Stoichiometry and Morphology Analysis of Thermally Deposited V2O5−x Thin Films for Si/V2O5−x Heterojunction Solar Cell Applications

In recent decades, dopant-free Si-based solar cells with a transition metal oxide layer have gained noticeable research interest as promising candidates for next-generation solar cells with both low manufacturing cost and high power conversion efficiency. Here, we report the effect of the substrate temperature for the deposition of vanadium oxide (V2O5−x, 0 ≤ X ≤ 5) thin films (TFs) for enhanced Si surface passivation. The effectiveness of SiOx formation at the Si/V2O5−x interface for Si surface passivation was investigated by comparing the results of minority carrier lifetime measurements, X-ray photoelectron spectroscopy, and atomic force microscopy. We successfully demonstrated that the deposition temperature of V2O5−x has a decisive effect on the surface passivation performance. The results confirmed that the aspect ratio of the V2O5−x islands that are initially deposited is a crucial factor to facilitate the transport of oxygen atoms originating from the V2O5−x being deposited to the Si surface. In addition, the stoichiometry of V2O5−x TFs can be notably altered by substrate temperature during deposition. As a result, experimentation with the fabricated Si/V2O5−x heterojunction solar cells confirmed that the power conversion efficiency is the highest at a V2O5−x deposition temperature of 75 °C.


Introduction
Crystalline Si (c-Si) solar cells are considered to be promising next-generation energy providers as one of the most mature technologies in the renewable energy market. However, from an economic viewpoint, c-Si solar cells are still inferior to existing energy sources based on fossil fuel. Consequently, much work remains to be done to drastically increase their market share, keeping in mind the pressing issue of global warming. The low economic feasibility of existing c-Si solar cells is primarily attributable to following factors: (1) their excessive consumption of materials (e.g., Si) and (2) relatively complicated and hightemperature fabrication process. Therefore, various cost-saving technologies have recently been explored. One representative technology is the production of solar cells using a thin Si wafer (<50 µm) to reduce the required amount of Si (i.e., the material that accounts for more than 40% of the cost of manufacturing a solar cell module) [1]. This technology is able to considerably lower the cost of materials for solar cell production. However, because Si has a low absorption coefficient in the long-wavelength region [2,3], the amount of light absorbed by a Si absorber with a reduced thickness also decreases. This, in turn, lowers the power conversion efficiency (PCE) of the solar cells [4,5]. Attempts to address this shortcoming This, in turn, lowers the power conversion efficiency (PCE) of the solar cells. [4,5] Attempts to address this shortcoming have resulted in the exploration of various technologies to improve the productivity while simplifying the manufacturing process. A possible approach to overcome these problems includes the introduction of novel materials which would enable the development of low-cost and high-efficiency solar cells after combining with c-Si. Thanks to these efforts, many novel materials with distinct optical and electrical characteristics such as perovskite [6,7], organic polymers [8,9], transparent conductive oxides [10], and transition metal oxides (TMOs) [11,12] have been developed. Recently, many researchers also actively investigated the realistic application of these materials to fabricate c-Si-based heterojunction solar cells (Si-HSCs), which could minimize the cost issue stemming from complicated and high-temperature fabrication processes such as the thermally diffused doping process to build a PN junction in a conventional c-Si-based solar cell [13,14].
Among them, TMOs (e.g., V2O5, MoO3, WO3, and TiO2) are useful as excellent carrierselective hetero-contact materials for n-type and p-type Si, both of which have work functions with a wide energy range (ΦTMO = 3~7 eV), as shown in Figure 1 [15,16]. They also offer excellent window layer characteristics that can minimize the decrease in PCE resulting from the parasitic absorption by the front layers of solar cells because of their large energy band gap (Eg > 3 eV) [17][18][19]. In addition, TMOs are promising for reducing the processing cost, as the layers can be formed via low-cost processes such as evaporation (e.g., thermal and electron-beam) or spin coating [20][21][22]. Recently, a high PCE of 22.5% was achieved using a Si/TMO HSC in which TMOs (e.g., V2O5, MoO3, and WO3) were employed as the hole-selective contact layer. This promising achievement demonstrates the high potential of Si/TMO HSCs as low-cost next-generation solar cells capable of yielding high efficiency [23]. Despite the aforementioned advantages, the carrier recombination velocity at the Si/TMO interface created by depositing the TMOs is generally high. This situation arises because of the incomplete formation of the silicon dioxide (SiO2) layer, which acts as a passivation layer for the Si surface, at the low temperatures at which the TMOs are processed. The resulting short lifetime of the photo-generated carriers lowers the PCE of the solar cell [24]. The development of a method that would enable the formation of highquality passivation layers for the Si surface in the form of a TMO deposition is, therefore, essential to fabricate high-efficiency Si/TMO HSCs.
