Advanced Photoelectrochemical Characterization: Principles and Applications of Dual-Working-Electrode Photoelectrochemistry

Abstract

The performance of a photoelectrode for photoelectrochemical solar fuel production can be enhanced by integrating an overlayer of electrocatalyst (EC) with the light-absorbing semiconductor (SC). However, the mechanisms through which the EC overlayer improves performance of composite photoelectrodes are not well understood. While the simple view is that the addition of the EC increases reaction kinetics, real systems are more complicated due to the existence of multiple, interacting components. Knowledge about the dynamic state of each critical component in complex photoelectrochemical (PEC) systems such as catalyzed photoelectrodes (and, e.g., dye-sensitized solar cells) could provide important insights. This information, however, is typically not directly accessible through a conventional three-electrode PEC cell setup with a single-working electrode (SWE) connected to the SC, a counter electrode, and a reference electrode. In this chapter we discuss the “dual-working-electrode” (DWE) PEC experimental technique that features an additional working electrode, which could be used to directly monitor or control the state of crucial components, such as the EC layer of a composite photoelectrode. We illustrate how this DWE PEC technique was employed to directly measure the in situ properties of SC|EC junctions in model water-oxidizing photoanodes and thus help answer the question of how the EC layer improves the PEC performance of composite photoelectrodes. We further discuss directions for future efforts in this area.

Appendix: Experimental Details

1.1 Au Thin-Film Contact

In our study, Au was chosen as the contact material because OER kinetics on Au are much slower than that of many catalysts, e.g., Ni(OH)₂ (with Fe impurities) and IrO _x , deposited on an Au substrate. A comparison of OER activities among Au, IrO _x , and Ni(OH)₂ in 0.1 M KOH solution is shown in Fig. 7.11.

Fig. 7.11

OER activity of Ni(OH)₂ (with Fe impurities) and IrO _x on an Au substrate in 0.1 M KOH. The working electrode potential is referenced to ${\mathrm{\mathcal{E}}}_{{\mathrm{O}}_2/{\mathrm{O}\mathrm{H}}^{-}}$ and all curves were collected at 50 mV/s. The OER activity of bare Au substrate is also shown for comparison

During the sample fabrication process, controlling the thickness of the Au layer is very important because the film needs to be thick enough to form a continuous conducting network but not too thick to hinder the free movement of ions into and out of the EC film (Doron-Mor et al. 2004). We use Au layers ~10–15 nm thick because the high surface tension of Au causes the film to “de-wet” from the surface and form a network of Au islands that is electrolyte permeable but electrically conductive. SEM images of a typical Ni(OH)₂ film (on an Au substrate) before and after the thin Au film deposition by thermal evaporation are shown in Fig. 7.12.

Fig. 7.12

SEM images of Ni(OH)₂ EC electrodeposited on Au substrate (a) before and (b) after Au film deposition. The scale bar is 200 nm in both images

Because the Au film evaporated on the epoxy surface surrounding the EC serves as the connecting area for WE2, it is desirable to make the epoxy surface as smooth as possible. It was found that curing the epoxy (Loctite Hysol 1C) in an oven at ~90 °C produced smoother surfaces compared to curing the epoxy at room temperature.

1.2 Sample Viability Tests

After samples for the DWE PEC experiment were fabricated, tests must be performed to ensure the quality of the device. One potential problem is that the Au contact film may prevent the electrolyte from permeating into the underlying EC film. Another potential problem is the Au film may form direct contact with the TiO₂ substrate if the EC film coverage is incomplete as a result of uneven deposition or due to the development of cracks during the drying process. If the EC film volume increases too much between the dry and wet states, it is also possible that the Au film near the edge of the epoxy surface become discontinuous which will make the EC film electrically inaccessible through WE2.

1.2.1 Test Cell and Measurement Setup

Photoanode samples were tested in a three-neck cell with an optically flat bottom for introducing light from a solar simulator at 1-sun intensity at the SC surface. The cell was constructed from fused silica to minimize parasitic adsorption of UV photons. Typically about 15 mL of 0.1 M KOH electrolyte was used, and the solution was stirred using a magnetic stir bar and sparged with pure O₂ to maintain a stable ${\mathrm{\mathcal{E}}}_{{\mathrm{O}}_2/{\mathrm{O}\mathrm{H}}^{-}}$ during PEC experiments. A Pt coil was used as the counter, and the reference was either Hg/HgO filled with 0.1 M KOH or a saturated calomel electrode.

