Two-Dimensional Multilayered Ferroelectric with Polarization-Boosted Photocatalytic Hydrogen Evolution

Abstract

Ferroelectric materials have attracted great attention for photocatalytic hydrogen (H₂) evolution due to their internal depolarization fields that promote carrier separation and directional migration. However, conventional inorganic ferroelectrics often suffer from wide band gaps and low conductivity, limiting their solar-to-hydrogen conversion efficiency. Here, we report a two-dimensional (2D) multilayered perovskite ferroelectric, [butylammonium]₂[ethylammonium]₂Pb₃I₁₀ (BAPI), which integrates robust spontaneous polarization (P_s) and excellent semiconductor properties to enable efficient photocatalysis. Under simultaneous light and ultrasonic excitation, BAPI/Pt (1 wt%) achieves a H₂ evolution rate of 1256 μmol g⁻¹ h⁻¹, which is twice that under light alone, due to dynamic polarization modulation that mitigates ionic screening and enhances internal electric fields. Notably, this enhancement vanishes when BAPI transitions to a centrosymmetric, nonpolar phase at 323 K, confirming the critical role of P_s. These findings offer a new pathway toward high-performance ferroelectric photocatalysts for solar hydrogen production.

1. Introduction

Photocatalytic hydrogen (H₂) evolution offers a promising route for converting intermittent solar energy into clean chemical fuel, enabling stable and high-density solar energy storage in the form of hydrogen [1,2,3,4]. Despite extensive progress, the performance of most conventional photocatalysts remains hampered by poor charge separation efficiency, leading to rapid recombination of photogenerated carriers and severely restricting the solar-to-hydrogen conversion efficiency [5,6]. In this context, polar materials, especially ferroelectrics, have emerged as compelling candidates for enhancing photocatalytic H₂ evolution, owing to the internal depolarization electric field generated by spontaneous polarization (P_s), which can facilitate directional separation of electron–hole pairs [7,8,9]. Compared with pyroelectrics, ferroelectrics offer stronger and tunable P_s, potentially delivering a more powerful built-in electric field and higher photocatalytic activity [10,11,12,13]. For instance, in single-domain BiFeO₃ nanosheets, photogenerated electrons and holes migrate along opposite P_s directions under the influence of the built-in depolarization field, leading to spatially selective photo-deposition of cocatalysts and improved photocatalytic activity [8]. Moreover, Liu et al. employed spatially resolved surface photovoltage spectroscopy to reveal that the potential difference along the polarization axis scales with particle size, indicating that the unscreened depolarization field, rather than polarization-induced band bending, is the main driving force for carrier separation [14]. Nevertheless, traditional inorganic ferroelectrics such as BiFeO₃, PbTiO_3, and BaTiO₃ often suffer from intrinsic limitations, including wide band gaps and low electrical conductivity, which significantly hinder their performance in visible-light-driven photocatalytic applications. Thus, there is a pressing need to develop new classes of ferroelectric materials that combine robust polarization with favourable semiconducting properties.

In recent years, metal halide perovskites have shown great potential for photocatalytic H₂ evolution, particularly via hydroiodic acid (HI) splitting, due to their outstanding optoelectronic properties, such as tunable band structures, high absorption coefficients, and long carrier diffusion lengths [15,16,17,18,19,20]. Among them, two-dimensional (2D) perovskites stand out, owing to their enhanced environmental stability, quantum-well-like electronic structures, and rich chemical tunability [21,22,23,24]. Notably, Fu et al. demonstrated the feasibility of water splitting using a halide-perovskite-based system composed of MoS₂-loaded (C₆H₅CH₂NH₂)₂PbI₄ as a hydrogen evolution photocatalyst and RuO_x-loaded WO₃ as an oxygen evolution photocatalyst, with the I₃⁻/I⁻ redox couple serving as an efficient charge shuttle—marking the first demonstration of overall water splitting in halide perovskite systems [25]. Importantly, the structural versatility of 2D perovskites provides an ideal platform for the integration of ferroelectricity via organic molecular engineering [26,27,28,29,30]. For example, Luo et al. designed and synthesized a series of 2D multilayered halide perovskite ferroelectrics, including EA₄Pb₃Cl₁₀ [31], (allyammonium)₂(EA)₂Pb₃Br₁₀ [32], (isobutylammonium)₂(EA)Pb₂Br₇ [33], and ([n-pentylaminium]₂[EA]₂Pb₃I₁₀ [34], by substituting conventional A-site cations such as methylammonium (MA⁺), formamidinium (FA⁺), or caesium (Cs⁺) with the larger ethylammonium (EA⁺) ion. These structural modifications break inversion symmetry and induce stable ferroelectric polarization while preserving high carrier mobility and visible-light absorption. As a result, these materials have demonstrated exceptional performance in photodetectors and photovoltaic devices [35,36]. However, despite their promising properties, the photocatalytic potential of 2D perovskite ferroelectrics has yet to be fully explored.