In this study, we firstly investigated the deposition temperature effect of TMO, specifically vanadium oxide (V2O5−x), to form a SiOy (0 ≤ y ≤ 2) passivation layer for Si surfaces. In many previous reports, the passivation effect of TMO for Si surfaces was studied, but they mostly focused on comparisons between different kinds of TMO such as V2O5, MoO3, and WO3, or high-temperature post-thermal annealing effects after TMO deposition [25,26]. In this report, V2O5−x was deposited at various Si substrate temperatures below 150 °C. Subsequently, the (1) stoichiometry, (2) morphology, and (3) Si surface passivation Despite the aforementioned advantages, the carrier recombination velocity at the Si/TMO interface created by depositing the TMOs is generally high. This situation arises because of the incomplete formation of the silicon dioxide (SiO 2 ) layer, which acts as a passivation layer for the Si surface, at the low temperatures at which the TMOs are processed. The resulting short lifetime of the photo-generated carriers lowers the PCE of the solar cell [24]. The development of a method that would enable the formation of highquality passivation layers for the Si surface in the form of a TMO deposition is, therefore, essential to fabricate high-efficiency Si/TMO HSCs.
In this study, we firstly investigated the deposition temperature effect of TMO, specifically vanadium oxide (V 2 O 5−x ), to form a SiO y (0 ≤ y ≤ 2) passivation layer for Si surfaces. In many previous reports, the passivation effect of TMO for Si surfaces was studied, but they mostly focused on comparisons between different kinds of TMO such as V 2 O 5 , MoO 3 , and WO 3 , or high-temperature post-thermal annealing effects after TMO deposition [25,26]. In this report, V 2 O 5−x was deposited at various Si substrate temperatures below 150 • C. Subsequently, the (1) stoichiometry, (2) morphology, and (3) Si surface passivation capability of the deposited V 2 O 5−x thin films (TFs) were investigated. Our results clearly revealed that the passivation characteristics of the Si surface could be noticeably improved by adjusting the temperature of the Si substrate, i.e., the deposition temperature, but that the temperature had to remain below 100 • C for the thermal deposition of V 2 O 5−x .
The effect of the V 2 O 5−x deposition temperature on the PCE of the Si/V 2 O 5−x HSC was evaluated by fabricating Si/V 2 O 5−x HSCs at different substrate temperatures. The HSC with the most enhanced PCE was fabricated at a deposition temperature of 75 • C, and this PCE was 25% higher than that of the sample that was not heated during the deposition. These results demonstrated that high-efficiency Si/TMO HSCs could potentially be fabricated by simply making a minor adjustment to the process temperature for TMO deposition.

Materials and Sample Preparation
A double-sided polished n-type CZ silicon wafer with a (100) orientation, thickness of 280 µm, and resistivity of 1.7-2.3 Ω-cm was cut to a size of 2 × 2 cm. The wafer was cleaned by sequential ultra-sonication with acetone, methanol, and distilled water (DI water) for 15 min each. Subsequent standard RCA cleaning with a mixture of NH 4 OH, H 2 O 2 , and DI water at a volume ratio of 1:1:5 served to remove residual organic contaminants from the wafer surface. The native oxides on the Si surface were removed by immersion in dilute HF (1 vol.%) for 1 min. The

Characterization
The dependence of τ e f f on the V 2 O 5−x deposition temperature was analyzed by measuring the τ e f f with a photo-conductance decay system (WCT-120, Sinton Instrument Inc., Boulder, CO, USA). The surface morphology of the deposited V 2 O 5−x TF was examined by acquiring atomic force microscopy (AFM) images in the tapping mode (Veeco Multi-Mode-V). The changes in the compositions of the V 2 O 5−x and the formed SiO x layers arising from oxidation and reduction reactions between the deposited V 2 O 5−x TF and the Si surface were studied using X-ray photoelectron spectroscopy (XPS, ESCALAB 250XI, Thermo-Fisher, Waltham, MA, USA). The conductivity of the V 2 O 5−x TF was derived using the sheet resistance measured via the TLM. V bi was obtained by measuring the capacitancevoltage (C-V) relationship using the Agilent E4980A LCR meter. Finally, the PCE of the fabricated HSCs was measured under the simulated air mass (AM) 1.5G condition after calibration using a standard silicon reference cell.

