A hole-selective hybrid TiO2 layer for stable and low-cost photoanodes in solar water oxidation | Nature Communications

Nature Communications volume 15, Article number: 9439 (2024) Cite this article

Metrics details

The use of conductive and corrosion-resistant protective layers represents a key strategy for improving the durability of light absorber materials in photoelectrochemical water splitting. For high performance photoanodes such as Si, GaAs, and GaP, amorphous TiO2 protective overlayers, deposited by atomic layer deposition, are conductive for holes via a defect band in the TiO2. However, when coated on simply prepared, low-cost photoanodes such as metal oxides, no charge transfer is observed through amorphous TiO2. Here, we report a hybrid polyethyleneimine/TiO2 layer that facilitates hole transfer from model oxides BiVO4 and Fe2O3, enabling access to a broader scope of available materials for practical water oxidation. A thin polyethyleneimine layer between the light absorber and the hybrid polyethyleneimine/TiO2 acts as a hole-selective interface, improving the optoelectronic properties of the photoanode devices. These polyethyleneimine/TiO2 modified photoanodes exhibit high photostability for solar water oxidation over 400 h.

Photoelectrochemical (PEC) water splitting is a promising route to low-cost and large-scale green hydrogen production. The durability of the light absorbers immersed in the aqueous electrolyte solution is a major factor in the cost of the resulting hydrogen1,2, and much of the efforts on simple semiconductor/electrolyte junction PEC systems have shifted towards incorporating protective overlayers to minimize (photo)corrosion of the materials. The use of such overlayers aids not only the durability, but also the optoelectronic properties of the absorber layer: the photovoltage in these coated systems is generated at a solid-solid interface, and various strategies such as contact selectivity can be used to improve the photovoltage3,4.

Amorphous titanium dioxide (a-TiO2) deposited by atomic layer depositions (ALD) has been widely employed for corrosion protection of PEC materials, primarily for photocathodes, since the conduction band of a-TiO2 is close in energy to the thermodynamic hydrogen evolution potential5,6. For smaller bandgap, high-efficiency photoanode materials, the deep valence band of the a-TiO2 overlayer represents a large barrier for hole injection that could preclude charge transfer. However, a defect band in the a-TiO2 deposited by ALD was shown to enable hole transfer, and photoanode materials such as crystalline Si, GaAs, and GaP modified with a-TiO2 have demonstrated excellent performance and stability for PEC water oxidation7.

Despite these demonstrations of photoanode stabilization using energy-intensive and/or high-cost semiconductors, the use of a-TiO2 layers on low-cost, easy-to-prepare photoanode materials, such as metal oxides8, has been rarely reported, likely due to high recombination at the photoanode/TiO2 interface from the relatively slow extraction of holes through the a-TiO2 defect band9,10. While metal oxides are investigated for large-scale water splitting due to their Earth abundance, ease of synthesis and perceived stability, they often suffer from (photo)corrosion11,12, and therefore efforts have been directed towards stabilizing them. For instance, McDowell et al. used a-TiO2 as a protective layer for BiVO4 photoanodes that showed stability for several hours13. However, the thickness of the protective layer was only 1 nm, and charge transfer could be achieved via tunneling. For long-term stability, there should ideally be no pinholes in the protective layer, and to have a good chance of achieving this, relatively thick layers of TiO2 will likely be required ( > 50 nm)14.

In this context, polyethyleneimine (PEI), which consists of repeating units of amine groups and two-carbon aliphatic spacers, was selected to alleviate unfavorable interface energetics and facilitate the hole-selective transfer. The non-conjugated PEI is known not only as a modifier of work function15,16 but also as a hole transfer channel17,18 since the amine moieties are easily oxidized, enabling hole transfer in PEC devices. We considered these unique properties of PEI to be promising for effectively addressing the aforementioned issues at the interface between metal oxides and a-TiO2.

Here, we report a hybrid PEI/TiO2 layer on both BiVO4 and Fe2O3 that not only protects these relatively small bandgap metal oxide photoanode materials from (photo)corrosion but also serves as a hole-selective contact. Although a PEI coating has been reported to lower the work function of semiconductor materials15, in this study, we embed the PEI in a-TiO2 during the ALD process by reacting the TiO2 precursor with the abundant amine functionalities of the polymer, yielding a highly defective PEI/TiO2 layer that transmits holes and blocks electrons. The BiVO4/PEI/TiO2 photoanode demonstrated an onset potential of 0.28 V vs. reversible hydrogen electrode (RHE), a photocurrent of 2.03 mA cm−2 at 1.23 vs. RHE, and stable PEC water oxidation for 400 h in pH 8 electrolyte solution.

We deposited a PEI layer onto the BiVO4 surface by spin-coating with a 4 wt.% aqueous PEI solution, followed by the deposition of nominally 100 nm a-TiO2 by ALD (BiVO4/PEI/TiO2) (Fig. 1). For comparison, we also prepared a photoanode without the PEI layer using the same procedure (BiVO4/TiO2). First, we compared the surface morphology of BiVO4/TiO2 and BiVO4/PEI/TiO2 by scanning electron microscopy (SEM). According to SEM measurements, the TiO2 layer was conformally deposited on the highly porous BiVO4 photoanodes, regardless of the presence of a PEI interfacial layer (Supplementary Fig. 1). The conformal and uniform layer of a-TiO2 was further confirmed by a pinhole test using cyclic voltammetry in a ferricyanide solution (Supplementary Figs. 2 and 3)14. The morphology of BiVO4/TiO2 and BiVO4/PEI/TiO2 photoanodes was slightly different. BiVO4/PEI/TiO2 exhibited lower porosity in comparison to BiVO4/TiO2. This result could be attributed to the PEI layer filling the pores within the porous BiVO4 structure (Supplementary Fig. 4). The thickness of the PEI layer on the BiVO4 surface was estimated as 50.5 nm through profilometer measurement on an FTO electrode (Supplementary Fig. 5).

Graphical illustration of a modified BiVO4 photoanode featuring a thin, insulating PEI layer between the BiVO4 and a hole-conductive hybrid PEI/TiO2 layer.

