Simultaneous engineering of cation disorder and morphology of molecular-ink derived AgBiS₂ photocathodes for solar water splitting

Abstract

This study introduced a strategy based on coordination chemistry for the simultaneous control of cation disorder and morphological refinement in ternary chalcogenide AgBiS₂ thin films. Cation disorder, a key parameter influencing optoelectronic properties such as light absorption, was effectively controlled by manipulating metal–ligand interactions within an n-butyl dithiocarbamate (DTC)-based molecular ink system. To further modulate crystallization kinetics, urea was incorporated as a cost-effective and environmentally benign Lewis base additive. Thermodynamic calculations and binding energy analyses revealed that urea preferentially coordinated with Bi³⁺ cations, thereby suppressing premature nucleation and promoting the growth of large and uniform grains without disrupting the DTC–metal coordination framework. This dual-control strategy enabled the fabrication of high-quality AgBiS₂ thin films with enhanced optical absorption and reduced grain boundary recombination, resulting in the first AgBiS₂-based photocathodes for photoelectrochemical water splitting.

---

Introduction

Cation-disorder engineering has been recognized as a versatile and novel approach for tailoring material properties without altering nominal composition or crystal structure. This method exploits the nonlinear relationship between cation arrangement and material performance by controlling the degree of order–disorder transitions in multicomponent systems.^1–3 Cation disorder can be described as the probabilistic distribution of cations over crystallographic sites. In the ordered phase, specific cations predominantly occupy designated sites, whereas in the disordered phase, site occupancy becomes more randomized, with each cation type having an approximately equal probability of occupying a given site.^4,5 Cation disorder engineering has been widely utilized in various applications, such as batteries,^6–8 catalysts^9,10 and photovoltaics.^11–13 However, its impact is not universally beneficial, as the effect depends on the material system and functional objectives. Therefore, the role of cation disorder should be carefully assessed and appropriately adjusted for specific applications. Consequently, increasing research efforts have focused on utilizing cation disorder as a tunable design parameter to optimize material properties tailored to functional demands of various applications.

Among multicomponent chalcogenides, AgBiS₂ has recently emerged as a promising material for optoelectronic applications owing to its environmentally friendly composition, suitable bandgap (about 1.0–1.4 eV), and strong light absorption characteristics (>10⁵ cm⁻¹). AgBiS₂ can be synthesized using various methods, including colloidal nanocrystal (NCs) synthesis,^12–14 molecular ink processing,^15–18 solvothermal synthesis^19,20 and chemical bath deposition,²¹ each offering different levels of control over its microstructure and composition. A particularly intriguing feature of AgBiS₂ is its pronounced sensitivity of optical and electronic properties to cation arrangement within its lattice. For instance, Wang et al. demonstrated that disordered AgBiS₂ nanocrystals exhibit an unusually high absorption coefficient of approximately 8 × 10⁵ cm⁻¹, enabling effective light harvesting even with absorber layers as thin as about 30 nm. This significantly reduces the charge-transport path and minimizes recombination losses, yielding power conversion efficiencies exceeding 9%.¹² Additionally, cation-disordered AgBiS₂ nanocrystals exhibit enhanced charge-carrier transport, attributed to a modified energy landscape and reduced trap density.¹³ In most AgBiS₂ NC systems, cation disorder engineering is achieved by tuning thermal treatment, where elevated annealing temperatures promote the formation of thermodynamically stable and homogeneous disordered phases. These findings highlight the importance of cation-disorder engineering to optimize the functional properties of AgBiS₂ for high-performance device integration.

Recent research on AgBiS₂ based optoelectronic devices, including solar cells, photodetectors, and photoelectrodes, has predominantly utilized colloidal NC synthesis,^12–14 which involves complex processes such as centrifugation and ligand exchange. Although these methods allow for precise control of particle size and composition, they have inherent drawbacks. Surface ligands that stabilize NCs can also act as electrical barriers to charge transport, thereby limiting device performance. Furthermore, because of the large Bohr radius of AgBiS₂ (in the order of tens of nanometers), the advantages of quantum confinement effects are significantly diminished in typical NC sizes, rendering the nanocrystal approach less beneficial. As an alternative, molecular-ink-based processing has emerged as a scalable and fabrication-friendly strategy for the production of large-scale thin films. For instance, Wu et al. demonstrated the facile synthesis of AgBiS₂ nanorods using thiourea- and DMSO-based solutions,¹⁶ whereas Joel van Embden et al. utilized metal-dithiocarbamate (DTC) complexes to fabricate ultrathin AgBiS₂ films with thicknesses below 40 nm.¹⁷ Our group recently reported that thermodynamically favorable cation-disordered AgBiS₂ phases could be achieved by modulating the coordination between metal cations and DTC ligands, yielding photodetectors with superior photocurrent and photoresponse.¹⁸ Despite these advancements, solution-processed AgBiS₂ films often exhibit porous morphologies and small crystal sizes, primarily because of the decomposition and volatilization of precursors and solvents during annealing. These suboptimal film qualities remain a significant barrier to achieving high-efficiency optoelectronic devices, highlighting the necessity for strategies to enhance film quality such as densification and large grain sizes.

To address these morphological limitations, additive engineering strategies involving Lewis base additives have recently gained significant attention as effective methods for controlling film morphology. Lewis base with lone-pair electrons, such as carbonyl (–C=O), amino (–NH₂) groups or cyanate (–O–C≡N) groups, can coordinate with electron-deficient metal centers to form transient complexes with metal cation precursors. This coordination alters solution reactivity and nucleation kinetics while simultaneously slowing crystal growth, thereby promoting the formation of more uniform and compact thin films. In perovskite solar cells, Lewis bases such as DMSO,²² urea,^23,24 thiourea,^24–26 and thiocyanate^27,28 have been widely employed. These additives have been shown to significantly improve film quality by increasing grain size, eliminating pinholes, and reducing trap states, leading to substantial enhancements in efficiency. Although such strategies have been explored in other systems, their applicability to AgBiS₂ has not yet been demonstrated. Furthermore, the interplay between coordination chemistry and cation-disorder engineering remains unexplored.

This study proposes a synthetic strategy for fabricating high-quality AgBiS₂ thin films with dense morphologies by integrating a thermodynamically driven cation-disorder engineering approach based on DTC coordination chemistry and employing urea as a Lewis base additive. This method not only enables the formation of thermodynamically favorable cation-disordered AgBiS₂ phases but also offers an integrated strategy for controlling both cation disorder and film morphology. Although these two parameters are critically influenced by solution-phase coordination chemistry, their interdependence remains largely unexplored. As such, achieving simultaneous modulation through a coordination chemistry-driven strategy is highly challenging, requiring careful consideration of factors such as metal–ligand binding energies, preferential coordination, and ligand exchange dynamics. These films were successfully employed as photocathodes for photoelectrochemical (PEC) water splitting, marking the first demonstration of AgBiS₂-based photocathodes.