Results and Discussion
In preparation for the τ e f f measurements to assess the effect of the V 2 O 5−x deposition temperature on the Si surface passivation, samples were fabricated as a sandwich structure by depositing V 2 O 5−x on both sides of the Si substrate at the following deposition temperatures: RT (i.e., no heating), 50, 75, 100, and 125 • C. The results are shown in Figure 2. As shown in Figure 2, the values of the samples measured at RT and 50 °C were comparable at 70 µs and 68 µs, respectively. However, at 75 °C, the dramatically increased to 132 µs. In contrast, the samples at 100 °C and 125 °C had values of 116 µs and 75 µs, respectively, indicating a decreasing trend at deposition temperatures beyond 75 °C. This result implies that the passivation effect of the V2O5−x TF can be altered by adjusting the deposition temperature, and that effective passivation was possible even at temperatures below 100 °C. Most importantly, this result reveals the existence of a specific deposition temperature at which the passivation effect is significantly enhanced. The improved Si surface passivation effect is expected to lead to a high short-circuit current (JSC) and open-circuit voltage (VOC), by decreasing the reverse saturation current (J0) when actual Si/TMO HSCs are fabricated [27]. According to a previous study, the Si surface passivation effect of a deposited V2O5−x TF arises from the formation of a SiOx layer at the V2O5−x/Si interface when the oxygen atoms in V2O5−x migrate to the Si surface during V2O5−x deposition [28]. This migration of oxygen (O) atoms to the Si surface originates from the more negative Gibbs formation energy (∆G) for SiO2 compared to that of V2O5−x. This is evident from the following chemical formula and the change in the Gibbs free energy (∆ 2 ) [25,29].
Therefore, when V2O5−x is deposited on the Si surface, the migration of the O atoms in the V2O5−x TF to the Si surface for SiO2 formation is induced by the negative ∆ 2 5 285 kJ/mol), as shown below. As shown in Figure 2, the τ e f f values of the samples measured at RT and 50 • C were comparable at 70 µs and 68 µs, respectively. However, at 75 • C, the τ e f f dramatically increased to 132 µs. In contrast, the samples at 100 • C and 125 • C had τ e f f values of 116 µs and 75 µs, respectively, indicating a decreasing trend at deposition temperatures beyond 75 • C. This result implies that the passivation effect of the V 2 O 5−x TF can be altered by adjusting the deposition temperature, and that effective passivation was possible even at temperatures below 100 • C. Most importantly, this result reveals the existence of a specific deposition temperature at which the passivation effect is significantly enhanced. The improved Si surface passivation effect is expected to lead to a high short-circuit current (J SC ) and open-circuit voltage (V OC ), by decreasing the reverse saturation current (J 0 ) when actual Si/TMO HSCs are fabricated [27]. According to a previous study, the Si surface passivation effect of a deposited V 2 O 5−x TF arises from the formation of a SiO x layer at the V 2 O 5−x /Si interface when the oxygen atoms in V 2 O 5−x migrate to the Si surface during V 2 O 5−x deposition [28]. This migration of oxygen (O) atoms to the Si surface originates from the more negative Gibbs formation energy (∆G) for SiO 2 compared to that of V 2 O 5−x . This is evident from the following chemical formula and the change in the Gibbs free energy (∆G Si−O 2 ) [25,29].
Therefore, when V 2 O 5−x is deposited on the Si surface, the migration of the O atoms in the V 2 O 5−x TF to the Si surface for SiO 2 formation is induced by the negative ∆G Si−V 2 O 5 (−285 kJ/mol), as shown below.
However, the spontaneity of this thermodynamic reaction would be expected to increase when the temperature increases [30]. This led us to predict that, in our study, an increase in the V 2 O 5−x deposition temperature would lead to a gradual increase in τ e f f owing to the formation of SiO y (0 ≤ y ≤ 2) on the Si surface. However, as shown in Figure 2, τ e f f increased until 75 • C, above which it decreased again at the higher temperatures of 100 • C and 125 • C. Therefore, the XPS profiles of these samples were measured to study the effect of the deposition temperature on the formation of the SiO y passivation layer at the V 2 O 5−x /Si interface. The Si 2p XPS profiles, shown in Figure 3, are the Gaussian profiles, which were deconvoluted to reveal peaks for the Si substrate (Si 0 , 99.2 eV), substoichiometric SiO z (Si 1+~1 00.15 eV, Si 2+~1 01.05 eV, Si 3+~1 01.75 eV), and stoichiometric SiO 2 (Si 4+~1 03.15 eV) [31,32]. This enabled the ratios of Si 0 , SiO z , and SiO 2 at the Si/V 2 O 5−x interface to be determined according to the deposition temperature by integrating the area of each separated peak. The ratio change between Si 0 and SiO y with deposition temperature is presented in Figure 4; the detail ratios are also listed in the embedded table.
