Understanding the effect of TiCl4 treatment at TiO2/Sb2S3 interface on the enhanced performance of Sb2S3 solar cells

The continuous search for low-cost and environment-friendly materials in photovoltaic applications has become a priority, as well as the understanding of the various strategies to boost the photovoltaic performance. In this work, we investigate the effect of TiCl4 treatment on a compact TiO2 layer used as an electron transport material (ETM) in Sb2S3 planar solar cells. After TiCl4 treatment, TiO2 exhibits higher crystallinity, lower density of hydroxyl groups acting as traps, and better surface coverage of the FTO substrate. Although no major structural changes are observed in Sb2S3 films grown on pristine or TiCl4 treated TiO2 films, there are differences in preferential growth of Sb2S3 (hk1) planes, sulfur-enrichment of the chalcogenide film, and superior substrate coverage after the TiCl4 treatment, leading to the decrease of interfacial trap states. The driving force for electron injection in the TiO2/Sb2S3 heterojunction is also favored by the shift on the VB and CB positions of TiCl4 treated TiO2. These findings are in agreement with the improved power conversion efficiency of the planar solar cell FTO/TiO2-Treated/Sb2S3/SbCl3/spiro-OMeTAD/Au.


Introduction
Antimony trisulfide (Sb 2 S 3 ) has been explored as a solar material due to its abundance in the earth's crust, low toxicity, suitable direct band gap energy (Eg) for solar radiation (1.7 eV), high absorption coefficient (α > 10 5 cm −1 at 450 nm), and good chemical stability.On the other hand, its existence in a stable phase with a low melting point (∼550 °C) favors the fabrication of Sb 2 S 3 thin films with high crystallinity and simple processing that can be synthesized at low temperatures (<350 °C) [1].
In relation to the architecture of third-generation solar cells, Sb 2 S 3 solar cells are usually assembled using a mesoporous or planar configuration [2][3][4].To date, most research has focused on mesoporous architecture, but planar structures are becoming more popular due to their ease of manufacturing and similar device performance.The electron transporting material (ETM) is considered a keystone in the planar solar cells since it could favor charge injection and facilitate transport from the absorber layer to the electrode.Additionally, ETM works as a hole-blocking layer, reducing hole recombination, which in turn prevents losses of photovoltage and promotes a better device performance [5,6].Among the different approaches, wide-bandgap metal oxides, such as TiO 2 , ZnO and SnO 2 , have received great interest for their intrinsic characteristics including proper electron conductivity, high dielectric constant, good chemical and thermal stability [7,8].In photovoltaic technologies, metal oxides must also have a proper band alignment and high electron mobility to achieve optimal device performance [9,10].
To date, TiO 2 is the most frequently used ETM to fabricate Sb 2 S 3 solar cells [11] mainly due to its chemical stability [12], optoelectronic properties [13], and compatibility with various deposition methods [14].Nevertheless, numerous trap states in the TiO 2 film might lead to inefficient charge transport and severe charge recombination [15].To overcome these drawbacks, titanium tetrachloride (TiCl 4 ) hydrolysis has been implemented in dye-sensitized solar cells (DSSC).This approach modifies the surface of the mesoporous TiO 2 film by improving the bonding between the TiO 2 nanoparticles [16][17][18][19][20]. Treatment with TiCl 4 , which adds an additional layer of TiO 2 , has shown to increase the average diameter of the particles and, at the same time, improve the amount of dye adsorbed on the anode of the DSSC [21].
TiO 2 compact layer treated with TiCl 4 has also played an important role in the photovoltaic performance of perovskite solar cells (PSC) [22][23][24].The post-heating process of TiO 2 films treated with TiCl 4 converts species from the TiCl 4 solution into TiO 2 surface crystals.Previous reports have suggested that TiCl 4 treatment also decreases gaps at the TiO 2 /perovskite interface since it is associated with an improvement in the wettability that eases the deposition of the perovskite precursor solution on the TiO 2 film surface [25].In addition, TiO 2 crystals derived from TiCl 4 treatment have been shown to improve the solar cell performance due to modification of the kinetic properties of TiO 2 such as injection, transport and recombination of charge derived from the appropriate conduction band edge positioning [22,[26][27][28].Specifically, Murakami et al [22] studied the mechanism of TiCl 4 treatment and showed that the conduction band edge of the TiO 2 compact layer was displaced to higher energy levels which improves charge separation at the TiO 2 /perovskite interface, increasing the photocurrent and photovoltaic performance of PSC [22].
