Deposition of inverse opal-like TiO2 thin film with enhanced photoelectrochemical activity by a spin-coating combined with a dip-coating method

TiO2 thin films with an inverse opal-like structure have attracted considerable attention recently owing to their high potential for a range of applications. In this study, we demonstrated the possibility to deposit TiO2 thin films with an inverse opal-like structure from TiO2 nanoparticle-based slurry paste using a conventional spin-coating process. In addition, we also showed that the photoelectrochemical (PEC) performance of as-fabricated inverse opal-like TiO2 films can be further improved by the dip-coating process. In particular, dip-coated and untreated inverse opal-like TiO2 films exhibit photocurrent densities of ∼66.5 μA cm−2 and ∼40.9 μA cm−2 at 1.23 V versus RHE, respectively. A detailed physicochemical analysis revealed that photocurrent density enhancement (∼38.5%) in dip-coated inverse opal-like films can be attributed to a variety of factors including improved interconnection between TiO2 nanoparticles, higher crystallinity, decreased light reflection, and reduced charge carriers recombination. We strongly believe that these findings will be useful in the development of highly efficient third-generation solar cells, photocatalytic systems, electrochromic devices, and gas sensors.


Introduction
Nanostructured TiO 2 is widely used for the construction of third-generation solar cells, catalysts, photoelectrochemical cells, gas sensors, and other devices as a cheap, chemically stable, and Earth-abundant material [1,2]. However, the overall performance of TiO 2 is influenced by a variety of factors, including crystal structure, surface area, bandgap energy, crystallinity, shape, size, etc. In this regard, inverse-opal structured titania films have gained popularity recently due to their improved light absorption, high surface area, and improved catalytic activity [3]. Inverse-opal structured titania is commonly produced using an infiltration method, which includes three steps: (a) formation of opal films composed of colloidal polystyrene (PS) or polymethyl methacrylate (PMMA) micro/nanospheres, (b) filling of voids with a titania precursor solution, and (c) removal of polymer particles through high-temperature treatment. However, this approach frequently causes pore collapse due to surface tension and is not very controllable in terms of void filling and uniform deposition over a large region [3]. Other popular methods for inverse opal-like TiO 2 film formation include electrodeposition [4,5] and atomic layer deposition [6,7]. Although these techniques are more efficient at filling voids in opal films, limited area deposition, and fabrication complexity severely limit their widespread applicability. Hence, the development of a user-friendly approach for a large-area deposition of inverse opal-like films is crucial from a technological perspective.
The spin coating method is viewed as a promising approach thanks to its compatibility with existing technical processes, operation simplicity, and ability to deposit relatively uniform and thickness-controllable films over large areas. For spin-coating deposition, one needs a viscous paste with excellent capability to evenly spread and adhere to the substrate under the action of centripetal forces. A possible candidate can be a TiO 2 nanoparticle-based slurry paste, which is frequently used to make porous electron-transporting materials in perovskite and dye-sensitized solar cells [8,9]. For example, a slurry paste mixture of TiO 2 nanoparticles and PS/ PMMA particles can be reliably deposited on conducting glass substrates and further converted to inverse opallike structures upon heat treatment. On the other hand, it is well-known that electron transport in TiO 2 nanoparticle-based films is hampered by the effects of grain boundary scattering, charge trapping, and recombination [10,11]. To address this issue, one can also consider finding a way to better connect TiO 2 nanoparticles between each other. Thus, the primary objective of this study is to investigate the possibility of preparing inverse opal-like films from TiO 2 nanoparticle-based slurry paste using the spin coating process. To the best of our knowledge, such a study has not been done before, emphasizing the novelty of the proposed methodology. Furthermore, we also proposed a novel dip-coating methodology to improve the connection between TiO 2 nanoparticles, which resulted in increased PEC activity of inverse opal-like TiO 2 films. Formed inverse opal-like films show great promise for use in third-generation solar cells such as dye-sensitized solar cells and perovskite solar cells for effective light utilization [12,13], highly efficient photocatalytic systems [14], electrochromic devices [15], and sensors [16].

