Bimetallic Au–Pd nanoparticles supported on silica with a tunable core@shell structure: enhanced catalytic activity of Pd(core)–Au(shell) over Au(core)–Pd(shell)

A facile ligand-assisted approach of synthesizing bimetallic Au–Pd nanoparticles supported on silica with a tunable core@shell structure is presented. Maneuvering the addition sequence of metal salts, both Aucore–Pdshell (Au@Pd–SiO2) and Pdcore–Aushell (Pd@Au–SiO2) nanoparticles were synthesized. The structures and compositions of the core–shell materials were confirmed by probe-corrected HRTEM, TEM-EDX mapping, EDS line scanning, XPS, PXRD, BET, FE-SEM-EDX and ICP analysis. The synergistic potentials of the core–shell materials were evaluated for two important reactions viz. hydrogenation of nitroarenes to anilines and hydration of nitriles to amides. In fact, in both the reactions, the Au–Pd materials exhibited superior performance over monometallic Au or Pd counterparts. Notably, among the two bimetallic materials, the one with Pdcore–Aushell structure displayed superior activity over the Aucore–Pdshell structure which could be attributed to the higher stability and uniform Au–Pd bimetallic interfaces in the former compared to the latter. Apart from enhanced synergism, high chemoselectivity in hydrogenation, wide functional group tolerance, high recyclability, etc. are other advantages of our system. A kinetic study has also been performed for the nitrile hydration reaction which demonstrates first order kinetics. Evaluation of rate constants along with a brief investigation on the Hammett parameters has also been presented.


Introduction
The concept of mutualism centers around the idea of integrating two metals in perfect correlation with each other, thereby exhibiting distinct superiority in chemical and physical properties over their monometallic counterparts. 1 In the past few years, bimetallic nanostructured materials have attracted enormous attention because of their multifunctional applications in diverse areas such as catalysis, 2,3 materials sciences (in optoelectronic and magnetic devices), 4 medical biotechnology, 5 etc. Among various bimetallic nanoparticles, the Au-Pd system has been one of the most studied systems because of its applicability in a number of catalytic reactions such as alcohol oxidation, 6 CO oxidation, 7 dehydration of formic acid, 8 dye degradation, 9 hydrodechlorination, 10 chemoselective hydrogenation 11 and so forth. However, the activity and selectivity of such bimetallic catalysts not only depend on metal systems alone but also on the structural motif or morphology of such materials. Thus, virtual control over material structures is one of the most desirable, yet challenging, areas in nanoparticle research. Among various nanostructured materials (e.g. alloy, core-shell, heterostructure, etc.), special attention has been paid to core-shell structures for their unprecedented success in the eld of catalysis. 12,13 In fact, there are multiple instances where coreshell Au-Pd nanoparticles exhibited improved catalytic efficacies compared to their alloy counterparts. 14,15 It is worth noting that the catalytic properties of a core-shell material vary not only with the size and morphology but also with the ordering of core and shell metals. Although, there have been multiple reports on the synthesis of core-shell Au-Pd NPs, [16][17][18][19] the ability to control the order of core and shell metals is still contemplated as one of the most difficult tasks in materials research. Incidentally, the majority of the core-shell Au-Pd systems reported till date are typically galvanic in nature, where the Au atom being more noble prefers to occupy the core and the Pd atom being less noble forms the shell. On the contrary, anti-galvanic core-shell nanoparticles with Pd core -Au shell structures are relatively few and limited in scope. 20,21 In fact, a number of theoretical studies involving molecular dynamics simulation 22,23 and density functional theory (DFT) 24 have envisaged that in a core-shell structure, the Pd core -Au shell arrangement is deemed more stable than the Au core -Pd shell arrangement. This theoretical outcome is further reinforced by a few experimental studies that demonstrated the synthesis of stable Au-Pd core-shell nanoparticles with Pd core -Au shell arrangement and their application as catalysts for different organic transformations like hydrogenation, 20 selective oxidation of alcohol, 21 photocatalytic oxidation, 25 etc.
Inquisitively, a majority of the Au-Pd core-shell catalysts reported to date are colloidal in nature, while from the catalytic perspective, supported NPs are more demanding as they facilitate catalyst separation, recovery and subsequent recycling. However, unlike colloidal nanoparticles, precise architectural control in a supported system is complicated as such systems predominantly produce Au core -Pd shell structures via a galvanic replacement pathway, [26][27][28] though from the catalytic perspective, nanoparticles with Pd core -Au shell structures could be more intriguing. Thus, it is of great interest to rationally design supported Au-Pd core-shell nanoparticles with tunable core and shell atoms and to explore their catalytic potentials. Several experimental methods are currently employed for the synthesis of such core-shell systems viz. the radiolytic method, 29 photocatalytic approach, 30 vapor deposition method, 20 sequential reduction approach, 31 surfactant-mediated method, 21 dendrimer-assisted method, 31 ligand-based stabilization 32 method, etc. Among these, the seed-mediated or sequential reduction approach that involved reduction of a shell metal onto a preformed metallic core was found to be the most attractive as it usually provided better control over the tunable core-shell property, 33 thereby controlling the activity and selectivity. It has been noted that in the seed-mediated process, the nature of the surface capping ligand oen plays the most crucial role in controlling the particle morphology. Thus, a judicious choice of the capping ligand also plays a signicant role 34 in this respect. Recently, we have reported the synthesis and catalytic activities of monometallic Pd-nanoparticles supported on silica stabilized by a phosphine ligand. 35 In this work, we have extended the scope of this system to a bimetallic platform by using a ligand stabilized less noble metal as the core which is coated with a more noble metal counterpart acting as the shell. The order is reversed through coordination of the same ligand with the more noble metal followed by coating with the less noble metal. Thus, by controlling the addition sequence of the metal salts, core-shell Au-Pd nanoparticles with Au core -Pd shell and Pd core -Au shell arrangements can be easily accessed. The synergistic potentials of the core-shell materials were explored for two important reactions viz. hydrogenation of nitroarenes to amines and hydration of nitriles to amides.