Results and discussion

We aimed to achieve the morphological development of AgBiS₂ thin films by employing molecular additives in a coordination chemistry-based approach that modulates cation disorder in the ternary chalcogenide AgBiS₂ using n-butyl DTC ligands. Urea was selected as the molecular additive because of its low cost, environmental benignity, and simple structure, which minimizes the risk of residual impurities. Its high solubility and chemical compatibility with the DTC-based molecular inks also enable uniform coordination and grain growth without disturbing the cation disorder control mechanism. As a Lewis base with electron-rich lone pairs, urea can coordinate with electron-deficient metal ions to modulate nucleation and crystallization kinetics, thereby enhancing crystallinity and promoting the growth of larger grains with fewer boundaries. To assess whether urea can effectively alter the crystallization behavior of AgBiS₂, we evaluated its coordination energetics with Ag⁺ and Bi³⁺. Given that cation disorder in AgBiS₂ is governed by DTC-based coordination chemistry, understanding the competitive binding affinity of urea toward metal cations is critical for elucidating its potential to simultaneously modulate cation arrangement and crystallization dynamics. Fig. 1a shows the numerically evaluated binding energies of Ag⁺ and Bi³⁺ ions with both DTC (the sulfur source) and urea additives as a function of the coordination number. Regardless of the coordinating ligand, the binding energy of Bi³⁺ ions was consistently and significantly lower than that of Ag⁺, suggesting preferential coordination between Bi³⁺ and both DTC and urea. This calculated binding energy is primarily governed by intrinsic factors such as the cation size and valence as well as the structural geometry of the coordinating ligands. Smaller and more highly charged cations form stronger electrostatic interactions with short linear ligands, whereas bulky or aromatic ligands cause steric hindrance that weakens coordination. Binding energy between Bi³⁺ and urea in a monodentate configuration was approximately −0.493 eV per atom, which is lower than that for Bi³⁺ with DTC (−0.410 eV per atom). This indicates that Bi³⁺ forms thermodynamically more favorable complexes with urea than with DTC. Following a previously established approach, cation disorder in AgBiS₂ can be tuned by modulating the concentration of DTC ligands in solution, thereby systematically controlling the degree of coordination between DTC and metal cations (Fig. 1b).

Fig. 1 (a) Calculated binding energy profile of Ag and Bi cations as a function of coordination number. (b) Conventional cation disorder engineering of AgBiS₂ based on coordination chemistry theory. (c) Schematic illustration of the optimized structure coordination between bismuth cations and urea or dithiocarbamate (DTC) ligands under a coordination number of one. (d) Proposed mechanism in Ag–Bi–urea–DTC molecular inks, where the incorporation of urea additives promotes facilitates morphological improvement of the AgBiS₂ film.

Urea coordinates through its electron-rich carbonyl group (C=O), which can donate metal cations and form stronger binding interactions than the CSS⁻ group in DTC. This is supported by the shorter coordination distance between urea and metal ions (∼1.99 Å) compared with that of DTC (∼2.6 Å), as shown in Fig. 1c. Optimized geometries of these metal–ligand complexes, including detailed bonding angles and distances, are provided in Fig. S1–S3 (SI). Urea exhibits a stronger binding affinity for Bi³⁺ ions than for both Ag⁺ or the DTC ligand, forming stable Bi–urea complexes in solution. This preferential coordination effectively reduces the concentration of Bi–DTC complexes available for crystal nucleation, thereby suppressing nucleation rates during the early stages of AgBiS₂ crystallization. By suppressing nucleation rates, urea extends the growth of crystals, facilitating the formation of larger grains. Consequently, the number of nucleation sites decreases while grain size increases. The overall mechanism is illustrated in Fig. 4d, where urea-driven coordination modulation of crystallization kinetics leads to enlarged grain size and improved film morphology, ultimately contributing to high-quality AgBiS₂ films. Because of its simple molecular structure and low molecular weight, urea undergoes thermal decomposition at relatively low temperatures during annealing, allowing nearby DTC ligands to coordinate and initiate the nucleation of AgBiS₂ crystals. Given that the amount of urea used was substantially lower than that of the sulfur source DTC, we theoretically assumed that it would not interfere with cation disorder engineering governed by DTC-based coordination chemistry.

To investigate the influence of urea on AgBiS₂ film morphology, surface characterization was conducted using scanning electron microscopy (SEM). As shown in Fig. 2a, AgBiS₂ films prepared under DTC-sufficient conditions (DTC36) without urea exhibited a highly porous morphology composed of aggregated nanoparticles with an average diameter of approximately 70 nm. In DTCX, X represents the molar ratio of DTC relative to the total concentration of metal cations (Ag⁺ or Bi³⁺) in solution. The porosity likely arises from gas release from the decomposition of DTC and solvent during annealing. Upon addition of 0.2 M urea, the grain size increased to ∼100 nm, and the film morphology became notably denser with reduced porosity (Fig. 2b). A further increase in urea concentration to 0.4 M resulted in larger grains (∼150 nm), although residual surface material (highlighted by red circles in Fig. 2c) and incomplete coverage were observed. These effects are likely attributed to the insufficient volatilization of excessive urea and associated gas formation during annealing. These observations corroborate the hypothesis that urea, acting as a Lewis base, coordinates with metal cations to modulate nucleation kinetics and facilitate grain growth. However, an excessive concentration of urea appears to induce non-uniform morphology, indicating the necessity for precise concentration optimization to achieve high film quality. Motivated by prior reports emphasizing the advantages of ultrathin (<100 nm) absorber layers for improved charge extraction and cost-effective device integration,^12,13 we targeted the fabrication of compact AgBiS₂ films within this dimensional regime. Despite the use of low concentration precursor solutions to fabricate ultrathin AgBiS₂ films, surface SEM images revealed substantial morphological improvement with the addition of 0.1 M urea, as evidenced by an increase in average grain size from 67.9 nm to 92.9 nm, corresponding to a 37% enhancement (Fig. S4).

Fig. 2 SEM surface images of AgBiS₂ films synthesized with varying concentrations of urea additives: (a) without urea, (b) 0.2 M urea, and (c) 0.4 M urea in Ag–Bi–DTC36 molecular inks. The scale bar corresponds to 500 nm.