Here, we prepared a 2D multilayer perovskite ferroelectric, [butylammonium]₂[EA]₂Pb₃I₁₀ (BAPI), which exhibits efficient photocatalytic H₂ evolution. Ferroelectricity of BAPI was confirmed by its characteristic square-shaped phase-voltage hysteresis loop and butterfly-shaped amplitude-voltage response under piezoresponse force microscopy. Emphatically, under simultaneous visible-light and ultrasonic excitation, BAPI exhibits a H₂ evolution rate up to 1256 μmol g⁻¹ h⁻¹, nearly double that observed under light irradiation alone. In contrast, when heated above the polar–centrosymmetry transition temperature (323 K), BAPI transitioned to a non-polar phase and demonstrated nearly identical H₂ production under light-only and light plus ultrasound conditions, underscoring the critical role of spontaneous polarization. Our findings provide new insights into the functional design of halide perovskite photocatalysts and establish a platform for exploring ferroelectric-semiconductor hybrids in efficient solar-driven H₂ evolution systems.

2. Results and Discussion

Red plate-like microcrystals of BAPI, with lateral dimensions of approximately 50 × 50 × 3 μm, were obtained via a rapid cooling process (see Supporting Information for details). As shown in Figure 1b, the powder X-ray diffraction (XRD) pattern of the as-grown BAPI microcrystals matches well with the simulated pattern, indicating the formation of a phase-pure product with high crystallinity [37]. Notably, the XRD pattern exhibits pronounced diffraction peaks corresponding to the {200} planes, suggesting a preferred orientation, consistent with the [(EA)₂Pb₃I₁₀²⁻]_∞ slabs. Single crystal XRD analysis reveals that BAPI crystalizes in the polar space group Cmc2₁, adopting a two-dimensional (2D) three-layered perovskite architecture in which alternating n-butylammonium bilayers and [(EA)₂Pb₃I₁₀²⁻]_∞ inorganic slabs are aligned along the crystallographic a-axis (Figure 1c). Emphatically, the positively charged centres arising from the terminal –NH₃⁺ groups and the negatively charged centres of [PbI₆]²⁻ octahedra exhibit a spatial displacement from the central axis, generating a spontaneous polarization (P_s) along the c-axis. According to the point charge model, the dipolar moment and P_s of BAPI along the c-axis are estimated to be −9.3507 × 10⁻²⁹ C∙m and 2.23 μC∙cm⁻², respectively (Figure S1, Tables S1 and S2). As temperature increases, BAPI undergoes a polar-to-nonpolar phase transition: from the Cmc2₁ phase to a nonpolar Pbca phase at 317 K, and subsequently to a centrosymmetric tetragonal I4/mmm phase above 361 K, as previously reported (Figure S2) [37].