After confirming the uniform deposition of the TiO2 protection layer, we evaluated the PEC performance and the stability of the photoanodes. Linear sweep voltammetry (LSV) was carried out in 0.5 M potassium phosphate (KPi) buffer solution (pH 8). First, the BiVO4/TiO2 photoanode exhibited a very low photocurrent density of 0.0008 mA cm-2 at 1.23 V vs. RHE under 1 sun illumination (Fig. 2a). Even after the modification with CoOOH co-catalyst (BiVO4/TiO2/CoOOH), the photocurrent density only remained in the microampere (μA cm−2) range at 1.23 V vs. RHE. The ALD TiO2 used here has been previously shown to transmit holes on silicon photoanodes19, and thus the typical leaky TiO2 is not suitable for BiVO4.

a–c LSV curves of BiVO4/TiO2 with and without CoOOH (a), bare BiVO4 and BiVO4/PEI/TiO2 (b), and BiVO4/CoOOH and BiVO4/PEI/TiO2/CoOOH c. e Concentration of dissolved metal ion from each photoanode during stability test. The error bars represent the standard deviations of triplicate experiments.

Previous studies have established that photoanodes modified with a-TiO2 and metal co-catalysts exhibited enhanced PEC performance, attributed to the sufficient charge extraction from the TiO2 to the co-catalysts20,21. However, even with various types of metal co-catalysts, we observed insufficient improvement and non-reproducibility in the PEC performance according to the work function of the used metal (Supplementary Fig. 6). This result indicates that the photogenerated holes are unable to be transferred through the amorphous TiO2 layer when the thick TiO2 is directly deposited on BiVO4. On the contrary, BiVO4/PEI/TiO2 showed photoanodic current even without co-catalysts: photocurrent density of 1.08 mA cm−2 at 1.23 V vs. RHE (Fig. 2b). Since co-catalysts are generally required for water oxidation7, the photoanodic current may (partly) originate from the oxidation of the PEI layer at the interface between BiVO4 and TiO2 layer22,23. Further details on the origin of the photocurrent in BiVO4/PEI/TiO2 will be discussed in a subsequent section.

We also prepared co-catalyst-modified photoanodes (BiVO4/CoOOH and BiVO4/PEI/TiO2/CoOOH) to facilitate efficient water oxidation (Supplementary Fig. 7 and 8). The CoOOH was selected due to its superior catalytic activity compared to FeOOH and NiOOH deposited via the immersion process (Supplementary Fig. 9). The BiVO4/PEI/TiO2/CoOOH photoanode exhibited a notable increase in photocurrent of 2.03 mA cm−2 (at 1.23 V vs. RHE), along with an onset potential of 0.28 V vs. RHE, estimated by extrapolating the linear rising region of the photocurrent to the x-axis (Fig. 2c). The applied bias photon-to-current efficiency (ABPE) showed maximum efficiencies of 0.15%, 0.61%, and 0.73% for the BiVO4, BiVO4/CoOOH, and BiVO4/PEI/TiO2/CoOOH, respectively (Supplementary Fig. 10). We also evaluated the photoconversion efficiency of BiVO4, BiVO4/PEI/TiO2, and BiVO4/PEI/TiO2/CoOOH electrodes through the incident photon-to-current efficiency (IPCE) measurement (Supplementary Fig. 11 and 12). In the measurement, the photoanodes with the overlayers exhibited much higher conversion efficiencies at longer wavelengths near the band gap.

Subsequently, we evaluated the stability of the photoanodes by chronoamperometry (CA) in PEC water oxidation. While the bare BiVO4 and BiVO4/CoOOH photoanodes rapidly degraded within 5 h, the BiVO4/PEI/TiO2/CoOOH maintained its PEC activity for an extended duration over 120 h, achieving 96% Faradaic efficiency for oxygen evolution as measured by gas chromatography (GC) (Fig. 2d and Supplementary Figs. 13 and 14). The SEM measurement revealed negligible changes in morphology compared to the photoanode before the stability test (Supplementary Fig. 15), and high-resolution XPS spectra showed a clear Ti 2p peak in BiVO4/PEI/TiO2/CoOOH even after the 120 h stability test (Supplementary Fig. 16). The stability of BiVO4/PEI/TiO2/CoOOH was also enhanced even under alkaline conditions (pH 11), in which BiVO4 is inherently unstable (Supplementary Fig. 17). Moreover, inductively coupled plasma-optical emission spectroscopy (ICP-OES) measurement demonstrated a negligible dissolution of Bi and V from BiVO4 to the electrolyte after the stability test of BiVO4/PEI/TiO2/CoOOH (Fig. 2e). The PEI/TiO2 layer therefore offers corrosion protection in addition to conductivity for holes (Table S1).

Next, we determined the elemental composition of the photoanodes before and after deposition of TiO2 and PEI layer by X-ray photoelectron spectroscopy (XPS) analysis (Fig. 3a). The peaks of Bi (4f, 4d, and 4p), V (2p), and O (1s) were observed in the bare BiVO4. The Bi and V peaks disappeared upon TiO2 deposition, and additional peaks of Ti (2p), N (1s), and C (1s) appeared, commonly originating from the TiO2 precursor tetrakis(dimethyl-amido)titanium (TDMAT)7. The disappearance of the Bi, V, and O peaks is attributed to the formation of a uniform and thick TiO2 protective layer on top of the BiVO4 surface. BiVO4/PEI/TiO2 also indicated a similar trend with BiVO4/TiO2 but showed negatively shifted C peaks at 285.6 eV and 287.4 eV, corresponding to C-NH2 and C-NHR, respectively, along with stronger N peaks (Supplementary Figs. 18 and 19)24,25. Especially, the high intensity of the N peak is presumed to originate from PEI and suggests the incorporation of PEI within the amorphous TiO2 layer.

a–c XPS (a), TOF-SIMS (b), and XRD (c) analysis of BiVO4, BiVO4/TiO2, and BiVO4/PEI/TiO2.