On the other hand, surface traps predominantly located at the TiO 2 surface have been found to be responsible for either charge recombination or limited charge transport [29].Moreover, the resulting TiO 2 surface crystals obtained from the TiCl 4 hydrolysis has the potential to impact on the interfacial traps density, resulting in a better charge transport and lower recombination after TiCl 4 treatment [30,31].Other perspectives have taken place with TiCl 4 treatment at low temperatures (130 °C) to achieve a TiO 2 coating layer with a nonstoichiometric titanium dioxide, TiO x .By implementing this approach, the surface traps were effectively passivated [23] and the surface roughness reduced [24].These results are interesting for flexible substrates, limited by the high temperature required for crystalline TiO 2 .
For Sb 2 S 3 sensitized solar cells based on TiO 2 with mesoporous architecture, TiCl 4 treatment has been widely used to obtain highly efficient devices [32][33][34][35].For instance, the best-performing Sb 2 S 3 device with TiCl 4 treated TiO 2 mesoporous layer exhibits a power conversion efficiency (PCE) of 7.5% and an open-circuit voltage (V oc ) of 711 mV [11].Still, there is a lack of understanding of how this treatment affects the growth of Sb 2 S 3 .Moreover, planar architectures are receiving a lot of attention due to cheaper manufacturing processes with similar results.To the best of our knowledge, TiCl 4 treatment has not been well investigated in Sb 2 S 3 solar cells with planar architecture.
In this work, a Sb 2 S 3 photovoltaic device with planar architecture based on TiCl 4 treated TiO 2 compact layer was fabricated.The inherent properties of TiO 2 originated from the TiCl 4 treatment were also clarified, and its influence in Sb 2 S 3 growth and device performance were systematically studied.It was found that the TiO 2 films treated with TiCl 4 demonstrated lower oxygen vacancies and density of hydroxyl groups.In addition, Sb 2 S 3 shows preferential growth of (hk1) planes, sulfur-enrichment, and improved surface coverage.The device using an ETM with TiCl 4 treatment obtained an enhanced PCE of 9% compared with that of pristine TiO 2 , which is mainly attributed to the reduction of interfacial trap density at the TiO 2 /Sb 2 S 3 interface.

Experimental section
A portion of fluorine doped tin oxide (FTO)-coated glass was etched with a diluted hydrochloric acid and zinc powder, cleaned sequentially by ultrasonic bath using diluted neutral detergent, deionized water, and ethanol for 15 min.Then, the substrate was UV/ozone treated for 20 min.A compact TiO 2 layer was deposited by spincoating a precursor solution reported elsewhere [36].Briefly, absolute ethanol, tetrabutyltitanate, and acetic acid were mixed in a volume ratio of 20:5:0.5.The solution was then spin-coated onto a patterned FTO substrate under ambient conditions at 4000 rpm for 40 seconds.The substrate was kept in a humid environment with a relative humidity of 50% overnight to allow for full hydrolysis of the TiO 2 precursor and a good polymerization of the Ti-O network.Afterward, the substrate was sintered at 550 °C for 30 min in air, resulting in the formation of a compact TiO 2 film on the FTO substrate.For TiCl 4 treatment, the substrates were immersed in 40 mM aqueous titanium tetrachloride solution at 70 °C for 30 min, followed by copiously washing with DI water and ethanol.After that, the samples were dried with clean N 2 and annealed at 450 °C for 30 min in air.

Fabrication of Sb 2 S 3 planar solar cells
For the Sb 2 S 3 precursor solution, 2 mmol of SbCl 3 was dissolved in a mixed solvent of 1.8 ml of dimethylformamide (DMF) and 0.2 ml of dimethylsulfoxide (DMSO) and stirred for 30 min.Thiourea (TU) was added into the previous SbCl 3 solution at the specific molar ratio of 1:1.8 (SbCl 3 :TU), and stirred for 30 min.The filtered Sb 2 S 3 precursor solution was spun at 4000 rpm for 40 s in a glove box under N 2 atmosphere.The samples were placed on a hot plate for 25 min at 100 °C followed by fast annealing at 180 °C for 2 min on another hot plate, then annealed to 225 °C and remained for another 25 min.Consequently, the samples were heated up to 265 °C for 10 min.Afterward, a solution consisting of 30 mg of SbCl 3 dissolved in 1 ml of ethanol was spincoated on Sb 2 S 3 films at 5000 rpm for 40 s and heated at 100 °C for 30 s.A spiro-OMeTAD solution [37] was spin-coated at 3000 rpm for 30 s, followed by an annealing process at 100 °C for 5 min.Finally, an 80 nm layer of gold was thermally evaporated on top.

