Optimising surface d charge of AuPd nanoalloy catalysts for enhanced catalytic activity

Understanding the catalytic mechanism of bimetallic nanocatalysts remains challenging. Here, we adopt an adsorbate mediated thermal reduction approach to yield monodispersed AuPd catalysts with continuous change of the Pd-Au coordination numbers embedded in a mesoporous carbonaceous matrix. The structure of nanoalloys is well-defined, allowing for a direct determination of the structure-property relationship. The results show that the Pd single atom and dimer are the active sites for the base-free oxidation of primary alcohols. Remarkably, the d-orbital charge on the surface of Pd serves as a descriptor to the adsorbate states and hence the catalytic performance. The maximum d-charge gain occurred in a composition with 33–50 at% Pd corresponds to up to 9 times enhancement in the reaction rate compared to the neat Pd. The findings not only open an avenue towards the rational design of catalysts but also enable the identification of key steps involved in the catalytic reactions.

. XPS spectra for monometallic and bimetallic catalysts.  metallic Pd and Pd-O, which were the dominant contributors to each spectrum, were discussed. A distinct overlap could be observed between the Pd 3d 5/2 and the Au 4d 5/2 components for the bimetallic catalysts. The Au 4d 5/2 intensity was calculated from the well-resolved Au 4f 7/2 intensity, and this value was subtracted from the above overlapped peak to determine the Pd 3d 5/2 intensity. The resulting value was used to calculate the Au:Pd ratio.
where A is the pre-exponential factor, R is the universal gas constant and T is the reaction temperature. The turn over frequency (TOF) for each catalyst was calculated on the basis of the estimated number of exposed palladium atoms. This value was calculated at less than 25% conversion, where X is the conversion and t is the reaction time. Although microstructural deviations occurred in the AuPd nanoalloys, we assumed that the value for the exposed surface atom dispersion (τ) calculated from equation (3) - (6), which was based on the similarities of the truncated octahedron shape of the nanoparticles (as evidenced in the aberration corrected STEM image) and the homogeneous distribution in the alloy, was accurate 3-6 .
N T = 16m 3 -33m 2 +24m-6 (4) Wherein, d particle is the diameter of particle, N T is the total atom number of each particle, N S is the surface atom number, and m is the number of atoms lying on an equivalent edge (corner atoms included). Note that the diameter of Au (d Au ) 20 and Pd (d Pd ) atom is 0.2884 and 0.2751 nm, respectively, an average atom diameter of 0.2818 nm for d AuPd is adopted.
The enthalpy of activation is determined as following: The TOF value can be expressed in the Eyring form 7,8 : where k B , h, ΔS 0* , and ΔH 0* are the Boltzmann constant, Planck constant, entropy of activation, and enthalpy of activation, respectively.
The activation energy E a could be related to ΔH 0* by the Temkin equation: where ΔH i and n i are the adsorption enthalpies and the reaction order of reactant i, respectively. Taking into account the approximate zero-order reaction kinetics for benzyl alcohol, E a is simplified to be close to ΔH 0* .
The entropy change in the activation step of the chemical reaction is closely related to the thermodynamics of the rate constant, which results in the following equation: The scale bar is 20 nm.
Supplementary Fig. 28. Synthesis and characterization of mercapto-functionalized ordered mesoporous silica SH-SBA-15. a, FT-IR spectra for pristine and mercapto-functionalized SH-SBA-15. In comparison to that of pristine SBA-15, the FTIR spectrum of SH-SBA-15 contained several absorbances in the 3000 -2800 cm -1 range, which were assigned to the C-H vibrations in the mercaptopropyl groups. This phenomenon provides further evidence of the modification of the organic functional groups on mesoporous silica SBA-15. However, the mercapto group is IR invisible under the current conditions, which may be due to the weak dipoles of the S-H groups that hinder detection of their modes by vibrational spectroscopy 12 . b, TG curves for pristine and mercapto-functionalized SH-SBA-15. The TG curve for pristine SBA-15 revealed a weight loss below 100 °C, corresponding to the physisorption of water. In comparison, the mercapto-functionalized SBA-15 exhibited the second distinct weight loss between 250 and 500 °C, which may be due to the loss of grafted mercaptopropyl groups on the mesoporous silica. The S content was estimated to be approximately 2.5 mmol S g -1 solid, which is consistent with the value obtained by elemental analysis. c, XPS spectrum for SH-SBA-15 in the S 2p region, showing one strong peak corresponding to the -SH group. d, N 2 sorption isotherms for pristine and mercapto-functionalized SH-SBA-15. Typical type IV curves can be observed for both samples, suggesting uniform mesopores. After modification of the mercaptopropyl groups, the BET surface area and pore volume decreased from 706 to 510 m 2 g -1 , and from 1.04 to 0.78 cm g -1 , respectively, and the pore size remained at 8.9 nm. The results are similar to those of mesoporous silica modified with organic moieties and reflects grafting of the moieties inside the pores 13 .
