In Situ Shell-Isolated Nanoparticle-Enhanced Raman Spectroscopy of Nickel-Catalyzed Hydrogenation Reactions.

Abstract Synthesis methods to prepare lower transition metal catalysts and specifically Ni for Shell‐Isolated Nanoparticle‐Enhanced Raman Spectroscopy (SHINERS) are explored. Impregnation, colloidal deposition, and spark ablation have been investigated as suitable synthesis routes to prepare SHINERS‐active Ni/Au@SiO2 catalyst/Shell‐Isolated Nanoparticles (SHINs). Ni precursors are confirmed to be notoriously difficult to reduce and the temperatures required are generally harsh enough to destroy SHINs, rendering SHINERS experiments on Ni infeasible using this approach. For colloidally synthesized Ni nanoparticles deposited on Au@SiO2 SHINs, stabilizing ligands first need to be removed before application is possible in catalysis. The required procedure results in transformation of the metallic Ni core to a fully oxidized metal nanoparticle, again too challenging to reduce at temperatures still compatible with SHINs. Finally, by use of spark ablation we were able to prepare metallic Ni catalysts directly on Au@SiO2 SHINs deposited on a Si wafer. These Ni/Au@SiO2 catalyst/SHINs were subsequently successfully probed with several molecules (i. e. CO and acetylene) of interest for heterogeneous catalysis, and we show that they could be used to study the in situ hydrogenation of acetylene. We observe the interaction of acetylene with the Ni surface. This study further illustrates the true potential of SHINERS by opening the door to studying industrially relevant reactions under in situ or operando reaction conditions.

UV-Vis spectroscopy of parent Au NPs. A UV-Vis absorption band at ~ 551 nm originating from the Localized Surface Plasmon Resonance (LSPR) of the Au NPs can be used to calculate both the average size and concentration of uncoated NPs (Figure 1b). 1 In line with the particle size distribution obtained from Transmission Electron Microscopy (TEM) measurements, the diameter of the Au NPs was calculated to be 83 nm.

Figure S3
Shell-Isolated Nanoparticle-Enhanced Raman Spectroscopy (SHINERS) activity and pinhole tests of Au@SiO 2 Shell-Isolated Nanoparticles (SHINs). The SHINERS-enhancement and quality of the SHINs was tested using Rhodamine 6G (a) and pyridine (b). Rhodamine 6G is a dye molecule with a high Raman cross-section often used to probe the SERS activity of plasmonic NPs and the subsequent loss in activity upon coating them with dielectric layers. [2][3][4][5][6] The spectra in (a) show that upon coating the Au NPs (yellow, top) with SiO 2 (grey, bottom) the Rhodamine 6G signal goes down. The spectra in (b) show pyridine tests: Pyridine is used to detect pinholes in the SiO 2 layer around the Au NPs. 3 Its lone pair can coordinate to the Au NP surface, which gives rise to chemical enhancement of the signal as well as traditional enhancement. Upon coating of the Au NPs (yellow, top) with SiO 2 (grey, bottom), the pyridine signal cannot be observed due to this loss of chemical enhancement.

Figure S4
Comparison of the different Ni-adsorbate stretching vibrations observed on the oxidized Ni/Au@SiO 2 samples. Dark blue: Oxidized Ni(Col), light blue: (smoothed to guide the eye, intensified for clarity) reduced Ni(Col), grey: reduced Ni(NO 3 ) 2 , brown: oxidized Ni(SA). Interestingly, the Raman band positions for the oxidized species differ, shifting from ~ 530 cm -1 on Ni(Col) to ~ 470 cm -1 on Ni(SA). Furthermore, for Ni(SA) an unknown band at 360 cm -1 can be observed. Note that the species found in the colloidal sample upon oxidation and reduction also differ slightly, indicating the existence of different types of nickel oxides. Furthermore, the width of the Raman bands is different for the different samples, which has been observed on other transition metals to be a measure of crystallite size and/or lattice strain, 7,8 with a broadening and blue shift of bands reported to be an indication of smaller NPs. 2,9,10 Therefore we can conclude that the different synthesis methods yield different types of Ni(O) NPs. Figure S5 shows the results of the phenylacetylene species adsorption experiments. First of all, it seems like phenylacetylene adsorbs onto the Ni(Col) surface as characteristic peaks for phenylacetylene are observed. Comparison with spectra obtained of phenylacetylene adsorption on Pt as we recently published 11 , shows similar characteristic vibrations, like ring vibrations and C=C stretching vibrations at 1000 cm -1 and 1580 cm -1 respectively. However, close inspection of the C≡C stretching region only reveals one peak at 2105 cm -1 . This peak originates from a C≡C with terminal hydrogen, like observed in the reference spectrum (red spectrum) and not chemisorbed onto a metal surface. The low intensity of the peak compared to the reference spectrum instead indicates physisorption of phenylacetylene on the Ni(Col)/Au@SiO 2 catalyst/SHINs, which may be due to the absence of metallic surface Ni.

