Development of δ-Bi₂O₃/Bi₂SiO₅ heterostructures based on biogenic silica for photocatalytic treatment from organic pollutants

The use and management of solid and liquid waste are one of the most important activities of the rice industry nowadays. The reuse of waste makes it possible to achieve a cyclical process and obtain value-added products. In 2019, annual rice production amounted to about 784 million tons, according to the reports of the UN Food and Agriculture Organization (Tyulyagin 2022). Huge amounts of solid waste are generated at various stages of raw rice processing, 20% of which is rice husk that contains ∼20% of silica (Pode 2016). This waste can serve as a source of biogenic amorphous silicon dioxide, which can serve as a basis for developing materials with catalytic properties. However, this application area of biogenic silica is understudied since silica-based supports of catalysts are currently obtained from mineral silicon-containing commercial compounds (Hanna et al. 2008; Zhong et al. 2011; Bailiche et al. 2013; Espro et al. 2016; Cui et al. 2016).

Fenton-like catalysts and photocatalysts were related to the use of biogenic silica from rice production waste only in a few works (Adam & Andas 2007; Adam et al. 2010; Gan & Li 2013; Ghime & Ghosh 2017; Arefieva et al. 2021a). Fenton-like iron-containing catalysts based on amorphous silicon dioxide from rice husk were obtained by the sol–gel technique and used in oxalic acid (Ghime & Ghosh 2017), Rhodamine (Gan & Li 2013), and phenol (Adam et al. 2010) degradation reactions. Arefieva et al. (2021b) showed that Fe-containing Fenton-like catalysts, based on biogenic silica from rice husk obtained by impregnation, could be used to degrade phenol. Arefieva et al. (2021b) and Arefieva et al. (2020) studied the possibility of using Fenton-like catalysts for the degradation of lignin from alkaline hydrolysates of rice husk by exposure to UV and sunlight irradiation. Photocatalytically active Bi₂O₃/Bi₂SiO₅ oxide heterostructures were obtained by a simple method of Bi(NO₃)₃ mechanical stirring and amorphous silicon dioxide from rice husk without using any solvents in the study (Arefieva et al. 2021a). The resulting Bi₂O₃/Bi₂SiO₅ heterostructures exhibited photocatalytic activity in the degradation of methyl orange (MO) under sunlight and UV irradiation (Arefieva et al. 2021a).

In recent years, photocatalysts based on bismuth(III) compounds have been actively studied due to simple methods of preparation, non-toxicity, and thermal stability. Bismuth(III) oxide forms four crystalline modifications: monoclinic α-Bi₂O₃, tetragonal β-Bi₂O₃, cubic γ-Bi₂O₃, and δ-Bi₂O₃. The band gap for α-Bi₂O₃, β-Bi₂O₃, and δ-Bi₂O₃ is 2.80, 2.48, and 3.01 eV, respectively. According to photocatalytic activity, Bi₂O₃ modifications can be arranged in the following order: β-Bi₂O₃ > α-Bi₂O₃ > δ-Bi₂O₃ (Cheng et al. 2010). Bismuth metasilicate (Bi₂SiO₅) and Bi₂O₃/Bi₂SiO₅ heterostructural photocatalysts are promising materials. Such catalysts show high photocatalytic activity during the degradation of various organic pollutants in the sunlight (Zhang et al. 2010; Chen et al. 2009; Wei et al. 2013; Liu et al. 2014; Lu et al. 2018).

The Bi₂O₃–SiO₂ binary system can include three complex oxide compounds with different Bi/Si atomic ratios, namely bismuth metasilicate Bi₂SiO₅, bismuth orthosilicate Bi₄Si₃O₁₂, and sillenite Bi₁₂SiO₂₀ (Takamori 1990; Fei et al. 2000a, 2000b; Onderka et al. 2017). The compounds are arranged in the following order according to the degradation rate of methyl violet under irradiation with visible light: Bi₂SiO₅ > Bi₄Si₃O₁₂ > Bi₁₂SiO₂₀ (Chen et al. 2017). The photocatalytic activity of bismuth silicates depends on the morphology, particle size, and the ratio of the initial components, which, in turn, depend on the method of their synthesis.