Materials and thin films deposition
Anhydrous isopropanol (99.5%), titanium (IV) isopropoxide (TIP, 97%), and other high-purity reagents were procured from Merck & Co. An aqueous suspension of plain polystyrene spheres (0.35 μm, 2.5% w/v) was purchased from Polysciences Inc. A slurry paste containing TiO 2 nanoparticles (∼20-26 nm) was prepared according to the literature [11]. The concentration of TiO 2 nanoparticles was ∼60 mg per 1 ml of the paste. First, a 400 μl suspension of polystyrene spheres was thoroughly dried in a glass vial before being combined with 200 μl of TiO 2 paste. A mixed suspension was further diluted with 50 μl of anhydrous ethanol, sealed, and stirred vigorously (900 rpm) for 24 h. Finally, thin films were formed on conducting side of precleaned fluorine-doped tin oxide FTO glasses (15 mm × 20 mm, 12-16 Ω·cm −2 ) by a spin coating method (500 rpm for 5 s followed by 3000 rpm for 20 s). Formed films were air-dried for two hours and then annealed at 500°C for 1 h to remove organic binders and polystyrene spheres.
In a typical dip-coating process, each of the pre-synthesized substrates was immersed in 5 ml of precursor (a mixture of 100 μl TIP in 10 ml isopropanol). The substrate remained in the solution for 30 min to ensure the even deposition of thin TiO 2 layer. Afterward, the substrate was withdrawn from the solution, and the nonconducting side of the glass was rinsed out with isopropanol. Air-dried substrates were annealed again in the air at 500°C for 1 h.

Characterization of thin films
Atomic force microscopy (AFM, SmartSPM 1000) and scanning electron microscopy (SEM, Crossbeam 540) were utilized to assess the morphology, surface texture, and roughness of prepared films. The structural properties of the prepared thin films were analyzed using X-ray diffraction (Rigaku SmartLab XRD) and Raman spectroscopy (Horiba LabRam Evolution system). Absorbance, reflectance, and photoluminescence study of the thin films were tested using an absolute quantum yield spectrometer (C9920-02, Hamamatsu Photonics) equipped with an integrating sphere. All measurements were measured at room temperature conditions.

Photoelectrochemical activity testing
LCS-100 solar simulator (100 W, 1.5 AM, Newport-Spectra Physics GmbH) calibrated with a reference silicon cell was utilized as a solar light source. A three-electrode cell setup with a platinum wire counter electrode, an Ag/AgCl reference electrode, and prepared TiO 2 working electrodes immersed in 0.1 M Na 2 SO 4 electrolyte was used to test the photoelectrochemical activity of prepared films. The average PEC activity results for several samples per batch were reported. Linear sweep voltammetry and chronoamperometry measurements were conducted using a portable potentiostat/galvanostat PalmSens4 (PalmSens BV). The measured potentials V versus Ag/AgCl were converted to reversible hydrogen electrode RHE scale by using the following equation: Where, E°A g/AgCl = 0.205 is the standard potential of the Ag/AgCl, and E Ag/AgCl is the measured potential against the reference Ag/AgCl electrode. Figure 1 depicts the deposition scheme of porous TiO 2 films. In the first step, a conventional spin-coating method was used to deposit a mixture of PS beads and TiO 2 slurry paste on FTO glass. The as-deposited film was dried for several hours in ambient conditions and then heat-treated for 1 h at 500°C to remove PS beads and organic components. During the annealing process, TiO 2 nanoparticles become connected and form a continuous network surrounding the voids left by the decomposed polystyrene particles. The heat-treated films were then immersed in a TIP-isopropanol solution to form an additional TiO 2 layer connecting TiO 2 nanoparticles. Finally, the dip-coated films were heat-treated again to form a crystalline TiO 2 layer. The results are highly controllable and reproducible, since the ratio between polystyrene beads and TiO 2 nanoparticles in paste solution is not changing, and nearly the same photocurrent density was obtained for several measurements.