Experimental
Materials 2-Diphenylphosphinoethyl-functionalized silica gel (0.7 mmol g À1 loading; 200-400 US mesh), HAuCl 4 and PdCl 2 were purchased from Sigma-Aldrich. All solvents, substrates and other chemicals were purchased from Acros Organics (purity > 98%), TCI chemicals and Merck (purity > 99%). Characterization X-ray diffraction patterns were acquired on a Rigaku Ultima IV Powder X-ray diffractometer using Cu-K N as the source (l ¼ 1.5406Å) with a scan rate of 2q ¼ 4 min À1 over a scanning angle of 5-80 . X-ray photoelectron spectroscopy (XPS) analysis was carried out using an FEI PHI 5000 Versa Prob II system. Surface area measurements were carried out at liquid nitrogen temperature on a Quantachrome Autosorb-iQ 2010 analyzer (ASIQA) using N 2 gas as the adsorbent. Surface area calculations are done from the isotherm using the Brunauer-Emmett-Teller (BET) model and pore size distributions were calculated using the Barrett-Joyner-Halenda (BJH) model. Samples were degassed at 100 C for 2 h before performing the experiments. A JEOL-JEM2010+ electron microscope (HRTEM) with an accelerating voltage of 200 keV was used to obtain high resolution transmission electron microscopy (HRTEM) images of the materials. Line scanning spectra were obtained using a special specimen stage atomic resolution scanning transmission electron microscope, JEOL-ARM200F equipped with an aberration corrector on the incident probe, which provides a nominal resolution of 0.1 nm at an acceptance semi angle of 26.5 mrad. A JEOL-JEM2010+ electron microscope (HRTEM) with an accelerating voltage of 200 keV was used to obtain high resolution transmission electron microscopy (HRTEM) images of the materials. A ZEISS SIGMA eld emission scanning electron microscope (FESEM) with a lanthanum hexaboride (LaB 6 ) lament and tungsten eld emission source (FES) was used to obtain high resolution eld emission microscopy (FESEM) images of the materials. The machine operates at an acceleration voltage on the order of magnitude of 0.5 to 30 kV and extreme vacuum of $10-6 Pa in the column of the microscope. The metal contents in the nanomaterials were measured using an ACROS, simultaneous inductively coupled plasma (ICP) spectrometer equipped with an R.F. generator (1.6 kW, 27.12 MHz), at wavelengths in the range of 130-770 nm. Catalytic hydrogenation reactions were carried out under appropriate reaction conditions in a sealed Teon coated glass tube mounted on an aluminum block synthesizer slot with an LED sensor (model: PPS-5511, dimensions: 465W, 335L, 295H, Make: Eyela, USA). Catalytic nitrile hydration reactions were performed accordingly under specic reaction conditions. Gas chromatographic analysis of the reactants and products was performed on a 7820A Agilent GC system equipped with a capillary column (L 30 m, I.D. 0.25 mm), tted with an FID and with n-dodecane as the internal standard. The mass spectroscopic analysis was performed using an MS-5965 Agilent analyzer.
Synthesis of monometallic Au nanoparticles (SiO 2 @Au) 2.0 g of 2-diphenylphosphinoethyl functionalized silica gel was weighed and taken in a 100 ml round bottom ask containing 20 ml distilled water and stirred for 30 minutes and sonicated at 50 C for 15 min to achieve complete dispersion. To the resultant mixture, 20 ml of 0.01 mmol HAuCl 4 solution was added and stirred for an hour. The color of the silica gel changed to canary yellow. To the resultant mixture, 4 ml (51.2% v/v) of hydrazine solution was added dropwise over a period of 15 minutes and the color instantly changed to bright green. The stirring was continued for 1 h to ensure complete reduction of HAuCl 4 . Subsequently, following the previous experiment, the resultant solid mass was washed multiple times, dried at room temperature and kept in a desiccator. The vacuum dried material was named SiO 2 @Au.
Synthesis of monometallic Pd nanoparticles (SiO 2 @Pd) 2.0 g of 2-diphenylphosphinoethyl functionalized silica gel was weighed and taken in a 100 ml round bottom ask containing 20 ml distilled water and stirred for 30 minutes and sonicated at 50 C for 15 min to achieve complete dispersion. To the resultant mixture, 20 ml of 0.01 mmol Na 2 PdCl 4 solution was added and stirred for an hour. The color of the silica gel changed to bright yellow. To the resultant mixture, 4 ml (51.2% v/v) of hydrazine solution was added dropwise over a period of 15 minutes and the color instantly changed to light brown. The stirring was continued for 1 h to ensure complete reduction of Na 2 PdCl 4 . Subsequently, this resultant solid mass was centrifuged and the residual ltrate was decanted. This process was performed thrice and the recovered solid mass was washed several times with small aliquots of distilled water, dried at room temperature and kept in a vacuum desiccator. The vacuum dried material was named SiO 2 @Pd.
Synthesis of bimetallic Au-Pd nanoparticles with Au core -Pd shell (Au@Pd-SiO 2 ) 2.0 g of SiO 2 @Au was taken in a 100 ml round bottom ask containing 20 ml distilled methanol and stirred for 30 minutes to achieve complete dispersion. To the resultant suspension, 20 ml, 0.01 mmol Na 2 PdCl 4 solution was added simultaneously and stirred for an hour. The color of the silica gel changed to pale yellow. To it 4 ml (51.2% v/v) hydrazine solution was added dropwise over a period of 15 minutes (rate ¼ 0.26 ml min À1 ) and the color instantly changed to grey. The stirring was continued for another 1 h to ensure complete reduction of NaPdCl 4 . The resultant solid mass was ltered and washed several times with small aliquots of distilled water, dried at room temperature and stored in vacuum desiccators. The vacuum dried material was named Au@Pd-SiO 2 .
Synthesis of bimetallic Au-Pd nanoparticles with Pd core -Au shell (Pd@Au-SiO 2 ) 2.0 g of SiO 2 @Pd was taken in a 100 ml round bottom ask containing 20 ml distilled methanol and stirred for 30 minutes to achieve complete dispersion. To the resultant suspension, 20 ml, 0.01 mmol HAuCl 4 solution was added simultaneously and stirred for an hour. The color of the silica gel changed to bright yellow. To it 4 ml (51.2% v/v) hydrazine solution was added dropwise over a period of 15 minutes (rate ¼ 0.26 ml min À1 ) and the color instantly changed to light brown. The stirring was continued for another hour to ensure complete reduction of HAuCl 4 . The resultant solid mass was ltered and washed several times with small aliquots of distilled water, dried at room temperature and stored in vacuum desiccators. The vacuum dried material was later named Pd@Au-SiO 2 .

