Chapter 2. Cu 2 O-BiVO 4 Photoelectrochemical Cells with Molecular Multilayers for Bias-Free
2.3. Results and Discussion
Fully solution-processable PEC cells for visible-light-driven overall water splitting were prepared by depositing HER and OER CMs on Cu2O and BiVO4 photoelectrodes, respectively, through LbL assembly (Figure 5). They were selected as a model photocathode and photoanode, respectively, and readily prepared by electrodeposition according to the method described in the literature.10, 31, 32 Despite their promising properties, they often suffer from low catalytic activity and stability under visible light illumination (Figure 5a). For example, Cu2O is quickly deactivated to CuO and Cu by self-oxidation and reduction reactions, respectively. BiVO4 has poor photocatalytic activity because of the fast recombination of photogenerated excitons in the bulk and sluggish water oxidation kinetics at its interface with electrolyte.10, 15, 33 To address these problems, conventionally, they were modified with various functional materials using material-specific and harsh processes that required toxic/hazardous chemicals and consumed large amounts of energy. In particular, the stability of Cu2O is too low, so it has been utilized after the deposition of protective film, such as Al-doped ZnO and TiO2, using vacuum deposition methods. In this study, we utilized LbL-assembled CMs, which can be readily prepared in an environmentally friendly manner. NiPOM with the molecular formula of [Ni4(H2O)2(PW9O34)2]10–
and CoPOM with that of [Co4(H2O)2(VW9O34)2]10– were used as molecular HER and OER catalysts, respectively (Figure 6),18, 19 and integrated into the respective CMs by using cationic PEI as an electrostatic adhesive (Figure 5b). Considering our recent report that the performance of various photoanodes can be significantly improved by modifying their surface with LbL-assembled OER CMs, it was anticipated that a Cu2O photocathode, a BiVO4 photoanode, and a bias-free PEC cell for overall water splitting could be readily fabricated using the LbL method in an environmentally friendly manner, without using toxic/hazardous chemicals or high-energy processes. The performance of both photoelectrodes was expected to significantly improve after the deposition of the respective CMs for
31
the following reasons: (1) enhanced catalytic charge transfer due to the deposition of the catalysts, (2) improved charge transport to the catalysts due to the formation of interfacial dipoles, and (3) increased stability due to the formation of protective coating layer (Figure 5c).
32
Figure 5. Schematic illustrations explaining (a) the inherent problems of a PEC cell composed of a Cu2O photocathode and BiVO4 photoanode, (b) the experimental procedures for addressing them, and (c) the expected results.
Figure 6. Cyclic voltammograms showing the HER and OER catalytic activity of NiPOM and CoPOM, respectively.
33
In this research, fabrication, characterization, and PEC measurement of Cu2O photocathode and (PEI/NiPOM)n multilayers were conducted by Mr. Hyunwoo Kim in School of Energy and Chemical Engineering, Ulsan National Institute of Science and Technology.
The fabrication of a Cu2O-based photocathode with the (PEI/NiPOM)n multilayers (i.e., HER CMs) was confirmed by scanning electron microscopy (SEM) and various spectroscopic analyses (Figure 7a–e). Dense Cu2O thin film with cubic and octahedral crystallites on a scale of several hundred nm was uniformly deposited by electrodeposition over the entire surface of a gold-coated fluorine- doped tin oxide (FTO) substrate (Figure 7a). The thickness of Cu2O was about 1 m when it was deposited at a charge density of 0.98 C cm–2. The deposition of the HER CMs using the LbL method led to the formation of a uniform and conformal coating on the rough Cu2O photocathode (Figure 7b).
Analyses with UV/visible spectrophotometry, quartz crystal microbalance (QCM), and ellipsometry showed that the deposited amounts of PEI and NiPOM gradually and linearly increased with the number of bilayers BL (n) (Figure 7a–d). The HER CMs yielded negligible absorbance (Figure 7b inset) compared to that of the bare Cu2O photocathode (Figure 7a inset). According to the QCM analysis, the areal molar densities of the deposited PEI and NiPOM were 5.17 × 10–9 and 2.35 × 10–9 mol cm–2, respectively, in terms of monomer concentration when 10 BL were deposited (Figure 7c). The average thickness of each BL was 1.8 (±0.31) nm (Figure 7d), slightly larger than the diameter of NiPOM (~1 nm). The uniform deposition of the HER CMs was also confirmed by XPS analysis, where additional peaks corresponding to N from PEI and Ni, W, and P elements from NiPOM were observed (Figure 7e) with a decreased Cu peak (data not shown).
