Room-Temperature Sputtered SnO2 as Robust Electron Transport Layer for Air-Stable and Efficient Perovskite Solar Cells on Rigid and Flexible Substrates

Extraordinary photovoltaic performance and intriguing optoelectronic properties of perovskite solar cells (PSCs) have aroused enormous interest from both academic research and photovoltaic (PV) industry. In order to bring PSC technology from laboratory to market, material stability, device flexibility, and scalability are important issues to address for vast production. Nevertheless, PSCs are still primarily prepared by solution methods which limit film scalability, while high-temperature processing of metal oxide electron transport layer (ETL) makes PSCs costly and incompatible with flexible substrates. Here, we demonstrate rarely-reported room-temperature radio frequency (RF) sputtered SnO2 as a promising ETL with suitable band structure, high transmittance, and excellent stability to replace its solution-processed counterpart. Power conversion efficiencies (PCEs) of 12.82% and 5.88% have been achieved on rigid glass substrate and flexible PEN substrate respectively. The former device retained 93% of its initial PCE after 192-hour exposure in dry air while the latter device maintained over 90% of its initial PCE after 100 consecutive bending cycles. The result is a solid stepping stone toward future PSC all-vapor-deposition fabrication which is being widely used in the PV industry now.

perovskite, TiO 2 and ZnO are mostly prepared by solution methods, which sacrifice film uniformity and limit large-scale device production. On the other hand, SnO 2 is regarded as an alternative to replace TiO 2 and ZnO as an effective ETL to achieve low-cost PSCs with improved stability 21,22 . SnO 2 has a wider bandgap and its electron mobility is two orders of magnitude higher than that of TiO 2 , making it a more suitable candidate for use in high performance devices 23 . Compare to TiO 2 and ZnO, SnO 2 is less hygroscopic in nature, has better thermal and UV stability, and possesses lower photocatalytic activity 24 . These properties prevent perovskite degradation and benefit PSC long-term stability. Despite that high-performance PSCs based on SnO 2 have been achieved, almost all reported SnO 2 films were prepared by spin-coating [25][26][27][28][29][30] , atomic layer deposition (ALD) 31,32 , plasma-enhanced ALD 33 , sol-gel process 34 , chemical bath deposition 35 , hydrothermal process 36 , and electrodeposition 37 . In fact, many of these methods involve high-temperature processing and annealing ranging from 100 °C up to 550 °C, which again increases fabrication complexity and cost and makes it incompatible with flexible substrates. In order to make PSC technology cost-effective in the future, a fabrication technique allowing vast production is absolutely necessary.
Considering the high reliability, maturity, and capability for large-scale production of sputtering technique in both industries and laboratories, SnO 2 prepared by magnetron sputtering for PSC application is however rarely reported 38 . And previously studied sputtered SnO 2 film was calibrated based on deposition time instead of thickness, which is a more accurate and reliable approach in principle since deposition time depends on a number of deposition parameters and equipment infrastructure. Moreover, previous study lacked investigation on device stability as well as thorough thin film characterization on their SnO 2 and the corresponding glovebox-processed spin-coated perovskite absorber, particularly morphology and crystallinity analysis by scanning electron microscopy (SEM), X-ray diffraction (XRD), and photoluminescence (PL). Therefore, it was hard to convince the effectiveness and compatibility of sputtered SnO 2 with perovskite solar cells. On the other hand, perovskite films are typically prepared by solution methods inside a glovebox filled with inert gas. Similarly, those solution methods limit film scalability while the use of glovebox increases production cost and complexity. Researchers have therefore put a great amount of effort to fabricate perovskite films and devices in ambient condition without sacrificing film quality, device performance, as well as stability 19,39 .
Here, we demonstrate room-temperature RF sputtered SnO 2 film as an effective and robust ETL and meanwhile take one step forward to implement it together with vapor-deposited perovskite absorber for air-stable and efficient PSCs on both rigid and flexible substrates. By this application, we are now only one step away from sputtered SnO 2 based all-vacuum-deposited perovskite solar cells which will eventually enable industrialization. To the best of our knowledge, there is no report of using sputtered SnO 2 for flexible PSCs application. Both SnO 2 thickness and deposition conditions including working pressure and gas environment were systematically investigated. Device stability and hysteresis were also carefully studied.