General procedure for hydrogenation
Hydrogenation of nitroarenes to corresponding amines was carried out by rst charging 2 mmol of the substrate in 4 ml of the solvent taken in a Teon coated glass tube mounted on a synthesizer slot. Initially 17.5 mg (0.07 mol%, 0.0014 mmol) of catalyst was weighed and subsequently added to the above reaction mixture. Thereaer, the synthesizer was heated to 80 C with constant stirring (500 rpm) for the desired period. Once the reaction reached completion, the synthesizer was cooled to room temperature. The reaction mixture was separated from the catalyst by syringe ltration and analyzed using a GC-MS. The separated catalyst was washed repeatedly with i PrOH and acetone, dried, and used further for recyclability tests.

General procedure for nitrile hydration
The catalytic hydration of nitriles to corresponding amides was carried out by similarly charging 0.5 mmol of the desired substrate in 4 ml of the solvent taken in a Teon coated glass tube mounted on a synthesizer slot. Instantly 12.5 mg (0.2 mol%, 0.001 mmol) of catalyst was added to the reaction mixture and the synthesizer was heated to 50 C with constant stirring (400 rpm) for the desired period. Aer completion of the reaction, the synthesizer was cooled to room temperature. The reaction mixture was carefully separated from the catalyst, ltered and analyzed using a GC-MS. The separated catalyst was recovered, washed repeatedly, dried, and used further for recyclability tests.

General procedure for heterogeneity tests
The reaction was performed with the desired substrate, solvent and the catalyst for a period of 15 minutes. The catalyst was carefully separated from the reaction mixture and the ltrate was later analyzed using a GC-MS. Subsequently, the solid mass was separated from the solution by ltration under hot conditions and the resultant ltrate was subjected to stirring for another hour under the same reaction conditions as eventually no increase in conversion was observed.

General procedure for recyclability tests
For recycling experiments, aer each run, the catalyst was washed several times with suitable solvents, dried to remove any occluded reactants and products physically attached to the surface of the catalytic material and reused for successive runs. Subsequently, the recovered catalyst was oven-dried at 95-100 C for an hour and transferred to a reaction vessel containing the desired substrate and solvent, and reusability experiments were performed under identical conditions. Likewise, all the recyclability tests were performed maintaining the same stoichiometric ratios.

