In Situ Decoration of ZnSnO₃ Nanosheets on the Surface of Hollow Zn₂SnO₄ Octahedrons for Enhanced Solar Energy Application

Abstract

Hierarchical ZnSnO₃/Zn₂SnO₄ porous hollow octahedrons were constructed using the method of combining the acid etching process with the in situ decoration technique for photovoltaic and photocatalytic applications. The composite was used as photoanode of the dye-sensitized solar cells (DSSCs), an overall 4.31% photovoltaic conversion efficiency was obtained, nearly a 73.1% improvement over the DSSCs that used Zn₂SnO₄ solid octahedrons. The composite was also determined to be a high-performance photocatalyst for the removal of heavy metal ion Cr (VI) and antibiotic ciprofloxacin (CIP) in single and co-existing systems under simulated sunlight irradiation. It was remarkable that the composite displayed good reusability and stability in a co-existing system, and the simultaneous removal performance could be restored by a simple acid treatment. These improvements of solar energy utilization were ascribed to the synergetic effect of the hierarchical porous hollow morphology, the introduction of ZnSnO₃ nanosheets, and the heterojunction formed between ZnSnO₃ and Zn₂SnO₄, which could improve light harvesting capacity, expedite electron transport and charge-separation efficiencies.

1. Introduction

Synthesis of ternary metal oxides with controllable sizes and shapes has been an active research field in the past two decades, because of their size- and shape- dependent physical, chemical, optical, electronic, and catalytic properties [1,2,3]. In particular, the two zinc stannates, Zn₂SnO₄ and ZnSnO₃, have caused considerable attention due to their wide applications in dye-sensitized solar cells (DSSCs) [4,5], gas sensors [6,7], lithium-ion batteries [8,9], and photocatalysts [10,11]. Zn₂SnO₄ and ZnSnO₃ are both photovoltaic and photocatalytic materials with band gaps of about 3.6 eV and 3.2 eV, and the Femi energy level of ZnSnO₃ is higher than that of Zn₂SnO₄ [12,13]. When ZnSnO₃ is coupled with Zn₂SnO₄, the heterojunction would form at the interface between the two ternary metal-oxide semiconductors, and the electrons would transfer from the conduction band of ZnSnO₃ to that of Zn₂SnO₄ until the system reaches equilibrium of Fermi energy level [13]. This process will promote the separation of photogenerated electron–hole pairs and thus increase the performance of the DSSCs and photocatalysts. Therefore, the composite consisting of Zn₂SnO₄ and ZnSnO₃ has been successfully synthesized to improve device performance [13]. For example, the ZnSnO₃/Zn₂SnO₄ acted as photocatalyst for degradation of Intracron Blue, and achieved high performance under UV illumination due to lower recombination of electron-hole pairs [14]. The mixed phases of ZnSnO₃ and Zn₂SnO₄ were used as photoanodes in DSSCs, and it was found the ferroelectric polarization would aid in obtaining relatively high fill factors and open-circuit voltage values in the devices [15].

The published research studies have demonstrated that the three-dimensional hierarchical architectures assembled from the one- or two- dimensional nanostructured building blocks are an interesting class of materials with an abundant variety of tunable physicochemical properties [16]. Hierarchical porous hollow microstructures composed of single-crystalline two-dimensional nanosheets can provide low material density, good surface permeability, high surface area and electron transport rate, as well as high light-harvesting efficiency for application areas of catalysis and photovoltaic devices [17,18]. In this structure system, the material with micrometer-sized diameters can act as light scattering centers to increase the light harvesting of the photoanodes and photocatalysts, which is one of the crucial factors of improving the photovoltaic conversion efficiency of DSSCs and photocatalytic activity [19]. The multiple-reflection inside hollow interior can provide more opportunities for the photoanodes and photocatalysts to trap the incident light for an enhanced light utilization [20,21]. The porous hollow structures possess a large surface area, and to a great extent enhance the mass transportation and the diffusion of liquid reactant in the active layers, meanwhile, the high porosity can provide rich channels with the sizes from nanometer to micron for the surface accessibility [22]. In addition, nanosheets can offer direct pathways for photogenerated electrons, which will result in enhanced electron transport rates [23]. Hence, the hierarchical porous hollow semiconductors were acknowledged as the ideal candidates for the high-performance DSSCs and photocatalysts. However, as far as we know now, there was scarcely any previous research on hierarchical porous hollow ZnSnO₃/Zn₂SnO₄ with the view of the photovoltaic and photocatalysis fields simultaneously.