As is evident from the embedded table in Figure 4, when the temperature was increased from RT to 75 • C, the ratio of Si 0 decreased from 11% to 4%, but those of SiO z and SiO 2 increased. The ratio of SiO 2 (i.e., Si 4+ ) increased particularly at 75 • C, confirming the formation of the highest quality SiO y passivation layer at 75 • C compared to those at RT and 50 • C. However, the increase in the SiO 2 ratio is small compared to that at 75 • C, whereas the increase in the Si 0 ratio is observed to be significant; consequently, at 100 • C, the total SiO y (i.e., sum ratio of SiO z~1.5 and SiO 2 ) formation ratio decreased as shown in Figure 4. At 125 • C, the further decrease in the total SiO y ratio is mainly attributed to the reduced ratio of SiO 2 .
Based on the Si 2p XPS profiles, the deposition temperature has a marked influence on the V 2 O 5−x passivation effect, even below 100 • C. However, as was previously reported, TMOs are well known to require relatively high deposition temperatures (>400 • C) to thermally induce the effective reduction of the TMOs to form SiO y on the Si surface [26,33]. Therefore, we assume that the enhanced passivation effect we observed in our work at such a low temperature most probably had morphological origins rather than the TMO reduction being facilitated by temperature. To confirm our assumption, the initially deposited 5 nm thick V2O 5−x layer was analyzed by AFM to study the morphological change at different deposition temperatures. However, the spontaneity of this thermodynamic reaction would be expected to increase when the temperature increases [30]. This led us to predict that, in our study, an increase in the V2O5−x deposition temperature would lead to a gradual increase in owing to the formation of SiOy (0 ≤ y ≤ 2) on the Si surface. However, as shown in Figure  2, increased until 75 °C, above which it decreased again at the higher temperatures of 100 °C and 125 °C. Therefore, the XPS profiles of these samples were measured to study the effect of the deposition temperature on the formation of the SiOy passivation layer at the V2O5−x/Si interface. The Si 2p XPS profiles, shown in Figure 3, are the Gaussian profiles, which were deconvoluted to reveal peaks for the Si substrate (Si 0 , 99.2 eV), sub-stoichiometric SiOz (Si 1+~1 00.15 eV, Si 2+~1 01.05 eV, Si 3+~1 01.75 eV), and stoichiometric SiO2 (Si 4+~1 03.15 eV) [31,32]. This enabled the ratios of Si 0 , SiOz, and SiO2 at the Si/V2O5−x interface to be determined according to the deposition temperature by integrating the area of each separated peak. The ratio change between Si 0 and SiOy with deposition temperature is presented in Figure 4; the detail ratios are also listed in the embedded table.   As is evident from the embedded table in Figure 4, when the temperature was increased from RT to 75 °C, the ratio of Si 0 decreased from 11% to 4%, but those of SiOz and SiO2 increased. The ratio of SiO2 (i.e., Si 4+ ) increased particularly at 75 °C, confirming the formation of the highest quality SiOy passivation layer at 75 °C compared to those at RT and 50 °C. However, the increase in the SiO2 ratio is small compared to that at 75 °C, whereas the increase in the Si 0 ratio is observed to be significant; consequently, at 100 °C, the total SiOy (i.e., sum ratio of SiOz~1.5 and SiO2) formation ratio decreased as shown in Figure 4. At 125 °C, the further decrease in the total SiOy ratio is mainly attributed to the reduced ratio of SiO2.
Based on the Si 2p XPS profiles, the deposition temperature has a marked influence on the V2O5−x passivation effect, even below 100 °C. However, as was previously reported, TMOs are well known to require relatively high deposition temperatures (>400 °C) to thermally induce the effective reduction of the TMOs to form SiOy on the Si surface [26,33]. Therefore, we assume that the enhanced passivation effect we observed in our work at such a low temperature most probably had morphological origins rather than the TMO reduction being facilitated by temperature. To confirm our assumption, the initially deposited 5 nm thick V2O5−x layer was analyzed by AFM to study the morphological change at different deposition temperatures.