XRD analysis was performed to assess whether morphological changes were accompanied by changes in crystallinity. As shown in Fig. 3a–c, all samples exhibited clear (111), (200), and (220) major orientations characteristic of cubic AgBiS₂, without any discernible peak shifts or secondary phases, regardless of urea addition. Across all DTC concentrations, samples containing 0.1 M urea exhibited no evidence of secondary phases such as Ag₂S or Bi₂S₃ (Fig. S5), confirming the successful formation of phase-pure cubic AgBiS₂ and effective thermal removal of urea-derived residues during post-annealing. Additionally, the introduction of urea did not interfere with the cation-disorder behavior governed by the existing DTC coordination environment. Under DTC-sufficient (DTC36) conditions, the main orientation peak of AgBiS₂ shifted toward higher angles regardless of the presence of urea, indicating a decrease in lattice constants (Fig. S6). This peak shift can be attributed to the contraction of Ag–S bond lengths in the cation-disordered phase compared to ordered phases.¹² These observations collectively suggest that the small amount of urea introduced, relative to DTC ligand concentration, does not significantly perturb the established coordination chemistry.

Fig. 3 (a–c) High-resolution X-ray diffraction (HRXRD) patterns of AgBiS₂ thin films prepared under DTC36 conditions, with and without urea additives. The patterns highlight crystal orientation peaks corresponding to the (111), (200), and (220) planes, respectively. (d and e) X-ray photoelectron spectroscopy (XPS) spectra of AgBiS₂ thin films illustrating the effects of varying DTC concentrations and the presence or absence of urea additives. The spectra correspond to the Ag 3d (d) and Bi 4f (e) core levels.

Although XRD analysis confirmed that the addition of urea did not influence the long-range crystallinity of AgBiS₂ films, further investigation was conducted to determine its effect on cation disorder and chemical states. The XPS profiles (Fig. 3d and e) indicate that the addition of urea did not significant alter core-level peak shapes. However, under DTC-sufficient conditions, the Ag 3d and Bi 4f peaks shifted toward lower binding energies, consistent with increased cation disorder and the corresponding rise in the Madelung potential reported in previous studies.¹² Notably, under DTC-deficient conditions (DTC6), a shoulder peak emerged at higher binding energies in the Bi 4f spectrum (Fig. 3e), indicating surface Bi–O bonding. Although oxygen was absent in the precursor formulation, this oxidation is believed to occur during post-deposition processing, consistent with previous reports on AgBiS₂ systems.²⁹ Upon addition of urea, the Bi–O shoulder peak intensified, likely because the carbonyl group (C=O) of urea acted as an internal oxygen source. Correspondingly, the intensity of the main Bi 4f peak decreased, further supporting the notion that urea promotes surface oxidation of AgBiS₂ under DTC-deficient conditions. These findings indicate the formation of an amorphous oxidation-induced phase not detectable by XRD spectra and unlikely to correspond to a crystalline Bi–O oxide phase.^21,30,31 Conversely, the S 2s core-level peak displayed a gradual shift to higher binding energy with increasing DTC concentration (Fig. S7), independent of urea addition. This trend is likely due to electron redistribution effects that differ from those observed in the metal cation signals. Collectively, these XRD and XPS results suggest that the use of 0.1 M urea neither compromises the crystallographic impact of AgBiS₂ films nor disrupt the coordination-mediated cation disorder modulation framework, while effectively improving morphological properties.

To complement these findings, we further investigated the optical properties of AgBiS₂ films. As depicted in Fig. 4a, optical photographs reveal that the urea-added films appear darker than those without urea, indicating enhanced optical absorption. Furthermore, the absorption coefficient was significantly higher under DTC36 than under DTC6, regardless of the addition of urea, indicating the crucial role of DTC coordination-based cation disorder engineering. This observation was further supported by UV-vis spectroscopy (Fig. 4b), which showed that DTC36 with a cation-disordered AgBiS₂ phase exhibited a high absorption coefficient exceeding 4 × 10⁵ cm⁻¹ at 350 nm, which increased to approximately 7 × 10⁵ cm⁻¹ upon the addition of urea, approaching the theoretical maximum.^12,32 This optical enhancement can be attributed to improved film morphology, including enhanced grain connectivity and surface uniformity. Under DTC6 conditions, a similar but modest improvement in absorption was observed, also attributed to changes in surface morphology. The optical band gaps of the films were further analyzed using Tauc plots (Fig. 4c), with a summary presented in Fig. 4d. Under DTC36 conditions, the bandgap was reduced to approximately 1.14 eV, whereas under DTC6 it remained wider at around 1.25 eV. This bandgap narrowing phenomenon is consistent with computational calculations and previous reports attributed the effect to a change of the band structure due to cation disorder.^13,18,33 Furthermore, the incorporation of urea led to a slight additional reduction in the bandgap under both DTC conditions (Fig. S8c and d), likely caused by internal stress relaxation associated with grain growth. Similar phenomena have been observed in perovskites and other semiconductor materials, where grain size enlargement results in reduced bandgap energy without significant changes to the lattice structure.^34–36

Fig. 4 (a) Optical photographs of AgBiS₂ thin film deposited on FTO substrate with and without the incorporation of 0.1 M urea additives. The scale bar corresponds to 1 cm. (b) Absorption coefficient and (c) Tauc plot under varing the relative DTC concentration ratio or whether urea additives are used or not. (d) Direct bandgap (E_g) box plot as a function of solution conditions. (e–h) Ultraviolet photoelectron spectroscopy (UPS) spectra of AgBiS₂ films prepared with different DTC concentration ratios. (e and f) Represent DTC6 conditions and (g and h) correspond to DTC36, both with the addition of 0.1 M urea. (i) Energy band diagrams of AgBiS₂ thin films based on DTC ligand concentrations, illustrating the relative positions of the valence band maximum and conduction band minimum as determined from UPS analysis.

To confirm whether the observed changes in optical properties were accompanied by changes in electronic band structure, ultraviolet photoelectron spectroscopy (UPS) was performed. The UPS spectra of DTC6 and DTC36 are shown in Fig. 4e–h, respectively. The work function was calculated by subtracting the secondary electron cutoff (SECO) value from the photon energy (He I, hν = 21.22 eV), and the valence band maximum (VBM) was estimated by linear extrapolation of the Fermi edge. Combining this information with optical bandgap values allowed the construction of energy band diagrams, as illustrated in Fig. 4i. Under DTC6 conditions, AgBiS₂ films exhibited electronic characteristics consistent with a slightly n-type semiconducting behavior, in agreement with previous studies.^37,38 In contrast, AgBiS₂ films prepared under DTC36 conditions demonstrated a subtle transition toward p-type semiconducting characteristics, indicating that cation disorder may induce modifications in the intrinsic carrier type. Such bipolar conduction behavior of AgBiS₂ has also been reported in the literature,³⁹ suggesting that variations in synthesis conditions and defect chemistry can significantly influence the native carrier polarity. Although cation disorder is well known to reduce bandgaps in compound semiconductors, its influence on native conductivity types remains underexplored. However, reports on nitrides and sulfides indicate that cation disorder can alter conductivity behavior.^40–42 The findings suggest that the emergence of donor- or acceptor-like defect states, resulting from nonperiodic cation configurations, may play a significant role in determining carrier polarity, warranting further investigations into the nature and tunability of these defects.