In our recent work, well-defined octahedral Zn₂SnO₄ was synthesized through a chemical solution route, and these Zn₂SnO₄ micro-crystals were converted to hollow octahedrons by an acid etching process. Whereafter, by means of an in situ growth method, ZnSnO₃ nanosheets were intentionally introduced on the surface of hollow Zn₂SnO₄ octahedrons to construct the hierarchical porous hollow ZnSnO₃/Zn₂SnO₄ octahedrons [12] (Scheme 1). In this paper, the as-prepared hollow ZnSnO₃/Zn₂SnO₄ octahedrons were used as photoanodes of DSSCs to evaluate their photovoltaic performance. Meanwhile, the photocatalytic properties of the composite under simulated sunlight irradiation were evaluated by the removal of antibiotic ciprofloxacin (CIP) and heavy metal ion Cr(VI) in single and co-existing systems. Significantly, the ZnSnO₃/Zn₂SnO₄ composite consisting of the two ternary metal oxides presented distinct advantages and had a higher photovoltaic conversion efficiency and better removal performance than solid and hollow Zn₂SnO₄ octahedrons. In addition, the morphology, crystal structure, optical properties and photocatalytic stability of the composite were investigated via various characterization methods, and the causes of enhanced photovoltaic and photocatalytic performance were proposed and discussed in detail.

2. Materials and Methods

2.1. Materials Synthesis

All chemicals were purchased from Shanghai Aladdin Bio-Chem Technology Co. Ltd. (Shanghai, China) and used without purification. Zn₂SnO₄ solid octahedron (S₁), Zn₂SnO₄ hollow octahedron (S₂), hierarchical ZnSnO₃/Zn₂SnO₄ hollow octahedron (S₃), and ZnSnO₃ nanosheets were prepared according to a reported procedure in our previous paper [12], and more details are described in the “Materials Synthesis” section in the Supplementary Materials.

2.2. Materials Characterization

The scanning electron microscopy (SEM, FEI NOVA NanoSEM230) and transmission electron microscopy (TEM, JEOL 3010) images were collected to obtain the morphology information of the samples. The compositions were analyzed by X-ray diffraction (XRD) (Rigaku, Tokyo, Japan) with Cu Ka radiation (λ = 1.54056 Å).

2.3. Preparation of the Electrode

An amount of 1 g of as-prepared sample powder and a certain amount of absolute ethanol were put into a mortar and stirred for 30 min. An amount of 4 g of the ethanol solution of two ethyl celluloses and 2 g of anhydrous terpineol were added, and the homogenization was performed using a sonicator (Vibra cell 72408, Bioblock Scientific, Wasserbad, Germany). The contents were concentrated by evaporator with 120 mbar at 38 °C and the pastes were finalized. Fluorine-doped tin oxide (FTO) conductive glass was cleaned with dilute acid and alkali solution, ethanol and deionized water several times in the ultrasonic cleaner. A doctor-blade technique was employed to coat paste onto the surface of FTO. The film was dried at 125 °C, annealed at 450 °C for 30 min and sintered at 500 °C for 15 min in the air to form the working electrodes. Typically, to improve the connection among ZnSnO₃/Zn₂SnO₄ hollow octahedrons, the as-prepared working electrodes were immersed into ZnSnO₃ sol for 30 min, then sintered at 500 °C for another 30 min.

A magnetron sputter platinum mirror acted as the counter electrode. Dye sensitization was accomplished by immersing a film in 0.3 mM dye solution (solvent mixture of tert-butyl alcohol and acetonitrile in a volume ratio of 1:1) at 80 °C over a set length of time. Here, the dye used in the DSSCs was cis-bis (isothiocyanato) bis (2,2′-bipyridyl-4,4′-dicarboxylato) ruthenium-(II) bis-tetrabutylammonium (also known as N719) [24]. The electrolyte contained 0.5 M tert-butylpyridine, 0.5 M LiI and 0.05M I₂ in acetonitrile. The FTO substrate, film, and counter electrode constituted a sandwich-like cell structure (Figure S1).