To further investigate the composition and structure of TiO2, we carried out time-of-flight secondary ion mass spectrometry (TOF-SIMS) and X-ray diffraction (XRD) analyses. In the TOF-SIMS measurement, the elemental composition of C, N, and Ti within the TiO2 layer on both BiVO4/TiO2 and BiVO4/PEI/TiO2 photoelectrodes was confirmed. Although all elements are present in both samples, the C and N signals are much higher in the BiVO4/PEI/TiO2 (Fig. 3b and Supplementary Fig. 20). Especially, BiVO4/PEI/TiO2 exhibited a significantly higher intensity of C and N near the TiO2 surface (Supplementary Fig. 20a–b), which is consistent with the XPS analysis. XRD analysis exhibited negligible changes in the phases of BiVO4 and an amorphous TiO2 with no diffraction peaks regardless of the presence of PEI (Fig. 3c). These results reveal that the PEI interfacial layer does not affect the amorphous nature of the TiO2 but does contribute to an increase in the amount of embedded C and N, altering the intrinsic properties of the a-TiO2.

The cross-sectional images and configurations of the photoelectrodes were investigated using high-resolution transmission electron microscopy (HRTEM) and energy dispersive X-ray (EDX) spectroscopy to analyze the configurations of the hybrid PEI/TiO2 layer. In the HRTEM measurement, the BiVO4/TiO2 photoanode showed a conformally deposited 100 nm TiO2 layer that extended deep into the BiVO4 pores (Fig. 4a), and EDX analysis also clearly revealed a uniform distribution of Ti and O throughout the entire BiVO4 surface (Fig. 4b–d and Supplementary Fig. 21a–b). However, we observed the formation of a 125 nm thick TiO2 layer in BiVO4/PEI/TiO2 and the presence of an interfacial PEI layer with a thickness ranging from 1 to 12 nm between BiVO4 and TiO2 (Fig. 4e and Supplementary Fig. 21c–d), and EDX analysis revealed the absence of Ti and O within the BiVO4 structure (Fig. 4f–h). This result indicates that the ALD precursor TDMAT could penetrate the PEI layer some tens of nanometers (to within a nanometer of the interface with BiVO4), resulting in a thicker ALD TiO2 layer containing embedded PEI polymer. As demonstrated in molecular layer deposition techniques, it is presumed that the TDMAT precursor chemisorbed onto the amine groups of the PEI polyelectrolyte26,27.

a–d Cross-sectional image of BiVO4/TiO2 (a) and elemental mapping of Bi, Ti, and O (b–d), respectively. e–h Cross-sectional image of BiVO4/PEI/TiO2 (e) and the corresponding mapping elements (f–h). White arrows indicate the interfacial PEI layer between BiVO4 and TiO2.

We next investigated the PEC performance of BiVO4/PEI/TiO2 as a function of the thickness of the PEI layer. We prepared PEI layers with different thicknesses on the BiVO4 surface by adjusting the concentration of PEI solution (1 wt.%, 2 wt.%, and 4 wt.%), followed by the deposition of a 100 nm TiO2 layer on top of them. The thicknesses of PEI layer also estimated using a profilometer on a FTO substrate were 22.1 nm (1 wt.%), 35.5 nm (2 wt.%) (Supplementary Fig. 22), and 50.5 nm (4 wt.%) (Supplementary Fig. 5). The thickness of PEI is likely different on BiVO4 surfaces, but it could be nevertheless tuned by adjusting the concentration of the PEI in solution. LSV measurements showed variations in PEC performance depending on the thickness of the PEI layer. The photoelectrode with a thicker PEI layer exhibited a higher photocurrent compared to the one with a thinner PEI layer (Supplementary Fig. 23). In addition, TOF-SIMS measurement indicated lower concentrations of C and N in TiO2 with a thinner PEI layer (Supplementary Fig. 24). These results imply that the amount of embedded PEI may influence the properties of the hybrid PEI/TiO2, thereby contributing the PEC performance. As previously mentioned, we considered the possibility that the photoanodic current of BiVO4/PEI/TiO2 originated from the oxidation of the thin interfacial PEI layer. However, the fact that only certain thicknesses of PEI coatings give photocurrent suggests that the hybrid PEI/TiO2 layer has sufficient catalytic activity to facilitate water oxidation without a co-catalyst, although the stability is limited.

To confirm that the phenomenon is generalizable to other materials, we extended our investigation to Fe2O3, a well-studied semiconductor photoanode known for having slow hole transfer. We deposited a 100 nm a-TiO2 layer onto an Fe2O3 photoanode with and without a PEI layer using the same method explained before. The XPS and SEM measurements showed the same trend compared to BiVO4 (Supplementary Fig. 25 and 26), and the Fe2O3/TiO2 also exhibited a very low photocurrent density in the microampere scale (0.0018 mA cm−2 at 1.6 V vs. RHE) similar to that of BiVO4/TiO2 (Supplementary Fig. 27a–b). However, we observed improved PEC performance and the dramatic shift in onset potential with a thicker PEI interfacial layer under the TiO2 protective layer (Supplementary Fig. 27b–c), which indicates that the thickness of PEI significantly influences PEC performance of the hybrid PEI/TiO2 layer.

It was assumed that the presence of PEI moiety in hybrid PEI/TiO2 could influence the intrinsic properties of amorphous TiO2. To elucidate the intrinsic properties of the hybrid PEI/TiO2, we conducted electron energy-loss spectroscopy (EELS) for BiVO4/TiO2 and BiVO4/PEI/TiO2 photoanode. This analysis aimed to characterize the defect band of a-TiO2, which is crucial for facilitating hole transfer from light absorbers to drive water oxidation. In EELS analysis, Cs-corrected TEM was used with a line scan technique for depth profiling analysis to investigate the oxidation state of Ti and O within both the amorphous TiO2 and the hybrid PEI/TiO2 layer (Fig. 5a, b and Supplementary Fig. 28 and 29). EELS analysis of BiVO4/TiO2 showed two primary peaks in the Ti L-edge (459.2 and 464.5 eV) and O K-edge spectrum (531.4 and 542.65 eV), and we verified that the position of each peak remained consistent regardless of the depth (Fig. 5c and Supplementary Fig. 30 and 31). However, in the case of BiVO4/PEI/TiO2, while there were no depth-dependent peak variations, we observed an overall shift of Ti peaks towards the low-energy region (Fig. 5d and Supplementary Fig. 32 and Table S2 and S3). The low-energy shift of Ti is known to occur due to the reduction of Ti4+ in TiO2, which suggests that the hybrid PEI/TiO2 exhibits a higher Ti3+ population with an excess electron compared to normal a-TiO228,29. The reduction of Ti was also confirmed through high resolution XPS analysis and Kelvin probe force microscopy (KPFM) measurements for both normal TiO2 and hybrid PEI/TiO2. The peak deconvolution of Ti 2p1/2 and 2p3/2 in both electrodes showed a slightly higher proportion of Ti3+ in the hybrid PEI/TiO2 (Supplementary Fig. 33)30,31. The KPFM measurements revealed work functions of 4.82 eV, 4.57 eV, and 4.41 eV for the BiVO4, BiVO4/TiO2, and BiVO4/PEI/TiO2 photoanodes, respectively (Supplementary Fig. 34). The lower work function value of hybrid PEI/TiO2 indicates the reduction of TiO2, which coincides with the results obtained from EELS analysis.