Next, we examined whether the morphological and optical properties of urea additives translated into improved device performance by conducting PEC measurements. Prior to evaluation, the AgBiS₂ film fabrication process was optimized to minimize defect-related trap states such as vacancies or antisite substitutions, which could act as major recombination centers within the bandgap. Energy dispersive X-ray spectroscopy (EDX) was employed to assess compositional uniformity and stoichiometry. As shown in Fig. S9 and S10, molecular inks with various Ag : Bi precursor molar ratios (1 : 1, 1 : 1.2, and 1 : 1.4) were synthesized and analyzed. Elemental mapping confirmed the homogeneous distribution of Ag, Bi, and S across the film surface under all conditions. However, quantitative EDX analysis in Table S1 revealed that films derived from the 1 : 1 precursor ratio exhibited a pronounced Ag-rich composition, similar to previous studies on AgBiS₂,^18,29 with an actual atomic ratio of approximately Ag : Bi = 1 : 0.75, equivalent to about 33% excess silver. This observation aligns with previously reported compositional trends for AgBiS₂ systems. To achieve an ideal stoichiometric Ag : Bi ratio in the final AgBiS₂ film, a slight excess of the Bi precursor was incorporated into the molecular ink. Among the tested formulations, a 1 : 1.4 = Ag : Bi ratio yielded the composition closest to the ideal 1 : 1 stoichiometry. Accordingly, the AgBiS₂ thin films prepared under these conditions were selected for PEC performance assessment.

Based on the previously characterized band structure obtained via UPS, a photocathode device was constructed using the architecture FTO/AgBiS₂/CdS/TiO₂/Pt, representing the first integration of AgBiS₂ into a full photocathode design. In this multilayer structure, the CdS layer functions as a conventional buffer layer, the TiO₂ layer acts as an electron transport layer, and the Pt layer serves as a hydrogen evolution co-catalyst. This device configuration has been widely adopted in chalcogenide-based photocathodes because of its proven efficacy in facilitating charge separation and promoting catalytic activity.^43–50 CdS, TiO₂, and Pt layers were sequentially deposited on the AgBiS₂ photoabsorbing layer via chemical bath deposition (CBD), atomic layer deposition (ALD), and ion sputtering, respectively. PEC performance was evaluated using linear sweep voltammetry (LSV) in a three-electrode setup under simulated solar illumination (AM 1.5 G) in a K-Pi buffer electrolyte at neutral pH. To mitigate undesirable dark current arising from insufficient passivation and direct electrical contact between the Pt cocatalyst and the conductive FTO substrate, an ultrathin (∼1 nm) TiO₂ interlayer was introduced between AgBiS₂ and CdS buffer layers. This nanometric barrier facilitated electron tunneling while simultaneously providing physical passivation to reduce interfacial recombination and electrical leakage. Moreover, the interlayer was expected to suppress Cd diffusion during chemical bath deposition and protect the AgBiS₂ surface from alkaline-induced degradation. Interfacial passivation layers such as TiO₂ have been widely adopted in various photoelectronic architectures to enhance both chemical robustness and carrier selectivity.⁵¹ In the present context, this strategy plays a crucial role in maintaining the structural and functional integrity of devices.

A schematic of the photocathode structure is shown in Fig. 5a. A cross-sectional SEM image of the completed stack (Fig. 5b) clearly distinguishes AgBiS₂ (∼70 nm) deposited under the DTC36, CdS (∼50 nm), and TiO₂ (∼70 nm) layers. Because of its ultrathin thickness (∼1 nm), the TiO₂ interlayer located between AgBiS₂ and CdS was not visually resolved in the image. As shown in the J–V curves in Fig. 5d, devices fabricated without the TiO₂ interlayer displayed a continuously increasing dark current, indicative of parasitic current paths. In contrast, devices incorporating the interlayer exhibited a suppressed dark current, underscoring the efficacy of the TiO₂ layer in passivating interfacial leakage. Increasing the TiO₂ interlayer thickness from 0 nm to 2.5 nm led to progressively poorer CdS deposition with the films exhibiting rough and discontinuous coverage due to weak adhesion between CdS and the underlying oxide surface (Fig. S11). This morphological deterioration hindered the formation of a uniform buffer layer and consequently limited the proper operation of the photoelectrode. The corresponding photoelectrochemical J–V curves in Fig. S12 also show a pronounced decrease in photocurrent density with increasing TiO₂ thickness, confirming that too thick interlayer impedes carrier transport across the interface. Therefore, the 1 nm TiO₂ interlayer used in this work represents an experimentally optimized thickness that achieves a balance between surface passivation and efficient carrier extraction. To investigate the origin of this dark current, XRD and XPS depth profiling were employed to assess the structural integrity and elemental diffusion before and after CdS deposition. In the XRD spectra (Fig. S11), only peaks corresponding to the substrate and cubic AgBiS₂ were detected, with no discernible reflections from CdS, presumably because of the low crystallinity or amorphous nature of the CBD-deposited CdS layer.

Fig. 5 (a) Schematic of the AgBiS₂ photocathode. (b) The cross-sectional SEM image of fabricated AgBiS₂ photocathode. The illustration is provided as an eye-guided representation to indicate the approximate positions of each layer and the scale bar is 200 nm. (c) Band diagram of AgBiS₂ (DTC36) photocathode showing the relative energy positions of each of the semiconductors calculated from UPS. (d) LSV curves of FTO/AgBiS₂/CdS/TiO₂/Pt, FTO/AgBiS₂/TiO₂(1 nm)/CdS/TiO₂/Pt, and FTO/AgBiS₂/TiO₂(1 nm)/CdS/TiO₂/Pt with 0.1 M urea. (e) LSV curves of FTO/Au/AgBiS₂/TiO₂(1 nm)/CdS/TiO₂/Pt photocathodes without or with 0.1 M urea. All measurements were performed in 0.5 M K-Pi buffer (pH 7.02).