Results and discussion
Synthesis and characterization of silica-supported bimetallic Au-Pd nanoparticles A schematic representation of the synthesis of silica-supported core-shell Au-Pd NPs with Au core -Pd shell (Au@Pd-SiO 2 ) and Pd core -Au shell (Pd@Au-SiO 2 ) has been shown in Scheme 1. A ligand-assisted sequential reduction approach has been followed. In the rst step, silica-supported monometallic Pd NPs (SiO 2 @Pd) and Au NPs (SiO 2 @Au) were synthesized by reacting commercially available phosphine-functionalized silica gel with corresponding metal salts in the presence of hydrazine hydrate as the reducing agent following a previously reported protocol. 35 The morphological properties and other characterization details of the monometallic materials have been presented in the ESI section ( Fig. 4, 5, S1 and S2 †). The HRTEM micrographs of SiO 2 @Au and SiO 2 @Pd ( Fig. S1b and d †) demonstrated the formation of monodisperse Au and Pd NPs of uniform size and near-spherical shape homogeneously distributed throughout the silica matrix. The size-distribution histograms corresponding to the TEM micrographs indicated a diameter of 4.5 nm for Au NPs and 5.12 nm for Pd NPs (Fig. S1a and c † inset) respectively. In the second step, these monometallic Pd or Au NPs were employed as seeds and subsequent coatings with Au and Pd shells were made by treating methanolic suspensions of the preformed monometallic cores with an aqueous solution of the corresponding metal salts followed by reduction with hydrazine hydrate. It may be mentioned that this type of surface-ligand modulated seed-growth method has been recently reported by Kluenker et al. as an excellent method of synthesizing bimetallic Au-Pd nanoparticles with tuned morphology ranging from alloy to core-shell structures. 34 Similarly, Esparza and co-workers synthesized core-shell nanoparticles using a similar technique and studied the development of Pd shells over Au cores using the Volmer-Weber model. 36 The representative TEM micrographs of the bimetallic materials Au@Pd-SiO 2 (Fig. 1a) and Pd@Au-SiO 2 (Fig. 1b) indicated the formation of spherical nanoparticles. The size distribution histograms enumerate that the majority of the particles exhibited an average diameter of 5.42 nm for Au@Pd-SiO 2 (Fig. 1a, inset) and 3.78 nm (Fig. 1b, inset) for Pd@Au-SiO 2 respectively. The HRTEM micrographs of Au@Pd-SiO 2 and Pd@Au-SiO 2 exhibited nearly spherical morphology with slight inhomogeneity in size and differences in contrasts ( Fig. 1c and d respectively) 37-39 extending throughout the lattices.
The high-magnication TEM images for single particles of Au@Pd-SiO 2 (Fig. 1e) and Pd@Au-SiO 2 (Fig. 1f) showed the presence of distorted lattices 37 indicating the presence of an interface between two different chemical compositions in the materials.
Usually, for a bimetallic system, the nature of crystal structure and d-spacing values tell whether a system adopts a coreshell or alloy structure. 40,41 The fast Fourier transform (FFT) patterns presented in Fig. 1g and h reveal a face-centred cubic (FCC) arrangement for Au@Pd-SiO 2 and Pd@Au-SiO 2 materials indicating core-shell structures for both the materials. 0784). 42 In fact, the presence of peaks corresponding to the (111), (200) and (220) planes of Au and Pd was also observed from PXRD analysis of the Au-Pd materials (Fig. 2g). However, the peaks due to (200) and (220) planes observed at 2q ¼ 47.11 , 65.65 and 46.87 , 65.26 for Pd@Au-SiO 2 and Au@Pd-SiO 2 respectively were less intense compared to the (111) planes observed at 2q ¼ 39.05 and 38.11 . It may be noted that the peaks assigned to (111), (200) and (220) planes for Au and Pd crystallize in the fcc structure. Consequently, their respective peaks corresponding to these diffraction planes overlap. 43 Apart from the crystal structure and d-spacing, lattice-mismatch is also an important parameter to identify the heteroepitaxial growth in a core-shell material. A large lattice-mismatch observed in a HRTEM usually leads to a non-coherent interface or a polycrystalline shell layer. 44 In this respect, the FFT pattern corresponding to a HRTEM micrograph provides a better idea as multiple secondary reections in the pattern usually act as clear evidence of potential lattice mismatches in the crystal structure.
In fact, the high-magnication HRTEM micrograph of Au@Pd-SiO 2 (Fig. 1e) corroborated this and exhibited large lattice mismatches between the (111), (200) and (220) planes of Au and Pd indicative of a polycrystalline morphology. This is conrmed from the FFT pattern which depicts multiple secondary reections for both Au and Pd. On the other hand, the high magnication HRTEM images of Pd@Au-SiO 2 displayed a well-segregated core-shell interface with a mononuclear Pd core within an Au shell. As evidenced by the FFT pattern, a uniform heteroepitaxial growth with minimal lattice mismatch was observed between the (111), (200) and (220) planes corresponding to Au and Pd. This is well conrmed by the pristine (no secondary reections) FFT pattern ( Fig. 1f and h) indicative of the typical monocrystalline nature of Pd@Au-SiO 2 (Fig. 1g). 44 Other than  Nanoscale Advances a minor epitaxial reection of the (111) crystal plane upon the (101) plane (Fig. 2e) the crystal planes were free of any stack ( Fig. 2d and f), fault or twinning at the bimetallic interfaces (Fig. 2). TEM-EDX mapping images of Au@Pd-SiO 2 ( Fig. 3a and b) and Pd@Au-SiO 2 ( Fig. 3d and e) showed that Pd and Au atoms are homogeneously distributed in both the materials. As seen from the TEM-EDX mapping image in Fig. 3b for the material Au@Pd-SiO 2 , the intensity of green dots representing Au atoms was higher at the centre while the intensity of white dots representing Pd atoms is more at the surface. A magnied cross-sectional view of an individual particle in the inset (Fig. 3b) presented a clear idea of the differences in intensities of the metal atoms and highlights large patches of Au atoms at the core accompanied by small patches of Pd atoms at the shell. The fact was further corroborated by a TEM-EDX line scanning prole diagram obtained along the edges (Fig. 4a) as well as across the core (Fig. S7a †) which shows a galvanic type structure with an Au rich centre and a Pd rich boundary (Fig. 4a). On the other hand, for the material Pd@Au-SiO 2 , the TEM-EDX mapping images (Fig. 3e) showed that the white dots representing Pd atoms are more intense at the centre than the green dots for Au atoms. The magnied cross-sectional image of a single particle (inset, 3e) demonstrated a dense and pristine spherical core composed of Pd atoms surrounded by a thin layer of Au atoms. As anticipated, the TEM-EDX line prole diagrams ( Fig. 4b and S7b †) also conrmed a Pd rich centre engulfed by an Au rich boundary, a typical anti-galvanic type structure. Similar to TEM-EDX mapping, FESEM-EDX images (Fig. S6 †) of both Pd@Au-SiO 2 and Au@Pd-SiO 2 showed the presence of Au and Pd atoms on the silica support.  Thus, by controlling the addition sequence of metal salts both galvanic (Au core -Pd shell ) and anti-galvanic type (Pd core -Au shell ) Au-Pd NPs could be accessed. N 2 adsorption-desorption isotherms were measured for the materials at 77 K and the physical parameters were compared with those for the parent support. The N 2 -sorption measurements for the materials with respect to the neat support showed an isotherm with the type IV hysteresis loop characteristic of mesoporous materials. A prominent decrease in the surface area compared to the neat support was observed for Au@Pd- Fig. 5 (a) N 2 -sorption isotherms of the nanoparticle incorporated silica-based materials: SiO 2 @Au, SiO 2 @Pd, Au@Pd-SiO 2 and Pd@Au-SiO 2 , (b) XPS spectra of (i) SiO 2 @Au, (ii) Pd@Au-SiO 2 and (iii) Au@Pd-SiO 2 showing the Au 4f 5/2 and 4f 7/2 states and (c) XPS spectra of (i) SiO 2 @Pd, (ii) Au@Pd-SiO 2 and (iii) Pd@Au-SiO 2 3d 5/2 and 3d 3/2 showing Pd 3d 5/2 and 3d 3/2 states. SiO 2 (Fig. 5a) whereas only a marginal decrease was recorded for the SiO 2 @Au, SiO 2 @Pd and Pd@Au-SiO 2 materials (Fig. 5a). Nevertheless, the materials exhibited a gradual decrease in surface area as well as pore volume on moving from the neat support through monometallic to bimetallic nanoparticles (Table 1). This steady decrease can be attributed to the generation of mono/bimetallic nanoparticles within the pores of the silica matrices.
The Au@Pd-SiO 2 and Pd@Au-SiO 2 nanoparticles were further subjected to XPS analysis (Fig. 5b and c) to conrm the metallic nature of Au and Pd. The XPS peak values are also compared to those of the monometallic counterparts i.e., SiO 2 @Au [ Fig. 5b(i)] and SiO 2 @Pd [ Fig. 5c(i) where Au was present as the core and Pd as the shell, peaks due to Au 4f 5/2 and 4f 7/2 were marginally shied to higher energies (ca. +0.2 eV), while in Pd@Au-SiO 2 , where Au was present as the shell, relatively higher shis were noticed (+0.3 and +0.5 eV for 4f 5/2 and 4f 7/2 respectively), consistent with a change in the electronic environment presumably due to easy charge transfer between the well-organized Au-Pd interfaces induced by the phosphine donors in the ligand. A similar type of effect resulting in increment of binding energies due to the presence of phosphine donor ligands has been reported in the literature. 35,46 Similar trends were also noticed for Pd, where shis of +0.3 and À0.3 eV were evident for 3d 3/2 and 3d 5/2 electronic states respectively. It may be worth noting that XPS analysis can also be used to determine the thickness of shells in core-shell materials up to a depth of 10 nm. 47 Accordingly, we have calculated the shell thickness of Au@Pd-SiO 2 and Pd@Au-SiO 2 and values of 2.54 and 1.46 nm were obtained respectively. However, compared to the shell thickness determined by TEM analysis (Fig. 1a and b) the values for the respective materials were 4.6 and 2.8% lower. The relatively high mismatch obtained in the case of Au@Pd-SiO 2 compared to Pd@Au-SiO 2 may be attributed to the inhomogeneous shell thickness and dissymmetric core-shell interfaces. Such an inhomogeneous shell with a core displaced from the particle center can oen lead to an underestimation of the shell thickness.
The amount of Au and Pd present on the synthesized materials was estimated by ICP-AES analysis and was found to be 0.91 and 0.78 for Au@Pd-SiO 2 and 0.89 and 0.82 wt% for Pd@Au-SiO 2 respectively (Table 1).
In one of our previous reports, we have laid special emphasis on the role of ligands with different donors viz. N-, P-and Satoms for the synthesis of Pd nanomaterials with different sizes, shapes and dimensions. Among the various ligands, the phosphine donor ligand demonstrated Pd NPs with exceptional stability, purity and dispersity. The Pd NPs exhibited sizes averaged at 6 nm and a d-spacing value of 0.223 nm with reticular planes extending throughout the structure without any stacking or fault. Additionally, XPS studies revealed that the interaction of P-donor atoms on the ligand-functionalized silica matrices induces a small positive charge on the Pd metal surface. A cumulative inuence of these factors can help utilize these Pd NPs as suitable cores for the deposition of a second metal. In this respect, a thick layer of Au was homogeneously deposited onto the preformed Pd cores leading to Pd core -Au shell nanoparticles.