Device and film characterization
Sb 2 S 3 film morphology and cross-sectional micrographs were taken by a Hitachi S5500 at an accelerating voltage of 1-5 kV.The topographic characterization of TiO 2 and Sb 2 S 3 thin films was carried out by a Bruker-Veeco Dimension Icon equipment using a ScanAsyst mode, and an image resolution of 512 or 1024 pixels.Crystalline structure was studied on a Rigaku Ultima IV diffractometer at 2°s −1 scanning rate with Cu Kα source.A Shimadzu Fourier Transform Infrared (FTIR) Spectrophotometer (IRTracer-100), including an ATR analyzer, was also implemented to measure the ETM samples.The Raman spectra were obtained with a WiTec microscope (Alpha 300) equipped with an excitation laser of 532 nm.The Raman mapping images were obtained in a 20 ×20 mm grid with an accumulation time of 1 s.A Shimadzu spectrophotometer (UV-3101PC) was used to measure the ultraviolet-visible (UV-vis) absorption spectra.XPS spectra were recorded at room temperature using a K-alpha+ spectrometer (Thermo Fischer Scientific Co.) equipped with a AlKα (1486.6 eV) monochromatic x-ray source, a dual-beam flood gun for charge neutralization and a 180°double focusing hemispherical analyzer operating in Constant Analyzer Energy mode (CAE).The measurement spot size was 400 μm and a base pressure of 1 × 10 -8 mbar was held in the analytical chamber.High-resolution spectra were recorded at 20 eV of pass energy for Ti2p, O1s and C1s signals with a step size of 0.1 eV.All spectra were processed with the Avantage software (v5.9925) provided by Thermo-Fisher Scientific Co.The curve fitting was performed with a Voight function and a Shirley-type background.All spectra were referred to the C-C/C-H contribution of the C 1s signal from adventitious carbon, set at 284.8 eV.Surface photovoltage (SPV) decays and work function of the samples were measured by using a Kelvin probe SKP-5050 system including a monochromatic light SPS040 system.Current density-voltage (J-V ) curves of the different solar cells were obtained in an Oriel Sol3A solar simulator under 1 sun with a 0.12 cm 2 active area.Incident-photonconversion-efficiency (IPCE) spectrum was measured by a Sciencetech system.Electrochemical impedance spectroscopy (EIS) was carried out on a Biologic VMP-300 workstation with an AC potential of 10 mV amplitude in dark conditions.

Effect of TiCl 4 treatment on TiO 2
The ETM systems composed of compact TiO 2 layers without and with TiCl 4 treatment (from now on TiO 2-Treated for the TiCl 4 treated TiO 2 ) deposited onto FTO substrates were evaluated separately.Scanning electron microscopy (SEM) and atomic force microscopy (AFM) images were acquired to elucidate the different morphologies as shown in figure 1.From the top-view SEM images (figures 1(a)-(c)), it is evident that the compact TiO 2 layer covers the FTO grains with a smooth surface.Nevertheless, a deeper inspection of the TiO 2 surface (figure S1(a)) exhibits the presence of pinholes with ∼10 nm diameter, which is completely covered after the TiCl 4 treatment (see figure S1(b)).This finding is in accordance with a previous study using TiCl 4 treatment, reporting that small particles tend to aggregate in the coarse areas of the pristine TiO 2 , resulting in a pinhole-free layer [38].On the other hand, AFM images in figures 1(e) and (f) shows that the root mean square roughness of TiO 2 (12.96 nm) and TiO 2-Treated (12.27 nm) systems, are lower than that of the bare FTO surface (24.02 nm). Figure S1(c) shows a magnified view of the AFM image of TiO 2-Treated sample.Small particles around 6 to 14 nm in diameter are distributed over larger crystals, they likely form due to the TiCl 4 treatment, evidencing its tendency to repair defects in the compact layer.
