Decorating TiO₂ Nanoparticle Thin Film with SnS_x (x < 1): Preparation, Characterization, and Photocatalytic Activity

Abstract

We report a study on the SnS_x (x < 1) decoration of porous TiO₂ nanoparticle thin films using the ionic layer adsorption and reaction (ILAR) method. UV-vis absorption measurements revealed a direct bandgap of 1.40–2.10 eV for SnS_x (with x = 0.85) and 3.15 eV for TiO₂. Degradation of rhodamine B molecules in aqueous solutions shows that coating with a Sn-to-Ti molar ratio of 2% improves the efficiency of the photocatalytic performance of titanium dioxide, but excessive coverage decreases it. We interpret the observed behavior as due to a delicate balance of many competing factors. The formation of intimate interfaces guaranteed by the ILAR growth technique and a nearly optimal alignment of conduction band edges facilitate electron transfer, reducing electron–hole recombination rates. However, the valence hole transfer from TiO₂ to SnS reduces the oxidative potential, which is crucial in the degradation mechanism.

1. Introduction

Water is utilized in innumerable chemical processes in biochemical, pharmaceutical, petrochemical, and other industries. Degrading and decomposing various organic and inorganic contaminants in the produced wastewater is paramount for human health and environmental concerns. As a sustainable and unselective process, heterogeneous photocatalysis employing solar light as a photon source has gained increasing attention in recent decades [1,2,3,4]. Titanium dioxide is one of the most widely used photocatalysts. Upon absorption of ultraviolet (UV) light with photon energy larger than the band gap value (3.2 eV), electron–hole pairs are generated. If they survive recombination, they can migrate to the surface and produce hydroxyl radicals and superoxide anions through redox reactions involving water and oxygen molecules. These oxidative and reductive species can decompose and mineralize various pollutants, such as residuals of medicines, pesticides, herbicides, dyes, phenols, hydrocarbons, and heavy metals, into CO₂, H₂O, and other simpler and harmless compounds [1,2,4,5].

Extensive research has focused on broadening solar light absorption range and diminishing electron–hole recombination rate to improve the photocatalytic efficacy of TiO₂. Doping TiO₂ with foreign species can reduce the bandgap, while creating appropriate nano-heterojunctions can facilitate charge transfer [2]. In this context, SnS is one of the best choices among IV–VI semiconductors. It possesses a bandgap of only 1.1–2.1 eV and an electron affinity similar to TiO₂ [6].

Synthesis of SnS nanoparticles can employ different methods, such as precipitation, solvothermal, microemulsion, and (successive) ionic layer adsorption and reaction ((S)ILAR) [7,8,9,10]. The most common and straightforward route to producing thin films is wet chemistry [11,12]. A supported thin film is necessary for practical applications of SnS@TiO₂ nanoparticles as photocatalysts. One effective method for forming such nano-heterojunctions with intimate contact involves coating a porous TiO₂ nanoparticle (NP) thin film with Sn²⁺ and S²⁻ precursor solutions using the SILAR method, followed by post-process annealing [13,14,15].

In this study, we report on the fabrication and characterization of SnS_x-decorated TiO₂ NP thin films and their photocatalytic activity toward rhodamine B (RhB) molecules, taken as a representative organic material. We demonstrate that, by selecting appropriate S²⁻ and Sn²⁺ precursor solution concentrations, it is possible to improve the photocatalytic performance of TiO₂.

2. Materials and Methods

TiO₂ nanoparticles (P25) were purchased from Evonik (Essen, Germany). All chemicals, including polyethylene glycol (PEG-20000, Merck, Taufkirchen, Germany), Triton X-100 (Union Carbide Company, Houston, TX, USA), acetylacetone (>99%, Sigma-Aldrich/Merck, St. Louis. MI, USA), anhydrous stannous chloride (SnCl₂, 98%, Alfa Aesar/Thermo Scientific, Waltham, MA, USA), sodium sulfide nonahydrate (Na₂S·9H₂O, Sigma-Aldrich/Merck), and rhodamine B (Thermo Scientific), were used as received without further purification.

TiO₂ sol (5 g in 25 mL distilled water), added with PEG-20000, acetylacetone, and Triton X-100, was sonicated for an extended period to break up large aggregates. Spin coating the sol onto a 0.15μm fluorine-doped tin oxide (FTO)-covered glass substrate (2.5 × 2.5 cm²) and subsequent calcination at 450 °C for 1 h yield a uniform thin layer. Further details can be found elsewhere [16]. Repeating the procedure several times could produce films with varying desired thicknesses. TiO₂ layers, with an average coverage of 2.2 $\pm$ 0.2 mg cm⁻², were routinely grown.