Notably, when the TiO₂ interlayer was absent, additional diffraction peaks emerged at ∼23.9° (strong) and ∼48.8° (weak) following CdS deposition. These peaks were attributed to the formation of secondary phases, possibly arising from undesired chemical reactions between the AgBiS₂ layer surface and the strongly alkaline (pH 10–11) CBD environment. These spurious phases were not observed in the samples with a TiO₂ interlayer, highlighting their protective capability. Notably, no detectable change in crystallinity was observed in the interlayer-protected films after CdS deposition, indicating that the TiO₂ layer preserved the chemical and structural fidelity of the AgBiS₂ interface under aggressive conditions. To further confirm the function of the interlayer in suppressing Cd incorporation, XPS depth profiling was performed on the films with and without the TiO₂ layer (Fig. S12). Based on an assumed uniform etch rate, complete removal of the 50 nm CdS layer was estimated to occur within 80–120 s. In the absence of an interlayer, a substantial Cd signal persisted at shallow etching depths, suggesting significant inward diffusion of Cd into the AgBiS₂ absorber. In contrast, the incorporation of the TiO₂ interlayer effectively mitigated Cd diffusion, as evidenced by the sharp decline in Cd concentration upon reaching the AgBiS₂ interface. Moreover, elemental analysis revealed that the surface atomic ratio of the CdS layer in the absence of an interlayer deviated significantly from the ideal 1 : 1 stoichiometry, which is consistent with Cd infiltration into the underlying AgBiS₂. Quantitative depth profiling (Fig. S13) further confirmed that within the estimated interface region (80–120 s etching time), samples containing the interlayer exhibited over a 50% reduction in Cd content compared to their unpassivated counterparts. These results underscore the efficacy of the TiO₂ layer in suppressing interdiffusion and preserving the interface composition. In the absence of an interlayer, a persistent dark current was observed owing to the direct contact between the Pt cocatalyst and conductive FTO substrate. Additionally, uncontrolled Cd diffusion is likely to induce midgap trap states, thereby limiting photocurrent generation. Upon introducing a very thin (1 nm) TiO₂ interlayer, the PEC performance was remarkably enhanced. At 0 V versus the reversible hydrogen electrode (RHE), corresponding to the theoretical thermodynamic equilibrium potential for hydrogen evolution reaction (HER), the urea-treated photocathode delivered a photocurrent density of 0.6 mA cm⁻², approximately 1.5 times higher than its untreated counterpart. A concurrent increase in onset potential from 0.53 V to 0.6 V vs. RHE was also observed, attributed to enhanced grain morphology, reduced interfacial recombination, and improved light absorption. Interestingly, the PEC activity was highly sensitive to DTC concentration in the precursor solution. Under DTC-deficient conditions (DTC6), the photocurrent response was negligible regardless of urea addition. In contrast, DTC-sufficient formulations (DTC36) yielded significant cathodic photocurrent characteristics for HER activity (Fig. S14). The anodic behavior of DTC6-based devices likely stems from the n-type character of AgBiS₂ under relatively low disorder, forming isotype junctions with the n-type CdS and TiO₂ layers. Conversely, DTC36-based films, exhibiting p-type semiconducting behavior owing to enhanced cation disorder, form type-II heterojunctions with CdS and TiO₂, enabling favorable band alignment for electron extraction and cathodic photocurrent generation. To further improve PEC efficiency, a thin Au underlayer was deposited on the FTO substrate prior to AgBiS₂ coating. Commonly utilized in photocathodes for water splitting, the Au interlayer functions as both a hole extraction layer and an optical back reflector, enhancing carrier separation and light harvesting.

As shown in Fig. 5e, the optimized FTO/Au/AgBiS₂/TiO₂(1 nm)/CdS/TiO₂/Pt structure exhibited a photocurrent density of approximately 1.2 mA cm⁻², representing a 20% increase compared to the device without urea. In addition, the onset potential increased to 0.68 V vs. RHE. To corroborate these enhancements in optoelectronic performance, incident photon-to-electron conversion efficiency (IPCE) measurements were performed at 0 V vs. RHE. As presented in Fig. S15, urea-treated devices demonstrated a significantly enhanced IPCE in the near-UV region, with further improvements observed upon the introduction of the Au back contact. However, photon absorption beyond 600 nm failed to translate into meaningful photocurrent, despite intrinsic bandgap of AgBiS₂ materials (about 1.0–1.3 eV) being sufficiently narrow to absorb long-wavelength light. This limitation can be attributed to the trap-assisted recombination of charge carriers generated by deeply penetrating long-wavelength photons.⁵² Although short-wavelength photons generate excitons near the film surface with short transport paths, long-wavelength photons penetrate deeper into the absorber and are more susceptible to recombination via shallow or deep trap states. As illustrated in Fig. S16, regardless of film thickness, photons with wavelengths >600 nm did not significantly contribute to the photocurrent, suggesting a fundamentally limited charge diffusion length in AgBiS₂. Additionally, residual organic byproducts or the incomplete removal of precursor species may exacerbate recombination losses and hinder efficient carrier extraction in the long-wavelength regime. Although AgBiS₂ has been predominantly investigated for photovoltaic applications,⁵³ and in one study, as a photoanode,⁵⁴ its application as a photocathode has not been previously reported. To the best of our knowledge, this is the first report of AgBiS₂ functioning in a photocathodic configuration.

Stability tests revealed that the optimized AgBiS₂ photocathode retained approximately 70% of its initial photocurrent after 6 h of continuous operation at 0 V vs. RHE (Fig. 6). During the initial 2 h, the device maintained nearly 90% of its original current density, but a gradual linear degradation was subsequently observed, indicating progressive photocorrosion. Although the device structure incorporated passivation layers, such as CdS and TiO₂, persistent decline in the photocurrent was likely attributed to photocorrosion of the AgBiS₂ photoabsorber under continuous illumination. This degradation trend is consistent with previous reports on AgBiS₂-based photoelectrodes and suggests that instability arises from intrinsic material limitations. In particular, Ag⁺, a soft cation with high ionic mobility, is prone to migration under an electric field or thermal stress, which can induce interfacial degradation and morphological instability. Moreover, the presence of intrinsic defects in the AgBiS₂ lattice promotes trap-assisted recombination pathways and reduces carrier lifetime, further accelerating device degradation over time. Another possible contribution to device degradation could arise from partial corrosion of the CdS and TiO₂ layers or detachment of the Pt co-catalyst during photoelectrochemical performance measurement operation.^55,56 Therefore, overcoming these degradation pathways and performance limitations will require suppression of trap states and reduction of non-radiative recombination through optimization of absorber thickness and band alignment with charge transport layers, thereby enabling more efficient utilization of long-wavelength charge carriers. Precise control of layer or catalyst detachment and bubble release dynamics is also essential for ensuring long-term operational stability.

Fig. 6 Photocurrent density versus time (hour) curve for the FTO/Au/AgBiS₂(DTC36, with 0.1 M urea)/TiO₂(1 nm)/CdS/TiO₂/Pt structured photocathode under 0 V_RHE. The inset shows a photograph of the device during operation showing H₂ bubble generation.

Conclusions

This study demonstrated a coordination chemistry-based strategy to simultaneously induce cation disorder and improve the morphology of AgBiS₂ thin films through the introduction of urea, a low-cost and environmentally benign Lewis base additive with electron-donating groups. Urea exhibits stronger binding affinity toward metal cations than sulfur-based DTC ligands, thereby preferentially coordinating with Bi³⁺ and Ag⁺ ions. This coordination suppresses nucleation and modulates crystallization kinetics, leading to significant grain enlargement and enhanced film coverage without disrupting the DTC-derived cation disorder framework. The resulting morphological improvement contributed to increased optical absorption, with an optimized urea concentration of 0.1 M yielding an absorption coefficient of up to 8 × 10⁵ cm⁻¹. These effects were systematically validated through computational binding energy calculations and comprehensive experimental analyses, including XRD, XPS, SEM, and UV-visible spectroscopy.