Regarding the crystallinity properties of TiO 2 layers with and without TiCl 4 treatment, XRD measurements were performed on TiO 2 films grown on FTO substrates.Figure 2   Tauc linear fit and the baseline was used to obtain reasonable values for band gaps [39].Nevertheless, the tendency of a slightly higher band gap for the TiCl 4 -treated TiO 2 sample is independent of the method used to obtain the band gaps, as can be seen in the inset of figure 2(b).
FTIR measurements reveal specific information about the bonding state of TiO 2 and TiO 2-Treated samples.Figure 2(c) shows visible differences in the main peak of the FTIR spectra in both systems.After TiCl 4 treatment, the main TiO 2 band at 795 cm −1 shifts to higher wavenumbers, i.e. 828 cm −1 , which is an important observation because the band's position depends on the reduced atomic mass and chemical bond strength of the constituents atoms [40].Choudhury [41] and Serna et al [42] reported that FTIR bands position of TiO 2 samples was affected by the nonstoichiometric nature of TiO 2 .Therefore, the shift in figure 2(c  and 530.0 eV in both films are related to O 2-in the oxide.The peak observed at 531.3 eV is commonly attributed to either -OH groups [45,46] or oxygen vacancies [47].However, recent studies suggest that the presence of vacancies does not significantly impact the electronic density of the oxygen lattice [48], indicating that it is more likely related to the -OH groups.The remaining contributions at 532.5 and 532.1 eV are attributed to organic species arising from adventitious contamination, which is in agreement with what was observed in the C1s region of both samples as can be seen in figure S3.To evaluate the change in the density of OH groups, the relative intensity of the OH signal was compared to that of O 2-in each of the samples.This analysis revealed that the OH/O 2-relative intensity ratio for the TiO 2 sample was 0.18, whereas for TiO 2-Treated was 0.15.Thus, a reduction of approximately 17% in the density of OH groups after the TiCl 4 treatment, which might lead to lower charge recombination, given that these groups act as trap states [49][50][51].

Effect of TiCl 4 treatment on Sb 2 S 3 growth
As the higher crystallinity, lower oxygen vacancies, lower density of hydroxyl groups, and better surface coverage of the FTO substrate in the TiO 2-Treated system could have an influence in the growth of the absorber layer, figure 4 compares the top-view FESEM of Sb 2 S 3 films deposited on TiO 2 (figure 4(a)) and TiO 2-Treated systems (figure 4(b)).On the TiO 2-Treated sample, the Sb 2 S 3 film exhibits a denser surface and good surface coverage of the TiO 2 compact layer.A more uniform Sb 2 S 3 film is a consequence of the wettability of the TiO 2 surface after the TiCl 4 treatment since pinholes in the Sb 2 S 3 layer are due to a lack of chemical affinity between the Sb 2 S 3 precursors and TiO 2 [52].XRD patterns of Sb 2 S 3 films grown over TiO 2 and TiO 2-Treated substrates are shown in figure 4(c).The crystal structure of the two systems can be ascribed to the orthorhombic phase stibnite of Sb 2 S 3 corresponding to the JCPDS 42-1393.For TiO 2-Treated , the XRD pattern exhibits a change in the intensity of some peaks, suggesting that the TiCl 4 treatment impacts on the growth and crystallinity of Sb 2 S 3 .Some peaks have higher intensity and others are weaker than the database reference diffraction pattern, revealing a possible preferred orientation.To quantify these changes, the texture coefficient (TC) of the principal diffraction planes was calculated from the following equation:  The TC analysis exhibits that after TiCl 4 treatment the (110), (220), and (130) planes decrease, whereas the (101) and (311) planes increase, as listed in table S1.This result suggests that the TiCl 4 treatment impacts in some degree the texture and preferential growth of Sb 2 S 3 film as has been previously reported [53].
Raman spectroscopy was performed to further evaluate the Sb 2 S 3 films deposited on both TiO 2 systems, as shown in figure 5. Figures 5(a) and (e) exhibits the spatial resolution Raman images using an optical microscope, consisting of light and dark zones.The punctual Raman spectra of light and dark zones are presented in figure 5(d), both spectra present the same bands but different intensities.The spectra exhibit nine bands (figures 5(d) and (h)), which is consistent with the reported literature for orthorhombic Sb 2 S 3 [54][55][56][57][58].