Results and Discussion
In this work, we demonstrated n-i-p planar structure of PSCs with optimized room-temperature-processed SnO 2 as ETL prepared by RF magnetron sputtering and vacuum-deposited perovskite film. Figure 1a shows the schematic of device structure of our PSCs: ITO-PEN/SnO 2 /MAPbI 3 /Spiro-OMeTAD/Au. Figure 1b,c display SEM images of complete device based on rigid FTO glass substrate and flexible ITO-PEN substrate respectively. The morphology and surface roughness of bare FTO ( Supplementary Fig. S1), 40 nm SnO 2 -coated FTO (Fig. 2a,b), and solution-processed SnO 2 (Fig. 2c,d) were studied by SEM and atomic force microscopy (AFM). Supplementary AFM reveals the root-mean-square (RMS) roughness and mean roughness of bare FTO glass were 7.736 nm and 6.096 nm respectively. In contrast, 40 nm SnO 2 -coated FTO had a RMS roughness and mean roughness of 5.488 nm and 4.358 nm respectively, while solution-processed SnO 2 had higher RMS roughness and mean roughness of 6.439 nm and 5.000 nm. It is reflected that sputtered SnO 2 film was uniformly deposited on FTO surface and SnO 2 grains were small enough to fill in the gaps between FTO grains, leading to a lower surface roughness than solution-processed SnO 2 , which is beneficial for the growth of vapor-deposited perovskite. To illustrate how surface roughness matters, perovskite was vapor-deposited on a 10 nm SnO 2 -coated FTO with RMS roughness and mean roughness of 7.199 nm and 5.708 nm respectively, measured by AFM ( Supplementary Fig. S2). It www.nature.com/scientificreports www.nature.com/scientificreports/ was clear that perovskite grain sizes on the 10 nm SnO 2 -coated FTO were significantly reduced ( Supplementary  Fig. S3a). Moreover, cross-sectional SEM shown in Supplementary Fig. S3b reveals that the perovskite grains and shape became more irregular on 10 nm SnO 2 -coated FTO, while single-crystal-thick perovskite grains with larger grain sizes and regular shapes were crystallized on 40 nm SnO 2 -coated FTO ( Supplementary Fig. S3c,d).
In addition, small perovskite grains tended to crystallize in the perovskite-SnO 2 interface on 10 nm SnO 2 -coated FTO, which could adversely affect the efficiency of carrier transport. Since materials were slowly deposited onto substrate down to a few angstroms per second, a rougher surface would hinder the ion migration and volumetric expansion as perovskite crystallization took place. In contrast, this problem becomes less significant when perovskite is prepared by solution methods because they allow ions within the precursors to easily spread all over the substrates without overcoming significant energy barriers. Therefore, this problem was not commonly discussed in depth before. In contrast, although perovskite grown on solution-processed SnO 2 film ( Supplementary Fig. S4) demonstrated comparable grain size with those grown on sputtered SnO 2 , the former perovskite films exhibit layered structure on most of the grains. This could be attributed to the higher roughness of solution-processed SnO 2 film, leading to inconsistent rate and degree of perovskite crystallization and ultimately high surface roughness. These could increase the probability of carrier recombination between perovskite absorber and the HTL.
The full ultraviolet photoelectron spectroscopy (UPS) spectrum of sputtered SnO 2 is shown in Supplementary  Fig. S5a. The sputtered SnO 2 film showed a secondary cutoff edge of 5.01 eV ( Supplementary Fig. S5b), indicating a work function W S of 5.01 eV. Supplementary Fig. S5c shows the valence band maximum (VBM) of the sputtered SnO 2 film is located at 3.07 eV, below E F . The bandgap of the sputtered SnO 2 film acquired from the Tauc plot ( Fig. 3a) was 3.72 eV, which is wider than that of ZnO and TiO 2 . A wider bandgap implies better hole blocking ability and can avoid absorption of high-energy photons which leads to small current loss 40 . Based on the above values, it can be calculated using the semiconductor band structure (E C = W S + VBM − E g ) that the E C of the sputtered SnO 2 film was 4.36 eV, which is deeper than that of TiO 2 and ZnO, both are 4.2 eV. The deeper conduction band of SnO 2 compared to TiO 2 and ZnO could enhance electron transfer from perovskite to the ETL. On the other hand, calculation showed E V of the sputtered SnO 2 film was 8.08 eV, which is much deeper than that of TiO 2 and ZnO, 7.4 eV and 7.6 eV respectively. The deeper valence band of SnO 2 can enhance hole blocking ability from perovskite to the ETL. Figure 3b shows the energy band diagram of each device component and the transportation of photo-generated electrons and holes.