To further substantiate the ferroelectric nature of BAPI, piezoelectric force microscopy (PFM) measurements were conducted on thin films of BAPI deposited on indium-tin-oxide (ITO)-coated glass substrates. PFM, a widely employed technique for probing nanoscale ferroelectric behaviour, provides insight into the local electromechanical response through two key signals: the PFM amplitude, representing the amplitude of the piezo-response, and the PFM phase, reflecting the phase shift between the applied oscillating voltage and the material’s mechanical deformation. The phase purity of the BAPI film was validated by XRD pattern (Figure S3). Topographical mapping and corresponding lateral PFM amplitude and phase images are presented in Figure 1d–f. The PFM amplitude image reveals well-defined, stripe-shaped ferroelectric domain structures, indicative of in-plane polarization anisotropy (Figure 1e). In terms of phase imaging, the lateral phase image exhibits the same pattern as that in the amplitude image, with alternating dark and bright regions representing domains with antiparallel P_s orientation. Given the in-plane orientation of P_s in the layered structure, a negligible out-of-plane PFM response was observed (Figure S4). To probe the switchable ferroelectric behaviour, a DC voltage was applied via the conductive PFM tip. The recorded local hysteresis loops, including a well-defined square phase–voltage loop and a characteristic butterfly-shaped amplitude–voltage loop, clearly confirm the presence of switchable polarization under an external electric field of ferroelectric functionality (Figure 1g). The emergence of robust, switchable in-plane P_s in this 2D hybrid halide system is particularly compelling. Such spontaneous polarization not only evidences intrinsic ferroelectricity, but also suggests potential for enhanced charge separation via internal depolarization fields—a promising feature for future optoelectronic and photovoltaic applications where efficient separation and transport of photo-induced carriers are critical.

Prior to evaluating the photocatalytic performance, the photophysical properties of BAPI were systematically investigated. As shown in Figure 2a, BAPI exhibits a strong absorption across the ultraviolet to visible spectrum, with an absorption edge extending to 628 nm. The optical band gap (E_g) was calculated to be 2.04 eV (Figure 2b), as derived from the Tauc equation, which is comparable to that of other reported 2D multilayered halide perovskites, such as (BA)₂(FA)Pb₂I₇ (2.03 eV) [38], [n-pentylaminium]₂[EA]₂Pb₃I₁₀ (1.86 eV) [34], [R/S)-1-(4-bromophenyl)-ethylammonium]₂FAPb₂I₇ (1.45 eV) [39], and (3-(aminomethyl)pyridinium)₂(MA)Pb₂I₇ (2.08 eV) [40], (FA = formidinum, MA = methylammonium). These results underscore BAPI’s favorable optical response in the visible range for potential photoactive applications. The energy level alignment of BAPI was further elucidated using the ultraviolet photoelectron spectroscopy (UPS, Figure 2c). The secondary electron cutoff was located at 16.87 eV, and the valence band onset was located 1.35 eV below the Fermi level, resulting in a valence-band maximum (VBM) at −5.70 eV versus the vacuum level. Given the E_g of 2.04 eV, the conductive-band maximum (CBM) was correspondingly calculated to be −3.66 eV versus the vacuum level. After converting these energy values to the electrochemical scale versus the reversible hydrogen electrode (RHE), the VBM and CBM of BAPI were estimated to be 1.20 V and −0.84 V versus RHE, respectively. Figure 2d shows a schematic of the energy levels of BAPI. Notably, the band alignment of BAPI is properly positioned for proton reduction (0.046 V versus RHE) and iodine oxidation (0.376 V versus RHE) in our designed system (6.06 mol L⁻¹ HI aqueous solution). This enables the BAPI to potentially facilitate a one-step excitation HI splitting reaction.

The photocatalytic hydrogen evolution performance of BAPI was evaluated using as-prepared microcrystals dispersed in 30 mL of a saturated BAPI hydroiodic acid (HI) solution (6.06 mol L⁻¹) under visible light irradiation (λ ≥ 420 nm, 180 mW cm⁻²). The solubility profile of BAPI in HI solution at different temperatures is presented in Figure S5. As previously reported, the photocatalytic splitting of HI inevitably leads to the accumulation of triiodide species (I₃⁻), which exhibit strong absorption up to 700 nm [18,24]. This parasitic absorption significantly reduces the availability of incident photons for BAPI excitation, thereby suppressing its photocatalytic efficiency. To mitigate this issue, hypophosphorous acid (H₃PO₂), a commonly used stabilizing agent for HI, was introduced into the reaction system. H₃PO₂ is electrochemically inert within the potential window of 0.2~0.8 V versus RHE [18], and selectively reduces I₃⁻ back to I⁻ without undergoing further oxidative side reactions [41,42]. This strategic addition effectively suppresses light shielding by I₃⁻ and maintains a consistent iodine redox environment conducive to sustained photocatalytic activity.