a, b Cs-corrected TEM images of BiVO4/TiO2 and BiVO4/PEI/TiO2 with the probing path of line scan. c, d The EELS spectra of the Ti L edge obtained from the certain region of BiVO4/TiO2 (c) and BiVO4/PEI/TiO2 (d).

We assumed that the increase of Ti3+ state could affect the electronic structure of the a-TiO2 state. Therefore, valence state XPS measurement was carried out to investigate the distinct structure of hybrid PEI/TiO2 compared to normal a-TiO2. BiVO4/TiO2 exhibited a valence band composed of O 2p orbitals, and a defect state was observed below 1.5 eV from the Fermi level, with a width of 0.75 eV (Supplementary Fig. 35a–b). These values coincide with the valence band and leaky state reported in previous studies7,32. However, BiVO4/PEI/TiO2 showed a widened defect which was twice as broad as that in normal a-TiO2 (Supplementary Fig. 35c–d). We conclude that the wide defect state is due to a reduction of Ti4+ by incorporated nitrogen. The long absorption tail observed in UV-Vis spectroscopy also supports the presence of the reduced Ti3+ state, as suggested in previous reports (Supplementary Fig. 36)30,33,34. Based on these results, we propose that the partial reduction of Ti4+ state in hybrid PEI/TiO2 widens the defect band, thereby enhancing conductivity through increased density of states (DOS) in TiO2.

In previous studies, PEI polyelectrolyte has been reported to lower the work function of semiconductors15, and it could be reasoned that a reduced work function of BiVO4 contributes to the formation of a hole-selective contact by enabling favorable band bending. However, the KPFM measurement revealed that there is a negligible difference in the work function of the BiVO4 photoelectrodes regardless of the thickness of the PEI layer (Supplementary Fig. 37). This result suggests that an alternative mechanism may be operative within our system.

We therefore investigated the carrier dynamics of the BiVO4 photoanode and hybrid PEI/TiO2 through dual-working electrode (DWE) analysis to evaluate hole transfer efficiency. We carried out operando open-circuit potential (OCP) measurement with the DWE to determine the energetics of the majority carriers of BiVO4 and TiO2 under dark and light conditions. In the measurement, BiVO4/TiO2 showed identical Fermi levels for BiVO4 and TiO2 under dark conditions, indicating Fermi level equilibration between the two semiconductor materials (Fig. 6a)19,35,36. We also observed a slight shift in the Fermi level of BiVO4 under illumination due to the accumulation of photogenerated electrons, and the Fermi level of BiVO4 decreased and returned to equilibrium with TiO2 after turning off the light. However, the photoelectrodes modified with the PEI interfacial layer revealed a significant difference in Fermi level equilibrium compared to BiVO4/TiO2. The Fermi level of BiVO4 did not equilibrate with that of TiO2, indicating that the insulating interfacial PEI layer hinders electron exchange (majority carriers) between the two semiconductors, and also exhibited a more significant increase than BiVO4/TiO2 under light conditions (Fig. 6b, c). Furthermore, we observed a prolonged decay of the Fermi level energy of BiVO4/PEI/TiO2 after turning off the light in DWE and KPFM measurements (Fig. 6b and Supplementary Fig. 38), again indicating the electron-blocking nature of the interfacial PEI layer. The tunneling efficiency of electrons (or holes) is a function of the barrier height37,38, and in our system the electrons in the conduction band of BiVO4 experience a larger barrier height than the holes in the valence band due to the energetic position of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) levels of the thin interfacial PEI layer39,40. We therefore achieve selective hole transfer via tunneling across the thin interfacial PEI layer.

a–c OCP potential of BiVO4/TiO2 (a), BiVO4/PEI/TiO2 (b), and BiVO4/PEI/TiO2/CoOOH (c) detected by DWE under dark and light conditions. The schematic illustrations above each graph depict the photoanode configurations used for DWE analysis.

In the DWE and KPFM analyses, although we confirmed that selective hole transfer is facilitated by the large electron barrier height of the interfacial PEI layer, we were concerned the oxidation of PEI by photogenerated holes contributed by the small hole barrier height. For instance, a sulfite oxidation measurement revealed that BiVO4/PEI/TiO2 shows a transient in the photocurrent even for the expected fast reaction kinetics of the sacrificial electron donor, while it disappeared in the presence of a CoOOH co-catalyst (Supplementary Fig. 39). This result implies either low catalytic activity or slow hole extraction of the hybrid PEI/TiO2, leading to the oxidation of the interfacial PEI layer during water oxidation, as we mentioned earlier. Therefore, we first conducted electrochemical impedance spectroscopy (EIS) for a comparative analysis between the BiVO4/PEI/TiO2 and BiVO4/PEI/TiO2/CoOOH electrodes to evaluate the catalytic activity of the hybrid PEI/TiO2 in water oxidation (Supplementary Fig. 40). The Nyquist plot was fitted by employing a series 3RC-equivalent circuit to determine the resistance and capacitance values in each frequency domain (Supplementary Fig. 41 and 42)19,41. BiVO4/PEI/TiO2 and BiVO4/PEI/TiO2/CoOOH exhibited nearly identical resistance and capacitance values in the high (HF) and medium-frequencies (MF), but we observed a significant difference of resistance in the low-frequency (LF) related to the charge transfer resistance into the electrolyte and therefore the water oxidation kinetics. While BiVO4/PEI/TiO2 maintained a higher resistance even at high applied potential, the resistance of BiVO4/PEI/TiO2/CoOOH decreased steeply after the onset potential (0.28 V vs. RHE) (Supplementary Fig. 40a–b). This result shows that BiVO4/PEI/TiO2 has relatively low catalytic activity, emphasizing the need for an effective cocatalyst on the hybrid PEI/TiO2 layer to achieve stable solar water oxidation. Our findings were confirmed in the stability test of BiVO4/PEI/TiO2 without co-catalyst. In the CA measurement, we observed that the current density of BiVO4/PEI/TiO2 reached nearly zero after 8 h (Supplementary Fig. 43), despite the surface morphology of the electrode remaining intact (Supplementary Fig. 44). In addition, XPS depth profiling measurements of the hybrid PEI/TiO2 revealed the presence of nitrogen species within the TiO2 structure (Supplementary Fig. 45). This result points to the fact that the primary cause is the decomposition of the interfacial PEI layer. The decomposition is likely due to slow hole extraction influenced by the absence of co-catalyst, as BiVO4/PEI/TiO2 demonstrated long-term stability with CoOOH. The slow hole extraction leads to changes or degradation of PEI at the interface between BiVO4 and TiO2, hindering selective hole transfer in the BiVO4/PEI/TiO2. Consequently, future studies should focus on incorporating more effective co-catalysts or replacing PEI with more stable alternatives to further enhance the hole extraction within the interfacial PEI layer.