The coating of TiO₂ layers with SnS_x was achieved using the ILAR method. In a typical reaction scheme, the sample was first immersed in the SnCl₂ aqueous solution with a molar concentration of [Sn²⁺] = Snμμμ (in mM), then rinsed and submerged into distilled water to remove unbounded Sn²⁺ ions, immersed in the Na₂S solution with a concentration of [S²⁻] = Sννν (in mM), and again washed. The immersion time was typically 30 s for each, unless otherwise specified, and the washing lasted 3 min. The so-prepared samples follow the SnμμμSννν nomination scheme. Both precursor solutions were magnetically stirred and kept at 75 °C. This adsorption and reaction process could be reiterated for several cycles. The samples were immediately heated at 180 °C in an oven with a nitrogen flow of 800 mL/min for 90 min. The SnS_x@TiO₂ films were brownish, but the hue depended on the amount and the composition. For heavily coated films, the color was nearly black.

Scanning electron microscopy (SEM) was performed with an ultra-high-resolution ZEISS Cross-Beam 350 (Zeiss, Oberkochen, Germany). For energy dispersive X-ray analysis (EDX), a beam of 15 keV was used to analyze a large surface area of 1.5 × 2.0 mm². X-ray diffraction (XRD) was employed to investigate the crystalline structure of the NP thin films using a Bruker D8 Advance diffractometer with Cu Kα radiation (λ = 1.5406 $Å$) (Bruker, Billerica, MA, USA).

X-ray photoelectron spectroscopy (XPS) was performed in a UHV chamber with a base pressure of 1 × 10⁻⁹ Torr utilizing a hemispherical electron energy analyzer with a constant pass energy of 100 eV. The excitation source was the Cu Kα emission. We calibrated the binding energy (BE) scale by setting graphite’s C 1 s peak to 284.3 eV and corrected the spectra for the analyzer transmission factor.

The absolute total UV-Vis-NIR (200–1250 nm) diffusive reflectance and transmittance were measured with a Cary-5E hemispherical spectrophotometer. According to the Tauc method, it is possible to determine the band gap energy E_g of a semiconductor with the following equation:

$${\left(Ah\nu \right)}^{\beta }=B(h\nu -E_g)$$

(1)

with A being the absorbance, B a constant, hν the photon energy, and β = 2 for direct band transitions.

We assessed the photocatalytic performance of the SnS_x@TiO₂ NP thin films by monitoring the change in concentration of RhB in an aqueous solution. The reaction occurred in a 600 mL glass beaker with an 8.2 cm inner diameter. The reaction cell contained a magnetically stirred 50 mL RhB solution with a concentration of 1 × 10⁻⁵ M and a pH of 8.5. A 300 W Xe lamp served as the source of photons, and the light, diaphragmed and incident through a JS1-type UV-transparent quartz cover, irradiated only the sample surface (ϕ = 2 cm). The radiance flux density was 21 mW cm⁻², measured with a power meter, and the optical path was 7 mm.

To determine the concentration of RhB using the Beer–Lambert method, an aliquot of 1.5 mL of the solution was withdrawn at specific time intervals to measure the absorbance spectrum with a UV-vis spectrometer. The transmittance signal in the 553–555 nm range was integrated, with water as a reference. After the analysis in less than 1 min, the sampled solution was promptly returned to the reaction cell.

3. Results

Figure 1 illustrates SEM images of a pure TiO₂ thin film (A and C) and one coated with SnS_x Sn100S600 (B and D). The TiO₂ film was uniform and porous. PEG-20000 contained in the sol evaporated during the sintering, leaving large pores in the film. The particle size ranged from approximately 14 nm to 30 nm, though many particles formed aggregates. The coating of SnS_x using the ILAR method did not make any significant morphology changes except for a slight increase in the NP size due to the intimate growth of tiny SnS on TiO₂.

The cross-sectional SEM images, as a typical one shown in Figure 2, revealed an average film thickness of about 11 μm. The TiO₂ layer mean coverage was 2.2 $\pm$ 0.2 mg cm⁻². By using the density of single crystalline TiO₂ of 4.23 g cm⁻³, we could easily estimate the thickness of a compact film as 2.2 mg cm⁻²/4.23 g cm⁻³ = 5.3 × 10⁻⁴ cm = 5.3 μm. Consequently, the film porosity was roughly 50%.

