Chun Pei,Shangjun Chen,Rongrong Song,Fei Lv,Ying Wan

Key Laboratory of Resource Chemistry of Ministry of Education,Shanghai Key Laboratory of Rare Earth Functional Materials,and Department of Chemistry,Shanghai Normal University,Shanghai 200234,China

Keywords:Gold catalyst Ordered mesoporous carbon Large pores Reduction Nitroarenes

ABSTRACT Simple encapsulation of 3 nm gold nanoparticles in ordered mesoporous carbon with large pores of 17 nm and thick pore walls of 16 nm was achieved by a metal–ligand coordination assisted-selfassembly approach.Polystyrene-block-polyethylene-oxide (PS-b-PEO) diblock copolymer with a large molecular weight of the PS chain and mercaptopropyltrimethoxysilane were used as the template and the metal ligand,respectively.Small-angle X-ray scattering,X-ray diffraction,transmission electron microscopy,and X-ray photoelectron spectroscopy showed that monodispersed aggregation-free gold nanoparticles approximately 3 nm in size were partially embedded in the large open pore structure of the ordered mesoporous carbon.The strong coordination between the gold species and the mercapto groups and the thick porous walls increased the dispersion of the gold nanoparticles and essentially inhibited particle aggregation at 600 °C.The gold nanoparticles in the ordered mesoporous carbon are active and stable in the reduction of nitroarenes involving bulky molecules using sodium borohydride as a reducing agent under ambient conditions (30 °C) in water.The large interconnected pore structure facilitates the mass transfer of bulky molecules.

1.Introduction

The catalytic reduction of nitro compounds using noble metal catalysts is one of the most cost-efficient and promising routes for the preparation of anilines or amines,which are key intermediates for the synthesis of fine chemicals,agrochemicals,pharmaceuticals,and polymers [1,2].Gold nanocatalysts represent a second generation of catalysts for such a reduction reaction,showing a higher selectivity than the platinum group metals [3,4].In order to improve the stability and catalytic activity of the gold nanocatalysts,many efforts have focused on stabilizing and dispersing gold nanoparticles on porous solid supports with high surface areas[5–9],with activated carbons being one of the most important industrial catalyst carriers [10].However,the pore system of activated carbons consists almost exclusively of micro-pores,which may cause a diffusion limitation particularly in catalytic conversion involving bulky molecules [11].This would lead to a compromise between catalytic activity and the mass-transport of reactants and products in or out of channels in the catalysts [12].

Ordered mesoporous carbon materials with large accessible pores greater than 10 nm,an interconnected pore structure,and a highly specific surface area may significantly improve the situation [13–15].However,the traditional synthesis methods such as deposition–precipitation (DP) and impregnation [16,17] cannot immobilize monodispersed Au nanoparticles on carbon materials due to their weak interaction with the inert support [18].The nanocasting method involving a gold-containing silica hard template can also be used to prepare carbon supported gold nanocatalysts,but the remove of the silica template with a strong acid or base may cause particle aggregation and environmental contamination [19,20].Sol-gel preparation is an alternative way to load gold nanoparticles on porous carbon,but residual organic stabilizers on the gold nanoparticles or supports may disturb the catalytic performance of the catalysts and cannot be ignored [21,22].

The one-step synthesis of a mesoporous carbon with ordered large pores supporting metal nanocatalysts could be preferable[23–25].For example,Wiesner et al.synthesized a Pt-containing polymeric carbonaceous material with open and large pores(17 nm) by the self-assembly of block copolymer (poly(isopreneblock-dimethylaminoet-hyl methacrylate),PI-b-PDMAEMA) with platinum nanoparticles stabilized by a thiol-containing ionic liquid[26].The carbon component mostly comes from pyrolysis of the PI block at relatively low temperatures.As a consequence,the catalyst had a carbonaceous framework with a low carbon content(18% (mass)) and a low surface area of only 18 m2·g-1indicating partially open mesopores,and a high sulfur residue (1% (mass))which might poison catalytic reactions.The same group,then used sulfur-free ligands such as cyclooctadiene,and added some nanoparticles,for example Nb2O5,to enhance the thermal stability of the catalyst[27].A moderate surface area of 145.3 m2·g-1and a uniform pore size distribution centered around 29 nm was obtained.However,the mesoporous metal oxide–carbon composite had only some degree of long-range order,and the Pt or PtPb particles were greater than 10 nm,which is larger than ideal[28].Small particles usually have a higher activity and selectivity than larger ones as a catalyst due to larger number of exposed defects and edges or corners.However,because the melting temperature of gold decreases steeply with decreasing particle size,the synthesis of aggregation-free gold nanoparticles with dimensions below 5 nm inside an ordered mesoporous carbon with large pores is more challenging and rare [29].