2.4. Photovoltaic Characterization

Electrochemical impedance spectroscopic (EIS) measurements were carried out with a PAR2273 workstation (Princeton Applied Research, Oak Ridge, TN, USA). The current–voltage (J-V) characteristics were measured with an active area of 0.16 cm² on an Oriel 92251A-1000 on a Keithley 236 source measurement unit. An Oriel 92251A-1000 sunlight simulator was used as the light source (AM1.5, 100 mW cm^−y). Intensity-modulated photocurrent/photovoltage spectra (IMPS/IMVS) were recorded on an electrochemical workstation (Zahner, Zennium, Kansas City, MO, USA). The incident photon-to-current efficiency (IPCE) was investigated on a PEC-S20 instrument (Peccell, Yokohama-shi, Japan). The film thickness was measured by a computerized profilometer (Dektak 3, Ulvac, Chigasaki, Japan). The dye-loading amount of photoanodes was ascertained by desorbing the dye into 10 mL of 0.1 M NaOH water solution and quantified by measuring N719 optical adsorption peak intensity.

2.5. Photocatalytic Measurements

The photocatalytic activities of the different photocatalysts were evaluated by the removal of CIP and Cr(VI) using a 500 W xenon lamp as an irradiation source. Photocatalyst (75 mg) was added into 150 mL of CIP (10 mg/L) or/and K₂Cr₂O₇ (50 mg/L) single contaminant or contaminant mixture. To achieve adsorption-desorption equilibrium, the suspension was continuously stirred for 24 h at room temperature in the dark. During the catalytic reaction, at the given intervals, 5 mL of suspension was collected and centrifuged with 4000× g rpm for 5 min to remove the photocatalyst. The concentration of Cr(VI) was estimated using the standard diphenylcarbazide analytical method at 540 nm with a UV-Vis spectrometer. The contents of CIP were analyzed by detecting the values of feature absorbance of 277 nm on a UV–Vis spectrometer. To study the reusable ability of the photocatalysts, the used samples were collected and washed with deionized water to remove the residual materials, followed by drying at 60 °C, and then the cycling experiments were conducted at optimal operating conditions.

3. Results and Discussion

3.1. Structure Characterization

The SEM image in the Figure S2a clearly demonstrated that the typical S₁ showed the octahedral morphology with an average size of ~1.2 μm and possessed smooth surfaces. After a simple acid etching process, these Zn₂SnO₄ octahedrons still maintained their original size and shape whilst also having rough surfaces (Figure S2b). As with design, S₁ converted into hollow structure, which could be confirmed by the TEM image of Figure S2c, to gain the sample S₂. As can be seen from the Figure S2d, the ZnSnO₃/Zn₂SnO₄ composite remained octahedron-like in structure, and the nanosheets of ~50 nm length and ~15 nm thickness distributed on each facet of the octahedron to form a hierarchical structure. Additionally, the TEM image revealed that the hollow structure was not damaged through the solvothermal process of ZnSnO₃ nanosheet growth (Figure S2e).

Figure S3 presented the XRD spectra of the three samples. All the diffraction peaks of S₁ and S₂ could be assigned to the cubic phase Zn₂SnO₄ (JCPDS: 24-1470), this result illustrated that no change of composition and crystal form transformation emerged during the acid etching process. For the cases of S₃, the additional XRD peaks characteristic of ZnSnO₃ (JCPDS No. 28-1486) confirmed that S₃ was ZnSnO₃/Zn₂SnO₄ composites.

To construct an energy band diagram for S₁, S₂ and S₃, the band gap energies (E_g) and flat-band potentials (E_fb) of the three samples should be determined. As shown in Figure 1a, the wavelength threshold of the pristine solid and hollow Zn₂SnO₄ octahedrons located at 340 nm and 360 nm, separately, and ZnSnO₃/Zn₂SnO₄ hollow octahedrons displayed the red-shifted light absorption in comparison with pure Zn₂SnO₄, which was helpful to extend the light absorption. The E_g was estimated according to the Kubelka-Munk function equation to be around 3.83 eV for S₁, 3.59 eV for S₂ and 3.43 eV for S₃ (the inset of Figure 1a). The positive slopes for Mott–Schottky plots indicated the characteristics of n-type semiconductors for the three samples, and the E_fb values were calculated to be −1.15, −1.15 and −1.29 V for S₁, S₂ and S₃ (vs. Ag/AgCl, pH = 7), respectively, from the tangential intercept of the curves (Figure 1b). Then, the conduction band (CB) potentials of S₁, S₂ and S₃ could be determined to be −0.93, −0.93 and −1.07 V (vs NHE, pH = 7), respectively, from the formula (Equation (1)) [25]:

$${\mathrm{E}}_{\mathrm{NHE}}={\mathrm{E}}_{\mathrm{Ag}/\mathrm{AgCl}}-{\mathrm{E}}^{\mathsf{\theta }}+0.059\mathrm{pH}$$