As illustrated in Figure 3, EDX detected no Sn and Si signals for pure TiO₂, sample Sn000S000, indicating that the film was thick enough to shield the underlying FTO-glazed glass substrate. For sample Sn010S060, the measured atomic ratio between Sn and Ti, χ, was 0.03, whereas that between S and Sn, x, was 0.85. For Sn100S600, χ increased to 0.08 but x remained the same.

We explored the impact of the concentration ratio of precursor solutions on the characteristics of SnS_x NPs, keeping the [Sn²⁺] constant (10 mM) and varying [S²⁻] from 10 to 100 mM. The results, summarized in

Table 1. SnS nanoparticles formed with a constant [Sn²⁺] = 10 mM and various [S²⁻].

	Sn-to-Ti Ratio χ ± 0.01	S-to-Sn Ratio x ± 0.05	Direct Band Gap Eg(SnSx) (eV) ± 0.05 eV
Sn010S010	0.03	0.20	2.10
Sn010S020	0.03	0.40	2.05
Sn010S040	0.03	0.60	2.00
Sn010S060	0.03	0.85	2.00
Sn010S100	0.03	0.85	2.00

, show that as the ratio α = [S²⁻]/[Sn²⁺] increased from 1 to 6, more S²⁻ ions reacted to the adsorbed Sn²⁺, resulting in a rise of x from 0.20 to 0.85. The lack of perfectly stochiometric SnS nanoparticle formation aligns with the findings of Nengzi et al., likely due to the oxidation of some Sn²⁺ ions either during the formation process or during the heating stage [13]. We observed a similar trend when extending the immersion time in the S²⁻ precursor solution, maintaining the same molar concentrations [S²⁻] = [Sn²⁺] = 10 mM. However, the maximum value of x was only 0.75. In the following, we focus on the cases of α = 6 for several [Sn²⁺] and the same immersion time (30 s), for which x remained constant at 0.85 $\pm$ 0.05. These SnS_x NPs could be considered either SnO-doped, SnO₂-doped, or mixed.

Figure 4 displays the XRD patterns of various SnS_x@TiO₂ films. The bottom-most curve is of P25 TiO₂ thin film annealed at 450 °C for 1 h. It mainly contains anatase polymorph, about 10% of the rutile phase, and the SnO₂ of the underlying FTO. Decoration with Sn020S120 did not yield discernible SnS structures (curve b). However, in the case of Sn100S600, features in 2θ between 30 and 32° became visible (curve c). Heavy SnS_x decoration using Sn100S600 (5 cycles) resulted in two well-distinct peaks at about 2θ = 30.8° and 31.8°, identified as SnS (curve d).

In Figure 5, we present a set of XPS spectra of Ti 2p, O 1s, Sn 3d, and S 2p levels for pristine TiO₂ (Sn000S000) and SnS_x-decorated TiO₂ (Sn100S600) films. The line shape of the Ti 2p spectrum did not show any appreciable changes upon the decoration of SnS_x. The same was true also for O 1s. The Sn 3d_5/2 and 3d_3/2 peak energies were 486.5 eV and 495 eV, respectively, and the S 2p BE was 261.2 eV. These XPS spectra confirm the formation of ZnS [13,15,17,18].

The upper panel of Figure 6 shows a series of absorbance curves, with α = 6, obtained using UV-vis total diffusive reflection and transmission spectroscopy for various NP thin films. In the visible range, absorption gradually increased as more SnS_x NPs coated on TiO₂, broadening the absorption spectrum. The lower panel displays the Tauc plot of (Ahν)² as a function of the photon energy hν for determining the direct bandgap energies E_g of the SnS_x@TiO₂ coating layers. The curve for a pure TiO₂ film yielded E_g(TiO₂) = 3.15 eV, which agrees with previous reports. It was still possible to infer this bandgap value even after SnS_x decoration. The absorption of the FTO-covered glass substrate was null for hν < 4 eV. All Tauc curves have been vertically offset for clarity. No absorption from SnS_x was discernible for Sn001S006. With increasing precursor concentration, E_g(SnS_x) decreased from 2.10 to 1.40 eV (see

Table 2. NPs formed with a constant [S²⁻] to [Sn²⁺] precursor concentration ratio of α = 6 (S-to-Sn ratio x = 0.85 $\pm$ 0.05).

	Sn-to-Ti Ratio χ ± 0.01	Direct Band Gap Eg(SnSx) (eV) ± 0.05 eV
Sn000S000	0.00	-
Sn001S006	0.01	-
Sn005S030	0.02	2.10
Sn010S060	0.03	2.00
Sn020S120	0.04	1.90
Sn100S600	0.08	1.40

).