Bismuth silicate with various morphologies can be obtained by hydrothermal (Wei et al. 2013), solvothermal (Dou et al. 2020a), template (Zhang et al. 2010; Liu et al. 2014; Dou et al. 2020b), and sol–gel (Zhang et al. 2010; Chen et al. 2013; Wu et al. 2017) methods. Among all the considered methods, the sol–gel method is the most accessible since it does not require additional equipment. The method is used for the preparation of nano- or submicro-powders (Wu et al. 2017) or thin film deposition (Chen et al. 2013; Wu et al. 2017). For example, Bi₂SiO₅ with nano- and submicro-particles is synthesized by the Pechini sol–gel method (the citrate method; method of polymer complexes) (Wu et al. 2017). All studies use commercial silicon-containing compounds. Biogenic silica is not used for synthesizing bismuth silicates. The advantage of biogenic silica is that it is a simple way to obtain its amorphous form, a quick transition from sol to gel with the formation of composites.

The aim of this work is to obtain Bi₂O₃/Bi₂SiO₅ heterostructure photocatalysts from bismuth nitrate and biogenic silica from rice husk using the sol–gel method and study the effect of bismuth and silicon oxide ratios on the composition, structure, and photocatalytic activity in MO degradation reaction of the as-obtained composites.

In this work, photocatalysts with a theoretical mass content of Bi₂O₃ ranging from 10 to 60% (Х-Bi-Si, Х = 10–60% Bi₂O₃) were obtained by a procedure proposed by Karthika et al. (2019).

The bismuth nitrate pentahydrate Bi(NO₃)₃·5H₂O was used as a precursor for the synthesis of Bi₂O₃. This reagent was pre-dried at 60 °C for 2 h, followed by dissolution by heating in a minimum volume of concentrated nitric acid. Minimum volume is required to dissolve a given substance at a given temperature. In the present work, it was determined experimentally.

Biogenic silica (the content of silicon dioxide is ∼99%) was obtained from rice husk according to a method described in the study (Zemnukhova et al. 2006). The rice husk was obtained in the production of rice of the Dolynny variety of the Far Eastern selection (Ussuriysk, Russia). The husk sample was meshed (∼8 mesh) to remove tiny fractions (bran siftings and dust). The raw materials were washed in distilled water, air-dried, and then hydrolyzed by 0.1 M hydrochloric acid by heating to 90 °С within an hour in a laboratory reactor. The temperature was controlled using an EKT Hei Con thermocouple (Heidolph, Germany). The mass ratio of solid to liquid was 1:13. Fibrous residue was filtered through a synthetic fabric filter with a pore size of 15 μm, repeatedly washed with distilled water, and air-dried at t = 25 °C for 48 h. Then, oxidative calcination was performed in a WiseTherm muffle furnace (DAIHAN, South Korea) at 600 °C for 3 h. The result was amorphous silicon dioxide in the form of a white powder. This powder was dissolved in a minimal volume of 1 M NaOH solution at 60 °C to obtain a sodium orthosilicate solution.

To form Х-Bi-Si samples, the Bi(NO₃)₃ dissolved in HNO₃ was added to the sodium orthosilicate solution with constant stirring to pH 6. Masses of the reagents and designations of the photocatalysts are given in Table 1S. Sol was formed while stirring, and this led to a white gel formation during a pH transition. The resulting gel was left to age in a mother liquor for 24 h. The solid precipitate was filtered, washed with distilled water until a neutral reaction, and then dried at 25 °C temperature for 48 h. The resulting powder was calcined in a muffle furnace WiseTherm (DAIHAN, South Korea) to 600 °C for 2.5 h and kept at this temperature for 4 h. The samples were designated depending on the weight content of bismuth oxide in the Bi(NO₃)₃–SiO₂ mixture, i.e., the samples containing 10, 25, 40, 50, and 60% were designated as 10-Bi-Si, 25-Bi-Si, 40-Bi-Si, 50-Bi-Si, and 60-Bi-Si, respectively.