As depicted in Figure 3a, pristine BAPI microcrystals exhibit a H₂ evolution rate of 230 μmol g⁻¹ h⁻¹ under visible light. Upon photodeposition of Pt nanoparticles as a co-catalyst, the H₂ evolution rate increases markedly, reaching a maximum of 680 μmol g⁻¹ h⁻¹ at a Pt loading of 1 wt%. However, further increasing the Pt content to 2 wt% leads to a reduced H₂ evolution rate (541 μmol g⁻¹ h⁻¹) attributed to a trade-off between the enhanced number of catalytically active sites and the parasitic light absorption by excess Pt—a phenomenon well-documented in semiconductor–metal hybrid systems [43,44]. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging reveals that Pt clusters (0.25–2 wt %) deposited on BAPI exhibit sizes in the range of 1.3~1.9 nm (Figure S6). Thus, all photocatalytic experiments in this work use a Pt loading of 1 wt%, unless otherwise specified. To elucidate the role of ferroelectricity in the observed photocatalytic enhancement, both piezocatalytic and piezo-photocatalytic experiments of BAPI/Pt were conducted. Under ultrasonic excitation in the absence of light, BAPI/Pt yields a modest H₂ evolution rate of 23 μmol g⁻¹ h⁻¹ (Figure S7). Strikingly, when subjected to simultaneous ultrasonic and light excitation, the H₂ evolution rate surges to 1256 μmol g⁻¹ h⁻¹, which is approximately 2 times the rate achieved under light irradiation alone. No detectable H₂ evolution was observed in control experiments conducted without light (Figure S8) or without BAPI microcrystals in the saturated BAPI aqueous HI solution under light plus ultrasound (Figure S9), confirming that the observed H₂ originates from the photocatalytic HI splitting over BAPI/Pt. The existence of I₃⁻ and complete HI splitting was confirmed by time-resolved UV–Vis spectra of the solution (Figure S10). Figure 3c presents the apparent quantum yield (AQY) of BAPI/Pt under combined light and ultrasound at various wavelengths. BAPI/Pt exhibits a relatively flat AQY in the range of 400~600 nm (6~8%), which sharply drops beyond 630 nm, consistent with the UV–vis absorption spectrum. These values exceed those reported for conventional inorganic ferroelectric photocatalysts (Table S3). Furthermore, piezo-photocatalytic H₂ evolution remains stable over four consecutive test cycles (4 h each), and no structural degradation is evident in the post-reaction XRD pattern (Figure S11), indicating the good structural durability of BAPI under operational conditions.

To further elucidate the role of P_s in enhancing photocatalytic activity, the influence of ultrasonic power was examined, along with a comparative study involving the non-polar phase of BAPI. Previous studies have demonstrated that P_s in polar materials can promote the spatial separation of photogenerated electron-hole pairs through the internal depolarization field [7,8,9,24,45]. However, in static environments, this field can be rapidly screened by free ionic species. Upon ultrasonic excitation, the collapse of cavitation microbubbles generates transient shock waves that periodically deform the BAPI lattice. This cyclic mechanical perturbation dynamically modulates the ferroelectric P_s, including repeated absorption and release of screening charges. Such charge oscillations diminish the electrostatic shielding of the internal field and reinforce the driving force for photogenerated carrier separation [46,47,48]. As a result, the H₂ evolution rate of BAPI/Pt increases progressively with ultrasonic power (Figure 4a), highlighting the synergistic effect of mechanical P_s modulation. To unambiguously confirm the role of P_s, piezo-photocatalytic H₂ production was tested at 323 K, where BAPI undergoes a structural phase transition into a nonpolar Pbca phase. Under these conditions, non-polar BAPI/Pt demonstrates nearly identical H₂ evolution rates under light alone and combined light and ultrasound (Figure 4b), corroborating the necessity of P_s for piezo-photocatalytic enhancement. A schematic representation of the electronic band structure and carrier separation process in polar BAPI/Pt is provided in Figure 4c. Upon visible light excitation, electrons are promoted from the valence band to the conduction band. The internal depolarization field arising from P_s directs the spatial migration of photogenerated electrons and holes toward opposite crystal facets, thereby suppressing charge recombination and enhancing the overall photocatalytic efficiency.