Being an effective ETL for PSCs of n-i-p planar structure, transmittance, thickness, and film quality are crucial. XRD (Fig. 3c) revealed that both sputtered and solution-processed SnO 2 films were polycrystalline but the former one exhibited better crystallinity. All XRD peaks for SnO 2 were indexable to the tetragonal SnO 2 structure, indicating the formation of pure SnO 2 crystals. In addition, sputtered SnO 2 film on FTO glass showed good www.nature.com/scientificreports www.nature.com/scientificreports/ transparency with transmittance close to 90% in the visible region ( Fig. 3d), while solution-processed SnO 2 had lower transmittance of about 80%. It is noteworthy the room-temperature sputtered SnO 2 here demonstrated even higher transmittance than other high-temperature-processed spin-coated SnO 2 films 41 . To obtain high-quality SnO 2 films, impact of thickness, sputtering working pressure, and the flow rates of O 2 and Ar during sputtering were systematically studied.
Since the thickness of ETL can critically affect cell performance, SnO 2 film was first sputtered under the same sputtering power of 60 W on FTO glass substrates kept at room temperature with four different thicknesses, 20 nm, 40 nm, 60 nm, and 80 nm, which took 8 min, 15 min, 23 min, and 30 min for sputtering, respectively. It can be inferred that the deposition rate was approximately 0.43 Å s −1 . After sputtering, SnO 2 -coated FTO substrates were transferred to the evaporator for perovskite fabrication via a two-step vapor deposition as described in the Method section. The as-deposited perovskite samples were annealed in ambient air condition with over 65% humidity. It has been reported that certain level moisture is helpful for perovskite crystallization but excessive moisture could be detrimental to perovskite 42 . To overcome the humidity problem, perovskite samples were annealed for a short period of time at elevated temperature to accelerate the perovskite crystallization process and meanwhile minimize perovskite film degradation in ambient condition 39,43,44 . As-deposited perovskite samples were annealed at 130 °C for 10 min instead of the conventional 100 °C for an hour. It turned out that this method is also workable for vapor-deposited perovskite films, not only for solution-processed perovskite films. The UV-vis absorption spectra shown in Fig. 4a of the vapor-deposited perovskite shows good absorption in the visible region. It also revealed that perovskite grown on sputtered SnO 2 exhibited higher absorption than that grown on solution-processed SnO 2 , which was attributed to its lower transmittance than sputtered SnO 2 . The absorption onset corresponded to an optical bandgap of 1.57 eV, estimated from the Tauc plot ( Supplementary  Fig. S6). The estimation matches well with the perovskite PL peak at 788 nm and the steady PL showed a more significant quench when depositing the perovskite film on sputtered SnO 2 compared to solution-processed SnO 2 (Fig. 4b). It supports that sputtered SnO 2 possessed more efficient electron transport ability. XRD of perovskite (Fig. 4c) presented the expected perovskite pattern, with intense signals at 14.1°, 28.4°, and 31.9° corresponding to the (100), (200), and (310) directions, respectively. It showed an extra peak of PbI 2 at 12.7° for perovskite that underwent 30 min prolonged annealing in humid air condition. Supplementary Fig. S7 shows the SEM of perovskite annealed for 30 min. The perovskite decomposing into PbI 2 hindered its crystallization by grain boundary expansion and grain cracking. As a consequence, the intensity of each perovskite peak was clearly reduced as shown in Fig. 4c. It confirms the effectiveness of short-time annealing processing at elevated temperature in ambient condition.
The J-V characteristics of devices based on FTO glass substrates with different SnO 2 thicknesses measured under AM1.5G illumination are shown in Fig. 5a. The device performance firstly increased and then decreased as www.nature.com/scientificreports www.nature.com/scientificreports/ the sputtered SnO 2 thickness increased. It is seen that devices with 40 nm SnO 2 yielded the best performance, with a PCE of 11.14%, a V OC of 0.934 V, a J SC of 22.91 mAcm −2 , and a fill factor (FF) of 52.1%, so 40 nm was taken as the optimum thickness. If the SnO 2 layer was too thin, it could not fully cover the FTO surface for effective electron transport. On the other hand, a too thick SnO 2 layer would induce a larger series resistance.