Elemental analysis of the samples was performed by energy-dispersive X-ray fluorescence (EDXRF) analysis on a Shimadzu EDX 800 HS spectrometer (Japan). IR absorption spectra were recorded within the range of 400–4,000 cm^–1 in potassium bromide on a Bruker Vertex 70 Fourier-transform spectrometer (Germany) to determine functional groups in the studied samples. X-ray powder diffraction (XRD) analysis of the samples was performed using СuK_α radiation on a Bruker D8 Advance diffractometer (Germany). Morphology and elemental composition of the catalysts' surfaces were studied using the Carl Zeiss Ultra+ scanning electron microscope (SEM) (Germany) equipped with a Thermo Scientific add-on unit from Oxford instruments (England) for energy-dispersive X-ray (EDX) analysis. Optical absorption ability and band gap studies were carried out by ultraviolet–visible diffuse reflectance spectroscopy (DRS) on a Shimadzu UV2600 spectrophotometer (Japan) with wavelengths between 200 and 800 nm using BaSO₄ as a reflectance reference. X-ray photoelectron spectra (XPS) were recorded on a Specs ultra-high vacuum photoelectron spectrometer (Germany) with a hemispherical electrostatic analyzer (radius of curvature 150 mm) and a MgK_α radiation source (source energy 1,253.6 eV), along with 10⁻⁶–10⁻⁷ Pa pressures in the analysis chamber. The spectra were processed using the CASA XPS program (CasaXPS Version 2.3.12). The calibration of the electron-binding energy scale was done using an internal standard technique for which C1s level (285.0 eV) was chosen. A spread function of the spectrometer in a mode of characteristic atomic-level registration, which was determined from Ag3d_5/2 band shape, had 1.2 eV half-width. Survey spectra were obtained with 50 eV pass energy and region spectra with 20 eV pass energy.

Currently, several methods have been proposed for determining the zero-charge point (pH): Frumkin-Atkinson, Zerensen-de Bruin, Parks-Bobyrenko, and Nechiporenko. In this paper, the Zerensen-de Bruin method was used (Noh & Schwarz 1989). The zero-charge point (ZCP) was determined by the following procedure. The studied X-Bi-Si catalysts, where X = 10–60% Bi₂O₃, were sequentially added in small portions (0.05 g) in certain time intervals (5–10 min) into 25 mL of distilled water in a potentiometric cell until constant values of glass electrode potential were reached.

Photocatalytic properties of the obtained samples were evaluated by an example of MO degradation (pH 6.8) reaction in the UV spectral region. The concentration of MO was 10 mg L^–1. The samples of X-Bi-Si, where X = 10–60% Bi₂O₃, were used as photocatalysts. Individual oxides, Bi₂O₃ and SiO₂ (biogenic silica), were used as control samples. The catalyst dosage was 1 g per 1 L of the MO solution. Photocatalytic decomposition in the UV region was conducted in a 100-mL quartz cell filled with 50 mL of the MO solution (10 mg L^–1) and 0.05 g of catalyst. The radiation source was a 100P/F UV lamp (radiation maximum is λ = 365 nm). The solution was irradiated at constant magnetic stirring (625 rpm) for 3 h.

The Bi content was determined using an atomic absorption spectrophotometer (AAS) SHIMADZU AA-7000 (Japan) equipped with a lamp with a hollow cathode containing bismuth. Conditions for determining bismuth were as follows: wavelength – 223.1 nm; bandwidth – 0.7 nm; lamp current – 10 mA; flame – 15.0 L min⁻¹ of air, and 2.2 L min⁻¹ of acetylene. All measurements were carried out without background correction. Detection limits of the element range from 0.15 to 20 mg L⁻¹.

Several experiments on the degradation of MO in the presence of a photocatalyst 50-Bi-Si were conducted to establish a mechanism of the decomposition process. 2 mmol of isopropanol (IPA), benzoquinone (BQ), hydrogen peroxide (H₂O₂), and ammonium oxalate ((NH₄)₂C₂O₄) were added into MO solutions to capture ^·OH (hydroxyl radicals), ^•O₂⁻ (superoxide radicals), electrons and holes, respectively. All solutions were kept under magnetic stirring in the UV spectral region, and further analyses were performed as described above.

Table 1 presents the theoretical and experimental chemical composition of the obtained samples. The theoretical composition was evaluated from the content of precursors in the reaction mixture, whereas an experimental one was determined from the EDXRF results followed by recalculation for oxides. The measured content of bismuth oxide in the as-obtained samples is higher than that of the theoretical one, apparently due to an incomplete transition of sodium orthosilicate to silicic acid sol with a change in solution pH. The samples contain sodium because gelation was carried out in an alkaline medium, which was created by sodium hydroxide. It should be noted that 60-Bi-Si contains a high amount of sodium since a large dosage of Bi(NO₃)₃ requires additional amounts of NaOH solution to be added into the mixture in order to adjust the pH of the medium to 6.