Here, it is important to recall that the crystal structure of Sb 2 S 3 is formed of infinite 1-D ribbons of polymerized (Sb 4 S 6 ) n units running parallel to the b axis.Each ribbon consists of two Sb(1)S 3 trigonal pyramids at the rims and two distorted Sb(2)S 5 square pyramids in the center.To explain the vibrational modes of Sb 2 S 3 , the contribution of Sb(1)S 3 trigonal pyramids as a unit is the one considered.Weak Sb-S and S-S bonds link each unit to four neighbor ribbons [54].The band at 306 cm-1 corresponds to symmetric Sb-S stretching vibration, whereas the bands at 283 and 300 cm −1 are attributed to antisymmetric stretching vibrations [56].On the other hand, the bands at 254 and 240 cm −1 correspond to symmetric bending vibration of S-Sb-S modes, while those at 211 and 191 cm −1 are associated with antisymmetric bending vibration [56].The bands at 156 and 130 cm −1 are assigned to the crystal lattice vibrations of Sb 2 S 3 .Although the dominant bands correspond to antisymmetric (283 cm −1 ) and symmetric (306 cm −1 ) stretching vibrations of Sb-S modes in Sb 2 S 3 , all the bands are present in the obtained Raman spectra of Sb 2 S 3 (figure 5(d) and (h)).
In general, the light zone spectrum presents bands with higher intensities than the dark zone, except for the band at 306 cm −1 (see figure 5(d)).The characteristic band of the light zone spectrum is located at 283 cm −1 , and the principal bands of the dark zone are 300 and 306 cm −1 .It is worth mentioning that the light and dark zones are the same for Sb 2 S 3 films deposited onto TiO 2 or TiO 2-Treated systems.However, the distribution of light and dark zones is visibly different, which could be correlated with the substrate-related differences on preferential orientation of the Sb 2 S 3 planes.Sereni et al and Karbish et al investigated oriented single crystals of stibnite and other metal chalcogenides by polarized Raman spectroscopy [55,58].The spectra of stibnite were measured with laser polarization parallel to the a, b and c axes of the crystal.The band intensity varies significantly with the polarization direction of the incident laser.As a matter of fact, the dark zones spectrum of figure 5(d) is similar to the polarization spectra found by Karbish et al [55] parallel to the a axis, where the symmetric stretching vibration band (306 cm −1 ) is more intense than the antisymmetric stretching vibration band (283 cm −1 ).The spectra of light zones, on the other hand, is comparable to the polarization spectra parallel to the b axis, where the symmetric stretching mode is lower than the antisymmetric stretching mode [55].Additionally, Maiti et al [54]  demonstrated that the relative intensity between 283 and 306 cm −1 bands is influenced by the S/Sb stoichiometry.Thus, the higher the sulfur concentration in Sb 2 S 3 film, the higher the intensity of the 306 cm −1 band compared with that at 286 cm −1 .They attributed these changes to the reduction of sulfur vacancies, shortening the bond between Sb and S atoms [54].These ideas suggest that Raman bands are also very susceptible to stoichiometry and light and dark zones correspond to the heterogeneous distribution of domains with different sulfur content, with light and dark zones having lower or higher sulfur content, respectively.
To further analyze the distribution of light and dark zones, confocal Raman mapping images of Sb 2 S 3 films deposited on TiO 2 (figures 5(b) and (c)) and TiCl 4 treated TiO 2 (figures 5(f) and (g)) were obtained.Blue regions on the maps represent the dark zones (filtered using the peak at 306 cm −1 ), whereas red regions correspond to the light zones (filtered using the peak at 283 cm −1 ).A deeper inspection reveals an appropriate match between optical and Raman images, light zones in optical images (figures 5(a) and (e)) correspond to the intense red regions of figures 5(c) and (g), respectively, meanwhile, dark zones of figures 5(a) and (e) agrees with the intense blue regions of figures 5(b) and (f), respectively.Moreover, Raman mapping image from the filter at 306 cm −1 , i.e., blue regions (figures 5(b) and (f)), are the negative of Raman mapping image of the filter at 283 cm −1 , red regions (figures 5(c) and (g)).Thus, Raman mapping images is a good technique for correlating physical to chemical-structural properties.