1.2.2 Impact of the Au Contact

Deposition of an Au film on the EC layer could have a few adverse effects; it will reduce the light intensity that reaches the SC|EC junction and obscure part of the EC surface. At increased thickness, the Au film could also impede the permeation of electrolyte into the porous Ni(OH)₂ film. The effect of optical absorption/reflection by the Au film was tested first and shown in Fig. 7.13 for both ECs used. In both cases, the light intensity was reduced more than 50 % by the Au film as evident from the reduced photocurrent. Furthermore, the redox capacity of the IrO _x catalyst was reduced more significantly than that of the Ni(OH)₂ catalyst, presumably due to surface activity of the IrO _x film versus bulk activity of the Ni(OH)₂ film.

Fig. 7.13

Cyclic voltammograms of TiO₂|EC samples at 1-sun illumination before and after Au film contact deposition. (a) TiO₂|IrO _x sample at scan rate of 50 mV/s. (b) TiO₂|Ni(OH)₂ sample at scan rate of 10 mV/s. Figure adapted from Lin and Boettcher (2014). Copyrighted by Nature Publishing Group

To verify that the Au contact on the Ni(OH)₂ surface does not hinder the free flow of ions into/out of the porous Ni(OH)₂ film, oxidation and reduction of the EC were carried out through both WE1 (TiO₂) and WE2 (Au). A typical cyclic voltammetry (CV) of a Ni(OH)₂-coated TiO₂ photoanode with a top Au contact is shown in Fig. 7.14.

Fig. 7.14

Oxidation and reduction of the EC on a DWE sample. The Ni(OH)₂ EC was oxidized via the TiO₂ electrode under 1-sun illumination by sweeping V _sem negative (from points 1 to 2) then positive (from points 2 to 3) and reduced again via the Au film in contact with the EC (by sweeping V _cat through points 4–5). The scan rate was 20 mV/s. The starting point of each scan is indicated by an arrow. Integration of the oxidation wave on the Au, the reduction wave on the Au, and the reduction wave on the TiO₂ all yielded ~13 mC/cm² after background correction, indicating that the same Ni(OH)₂/NiOOH species are oxidized/reduced by both TiO₂ and Au electrodes. Figure adapted from Lin and Boettcher (2014). Copyrighted by Nature Publishing Group

The observation that oxidation and reduction of the Ni(OH)₂ catalyst can be accomplished via both WE1 and WE2 without distortion of the reduction wave indicates that the thin Au film remains permeable to ions in the electrolyte. The cathodic shift of the NiOOH reduction wave measured via the TiO₂ (WE1) relative to that via Au (WE2) is due to the photovoltage V _ph generated by the TiO₂|Ni(OH)₂/NiOOH junction. The CV of Ni(OH)₂ via WE2 is identical to those obtained by depositing Ni(OH)₂ directly on Au substrate, which confirms that the electrical contact between WE2 and the EC film is intact.

Similar tests can be performed on any EC that shows either a bulk redox wave or surface capacitive charging and discharging behavior.

1.2.3 Verifying the Independence of the Two WEs

After verifying the Au film permeability and WE2 contact integrity, the DWE sample was further tested for any possible direct Au–TiO₂ electrical short. Current-voltage curves through the SC|EC junction (between WE1 and WE2) were collected while the sample is in solution for all DWE samples, and any sample that exhibited high reverse current density (such as the green curve in Fig. 7.15, cracked EC) were excluded.

Fig. 7.15

Representative J–V curves of TiO₂|EC|Au and control TiO₂|Au samples. The control TiO₂|Au sample without catalyst showed larger, symmetric J–V curve, while the IrO _x -coated sample showed rectifying J–V curve. Samples with possibly cracked Ni(OH)₂ catalyst film showed increased reverse current, while sample with intact EC layer exhibited low reverse current. Figure adapted from Lin and Boettcher (2014). Copyrighted by Nature Publishing Group

The control samples with an Au film deposited directly on the TiO₂ substrate (red curve in Fig. 7.15) show high reverse current due to the relatively low Schottky barrier between Au and TiO₂. Based on the near symmetric J–V curve of TiO₂|Au junction, we hypothesize that high reverse current seen in some of the samples resulted from cracks in the EC film. Examination of samples with electrodeposited Ni(OH)₂ films revealed that some EC films developed cracks during the drying process (Fig. 7.16b), which probably caused direct Au–TiO₂ contact to form upon deposition of the Au layer.

Fig. 7.16

SEM images of electrodeposited Ni(OH)₂ catalyst on TiO₂ substrate. (a) Sample with continuous EC film. (b) Sample with cracked EC film that could lead to direct Au–TiO₂ contact. The scale bar is 500 nm in both images