Recently,we have developed a self-assembly strategy to immobilize 4 nm gold nanoparticles in ordered mesoporous carbon with large pores of 12 nm.We can now update this self-assembly approach and report the encapsulation of 3 nm-gold nanoparticles in an ordered mesoporous carbon with 17 nm-mesopores (Au/OMC-17).In this approach,a polystyrene-block-polyethylene oxide(PS-b-PEO) diblock copolymer,mercaptopropyltrimethoxysilane(MPTMS),HAuCl4,and low-polymerized phenolic resin were used as the template,coordination agent,gold precursor,and carbon source,respectively.Metal-ligand coordination-assisted selfassembly leads to a face-centered cubic organic–inorganic mesostructured hybrid material.Heat treatment of this material under flowing inert gas at 600 °C simultaneously decomposed the diblock copolymer,reduced the metal precursor to nanoparticles,and carbonized the resin component,while transforming the material into a mesoporous structure.The catalysts produced have a highly ordered 3D cubic mesostructure,high surface areas(403–434 m2·g-1),large pore volumes (0.29–0.36 cm3·g-1),and tunable large pore sizes (7–17 nm).Monodispersed aggregation-free gold nanoparticles approximately 3 nm in size were separated by the thick pore walls,and had a high thermal stability at 600 °C.The gold catalyst with large pores (Au/OMC-17) showed a higher catalytic activity than the gold catalyst with small 4 nm mesopores(Au/SC-4) in the reduction of nitroarenes with a large steric hindrance using sodium borohydride (NaBH4) as a reducing agent under ambient conditions (30 °C) in water.The large interconnected pore structure of Au/OMC-17 facilitates the mass transfer of bulky molecules.The gold catalysts are stable after eight runs with no obvious gold leaching or activity loss.

2.Materials and Methods

2.1.Materials

Styrene (St,＞99% (mass)),monomethoxy poly(ethylene oxide)(PEO,99% (mass),2-bromoisobutyryl bromide (99% (mass)),tetrahydrofuran (THF,＞99% (mass)),N,N-dimethylformamide(DMF,＞99% (mass)),CuBr (99% (mass)),pyridine (99% (mass)),N,N,N′-N′,N′-pentamethyldiethylenetriamine (PMDETA,99% (mass)),mercaptopropyltrimethoxysilane (MPTMS,98% (mass)),Pluronic F127 (EO106PO70EO106,99% (mass)),chloroauric acid tetrahydrate(HAuCl4·4H2O,Au minimum 47.8% (mass)),4-nitrophenol (4-NP,＞99% (mass)),and 4-(tert-butyl)-2-nitrophenol (4-TB-2-NP,＞99%(mass)),2,6-dimethylnitrobenzene(2,6-DMNB,＞99%(mass))were purchased from Sigma-Aldrich Co.The styrene and THF were dried over calcium hydride and distilled under reduced pressure before use.The water used in all experiments was deionized.All the other reagents were of analytical grade and were purchased from the Shanghai Chemical Company and used as received without any further purification.

2.2.Synthesis of 3 nm gold nanoparticles encapsulated in ordered mesoporous carbon with large pores

PS-b-PEO diblock polymers with different average molecular weights(Mn)were prepared by a simple atom transfer radical polymerization(ATRP)method[30].Details of the synthesis of copolymers are given in the Supplementary Material.In a typical synthesis of Au/OMC-17,0.2 g PS-b-PEO (Mn=30630 g·mol-1),0.5 ml of HAuCl4solution (gold concentration in THF:48.5 mmol·L-1),0.098 g MPTMS,0.1 g HCl solution (2.0 mol·L-1),and 5.0 g of phenolic resin solution (20% (mass) in THF) were mixed in 10.0 g THF at 40 °C.After stirring for 1 h,the solution was poured into multiple dishes.Films were cast by evaporation of the solvent at 40 °C and were further annealed at 100 °C for 24 h.Subsequent heat treatment was carried out under highpurity nitrogen with heating rates of 1 °C·min-1from room temperature to 350°C and then 5°C·min-1to 600°C,where it was kept for 3 h.The obtained black sample was denoted Au/OMC-17.By tuning the molecular weight of PS-b-PEO to 20,750 and 16070 g·mol-1,catalysts with pore sizes of 12.1 nm (Au/OMC-12)and 7.3 nm (Au/OMC-7),respectively,were obtained.Gold free samples (OMC-m,where m is the pore size in nm) were prepared using the same synthesis procedure in the absence of gold.A reference gold catalyst with a small pore size (Au/SC-4) was also synthesized using commercial Pluronic F127 as the structuredirecting agent using the established published procedure(Supplementary Material) [31].A commercial gold catalyst (Au/C) was purchased from Haruta Gold Incorporated.