The total average spectrum of Raman mapping images, which integrate the spectra acquired in all the points, represents trustworthy Raman phenomena of the whole area and not just a single point.Figure 5(h) displays total average spectra of Sb 2 S 3 thin films deposited onto pristine TiO 2 and TiCl 4 treated TiO 2 .For both films, the Raman bands are almost identical, and the predominant band is the 300-306 cm −1 .However, the signal strength of the peak 300-306 cm −1 is stronger in the Sb 2 S 3 film deposited onto the TiO 2-Treated substrate.The previous results confirm that TiCl 4 treatment directs the growth mechanism and S-enrichment of Sb 2 S 3 films with respect to pristine TiO 2 substrate.

Effect of TiCl 4 treatment on photovoltaic performance
To understand the effect of TiCl 4 treatment on the photovoltaic performance of the devices, TiO 2 /Sb 2 S 3 planar solar cells based on pristine TiO 2 and TiO 2-Treated ETM systems were fabricated.The cross-sectional SEM image in figure 6(a) exhibits the typical FTO/TiO 2-Treated /Sb 2 S 3 /SbCl 3 /spiro-OMeTAD/Au device configuration and the thickness of each layer is listed in table S2. Figure 6(b) shows the J-V curves of the champion Sb 2 S 3 devices using TiO 2 and TiO 2-Treated ETM under 100 mW cm −2 illumination, whereas the photovoltaic parameters are resumed in the inset of figure 6(b).Briefly, the champion device fabricated with TiO 2-Treated yields a higher PCE of 5.07% compared with that of pristine TiO 2 (4.57%).In addition, the IPCE spectra and integrated J sc curves of both devices are presented in figure 6(c).The IPCE spectra are very similar, although the device with TiO 2-Treated exhibits a higher PCE compared with the TiO 2 device.Moreover, the integrated J sc values of the device based on TiO 2-Treated is 16.76 mA cm − 2 , while the based on TiO 2 is 15.22 mA cm − 2 , which agrees well with the J sc values from J-V measurements.
The band structure schematic diagram at TiO 2 /Sb 2 S 3 interface was obtained by XPS and Kelvin probe measurements.The valence band of pristine TiO 2 and TiCl 4 treated TiO 2 is determined by a linear extrapolation of the initial onset of XPS spectra, see figure S4.As the binding energy scale is referenced to the Fermi energy level (E f ), the valence band maximum (VBM) values are found to be ∼2.67 and ∼2.41 eV below the E f of TiO 2 and TiO 2-Treated films, respectively; these values are close to those reported [59].On the other hand, the E f of pristine TiO 2 (−5.01 eV) and TiCl 4 treated TiO 2 (−4.93 eV) films were calculated from Kelvin probe measurements, using the vacuum energy position as a reference.Considering an optical E g of 3.3 eV in both ETM systems (figure 2(d)), the energy level diagram of the two heterojunctions can be depicted (figure 6(d)).After TiCl 4 treatment, the VBM and conduction band maximum (CBM) positions of TiO 2 film are lower in energy than those of pristine film, which in turn is favorable for enhancement of the electron injection from Sb 2 S 3 [60].As explained before, changes in surface chemistry will cause these shifts.
Other techniques are required to understand the relevance of TiCl 4 treatment in determining the interfacial phenomena.Nyquist plots of TiO 2 and TiO 2-Treated devices are presented in figure 7(a), derived from EIS measurements under dark conditions and at V oc bias.Each plot exhibits a distinct semicircular shape, which is a characteristic feature of Sb 2 S 3 planar solar cells [61,62].To analyze the data, an equivalent electrical circuit was employed, as shown in the inset of figure 7(a).The high-frequency resistance (R s ) represents the series resistance, encompassing both contacts and film resistance, while the parallel circuit is mainly associated with the recombination resistance (R rec ) and a constant phase element (CPE) linked to the chemical capacitance of the active layer [62].In contrast to the TiO 2 device, the TiO 2-Treated device displays significantly higher R rec values across the entire range of applied frequencies (figure 7(a)), implying a lower recombination rate.The recombination phenomena are likely taking place at the ETM/Sb 2 S 3 interface or inside Sb 2 S 3 itself.Although no major structural changes are observed in the Sb 2 S 3 film, there are differences in preferential growth of (hk1) planes, sulfur-enrichment, and surface coverage after the TiCl 4 treatment.As a result, TiCl 4 treatment significantly influences the TiO 2 /Sb 2 S 3 interface because it alters the properties of both components.
