Electronic Structure Evolution with Composition Alteration of RhxCuy Alloy Nanoparticles

The change in electronic structure of extremely small RhxCuy alloy nanoparticles (NPs) with composition variation was investigated by core-level (CL) and valence-band (VB) hard X-ray photoelectron spectroscopy. A combination of CL and VB spectra analyses confirmed that intermetallic charge transfer occurs between Rh and Cu. This is an important compensation mechanism that helps to explain the relationship between the catalytic activity and composition of RhxCuy alloy NPs. For monometallic Rh and Rh-rich alloy (Rh0.77Cu0.23) NPs, the formation of Rh surface oxide with a non-integer oxidation state (Rh(3−δ)+) resulted in high catalytic activity. Conversely, for alloy NPs with comparable Rh:Cu ratio (Rh0.53Cu0.47 and Rh0.50Cu0.50), the decreased fraction of catalytically active Rh(3−δ)+ oxide is compensated by charge transfer from Cu to Rh. As a result, ensuring negligible change in the catalytic activities of the NPs with comparable Rh:Cu ratio to those of Rh-rich and monometallic Rh NPs.


ABSTRACT
The change in electronic structure of extremely small Rh x Cu y alloy nanoparticles (NPs) with composition variation was investigated by core-level (CL) and valence-band (VB) hard X-ray photoelectron spectroscopy. A combination of CL and VB spectra analyses confirmed that intermetallic charge transfer occurs between Rh and Cu. This is an important compensation mechanism that helps to explain the relationship between catalytic activity and composition variation of the Rh x Cu y alloy NPs. For monometallic Rh and Rh-rich alloy (Rh 0.77 Cu 0.23 ) NPs, the formation of Rh surface oxide with a non-integer oxidation state (Rh (3−δ )+ ) resulted in higher catalytic activity. Whereas, for alloy NPs with comparable Rh:Cu ratio (Rh 0.53 Cu 0.47 and Rh 0.50 Cu 0.50 ), the decreased fraction of catalytically active Rh (3−δ )+ oxide is compensated by charge transfer from Cu to Rh. As a result, ensuring negligible change in the catalytic activities of the NPs with comparable Rh:Cu ratio to those of Rh-rich and monometallic Rh NPs.

Initial characterization Transmission Electron Microscopy
Transmission-electron microscopy (TEM) images were recorded by using Hitachi HT7700 with the acceleration voltage of 100 kV. TEM images and respective size distribution histogram are shown in Figure SI 1. On the basis of TEM results and their standard deviation the size of NPs studied in this work can be treated as almost the same. Therefore, size effects are not considered in the discussion related to the electronic structure modification with compositional alteration.

Three-way catalytic activity measurement
The synthesized Rh x Cu y alloy NPs were supported on γ-Al 2 O 3 catalysts by wet impregnation. The amount of NPS was 1 wt % of γ-Al 2 O 3 . Temperature dependent three-way catalytic activity was measured on a gas mixture of CO: 5046 ppm, O 2 : 4920 ppm, H 2 : 1758 ppm, CO 2 : 12.45 %, C 3 H 6 : 467 ppm, NO: 1161 ppm and He at a flow rate of 200 ml min −1 with a sample amount of 0.2 g (SV=60,000 mL/gh), where SV is space velocity. Temperature dependent catalytic activity data are shown in Figure SI 2 for (a) NO x , (b) CO and (c) hydrocarbon conversions to less harmful N 2 , CO 2 and water respectively. Please note that catalytic activity data for Rh 0.53 Cu 0.47 are very similar to the data obtained for Rh 0.50 Cu 0.50 NPs and was omitted in order to avoid visual overload in Figure SI 2. Additionally, the NPs with x 0.6 are intuitively fall into category of Rh-rich alloys and hence, the inclusion of these samples will strengthen discussion but hardly will provide qualitatively new information, hence not shown here.

Core level and valence band HAXPES
Influence of polyvinylpyrrolidone (PVP) Alloy NPs were prepared by solid-solution synthesis and covered with an polyvinylpyrrolidone (PVP) to prevent NPs from agglomeration. The ratio of PVP to the overall metallic content was estimated as large as 20 wt %, resulting in the thickness of protective agent about few nm. For NPs of 2 nm size this is a substantial fraction. The overall effect: (i) chemical reactivity and (ii) its insulating nature, of PVP should not be underestimated. Figure SI 3 and Figure SI 4 show N 1s and C 1s core-level (CL) spectra for NPs and including bulk references (C 1s CL spectra). Acquired CL HAXPES data showed similar trend for all samples, reassuring that influence of the protective agent on electronic structure of NPs is the same. Therefore, all differences which are discussed in the manuscript, are originate from the change in electronic structure as a function of compositional alteration.

Cu 2p CL fit results
Fit results of Cu 2p CL HAXPES data clearly indicate an oxidation of Cu (see Figure SI 5 shadowed area indicative for Cu 2+ ) for alloy NPs with comparable Rh:Cu ratio (2 nd group). This finding supports presence of charge transfer mechanism from Cu to Rh as suggested based on the fit results of Rh 3d CL HAXPES data shown in the main manuscript. Whereas, Rh-rich alloy NPs exhibit only partial oxidation of Cu, refer to Figure SI 5 (b).

Valence band linear combination fit results
Linear combination fit (LCF) results by fitting each sample spectrum with weighted mixtures (accounting for photo-ionization cross-section) of the bulk Rh and Cu references are shown in Figure SI 6. Our data support presence of: (i) partial oxidation of Rh for monometallic and Rh-rich alloy NPs (1 st group) and charge transfer mechanism from Cu to Rh for alloy NPs with comparable Rh:Cu ratio (2 nd group). Additionally, LCF analysis failed to reproduce experimental VB spectra, indicating a chemical mixture of Rh and Cu on the atomic level during synthesis. In the energy range near E F the mismatch between experimental and LCF spectra are most pronounced for alloy NP's with comparable Rh:Cu ratio, supporting intermetallic charge transfer mechanism.   Figure SI 6. VB fit results supporting presence of: (i) partial oxidation of Rh for monometallic and Rh-rich alloy NPs (1 st group) and charge transfer mechanism from Cu to Rh for alloy NPs with comparable Rh:Cu ratio (2 nd group). Linear combination fit (LCF) analysis based on the bulk reference compounds failed to reproduce experimental VB spectra, indicating a chemical mixture of Rh and Cu on the atomic level during synthesis. In the energy range near E F the mismatch between experimental and LCF spectra are most pronounced for alloy NP's with comparable Rh:Cu ratio, supporting intermetallic charge transfer mechanism.