(1)

where E^θ and E_Ag/AgCl represent standard electrode potentials and the Ag/AgCl electrode potential at pH = 7 (0.197 V), respectively. According to the formula (E_VB = E_CB + E_g), the valence band (VB) potentials of S₁, S₂ and S₃ were calculated to be 2.90, 2.66 and 2.36 V, respectively. Based on the E_fb, E_CB and E_VB values of the three samples (Table S1), their energy band constructions were depicted in Scheme 2a. For the three materials, the bottom of CB and the top of VB were lower than lowest unoccupied molecular orbital (LUMO, −1.5 eV vs. NHE) and the highest occupied molecular orbital (HOMO, +1.1 eV vs. NHE) energy levels of N719 dye, respectively, and match well with the redox couple in the electrolyte, ensuring the operation of the DSSCs [26].

3.2. Photovoltaic Characteristics of Zn₂SnO₄ and ZnSnO₃/Zn₂SnO₄

3.2.1. Photovoltaic Characteristics of Zn₂SnO₄

Four DSSCs based on the Zn₂SnO₄ solid octahedrons photoanodes with the same thickness of ~10 µm were sensitized with various sensitization durations to examine the optimum immersion time in N719 dye solution (Table S2). Under 1 sun AM 1.5 illumination, the short-circuit current (J_sc) increased with the sensitization time in the initial 6 h, while the open-circuit voltage (V_oc) and fill factors (FF) remained stable, then the photovoltaic conversion efficiency (η) reached the maximum 2.35% at 6 h. However, a decrease in the FF, J_sc and η were observed at 8 h later probably because of the formation of aggregates, as in the case of Zn₂SnO₄ [27]. Reduced V_oc was largely due to the formation of Zn²⁺/dye complex. The immersion time was longer, a complex layer could be observed and formed a thick covering layer, which was therefore inactive for electron injection [28]. These results suggested that the optimal sensitization time was six hours in the acidic dye molecules. To explore the effect of film thickness on η of DSSCs, solid Zn₂SnO₄ films with the different thicknesses of 6, 9, 12, 15 and 18 μm were prepared and the detailed photovoltaic parameters of DSSCs based on these photoelectrodes are summarized in Table S3. In effect, when the thickness varied from 6 μm to 12 μm, the thicker film could adsorb much more dye to enhance the light harvesting ability of the photoanode, and then J_sc significantly increased from 4.17 to 5.65 mA cm⁻². As the film thickness increased continuously to 15 μm or thicker, the more trapping sites and electron recombination centers would generate, so the more serious electron recombination would lower the electrons concentration, and hence lead to the decline of J_sc, η [29] and V_oc [30]. For the optimized DSSCs based on Zn₂SnO₄ solid octahedrons, the η, V_oc, J_sc and FF were 2.49%, 0.64 V, 5.65 mA cm⁻² and 68.79%, respectively.

To study the effect of the hollow structure on the photovoltaic properties of DSSCs, Zn₂SnO₄ hollow octahedron photoelectrodes were prepared by depositing 12 μm film on FTO glass, and photovoltaic properties were tested by constructing DSSCs. The characteristic J-V curves and their photovoltaic parameters are given in Figure 2a. It can be seen that the V_oc and FF of S₂-based DSSCs were similar to those of S₁-based DSSCs, the distinct photovoltaic behavior of S₂ was larger J_sc (7.61 mA cm⁻²) compared to S₁ (5.65 mA cm⁻²). The relatively high J_sc of S₂ can be considered following the three combined effects: (1) Higher dye loading, due to larger specific surface area. The measured dye-loading amount for the S₂ film was around 4.23 × 10⁻⁷ mol·cm⁻² μm⁻¹, and was more than that of the S₁ film (2.87 × 10⁻⁷ mol·cm⁻² μm⁻¹). The surface area of S₂ was measured as 37.7 m² g⁻¹ (Figure S4), 70.6% higher than that of S₁ (22.1 m² g⁻¹). So, the larger specific surface area enabled S₂ to adsorb more dye to improve J_sc of DSSCs. (2) Stronger light-harvesting efficiency, due to the hollow structure. The corresponding UV–vis absorption spectra of solid and hollow Zn₂SnO₄ octahedrons displayed a strong absorption peak in the ultraviolet region, and the absorption intensity of the latter was stronger than that of the former (Figure 1a). Because of the same crystal structure, the enhanced light absorption of S₂ was presumably due to its hollow structure, as explained in Figure S5. Obviously, in case of Zn₂SnO₄ hollow octahedrons, the empty internal structure could offer more interfaces to considerably increase the chance of light reflections, leading to the more efficient utilization of the incident light and increased light-harvesting efficiency, and then more photogenerated electron-hole pairs would generate when S₂ was under light irradiation [31,32]. (3) Efficient electrolyte diffusion into the interior of the hollow octahedrons, due to mesoporous wall. The results of N₂ adsorption/desorption confirmed the existence of the mesoporous microstructure for Zn₂SnO₄ hollow octahedrons (Figure S4). The pore-size distribution indicated that the average pore size for S₂ was about 3.6 nm, which would promote electrolytes to transport in the photoelectrode and enter the internal area of hollow octahedrons to increase the contact probability between electrolyte and electrodes to improve the η of DSSCs.