The AFM results ( Figure 5) indicated that the deposited V2O5−x of 5 nm thickness produced isolated islands. Before obtaining the AFM results, we actually expected that the surface coverage of these V2O5−x islands would continue to increase with the substrate temperature because the temperatures at which we conducted the measurements were in the low-temperature regime [34][35][36]. However, as shown in Figure 5f, the surface coverage by V2O5−x increased only until 75 °C, above which the islands again became more distinct. The uniformity of the V2O5−x islands also followed this trend of coverage and revealed the lowest root mean square roughness (RMS) at 75 °C. Based on these AFM results, we confirmed that, even in this low-temperature regime, the morphology of the V2O5−x TF was significantly affected. The AFM results ( Figure 5) indicated that the deposited V 2 O 5−x of 5 nm thickness produced isolated islands. Before obtaining the AFM results, we actually expected that the surface coverage of these V 2 O 5−x islands would continue to increase with the substrate temperature because the temperatures at which we conducted the measurements were in the low-temperature regime [34][35][36]. However, as shown in Figure 5f, the surface coverage by V 2 O 5−x increased only until 75 • C, above which the islands again became more distinct. The uniformity of the V 2 O 5−x islands also followed this trend of coverage and revealed the lowest root mean square roughness (RMS) at 75 • C. Based on these AFM results, we confirmed that, even in this low-temperature regime, the morphology of the V 2 O 5−x TF was significantly affected.
Based on this morphological sensitivity of the initial V 2 O 5−x islands to the deposition temperature, we surmised that the formation of the SiO x passivation layer at the V 2 O 5−x /Si interface was mainly influenced by the V 2 O 5−x morphology. To clarify our theory, we further investigated the change in the V 2 O 5−x morphology in terms of the aspect ratio (AR) of the initial V 2 O 5−x islands. Figure 6a-e shows the height (H) and width (W) of the V 2 O 5−x islands measured from the cross-sectional profiles of the AFM images to derive their AR at the different deposition temperatures. The obtained ARs are plotted in Figure 6f; this graph shows that the V 2 O 5−x islands formed at 75 • C had the lowest AR, 0.093, compared to those for the other temperatures. The advantage of such a low AR is that it can offer an expanded active region for supplying O atoms to the Si surface compared to those with higher ARs for the same amount of deposited V 2 O 5−x (Figure 6g,h). Therefore, a lower AR would be expected to result in the formation of a more effective SiO y passivation layer over a larger Si surface resulting from V 2 O 5−x deposition.
A possible mechanism whereby the thermally evaporated atoms are deposited is illustrated in Figure 7 as an attempt to explain the variation in the AR with the deposition temperature. Accordingly, the formation of the V 2 O 5−x TF occurs in five steps: (a) First, the solid-state V 2 O 5−x in the evaporation boat is decomposed into V and O atoms, respectively, in the vapor state during thermal deposition and (b) these atoms diffuse to the substrate.  Based on this morphological sensitivity of the initial V2O5−x islands to the deposition temperature, we surmised that the formation of the SiOx passivation layer at the V2O5−x/Si interface was mainly influenced by the V2O5−x morphology. To clarify our theory, we further investigated the change in the V2O5−x morphology in terms of the aspect ratio (AR) of the initial V2O5−x islands. Figure 6a-e shows the height (H) and width (W) of the V2O5−x islands measured from the cross-sectional profiles of the AFM images to derive their AR at the different deposition temperatures. The obtained ARs are plotted in Figure 6f; this graph shows that the V2O5−x islands formed at 75 °C had the lowest AR, 0.093, compared to those for the other temperatures. The advantage of such a low AR is that it can offer an expanded active region for supplying O atoms to the Si surface compared to those with higher ARs for the same amount of deposited V2O5−x (Figure 6g,h). Therefore, a lower AR would be expected to result in the formation of a more effective SiOy passivation layer over a larger Si surface resulting from V2O5−x deposition. A possible mechanism whereby the thermally evaporated atoms are deposited is illustrated in Figure 7 as an attempt to explain the variation in the AR with the deposition temperature. Accordingly, the formation of the V2O5−x TF occurs in five steps: (a) First, the solid-state V2O5−x in the evaporation boat is decomposed into V and O atoms, respectively,  In this process, assuming that the operating power, deposition distance, and vacuum conditions are the same, then the quality of the TF in terms of the stoichiometry, uniformity, and coverage would be most significantly affected by the distance the atoms are transported and the V2O5−x nucleation rate on the surface in Figure 7d [39,40]. Based on this consideration, the observed AR variation with the deposition temperature can be explained