Table 2 shows the results of the EDX analysis of surface layers (approximately 1 μm) for the studied catalysts. It can be seen that bismuth content rises at the surface with an increase in the amount of its oxide added into the samples (from 10 to 50%). The 60-Bi-Si sample has a low content of bismuth (6%) and a rather high content of sodium on the surface. Thus, excess sodium hydroxide in the solution leads to the formation of the sample with low bismuth content and high sodium content both in volume and on its surface. In this case, sol–gel transition does not occur and mixed composite structures are not formed. In addition, silicon content on the surface of 50-Bi-Si and 60-Bi-Si is quite low, which is also due to the peculiarities of the gelation process.

Fig. 1S shows the Fourier transform infrared (FT-IR) result for the studied photocatalysts. FT-IR spectra of 10-Bi-Si and 25-Bi-Si samples are identical (Fig. 1Sa). In the FT-IR spectra, absorption bands correspond to bending (469 cm^–1) and stretching (symmetric and asymmetric) bands (804 and 1,099 cm^–1) of Si–O–Si siloxane bonds (Plyusnina 1976). With a higher content of bismuth oxide in the samples 40-Bi-Si (59.1%, Table 1), 50-Bi-Si (65.4%, Table 1), and 60-Bi-Si (42.9%, Table 1) (Figs. 1Sb and 1Sc), absorption bands appear at 1,385 and 619 cm^–1, specific to the Bi–O bond at 850 cm^–1, corresponding to Bi–O–Si vibrations (Dou et al. 2020a). In the samples 10-Bi-Si and 25-Bi-Si, the content of bismuth oxide is 22 and 31.7% (Table 1), respectively. The sample 40-Bi-Si contains 59.1% bismuth oxide and 50-Bi-Si contains 65.4% bismuth oxide (Table 1). It can be noted that FT-IR spectra of the 60-Bi-Si sample contain absorption bands at 1,640 and 3,450 cm^–1, indicating the presence of OH groups, which are caused by the presence of sorbed water (Fig. 1S).

According to XRD data (Table 3; Figs. 2S and 3S), 10-Bi-Si and 25-Bi-Si samples are in an amorphous–crystalline state, while 40-Bi-Si, 50-Bi-Si, and 60-Bi-Si are in a crystalline state. A diffuse peak corresponding to an amorphous structure is observed near 2θ ≈ 20° on XRD patterns of 10-Bi-Si and 25-Bi-Si samples. The samples contain cristobalite and 60-Bi-Si is a moganite. Crystallization of silica is due to a high sodium content. Bismuth silicate Bi₂SiO₅, which exhibits the highest photocatalytic activity among binary Bi₂O₃–SiO₂ systems, was formed in all the samples (Chen et al. 2017). 10-Bi-Si and 50-Bi-Si samples also contain photoactive cubic modification δ-Bi₂O₃. δ-Bi₂O₃ is one of the best-known oxygen ion conductors. The reason for extraordinarily high oxygen mobility in δ-Bi₂O₃ is more oxygen vacancies in the δ-phase naturally possesses (Dreyer et al. 2016).

Figure 1(b) shows a high-resolution XPS peak of O 1s. The O 1s spectrum can be deconvoluted into two peaks. The peak at 532.4 eV is connected with the O 1s orbital of lattice oxygen of SiO₂ (Miller & Linton 1985), while the peak at 530.0 eV is the characteristic of oxygen of the Bi–O bond (Chen et al. 2017; Lu et al. 2018). As demonstrated in Figure 1(c), the binding energies at 158.6 and 163.9 eV can be ascribed to Bi 4f7/2 and Bi 4f5/2, respectively. These values of binding energies might belong to Bi³⁺ in both Bi₂SiO₅ and Bi₂O₃ (Lu et al. 2018; Dou et al. 2020a).

Figure 1(d) depicts the Si 2p XP spectra; the binding energies of 103.6 and 102.0 eV are attributed to Si(IV) of SiO₂ (Xu et al. 2013) and Bi₂SiO₅ (Chen et al. 2017), respectively.