The sputtering working pressure is another critical factor to determine the sputtered film quality. We fabricated PSCs based on 40 nm SnO 2 sputtered under the same power of 60 W but three different working pressures, 0.25 Pa, 0.5 Pa, and 1.0 Pa. The J-V characteristics of respective devices are shown in Fig. 5b. The device performance, particularly the V OC , decreased as working pressure increased. It is seen that devices with SnO 2 sputtered at 0.25 Pa yielded the best performance, with a PCE of 12.18%, a V OC of 0.948 V, a J SC of 22.34 mAcm −2 , and an FF of 57.5%, so 0.25 Pa was taken as the optimum working pressure. An optimum working pressure is important so that the mean free path of gas molecules (O 2 and Ar) is comparable to the distance between the target and substrates. A high working pressure will reduce the mean free path of molecules. In other words, there will be so much scattering that electrons will not have enough time to gather enough energy between collisions to ionize the atoms on the target. As a result, it causes a less uniform film deposition over the substrates. The reduced V OC is attributed to uneven deposition due to too high working pressure. Sputtering under working pressure below 0.25 Pa was attempted, however, the plasma became unsustainable and unstable around 0.2 Pa due to too low gas molecule concentration. In order to yield a self-sustaining plasma, each electron has to generate enough secondary emission. Therefore, 0.25 Pa was concluded to be the optimum working pressure for SnO 2 sputtering without compromising film quality and device performance.   Table 1.  Fig. 5f.
To investigate the stability, both champion devices were left in room-temperature dry air with 30% humidity in dark for 192 hours. The PCE of each device was measured every 24 hours. It is found after 192 hours that the sputtered SnO 2 based champion device could retain over 93% of its initial PCE while the solution-processed SnO 2 based champion device could only retain 77% of its initial PCE. Supplementary Fig. S8 shows the J-V characteristics of both devices measured after 192 hours of stability test. The PCE of the sputtered SnO 2 based device dropped to 11.91% with a V OC of 0.929 V, a J SC of 21.99 mAcm −2 , and an FF of 58.3% while the PCE of the solution-processed SnO 2 based device dropped to 8.34% with a V OC of 0.853 V, a J SC of 19.75 mAcm −2 , and an FF of 49.5%. Supplementary Fig. S9a shows the evolution of their PCEs throughout the monitored period. The performance loss can be attributed to the degradation of the perovskite film within an unencapsulated device in 65% humidity. The highly hygroscopic and deliquescent properties of Li-TFSI used as dopant in Spiro-OMeTAD can also be regarded as a contribution to perovskite degradation due to penetration of water molecules 45 . In addition, the perovskite film could deteriorate rapidly in the presence of TBP as a polar solvent. Therefore, TBP may be a good solvent for perovskite, which means it can corrode the perovskite layer 46 . It is worth mentioning that the perovskite films, HTL films, and all devices were annealed, prepared, unencapsulated, and characterized in ambient condition with humidity greater than 65%. It convinces the viability of air processability of perovskite films, HTL films, and device characterization without the use of a glove box. To further illustrate this assertion, 40 sputtered SnO 2 based PSCs in one single batch using the optimized SnO 2 parameters and 40 solution-processed SnO 2 based PSCs in one batch were fabricated by the same procedures. Statistics as shown in Supplementary Fig. S9b convince the reproducibility and consistency of device performance. Both batches of devices demonstrated normal distribution of PCEs. It can be concluded that sputtered SnO 2 based devices showed higher average PCE and sputtered SnO 2 is therefore more effective and beneficial to enhance PSC device performance and stability.
