. Yao, T., Cui, T., Wang, H., Xu, L., Cui, F., Wu, J., 2014. A simple way to prepare Au@polypyrrole/Fe3O4 hollow capsules with high stability and their application in catalytic reduction of methylene blue dye. Nanoscale 6, 7666-7674. 26

Yatabe, T., Jin, X., Yamaguchi, K., Mizuno, N., 2015. Gold nanoparticles supported on a layered double hydroxide as efficient catalysts for the one-pot synthesis of flavones. Angew. Chem., Int. Ed. 54, 13302-13306. Yin, Y., Jiang, W.-J., Liu, X.-Q., Li, Y.-H., Sun, L.-B., 2012a. Dispersion of copper species in a confined space and their application in thiophene capture. J. Mater. Chem. 22, 18514-18521. Yin, Y., Tan, P., Liu, X.-Q., Zhu, J., Sun, L.-B., 2014. Constructing a confined space in silica nanopores: an ideal platform for the formation and dispersion of cuprous sites. J. Mater. Chem. A 2, 3399-3406. Yin, Y., Xue, D.-M., Liu, X.-Q., Xu, G., Ye, P., Wu, M.-Y., Sun, L.-B., 2012b. Unusual ceria dispersion formed in confined space: a stable and reusable adsorbent for aromatic sulfur capture. Chem. Commun. 48, 9495-9497. Yonemoto, B.T., Hutchings, G.S., Jiao, F., 2014. A general synthetic approach for ordered mesoporous metal sulfides. J. Am. Chem. Soc. 136, 8895-8898. Yoon, M., Choi, J., Cho, J., 2013. Multifunctional colloids with reversible phase transfer between organic and aqueous media via layer-by-layer assembly. Chem. Mater. 25, 1735-1743. Yu, N., Ding, Y., Lo, A.-Y., Huang, S.-J., Wu, P.-H., Liu, C., Yin, D., Fu, Z., Yin, D., Hung, C.-T., Lei, Z., Liu, S.-B., 2011. Gold nanoparticles supported on periodic mesoporous organosilicas for epoxidation of olefins: Effects of pore architecture and surface modification method of the supports. Microporous Mesoporous Mater. 143, 426-434. Yue, M.B., Sun, L.B., Cao, Y., Wang, Y., Wang, Z.J., Zhu, J.H., 2008. Efficient CO2 capturer derived from as-synthesized MCM-41 modified with amine. Chem. Eur. J. 14, 3442-3451. Zhang, Q., Fan, X., Wang, H., Chen, S., Quan, X., 2016. Fabrication of Au/CNT hollow fiber membrane for 4-nitrophenol reduction. RSC Adv. 6, 41114-41121. Zhang, X., Su, Z., 2012. Polyelectrolyte-multilayer-supported Au@Ag core-shell nanoparticles with 27

high catalytic activity. Adv. Mater. 24, 4574-4577. Zhao, D., Feng, J., Huo, Q., Melosh, N., Fredrickson, G.H., Chmelka, B.F., Stucky, G.D., 1998a. Triblock copolymer syntheses of mesoporous silica with periodic 50 to 300 angstrom pores. Science 279, 548-552. Zhao, D., Feng, J., Huo, Q., Melosh, N., Fredrickson, G.H., Chmelka, B.F., Stucky, G.D., 1998b. Triblock copolymer syntheses of mesoporous silica with periodic 50 to 300 angstrom pores. Science 279, 548-552. Zhu, W., Michalsky, R., Metin, O., Lv, H., Guo, S., Wright, C.J., Sun, X., Peterson, A.A., Sun, S., 2013. Monodisperse Au nanoparticles for selective electrocatalytic reduction of CO2 to CO. J. Am. Chem. Soc. 135, 16833-16836.

28

Scheme 1 The SPR strategy for fabrication of (A) aggregated Au NPs in conventional mesopores and (B) well-dispersed Au NPs in confined spaces (template-occluded mesopores).

29

Fig. 1 (A) Low-angle and (B) wide-angle XRD patterns of AS, CS, AuAS, and AuCS samples.

30

Fig. 2 Dark-field TEM images of (A) 1.0AuAS and (B) 1.0AuCS. EDX elemental mapping of Si on (C) 1.0AuAS and (D) 1.0AuCS as well as Au on (E) 1.0AuAS and (F) 1.0AuCS. Insets in Figures A and B are photographs of 1.0AuAS and 1.0AuCS, respectively.

31

Fig. 3 HRTEM images of 1.0AuAS viewing (A) normal to the pore axis and (B) along the pore axis of host AS.

32

Fig. 4 (A) N2 adsorption/desorption isotherms and (B) PSDs of AS, CS, 0.5AuAS, 1.0AuAS, 1.0AuCS, and 2.0AuAS.

33

Fig. 5 XPS peak fitting of Au 4f7/2 and 4f5/2 spectra of (A) 1.0AuAS, (B) 1.0AuCS, (C) 0.5AuAS, and (D) 2.0AuAS.

34

Fig. 6 Diffuse reflectance UV-vis spectra of different samples.

35

Fig. 7 (A) The reaction equation for the catalytic reduction of 4-NP to 4-AP. (B) The plots of ct/c0 against the reaction time over different samples. (C) The reusability of 1.0AuAS as the catalyst for the reduction of 4-NP. Insets in Figure B are the photographs of 4-NP solutions before and after reaction. Reaction conditions: 4-NP amount was 0.02 mL (0.005 M), NaBH4 amount was 0.2 mL (0.2 M), solvent amount was 2.5 mL, and catalyst amount was 2 mg. The reaction time for the recycle test was 24 min. 36

Fig. 8 (A) The reaction equation for the catalytic reduction of MB to LMB. (B) The plots of ct/c0 against the reaction time over different samples. (C) The reusability of 1.0AuAS as the catalyst for the reduction of MB. Insets in Figure B are the photographs of MB solutions before and after reaction. Reaction conditions: MB amount was 2.5 mL (25 ppm), NaBH4 amount was 0.2 mL (0.2 M), solvent amount was 2.5 mL, and catalyst amount was 1 mg. The reaction time for the recycle test was 5 min.

37

Fig. 9 (A) TG and (B) DTG curves of AS, 1.0AuAS, and 1.0AuCS.

Graphical Abstract

38

Highlights 

The confined space hinders the aggregation of Au during reduction and maintains the catalytic 39

activity. 

The solid-phase reduction strategy avoids the competitive adsorption of solvent molecules and restricts the diffusion of Au species.



Highly dispersed Au species are highly active for catalytic reduction reactions.



The preservation of template is beneficial to not only the dispersion of Au NPs, but also the enhancement of catalytic activity.

40