To quantitatively evaluate the density of trap states within the Sb 2 S 3 films, space charge limited current technique was conducted.To achieve this, electron-only devices were fabricated consisting of FTO/TiO 2 (without and with TiCl 4 treatment)/Sb 2 S 3 /PCBM/Au.In this structure, electron-only devices are created by selecting electrodes that only inject electrons into the semiconductor, i.e.TiO 2 and PCBM are n-type semiconductors.Figures 7(b) and (c) shows the resulting dark J-V curves exhibiting distinctive inflection points dividing two regions: an ohmic region and a trap-filled limit (TFL) region [63].The density of trap-state (N traps ) was calculated according to [64], by finding the onset voltage of the TFL region (V TFL ), as shown in figures 7(b) and (c).The TiCl 4 -treated device has a lower V TFL value of 0.14 V compared with the device without TiCl 4 treatment, which has a higher V TFL value of 0.49 V.The N traps value is found to be 11.41 ×10 15 cm −3 for the device based on TiO 2 , whereas the value decreased to 3.26 × 10 15 cm −3 after TiCl 4 treatment.
Surface photovoltage (SPV) spectroscopy also gives information about the interfacial phenomena upon the introduction of TiCl 4 treatment on the TiO 2 ETM. Figure 8(a) exhibits the SPV decays of TiO 2 /Sb 2 S 3 and

Conclusions
After TiCl 4 treatment, the compact TiO 2 ETM shows higher crystallinity, lower oxygen vacancies, lower density of hydroxyl groups acting as traps, and better surface coverage of the FTO substrate.Moreover, the VBM and CBM positions are lower in energy than those of pristine film, increasing the driving force for electron injection from Sb 2 S 3 .Although no major structural changes are observed in Sb 2 S 3 films, there are differences in preferential growth of (hk1) planes, as well as sulfur-enrichment, and superior surface coverage.These changes decrease the density of interfacial trap states and are associated with enhanced photovoltaic performance of devices based on TiCl 4 -treated TiO 2 .
(a) exhibits peaks located at 2θ = 25.49°and25.37°for TiO 2 and TiO 2-Treated thin layers, respectively, corresponding to the characteristic (101) plane of anatase phase (JCPDS 01-070-6826).Another small peak around 48.05°is present only on TiO 2-Treated system, which can be related to the (200) plane of anatase.Further, no more peaks of crystalline TiO 2 or other impurity phases were observed, except of those belonging to FTO (JCPDS 46-1088).The peak associated with the (101) crystalline plane of anatase increases in intensity and shifts to lower values due to the TiCl 4 treatment.As depicted in figure 2(b), the structural differences do not cause a significant change in the absorption properties of TiO 2 substrates.Moreover, the inset of figure 2(b) depicts the Tauc plot of TiO 2 and TiO 2-Treated systems, showing almost identical optical band gaps of 3.24 and 3.30 eV, respectively.Here, it is important to mention that due to the interference fringe in the Tauc plot, a vertical line perpendicular to the point of intersection of the
) indicates reduction in oxygen vacancies and stronger bonding between Ti and O atoms after TiCl 4 treatment.Differences attributed to TiCl 4 treatment were also investigated by Raman spectroscopy.It is well reported that TiO 2 crystallite dimensions impact on the position and full-width half maximum of the most intense Raman band located around 144 cm −1[43].Although Raman changes are difficult to see in films depicted in figure 2(d), a blue shift and broadening of the peak located around 144 cm −1 can be observed for the TiO 2-Treated sample in the inset of figure 2(d).These spectral modifications can be associated with changes in the crystal size, i.e., the nanoparticle size becoming smaller[44].These results are in good agreement with the morphological characterization obtained from SEM and AFM images, which exposed the formation of small TiO 2 particles from the TiCl 4 treatment.To verify changes in the chemical nature of the ETM systems deposited on FTO substrates, high resolution Ti2p, O1s and C1s XPS spectra were collected.The XPS wide survey spectra of TiO 2 and TiO 2-Treated samples are also shown in figureS2.Figure3shows the core levels of TiO 2 and TiO 2-Treated samples, where the characteristics Ti2p and O1s peaks are evident.In figures 3(a) and (b), the binding energies of Ti2p 1/2 and Ti2p 3/2 for TiO 2-Treated sample are located at 464.4 and 458.7 eV, respectively, and for TiO 2 at 464.4 and 458.7 eV, respectively.These values of binding energy indicate that only one chemical state, i.e.Ti 4+ , is dominant [45-47].In contrast, the O1s spectrum can be fitted by three peaks at 529.9, 531.3 and 532.5 eV for TiO 2 sample (figure 3(c)), and 530.0, 531.3 and 532.1 eV for TiO 2-Treated sample (figure 3(d)).The contributions around 529.9

Figure 4 .