2.3.Characterization

Proton nuclear magnetic resonance(1H NMR)spectra were collected on a 400-MHz spectrometer (Bruker,Germany) at 25 °C using tetramethylsilane as an internal standard and CDCl3as a solvent.Fourier transform infrared(FT-IR)spectra were collected on a Nicolet Fourier spectrophotometer (USA) using KBr pellets.Gel permeation chromatography (GPC) measurement was recorded on an Agilent 1100 gel permeation chromatography unit using THF as an eluent.Small-angle X-ray scattering (SAXS) measurements were performed on a Nanostar U SAXS system(Bruker,Germany) using Cu Kα radiation (40 kV,35 mA).X-ray photoelectron spectroscopy (XPS) measurements were taken on a Perkin-Elmer PHI 5000CESCA system with a base pressure of 133.3×10-9Pa,and X-ray diffraction(XRD)measurements were made on a Rigaku D max-3C diffractometer using Cu Kα radiation (40 kV,20 mA,λ=0.15908 nm).N2sorption isotherms were obtained on a Micromeritics TriStar II 3020 analyzer at 77 K.The specific surface areas(SBET)and the pore volumes and pore size distributions were calculated by the Brunauer-Emmett-Teller (BET) method and the Broekoff-de Boer (BdB) model,respectively.Transmission electron microscopy (TEM) images of the samples were recorded on a JEM 2100 microscope operating at 200 kV.A JEM 2100F microscope equipped with a field emission gun operating at 200 kV was used for the scanning transmission electron microscopy-high angle annular dark field (STEM-HAADF) imaging.The size distribution of the gold nanoparticles in each catalyst was determined by randomly counting at least 200 individual Au nanoparticles with an approximate error of ±5%.The gold content of each catalyst was determined by inductively coupled plasma-atomic emission spectrometry (ICP-AES,Varian VISTA-MPX).

2.4.Selective reduction of nitroarenes

Reduction reactions were performed in a glass reactor using NaBH4as a reducing agent.Typically,3.5 ml of 4-NP(0.0193 mmol)aqueous solution was placed in a glass reactor,followed by the addition of 100 ml of freshly prepared aqueous NaBH4(176 mmol·L-1) and the supported Au catalyst (20 mg,Au/OMC-17) under stirring at 800 r·min-1at 30 °C.The reaction was monitored by UV–Vis spectrophotometer and high-performance liquid chromatography(HPLC,Agilent 1100 series)using a C18 reversedphase column [ZORBAX Eclipse Plus C18 (4.6 × 100 mm2)] and a diode-array UV–Vis detector [32].The conversion and selectivity determined on the basis of the UV–Vis and HPLC analyses are in good agreement with each other,within an error of ±5%.The catalytic results are given in terms of the apparent rate constant(kapp),initial reaction rate (r0,mmol of reacted reactant per mmol of Au and time of the initial reaction period in minute),and turnover Frequency (TOF,molecules of nitroarene converted per surface atom of Au per minute).The absorbance values,A0and At,are the 4-NP absorptions at concentrations of c0and ct,respectively,where c0is the initial concentration and t is the time.The kappvalue was estimated from the slope of the linear relationship between ln(ct/c0) and t.The TOF value was calculated at a conversion below 20%and was reproducible to within±8%.Here,we used the method established by Rossi et al.to calculate the exposed surface area of the dispersed Au particles [33].A hot infiltration test and a solid-trapping experiment using mercapto-functionalized ordered mesoporous silica (SH-SBA-15) were conducted to determine the presence of any soluble active gold during the reaction.For the form test,the solid catalyst was filtered out of the reaction solution after 2 min-reaction and the filtrate was monitored for continued activity.For the solid-trapping experiment,SH-SBA-15 and the supported gold catalysts with an SH:Au (mol:mol) ratio of 35 were placed in the reaction batch together.