It is known that the initial stage of heterogeneous catalytic processes is the sorption of substances on the surface of a solid. Solution pH has a crucial effect on the adsorption capacity of dyes. A change in the solution pH can modify the surface charge of an adsorbent and a dye ionization degree. According to the results shown in Figure 3, the pH_pzc of the photocatalysts was specified to be more than 7. It means that in the studied MO solution (pH 6.8), the surface is positively charged. MO is an anionic dye and can maintain its anionic configuration in the pH range of 3–14 (pKa = 3.8). This contributed to an interaction of the positive surface of all the studied samples with anionic dye (Rekavandi et al. 2019; Khan et al. 2022).

The low photocatalytic activity of 10-Bi-Si and 25-Bi-Si materials can be explained by their amorphous state (Table 3, Fig. 2S) and a high proportion of silicon dioxide (75 and 67%, respectively, (Table 1)), which, as a typical dielectric, blocks the generation of electron-hole pairs. The appearance of photocatalytic activity in 40-Bi-Si and 50-Bi-Si samples is associated with their transition to crystalline state (Table 3; Figs. 2S and 3S) and the formation of crystals on the surface (Figure 2(f) and 2(h)). The higher photoactivity of the 50-Bi-Si sample compared to 40-Bi-Si is explained by the formation of a two-phase system: δ-Bi₂O₃ and Bi₂SiO₅. In this system, Bi₂SiO₅ is a typical n-type semiconductor, and δ-Bi₂O₃ is a p-type semiconductor. This contributes to the formation of the p–n heterostructure and, as a consequence, efficient separation of photogenerated holes and electrons (Lu et al. 2018). The 60-Bi-Si sample has an insignificant photocatalytic activity due to an undeveloped surface and a low bismuth content (Table 2, Figure 2(j)).

The plots of against are shown in Figure 5(b) and 5(c). The analysis shows that these dependences are characterized by the presence of two linear segments, which correspond to different electronic levels allowing us to absorb photons. The E_g values are 2.87 eV (n= 2) and 3.80 eV (n= 1/2) for δ-Bi₂O₃ and 3.29 eV (n= 2) and 3.89 eV (n= 1/2) for Bi₂SiO₅. The obtained results are consistent with the reported values (Agasiev et al. 1986; Cheng et al. 2010; Dreyer et al. 2016; Gu et al. 2018).

From the plot of (ahν)²–hν (Figure 5(c)), we estimated that band gaps of Bi₂SiO₅ and δ-Bi₂O₃ were 3.89 and 3.8 eV, respectively. X values for Bi₂SiO₅ and Bi₂O₃ were 6.12 and 5.91 eV, respectively. Valence band positions of δ-Bi₂O₃ and Bi₂SiO₅ were 3.31 and 3.56 eV, respectively. Conduction band positions of δ-Bi₂O₃ and Bi₂SiO₅ were −0.48 and −0.33 eV, respectively.

When p-type δ-Bi₂O₃ and n-type Bi₂SiO₅ contacted each other, a p–n heterojunction was formed at an interface of a phase (Lu et al. 2018). In this process, energy levels of conduction (CB) and valence (VB) bands of p-type δ-Bi₂O₃ tended to ascend, while levels of n-type Bi₂SiO₅ tended to descend until a Fermi-level equilibrium was reached (Low et al. 2017). Because of the formation of the p–n heterojunction, a region close to the p–n interface was charged, creating an internal electric field (Low et al. 2017; Sun et al. 2017).

Under UV-light irradiation, both Bi₂SiO₅ and δ-Bi₂O₃ can absorb photons giving rise to electrons and holes. Because the CB of δ-Bi₂O₃ (−0.48 eV) is more negative than O₂/^•O₂⁻ potential (−0.33 eV vs. normal hydrogen electrode (NHE)), photogenerated electrons can react with O₂ to generate ^•O₂⁻, which is an active species with strong oxidizing ability and can react with MO. Moreover, a high position of CB of δ-Bi₂O₃ and internal electric field at the interface lead to a transfer of photogenerated electrons from δ-Bi₂O₃ to Bi₂SiO₅. Meanwhile, photogenerated holes from a valence band of Bi₂SiO₅ can be easily transferred to a valence band of δ-Bi₂O₃ and then directly reacted with MO. Therefore, photogenerated holes and electrons can be separated effectively and photocatalytic activity is enhanced a lot.

Valence band positions of δ-Bi₂O₃ (3.31 eV) and Bi₂SiO₅ (3.56 eV) are more positive than those of OH^•/H₂O (2.27 eV) and OH^•/OH⁻ (1.99 eV) potentials before connection, indicating that ^•OH can be generated by the oxidation reaction of h⁺.