To verify whether these enhancements translate into practical device performance, we fabricated an FTO/Au/AgBiS₂/TiO₂ (1 nm, interlayer)/CdS/TiO₂/Pt structured photocathode and performed photoelectrochemical water splitting for HER evaluation. The introduction of an ultrathin (1 nm) TiO₂ interlayer between AgBiS₂ photo absorbing and CdS buffer layers effectively mitigated interfacial issues, such as undesirable Cd diffusion and secondary phase formation during chemical bath deposition. Consequently, the device incorporating urea-treated and cation disordered AgBiS₂ exhibited a photocurrent density of approximately 1.2 mA cm⁻² at 0 V vs. RHE and an onset potential of 0.68 V vs. RHE, representing an approximately 20% enhancement compared to its untreated counterpart. Furthermore, the optimized photocathode retained 70% of its initial photocurrent after 6 h of continuous operation at 0 V vs. RHE, indicating appreciable operational stability. These findings validate the dual modulation of cation disorder and film morphology as a viable approach for improving optoelectronic performance of molecular ink-driven AgBiS₂-based photocathodes. This study represents the first demonstration of a functional PEC photocathode based on AgBiS₂ that exhibits excellent HER performance in terms of photocurrent generation and long-term durability.

Experimental

Materials

All chemicals were used as received, without additional purifications: silver chloride powder (AgCl, 99.999%, Sigma-Aldrich), bismuth(III) chloride powder (BiCl₃, 99.99%, Sigma-Aldrich), urea (CO(NH₂)₂, ≥99.0%, Sigma-Aldrich), carbon disulfide (CS₂, ≥99.9%, Sigma-Aldrich), N-butylamine (CH₃(CH₂)₃NH₂, 99.5%, Sigma-Aldrich), pyridine (C₂H₅N, anhydrous, 99.8%, Sigma-Aldrich).

Preparation of Ag–Bi–DTC–urea molecular inks

All DTC solution experiments were conducted in a fume hood. N-butyl DTC solution was prepared by reacting with n-butylamine carbon disulfide (CS₂). Pyridine 2.85 mL (2.13 mL) was mixed with n-butylamine 0.09 mL (0.53 mL) in a glass vial and stirred in an ice bath for 15 min. 0.06 mL (0.34 mL) of CS₂ was injected into the vial under stirring to prepare DTC6 (or DTC36) inks. During this process, n-butylamine reacted exothermically with CS₂, and the solution color changed from transparent to yellow. At higher DTC concentrations, DTC36 exhibited a deeper yellow-orange hue compared with the lighter yellow of DTC6. The DTC solution was stirred for 5 min after injecting CS₂, removed from the ice bath, and allowed to warm naturally to room temperature for 5 min.

All DTC solutions were used to dissolve silver chloride (AgCl) and bismuth chloride (BiCl₃), maintaining a certain metal-to-DTC molar ratio (1 : 6 or 36) in an N₂-filled glove box, at a metal concentration of 0.05 M for both Ag and Bi–DTC solutions. After overnight stirring, the fully dissolved Ag–DTC solution was transparent yellow, whereas the Bi-DTC solution was transparent orange, with no precipitates. The mixture was mixed overnight in a 1 : 1 volume ratio to determine the stoichiometric ratio with fully dissolved Ag–DTC and Bi–DTC solutions to obtain a homogeneous Ag–Bi–DTC solution. The metal concentration ratio was used to control the stoichiometric ratio of AgBiS₂ by adjusting the relative volumes of Ag–DTC and Bi–DTC solutions. After preparing a homogeneous Ag–Bi–DTC solution, the desired molar concentrations of urea additives were added. The solution was stirred overnight to obtain a homogeneous orange Ag–Bi–DTC–urea solution without any observable precipitates. Urea dissolved quickly and did not induce further color changes.

Fabrication of the AgBiS₂ thin film

FTO (TEC 8)-coated substrates (or ITO substrates) were cleaned via sequential sonication with acetone, isopropanol (IPA) and deionized (DI) water for 10 min each, followed by UV-ozone treatment for 40 min. The fabrication and annealing of the AgBiS₂ thin films were conducted in an N₂-filled glove box. The prepared inks were spin-coated onto the cleaned substrates at 5000 rpm for 30 s. Deposited films were annealed in two steps: 100 °C for 2 min and at 250 °C for 2 min. This process was repeated six times, and the films were post-annealed at 250 °C for 10 min.

Fabrication of AgBiS₂ photocathode

CdS layers were deposited via chemical bath deposition (CBD) by immersing samples in a bath solution of cadmium sulfate (CdSO₄, 99%, Sigma-Aldrich), thiourea (99%, Sigma-Aldrich), deionized water, and ammonium solution (NH₄OH, 25–30 wt%. Duksan, Korea) for 11 min and 30 s at 56.5 °C. Samples were then rinsed using deionized water. TiO₂ layers were deposited via customized atomic layer deposition using tetrakis(dimethylamino) titanium (TDMAT, Easychem, Korea) and deionized water (H₂O) as the Ti and O sources, respectively. A TiO₂ layer was deposited for 1000 cycles at 120 °C. An ion sputter coater (G20 ion sputter coater, GSEM, Korea) was used to deposit the Pt co-catalyst for 60 s at a current of 10 mA. For samples incorporating the TiO₂ interlayer, a layer with a nominal thickness of 1 nm (corresponding to 15 ALD cycles) was deposited via atomic layer deposition (ALD) prior to CdS deposition. Subsequently, the surface was rendered hydrophilic by UV-ozone treatment for 20 min. For FTO/Au substrates, Au (3 mm Dia × 3 mm Th pellets 4 N, iTASCO, Korea) was deposited on cleaned FTO substrates using a thermal evaporator (DDHT-SP011, Daedong High Technologies, Korea), producing a Au layer of 70 nm.

Material characterizations

Surface morphologies and cross-sectional profiles were analyzed using field-emission scanning electron microscopy (FE-SEM, JSM-IT800, JEOL Ltd, Japan). Phase identification and evaluation of cation disorder were conducted by high resolution XRD (Smartlab, Rigaku, Japan) using Cu-Kα radiation (λ = 0.15406 nm). Surface chemical analyses and depth profiles were obtained using XPS (ESCALAB 250Xi, Thermo Scientific, UK). Absorption coefficients were measured using a UV-vis spectrophotometer (UV-3600i Plus, Shimadzu, Japan). Band structure analyses were performed using ultraviolet phvotoelectron spectroscopy (NEXSA G2, Thermo Fisher Scientific, USA).