Recycle tests were performed under the same reaction conditions except for the use of a recovered catalyst.The solid catalysts were separated from the reaction mixture after complete reaction and were thoroughly washed with plenty of ethanol and water.After drying under a vacuum at 80 °C,the catalyst was used for the next reaction.In order to ensure the same amount of the catalyst was used in each run,several parallel reactions were carried out.The aqueous solution after each cycle was collected to determine the gold leaching concentration.

3.Results and Discussion

3.1.Synthesis of 3 nm gold nanoparticles encapsulated in ordered mesoporous carbon with large pores by self-assembly

3 nm gold nanoparticles encapsulated in ordered mesoporous carbon with large pores were synthesized by a coordination-assisted self-assembly approach [23].Mercaptopropyltrimethoxysilane(MPTMS),HAuCl4,polystyrene-block-polyethylene oxide (PS-b-PEO)diblock copolymer,and low-polymerized phenolic resin were used as coordination agent,metal source,structure-directing agent,and carbon precursor,respectively.The PS-b-PEO diblock copolymer was prepared by a simple atom transfer radical polymerization(ATRP) between macroinitiator PEO-Br and styrene (Scheme S1,see Supplementary Material).Three diblock copolymers with different average molecular weights(Mn)were synthesized by changing the polymerization reaction time.The chemical structures and compositions of PEO-Br and the diblock copolymers were verified by Fourier transform infrared (FT-IR) spectra (Fig.S1),1HNMR spectra,and GPC traces (Fig.S2).The FT-IR spectrum of PEO-Br has an intense absorption band at 1110 cm-1originating from the EO units,and a weak band at 1733 cm-1assigned to the carbonyl (-C＝O) stretch groups formed by the acetylation reaction.The characteristic new bands at 1450,1493,1601 and 698 cm-1in the FT-IR spectrum of PS-b-PEO indicate the presence of styrene units in the diblock copolymer [34].The proton resonance signals around 1.95 and 3.66 ppm are respectively assigned to the aliphatic protons of the methyl groups and the ethylene oxide units of PEO-Br.In the spectrum of PS-b-PEO,additional signals at around 6.46–6.60 and 7.04–7.20 ppm were observed,corresponding to the protons on the benzene ring of PS.The compositions of these diblock copolymers can be approximately formulated as PS106-b-PEO125,PS150-b-PEO125,and PS245-b-PEO125estimated by their Mnvalues.GPC traces show a polydispersity index (PDI) of approximately 1.10 for all the PS-b-PEO diblock copolymers,indicating a narrow molecular weight distribution (Table S1).These results demonstrated the successful preparation of the diblock copolymer PS-b-PEO using the ATRP method.

The SAXS patterns of both the pristine OMC-17 and the supported gold catalyst Au/OMC-17 synthesized with PS245-b-PEO125show a first-order peak at a similar q value of 0.23 nm-1,along with two higher-order peaks at higher q values of 0.43 nm-1and 0.78 nm-1(Fig.1).This sequence is expected for a face-centered cubic ordered mesostructure with the space group Fm3-m [35].These findings suggest the successful co-assembly of the diblock copolymer,MPTMS and phenolic resins,and negligible disturbance of the ordered mesostructure of the material in the presence of gold.The lattice parameter (a0) of Au/OMC-17 was calculated to be 47.3 nm,indicating that the mesostructured has a large unit cell.Similar SAXS patterns were observed for OMC-12,OMC-7,Au/OMC-12,and Au/OMC-7,which were templated by PS-b-PEO with a relatively small Mn.The lattice parameter value a0of the catalysts gradually decreases to 37.5 nm for OMC-12 and Au/OMC-12,and 29.4 nm for OMC-7 and Au/OMC-7.A reduction in the PS fractions of the diblock copolymers does not distinctly affect the mesostructure of the catalysts regardless of the presence of gold.The highly ordered 3D cubic,mesoporous structure is retained as shown in Fig.1.

The N2adsorption–desorption isotherms of both OMC-17 and Au/OMC-17 show typical type-IV curves(Fig.2A).A sharp capillary condensation step in the relative pressure range 0.42–0.95 was observed for these samples,indicating large and uniform mesopores.The large H2-type hysteresis loops with delayed capillary evaporation located at middle relative pressures imply caged mesopores with a window size below 4.0 nm.The most probable pore size measured from the adsorption branch is as large as～17.0 nm for both samples (Fig.2B),which is in good agreement with the SAXS results.The BET surface areas and pore volumes are in the range 491–434 m2·g-1and 0.39–0.36 cm3·g-1,respectively (Table 1).Similar isotherms are also detected for OMC-7,Au/OMC-7,OMC-12,and Au/OMC-12(Fig.2C).The mean pore sizes calculated from the BdB sphere model decrease to about 12.0 nm for OMC-12 and Au/OMC-12,and to about 7.0 nm for OMC-7 and Au/OMC-7(Fig.2D).These results also indicate that the incorporation of gold nanoparticles in the pore walls has a negligible effect on the uniformity of the pores.The wall thickness of the supported gold catalysts was very thick,ranging from 16.3 to 13.3 nm and decreased with the decrease in molecular weight of the PS chains[34].