by (1) the distance the atoms are transported and (2) nucleation rate. First, in terms of the transportation distance of the atoms, at low substrate temperatures, the amount of TE transferred from the atoms to the substrate would be relatively large to achieve thermal equilibrium between TEVO and TEsub. This would result in the rapid adsorption of V and O atoms onto the surface when they reach the substrate, but the distance the atoms are transported would decrease owing to excessive TE loss. Therefore, at low substrate temperatures, the AR of the V2O5−x islands would be expected to be high. In contrast, at higher substrate temperatures, TE transfer from the atoms to the substrate decreases. Hence, the depositing atoms could be transported for a comparatively longer distance, which would lower the AR of the V2O5−x islands. However, as presented in Figure 6f, the AR of the V2O5−x islands gradually decreased only until 75 °C, after which it increased again at higher temperatures (i.e., 100 and 125 °C). On the basis of this behavior, we assume that the nucleation rate for the V2O5−x would decrease at temperatures above 75 °C, at which more metallic V-rich V2O5−x islands would form to increase the AR [40,41]. To confirm this assumption, additional XPS measurements were performed to investigate the stoichiometry of the V2O5−x TFs deposited at different temperatures and the results are presented in Figures  8 and 9. The results confirmed that deposition temperature notably affected the composition of the vanadium oxidation states (VOS) in the deposited V2O5−x TFs. In this process, assuming that the operating power, deposition distance, and vacuum conditions are the same, then the quality of the TF in terms of the stoichiometry, uniformity, and coverage would be most significantly affected by the distance the atoms are transported and the V 2 O 5−x nucleation rate on the surface in Figure 7d [39,40]. Based on this consideration, the observed AR variation with the deposition temperature can be explained by (1) the distance the atoms are transported and (2) nucleation rate. First, in terms of the transportation distance of the atoms, at low substrate temperatures, the amount of TE transferred from the atoms to the substrate would be relatively large to achieve thermal equilibrium between TE VO and TE sub . This would result in the rapid adsorption of V and O atoms onto the surface when they reach the substrate, but the distance the atoms are transported would decrease owing to excessive TE loss. Therefore, at low substrate temperatures, the AR of the V 2 O 5−x islands would be expected to be high. In contrast, at higher substrate temperatures, TE transfer from the atoms to the substrate decreases. Hence, the depositing atoms could be transported for a comparatively longer distance, which would lower the AR of the V 2 O 5−x islands. However, as presented in Figure 6f, the AR of the V 2 O 5−x islands gradually decreased only until 75 • C, after which it increased again at higher temperatures (i.e., 100 and 125 • C). On the basis of this behavior, we assume that the nucleation rate for the V 2 O 5−x would decrease at temperatures above 75 • C, at which more metallic V-rich V 2 O 5−x islands would form to increase the AR [40,41]. To confirm this assumption, additional XPS measurements were performed to investigate the stoichiometry of the V 2 O 5−x TFs deposited at different temperatures and the results are presented in Figures 8 and 9. The results confirmed that deposition temperature notably affected the composition of the vanadium oxidation states (VOS) in the deposited V 2 O 5−x TFs. Materials 2022, 15, x FOR PEER REVIEW 10 of 14  In Figure 8, the V 2p XPS results exhibit no change in the peak positions at the tested temperatures; the peaks are assigned to V2O5−x. Figure 8 shows that the deposition of V2O5−x on Si occurs via the formation of various VOS, identified as V 5+ with a high binding energy, and V 4+ and V 3+ with lower binding energies, respectively [7,42,43]. The measured XPS profiles were deconvoluted to distinguish the individual VOS within the V2O5−x TFs, followed by calculation of the peak areas to compare the composition ratios of the three VOS, namely, V 5+ , V 4+ , and V 3+ (Figure 9). The results from the Figure 9 showed that, as the deposition temperature increases from RT to 75 °C, the ratio of V 5+ is preserved at approximately 88%, whereas that of V 4+ increases from 6.9% to 10.5%. With respect to V 3+ , the  In Figure 8, the V 2p XPS results exhibit no change in the peak positions at the tested temperatures; the peaks are assigned to V2O5−x. Figure 8 shows that the deposition of V2O5−x on Si occurs via the formation of various VOS, identified as V 5+ with a high binding energy, and V 4+ and V 3+ with lower binding energies, respectively [7,42,43]. The measured XPS profiles were deconvoluted to distinguish the individual VOS within the V2O5−x TFs, followed by calculation of the peak areas to compare the composition ratios of the three VOS, namely, V 5+ , V 4+ , and V 3+ (Figure 9). The results from the Figure 9 showed that, as the deposition temperature increases