3. Materials and Methods

3.1. Materials

Lead (II) acetate trihydrate (Pb(Ac)₂·3H₂O, 99.5%, Aladdin, Shanghai, China), butylamine (BA; ≥99.5%, Aladdin, Shanghai, China), Barium sulfate (BaSO₄, AR, Aladdin, Shanghai, China), ethylamine (EA, 68.0–72.0% in H₂O, Adamas, Shanghai, China), Hydroiodic acid (HI, 55–57 wt% aqueous solution, Adamas-beta, Shanghai, China), H₂PtCl₆·6H₂O (AR, Alfa, Shanghai, China), and H₃PO₂ (50 wt% solution in water, Adamas, Shanghai, China) were used. All the chemicals were bought and used without further purification.

3.2. Synthesis of BAPI Microcrystals

Red plate-like microcrystals of BAPI were synthesized via a fast-cooling method from a hydroiodic acid solution. Specifically, 9.76 g of Pb(Ac)₂·3H₂O (25.7 mmol) was dissolved in 60 mL of a mixed aqueous solution of HI and H₃PO₂ in a 4:1 volume ratio under stirring at room temperature. Subsequently, 0.94 g of BA (12.85 mmol) was added, resulting in the formation of a yellow precipitate. EA (1.37 g, 21.3 mmol) was then introduced dropwise into the reaction mixture, gradually changing the precipitate color from yellow to red. The suspension was heated under stirring until a clear solution was obtained. The clear solution was then cooled in an ice bath under continuous stirring, forming red plate-like microcrystals of BAPI. The resulting crystals were collected by vacuum filtration and dried in a vacuum oven at 60 °C for 6 h to yield the dry polycrystalline BAPI product. The filtered saturated mother liquor was retained and used in subsequent photocatalytic experiments.

3.3. Material Characterization

Powder diffraction patterns were recorded over a 2θ range of 5~45°, using a step size of 0.05° and a scan rate of 2° min⁻¹. The height of BAPI plates was estimated using a step profiler. Piezoelectric force microscope (PFM) was carried out using a Bruker Dimension Edge atomic force microscope at room temperature. Both amplitude and phase signals were collected in lateral mode to evaluate in-plane ferroelectric domain structures. Ultraviolet–visible (UV–vis) diffuse reflectance spectrum was measured by a UV–vis spectrophotometer (Jindao UV-2450, Shanghai, China), equipped with an integrating sphere. BaSO₄ was employed as a 100% reflectance reference. The optical band gaps (E_g) were estimated using the Tauc equation for a direct transition:${\left[hvF\left(R_{\infty }\right)\right]}^2=A(hv-E_g)$, where h is Planck’s constant, v is the vibration frequency of the photon, A is the proportionality constant, and $F\left(R_{\infty }\right)$ is the Kubelka–Munk function. The Kubelka–Munk function is defined as F(R_∞) = ${K/S=(1-R_{\infty })}^2/{2R}_{\infty }$, where K is the absorption coefficient, S is the back-scattering coefficient, and $R_{\infty }$ represents the reflectance of an optically thick sample. Work functions and valence band maxima were determined via ultraviolet photoelectron spectrum (UPS) using a He I excitation source (21.22 eV). The valence band energy (E_v) of BAPI was calculated to be 5.70 eV by subtracting the width of the He I UPS spectra from the excitation energy [21.22 − (16.87 − 1.35)]. Given the E_g of 2.04 eV, the conductive-band minimum (E_c) is thus estimated at 3.66 eV from E_V − E_g. The E_V and E_c values of BAPI in electron volts are converted to electrochemical energy potentials in volts according to the reference standard, for which 0 V versus RHE equals—4.44 eV versus the vacuum level [49]. The actual loading of Pt cocatalyst in the BAPI/Pt composites was quantified gravimetrically. The nominal loading percentages of 0.25 wt% 0.5 wt% 0.75 wt%, 1 wt%, and 2 wt% corresponded to measured values of 0.22 wt%, 0.34 wt% 0.58 wt%, 0.83 wt%, and 1.12 wt%, respectively. The standard I₃⁻ solutions were prepared by diluting a standard iodine solution with 0.02 mol L⁻¹ of KI solution to the desired concentration. At certain time intervals, 0.5 mL of solution was withdrawn and diluted 1:50 with deionized water. After centrifuging to remove the deposited PbI₂, clear solutions were obtained and analyzed using a UV/Vis/NIR spectrometer. Crystallographic data for BAPI can be obtained from the CCDC under the following numbers: 1912877, 1912878, and 1912879.