AgBiS₂ photocathode performance evaluation

Photoelectrochemical measurements were performed with conventional three-electrode configurations using a potentiostat (PP211, Zahner, Germany) in a 0.5 M K-Pi buffer (pH = 7.02) solution electrolyte. An Ag/AgCl (sat. KCl) and a coiled Pt coil were used as the reference and counter electrodes, respectively. The scan rate of the J–V curves was 10 mV s⁻¹. Sunlight simulations and 1-sun calibrations were conducted using a commercial AM 1.5 G solar simulator (LH300, Newport Corporation, USA). Photocurrent density was calculated by dividing the photocurrent with the measured active area of each device. For all PEC measurements, the applied potentials were converted to RHE values using the following equation: E_RHE = E_Ag/AgCl + 0.05916 pH + 0.197. IPCE measurements were performed using an electrochemical workstation (Zennium, Zahner, Germany) and a potentiostat (PP211, Zahner, Germany) with a monochromatic light source (TLS03, Zahner, Germany).

Computational details

All calculations were performed using Gaussian 09 software package.⁵⁷ Geometry optimization of the ground state and vibrational frequency calculations for Raman spectroscopic analysis were performed at the HF/SDD level of theory, and the energy minima were confirmed by calculations using the zero imaginary modes of vibrations. The binding energy (E_b) between DTC and Ag⁺ or Bi³⁺ was calculated from the ground state energies, E_b = (E_complex − E_ion-n × E_{DTC or urea})/m, where E_complex is the total energy of the DTC–Ag⁺, urea–Ag⁺, DTC–Bi³⁺ or urea–Bi³⁺ complex; E_ion and E_{DTC or urea} denote the energies of the ions (Ag⁺ or Bi³⁺) and DTC or urea, respectively; and n and m are the number of DTC (or urea) molecules and total number of atoms, respectively. All calculations were performed in pyridine using a polarizable continuum model.⁵⁸

Author contributions

N. Ha conceptualized and conducting the overall experiment analysed the data, wrote the manuscript and investigated. J. Jung and D. Lee proceeded computational calculation. J. Park, H. Shin and S. U. Lee conducted supervision. W. Yang supervised this experiment, directed the research, contributed to the writing of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data will be provided upon reasonable request.

Supplementary information (SI): calculated optimized model information, SEM, XRD, XPS and UV-Vis spectra, supplementary figures and tables supporting the main text. See DOI: https://doi.org/10.1039/d5ta07632c.

Acknowledgements

This work was supported by the National R&D Program of the National Research Foundation of Korea (NRF), funded by the Ministry of Science and ICT (Grant no RS-2025-05482972).