Hydrogenation of nitroarenes
Aromatic chloroamines are widely used as important intermediates in the chemical industry for the production of pharmaceuticals, agrochemicals, polymers and dyes. 48 Currently, access to those amines occurs mainly through reduction of appropriate nitro compounds using molecular H 2 as the reductant. Although molecular H 2 is an environmentally benign reducing agent, in some cases, particularly under high pressure, the use of gaseous H 2 could make the system potentially explosive and thus limit its industrial scale utility. 49 Moreover, gaseous hydrogen at higher pressures can also lead to hydrodechlorination reaction through C-Cl bond scission to produce undesired anilines in high yields thereby affecting the selectivity. 35 Another important route to access chloroamines is transfer hydrogenation of nitroarenes which utilizes hydrogen supplied from a co-solvent, usually an alcohol such as 2-propanol. However, the production of hydrogen from 2-propanol usually requires higher temperatures and longer reaction times as observed in recent reports. 50 On the contrary, addition of hydrogen from sources other than molecular hydrogen and alcohols (e.g. ammonia borane, 51 formic acid, 52 hydrazine hydrate, 53,54 etc.) has emerged as much safer alternatives in recent times. Among those hydrogen sources, hydrazine hydrate is particularly promising as it produces N 2 as the sole byproduct thereby ensuring higher safety.
Literature evidence suggests that bimetallic systems involving the use of a second metal with Pd or Au oen improve their catalytic efficacies for various organic transformations. 55,56 However, such improvements are usually at the crucial cost of selectivity. Today, a number of catalytic systems, including metal nanoparticles of Pd or Au (e.g. PdNP@CeO 2 , 57 AuNP@TiO 2 , 58 etc.) are known to promote this hydrogenation. However, such systems usually suffer from unsatisfactory chemoselectivity, particularly in cases of functionalized nitro substrates that contain groups like carbonyls, halogens, etc. owing to the side reactions of over-hydrogenation/dehalogenation. 53,59 Thus, the search for more promising heterogeneous catalysts that can produce targeted amines through safer hydrogen sources without compromising with the chemoselectivity is highly desirable. Hence, we have tested the hydrogenation prowess of our bimetallic Au-Pd materials, taking 4-chloronitrobenzene (4-CNB) as the model substrate (2 mmol) and ethanol (5 ml) as the solvent with hydrazine hydrate (1 equivalent) at 60 C for 1 h (Scheme 2).
To illustrate the synergistic effects, we have also tested the catalytic efficacies of the corresponding silica-supported monometallic Au and Pd NPs. As shown in Fig. 6, the monometallic Pd catalyst, Pd@SiO 2 , gave only 22% of 4-chloroaniline (4-CAN) with a selectivity of 100%, while the monometallic Au catalyst, Au@SiO 2 , showed slightly higher activity (conversion: 44%) with a moderate drop in selectivity (80%). Interestingly, use of bimetallic catalysts (Pd@Au-SiO 2 or Au@Pd-SiO 2 ) signicantly improved the activity, demonstrating the synergistic inuence between the two metals. Under the same set of experimental conditions, the Pd@Au-SiO 2 catalyst having Pd core -Au shell structure displayed superior activity and selectivity over the Au@Pd-SiO 2 catalyst having the Au core -Pd shell structure. This higher activity of the Pd@Au-SiO 2 catalyst could be attributed to the uniform and well-dened Au-Pd bimetallic interfaces compared to the non-uniform interfaces in Au@Pd-SiO 2 . 34 Homogeneous interfaces may lead to pronounced ligand, ensemble and strain effects 60 that govern the mechanism of synergism in a bimetallic material. Higher synergy will promote greater activity while lower synergy produces a diminishing effect on activity.
Considering the superior catalytic performance of Pd@Au-SiO 2 over others, we have applied it for further optimization studies. Indeed, a temperature optimization study with Pd@Au-SiO 2 (Fig. 7a) revealed that both the activity and selectivity are temperature dependent. It has been seen that an increase in temperature resulted in an increase in conversion, although a little drop in selectivity was noticed. The optimum temperature was found to be 80 C when 76% conversion was achieved with a 4-CAN selectivity of 96%.
Interestingly, by doubling the hydrazine hydrate concentration (from 1 mmol to 2 mmol) the conversion could be increased to 88% while maintaining a selectivity of 96%. Indeed, apart from temperature and hydrazine-hydrate concentration, the choice of solvent also had some vital impact on our catalytic system (Fig. 7b). Among the solvents screened, ethanol was found to be the most effective, while THF was the least effective. To determine the best catalytic conditions, additional screenings such as catalyst : substrate ratio, effect of hydrogen source, etc. were performed with our model system (Table S1 †). Important conclusions can be drawn from these studies. Firstly, on increasing the catalyst : substrate ratio, a decrease in both conversion (72%) and selectivity (90%) was observed (Table S1, † entry 8). Secondly, an increase in the reaction time led to almost stoichiometric conversion (98%) while at the same time brought down the selectivity (84%) owing to extended exposure to longer time intervals and subsequent circumcision of the C-Cl bond from the preformed    (Table S1, † entry 7). Finally, the reaction was also screened in the presence of different hydrogen sources under similar experimental conditions to explore the rates of reduction exhibited by these agents (Table S1, † entries 1 -4, 6, 9-12).