3.2.2. Photovoltaic Characteristics of ZnSnO₃/Zn₂SnO₄

To investigate the influence of the ZnSnO₃ nanosheets on the electron transfer and recombination behaviors, the IMPS/IMVS measurements of S₂ and S₃ were performed. The IMPS/IMVS measurements could provide direct evidence of the electron transport time (τ_t) and electron recombination time (τ_r) through the equations [33]:

$${\tau }_t=1/\left(2\pi f_t\right)$$

(2)

$${\tau }_r=1/\left(2\pi f_r\right)$$

(3)

where f_t (f_r) is the minimum frequency of the IMPS(IMVS) imaginary component. Here, τ_t and τ_r characterize the electron transport properties in photoanode films and recombination with redox species in the electrolyte, respectively. Hence, smaller τ_t and larger τ_r mean a faster transport rate and slower recombination rate. Examinations of Figure 2b,c clearly showed that the τ_t (τ_r) of the DSSCs based on S₃ photoanodes was shorter (longer) than that of S₂ over light intensities from 30 to 150 W m⁻². This finding was a consequence of the following reasons: first, the engineered interfacial design of band-structure-matched ZnSnO₃/Zn₂SnO₄ composite photoanodes. A schematic illustration of the energy levels of ZnSnO₃ and Zn₂SnO₄ is shown in Scheme 2b, the CB edge of ZnSnO₃ was determined to be −0.99 V (Figure S6), which was more negative than that of Zn₂SnO₄ (−0.93 V, Table S1), such an energy band matching composite photoanode would serve as the “bridge” to expedite electron transport from excited ZnSnO₃ to FTO resulting in shorter τ_t [34]. The band-structure-matched ZnSnO₃/Zn₂SnO₄ photoanodes, meanwhile, to some extent avoid any extra internal traps, and could effectively suppress the recombination between the electron in the CB of composite photoanodes and I₃⁻ in the electrolyte, which rationalizes the reasons for the increase in τ_r. Second, the two-dimensional ZnSnO₃ nanosheets in the composite photoanodes caused direct electron transport paths and were beneficial for faster electron transport. Besides, it is well known that the less trap sites there are, the slower charge recombination occurs [35]. Compared to the nanoparticles with abundant grain boundaries, ZnSnO3 nanosheets exhibited a reduced recombination loss, which can contribute to longer τ_r. Third, interconnected contact between the ZnSnO₃ nanosheets of neighboring octahedrons made the contact area larger and prolonged the electron transport channel of the injected electrons, so lead to fast electron transport to the back contact, thus a lower recombination rate [36]. These results verified the reasonableness of the promoting the electron transfer of the ZnSnO₃ nanosheets in the composite film.

To evaluate the scattering effect of the ZnSnO₃ nanosheets, the diffuse reflection spectra of the films made from S₂ and S₃ were measured (Figure 3a). Compared to the Zn₂SnO₄ film, the composite film had higher diffuse reflection ability in the wavelength region from 380 to 800 nm, suggesting that the vertical ZnSnO₃ nanosheets would lead to more significant incident light scatter within the composite films. A possible explanation can be that S₃, owing to the surface ZnSnO₃ nanosheets, can scatter the incident light to all directions, as schematically illustrated in Scheme 3. Thereby, the composite film utilized the incident light to a higher degree than that in the case of S₂. At the same time, this scattering capacity could provide more chances for the N719 molecules on the composite area to trap photons, and then turn into electrons [26]. Otherwise, compared with S₂, the absorption edge (~390 nm) of S₃ exhibited red shift (Figure 1a), implying more holes and electrons can be generated after ZnSnO₃ incorporation, which probably was attributed to the interaction between ZnSnO₃ and Zn₂SnO₄. These results confirmed the positive impact of the ZnSnO₃ nanosheets in the composite film on the light harvesting efficiency.