3.4. Photocatalytic H₂ Evolution Measurements

Photocatalytic H₂ evolution reactions were carried out in a sealed gas-circulation system (CEL-SPH2N, CEAULIGHT, Beijing, China) using a quartz reactor with a top-irradiation configuration and an illumination area of 15.9 cm². In a typical experiment, 100 mg of BAPI microcrystals was suspended in 30 mL of saturated HI solution containing H₃PO₂ (volume ratio HI:H₃PO₂ = 4:1). During the photocatalytic measurements, the reactant solution was maintained at 298 K using circulating cooling water (Figure S12). For simultaneous light irradiation and ultrasound experiments, the reaction vessel was placed in an ultrasonic machine with a 2 cm gap between the vessel bottom and the cavity, while cooling water circulation was maintained. To avoid excessive temperature rise, ice bags were added to the ultrasonic machine every hour. The reaction solution was continuously stirred and maintained at 298 K using a water circulation system. Pt cocatalyst was deposited in situ by photoreduction. Specifically, 0.7964 g of H₂PtCl₆·6H₂O was dissolved in 200 mL of aqueous solution (H₂O:H₃PO₂ = 4:1) to yield a 1.5 mg/mL stock solution based on Pt content. Desired volumes of the Pt precursor (166 μL, 333 μL, 500 μL, 666 μL, and 1333 μL for 0.25 wt%, 0.5 wt%, 0.75 wt%, 1 wt%, and 2 wt% BAPI, respectively) were added to the BAPI suspension under vigorous stirring. The reactor was subsequently evacuated to −0.1 MPa to remove dissolved gases. Photoreduction was carried out under visible light irradiation (≥420 nm, 180 mW cm⁻²) for 1 h using a 300 W xenon lamp (CEL-HXBF 300, CEAULIGHT, Beijing, China). The resultant BAPI/Pt photocatalysts were used without further processing. Hydrogen evolution was monitored online using a gas chromatograph (GC-2060, thermal conductivity detector, Ar carrier gas). For cycling tests, the photocatalytic reaction was repeated in 4-h cycles, with complete re-evacuation and re-illumination between each cycle.

3.5. Calculations of AQYs

The AQY values of BAPI/Pt were determined under identical photocatalytic conditions, using monochromatic irradiation achieved with band-pass filters centered at 420, 450, 500, 550, 600, 630, and 650 nm. The illuminated area was fixed at 15.9 cm². Photon flux was quantified using a calibrated spectroradiometer (CEL-NP2000, CEAULIGHT). AQY values were calculated based on the amount of hydrogen produced over 30 min using the following equation:

$$\mathrm{AQY}={\displaystyle \frac{2\times thenumberofevolved{H}_2}{thenumberofincidentphotons}}\times 100\%$$

(1)

The factor of 2 accounts for the stoichiometry of hydrogen evolution (2 electrons per H₂ molecule). The calculated AQYs were used to evaluate the wavelength-dependent photoresponse of BAPI/Pt and to verify consistency with its optical absorption spectrum.

4. Conclusions

In summary, we demonstrate the application of 2D halide perovskite ferroelectric for photocatalytic H₂ evolution, combining robust P_s with a broad light absorption range and favourable semiconducting properties. Benefiting from the internal depolarization field, BAPI exhibits excellent H₂ evolution performance. Under simultaneous light and ultrasonic excitation, dynamic modulation of polarization further mitigates ionic screening, yielding a significantly enhanced photocatalytic activity. The BAPI/Pt system achieves a H₂ evolution rate of 1256 μmol g⁻¹ h⁻¹, over two times higher than under light-only conditions. Temperature-dependent phase transition experiments confirm the direct role of ferroelectricity in photocatalytic enhancement, as the nonpolar phase exhibits no such improvement. In addition, BAPI exhibits excellent stability and high AQYs, highlighting its promise as a platform material. This work not only illustrates the synergy between ferroelectricity and photocatalysis in 2D halide perovskites, but also opens new opportunities for designing ferroelectric semiconductors as high-efficiency photocatalysts for solar H₂ production.