Figure 4. (a), (b) Top view SEM images of the Sb 2 S 3 film deposited on the pristine TiO 2 and TiO 2-Treated substrate, respectively.The insets expose higher magnifications of the films.(c) XRD pattern and (d) absorption coefficient of Sb 2 S 3 films deposited on pristine TiO 2 and TiO 2-Treated substrates.Inset in (d) corresponds to Tauc plots.
hkl) is the experimental intensity of the (hkl) diffraction plane, I 0 (hkl) is the powder diffraction intensities of stibnite Sb 2 S 3 according to JCPDS 42-1393, and n is the number of peaks used in the calculation.

Figure 6 .
Figure 6.(a) Cross-sectional SEM image of Sb 2 S 3 solar cell using TiO 2-Treated layer.(b) J-V curves and (c) IPCE spectra of the champion Sb 2 S 3 devices fabricated with TiO 2 and TiO 2-Treated ETM.(d) Schematic representation of band diagram at the interface between ETM systems and Sb 2 S 3 film.

TiO 2 -
Treated /Sb 2 S 3 samples under 690 nm illumination.Here, it is important to clarify that the SPV is defined as the difference between the contact difference potential (CPD) under light and dark conditions.The SPV decay exhibits a fast region (band-to-band recombination) and an exponential region (interfacial recombination)[37].The recombination times at the ETM/Sb 2 S 3 interface are determined by fitting the SPV decays to a biexponential model[65], estimating 1351.3 and 369.0 s for devices with and without TiCl 4 treatment, respectively.Consequently, this difference highlights the distinct behavior of the TiO 2-Treated /Sb 2 S 3 system, exhibiting a pronounced reduction in trap states at the interface between the Sb 2 S 3 and TiCl 4 -treated TiO 2 layer.Additionally, SPV spectra of Sb 2 S 3 grown on TiO 2 and TiO 2-Treated systems are shown in figure 8(b).These spectra assess the normalized SPV signals below and above the band gap energy of the Sb 2 S 3 as a function of photon energy light.SPV investigates the photo-induced transitions of electrons between valence and conduction band.When using sub-band gap light, both surface and defect states associated with traps become apparent.In the spectral region close to the band gap, photons can induce free charges from shallow states of the band gap, often referred to as 'tail states', to one band.SPV spectra closely resemble the absorption spectrum of Sb 2 S 3 with a similar onset around 1.5 eV (figure4(d)), but with a gradual rise in the SPV signal.When the SPV reaches a saturation point, a knee shape is formed, the inflexion point of that knee curve is ascribed to the electronic E g .As shown in figure 8(b), both Sb 2 S 3 films exhibit an electronic E g value of ∼ 1.8 eV, which is similar to the optical E g obtained from Tauc Plot in figures 4(d).
Figure 8(b) also shows that both samples exhibit signals below electronic E g , indicating the presence of tail states close to the band gap.Compared with the TiO 2-Treated /Sb 2 S 3 sample, the SPV spectrum of the TiO 2 /Sb 2 S 3 sample reveals higher intensity signal below the E g of Sb 2 S 3 , which is associated with the shallow trap states.These findings validate that trap states are responsible for SPV signals below the absorption edge and can be passivated after the TiCl 4 treatment of the TiO 2 ETM layer.

Figure 8 .
Figure 8. Normalized SPV decays (a) and SPV as a function of photon energy (b), of Sb 2 S 3 films deposited on TiO 2 and TiO 2-Treated substrates.