The broad diffraction peak of the XRD pattern located at 38.2°of Au/OMC-17 is identified as the(111)plane of metallic gold(JCPDS card no.04-0784,Fig.S3).The other diffraction planes of the gold nanoparticles are relatively weak and not very evident.The average particle size was estimated to be about 3.0 nm using the Debye-Scherer formula,and the gold content assessed by ICP is 0.5% (mass).The small particle size and low gold content result in the broad diffraction peaks in the XRD patterns.Bigger gold nanoparticles were found for Au/OMC-12 and Au/OMC-7 with a smaller pore size and similar gold content (Table 1),as shown by the relatively narrow and intense diffraction peaks located at 38.2°.The larger pore size and the thicker pore walls of Au/OMC-17 may lead to the formation of smaller gold nanoparticles [36].

Fig.1.(A)SAXS patterns of the pristine OMC and(B)the supported gold catalysts.A:(a)OMC-17,(b)OMC-12,and(c)OMC-7.B:(a)Au/OMC-17,(b)Au/OMC-12,and(c)Au/OMC-7,and (d) Au/OMC-17 after 8 runs (Au/OMC-17-R) in the reduction of 4-NP.

TEM and HAADF-STEM images of Au/OMC-17 show wellordered mesoscopic arrays with uniform and spherical large mesopores viewed along the [211] and [110] directions (Fig.3).Indistinct interruption of the ordered packing is observed in the presence of the gold species.The lattice parameter a0estimated from TEM images is about 43.5 nm.These findings are consistent with the analysis of the SAXS patterns and confirm the highly ordered cubic structure with Fm3-m symmetry of the gold catalyst[29].Notably,gold nanoparticles with an approximate size of 3.0 nm are observed in the ordered channels of Au/OMC-17.They are well-dispersed with a negligible number of large agglomerates.

These results show that the gold nanoparticles encapsulated in the large open pores are highly stable and accessible.The formation of gold nanoparticles is also confirmed by XPS spectra(Fig.4).The spectrum of the Au 4f core level of Au/OMC-17 is characterized by two high-intensity peaks with binding energies of approximately 84.2 and 87.9 eV,which are respectively assigned to the 4f7/2and 4f5/2core levels of metallic Au0[23,37].The presence of the oxidized gold species is minor (＜15%).These findings imply the generation of metallic gold nanoparticles due to the reduction of the gold species during the calcination process [25].No peaks of any S species can be detected in XPS spectra of Au/OMC-17,which indicates the completely elimination of the metal-ligands and the depletion of sulfur in the final catalysts(Fig.S4).

For comparison,a gold catalyst with small pore size of 4.2 nm was synthesized using triblock copolymer F127 as the template(Au/SC-4).The catalyst has a similar surface area and pore volume to that of the Au/OMC catalysts (Table 1,Fig.S5) and TEM images showed that the most probable size of the gold nanoparticles was about 9.0 nm (Fig.S6).The Au 4f XPS spectrum also indicates the metallic state of the gold nanoparticles (Fig.S7).The Au/C catalyst is commercially available and consists of 6.8 nm gold nanoparticles on activated carbon.The detailed characterization was shown in supplementary material (Figs.S8,S9).

Table 1 Structural and textural parameters of pristine OMC and supported gold catalysts

