Understanding the Performance of NiO Photocathodes with Alkyl-Derivatized Cobalt Catalysts and a Push–Pull Dye

Mesoporous NiO photocathodes containing the push–pull dye PB6 and alkyl-derivatized cobaloxime catalysts were prepared using surface amide couplings and analyzed for photocatalytic proton reduction catalysis. The length of the alkyl linker used to derivatize the cobalt catalysts was found to correlate to the photocurrent with the highest photocurrent observed using shorter alkyl linkers but the lowest one for samples without linker. The alkyl linkers were also helpful in slowing dye–NiO charge recombination. Photoelectrochemical measurements and femtosecond transient absorption spectroscopic measurements suggested electron transfer to the surface-immobilized catalysts occurred; however, H2 evolution was not observed. Based on UV–vis, X-ray fluorescence spectroscopy (XRF), and X-ray photoelectron spectroscopy (XPS) measurements, the cobalt catalyst appeared to be limiting the photocathode performance mainly via cobalt demetallation from the oxime ligand. This study highlights the need for a deeper understanding of the effect of catalyst molecular design on photocathode performance.

on the electrodes, likely occupying any remaining surface sites not blocked by PB6 or octanoic acid. When this electrode was soaked in methanol for five minutes, to remove any loosely bound Co, the loading of Co measured by XRF did not change, suggesting that Co was indeed covalently attached to the electrode and not loosely associated on the surface. Figure S1. Octanoic acid and L4Sil alkyl linkers used to prepare NiO-PB6-Al2O3octanoicacid, NiO-PB6-Al2O3-L4Sil, and NiO-PB6-Al2O3-L4Sil-Co.
Photocathodes with the L4Sil silatrane linker ( Figure S1) were prepared in a slightly modified method. NiO-PB6-Al2O3 was prepared as stated above. The silatrane was attached via a preacidification method 6 by soaking the electrode in 0.1 M phosphate buffer at pH 2 for one hour followed by air drying. The pre-acidified electrode was then soaked in 8 mM L4Sil in acetonitrile overnight, rinsed with solvent and air dried, affording NiO-PB6-Al2O3-L4Sil electrodes. The surface amide coupling procedure was performed as described above to give the NiO-PB6-Al2O3-L4Sil-Co electrode.
The photocathodes with ALD of TiO2 were prepared by applying four cycles of TiO2 over NiO-PB6-Al2O3-L4-Co electrodes at 150 °C using TiCl4 and water precursors pulsed at 0.1 s and then purged with N2 for 15 seconds. The thickness of the layer was 0.4 nm as measured by ellipsometry on a silicon wafer. This procedure afforded the NiO-PB6-Al2O3-L4-Co-TiO2 electrodes. This method was inspired by previously published work. 7 Photocathode Loadings. PB6 loadings were obtained by UV-Vis spectroscopic measurements on the electrode. Surface loadings (Γ) were found using the following formula: Γ(mol cm −2 ) = A(λ)/ (1000ε), where A is the absorbance at wavelength λ, and ε is the molar extinction coefficient at wavelength λ. 8 PB6 exhibits a signature peak at 530 nm in the UV-Vis spectrum having a molar extinction coefficient of 33,600 M -1 cm -1 , which was used to find the loadings (Table S1). 1 Co loadings were found by XRF spectroscopic measurements. To do this, known quantities of a Co methanol solution were drop casted onto NiO electrodes. The XRF cobalt signal was measured on these electrodes and used to construct a calibration curve ( Figure S2). The calibration curve was then used to find the Co loadings on the photocathodes based on the XRF cobalt signal obtained during measurements (Table S1).  Photoelectrochemistry. Chronoamperometric (CA) measurements were performed in an H-cell, with the working and auxiliary compartments separated by a Nafion membrane; each compartment contained 10.0 mL of 0.1 M MES buffer at pH 5. The photocathode working electrode and Ag/AgCl(sat´d NaCl) reference electrode were placed in the working chamber, while the Pt wire auxiliary electrode was placed in the auxiliary chamber. The H-cell was degassed with N2 prior to measurements. During the CA experiments, a 0.1 V vs. NHE bias was applied to the system. A white LED lamp was used to illuminate the photocathodes and was placed 10 cm from the photocathode, having approximately 100 mW/cm 2 .