Results and discussion
The morphology of the prepared films was examined using the SEM. Figures 2(a)-(d) show low-and highresolution SEM micrographs of bare (as-fabricated) and dip-coated (after the dip-coating process) porous TiO 2 films. The formation of inverse opal-like TiO 2 films with a mean pore size of ∼300-350 nm is easily observed. The pore size distribution is consistent with the size of the PS beads (∼350 nm), implying that they were completely removed during the heat-treatment process. The inverse opal-like structure is formed throughout the entire volume of TiO 2 films as shown in figure S1 (Supplementary Information). The average thickness was found to be ∼1.4 ± 0.1 μm. In general, there is no visual difference between the bare and dip-coated thin films; thus, the prepared samples should be thoroughly analyzed by other methods.
The XRD patterns of bare and dip-coated inverse opal-like TiO 2 films deposited on FTO substrates are shown in figure 3(a). The 2θ peaks at 25.3°and 48.1°can be readily assigned to characteristic (101) and (200) planes of anatase TiO 2 (JCPDS card no. 21-1272) [11,17]. The intensity of peaks has slightly increased, indicating an improvement in crystallinity. The mean crystallite sizes calculated from Debye-Scherrer's equation were ∼13.7 nm and ∼19.3 nm for bare and dip-coated TiO 2 films respectively. Raman analysis further confirmed the formation of the anatase TiO 2 phase and improvement in crystallinity ( figure 3(b)). In general, anatase TiO 2 has a tetragonal crystal structure and belongs to the D 4h point group with six Raman active modes (E g , B 1g , and A 1g ) [18]. These Raman active modes appeared at 136.4 (E g ), 196.5 (E g ), 387.0 (B 1g ), 545.6 (B 1g and A 1g ), and 672.2 (E g ) cm −1 . The peak at 545.6 cm −1 is usually attributed to the combination of B 1g and A 1g , as they cannot be resolved at room temperature. The E g mode is primarily caused by the symmetric stretching vibration of O-Ti-O, while B 1g and A 1g modes are mainly caused by the symmetric and anti-symmetric bending vibration of O-Ti-O respectively [19]. More intense E g mode at 136.4 cm −1 confirms the improvement in crystallinity after the dip-coating process [20].
Linear sweep voltammetry (LSV) and chronoamperometry were used to assess the PEC activity and stability of prepared films. Typically, the PEC activity of the films is compared using the theoretical oxidation potential of water (1.23 V versus RHE). The LSV curves of the films measured in the dark and under a solar simulator are shown in figure 4(a). The photocurrent density was negligible for both samples under dark conditions. On the other hand, the photocurrent density changed significantly when exposed to simulated solar light. For example, measured photocurrent densities were ∼40.9 μA cm −2 and ∼66.5 μA cm −2 at 1.23 V versus RHE for bare and dip-coated films respectively. Thus, a simple dip-coating procedure can increase the PEC activity of porous inverse opal-like TiO 2 films by ∼38.5%. Chronoamperometry analysis ( figure 4(b)) at 1.23 V versus RHE revealed that the photocurrent density is stable in both samples. For comparison, the photocurrent density of dip-coated TiO 2 film (∼66.5 μA cm −2 at 1.23 V versus RHE) was higher as compared to TiO 2 film produced by pulsed laser deposition (6.1 μA cm −2 ) [21], pristine TiO 2 nanotubes (51.76 μA cm −2 ) [22], and TiO 2 nanosheets array (36 μA cm −2 ) [23]. It should be noted that the proposed dip-coating methodology is cheaper and less energy-consuming as compared to a well-known TiCl 4 treatment used for TiO 2 photoelectrodes modification in dye-sensitized solar cells [24,25]. In general, the dip-coating process is beneficial to form a thin 'bridge' TiO 2 layer connecting TiO 2 nanoparticles and FTO substrate, which in turn can improve the transporting of charge carriers in films. This statement is well consistent with the obvious decrease in onset potential from 0.19 V versus RHE (bare TiO 2 ) to 0.08 V versus RHE (dip-coated TiO 2 ).