The immobilization of 3 nm gold nanoparticles in ordered mesoporous carbon with large pores is primarily attributed to the self-assembly of the large molecular weight diblock copolymer with gold ions,thiol-containing silane,and phenolic resin.An illustration of the fabrication of the gold catalyst is shown in Fig.5.The thiol-containing silane reacts with resin and the thiol group of the silane serves as a ligand to stabilize gold ions [38].Evaporation of solvents induces the self-assembly of diblock copolymers with PS block inside the cores and PEO block surrounding the shells.The gold-containing composites with abundant hydroxyl groups selectively swell the PEO block,while the PS blocks aggregate to spheres.The as-made hybrid material has a face-centered cubic ordered mesostructure with spheres consisting of PS in a matrix of PEO,resin,silicate and gold precursor.The formation of the mesostructure is originated by the self-assembly of PS-b-PEO to ordered 3D mesostructure,and the simultaneous hydrogenbonding interaction between hydrophilic EO segments of the template and hydroxyl groups of resins and silicates surrounding them[39–41].In the present self-assembly,the acidic solution is critical to balance the hydrolysis process of precursor and assembly.Under acidic conditions,hydrolysis of MPTMS is faster than condensation.Small fragments of low-polymerized resin and partially hydrolyzed silanols would participate in the co-assembly.Finally,once the mesostrcuture assembles the condensation of the silicate and polymers facilitate the stabilization of solid materials.However,if silicate condenses to large framework prior to the complete hydrolysis for example,under the basic conditions,the coassembly would be disturbed,and phase separation would happen.The TEM image of the catalyst at the growth stage of 350 °C(Fig.S10) indicates that heat treatment of the as-made hybrid material at 350°C removes most of the PS blocks,large open mesopores,gold precursors are thermally decomposed and small gold nanoparticles are then formed.During calcination at 600°C,metallic gold nanoparticles are maintained in the absence of aggregation to large particles and moving to outside mesostructure.The PEO block mixed with the resin and silicate precursors is decomposed and pore walls consisting of amorphous graphitic-like carbon materials with a pinch of silica are formed [38].Because the gold precursor preferentially resides in the PEO domains,these pore walls are expected to be decorated with gold nanoparticles.The pore walls after carbonization are composed by at least two parts.One part is from the decomposition of phenolic resins and PEO blocks.A relatively high ratio of phenolic resin to the template increases the thickness of the pore walls because the PEO segments easily decompose to release small molecules.The large,hydrophobic PS segments with sp2-hybridized-carbon can be partially converted into carbon residuals,which further increase the thickness of the inner walls of the large pores of the mesoporous carbon[42].The thick porous walls not only increase the dispersion of the gold nanoparticles but also inhibit the mobility and agglomeration of the metal particles.These findings are similar to the Au nanoparticles riveted to carbon in Au/TiO2[43]and the Pt nanoparticles riveted to carbon in Pt/Vulcan XC-72 [44].The formation of 3 nm-gold nanoparticles is highly related to the strong bonding between the thiol group in the silane and the gold species,which has been proven in our previous studies[23,25,32,38].The coordination between gold species and thiol group prevents the reduction of gold at low temperatures during pyrolysis.The aggregation of gold nanoparticles is inhibited by the confinement of the relatively rigid thick pore walls,and the particle size of the gold nanoparticles can be well controlled [29,31].The monodispersed gold nanoparticles encapsulated in the pore walls may be riveted by the carbonaceous material preventing their mobility and agglomeration.We have previously found that dispersed gold nanoparticles with uniform sizes between 3 and 9 nm were obtained at a high pyrolysis temperature of 900 °C [25,32].Therefore,in this case,the inhibition of aggregation of gold nanoparticles in large mesopores can also be expected due to the confinement of the relatively rigid thick pore walls at a higher temperature.Moreover,the pore sizes of the mesoporous carbon supports are highly dependent on the size of the PS blocks and are simply changed by changing the molecular weights of the PS blocks.As a result,gold nanoparticles about 3 nm in size encapsulated in ordered mesoporous carbon with pores as large as 17 nm can be formed.

Fig.2.(A,C)N2 adsorption–desorption isotherms and(B,D)pore size distribution curves of(A,B)pristine OMC:(a)OMC-17,(b)OMC-12,and(c)OMC-7,and(C,D)supported gold catalysts:(a) Au/OMC-17,(b) Au/OMC-12,(c) Au/OMC-7,and (d) Au/OMC-17-R.

Fig.3.(a),(c),(d) TEM and (b) HAADF-STEM images of the (a),(b),(c) fresh Au/OMC-17 and (d) Au/OMC-17-R catalysts viewed along the (a),(b) [211] and (c),(d) [110]directions.The insets in (c) and (d) are the metal particle size distribution histograms.