Based on these findings, we investigated the PEC performance of the Cu2O photocathodes for visible-light-driven hydrogen evolution in the presence and absence of the HER CMs. The performance was evaluated by linear sweep voltammetry (LSV) under periodic visible-light illumination (Figure 8a, b and Figure 9). It seemed that the bare Cu2O photocathode generally exhibited a higher current density than its modified counterpart. However, the bare Cu2O exhibited non-zero current density under dark conditions even at a low applied bias of 0.35 V vs. reversible hydrogen electrode (RHE), indicating the presence of unwanted side reactions. It is well-known that Cu2O can be converted and deactivated to Cu and CuO under light irradiation in the presence of water:34-36
Cu2O + H2O + 2e− ⇌ 2Cu + 2OH− Eo= 0.82 V 𝑣𝑠. RHE 2CuO + H2O + 2e−⇌ 2Cu2O + 2OH− Eo= 0.95 V 𝑣𝑠. RHE
34
Figure 7. Formation of the HER CMs on the Cu2O photocathode. (a, b) Cross-sectional SEM images and (insets) UV-vis absorption spectra of (a) the Cu2O photocathodes (a) on FTO and (b) the HER CMs deposited on a slide glass. Raw and false-colored images were shown together for comparison.
The deposition of the HER CMs was also confirmed by (c) QCM, (d) ellipsometry, and (e) XPS.
35
Figure 8. Improved PEC performance of the Cu2O photocathode after the deposition of the HER CMs. (a) LSV curves for the Cu2O photocathodes before and after the deposition of HER CMs under periodic light illumination. (b) The effect of the number of HER BLs on the deactivation potential (Cu2O/Cu) and current density (JCu2O/Cu) of the Cu2O photocathode. (c) Chronoamperograms and SEM images showing the improved stability of the Cu2O photocathode after the deposition of 15 BL of the HER CMs.
36
Figure 9. LSV curves showing the effect of the number of HER BLs on the performance and stability of the Cu2O photocathode. LSV curves were measured under periodic and continuous illumination.
37
We also found that the LSV curve measured under light illumination exhibited a local maximum peak. Given a cathodic potential sweep for LSV analysis, this peak was attributed to the deactivation of Cu2O through its reduction to Cu. To evaluate the degree of deactivation, we defined the first cathodic peak potential as a deactivation potential (Cu2O/Cu) and dark current density measured at an applied bias of 0.35 V vs. RHE as a deactivation current density (JCu2O/Cu). The deposition of the HER CMs led to a decrease of both light and dark current densities. As the number of BL increased, JCu2O/Cu approached zero and Cu2O/Cu exhibited a cathodic shift (Figure 8a, b), implying the suppression of Cu2O reduction to Cu.37 Interestingly, the deposition of more BL also resulted in a decrease in the intensity of a cathodic and anodic transient spike in the photocurrent (Figure 8a and Figure 9).
According to the literature, a reduction of these spikes indicates suppression of exciton recombination on the surface and in the bulk of a photoelectrode.33, 38 These results indicate that the Cu2O photocathode with the HER CMs exhibited improved PEC performance in terms of both catalytic activity and stability.