4-CAN
Aer systemic optimization of a number of reaction variables, it was observed that a reaction time of 1 h coupled with a temperature of 80 C, 0.28 mol% of Pd@Au-SiO 2 [Au (0.13 mol%), Pd (0.15 mol%)], 2 mmol of model substrate (4-CNB) and 2 equivalents of hydrazine hydrate in ethanol (5 ml) demonstrated best results and were chosen for substrate scope evaluation.
To explore the scope and limitations of the present methodology, different nitroarenes were examined under the optimized reaction conditions and the results are presented in Table 2. Nitroarenes containing electron-donating, electron-neutral as well as electron-withdrawing substituents mostly at the para-position could provide desired anilines in excellent yields and selectivity. Indeed, this high conversion and selectivity were also maintained with halogen-substituted nitroarenes without any noticeable dehydrohalogenated products (2a-e). Thus, our result is quite satisfactory as we could achieve the desired amines with selectivity >95% without compromising with the activity.
In fact, most of the reported catalytic systems showed either relatively low conversion, 61 or selectivity <90%. 62,63 Remarkably, the subtle nature of the catalyst was unveiled with a slightly challenging substrate like 2,5-dichloronitrobenzene, as both the -Cl groups were retained in the product forming 2,5dichloroaminobenzene as the sole product (2d). Interestingly, substituted dinitrobenzene compounds like 1-chloro-2,5- Table 2 Substrate scope of Au-Pd catalyzed hydrogenation reactions a,b,c a Reaction conditions: substrate (2 mmol), ethanol (5 ml), Pd@Au-SiO 2 [Au (0.13 mol%), Pd (0.15 mol%)], hydrazine-hydrate (2 equivalents) at 80 C for 1 h. b Determined by GC-MS analysis and aniline (AN) was formed as the sole by-product. c MS-spectra available in the ESI. dinitrobenzene underwent hydrogenation to furnish a monoaminated product instead of the diamine. The reaction was regioselective for 3-chloro-4-nitroaniline (2e) instead of 2chloro-4-nitroaniline (2f) which might be attributable to the intramolecular hydrogen bonding observed in the case of the 2chloro derivative. Functionalized nitroarenes containing reducible moieties are considered to be difficult substrates to hydrogenate chemoselectively. 64,65 However, we are gratied to see that our bimetallic Au-Pd catalyst can effectively perform hydrogenation of compounds like 4-nitrobenzaldehyde and 4nitroacetophenone and produce 4-aminobenzaldehyde (commonly known as PABA, used as a biodegradable pesticide and as a UV-lter in sunscreens) 66 (2j), and 4-aminoacetophenone (2k) almost quantitatively with selectivity >98%, without affecting the easily reducible carbonyl groups. Similarly, 4-nitrobenzoic acid was converted to 4-aminobenzoic acid (2l) with moderate conversion and excellent selectivity.
In order to ascertain the superiority of our catalyst, a comparative study was performed with reported Au-or Pd NP based catalysts immobilized onto different supports (Table 3). Clearly, Pd-based catalysts are of prominent interest in this regime. Most of these catalysts, just like in our case do not involve the use of bases for activation. A number of these catalysts involve high temperatures (e.g., Pd/AlPO 4 /SiO 2 , Au@Pt/ TiO 2 , Au-Pd/Mo 2 C), prolonged reaction times (e.g., Co@Pd/NC, Pd/ZrP), high metal loadings (e.g., Au-Pd/Mo 2 C), drastic conditions (e.g., Au-Sn/SiO 2 ) and deleterious solvents (e.g., Au-Sn/SiO 2 , Au-Ni/TiO 2 ). Although, the activity and selectivity of catalysts such as G-NiPd, PdNi/mCN and PdCu@MWCNT seem promising, lengthy and multistep synthetic procedures hinder their industrial applicability. Use of moisture sensitive precursors further adds to the complicacy in their handling procedures. Hence, it is apparent from Table 3 that Pd@Au-SiO 2 is much more superior to the reported catalysts.