from RT to 75 °C, the ratio of V 5+ is preserved at approximately 88%, whereas that of V 4+ increases from 6.9% to 10.5%. With respect to V 3+ , the In Figure 8, the V 2p XPS results exhibit no change in the peak positions at the tested temperatures; the peaks are assigned to V 2 O 5−x . Figure 8 shows that the deposition of V 2 O 5−x on Si occurs via the formation of various VOS, identified as V 5+ with a high binding energy, and V 4+ and V 3+ with lower binding energies, respectively [7,42,43]. The measured XPS profiles were deconvoluted to distinguish the individual VOS within the V 2 O 5−x TFs, followed by calculation of the peak areas to compare the composition ratios of the three VOS, namely, V 5+ , V 4+ , and V 3+ (Figure 9). The results from the Figure 9 showed that, as the deposition temperature increases from RT to 75 • C, the ratio of V 5+ is preserved at approximately 88%, whereas that of V 4+ increases from 6.9% to 10.5%. With respect to V 3+ , the increase in the V 4+ state likely arises from the decrease in the ratio of the V 3+ state from 4.7% at RT to 1.2% at 75 • C, seemingly as a result of the oxidation of V 3+ to V 4+ .
However, as the deposition temperature increases to 100 • C, the V 5+ state, which remained at approximately 88% of the VOS up to 75 • C, sharply decreased to 83% at higher temperatures. Simultaneously, the ratio of V 4+ increased from 10.5% to 14.1%, and that of V 3+ , from 1.2% to 2.7%. Thus, the VOS in the TF underwent reduction to lower states. This trend was more pronounced when the deposition temperature was increased to 125 • C. This implies that the nucleation of V 2 O 5−x above 75 • C is limited because of its high temperature [40]. Therefore, as indicated in Figure 9 and the embedded table, the ratio of lower VOS would increase as a result of the significant loss of volatile O atoms from the surface followed by the formation of V 2 O 5−x islands with high AR, as shown in Figure 6.
With these results, we demonstrated that the optimal deposition temperature can promote the passivation capability of the deposited V 2 O 5−x TF owing to the effective formation of the SiO y layer at the V 2 O 5−x /Si interface because of the low AR of the initial V 2 O 5−x islands. However, this facilitated formation of SiO y could also improve the electrical conductivity of the V 2 O 5−x TF itself by inducing the generation of O deficiencies in the V 2 O 5−x TF [25,44]. Therefore, using the TLM, the electrical conductivities were also measured as a function of the deposition temperature, and the results are shown in Figure S1. The electrical conductivities measured above 75 • C were higher compared to those at lower temperatures and this was mostly attributed to the higher occurrence of lower VOS within the V 2 O 5−x TFs above 75 • C.
Finally, the relationship between the V 2 O 5−x deposition temperature and the performance of the solar cell was investigated by fabricating Si/V 2 O 5−x HSCs at different substrate temperatures, as shown in Figure 10. Figure 8b shows the current density vs. voltage (J-V) curve for the fabricated HSCs, and Table 1 summarizes the HSC performance parameters. These results clearly confirmed that the HSC fabricated at 75 • C had the highest PCE value of 3.3%, compared to those fabricated at the other temperatures. As is clear from Table 1, the overall solar cell performance of the HSC fabricated at 75 • C is higher compared to that at the other temperatures. Most notably, in terms of FF, the 75 • C sample had the highest value of 37.8%. For the Si-based solar cell, effective passivation on high defect states on the surface is crucial to improve R sh . In addition, providing reduced sheet resistance of contact layers is also highly important to decrease R s . Therefore, as shown in Figure 2 and Figure S1, the enhanced passivation effect at 75 • C and improved V 2 O 5−x conductivity were the main factors for the improve PCE with the 75 • C sample [45,46]. However, the morphology of the V 2 O 5−x TF presumably also affected the value of R s considering that the smoothest surface morphology was produced at 75 • C, thereby improving the metal contact resistance in the HSCs. increase in the V 4+ state likely arises from the decrease in the ratio of the V 3+ state from 4.7% at RT to 1.2% at 75 °C, seemingly as a result of the oxidation of V 3+ to V 4+ . However, as the deposition temperature increases to 100 °C, the V 5+ state, which remained at approximately 88% of the VOS up to 75 °C, sharply decreased to 83% at higher temperatures. Simultaneously, the ratio of V 4+ increased from 10.5% to 14.1%, and that of V 3+ , from 1.2% to 2.7%. Thus, the VOS in the TF underwent reduction to lower states. This trend was more pronounced when the deposition temperature was increased to 125 °C. This implies that the nucleation of V2O5−x above 75 °C is limited because of its high temperature [40]. Therefore, as indicated in Figure 9 and the embedded table, the ratio of lower VOS would increase as a result of the significant loss of volatile O atoms from the surface followed by the formation of V2O5−x islands with high AR, as shown in Figure 6.