3.2.Catalytic activities of the supported gold catalysts for the reduction of nitroarenes involving bulky molecules

The reduction of nitroarenes with bulky molecules was selected as a model reaction to assess the catalytic activity of the Au/OMC-17 catalysts.The reactions were performed using sodium borohydride (NaBH4) as a reducing agent under ambient conditions(30 °C) in water.Sodium borohydride works as a reductant in the reduction of nitrobenzenes.The reduction of 4-nitrophenol (4-NP) with small steric hindrance was studied first.The reduction of 4-NP over Au/OMC-17 was monitored by UV–Vis absorption spectroscopy.A decrease of the absorbance at 400 nm,along with an increase of the peak of at 300 nm was observed with increasing time.At the same time,two isosbestic points were found at 278 and 315 nm(Fig.S11).These findings imply that the reduction proceeded without producing byproducts.4-NP was almost completely reduced by Au/OMC-17 within 18 min.Control experiments showed no transformation of 4-NP in the absence of the catalyst.When Au-free OMC-17 was added to the reaction batch,only a decrease of the absorbance at 400 nm occurred but no new absorbance bands were formed (Fig.S12).These phenomena confirm that Au nanoparticles are essential for the reduction reactions.The B-H bonds incleavage on the Au surface and generate the active M-H species on the metal surface.This active M-H species then reacts with nitrobenzenes to produce anilines[45].A linear relationship of ln(ct/c0) versus time (ctand c0are the 4-NP concentrations at times t and 0,respectively) was observed for the Au/OMC-17 catalyst (Fig.6A),indicating a pseudo-first-order reaction [45].The kappvalue calculated from the slope of the linear relationship between ln(ct/c0) and the reaction time is 0.35 min-1,corresponding to a TOF value of 30.5 min-1.These results are similar to results on previously reported supported gold catalysts under similar reaction conditions (Table S3).

Fig.4.XPS spectra in the Au 4f level of(a)fresh Au/OMC-17 and(b)Au/OMC-17-R.

In contrast,the Au/SC-4 catalyst with small pore sizes is involved in a two-step process in the reduction of 4-NP.The initial slow step is possibly due to the inaccessibility of 4-NP to the gold nanoparticles encapsulated inside the long channels with small pore size (Fig.6A).The reduction occurs until the substrate molecules reach the gold surface (Table 2).These findings reveal the obvious diffusion limitation for the Au/SC-4 catalyst in the reduction of 4-NP.

A hot filtration test was used to assess the presence of any leached active metal during the reaction.When Au/OMC-17 was filtered,after approximately 80% conversion,the residue solution did not show any change in concentration (Fig.6B).This result indicates negligible Au leaching during the reduction.We also performed a solid-trapping experiment using mercaptofunctionalized ordered mesoporous silica (SH-SBA-15) to distinguish homogeneous or heterogeneous contributions to the catalysis (Fig.6C).SH-SBA-15 has been postulated to be an effective and selective poison of metallic homogeneous species or nanoparticles and can therefore be used in the reaction to trap any gold species leached from the catalysts (Fig.S13).A control experiment was then performed by adding SH-SBA-15 in the initial stage of the reaction with an SH:Au ratio of approximately 35.Insignificant changes in the conversion plot in the presence SHSBA-15 were detected in the reduction of 4-NP catalyzed by Au/OMC-17,excluding the possibility of leached Au species and providing evidence of the heterogeneity of the Au catalyst [46].Similar to Au/OMC-17,Au/SC-4 shows an insignificant reduction of the conversion of 4-NP in the presence of SH-SBA-15.In contrast,the reference commercial gold catalyst(Au/C)undergoes an almost quenching of activity under the same reaction conditions,implying serious leaching of Au species (Fig.6C and Table 2).

The reduction of tert-butyl substituted nitroarene(4-tert-butyl-2-nitrophenol,4-TB-2-NP)with large steric hindrance catalyzed by Au/OMC-17 was also tested (Fig.6D and Table S2).A complete reduction of 4-TB-2-NP over Au/OMC-17 was achieved within 20 min.A linear relationship between ln(ct/c0) and the reaction time was obtained.A comparable apparent rate constant kappand the corresponding TOF value were calculated to be 0.44 min-1and 27.2 min-1,respectively.However,the reduction of 2,6-DMNB over Au/SC-4 proceeds extremely slowly and less than 20% of the 2,6-DMNB was converted in almost 2 h.These results clearly demonstrate that the mass transfer effect is absent in Au/OMC-17 even for bulky substrates,and there are more serious diffusion limitations for catalysts with small a pore size in the reduction of large substrates molecules.Since these two Au nanocatalysts have similar surface areas,their catalytic activities appear to be controlled by their pore structure(porosity and tortuosity) which could induce different diffusion limitations in the reduction reactions [47,48].The three-dimensional large pore mesoporous network of the OMC facilitates the diffusion of the reactants and products,therefore increasing the contact probability between the substrates and the Au active sites.On contrary,the accessibility of the substrates to the gold nanoparticles embedded inside pores is seriously hindered by the smaller onedimensional pores of Au/SC-4.The above results clearly address the advantages of the large pore size of Au/OMC-17 in the reduction of nitroarenes involving bulky molecules.