We investigated the effects of decorating TiO₂ with SnS_x on the photocatalytic activity of thin films. The upper panel of Figure 7 depicts the time evolution of the RhB absorption for sample Sn005S030. The middle panel illustrates the evolution of RhB concentration change over time C(t)/C(0) under Xe lamp illumination for α = 6 (x = 0.85). The figure also shows the direct photolysis curve for a bare glass sample. At low coverages, the SnS_x coating seemed to speed up the degradation process compared to the pure TiO₂ film. However, at higher concentrations, there was a decrease in photocatalytic efficiency. It is important to note that the concentration change follows zero-order kinetics. The linear fitting of C(t)/C(0) − t yielded the RhB degradation rate constant κ. In the lower panel of Figure 7, we plot κ versus the S-to-Sn ratio χ, whose highest value corresponded to x = 2%.

4. Discussion

The XRD, EDX, XPS, and UV-vis absorption results all confirm that the synthesized thin films using the ILAR method were of SnS_x@TiO₂ NPs. The large E_g(SnS_x) values observed at small χ can be related to the quantum size effect of the SnS_x NPs [6,19]. The continuous evolution of the bandgap energies reported in Table 2 supports this interpretation.

It is well known that a homogeneous RhB solution exhibits considerably high stability under visible light excitation [20]. Photolysis occurs only under UV irradiation but with a negligible quantum efficiency [20]. This explains why the direct photolysis of this dye is a highly inefficient process (see Figure 7).

We roughly estimate the conduction band minimum (CBM) and the valence band maximum (VBM) of SnS by using the following empirical formula:

CBM(SnS) = X(SnS) − E_g(SnS)/2 + E₀

(2)

VBM(SnS) = CBM(SnS) + E_g(SnS)

(3)

The Mulliken electronegativity, X(SnS) = 5.17 eV, is estimated as the geometric mean of the Sn and S electronegativities, with values of 4.3 and 6.22 eV, respectively. E₀ = −4.44 eV is the vacuum level relative to the normal hydrogen electrode (NHE). With a bandgap energy of E_g(SnS) = 2.00 eV for SnS, the CBM(SnS) and VBM(SnS) potentials are approximately −0.27 V and 1.73 V, respectively, on the NHE scale. When the bandgap energy is reduced to 1.40 eV, the potentials shift to 0.03 V and 1.43 V, agreeing with previously reported values [10,21]. Anatase TiO₂ has a CBM around −3.95 V relative to the vacuum level or −0.50 V relative to the NHE [22].

Figure 8 schematically depicts the potential diagram of the VBM and CBM of both TiO₂ and SnS, the ground state and the first excited state of rhodamine B, *OH radical, as well as the relevant redox reactions. It is worthwhile mentioning that the bandgap energy of SnS reported in the literature varies for several tenths of eV, depending on the fabrication procedure and other factors [6].

In the presence of a semiconductor catalyst, the photo-induced degradation process can start through two primary reaction pathways: photoexcitation of either adsorbed dye molecules or semiconductor photocatalyst. When a RhB molecule on the surface absorbs light, the excited electron may recombine back to the ground state. Alternatively, it may transfer to an adjacent adsorbed oxygen molecule, forming an *O₂⁻ superoxide ion, or to the conduction band of the semiconductor catalyst. Due to the proximity of the *O₂⁻, the excited electron has a high probability of transferring back to the (RhB)⁺ ion, leading to the regeneration of the dye (O₂ sensitization) [23]. If the charge injected in the conduction band diffuses away, then (RhB)⁺ can initiate the N-deethylation and subsequent decomposition [23]. It is important to note that the excitation of RhB is a resonance process with the maximum cross-section being at 564 nm so that only light with a wavelength in a very narrow range can be absorbed. The fact that the photodegradation of RhB by TiO₂ under visible light illumination is negligible indicates that this mechanism is not operative for TiO₂.

In photocatalysis, when a photon with energy greater than the band gap is absorbed, an electron in the valence band is excited into the conduction band. If the photogenerated charges survive the recombination, the electron can migrate to the surface and diffuse in the semiconductor matrix, reducing an adsorbed oxygen molecule to a superoxide anion. Meanwhile, the hole left in the valence band can either directly oxidize a dye molecule or react with an adsorbed OH⁻ or H₂O molecule, generating a strong oxidant hydroxyl radical *OH [5]. *O₂⁻ and *OH are fundamental components in the degradation and decomposition mechanisms of nearly all organic and inorganic pollutant molecules [1,2,5]. The specific pathways and mechanisms for photodegradation and decomposition of RhB have been well described in the literature [20,23,24].