With these results, we demonstrated that the optimal deposition temperature can promote the passivation capability of the deposited V2O5−x TF owing to the effective formation of the SiOy layer at the V2O5−x/Si interface because of the low AR of the initial V2O5−x islands. However, this facilitated formation of SiOy could also improve the electrical conductivity of the V2O5−x TF itself by inducing the generation of O deficiencies in the V2O5−x TF [25,44]. Therefore, using the TLM, the electrical conductivities were also measured as a function of the deposition temperature, and the results are shown in Figure S1. The electrical conductivities measured above 75 °C were higher compared to those at lower temperatures and this was mostly attributed to the higher occurrence of lower VOS within the V2O5−x TFs above 75 °C.
Finally, the relationship between the V2O5−x deposition temperature and the performance of the solar cell was investigated by fabricating Si/V2O5−x HSCs at different substrate temperatures, as shown in Figure 10. Figure 8b shows the current density vs. voltage (J-V) curve for the fabricated HSCs, and Table 1 summarizes the HSC performance parameters. These results clearly confirmed that the HSC fabricated at 75 °C had the highest PCE value of 3.3%, compared to those fabricated at the other temperatures. As is clear from Table 1, the overall solar cell performance of the HSC fabricated at 75 °C is higher compared to that at the other temperatures. Most notably, in terms of FF, the 75 °C sample had the highest value of 37.8%. For the Si-based solar cell, effective passivation on high defect states on the surface is crucial to improve Rsh. In addition, providing reduced sheet resistance of contact layers is also highly important to decrease Rs. Therefore, as shown in Figures 2 and S1, the enhanced passivation effect at 75 °C and improved V2O5−x conductivity were the main factors for the improve PCE with the 75 °C sample [45,46]. However, the morphology of the V2O5−x TF presumably also affected the value of Rs considering that the smoothest surface morphology was produced at 75 °C, thereby improving the metal contact resistance in the HSCs.

Conclusions
In this report, we presented our systematic comparative study to explore the effect of the deposition temperature on the ability of the V 2 O 5−x TF to passivate the Si surface. Our results confirmed that the τ e f f can be considerably enhanced even with a minor elevation in the temperature at which the V 2 O 5−x TF is deposited below 100 • C. The XPS measurements showed that formation of the SiO x passivation layer at the Si/V 2 O 5−x interface is highly facilitated even at a temperature as low as 75 • C, which is an unprecedentedly low temperature for TMO thermal treatment in general. To investigate the origin of the improved passivation effect at such a low temperature, the morphologies of the initial V 2 O 5−x islands were analyzed. The AR of the V 2 O 5−x islands was found to be highly sensitive to even minor changes in the deposition temperature below 125 • C. In addition, a detailed study of the VOS with XPS revealed that the stoichiometry of the V 2 O 5−x TF was also notably affected by the deposition temperature. Consequently, a specific temperature, 75 • C, was found to produce V 2 O 5−x islands with the lowest AR, which offers the most-expanded V 2 O 5−x -active region to supply O atoms to the Si surface for effective SiO y layer formation. As a result, the PCE of fabricated Si/V 2 O 5−x HSCs was noticeably higher at our optimized deposition temperature compared to the PCE of the other HSCs.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/ma15155243/s1, Figure S1: (a) Measured sheet resistance of VO TFs according to Si substrate temperature extracted from the ohmic current-voltage response in TLM measurements and (b) derived conductivities of each VO TFs; Figure S2: Experimental Mott-Schottky plots and linear fittings for the samples from the different deposition temperature. Reference [47] are cited in the supplementary materials.