Based on our characterization and analysis, we elucidate the underlying mechanism of the selective hole transfer mediated by the interfacial PEI and the hybrid PEI/TiO2. In the absence of the PEI interfacial layer, an unfavorable band alignment between the metal oxide and a-TiO2 leads to sluggish hole transfer, rendering PEC water oxidation unfeasible (Fig. 7a). The introduction of the interfacial PEI layer and the hybrid PEI/TiO2 not only mitigates this energetic misalignment but also promotes hole-selective transfer, facilitated by the electron-blocking nature of the PEI and the introduction of new defect states within the hybrid PEI/TiO2 layer (Fig. 7b and Supplementary Fig 46).

a, b The band energetics for metal oxide/TiO2 (a) and metal oxide/PEI/TiO2 (b). The band positions of the metal oxide are based on a BiVO4 photoanode. The Fermi level of semiconductors and defect band position were obtained by DWE, KPFM, and valence state XPS analysis.

In summary, we investigated hybrid PEI/TiO2 as a protection layer for metal oxide photoanodes that significantly enhances the stability of these materials for solar water oxidation. The hybrid TiO2 layer was formed on the interfacial PEI layer during the ALD process, and we observed distinct energetics associated with a new defect state resulting from the partial reduction of Ti4+ in TiO2. Furthermore, the unique properties of the PEI layer (e.g., preventing unfavorable band bending) facilitated selective-hole transfer between the metal oxide and TiO2 layer. Addressing the challenges related to unfavorable hole-selective contact, the metal oxide photoanode modified with the hybrid TiO2 demonstrated feasibility in solar water oxidation with long-term stability.

Tin(II) chloride dihydrate (SnCl2 · 2H2O), isopropyl alcohol, bismuth(III) nitrate pentahydrate (Bi(NO3)3 · 5H2O), vanadyl acetylacetonate (VO(acac)2), methanol, iron(III) chloride hexahydrate (FeCl3 · 6H2O), sodium nitrate (NaNO3), polyethyleneimine (branched), hydrochloric acid (HCl), Tetrakis(dimethylamino)titanium (TDMAT), iron(II) sulfate heptahydrate (FeSO4 · 7H2O), nickel(II) nitrate hexahydrate (Ni(NO3)2 · 6H2O), cobalt(II) nitrate hexahydrate (Co(NO3)2 · 6H2O), sodium hydroxide (NaOH), and potassium phosphate monobasic were purchased from Sigma-Aldrich. All chemicals were used as received without further purification.

BiVO4 was fabricated on a fluorine-doped tin oxide (FTO) substrate with SnO2 hole blocking layer42. First, 0.1 M of SnCl2 · 2H2O (0.226 g, 1.00 mmol) was dissolved in 10 mL of isopropyl alcohol with stirring for 1 h and kept for 1 day under ambient conditions. A SnO2 layer was then spin-coated onto a cleaned FTO substrate at 3000 rpm for 20 s, followed by annealing at 500 °C for 1 h under ambient air conditions. BiVO4 was synthesized on the SnO2-coated FTO substrate using a metal-organic deposition method. A 0.5 M of Bi(NO3)3 · 5H2O (1.213 g, 2.50 mmol) in 4.8 mL of acetic acid and a 0.27 M of VO(acac)2 (0.357 g, 1.35 mmol) in 5 mL of methanol (aged for 3 days) were prepared, respectively. After dissolving each precursor in solvents, the V solution was added to the Bi solution with 1.1 Bi/V molar ratio (e.g. 2.961 mL of Bi solution was mixed with 5 mL of V solution). The mixed precursor solution was spin-coated on the substrate at 2000 rpm for 20 s, and the annealing process was carried out at 480 °C for 20 min under air condition. The spin-coating process was repeated to obtain the desired thickness of BiVO4.

Sn-doped Fe2O3 was synthesized by hydrothermal method on cleaned FTO substrate43. Briefly, 20 mL of precursor solution (D.I) containing 0.15 M FeCl3 · 6H2O (0.8109 g, 3.00 mmol) and 1.0 M NaNO3 (1.70 g, 20.0 mmol) was transferred to Teflon-lined stainless-steel autoclave (50 mL of volume). FTO was placed at the bottom of the autoclave with conductive side facing up, and heat-treatment was conducted at 95 °C for 4 h to form β-FeOOH nanowires on the FTO substrate. After heating process, β-FeOOH was sintered in ambient air condition at 550 °C for 2 h (ramping rate of 4 K min−1) and then annealed at 800 °C for 20 min.