Reusability and heterogeneity
As reusability is one of the most important features of any supported nanomaterial, we have performed a recyclability test of our catalyst, Pd@Au-SiO 2 , taking p-CNB as the substrate under optimized conditions. Aer performing the reaction, the catalyst was separated by centrifugation, washed with i PrOH-H 2 O (1 : 1), dried and then used for subsequent runs. Our results showed that the catalyst could be used at least 5 times without noticeable changes in activity and selectivity. The slight drop in activity might be due to unavoidable handling loss of the sample during recycling experiments. For complete evaluation of reusability, systematic recycling experiments for all the catalysts under the optimized set of conditions were performed which demonstrated very little change in activity or selectivity ( Fig. 8a and b). The powder X-ray diffraction pattern of the material obtained aer the 5th run showed all three prominent reections corresponding to the (111), (200) and (211) planes of fcc Au and Pd (Fig. 8d). In fact, TEM (Fig. 8e) and TEM-EDX studies (Fig. 8f) of the recovered catalyst (Pd@Au-SiO 2 ) aer the 5th run showed that there is virtually no change in the surface morphology suggesting that the catalyst could be reused for more runs. To check the heterogeneity of our catalyst, the hot ltration test was performed using 4-CNB as a substrate. The reaction was performed with Pd@Au-SiO 2 as the catalyst under a hydrazine-hydrate concentration of 2 equivalents and a temperature of 80 C in ethanol. The reaction time was shortened from 1 h to a period of 15 minutes. Aer completion of the reaction, the catalyst was carefully separated from the reaction mixture and the ltrate was subsequently analyzed using a GC-MS which showed a conversion of 49% with a selectivity of 100%. Interestingly, when the ltrate was stirred for one more hour under similar conditions, no increase in conversion was observed (Fig. 8c) suggesting that there was no leaching of Au or Pd into the solution. In fact, the ICP-AES analysis for the catalyst aer the h cycle exhibited almost negligible loss of Au (0.011 wt%) and Pd (0.019 wt%) into the solution thus demonstrating the heterogeneous nature of the core-shell material.

Nitrile hydration
Transition metal mediated hydration of nitriles to amides is one of the most atom-economical methods for the conversion of aromatic nitriles to corresponding amides. 76,77 A number of noble-metal based catalysts including nanoparticles of Au and Pd are known for this transformation. 78 However, to the best of our knowledge, there is not a single report on the use of bimetallic Au-Pd catalysts for the hydration of nitriles. Thus, with the catalyst at hand, we have tested the activity of our bimetallic core-shell materials for the nitrile hydration reaction. The hydration of benzonitrile was used as a test reaction for initial screening studies and the reactions were conducted taking aqueous i PrOH (1 : 1 mixture) as solvent at 50 C for 1 h (Scheme 3). Like hydrogenation, in this case also we have compared the catalytic efficacies of the monometallic Au and Pd NPs. As seen in Fig. 9, the monometallic Au or Pd NPs show more or less comparable results, however the catalytic activity signicantly improved when we used bimetallic catalysts Pd@Au-SiO 2 or Au@Pd-SiO 2 , coherent with the synergistic inuence between Au and Pd. Like hydrogenation, the Pd core -Au shell catalyst (Pd@Au-SiO 2 ) displayed superior activity over the Au core -Pd shell catalyst (Au@Pd-SiO 2 ).
It needs to be mentioned that because of the poor solubility of nitriles in water, mixed aqueous solvents are oen preferred in the nitrile hydration reaction. To nd the best solvent system for our reaction, a range of mixed aqueous media were screened (Table S2 †). Almost comparable results were obtained with MeOH-H 2 O and EtOH-H 2 O at 50 C (entries 2 and 5). However, anticipating better solubility of nitriles in i PrOH over MeOH, we have chosen the i PrOH-H 2 O mixture for further optimization studies (entry 7).
As seen from Table S2, † (entries 7-10), the catalytic activity is also dependent on the ratio of i PrOH and H 2 O. The best condition was obtained when i PrOH and H 2 O were used in a 1 : 1 ratio. On decreasing the i PrOH-H 2 O ratio the reactivity declined, presumably due to a decrease in the solubility of substrates with decreasing i PrOH count in the reaction mixture. An increase in the reaction time resulted in a pragmatic Fig. 9 Comparison of conversion and selectivity obtained for different catalysts: SiO 2 @Au (Au: 0.2 mol%); SiO 2 @Pd (Pd: 0.2 mol%); Au@Pd-SiO 2 (Au: 0.12 mol%, Pd: 0.08 mol%); Pd@Au-SiO 2 (Au: 0.09 mol%, Pd: 0.11 mol%) (metal concentrations in the catalysts were determined by ICP-AES analysis). increase in conversion (Table S3, † entries 3-5). Furthermore, on increasing the catalyst : substrate ratio, a subsequent increase in conversion was observed (Table S3, † entry 6). In addition to routine screening, a temperature optimization study was performed to explore the effects of temperature on conversion and selectivity (Table S3 †). It was observed that a marginal increase in temperature to 60 C leads to almost quantitative conversion (95%) without any variation in selectivity (Table S3, † entry 5). Any further increase in the temperature had no adverse effect on either activity or selectivity (Table S3, † entries 7 and 8). In fact, no trace of acid by-product was recorded in any of the performed optimization experiments.
With the optimized reaction conditions ( i PrOH : H 2 O (1 : 1), 60 C, 1 h, 0.5 mmol of substrate and 0.2 mol% of Pd@Au-SiO 2 [Au (0.09 mol%), Pd (0.11 mol%)]), we further examined the scope of our catalyst for the hydration of a range of nitriles to amides and the results are presented in Table 4. It was evident that most of the ortho-substrates demonstrated low conversions compared to the meta-and para-analogues. A similar trend was observed in a representative study by Cadierno et al. irrespective of the nature of the aromatic ring (aryl, pyridyl, piperidyl, thienyl, etc.). 79 Electron-negative functionalities (4d-f and 4j) exhibited good-to-excellent conversions (80-99%) with excellent selectivity (99%). Conversely, nitriles containing electron donating groups like -CH 3 (4g) and -OCH 3 (4i) exhibited moderate conversions with excellent selectivity >99%. Likewise, the catalyst was also screened for nitrogen-containing heteroaromatic substrates (4a-c). Depending on the relative position of the -CN group and the heteroatom, noticeable differences in the activities were observed. Almost quantitative conversions were achieved for 2-cyanopyridine and 4-cyanopyridine furnishing picolinamide (4b) and isonicotinamide (4c), respectively, aer about 30 min. However, a much longer reaction time (1 h) was required to obtain nicotinamide (4a) from 3cyanopyridine in high yield. Plausible reasons for this difference in reactivity incline towards resonance effects since in (4a) the nitrile carbon has a reduced electrophilicity (negative charges are always located on the carbon atoms of the pyridine ring vs. N atoms in the case of 4b and 4c), and therefore the nucleophilic addition of water at this position is less favoured. 79 Nicotinamide (4a), a form of vitamin B 3 , is used as a common dietary supplement 80 while picolinamide (4c), used as a template for the synthesis of molecular imprinting polymers, 81 is particularly important for pharmaceutical and industrial applications. Smooth conversion as well as selectivity were recorded in the case of haloaromatic substrates (-F, -Cl and -Br) (4d-f). Credibly, our catalyst could also tolerate benzylic nitrile derivatives (4l-m) yielding the desired product with moderate-to-good conversion and excellent selectivity.