Mercapto-group functionalized mesoporous silica SBA-15 (SH-SBA-15) was synthesized by grafting mercaptopropyl groups on pre-prepared mesoporous silica SBA-15 13,14 . Pure mesoporous silica SBA-15 was synthesized by the hydrothermal method. The initial solutions contained 2.08 g of TEOS, 1.0 g of P123, 30 g of HCl (2 M), and 7.5 g of H 2 O 15 . The hydrothermal temperature and time were 100 °C and 24 h. After surfactant removal at 550 °C under air, the mesoporous silica SBA-15 carrier was obtained. A toluene (120 mL) suspension of SBA-15 (4.1 g) was then mixed with 10.0 g of 3-thiolpropyltrimethoxysilane at reflux temperature for 48 h; 1.8 mL of water was added to promote cross-linking, and the mixture was heated at reflux for an additional 24 h. The solids were then filtered and washed with copious amounts of toluene, hexanes, and methanol to remove unreacted silanes. The recovered solids were Soxhlet extracted with dichloromethane at reflux temperature for 24 h. The resulting white solids were collected, dried at room temperature overnight and subsequently at 150 °C for 3 h under vacuum, and stored in a vacuum dryer. c Au:Pd atomic ratio estimated from the XPS spectra. A distinct overlap could be observed between the Pd 3d 5/2 and the Au 4d 5/2 components for the bimetallic catalysts. The Au 4d 5/2 intensity was calculated from the well-resolved Au 4f 7/2 intensity, and this value was subtracted from the above overlapped peak to determine the Pd 3d 5/2 intensity. The resulting value was used to calculate the Au:Pd ratio. Au 50 Pd 50 (111) p (44) 2 2 331 a vacuum spacing of 15 Å was set to reduce the interaction between repeating slabs. b computational details: All periodic spin-polarized density functional theory (DFT) calculations were carried out using the Vienna Ab-initio Simulation Package (VASP) [17][18][19][20] . The interaction between ion cores and valence electrons was described by the projector-augmented wave (PAW) method 21 , and the exchange-correlation functional was GGA-PBE 22,23 . The solution of the Kohn-Sham equations was expanded in a plane wave basis set with a cutoff energy of 450 eV. The Brillouin zone sampling was performed using a Monkhorst-Pack grid 24 , and the electronic occupancies were determined in light of a Methfessel-Paxton scheme with an energy smearing of 0.2 eV 25 . In all the calculations, a force-based conjugated gradient method was used to optimize the geometries 26 . Saddle points and minima were considered to be converged when the maximum force in each degree of freedom was less than 0.03 eV Å -1 . The optimized lattice parameter of Pd and Au is 3.956 and 4.173 Å, respectively, which is similar to the values from experiments and other DFT calculations 27,28 .
The adsorption energy of the adsorbate species, E ads , was calculated with the formula: where E adsorbate + slab is the total energy of the relaxed adsorbate-surface system, while E slab and E adsorbate are the total energy of the relaxed bare surface and gas phase adsorbate, respectively 29 . Hence, the adsorption energy was defined as negative if the total energy decreased when the adsorbate was brought from infinity and placed onto the surface.
Supplementary Table 3. The values (in units of eV) of the measured core-level binding energy shift ΔE(i), the work function change ΔФ, the final-state relaxation energy shift ΔE r , the change of the one-electron energy of the core-level where Δϵ(i) is the change of the Hartree-Fock one-electron energy of the core level i, ΔE F is the change of the Fermi energy, and ΔE r is the change of relaxation energy in the presence of the core hole between pure metal and alloy.
The change of the Hartree-Fock one-electron energy of the core level Δϵ(i), which is directly related to the charge transfer in the initial ground state, can be written as     val j j n n (13) where F 0 (i, j) is the screened Coulomb integrals between the core electron i and the valence electron j, and Δn j is the change in the number of the valence electron j. The following values for the Coulomb integrals in the solids are used: F 0 (3d, 4d) = 18.4 eV and F 0 (3d, 4s) = 12.9 eV for solid Pd 32 , and F 0 (4f, 5d) = 16.1 eV and F 0 (4f, 6s) = 12.8 eV for solid Au 33 .
is a Madelung-like potential energy associated with the total charge transfer in the surrounding host atoms. Using the metallic radii of Pd and Au (r Pd = 0.138 nm and r Au = 0.144 nm) 28 , (e 2 /r Pd ) = 10.5 eV for the Pd atom and (e 2 /r Au ) = 10.1 eV for the Au atom is obtained.
The first term in the right-hand side of Equation (12) arises from the redistribution of valence electrons, and the second term is associated with the charge changes on other lattice sites due to the total charge transfer in or out of the parent atomic site. In the case of Au (or Pd) atom, the Equation (12) can be rewritten as: We first assumed that ΔE F = -ΔФ, the difference of the experimental work functions of the metal and the alloy. This is equivalent to assuming that the surface dipole barrier potential contribution to the work function does not change between pure metal and the alloy. The term △E r is the difference of the relaxation energy in the presence of the core hole between alloy and metal, so it is related to the final-state screening effect. A the oretical calculation 34 on alloy systems based on the pseudopotential linear response method in the presence of the core hole revealed that △E r could be as large as 2 eV, which was the same order of magnitude as the initial-state effect Δϵ(i) and therefore could not be  (17) δn Au (δn Pd ) is the amount of the total interatomic charge transfer into the Au (Pd) atom.

Supplementary Methods
Preparation of phenolic resins. The carbon precursors (low molecular weight, soluble phenolic resins) were prepared from phenol and formaldehyde in a base-catalyzed process. In a typical procedure, 8.0 g of phenol were melted at 42 -45 °C in a flask and mixed with 0.34 g of a 20 wt% aqueous sodium hydroxide (NaOH) solution under stirring. After 10 min, 5.24 g of formalin (37 wt% formaldehyde) were added. Then, the mixture was heated to 70 °C. After additional stirring for 1 h at this temperature, the mixture was cooled to room temperature. The pH value was adjusted to ~ 7.0 with a 2 M HCl solution. Then, water was removed by vacuum evaporation below 45 °C. The water-and ethanol-soluble phenolic resins were dissolved in ethanol (20 wt%) for further use.