Femtosecond transient absorption spectroscopy (fsTAS)
. fsTAS measurements were performed as previously described. 4 Briefly, a Coherent Libra Ti:sapphire amplifier (1.5 mJ, 3kHz, 800 nm, fwhm 40s) was split into pump and probe beams. The pump beam was directed into optical amplifiers (TOPAS-Prime and NIRUVVIS, Light Conversion) to generate 560 nm excitation wavelength. The beam was centered on the dry photocathode film with a pump power of 3.3 mW (1.1 µJ/pulse). An optical delay line was used to record the transient absorption spectra at different time points by scanning the delay of the probe beam relative to the pump beam (from -10.5 ps to 8 ns). Four scans were collected for each sample and averaged. Prior to analysis, data was corrected for spectral chirp and global analysis was performed using a home-written MATLAB script. Note that only one layer of screen printed NiO was used to prepare the fsTAS samples to provide more transparent electrodes.
Clark electrode H 2 detection. Clark electrode H2 detection measurements were performed as previously described, 4 which was based off of a recent study that showed how to measure low concentrations of dissolved H2. 9 Briefly, the Clark electrode was calibrated prior to measurements by measuring the Clark electrode signal in 0.1 MES buffer at pH 5 with known concentrations of dissolved H2. The same H-cell setup was used as in the photoelectrochemical measurements. The volume of MES buffer in the H-cell was increased, to have minimum amount of headspace in the working chamber in order to increase the chances of measuring dissolved H2 in the electrolyte. Prior to CA measurements, the Clark electrode baseline signal was measured in the electrolyte. Then, it was removed from the electrolyte and chronoamperometry was performed at 0.1 V vs. NHE for one hour under illumination (as described above in the photoelectrochemical section). After, the Clark electrode was placed back in the electrolyte and the signal remeasured. The Clark electrode was not placed in the electrolyte during CA to avoid interference from the light and electrochemical setup, as this does not give reliable Clark electrode results. 9 If H2 had been detected, the signal should have significantly increased from the baseline signal. However, we were not able to detect any significant signal increase from any of the photocathodes tested.           Figure S13. ATR-FTIR spectra of Co powder (black, A-D) and NiO-L4-Co (pink, A), NiO-L8-Co (pink, B), NiO-L11-Co (pink, C), and NiO-Co (pink, D). All samples exhibit the characteristic oxime ligand stretch at 1225 cm -1 . (E) ATR-FTIR spectra of PB6 powder (grey) and NiO-PB6 (pink). Figure S14. XPS spectra of the N 1s region of NiO-PB6-Al2O3 (pink), NiO-PB6-Al2O3-Co (burgundy), NiO-PB6-Al2O3-L4 (light blue), NiO-PB6-Al2O3-L4-Co (royal blue), and NiO-oxime (gray). The NiO-oxime ligand electrode was prepared by soaking a NiO electrode in a 0.1 mM oxime ligand (structure shown above) methanol solution for two hours, rinsed with methanol, and air dried. NiO-PB6-Al2O3 exhibits a peak in the N 1s, which is due to the N content in PB6. NiO-PB6-Al2O3-L4 exhibits a slightly broader peak in the N 1s, which is due to the N content in PB6 and in the L4 linker. The samples with Co show an even larger peak, due to the additional N content from the oxime ligand.

XPS deconvolution:
The following tables and figures show the deconvoluted components for the N 1s region for the oxime-functionalized NiO, NiO soaked in MES, and the cobaloximefunctionalized sample after 1 hour electrolysis. There is an overall increase in area in the region of 400 eV due to the oxime, while the relative ratios of the areas of the MES components do not change. XPS ratios are determined by peak area, not peak intensity.