Kinetic studies
The kinetics for the conversion of benzonitrile to benzamide was investigated by monitoring the reaction at different time intervals and plots of conversion versus time are presented in Fig. 10 which in most cases followed a parabolic curve. A linear t of rst order was observed in each case as was evidenced from the stacked plots of C 0 /C t versus time (Fig. 10a) and ln(C 0 / C t ) versus time (Fig. 10b), where C 0 and C t are the concentrations of benzonitrile at time 0 and t respectively. From the slope of the plot, the rst order rate constants were calculated and found to correspond to the following order of efficacies, Pd@Au-SiO 2 (4.448 Â 10 À4 s À1 ) > Au@Pd-SiO 2 (3.617 Â 10 À4 s À1 ) > SiO 2 @Au (2.979 Â 10 À4 s À1 ) > SiO 2 @Pd (2.536 Â 10 À4 s À1 ). The relationship between the relative rates and the Hammett parameter (s) for the hydration of various nitriles by Pd@Au-SiO 2 was examined.
The order of reactivity for hydration of benzonitrile was 4-NO 2 (k X /k H ¼ 2.71) > 4-F (1.29) > 4-H (1.00) > 4-CH 3 (0.82) > 4-OCH 3 (0.82). The relative rates [log(k X /k H )] were plotted against the substituent constant s (Fig. 10c). There was a fairly good linearity between log(k X /k H ) and s (r 2 ¼ 0.99), and the s slope of the linear line in Fig. 10c gives a Hammett, r (rho) value of +0.79. The typical positive r value indicated that the reaction is largely favoured by the presence of electron withdrawing groups at the para-position.

Reusability and heterogeneity
In order to test the heterogeneity of the material, Pd@Au-SiO 2 reusability tests were performed using benzonitrile as the model substrate. To evaluate this property the catalyst was carefully separated from the reaction mixture by centrifugation, ltration and subsequent washing with i PrOH-H 2 O (1 : 1), and it was dried and reused for successive runs (Fig. 11). The tests were performed maintaining the previously used stoichiometric ratio (Table S3, † entry 5). This result suggested that the catalyst could be used for ve successive runs without compromising much with the activity. A marginal loss in selectivity was observed in subsequent runs which could be assigned to the gradual blockage of some of the active sites in the silica matrix as well as due to unavoidable weight losses encountered during recycling.
Reusability tests were also performed for the other three catalysts under the optimized reaction conditions which exhibited minimal change in activity or selectivity. The extent of Au/Pd leaching into the solution has been estimated using ICP-AES analysis of the ltrate aer the 5th run. The analysis  To check the heterogeneity of the catalyst, hot ltration test was performed and benzonitrile was used as the model substrate. The reaction was stopped aer 15 minutes and the conversion was recorded to be 24% (analyzed by GC-MS). The reaction mixture was separated by ltration under hot conditions and the ltrate was allowed to stir continuously for another hour. The conversion was found to be almost constant which conrmed the heterogeneous nature of the catalyst. The morphology of the catalyst aer the h run was examined using HRTEM analysis (Fig. 11c) which exhibited a ne size distribution (inset) just like the fresh material. The HRTEM micrograph showed that the surface morphology of the catalyst remained almost intact without any particle agglomeration. This result is in agreement with the notion that the material is catalytically stable to be used for yet more catalytic runs.
To apprehend the advantages of our catalyst, a comparative study was carried out with literature reported Au-or Pd NP based catalysts immobilized onto different supports (Table 5). Although, the activity and selectivity of some of those catalysts are very similar, in most cases the metal content was radically high ranging from 2-10 mol%. Moreover, most of the reported systems are monometallic in nature and either required long intervals [e.g., Pd(OAc) 2 ] 83 or high temperatures (e.g., Pd-PVP, Pd/C-500ox), 84,88 thus confronting with the sustainability issue. Hence, it is apparent from Table 5 that our bimetallic Pd@Au-SiO 2 catalyst is much more superior to the reported catalysts.

Conclusion
In summary, we have methodically synthesized Au-Pd coreshell nanoparticles supported on 2-diphenylphosphinoethyl functionalized silica gel by a sequential reduction technique. Merely, by controlling the addition sequence of metal salts, both galvanic (Au core -Pd shell ) and anti-galvanic (Pd core -Au shell ) type core-shell nanoparticles were synthesized. We have mainly focused our study on two intriguing aspects, rstly, the differences in structural morphology of core-shell nanoparticles introduced through fabrication via the preemptive ligand assisted method and secondly, the catalytic performances of the materials have been measured in terms of their respective synergistic effects for two signicant reactions viz. hydrogenation of nitroarenes to amines and hydration of nitriles to amides under mild conditions. Remarkably, with one of our catalysts, we could achieve high conversion as well as excellent selectivity for both the aforementioned reactions. The protocol reported herein for the synthesis of Au-Pd core-shell nanoparticles anchored to functionalized silica for efficient hydrogenation to amines and nitrile hydration to amides may open new avenues in synthetic methodologies, and environmental and catalytic research in future.

Conflicts of interest
There are no conicts to declare.