Flexibility is a desirable feature of thin film solar cells for a variety of applications, such as portable power sources, building-integrated photovoltaics, clothing and textiles, power-generating fabrics, and electronics with light-weight curved surface. One of the main advantages in the use of sputtered SnO 2 is that no sintering or annealing step is required thus flexible plastic substrates can be used. The other major fabrication steps (vapor deposition of perovskite, spin-coating of Spiro-OMeTAD, and thermal evaporation of Au) are also carried out in room temperature condition, which means this fabrication process as a whole is compatible with flexible substrates. To date, majority of the reported flexible PSCs employ spin-coated TiO 2 , ZnO, or PCBM as ETL [47][48][49][50][51] , while very few of them used solution-processed SnO 2 as ETL. To illustrate the compatibility of sputtered SnO 2 with flexible substrates for perovskite photovoltaic, in our work rigid FTO glass substrate was therefore replaced by flexible substrate, namely indium-doped-tin-oxide-coated polyethylene naphthalate (ITO-PEN) (Fig. 6a). Supplementary  Fig. S10a shows the XRD spectrum of vapor-deposited perovskite grown on flexible ITO-PEN. It presented an  www.nature.com/scientificreports www.nature.com/scientificreports/ expected perovskite spectrum with sharp signal intensities and without PbI 2 residue, showing that vapor deposition and post-annealing treatment of perovskite on flexible substrates did not induce any perovskite degradation. 20 devices were fabricated using the same preparation procedures and their PCE distribution is summarized in Supplementary Fig. S10b. Figure 6b shows the J-V characteristics of the champion device on flexible ITO substrate. It yielded a PCE of 5.88% with a V OC of 0.932 V, a J SC of 8.91 mAcm −2 , and an FF of 70.8% when measured under reverse voltage scanning and a PCE of 5.48% with a V OC of 0.898 V, a J SC of 8.71 mAcm −2 , and an FF of 70.1% when measured under forward voltage scanning. Therefore, the device exhibits a small hysteresis. The discrepancies between J SC 's appeared in the studied devices measured under reverse and forward scannings might be the result of slow response of photocurrent and higher defect density 52 . It has been confirmed that higher defect density significantly contributes to the J-V hysteresis and degradation of photovoltaic parameters 52,53 . In comparison with the champion device based on rigid FTO glass substrate, the flexible devices yielded a substantial loss in J SC but a significant improvement in FF. The loss of J SC could be accounted to the lower transmittance through ITO-PEN substrates compared to FTO glass substrates and the threefold increase in device area from 3.14 mm 2 to 10 mm 2 . On the other hand, the improvement in FF could be caused by the smoother ITO electrode surface, allowing flatter coverage of SnO 2 film and eventually more compact interface between SnO 2 film and perovskite for even more effective electron extraction and reduced series resistance.
Mechanical flexibility of devices under bending stress is of great importance concerning flexible and/or wearable device applications. Bending tests showed how well the device performance retained after being bent repeatedly to decreasing radii of curvature. The identical flexible PSC was bent by mechanical force with 10 different radii of curvature in one bending cycle. After each round of bending, the device performance was measured repetitively. The impact of bending on device PCE is presented in Fig. 6c. Less than 12% drop in PCE was observed. This result indicated sputtered SnO 2 film is an effective and robust ETL for flexible PSC application. The impact of mechanical bending via multiple cycles of bending test was further evaluated. A total of 100 consecutive bending cycles at radius of 2 cm were performed. As shown in Fig. 6d, the device sustained over 90% of its initial PCE. After the bending test, V OC , J SC , and FF of the flexible device respectively dropped from 0.930 V to 0.929 V, from 8.76 mAcm −2 to 8.72 mAcm −2 , and from 71.4% to 70.3%. Consequently, the PCE reduced from 5.82% to 5.69%. Although the flexible devices based on sputtered SnO 2 demonstrated lower PCE than devices based on TiO 2 , ZnO, and PCBM, this is a pioneering work proving that sputtered SnO 2 is effective and robust on both rigid and flexible substrates for perovskite photovoltaics. Sputtering technique is desirable for upscaling device area with uniform film deposition while flexible devices are especially attractive for a variety of consumer-driven products. www.nature.com/scientificreports www.nature.com/scientificreports/

Conclusion
We have pioneered the optimization and implementation of radio frequency magnetron sputtered SnO 2 as electron transport layer for vapor-deposited-MAPbI 3 -based perovskite solar cells on both rigid and flexible substrates. It was demonstrated that neither mesoporous scaffold nor any high-temperature processing procedures were required to achieve efficient and air-stable devices without the use of a glove box. It is noteworthy that in the current device structure there was no backside passivation and all devices were not packaged, so the entire fabrication and characterization processes were subject to ambient condition of humidity greater than 65%. Despite the air processing in humid environment and perovskite annealing on flexible substrates, PSCs of 12.82% PCE on rigid glass substrates and 5.88% PCE on flexible substrates were achieved. We have also shown that the viability and repeatability of acquiring high-quality vapor-deposited perovskite films with large grain sizes and smooth morphology on sputtered SnO 2 film via short-time annealing at elevated temperature processing in ambient condition, proving its compatibility with vapor-deposited perovskite films. More importantly, sputtered SnO 2 based devices were demonstrated to have better device photovoltaic performance and stability than solution-processed SnO 2 based devices. Such successful implementation of robust sputtered SnO 2 films on flexible devices could serve as a promising route for future development and application of sputtered SnO 2 film into large-scale cost-effective all-vacuum-deposited flexible perovskite photovoltaics.