After the p–n heterojunction formation, E_VB of δ-Bi₂O₃ slightly reduced but its value remained favourable for a sufficient oxidation capability for OH^•/H₂O and OH^•/OH⁻ conversions. Therefore, MO can be degraded by the O₂/^•O₂⁻, h⁺, or OH^• oxidation pathway.

To disclose the photocatalytic mechanism of the as-obtained 50-Bi-Si photocatalyst, reaction rates were determined in the presence of reactive species (^•O₂⁻, h⁺, e⁻, or ^•OH) involved in the photocatalytic decomposition process. IPA, BQ, H₂O₂, and (NH₄)₂C₂O₄ were introduced as ^•OH, ^•O₂⁻, electrons, and holes, respectively.

Usually, the kinetics of the photocatalytic degradation of organic pollutants is determined using the Langmuir–Hinshelwood model since it assumes that a degradation reaction occurs between generated reactive oxygen species and adsorbed pollutant molecules. As the second-order kinetic model typically fits the Langmuir–Hinshelwood model, it can be simplified to a pseudo-first-order kinetic model by considering that the adsorbed amount of pollutant molecules is independent of its concentration in a medium (Mota et al. 2020).

As shown in Figure 7, ^•O₂⁻ and e⁻ participate in a degradation process. Electrons play a dominant role, and ^•O₂⁻ takes a second role since the rate constant of the MO degradation reaction in the presence of hydrogen peroxide is three times less (Table 2S). H⁺ and ^•OH do not play a role in the degradation process. Reaction rate constants in the presence of these scavengers do not differ from each other and the MO degradation constants in the absence of scavengers.

The AAS method was used to determine bismuth concentration in aqueous solutions of MO after the third test cycle to assess the stability of the photocatalyst in aqueous solutions. Bismuth content in the MO solution after each photocatalysis cycle did not exceed 0.2 mg L^–1 (0.2 mg L^–1 is the limit of detection of bismuth by the AAS method). Thus, the obtained results indicate that the studied catalysts are stable in aqueous solutions. The content of bismuth does not exceed 0.2 (0.2 is the limit of detection of bismuth)

Photocatalytically active δ-Bi₂O₃/Bi₂SiO₅ heterostructures were synthesized by the sol–gel method using bismuth nitrate and biogenic silica produced from rice husk. The influence of the content of bismuth oxide in the samples on their composition, structure and photocatalytic properties was shown.

It was established that the obtained samples contain photoactive Bi₂SiO₅ and, in some cases, δ-Bi₂O₃ and Bi₁₂SiO₂₀. The as-obtained samples exhibit photocatalytic activity in the oxidation of MO under UV irradiation. Extensive characterization revealed the relationship between the content of bismuth oxide in materials and their photocatalytic properties. A degree of UV-light-assisted degradation of MO degradation in a neutral medium (pH 6.8), in the presence of the samples containing 59–65% Bi₂O₃, reaches 22–45% in 60 min. Hence, enhanced photocatalytic performance can be ascribed to the formation of δ-Bi₂O₃ and, probably, the fabrication of δ-Bi₂O₃/Bi₂SiO₅ heterojunction. The as-obtained photocatalysts remain stable in aqueous solutions, and bismuth leaching after photocatalytic tests does not exceed 0.2 mg L^–1. This study evaluated the roles of different reactive oxygen species on the degradation kinetics of MO on Bi₂O₃/Bi₂SiO₅ using radical scavengers. It was found that electrons were the main active species for the degradation of MO by UV irradiation, whereas ^•O₂⁻ played a supporting role during the photocatalytic process.

Thus, the heterostructures obtained by a simple sol–gel method using biogenic silica are environmentally acceptable, easily separated from the solution, and exhibit photocatalytic activity without acidification and the addition of oxidants. Based on the mentioned above, the as-obtained δ-Bi₂O₃/Bi₂SiO₅ samples are promising photocatalysts for water treatment from persistent organic pollutants in natural conditions.

This work was supported by the state task of the Institute of Chemistry of the FEB RAS FWFN(0205)-2022-0002 (topic 1, section 3); the UV–Vis studies were performed with the financial support of the Russian Science Foundation under Grant number 22-23-00912. The study of the morphology of the surface of the samples was carried out on the equipment of the Center for Collective Use of the Far Eastern Federal University No. 200556 (Vladivostok).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.