Table 2 Catalytic performance of the supported gold catalysts for the reduction of nitroarenes

Fig.5.Schematic of the synthesis of 3 nm gold nanoparticles encapsulated in ordered mesoporous carbon with large pores.

Fig.6.(A)Plot of ln(ct/c0)as a function of time in the reduction of 4-NP over(a)Au/OMC-17,(b)Au/SC-4,and(c)Au/C.ct and c0 are the concentrations of 4-NP at the specified time and initially,respectively.(B)Plot of ln(ct/c0)as a function of time in the reduction of 4-NP over Au/OMC-17 in the hot infiltration experiment.The Au/OMC-17 catalyst was removed by hot filtration after 2 min reaction.(C) Plot of ln(ct/c0) as a function of time in the reduction of 4-NP over (a) Au/OMC-17,(b) Au/SC-4,and (c) Au/C in the presence of SH-SBA-15.(D) Plot of ln(ct/c0) as a function of time in the reduction of (a) 4-tert-butyl-2-nitrophenol (4-TB-2-NP) over Au/OMC-17 and (b) 2,6-DMNB over Au/SC-4.

3.3.Stability and reusability of Au/OMC-17

The stability and reusability of the Au nanocatalysts were also investigated.The linear relation of ln(ct/c0)versus time with a similar apparent rate constant for the selective reduction of 4-NP is observed for the re-used catalyst in successive runs (Fig.7).The reaction mother liquors after separation from the solid catalyst after each run were collected and mixed.The Au content in the final mixture was below the detection limit of ICP-AES analysis.These results further prove there was negligible metal leaching during the repeated reactions.The means that the contribution of leached Au species to the performance of the catalyst can be excluded and that the Au/OMC-17 catalyst is stable and reusable.SAXS patterns (Fig.1B),N2sorption isotherms (Fig.2C,2D),and XRD patterns (Fig.S3) of the catalyst after each of the 8 runs are similar to that of the fresh catalyst.A TEM image of the used cata-lyst shows ordered large mesopore arrays with highly dispersed nanoparticles.No obvious aggregation of the nanoparticles was detected (Fig.3D).The XPS spectrum of the used catalyst shows an insignificant difference to that of the fresh catalyst,indicating that their gold contents were very close (Fig.4).These findings show that the nanoparticle size and oxidation state,and the ordered 3D cubic mesostructured with uniform large pore mesopores of the catalyst are retained after repeated reactions.

Fig.7.Apparent rate constant and catalytic conversion of 4-NP catalyzed by the recovered Au/OMC-17 in successive runs.Reaction conditions:20 mg of catalyst,0.0193 mmol of 4-NP,17.6 mmol of NaBH4,100 ml of water,and 30 °C.

4.Conclusions

We have described a novel metal–ligand coordination selfassembly approach to encapsulate 3 nm gold nanoparticles in ordered mesoporous carbon with large pores (17 nm) and thick pore walls (16 nm).The strong bonding between the gold ions and the thiol groups is essential for the stabilization of the gold species during self-assembly.The large pores and thick pore walls of the ordered mesoporous carbon separate the monodispersed Au nanoparticles and increase the thermal stability of the gold nanocatalyst.The aggregation of the nanoparticles is inhibited at a calcination temperature of 600 °C.The assembly of the PS-b-PEO diblock copolymer determines the ordered mesostructure as well as the large mesopores,and is not affected by the presence of gold species.The resulting Au nanocatalysts with threedimensional large mesopores(17 nm)show higher catalytic activities and stability in the reduction of bulky nitroarenes than a similar gold catalyst but with a small pore size(4 nm).The large open pores facilitate the access of bulky molecules to the active sites of the Au catalyst.This class of materials provides new opportunities to overcome the limitations of diffusion-constrained reactions for processing bulky molecules.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (22025204,92034301,21773156,and 51932005),the Shanghai Sci.&Tech.and Edu.Committee(19070502700),and the Innovation Program of the Shanghai Municipal Education Commission (2021-01-07-00-02-E00119).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.10.004.