Encouraged by these findings, we investigated the effect of the HER CMs on the stability of the Cu2O photocathode by monitoring the changes in current-density and morphology for a prolonged time. As shown in Figure 8c, the Cu2O photocathode with 15 BL exhibited a more stable photocurrent density of about 0.2 mA cm–2 for more than 1 hour compared to the bare counterpart at an applied bias of 0.35 V vs. RHE. Ex situ SEM analysis showed that upon the PEC test, the roughness of the bare Cu2O photocathode was rapidly increased due to the formation of nanoparticles, which can be attributed to the self-reduction of Cu2O to Cu.37 In contrast, the Cu2O photocathode with the HER CMs had a negligible morphology change even after 1 hour of the PEC test. The improved stability of the Cu2O photocathode after the deposition of the HER CMs was also confirmed by ex situ XPS analysis (Figure 10). XPS spectra of the bare Cu2O showed a significant shift of the peaks corresponding to Cu 2p 1/2 and Cu 2p 3/2, implying considerable change in the binding energy of Cu due to the deactivation of Cu2O by self-redox reactions. On the contrary, there was negligible change in XPS spectra of the Cu2O with the HER CMs, demonstrating the improved stability of Cu2O by the HER CMs. Although we were able to significantly improve the stability of Cu2O photocathode by the deposition of HER CMs using a simple, green solution process, further improvements are required for practical application.
Nevertheless, the improved stability is quite impressive compared to relevant studies by other groups, which reported the formation of inorganic passivation layers using vacuum deposition methods.34-36
38
Figure 10. Ex situ XPS spectra of the bare Cu2O and the Cu2O with 10 BL of the HER CMs before and after the PEC hydrogen evolution test at an applied bias of 0.35V vs. RHE for 1h. (Black dotted line for Cu2O, Blue dotted line for Cu2O-Ni10BL)
Figure 11. Performance of Cu2O photocathodes modified with Pt nanoparticles (Cu2O-Pt). (a) Chronoamperograms of various Cu2O photocathodes—the bare Cu2O, Cu2O with Pt nanoparticles, and Cu2O with the HER CMs-are shown for comparison. (b, c) SEM images of the Cu2O-Pt photocathode before and after the PEC test at an applied bias of 0.35 V vs. RHE for 1 h.
39
As a control, we also prepared a Cu2O photocathode modified with Pt nanoparticles, a well- known HER catalyst, by the electrodeposition method (Cu2O–Pt) and compared its performance with that of the Cu2O photocathode with the HER CMs (Figure 11). Although the Cu2O–Pt photocathode showed a much higher initial photocurrent density than the Cu2O with the HER CMs, it quickly deactivated, as shown in Figure 11a. SEM images showed that its deactivation was caused by both the detachment of Pt nanoparticles and conversion of Cu2O microcrystals to inactive Cu nanoparticles through self-redox reactions, which was also found in the bare Cu2O photocathode (Figure 11b, c).
These results show that the Cu2O photocathode with the HER CMs exhibits higher HER performance and stability than conventional Pt catalysts.
Electrochemical impedance spectroscopy (EIS) and Mott-Schottky (M-S) analysis were carried out to elucidate the underlying mechanism for the observed performance improvement for the Cu2O photocathode with the HER CMs. The impedance spectra were well-fitted using a 2-RC equivalent circuit model regardless of the presence of the HER CMs (Figure 12a, b and Table 1). There was a negligible difference between the Rs values, which represents the series resistance related to ionic conduction through an electrolyte and electronic conduction through an external circuit. One can expect that the deposition of HER CMs facilitates catalytic charge transfer at the photocathode–electrolyte interface, resulting in a significant decrease of catalytic charge transfer resistance (R2) in the Nyquist plot (Figure 12a) and a shift of a low frequency peak for the catalytic reaction to a higher frequency in the Bode plot (Figure 12b). Unexpectedly, there was also a considerable decrease of the R1 value, which is related to the transport of charge carriers from a bulk photocathode to catalysts. M-S analysis was carried out to study the origin of the improved charge transport by comparing the charge carrier density (NA), flat-band potential (EFB), and Helmholtz-layer capacitance (CH) of the Cu2O photocathode before and after the deposition of the HER CMs. They can be estimated using the following M-S equation:39
1 𝐶2= 1
𝐶𝐻2+ 1 𝐶𝑠𝑐2
1
𝐶𝑠𝑐2 = 2
𝜖𝜖0𝑒0𝑁𝐴(𝐸 − 𝐸𝐹𝐵−𝑘𝑇 𝑒0)
where C and Csc are the overall and space charge capacitance, and are the electrode and free space permittivity, e is the elementary electric charge, E is the applied bias, k is the Boltzmann constant, and T is the absolute temperature. The negative slope of the M-S plot suggests that Cu2O exhibited p-type
40
semiconductivity. While there was a negligible effect on NA, the deposition of the CMs led a shift of the EFB of the photocathode from 0.50 to 0.39 V (Figure 12c). It is thought that the interfacial dipoles formed by the deposition of cationic polyelectrolytes and anionic NiPOM affect the band-structure near the photocathode–electrolyte interface40 and facilitate charge transport to the catalysts. In addition, we observed a significant decrease of CH, which can be attributed to the burial of more conductive Cu2O after the deposition of the less conductive CMs.41 Taking together the results of EIS and M-S analysis, we can conclude that the deposition of HER CMs improved the catalytic activity of the underlying photocathode by facilitating charge transport/transfer processes as well as its stability by rapidly scavenging photogenerated electrons and reducing its exposure to the electrolyte.