To evaluate the beneficial impact of the addition of the ZnSnO₃ nanosheets on the photovoltaic conversion properties, the composite was applied as photoanodes of DSSCs. Four ZnSnO₃/Zn₂SnO₄ DSSCs with the same thickness (~12 µm) were fabricated with various sensitization durations. The η gradually increased with the sensitization time and reached the maximum at 4 h (Table S4). Figure 2a shows the J–V curves of the DSSCs based on the optimized composite photoanode film DSSCs. Compared with DSSCs based on S₂, the addition of ZnSnO₃ nanosheets to Zn₂SnO₄ cells resulted in considerable improvement in V_oc, J_sc and FF, so the DSSCs with S₃ exhibited an enhanced η (4.31%) value by around 27.9% as compared with S₂ DSSCs (3.34%). It can be seen that the V_oc of the S₃ (0.66 V) was higher than that obtained from the S₂ (0.64 V). The V_oc is determined by the chemical potential of the electrolyte and the Fermi level of the semiconductor oxide [37]. As seen in Scheme 2a, the S₃ showed a more negative E_fb value, so the difference between the Fermi level of S₃ and the redox potential of I⁻/I₃⁻ would be larger, and the corresponding V_oc will be higher. On the other hand, through the analysis of the IMVS and IMPS dates above, the introduction of ZnSnO₃ nanosheets in the Zn₂SnO₄ film can inhibit electron recombination and prolong electron lifetime, which could increase the V_oc. So, the enhanced V_oc with ZnSnO₃ nanosheets added should offer the combined impact of the elevated CB edge and the suppressed interfacial charge recombination.

The J_sc can be estimated using the following formula [38]:

$$J_{sc}=q{I}_0{\eta }_{lh}{\eta }_{inj}{\eta }_{cc}$$

(4)

where q and I₀ are the elementary charge and light flux. Hence, the J_sc is directly proportional to the light harvesting efficiency (η_lh), electron injection efficiency (η_inj) and electron collection efficiency (η_cc) of DSSCs. First, the η_lh is determined by both the dye loading amount and/or the light scattering capability [39]. The absorption quantity of N719 on S₃ with the surface area of 41.3 m⁻² g⁻¹ (4.49 × 10⁻⁷ mol·cm⁻² μm⁻¹) was slightly larger than that of S₂ (4.23 × 10⁻⁷ mol·cm⁻² μm⁻¹). Meanwhile, as discussed above, the addition of the ZnSnO₃ nanosheets favored enhanced light-harvesting efficiency and originated from light scattering and the red shift of the absorption edge. To further obtain the detailed information on the light harvest of the DSSCs, the IPCE measurements were performed, as shown in Figure 3b. The IPCE value of the S₃-cell was higher than that of the S₂-cell in whole wavelength range from 400 to 800 nm, indicating that the quantum efficiency of the S₃-cell was greater than that of the S₂-cell in the visible-light wavelength region [40]. The higher IPCE values for the S₃-cell at short wavelengths reflected on its stronger dye-loading capacity. In the longer wavelength region from 600 to 750 nm, the increased IPCE values of the S₃-cell could be explained by the enhanced light scattering efficiency, which promotes the light harvesting of the N719 dye in this region. Further, the IPCE measurements also validated the estimated J_sc values (Figure S7) and correlated well with those obtained from the J-V curves (Figure 2a). Those studies indicated superior light harvesting efficiency of S₃ films compared to S₂ films. Second, φ_inj is the probability of electron injection from the dye-excited state into the CB of the semiconductor. The CB was moving upward in the S₃ film, which meant that charge injection from LUMO of N719 dye to the CB of S₃ would be limited, thus reducing φ_inj. Third, the η_cc can be estimated according to τ_t and τ_r [33].

$${\mathsf{\eta }}_{\mathrm{cc}}=1-{\mathsf{\tau }}_{\mathrm{t}}/{\mathsf{\tau }}_{\mathrm{r}}$$

(5)

Figure 2d showed the η_cc of the S₃-cell was superior to the S₂-cell under various light intensities, which is attributed to the incorporation of ZnSnO₃ nanosheets enhancing the electron transfer. In summary, although the φ_inj values would proportionally decrease, the S₃ showed an increased J_sc because the η_lh and η_cc were raised more significantly than the decrease in φ_inj.