Polyethyleneimine (PEI) layer was deposited by spin-coating method on metal oxide photoanodes. PEI (Mw ~25,000) was dissolved in D.I as various concentration (wt.%) to control a thickness of PEI layer (e.g. to prepare a 4% solution, 0.2 g of PEI was dissolved in 4 ml of D.I), and pH of PEI solution was adjusted as 4.2 by 1 mL of 3 M HCl solution. The PEI solution was spin-coated on photoanodes at 4000 rpm for 20 s, followed by drying in oven at 70 °C for 30 min. The concentration and the spin-coating rate could be modulated to adjust the thickness of PEI layer.

Amorphous TiO2 was deposited on the photoanodes by atomic layer deposition (ALD) technique (R200, Picosun). Tetrakis(dimethylamino)titanium (TDMAT, 99.99%, Aldrich) and H2O were used as precursors for Ti and O, respectively. Each photoanode was placed in ALD chamber at 120 °C, TDMAT heated at 85 °C was put with a 1.6s pulse, followed by N2 purge with a 6.0 s. H2O was kept at room temperature with a 0.1 s pulse, followed by a 6.0 s N2 purge. The thickness of TiO2 layer was confirmed with Si substrate using alpha-SE ellipsometer (J.A. Woollam Co.), and 0.55 Å of TiO2 layer was formed per TDMAT-H2O cycle, approximately.

For deposition of FeOOH, NiOOH, and CoOOH co-catalysts, 10 mM of iron(II) sulfate heptahydrate, nickel(II) nitrate hexahydrate, or cobalt(II) nitrate hexahydrate were dissolved in D.I water. The pH was adjusted to 4.5 for the iron solution and 7.3 ~ 7.4 for the nickel and cobalt solution using 0.1 M NaOH. The photoelectrodes are soaked in the precursor solution for 3.5 h and then washed by D.I and dried by N2 gun, gently.

For a BiVO4/TiO2 photoanode, 5 nm of metal layers was deposited onto the photoanode surface by Leica EM ACE600 magnetron sputter. Ni (100 mA working current and 2.0 E−2 mbar), Pt (35 mA working current and 5.0 E−2 mbar), and Au (30 mA working current and 5.0 E−2 mbar) were deposited under certain conditions.

Photoelectrochemical (PEC) characterizations were carried out by SP-200 Bio-Logic potentiostat in a three‐electrode configuration under AM 1.5G illumination. the photoanode, Pt wire, and Ag/AgCl electrode were used as working, counter, and reference electrode, respectively. For the measurement of PEC performance, 0.5 M potassium phosphate (KPi) was used as the electrolyte under back-side (BiVO4) and front-side (Fe2O3) illumination, and epoxy resin (Loctite Epoxide-resin EA 9461 and EA 9466) was used to obtain certain surface area of the photoanodes (0.2 ~ 0.25 cm−2). IPCE measurement was conducted using a home-built double monochromator with a halogen light source. The light intensity of the monochromator was calibrated using a Si diode. Electrochemical impedance spectroscopy (EIS) was measured by SP‐300 with 10 mV AC voltage amplitude and frequency range from 0.2 Hz to 1 MHz under the illumination of white light LED (SP-12-W5, cool white Luxeon Rebel). Numerical fitting of EIS data was conducted by Zview software.

A surface and cross-sectional morphology were analyzed with a Hitachi SU-7000 field-emission scanning electron microscope (SEM) and Zeiss Gemini 450 SEM. The Bi and V contents in electrolytes after stability test were evaluated with a Varian inductively coupled plasma-optical emission spectrometer (ICP-OES). High-resolution X-ray photoelectron spectroscopy (XPS) spectra of the photoanodes was obtained by Thermo Fisher K-Alpha XPS instrument, and valence state XPS analysis to confirm leaky state was carried out with Physical Electronics (PHI) Quantum 2000 X-ray photoelectron spectrometer featuring monochromatic Al-Kα radiation, generated from an electron beam operated at 15 kV and 35 W. The energy scale linearity of the instrument was established through calibration with a reference sample of Au. Time-of-flight secondary ion mass spectrometry (TOF-SIMS) and X-ray crystallography (XRD) were conducted to investigate the depth-profiling and crystallinity of photoanodes through IONTOF TOF-SIMS-5 and Rigaku Smartlab diffractometer, respectively. The elemental composition in the cross-sectional direction was obtained by high-resolution transmission electron microscopy (HRTEM) with a JEOL JEM 2010 transmission electron microscope. Electron energy-loss spectroscopy (EELS) was used to compare oxidation state, and the measurement was carried out using a Cs-corrected TEM with JEOL JEM-ARM300 transmission electron microscope to ensure high-resolution. HAADF-STEM and EELS resolution are 0.058 nm and 0.3 eV with 300 kV acceleration voltage, respectively. The samples for TEM measurement were prepared by focused ion beam (FIB) milling. The absorbance measurement was carried out by UV-Vis spectroscopy with CRAIC 20/20/PV UV-Vis microspectrometer.

Kelvin probe force microscopy (KPFM) was carried out to confirm work function (WF) of each photoanode. An Asylum Research AFM (MFP-3D) was used to measure the work function of the samples. The probe used for the measurement was a AC240TM-R3. For calibration of the work function of the tip a highly ordered pyrolytic graphite (HOPG) has been used which has a reported work function of ~ 4.6 eV44. To achieve a fresh HOPG surface a piece of scotch tape was used to pull off a few top layers of the graphite and exposing a fresh clean surface for the calibration. The HOPG used was purchased from MikroMasch (Grade: ZYA). The fresh HOPG surface changes its work function in a time window of several tens of minutes when exposed to air. Therefore, the HOPG was measured against an Aluminium metal mirror with native Al2O3 layer on the surface. The work function of the Al/Al2O3 (~3.90 eV) was stable for several hours and even days45. Open-source Gwyddion software package as well as the Asylum Research build in software were used to further analyze the AFM pictures and determine the average work function of the surface.

Initially, a 10 nm-thick permeable Au layer was deposited using sputtering onto each photoanode surface, which had been covered with epoxy resin (Epoxide-resin EA 9461, Loctite). To establish an additional contact for surface potential measurement of the TiO2 layer, a front contact was fabricated on the Au layer of the epoxy surface. The front contact was connected with a Cu foil and the Au layer using Ag paste and sealed by additional epoxy to protect the contact from the electrolyte. Open circuit potential (OCP) measurement was carried out using a dual-working electrode (DWE) through Bio-Logic SP‐300 potentiostat with V2-controlled PEIS measurements. A first and second working electrode (WE1 and WE2) were connected to the BiVO4 back contact and the TiO2 front contact, allowing to monitor changes in Fermi level and quasi electron Fermi level under dark and light illumination.