Device fabrication. The substrates were sequentially washed with acetone, isopropanol, and deionized water. The sheet resistance of FTO is 15 Ω □ −1 and the thickness of glass and FTO are 1.6 mm and 420 nm respectively. The average transmittance of FTO glass in the visible region is 85%. The ITO-PEN has a sheet resistance of 15 Ω □ −1 , a thickness of 0.125 mm, and 78% transmittance in the visible region. SnO 2 was deposited on FTO glass and ITO-PEN by radio frequency magnetron sputtering in room temperature. The clean substrates were transferred to a vacuum chamber and evacuated to a pressure of 4 × 10 −4 Pa for SnO 2 sputtering. The substrates were mounted on a rotating platform, 10 cm above the SnO 2 target (China Rare Metal Co. Ltd.). The sputtering atmosphere was consisted of O 2 and Ar. When 4 × 10 −4 Pa was reached, O 2 (99.99%) and Ar (99.99%) were pumped into the chamber. The gas flow rates of O 2 and Ar were controlled by gas-flow meters and the gas flow ratio of O 2 and Ar was set 1 sccm and 50 sccm, 5 sccm and 50 sccm, or 10 sccm and 50 sccm respectively. The working pressure for sputtering was maintained 0.25 Pa, 0.5 Pa, or 1.0 Pa. The SnO 2 target was sputtered with a sputtering power of 60 W. The sputtered SnO 2 thickness was set as 20 nm, 40 nm, 60 nm, or 80 nm at a deposition rate of 0.43 Å s −1 . Solution-processed SnO 2 films were prepared by spin-coating 0.1 M precursor solution of SnCl 2 · 2H 2 O in ethanol at 3000 rpm for 30 seconds on clean FTO substrates. The SnO 2 thin films were finally heated in air at 180 °C for 1 hour. The MAPbI 3 perovskite was fabricated by a 2-step vapor deposition. The vapor deposition rate was controlled using a quartz sensor and calibrated after measuring the thickness of PbI 2 and MAI films. The sources were located at the bottom of the chamber with an angle of 90° with respect to the SnO 2 -coated substrates. The distance between source and substrate was 20 cm. The evaporation rate of both PbI 2 and MAI was maintained in a range of 1.5-2.0 Å s −1 . 120 nm PbI 2 and 280 nm MAI were evaporated to generate a resultant 400 nm MAPbI 3 film. The as-deposited films were annealed at 130 °C for 10 minutes in ambient condition of 65% humidity. The perovskite films were then covered by Spiro-OMeTAD, which composed of 80 mgmL −1 chlorobenzene, 17.5 μL Li-TFSI (520 mgmL −1 acetonitrile), and 28.5 μL TBP, was spin-coated at 3000 rpm for 30 s. The films were left in a desiccator overnight. To complete the devices, 100 nm gold was deposited by thermal evaporation at 1 Å s −1 as an electrode. The device area on FTO glass and ITO-PEN were 0.0314 cm 2 and 0.1 cm 2 , respectively. Device measurements. The AM1.5G solar spectrum was simulated by an Abet Class AAB Sun 2000 simulator with an intensity of 100 mWcm −2 calibrated with a KG5-filtered Si reference cell. The current-voltage (I-V) data were measured using a 2400 series sourcemeter (Keithley, USA). I-V sweeps (forward and reverse) were performed between −1.2 and +1.2 V, with a step size of 0.02 V and a delay time of 100 ms at each point.

Material characterization.
Field-emission scanning electron microscopy (JEOL JSM-7100F) and X-ray diffraction method (Bruker D8 X-ray diffractometer, USA) utilizing Cu K α radiation were used to study the thickness, morphology, roughness of the films, and phase characterization. The optical absorption and steady-state photoluminescence spectra were recorded on a Lambda 20 spectrophotometer (Perkin Elmer, USA) and InVia (Renishaw) micro raman/photoluminescence system, respectively. Ultraviolet photoelectron spectroscopy (Axis Ultra DLD) was used to determine the valence band maximum of SnO 2 films. Scanning probe microscopy (NanoScope III) (Digital Instruments) was used to characterize the surface roughness of films.