41
Table 1. The fitting results for the EIS spectra (Figure 12a) of the Cu2O photocathode with and without 15 BL of the HER CMs.
R (Ω) RS R1 R2
Cu2O 30.26 79.5 992.3
Cu2O-15BL 32.73 38.25 589.3
42
Figure 12. Analysis of the underlying mechanism for the improved PEC performance of the Cu2O photocathode after the deposition of the 15 BL of the HER CMs. (a, b) EIS analysis was carried out under visible light illumination and represented in the form of (a) Nyquist and (b) Bode plots. (c) The charge carrier densities (NA), flat-band potentials (EFB), and Helmholtz-layer capacitance (CH) of the Cu2O photocathodes were estimated before and after the deposition of the HER CMs using M-S analysis.
43
The improved charge separation kinetics after the deposition of the HER CMs was indirectly demonstrated by calculating the charge separation efficiency (sep) through comparison between the photocurrent densities in the presence and absence of Na2S2O8, an efficient electron scavenger. sep was calculated using the following equation:42, 43
∅𝑠𝑒𝑝= 𝐽𝐻2𝑂 𝐽𝑠𝑐𝑎𝑣𝑒𝑛𝑔𝑒𝑟
where JH2O and Jscavenger are the photocurrent densities measured without and with the corresponding scavengers, respectively. As shown in Figure 13, the charge separation efficiency of the Cu2O photocathode was significantly improved after the deposition of the HER CMs.
To build a bias-free PEC cell, we also prepared a photoanode for solar water oxidation by modifying the surface of nanoporous BiVO4 with (PEI/CoPOM)n multilayers (i.e., OER CMs). SEM and transmission electron microscopy (TEM) images showed that the OER CMs were uniformly and conformally coated even on a highly porous BiVO4 photoanode (Figure 14a, b). The OER CMs were readily distinguished from the underlying BiVO4 photoanode by contrast difference in the TEM images.
The morphological change of BiVO4 photoanodes after the modification was also investigated using high-resolution TEM (HRTEM), scanning TEM (STEM), and elemental mapping analyses. As shown in Figure 15, the OER CMs were conformally and uniformly coated even on the surface of nanoporous BiVO4 photoanodes. Considering that it might be more difficult to uniformly coat nanoporous BiVO4
photoanodes than relatively flat Cu2O photocathodes with the corresponding CMs, it is thought that the HER CMs could also be readily deposited on the surface of Cu2O photocathodes. The gradual and linear growth of the OER CMs was confirmed again with UV/visible absorbance spectroscopy, QCM, ellipsometry, and XPS analysis (Figure 16). According to ellipsometry, the average thickness of each BL on the BiVO4 photoanode was 4.4 (±0.7) nm, much thicker than that on the Cu2O photocathode.
Given that the structure and size of CoPOM are very similar to those of NiPOM, the observed difference in the average BL thickness is attributed to the morphological difference of BiVO4 and Cu2O photoelectrodes: 4.4 nm for nanoporous BiVO4 vs. 1.8 nm for dense Cu2O.
44
Figure 13. (a) Comparison between the charge separation efficiency for the bare Cu2O and the Cu2O with 15 BL of the HER CMs. (b) Chronoamperograms measured for the calculation of the charge separation efficiency. 0.1 M sodium persulfate (Na2S2O8) was used as an electron scavenger.