In addition to the formation of a thin 'bridge' TiO 2 layer, photocurrent density enhancement can be also associated with a variety of other factors such as bandgap alteration, changes in surface roughness, reduction of charge carrier recombination rate, and change in film reflectivity. Therefore, these factors have been investigated further using AFM and PL spectroscopy. Figures S2 and S3 (supplementary information) depict the optical bandgaps calculated from Tauc's plots using a recently revised protocol [26]. Optical bandgap values are estimated to be ∼3.21 eV and ∼3.24 eV for bare and dip-coated TiO 2 films respectively. The results show that the bandgap values do not change significantly and therefore can be excluded from further consideration. The prepared samples were then subjected to AFM analysis to estimate changes in surface roughness. Figures S4 and  S5 (supplementary information) show the 3D topography images of bare and dip-coated TiO 2 films taken on a 10 × 10 μm area. The root mean square roughness R rms of bare and dip-coated TiO 2 films were found to be ∼169 nm and ∼220 nm, respectively. Increased surface roughness can be further linked to a decrease in film reflectivity because of the light scattering effects. The reflectance analysis of both samples is shown in figure 5(a), and a broad peak in the UV range is typically associated with bandgap absorption of TiO 2 . One can notice that the reflectance of dip-coated TiO 2 is reduced, and the area over the curves (from 280 to 400 nm) was calculated to quantify the changes in reflectance. Calculations revealed ∼7.46% reduction in reflectivity as compared to bare TiO 2 which is likely happening as a result of a surface roughness increase. In general, such a reduction in reflectivity is not sufficient to yield significant photocurrent enhancement. Hence, the major factor of enhanced PEC activity can be considered the formation of a 'bridge' TiO 2 layer and better crystallinity of the dip-coated samples. On the other hand, a 'bridge' TiO 2 layer can also facilitate better charge transporting properties and reduced recombination of electron-hole pairs. To verify this assumption, steady-state PL(λ exc. = 280 nm) was measured as it is shown in figure 5(b). In general, the broad peak in the UV range (∼315-380 nm) is usually attributed to direct electron-hole recombination [11,27]. The intensity of dip-coated TiO 2 was noticeably lower, implying a reduced charge carrier recombination. Hence, the enhanced PEC activity of the dip-coated samples can be typically attributed to the formation of a 'bridge' TiO 2 layer, reduced charge carrier recombination, improved crystallinity, and reduced reflectivity. It should be also mentioned that the PEC activity of inverse opal-like TiO 2 films can be further altered by using PS beads with proper size and concentration, by varying the thickness of the films, and by optimizing the number of dip-coating cycles. It is intended to address these issues in a comprehensive study in the near future.

Conclusions
In conclusion, we proposed a facile spin-coating methodology for the deposition of TiO 2 nanoparticle-based inverse opal-like films. We also demonstrated that a dip-coating treatment of prepared films facilitates the formation of a thin TiO 2 connecting layer which in turn significantly improves the PEC activity of the films. In particular, the photocurrent density of dip-coated films was higher to ∼38.5% as compared to non-treated films. A thorough analysis revealed that the increased PEC activity was attributed to improvements in film crystallinity, a decrease in film reflectivity, a decrease in the recombination rate of charge carriers, and a better connection between the TiO 2 nanoparticles/FTO substrate. It is important to note that the proposed methodology can be easily extended to the fabrication of undoped/doped inverse opal-like metal oxide films, making it extremely useful for producing low-cost porous films for practical applications.