A comparison of photovoltaic performances between our assembled DSSCs and previously reported DSSCs is summarized in Table S5, including J_sc, V_oc, FF and η. It can be seen that, the η of the DSSC based on ZnSnO₃/Zn₂SnO₄ hollow octahedrons was not higher than that of Ba²⁺ ion doped Zn₂SnO₄ nanocrystalline [41], the cause was probably that the photo sensitizer was a mixture including N719 dye and D131 dye instead of only N719. The DSSC based on ZnSnO₃/Zn₂SnO₄ hollow octahedrons showed superior photovoltaic performances compared to metallic silver decorated Zn₂SnO₄ spheres [42], Au inlaid Zn₂SnO₄ /SnO₂ hollow rounded cubes [43], the binary metal oxide MgO passivation layers modified Zn₂SnO₄ nanoparticles [44], and ZnO/Zn₂SnO₄ core shell photoanodes [45]. In addition, the higher FF showed the advantage of ZnSnO₃/Zn₂SnO₄ hollow octahedrons in practical application [46].

3.3. Photocatalytic Performance

In order to further demonstrate the beneficial effect of the ZnSnO₃ nanosheets on photocatalytic activity, S₂ and S₃ were used to perform photocatalytic removal of CIP and Cr(VI) in single and co-existing systems under simulated sunlight. A blank control experiment was carried out to confirm that CIP would hardly self-degrade under the condition of no photocatalyst (Figure 4a). Initially, only 64.3% of CIP was degraded by S₂ within 105 min of light irradiation. With the addition of ZnSnO₃, the catalyzed degradation efficiency of the composites reached 94.3% within same irradiation time length. As shown in Figure 4b, the photodegradation of CIP catalyzed by S₂ and S₃ followed a pseudo-first-order rate law (Table S6), ln (C₀/C_t) = kt, where k is the degradation rate constant, C₀ and C_t are the concentrations of the pollutant solution before light irradiation and at irradiation time t, respectively. The k values of S₂ and S₃ were 0.010 and 0.026 min⁻¹ (Figure 4b), respectively, and the photodegradation rate constant of S₃ was around 2.6 times higher than that of S₂. The solution of Cr(VI) was also treated to investigate the photoreduction performance of photocatalyst. As shown in Figure 4c, the reduction of Cr(VI) was almost negligible in the absence of the photocatalyst, and the Cr(VI) concentration was decreased by the processes of simulated sunlight. In 200 min, 60.5% and 87.1% photoreduction efficiency was observed using S₂ and S₃, respectively. The kinetic constant 0.010 min⁻¹ of S₃ was numerically about 2.50 times as high as S₂ at 0.004 min⁻¹ (Figure 4d). All these showed, compared with the S₂ sample, that the S₃ photocatalyst displayed stronger removal efficiency under illumination, which verified that the addition of ZnSnO₃ nanosheets effectively enhanced system photocatalytic performance. For further research, the simultaneous photocatalytic CIP oxidation and Cr(VI) reduction over S₃ were performed in the co-existing systems containing CIP (10 mg/L)/K₂Cr₂O₇ (50 mg/L). S₃ only removed a small amount of CIP and Cr(VI) under a no-light condition, which showed no simulated sunlight response (Figure 5a). Compared with the photodegradation results in Figure 4 and Figure 5b, the higher CIP removal performance presented after the introduction of Cr(VI) in the system, and the k value increased from 0.026 min⁻¹ without Cr(VI) to 0.033 min⁻¹ with co-existence of Cr(VI). Meanwhile, the removal rate of Cr(VI) increased with co-existence of CIP, and the k value of S₃ in the mixed solution was 1.70 times as high as in the absence of CIP. The results of the experiment suggested the synergistic effect between photocatalytic CIP oxidation and Cr(VI) reduction.

The quenching experiments were conducted to further explore the removal mechanism of Cr(VI) and CIP over S₃ (Figure S8). In general, ammonium oxalate (AO), isopropyl alcohol (IPA), dimethyl sulfoxide (DMSO) and benzoquinone (BQ) were used to quench the active species photogenerated holes (h⁺), hydroxyl radicals (OH), photogenerated electrons (e⁻) and superoxide radicals (O₂⁻), respectively [47]. It could be observed that the photodegradation efficiencies of CIP decreased evidently with the addition of AO and BQ to the co-existing system, reflecting that h⁺ and O₂⁻ had a significant effect in the removal reaction of CIP over S₃. In contrast, when the IPA was introduced into this system, there was only a negligible suppression, indicating the influence of ·OH could be ignored. Therefore, h⁺ and O₂⁻ were the predominant reactive species in the CIP photocatalytic degradation process. Meanwhile, the photoreduction efficiencies of Cr(VI) obviously decreased from 92.9 to 47.6% with the addition of DMSO, it could be deduced that Cr(VI) was reduced by e⁻ and the photogenerated electrons served as the main reaction species of the photocatalytic reduction of Cr(VI).