The authors declare that the data supporting the findings of this study are available within the paper and its supplementary information files, including the source data file. Source data are provided with this paper.

Sathre, R. et al. Opportunities to improve the net energy performance of photoelectrochemical water-splitting technology. Energy Environ. Sci. 9, 803–819 (2016).

Article CAS Google Scholar

Vilanova, A. et al. The route for commercial photoelectrochemical water splitting: a review of large-area devices and key upscaling challenges. Chem. Soc. Rev. 53, 2388–2434 (2024).

Article CAS PubMed Google Scholar

Schleuning, M. et al. The role of selective contacts and built-in field for charge separation and transport in photoelectrochemical devices. Sustain. Energy Fuels 6, 3701–3716 (2022).

Article CAS Google Scholar

Wick-Joliat, R. et al. Stable and tunable phosphonic acid dipole layer for band edge engineering of photoelectrochemical and photovoltaic heterojunction devices. Energy Environ. Sci. 12, 1901–1909 (2019).

Article CAS Google Scholar

Paracchino, A. et al. Highly active oxide photocathode for photoelectrochemical water reduction. Nat. Mater. 10, 456–461 (2011).

Article ADS CAS PubMed Google Scholar

Mei, B. et al. Crystalline TiO2: a generic and effective electron-conducting protection layer for photoanodes and -cathodes. J. Phys. Chem. C. 119, 15019–15027 (2015).

Article CAS Google Scholar

Hu, S. et al. Amorphous TiO2 coatings stabilize Si, GaAs, and GaP photoanodes for efficient water oxidation. Science 344, 1005–1009 (2014).

Article ADS CAS PubMed Google Scholar

Corby, S. et al. The kinetics of metal oxide photoanodes from charge generation to catalysis. Nat. Rev. Mater. 6, 1136–1155 (2021).

Article ADS CAS Google Scholar

Scheuermann, A. G. et al. Effects of catalyst material and atomic layer deposited TiO2 oxide thickness on the water oxidation performance of metal–insulator–silicon anodes. Energy Environ. Sci. 6, 2487–2496 (2013).

Article CAS Google Scholar

Scheuermann, A. G. et al. Design principles for maximizing photovoltage in metal-oxide-protected water-splitting photoanodes. Nat. Mater. 15, 99–105 (2016).

Article ADS CAS PubMed Google Scholar

Lee, D. K. et al. Enhancing long-term photostability of BiVO4 photoanodes for solar water splitting by tuning electrolyte composition. Nat. Energy 3, 53–60 (2017).

Article ADS CAS Google Scholar

Zhang, S. et al. Dissolution of BiVO4 photoanodes revealed by time-resolved measurements under photoelectrochemical conditions. J. Phys. Chem. C. 123, 23410–23418 (2019).

Article CAS Google Scholar

McDowell, M. T. et al. Improved stability of polycrystalline bismuth vanadate photoanodes by use of dual-layer thin TiO2/Ni coatings. J. Phys. Chem. C. 118, 19618–19624 (2014).

Article CAS Google Scholar

Moehl, T. et al. Investigation of (leaky) ALD TiO2 protection layers for water-splitting photoelectrodes. ACS Appl. Mater. Interfaces 9, 43614–43622 (2017).

Article CAS PubMed Google Scholar

Zhou, Y. et al. A Universal method to produce low–work function electrodes for organic electronics. Science 336, 327–332 (2012).

Article ADS CAS PubMed Google Scholar

Kim, Y. H. et al. Polyethylene imine as an ideal interlayer for highly efficient inverted polymer light‐emitting diodes. Adv. Funct. Mater. 24, 3808–3814 (2014).

Article CAS Google Scholar

Huang, W. et al. Metal composition and polyethylenimine doping capacity effects on semiconducting metal oxide-polymer blend charge transport. J. Am. Chem. Soc. 140, 5457–5473 (2018).

Article CAS PubMed Google Scholar

Wei, Z.-Q. et al. Branched polymer-incorporated multi-layered heterostructured photoanode: precisely tuning directional charge transfer toward solar water oxidation. J. Mater. Chem. A 8, 177–189 (2020).

Article CAS Google Scholar

Cui, W. et al. Operandoelectrochemical study of charge carrier processes in water splitting photoanodes protected by atomic layer deposited TiO2. Sustain. Energy Fuels 3, 3085–3092 (2019).

Article CAS Google Scholar

Digdaya, I. A. et al. Interfacial engineering of metal-insulator-semiconductor junctions for efficient and stable photoelectrochemical water oxidation. Nat. Commun. 8, 15968 (2017).

Article ADS CAS PubMed PubMed Central Google Scholar

Nunez, P. et al. Characterization of electronic transport through amorphous TiO2 produced by atomic layer deposition. J. Phys. Chem. C. 123, 20116–20129 (2019).

Article CAS Google Scholar

Min, K. et al. Oxidation-stable amine-containing adsorbents for carbon dioxide capture. Nat. Commun. 9, 726 (2018).

Article ADS PubMed PubMed Central Google Scholar

Bender, M. T. et al. Electrochemical dehydrogenation pathways of amines to nitriles on NiOOH. JACS Au 2, 1169–1180 (2022).

Article CAS PubMed PubMed Central Google Scholar

Zhu, H. et al. Insitu growth of Rh nanoparticles with controlled sizes and dispersions on the cross-linked PVA–PEI nanofibers and their electrocatalytic properties towards H2O2. RSC Adv. 4, 794–804 (2014).

Article ADS CAS Google Scholar

Simoes Dos Reis, G. et al. Facile synthesis of sustainable activated biochars with different pore structures as efficient additive-carbon-free anodes for lithium- and sodium-ion batteries. ACS Omega 7, 42570–42581 (2022).

Article CAS PubMed PubMed Central Google Scholar

Zhao, Y. et al. Molecular layer deposition for energy conversion and storage. ACS Energy Lett. 3, 899–914 (2018).