45
Figure 14. Formation of the OER CMs and their influence on the PEC performance of the BiVO4
photoanodes. (a, b) Cross-sectional SEM and (insets) TEM images of the BiVO4 photoanode (a) before and (b) after the deposition of 10 BL of the OER CMs. (c) LSV curves and (d) summarized results showing the effect of the OER CMs on the PEC performance of the BiVO4 photoanode. (e) EIS analysis was carried out to elucidate the mechanism underlying the performance improvement. (f) LSV curves showing the superior performance of the OER CMs on the BiVO4 photoanode compared to well-known OER catalysts, such as CoPi and NiOOH.
46
Figure 15. TEM, STEM and elemental mapping analyses of BiVO4 photoanodes before and after the modification with 10 BLs of the OER CMs.
Figure 16. Formation of the OER CMs was investigated with (a) UV-vis absorption spectroscopy, (b) QCM analysis, (c) ellipsometry, and (d) XPS.
47
The PEC performance of the BiVO4 photoanode for visible-light-driven water oxidation was measured under visible-light illumination in the presence and absence of OER CMs (Figure 14c, d and Figure 17). Regardless of the presence of the CMs, back-side illumination (substrate–electrode side) resulted in a much higher photocurrent density than front-side illumination (electrolyte–electrode side) (Figure 17), due to the suppression of exciton recombination.44 Similar to the case of the Cu2O photocathode, the BiVO4 photoanode exhibited an excellent performance when the OER CMs were deposited. The PEC performance of the BiVO4 photoanode was highly dependent on the number of BL.
The best performance was observed when 10BL of the OER CMs were deposited. After the deposition, the onset potential for solar water oxidation was shifted from 0.63 to 0.24 V vs. RHE, and the photocurrent density at 1.23 V vs. RHE was increased from 0.78 to 2.34 mA cm–2. It is noteworthy here that the observed cathodic shift in the onset potential of about 400 mV is one of the largest cathodic shifts for the BiVO4 photoanode. While the onset potential remained almost constant for the BiVO4
photoanode with the OER CMs, the photocurrent density rapidly increased up to 10 BL and decreased thereafter.
According to EIS analysis (Figure 14e and Table 2), the considerably improved performance of the BiVO4 photoanode after the deposition of the OER CMs resulted from (1) the improved charge transport to the interface and (2) enhanced catalytic activity, similar to the case of the Cu2O photocathode with the HER CMs. sep for the BiVO4 photoanodes in the presence and absence of the OER CMs was calculated using a method similar to the case of the Cu2O photocathodes. Note that Na2SO3 was used as a hole scavenger for BiVO4. As shown in Figure 18, sep of the BiVO4 photoanode was significantly improved at both a low and high applied bias after the deposition of the OER CMs.
The underlying mechanism for the improved sep is unclear at this stage, requiring further studies using spectroscopic analysis such as transient absorption and photoluminescence spectroscopies. As a result, our OER CMs exhibited superior performance for solar water oxidation compared to well-known OER catalysts (Figure 14f), such as cobalt phosphate (CoPi) and nickel oxyhydroxide (NiOOH). In this study we will focus on the fabrication of fully solution-processable PEC devices with the CMs and additional studies on the principles underlying the outstanding performance we observed will be reported soon.
48
Figure 17. LSV curves showing the effect of (a) the number of OER BL and (b) illumination direction on the performance of the BiVO4 photoanode.
Table 2. The fitting results for the EIS spectra (Figure 14e) of the BiVO4 photoanode with and without OER CMs.
RS(Ω) R1(Ω) CPE1 (F) R2(Ω) CPE2 (F)
BiVO4 26.51 360.3 7.66 x 10-5 2.00 x 105 7.95 x 10-5
BiVO4-5BL 31.95 173.3 4.58 x 10-5 988.4 1.75 x 10-4
BiVO4-10BL 21.58 123.2 4.00 x 10-5 727.6 4.39 x 10-4
BiVO4-15BL 24.42 134.2 6.90 x 10-5 832.4 3.25 x 10-4