Notes and references

1. C. C. Lin, Y.-T. Tsai, H. E. Johnston, M.-H. Fang, F. Yu, W. Zhou, P. Whitfield, Y. Li, J. Wang and R.-S. Liu, J. Am. Chem. Soc., 2017, 139, 11766.
2. P. F. Ndione, Y. Shi, V. Stevanovic, S. Lany, A. Zakutayev, P. A. Parilla, J. D. Perkins, J. J. Berry, D. S. Ginley and M. F. Toney, Adv. Funct. Mater., 2014, 24, 610.
3. A. Abdellahi, A. Urban, S. Dacek and G. Ceder, Chem. Mater., 2016, 28, 5373.
4. A. Simonov and A. L. Goodwin, Nat. Rev. Chem., 2020, 4, 657.
5. J. Jung, N. Ha and W. Yang, ACS Appl. Mater. Interfaces, 2025, 17, 16410.
6. W. Yin, A. Grimaud, G. Rousse, A. M. Abakumov, A. Senyshyn, L. Zhang, S. Trabesinger, A. Iadecola, D. Foix, D. Giaume and J. M. Tarascon, Nat. Commun., 2020, 11, 1252.
7. M. Kim, D. Kim, Y. Wen, M. Kim, H. M. Jang, H. Li, L. Gu and B. Kang, Joule, 2019, 3, 1064.
8. A. R. Symington, J. Purton, J. Statham, M. Molinari, M. S. Islam and S. C. Parker, J. Mater. Chem. A, 2020, 8, 19603.
9. C. Gervas, M. D. Khan, C. Zhang, C. Zhao, R. K. Gupta, E. Carleschi, B. P. Doyle and N. Revaprasadu, RSC Adv., 2018, 8, 24049.
10. S. Zuo, Z. P. Wu, G. Zhang, C. Chen, Y. Ren, Q. Liu, L. Zheng, J. Zhang, Y. Han and H. Zhang, Angew Chem. Int. Ed. Engl., 2024, 63, e202316762.
11. W. Chen, D. Dahliah, G.-M. Rignanese and G. Hautier, Energy Environ. Sci., 2021, 14, 3567.
12. Y. Wang, S. R. Kavanagh, I. Burgués-Ceballos, A. Walsh, D. O. Scanlon and G. Konstantatos, Nat. Protoc., 2022, 16, 235.
13. M. Righetto, Y. Wang, K. A. Elmestekawy, C. Q. Xia, M. B. Johnston, G. Konstantatos and L. M. Herz, Adv. Mater., 2023, 35, e2305009.
14. M. Bernechea, N. Cates, G. Xercavins, D. So, A. Stavrinadis and G. Konstantatos, Nat. Protoc., 2016, 10, 521.
15. E. Gu, X. Lin, X. Tang, G. J. Matt, A. Osvet, Y. Hou, S. Jäger, C. Xie, A. Karl and R. Hock, J. Mater. Chem. C, 2018, 6, 7642.
16. Y. Wu, L. Wan, W. Zhang, X. Li and J. Fang, CrystEngComm, 2019, 21, 3137.
17. J. van Embden and E. Della Gaspera, ACS Appl. Mater. Interfaces, 2019, 11, 16674.
18. N. Ha, G. Lee, J. Park, J. H. Lee, J. Jung, S. V. Barma, J. Kim, J. H. Kim, J. K. Kim and S. J. Kwon, Adv. Energy Mater., 2025, 15, 2402099.
19. S. Sugarthi, G. Bakiyaraj, R. Abinaya, M. Navaneethan, J. Archana and M. Shimomura, Mater. Sci. Semicond. Process., 2020, 107, 104781.
20. W. Wang, S. Bao, C. Gao, T. Huang, Y. Zhu and R. Xu, Mater. Sci. Eng., B, 2024, 310, 117739.
21. F. Yao, L. Jiang, Y. Qi, R. Li, Y. Li, Y. Xu, H. Liu and Q. Lin, Appl. Mater. Today, 2022, 26, 101262.
22. Y. Jo, K. S. Oh, M. Kim, K. H. Kim, H. Lee, C. W. Lee and D. S. Kim, Adv. Mater. Interfaces, 2016, 3, 1500768.
23. Q. Cai, C. Liang, Z. Lin, W. Zhang, G. Shen, H. Dong, X. Xu, H. Wang, C. Mu and G. Xing, Sustain. Energy Fuels, 2022, 6, 2955.
24. C. M. Hsieh, Y. S. Liao, Y. R. Lin, C. P. Chen, C. M. Tsai, E. Wei-Guang Diau and S. C. Chuang, RSC Adv., 2018, 8, 19610.
25. X. Yang, H. Xiang, J. Huang, C. Zhou, R. Ran, W. Wang, W. Zhou and Z. Shao, J. Colloid Interface Sci., 2022, 628, 476.
26. J. V. Patil, S. S. Mali and C. K. Hong, Nanoscale, 2019, 11, 21824.
27. S. Jin, Y. Wei, X. Yang, D. Luo, Y. Fang, Y. Zhao, Q. Guo, Y. Huang, L. Fan and J. Wu, Org. Electron., 2018, 63, 207.
28. R. Zhang, M. Li, Y. Huan, J. Xi, S. Zhang, X. Cheng, H. Wu, W. Peng, Z. Bai and X. Yan, Inorg. Chem. Front., 2019, 6, 434.
29. T. Nakazawa, D. Kim, Y. Oshima, H. Sato, J. Park and H. Kim, ACS Omega, 2021, 6, 20710.
30. T. Manimozhi, S. Kavirajan, K. Kamala Bharathi, E. Senthil Kumar and M. Navaneethan, J. Mater. Sci.: Mater. Electron., 2022, 33, 12615.
31. M. Jiang, Y. Ding, H. Zhang, J. Ren, J. Li, C. Wan, Y. Hong, M. Qi, B. Mei and L. Deng, J. Solid State Electrochem., 2020, 24, 2487.
32. L. Mehdaoui, R. Miloua, M. Khadraoui, M. Bensaid, D. Abdelkader, F. Chiker and A. Bouzidi, Phys. B Condens. Matter, 2019, 564, 114.
33. F. Vines, G. Konstantatos and F. Illas, Phys. Chem. Chem. Phys., 2017, 19, 27940.
34. M. T. Ahmed, S. Islam and F. Ahmed, Adv. Condens. Matter Phys., 2022, 2022, 9535932.
35. P. Jain and P. Arun, Thin Solid Films, 2013, 548, 241.
36. K. P. Dhakal, S. Roy, H. Jang, X. Chen, W. S. Yun, H. Kim, J. Lee, J. Kim and J.-H. Ahn, Chem. Mater., 2017, 29, 5124.
37. Y. Wang, L. Peng, Z. Wang and G. Konstantatos, Adv. Energy Mater., 2022, 12, 2200700.
38. I. Burgués-Ceballos, Y. Wang, M. Z. Akgul and G. Konstantatos, Nano Energy, 2020, 75, 104961.
39. H. J. Kim, J. Y. Park, Y. J. Choi, S. K. Kim, T. Yong, W. Lee, G. Seo, E. J. Lee, S. Choi and H. R. You, Adv. Energy Mater., 2025, 15, 2404552.
40. E. Sirotti, L. I. Wagner, C. M. Jiang, J. Eichhorn, F. Munnik, V. Streibel, M. J. Schilcher, B. März, F. S. Hegner and M. Kuhl, Adv. Energy Mater., 2024, 14, 2402540.
41. K. Ye, S. C. Siah, P. T. Erslev, A. Akey, C. Settens, M. S. B. Hoque, J. Braun, P. Hopkins, G. Teeter and T. Buonassisi, Chem. Mater., 2019, 31, 8402.
42. S. Ke, J. S. Mangum, A. Zakutayev, A. L. Greenaway and J. B. Neaton, Chem. Mater., 2024, 36, 3164.
43. S. Moon, J. Park, H. Lee, J. W. Yang, J. Yun, Y. S. Park, J. Lee, H. Im, H. W. Jang, W. Yang and J. Moon, Adv. Sci., 2023, 10, e2206286.
44. D. Huang, L. Li, K. Wang, Y. Li, K. Feng and F. Jiang, Nat. Commun., 2021, 12, 3795.
45. J. Park, W. Yang, Y. Oh, J. Tan, H. Lee, R. Boppella and J. Moon, ACS Energy Lett., 2019, 4, 517.
46. J. Park, K. H. Kim, D. Kim, J. K. Kim and W. Yang, Sustain. Energy Fuels, 2024, 8, 481.
47. P. Adams, R. Schnyder, T. Moehl, J. Bühler, A. L. Alvarez, M. Dimitrievska, K. McKenna, W. Yang and S. D. Tilley, Adv. Funct. Mater., 2024, 34, 2310596.
48. S. Moon, J. Park, H. Lee, J. W. Yang, J. Yun, Y. S. Park, J. Lee, H. Im, H. W. Jang and W. Yang, Adv. Sci., 2023, 10, 2206286.
49. X. Zhang, W. Yang, W. Niu, P. Adams, S. Siol, Z. Wang and S. D. Tilley, ChemSusChem, 2021, 14, 3967.
50. W. Yang, R. R. Prabhakar, J. Tan, S. D. Tilley and J. Moon, Chem. Soc. Rev., 2019, 48, 4979.
51. C. Zhou, Z. Xi, D. J. Stacchiola and M. Liu, Energy Sci. Eng., 2022, 10, 1614.
52. W. Niu, T. Moehl, P. Adams, X. Zhang, R. Lefèvre, A. M. Cruz, P. Zeng, K. Kunze, W. Yang and S. D. Tilley, Energy Environ. Sci., 2022, 15, 2002.
53. S. Akhil and R. G. Balakrishna, J. Mater. Chem. A, 2022, 10, 8615.
54. J. Y. Park, G. Park, S. Y. Bae, H. J. Kim, D. H. Lee, S. Ko, S.-K. Kim, G. Lee, H. R. You and H. Choi, ACS Appl. Energy Mater., 2023, 6, 3872.
55. J. Tan, W. Yang, Y. Oh, H. Lee, J. Park, R. Boppella, J. Kim and J. Moon, Adv. Energy Mater., 2019, 9, 1900179.
56. J. Park, M. J. Kim, Y. Kim, S. Lee, S. Park and W. Yang, ACS Energy Lett., 2024, 10, 212–237.
57. M. Frisch, G. Trucks, H. Schlegel, G. Scuseria, M. Robb, J. Cheeseman, G. Scalmani, V. Barone, B. Mennucci and G. Petersson, Gaussian 09, Revision D. 01, Gaussian, Inc, Wallingford, CT, 2013.
58. M. Cossi, N. Rega, G. Scalmani and V. Barone, J. Comput. Chem., 2003, 24, 669.

---