In Situ SHINERS and DRIFTS of Acetylene Species on Ni(Col)/Au@SiO2
To investigate the interaction between phenylacetylene and the nickel colloids in more detail, the colloidal Ni catalysts were deposited on a DAVICAT SiO 2 support following the procedure by Casavola et al. 12 and phenylacetylene adsorption on the Ni/SiO 2 catalyst was investigated using Diffuse Reflectance InfraRed Fourier Transform Spectroscopy (DRIFTS). This technique allows us to derive complementary information, especially regarding C-H stretching vibrations, which are insufficiently enhanced in the SHINER spectra to use for characterization. Additionally, this experimental setup does not Ni/SiO 2 a second peak can be observed in the shape of a shoulder. Note that due to the IR inactivity of the C≡C stretching mode the peak in the red reference spectrum is very weak. The intensity of this peak in the other spectra indicate that either more phenylacetylene is present in total, giving a stronger signal, or that upon interaction with SiO 2 and Ni the stretching mode changes to such an extent that it becomes more IR-active.
require Au NPs to enhance the signal and employs a different in situ cell, the both of which allow for full reduction of the Ni NPs at the temperature determined by Vrijburg et al. 13 After reduction of the catalysts at 500 °C for 1 h in a 1:1 H 2 :He gas feed, the Ni/SiO 2 catalysts were subjected to phenylacetylene vapor saturated in N 2 . A blank experiment with just SiO 2 was carried out as well. The DRIFT spectra are displayed in Figure S5b-d, showing the different spectral regions for characteristic phenylacetylene vibrations that can be observed with IR. First of all, in the (C≡)C-H stretching region in Figure S5b, we can see a narrow, sharp peak for the reference DRIFT spectrum of free phenylacetylene (red). On a blank, dried SiO 2 reference sample, we see some broadening of this peak (green), whereas on Ni/SiO 2 (blue) we see distinct new maxima at lower wavenumbers, indicating interaction between the acetylene group and the Ni surface. The (C=)C-H stretching region in Figure S5c shows the C-H vibrations from the aromatic ring. Compared to the red reference spectrum for free, gas phase phenylacetylene, we see new peaks in both the SiO 2 and Ni/SiO 2 samples at lower wavenumbers. This either indicates interaction with the phenyl ring, the occurrence of dissociative adsorption in which the acetylene group becomes more ethylene-like, or a combination of both. Finally, in the C≡C stretching region displayed in Figure S5c, we see the symmetric stretching vibration of the acetylene group at around 2115 cm -1 . For the reference spectrum this vibration is very weak (spectrum multiplied by 10 4 × for clarity), as this vibration does not fulfill the selection rules for IR activity (change in dipole moment). 14 However, when phenylacetylene is measured on SiO 2 and Ni/SiO 2 this peak becomes much more intense due to distortion of the bond and related increased IR activity. Furthermore, the peak shifts to lower wavenumbers (2110 cm -1 ) and an extra peak is observed on the Ni/SiO 2 sample as a shoulder, all indicating direct interaction and chemisorption of phenylacetylene through the acetylene group on activated, metallic Ni catalysts.
Based on these DRIFTS results, we can say that in the SHINER spectrum in Figure S5a, no active Ni surface was present or accessible to phenylacetylene, resulting in only physisorption on either unreduced Ni surface or the SiO 2 shell. Comparison of the metal-adsorbate stretching region in the SHINER spectra in Figures 2 of the main text and Figure S5a show a broad band around 500 cm -1 , that points towards the presence of nickel oxide species. This is in line with incomplete reduction of the Ni(Col)/Au@SiO 2 catalyst/SHINs, and further confirms the difficulty of using the colloidal deposition method for the preparation of active Ni/Au@SiO 2 samples for in situ SHINERS studies.

Preparation of Ni/SiO2 samples for DRIFTS experiments
Ni/SiO 2 samples were prepared by the method of Casavola et al. 12 In short, ~800 mg DAVICAT® SI 1302 Silica Powder was mixed with a dispersion of 60 mg Ni colloids in 3 mL toluene (99+%, ACROS Organics) and 5 mL 1-octadecene (>90%, Sigma-Aldrich) while stirring. The NP/SiO 2 mixture was degassed and put under vacuum. After evaporation of toluene, the mixture was heated to 120 °C and kept at this T for 30 min. The mixture was flushed with N 2 three times and heated to 300 °C for 60 min. After cooling down to room temperature, the Ni/SiO 2 catalyst was washed alternatingly with n-hexane (99+%, ACROS Organics) and acetone (99.6%, ACROS Organics), several times. The samples were dried at 60 °C overnight, then 120 °C for 3 h and finally at 80 °C under vacuum for 3 h. A TEM sample is included in Figure S6.
DRIFTS experiments were carried out on a Bruker Tensor 27 FT-IR spectrometer equipped with an MCT detector, a Praying Mantis diffuse reflectance accessory, and a high temperature reactor cell with a KBr window. The Ni/SiO 2 powder samples were loaded into the sample cup of the reactor cell, packed on a quartz wool bed. The reactor cell was heated by an automatic temperature controller (Harrick ATC-02402). Spectra were recorded at a spectral resolution of 4 cm -1 over a range from 4000-600 cm -1 .

Figure S6
Transmission Electron Microscopy (TEM) images of Ni(Col)/SiO 2 . Due to the similar structure of DAVICAT SiO 2 and the colloidal Ni NPs the Ni NPs are hard to observe. Figure S7. (a) Shell-Isolated Nanoparticle-Enhanced Raman (SHINER) spectra of acetylene adsorption on Au@SiO 2 SHINs after undergoing the same treatment as impregnated Ni/Au@SiO 2 samples. No Raman peaks due to adsorption of acetylene on Au were observed. Only the formation of coke (between 1200-1600 cm -1 ) was observed. (b) Pyridine/pinhole test on NiCl 2 /Au@SiO 2 catalyst/SHINs after reduction. Raman peaks pointing to the presence of pyridine adsorbed on a bare Au surface can be observed at 1000 cm -1 , indicating the SiO 2 shell now contains pinholes.