Based on the above results and discussion, a photocatalytic mechanism was proposed (Scheme 2b). When the light irradiated on the photocatalyst, the ZnSnO₃/Zn₂SnO₄ composite could be excited by the photon, and equivalent energy would generate the electron-hole pairs. The conduction band of ZnSnO₃ (−0.99 V) was more negative than that of the hollow Zn₂SnO₄ (−0.93 V), the photogenerated electrons would easily inject from the former into the latter through the heterostructure between ZnSnO₃ and Zn₂SnO₄. Therefore, the photogenerated electron-hole pairs were spatially separated by this heterostructure to restrain the recombination of charges. Cr(VI) was a strong oxidant and could serve as an electron acceptor to consume these generated electrons. Because the E_CB of Zn₂SnO₄ was more negative than the O₂/·O₂⁻ potential (−0.33 eV) [48], and the electrons would react with the dissolved O₂ to produce O₂⁻. Meanwhile, the valence band of Zn₂SnO₄ (2.66 V) was more positive than that of ZnSnO₃ (2.23 V), and the holes would migrate simultaneously from VB of Zn₂SnO₄ to ZnSnO₃ in the opposite direction, whereas, h⁺ at VB of ZnSnO₃ was not more positive than the H₂O/·OH (2.40 eV) [48], and not enough to oxidize H₂O to ·OH. Hence, the O₂⁻ and h⁺ acted as the main oxidants to decompose CIP into small molecules. So, in this co-existing system, the faster photoreduction rate of Cr(VI) and photodegradation efficiency of CIP might be due to the synchronized actions of the CIP as an electron donor and Cr(VI) as an electron acceptor. The oxidation process of CIP and the reduction process of Cr(VI) consumed the photo-excited holes and electrons, respectively. These two processes effectively promoted the separation of photogenerated electrons and holes, and resulted in much more electrons for Cr(VI) reduction and holes for CIP oxidation, and the recombination of photogenerated carriers would be prohibited so as to enhance the photocatalytic efficiency [47].

To evaluate the reusable ability of S₃ in the co-existing systems, the recycling experiment was conducted under the same conditions. As shown in Figure 5c, after the three cycles of operation, the removal efficiencies changed from 92.7% to 86.1% for Cr(VI) and 91.1 to 83.2% for CIP, and the photocatalytic performance showed less than 9% loss, which confirmed that the ZnSnO₃/Zn₂SnO₄ composite possessed outstanding reusability. To further investigate the repeatable removal performance, the number of repeated uses was increased from three to nine, and the photocatalytic efficiency decreased to 71.8% for Cr(VI) and 75.3% for CIP. Borrowed from the previous method [49], after three consecutive cycles, S₃ was cleaned with 0.1 mM HCl once to remove the adsorbed pollutant molecules, it was found that S₃ with the acid-treatment preserved 88.5 and 89.2% of the initial efficiency for Cr(VI) and CIP in the ninth cycle, and greater than 15.6 and 10.4% photocatalytic efficiency without the acid-treatment, respectively (Figure 5d). So, the hierarchical ZnSnO₃/Zn₂SnO₄ porous hollow octahedrons could be applied in the simultaneous removal of antibiotic and heavy metal pollutants from contaminated water for environmental protection.

4. Conclusions

To summarize, hierarchical ZnSnO₃/Zn₂SnO₄ porous hollow octahedrons were designed by structure and composition evolution. According to IMVS/IMPS, reflectance spectra, IPCE and Mott-Schottky plots measurements, ZnSnO₃ nanosheets introduced on the surface of hollow Zn₂SnO₄ octahedrons could provide a direct pathway for electron transport, act as light-scattering centers to improve light harvesting, and induce the flat-band potential to shift negatively. As a result, as-prepared ZnSnO₃/Zn₂SnO₄ hollow octahedrons could enhance the overall photovoltaic conversion efficiency by 73.1% compared to that of the Zn₂SnO₄ solid octahedrons. Meanwhile, the composite also exhibited higher removal activity of Cr(VI) and CIP in single and co-existing systems under simulated sunlight irradiation. In addition, the composite photocatalyst showed good durability and stability in a co-existing system, suggesting its potential practical applications in the simultaneous removal of heavy metal and antibiotic pollutants from contaminated water.