SnS has a bandgap in the visible range, allowing it to absorb a significant portion of the solar light spectrum. Its photocatalytic activity has been extensively investigated for various dyes and pollutant molecules [13,14,15,17]. The efficacy of photocatalysis depends on several critical factors, including the actual band gap energy, the positions of band edges relative to the relevant redox potentials, the rate of electron–hole recombination, and the energy levels and degradation pathways specific to the molecules involved. Additionally, the performance is influenced by the crystalline structure, morphology, loading amount, particle size, and specific surface area of the photocatalyst nanoparticles, as well as factors like pollutant concentration, solution pH, and the spectrum and intensity of the light source. Comparing its performance directly with that of a wide bandgap semiconductor like TiO₂, even under very similar experimental conditions, may provide limited insights due to the significant differences in their photocatalytic mechanisms and properties.

Zhang et al. synthesized crystalline SnS nanoparticles using an anaerobic reaction method and found limited photoactivity toward RhB dye under visible light [25]. Mittal et al. fabricated TiO₂ and SnS nanoparticles with the precipitation technique. They observed that, in the presence of either TiO₂ or SnS suspended nanoparticles, RhB concentration follows pseudo-first-order kinetics and that TiO₂ exhibited higher photocatalytic activity under Xe UV-vis light compared to SnS [17]. In Figure 7, we observe similar trends in the degradation behavior of RhB upon Xe light illumination. We also notice differences in the reaction kinetic order compared to Mittal et al.’s study.

As indicated in the potential diagram of Figure 8, the VB edge of TiO₂ lies deeper than SnS. Therefore, the oxidizing power of a valence hole for RhB is stronger in TiO₂ than in SnS. Likewise, a VB hole in TiO₂ can react with an adsorbed water molecule, producing a hydroxyl radical, but this reaction route is not allowed in SnS. It is worthwhile mentioning that in our case, photocatalysis occurs only for the dye molecules adsorbed on the top few layers of the thin film. The reaction is active adsorption site limited, following pseudo-zero-order kinetics, and requires a longer irradiation time to reduce the concentration of RhB.

TiO₂ and SnS have CB edge potentials quite close to each other. The formation of nano-heterojunction with intimate contacts facilitates electron transfer across the interface from one semiconductor to another, effectively reducing the electron–hole recombination rate. However, hole transfer from the VB of TiO₂ to that of SnS has dual effects: it enhances charge survival but reduces oxidizing potential. The balance of these two competing effects depends on the details of the band alignment.

We attribute the enhancement of the photocatalytic activity of a small amount of SnS-decorated TiO₂ mainly to reduced electron–hole recombination rates in both semiconductors. With increasing χ, more photocatalysis occurs on the SnS surface, which has a lower efficiency than that on TiO₂. The reduction in SnS bandgap energy increases the generation of electron–hole pairs, but the upward shift of the VBM decreases the oxidative power of holes to some extent.

Studies conducted by Jiang et al. on SnS@TiO₂ nanobelts and by Yang et al. on amorphous SnS@TiO₂ NPs also showed improvements in the photocatalytic efficiency of various dyes compared to pure TiO₂ under different light illuminations [15,24]. Mittal et al. reported similar behavior of RhB bleaching rates with various χ as that observed in Figure 7 [17].

These findings collectively describe the complex interplay of semiconductor properties, coating ratios, and experimental conditions in determining photocatalytic efficacy for pollutant degradation.

5. Conclusions

We have coated porous TiO₂ nanoparticle thin films with SnS_x and characterized their properties with SEM, XRD, XPS, and UV-vis absorption techniques. We found that the effectiveness of SnS_x@TiO₂ nanoparticle thin films for photocatalytic applications hinges on finding a balance in the SnS_x content. Decoration with a 2% molar concentration of SnS_x enhances the photocatalytic activity towards rhodamine B, whereas excessive SnS_x reduces efficiency.

The observed photocatalytic behavior suggests a delicate balance of several competing factors. The ILAR growth technique ensures intimate contact at interfaces. Optimal alignment of CBMs can facilitate electron transfer, reducing electron–hole recombination rates and promoting the utilization of more photogenerated charges. However, when a hole transfers from the VB of TiO₂ to that of SnS, it decreases the oxidative potential, which is crucial for the pollutant degradation process.