Article ADS CAS Google Scholar

Zhao, Y. et al. Atomic/molecular layer deposition for energy storage and conversion. Chem. Soc. Rev. 50, 3889–3956 (2021).

Article CAS PubMed Google Scholar

Gao, Q. et al. Direct evidence of lithium-induced atomic ordering in amorphous TiO2 nanotubes. Chem. Mater. 26, 1660–1669 (2014).

Article CAS Google Scholar

Tian, M. et al. Structure and formation mechanism of black TiO2 nanoparticles. ACS Nano 9, 10482–10488 (2015).

Article CAS PubMed Google Scholar

Oh, S. et al. Band restructuring of ordered/disordered blue TiO2 for visible light photocatalysis. J. Mater. Chem. A 9, 4822–4830 (2021).

Article CAS Google Scholar

Dong, P. et al. Solar-driven photocatalytic removal of NO over a concrete paving eco-block containing black TiO2. J. Mater. Chem. A 11, 25429–25440 (2023).

Article CAS Google Scholar

Shen, X. et al. Tuning intermediate bands of protective coatings to reach the bulk‐recombination limit of stable water‐oxidation GaP photoanodes. Adv. Energy Mater. 12, 2201314 (2022).

Article CAS Google Scholar

Chen, X. et al. Increasing solar absorption for photocatalysis with black hydrogenated titanium dioxide nanocrystals. Science 331, 746–750 (2011).

Article ADS CAS PubMed Google Scholar

Zhu, Q. et al. Stable blue TiO2−x nanoparticles for efficient visible light photocatalysts. J. Mater. Chem. A 2, 4429–4437 (2014).

Article CAS Google Scholar

Lin, F. et al. Adaptive semiconductor/electrocatalyst junctions in water-splitting photoanodes. Nat. Mater. 13, 81–86 (2014).

Article ADS CAS PubMed Google Scholar

Nellist, M. R. et al. Semiconductor-electrocatalyst interfaces: theory, experiment, and applications in photoelectrochemical water splitting. Acc. Chem. Res. 49, 733–740 (2016).

Article CAS PubMed Google Scholar

Zhou, X. et al. Tunneling diode based on WSe2 /SnS2 heterostructure incorporating high detectivity and responsivity. Adv. Mater. 30, 1703286 (2018).

Article Google Scholar

Gao, F. et al. High‐performance van der Waals metal‐insulator‐semiconductor photodetector optimized with valence band matching. Adv. Funct. Mater. 31, 2104359 (2021).

Article ADS CAS Google Scholar

Li, P. et al. High-efficiency inverted polymer solar cells controlled by the thickness of polyethylenimine ethoxylated (PEIE) interfacial layers. Phys. Chem. Chem. Phys. 16, 23792–23799 (2014).

Article CAS PubMed Google Scholar

Davidson-Hall, T. et al. The role of polyethylenimine in enhancing the efficiencyof quantum dot light-emitting devices. Nanoscale 10, 2623–2631 (2018).

Article CAS PubMed Google Scholar

Yang, W. et al. Operando analysis of semiconductor junctions in multi‐layered photocathodes for solar water splitting by impedance spectroscopy. Adv. Energy Mater. 11, 2003569 (2021).

Article CAS Google Scholar

Choi, Y. et al. Atomically-dispersed cobalt ions on polyphenol-derived nanocarbon layers to improve charge separation, hole storage, and catalytic activity of water-oxidation photoanodes. J. Mater. Chem. A 9, 13874–13882 (2021).

Article CAS Google Scholar

Ling, Y. et al. Sn-doped hematite nanostructures for photoelectrochemical water splitting. Nano Lett. 11, 2119–2125 (2011).

Article ADS CAS PubMed Google Scholar

Fernandez Garrillo, P. A. et al. Calibrated work function mapping by Kelvin probe force microscopy. Rev. Sci. Instrum. 89, 043702 (2018).

Article ADS PubMed Google Scholar

Eastment, R. M. et al. Work function measurements on (100), (110) and (111) surfaces of aluminium. J. Phys. F: Met. Phys. 3, 1738 (1973).

Article ADS CAS Google Scholar

Download references

This work was supported by the University Research Priority Program LightChEC of the University of Zurich, the Swiss National Science Foundation (Project #184737 and #208451), the Basic Science Research Program (2021R1A2C2013684), and the Regional Leading Research Center (RLRC) (RS-2023-00217778) funded by the National Research Foundation (NRF) of Korea.

Department of Chemistry, University of Zurich, Winterthurerstrasse 190, 8057, Zurich, Switzerland

Sanghyun Bae, Thomas Moehl, Erin Service, Pardis Adams, Zhenbin Wang & S. David Tilley

School of Energy and Chemical Engineering, Ulsan National Institute of Science and Technology (UNIST), Ulsan, 44919, Republic of Korea

Sanghyun Bae, Minjung Kim, Yuri Choi & Jungki Ryu

Center for Renewable Carbon, Ulsan National Institute of Science and Technology (UNIST), Ulsan, 44919, Republic of Korea

Jungki Ryu

You can also search for this author in PubMed Google Scholar

S.B., J.R. and S.D.T. conceived the concept. S.B. performed the materials design, characterization, and photoelectrochemical experiments and analyses. T.M. carried out KPFM analysis. T.M. and E.S. performed EIS analysis. M.K and Y.C. conducted TOF-SIMS measurement. P.A. and Z.W. performed XPS and XRD analysis. S.B., J.R. and S.D.T. wrote the manuscript. All authors discussed the results and participated in writing the manuscript.

Correspondence to Jungki Ryu or S. David Tilley.

The authors declare no competing interests.

Nature Communications thanks Joan Morante and the other, anonymous, reviewers for their contribution to the peer review of this work. A peer review file is available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

Bae, S., Moehl, T., Service, E. et al. A hole-selective hybrid TiO2 layer for stable and low-cost photoanodes in solar water oxidation. Nat Commun 15, 9439 (2024). https://doi.org/10.1038/s41467-024-53754-9

Download citation

Received: 19 March 2024

Accepted: 17 October 2024

Published: 01 November 2024

DOI: https://doi.org/10.1038/s41467-024-53754-9

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative