Regulating Lewis Acid‐Base Interactions to Enhance Stability of Tin Oxide for High‐Performance Perovskite Solar Cells

Perovskite solar cells are an attractive technology for renewable energy production. However, stability issues with the electron transport layer (ETL), particularly the colloidal tin oxide (SnO2) solution, can impact cell efficiency. In this study, a novel acidization treatment is introduced to reactivate long‐time stored SnO2 solutions, which previously led to low‐efficiency perovskite solar cells. The acidization treatment results in enhanced conductivity of the SnO2 layer, improved perovskite film quality, and ultimately increased efficiency. These findings show that a 1‐month stored SnO2 solution treated with acetic acid produces a device with a photoelectric conversion efficiency (PCE) of 20.9%, compared to 13.5% efficiency without treatment. With the addition of PEAI, the champion efficiency of the acetic acid‐treated device is 22.3%. This study provides a simple and effective engineering approach to fabricating high‐performance and stable ETLs for perovskite solar cells.


Introduction
10][11] Han et al. demonstrated the addition of 5ammoniumvaleric acid (5-AVA) stabilizer at the grain boundary to stabilize perovskite solar cells for 9000 h. [12]Perovskite solar cells have also exhibited comparable stability to other solar cells in extreme environments.Chen et al. combined CsPbI 2 Br and DOI: 10.1002/admi.202300328carbon nanotubes to develop the first perovskite solar cell that can operate at high temperatures up to 200 °C, surpassing the performance of Si and CuInGaSe solar cells. [13]Additionally, Huang et al. reported a surface modification method for the perovskite layer to produce a water-insoluble lead oxysalt passivation, resulting in perovskite materials with superior stability even after 60 s of water storage. [14][20] You et al. reported a simple and cost-effective method for preparing SnO 2 ETL from commercial SnO 2 colloidal solution. [21]Despite achieving nearly 20% efficiency, this method still has some defects that need to be addressed.To improve the stability and performance of SnO 2 ETLs, various modifications have been proposed, such as passivation using black phosphorus quantum dots (BPQDs), surface defect passivation using KCl, and carrier transfer mobility adjustment using ammonia halide.Tan et al. effectively passivated the surface defect state by adding KCl to the SnO 2 surface, and the increased K can also effectively passivate the grain boundary. [22]After effective doping, the V oc of the device was increased by nearly 0.05 V and the efficiency was increased by 2%.Li et al. adjusted the carrier transfer mobility of SnO 2 ETL by adding an appropriate amount of ammonia halide to SnO 2 precursor solution. [23]Although these modifications have shown significant improvements, little attention has been paid to the stability of the SnO 2 colloidal solution itself.
In this study, we explored the reactivation of long-stored SnO 2 colloidal solutions for the preparation of high-efficiency perovskite solar cells using various acid treatments.We found that treating a 30-day stored SnO 2 colloidal solution with acetic acid at a 1:1 molar ratio resulted in an impressive power conversion efficiency (PCE) of 19.11% for the resulting sample, which we named A-SnO 2 .In comparison, the PCEs of new SnO 2 colloidal solution (N-SnO 2 ) and 30-day stored old SnO 2 colloidal solution (O-SnO 2 ) based devices were 17.29% and 13.5%, respectively.We also treated the long-stored SnO 2 colloidal solutions with inorganic acids, such as HCl and HNO 3 , and weak acidic organic acids, such as succinic acid.Our further studies showed that the addition of acidic solutions can disrupt the state of SnO 2 particles in the colloidal solution, thereby avoiding storage-related issues.Additionally, we observed that the SnO 2 ETL prepared from the acidified solution had a rougher surface, which facilitated the film formation of the perovskite layer.This work presents a novel, simple, and practical method to reactivate inefficient long-stored SnO 2 colloidal solutions, which can be used to achieve high efficiency and long-term stability in perovskite solar cells.

Results and Discussion
The compactness, surface roughness, and hydrophilicity of the ETL will affect the subsequent preparation of the perovskite layer.Figure 1 shows the morphology of different ETLs. Figure 1a displays the morphology of the N-SnO 2 ETL, which was prepared using a newly diluted SnO 2 colloidal solution, as observed by scanning electron microscopy (SEM).The image shows that the SnO 2 colloidal particles are evenly distributed on the surface of FTO, resulting in a relatively compact and uniform ETL.In contrast, Figure 1b shows the morphology of the O-SnO 2 ETL, which was prepared using a 30-day stored SnO 2 colloidal solution.The SnO 2 nanoparticles tend to agglomerate at the bottom of FTO, and the top position has almost no SnO 2 nanoparticles, similar to pure FTO (Figure S1, Supporting Information).This may cause some pinholes in the top part, resulting in a less compact and uniform film than that of N-SnO 2 ETL. Figure 1c shows the morphology of the A-SnO 2 ETL, which was prepared using an acetic acidtreated 30-day stored SnO 2 colloidal solution.The acid-treated sample has a more roughened surface with a mesoporous-like structure, different from the N-SnO 2 ETL.The addition of acid could have destroyed the dispersion of SnO 2 in solution, leading to this mesoporous-like ETL structure.Figure S2 (Supporting Information) shows a picture of the solution after treatment, which changed from transparent to translucent.We measured the acidity and alkalinity of the diluted SnO 2 solution and observed a pH value close to 11.By adding acidification treatment, we reduced the pH value of the solution to nearly 7.This outcome suggests that the acidification treatment effectively altered the environment of the SnO 2 nanoparticles.To gain further dif-ferences in the external groups of the SnO 2 nanoparticles, we conducted FTIR testing on the resulting sample.Based on the Figure S3 (Supporting Information), it is evident that both the N-SnO 2 and A-SnO 2 samples exhibit a minor absorption peak ≈600 cm −1 , which can be attributed to the absorption of Sn-O or O-Sn-O.However, this peak is absent in the O-SnO 2 sample. [24]imilarly, at 1100 cm −1 , both N-SnO 2 and A-SnO 2 exhibit a distinct absorption peak, which is typically associated with the signal observed when nanoparticles are encapsulated within the surrounding aqueous solution.However, in the curve of O-SnO 2 samples, no absorption peak is observed at this wavenumber.The FTIR test results indicate that after acidification treatment, the external functional groups of SnO 2 nanoparticles were successfully restored to a state similar to that of N-SnO 2 nanoparticles.Although the surface of the ETL is rough, the uniformity of the sample surface is excellent, which may be more suitable for the uniform distribution of the perovskite solution.To exclude the effect of the FTO substrate on SnO 2 film formation, we further investigated the spin-coating of the solution on a flat Si at different storage times, and the surface roughness was measured using atomic force microscopy (AFM).Figures 1d,e displays the SnO 2 films obtained by spinning a newly configured SnO 2 solution and a 30-day stored solution, respectively.The comparison of the coordinate rulers shows that the sample obtained with the newly configured SnO 2 solution is flat, whereas the sample obtained with the 30-day stored solution is coarser.We also measured the particle size change of SnO 2 in the solution after storage for different times using a hydrodynamic test.Figure 1f shows that the SnO 2 particles in the solution stored for 30 days became significantly smaller, with a diameter of only 1.4 nm, while the diameter of SnO 2 in the newly configured solution was 3.6 nm.The smaller volume of SnO 2 nanoparticle makes it easier to agglomerate in the subsequent spin-coating and annealing process, which can destroy the properties of SnO 2 films.
[27]   showing a value of 15.8°.Figures 2b,c displays contact angle test results for the O-SnO 2 and A-SnO 2 samples, respectively, with O-SnO 2 having the highest value at 20.1°and A-SnO2 having the lowest value at 12.8°.Lower contact angle values indicate better hydrophilic performance for different ETLs.X-ray diffraction (XRD) measurements were used to test the crystallization of different SnO 2 samples, and the XRD characteristic peaks of all samples were found to be almost identical (Figure 2d), corresponding to the characteristic peaks of SnO 2 (JCPDS 71-0652). [28]To further understand the effect of storage time and acidification treatment on SnO 2 , we analyzed different samples using X-ray photoemission spectroscopy (XPS) (Figure 2e).The Sn XPS peak of the O-SnO 2 sample shifted to a higher energy potential, and the Sn XPS characteristic peak of the A-SnO 2 sample also shifted to a higher energy potential.This trend indicates that both long-time storage and acidification induced a negative charge in the vicinity of the Sn atom.The shift of the XPS peak indicates that A-SnO 2 has a relatively high electron  affinity, which means it can better accept electrons.When A-SnO 2 is applied to the perovskite electron transport layer, its electron affinity may lead to easier electron transfer from adjacent atoms to Sn ions, thereby improving device performance. [29,30]aman spectroscopy was also used to analyze the samples (Figure 2f).All three samples displayed obvious E g characteristic peaks at 472 cm −1 .Additionally, all samples displayed obvious surface phonon resonance peaks of IF 351, as they are composed of small-size nanoparticles. [31]The O-SnO 2 sample exhibited the highest IF 351 peak, indicating that the SnO 2 particles stored for a long time had become smaller.Transmission test curves for different samples are displayed in Figure S4 (Supporting Information).The sample prepared using N-SnO 2 had the highest transmittance due to its ultra-thin thickness and excellent uniformity.The A-SnO 2 sample had the lowest transmittance due to its rough surface, which is also supported by the SEM image results.
The uniformity of grain and film compactness are important parameters for measuring the quality of the perovskite layer.Figure 3a depicts the SEM image of the perovskite layer on N-SnO 2 ETL, indicating that the film is compact, and the grain size of perovskite is quite large.However, the perovskite grain size distribution is non-uniform, and some pinholes still exist in the perovskite layer, which reduces the device's efficiency.Figure 3b illustrates the morphology of the perovskite layer on O-SnO 2 .Unlike the N-SnO 2 sample, the perovskite grain boundary is unclear, and the perovskite's surface is also not smooth.We consider that the poor morphology may be due to the surface of O-SnO 2 , which disrupts the perovskite films' growth.Figure 3c shows the SEM image of the perovskite layer on A-SnO 2 ETL, indicating that the film is sufficiently compact.The perovskite grain size on A-SnO 2 ETL is larger and more uniform than that on N-SnO 2 or O-SnO 2 ETL, which may be due to the mesoporous-like ETL that provides better performance in growing big and uniform perovskite grains during the annealing process.Acetate ions can also bond with free Pb ions, which further improve the perovskite films' formation during the annealing process and prepares more compact perovskite films.Figure S5 (Supporting Information) shows the absorption curves of perovskite films deposited on different ETLs.Samples based on N-SnO 2 and O-SnO 2 ETLs have similar absorbance, while the sample based on A-SnO 2 ETL has higher absorbance in the 400-550 nm range, which is consistent with the transmission test.Meanwhile, the absorption edge of all samples is ≈800 nm, which is the same as that of FAPbI 3 type perovskite reported in the literature. [32,33]To obtain a clearer cross-sectional SEM morphology of the perovskite layer and A-SnO 2 ETL, we prepared the A-SnO 2 on the glass and then spincoated the perovskite layer on it, as shown in Figure 3d.In this image, the A-SnO 2 layer's thickness is close to 50 nm, and its cross-sectional morphology is more like the mesoporous structure than the planar type, and the perovskite layer has a thickness of ≈500 nm and good compactness.Figure S6 (Supporting Information) shows the cross-sectional morphology of the perovskite layer on FTO, which has a similar morphology to most works reported before, with nearly 500 nm thickness of the perovskite layer.
The current density-voltage (J-V) curves of different samples are shown in Figure 4.All samples were tested under AM 1.5G illumination from 1.2 to 0 V, with a scan rate of 100 mV s −1 and a step size of 0.04 V. Parameters such as open circuit voltage (V OC ), short circuit current density (J SC ), filling factor (FF), and power conversion efficiency (PCE) values for the best samples with different ETLs can be found in Table 1. Figure 4a displays  From the IPCE results, it can be seen that the values of N-SnO 2 and O-SnO 2 in the visible light region are both lower than those of A-SnO 2 .This may be due to the rougher surface of A-SnO 2 , achieving some light trapping effects.Figure 4b displays the efficiency distribution of 20 samples for the three ETLs.The A-SnO 2based devices have the highest efficiency and the smallest efficiency distribution variance, while the samples stored for 30 days have not only the lowest efficiency but also the largest device efficiency distribution variance.We also plotted the efficiency distribution histogram and Gaussian distribution of different devices in Figure S10 (Supporting Information).Hysteresis index (HI), which can be calculated as (PCE reverse -PCE forward )/PCE reverse , is a useful parameter for measuring device hysteresis. [34]Figure 4c shows the different scan direction curves of different ETL-based devices, from which we can calculate an HI of 0.21 for N-SnO 2 , 0.16 for O-SnO 2 , and 0.06 for A-SnO 2 -based devices.Figure 4d shows the 1000-h efficiency decay diagram of unpackaged devices stored in a dryer with a humidity of less than 20% and a temperature of 25 °C.The efficiency of the device prepared using O-SnO 2 ETL decreased by more than 50% after 200 h, while the device prepared using N-SnO 2 ETL decreased by ≈20% after 600 h of continuous irradiation.However, the device prepared using A-SnO 2 ETL remained stable, retaining an efficiency of 84% after 1000 h of continuous irradiation.Maxed power point (MPP) test also used in here to show the stability of different devices (Figure S11, Supporting Information), while the A-SnO 2 ETL based de-vice only has 0.49% decrease after 600 s measurement.Figure 4e depicts the variation in PCE of devices with different storage time ETLs and acidified ETLs.As the storage time of the solution increased, the efficiency of perovskite solar cells prepared using the stored SnO 2 solution decreased significantly.The devices prepared using the one-week stored SnO 2 solution achieved an efficiency of ≈15%, whereas the efficiency reduced to 13.5% with a 1-month stored solution.However, irrespective of the storage duration of the SnO 2 solution, acidification significantly improved the efficiency of the perovskite solar cell prepared with the treated solution, as shown by the blue curve.All devices achieved an efficiency of ≈20%, which was even higher than that of the device prepared using fresh SnO 2 solution (Figure S12, Supporting Information).Moreover, the perovskite solar cell prepared using the acidified solution exhibited stable and high efficiency even after long-term storage (Figure S13, Supporting Information).We believe that this stable performance could be due to the acidification breaking down the original solution state initially, and the longterm storage having no secondary effects.Figure 4f illustrates the J-V curves of devices prepared using 30 days stored SnO 2 solutions with different acid treatments.It can be observed from the figure that the device treated with acetic acid exhibited the highest efficiency.With the adding of PEAI, the efficiency of champion device is 22.3%.Additionally, we prepared a 0.5 cm 2 A-SnO 2based solar cell with a measured efficiency of 17.68% (Figure S14, Supporting Information).
To investigate the state of carrier transport in the device, we conducted optical and electrochemical tests on different ETLbased samples.We distinguished the photocarrier transport ability of different ETLs by photoluminescence (PL) spectroscopy, and the results are presented in Figure 5a.The sample prepared with O-SnO 2 has the highest characteristic peak, indicating that the excited carrier cannot be effectively transferred from the perovskite layer to the ETL, and its carrier transport capacity is the worst.In contrast, the value of the acidified A-SnO 2 sample is the lowest, even at 20 times magnification, indicating that it has the best carrier transport capacity.In order to further explore the internal situation of the device, we conducted time-resolved photoluminescence (TRPL) spectroscopy testing, as shown in Figure S15 (Supporting Information).We use second order functions to fit the lifetime of different ETLs, the fast decay lifetime ( 1 ) and slow decay lifetime ( 2 ).The average  of the N-SnO 2 is 515 ns and the A-SnO 2 ETL has relatively short average  of 618 ns, the O-SnO2 ETL-based device has the average  of 1040 ns.Through this test, we can also see that the acidified sample has better carrier transport ability.When device structures are similar, space charge limited current (SCLC) testing proves to be an effective method for evaluating the state of device defect states.We investigated the trap density of different ETL-based devices by performing the test on electron-only devices with the order of FTO/SnO2/perovskite/PCBM/Ag.The trap density (N t ) was determined by the trap-filled limit voltage (V TFL ) using the following equation: where d is the thickness of perovskite layer,  is the dielectric constant of perovskite, and  0 is the dielectric constant in vacuum). [35,36]The value of V FTL directly affects the value of N t .As shown in Figure 5b, the O-SnO 2 ETL-based device has the largest V TFL value, indicating that it has the largest trap density.The A-SnO 2 ETL-based device, on the other hand, has the smallest value of V FTL , indicating that it has the lowest trap density.Lower defect density of states can effectively avoid carrier loss during transmission and improve device performance.In the perovskite solar cell test, EIS test and TRPL test are common test methods used to evaluate solar cell performance and understand its working mechanism.Figure 5c shows the Nyquist plots and fitted curves of different perovskite solar cells with different ETLs tested under AM 1.5G illumination, and the value of fitted curves could be found in Table S1 (Supporting Information).The semicircle in the high-frequency region represents the transfer resistance (R ct ) of the device, while the semicircle in the low-frequency region represents the recombination impedance (R rec ) of the device. [37] smaller value of

Conclusion
In summary, the results of this study suggest that acetic acid treatment can be a simple and effective method to improve the performance of SnO 2 -based perovskite solar cells.The acidification process can passivate the defect sites in the SnO 2 ETLs and improve their conductivity, leading to better electron transport and reduced trap density.This, in turn, results in a higher efficiency of the solar cell.The use of acetic acid in the perovskite precursor solution can also improve the quality of the perovskite layer and lead to better light usage performance.Overall, the A-SnO 2 ETLbased device prepared in this study showed the best performance, with a PCE of 20.9%, indicating the potential of this method for the development of high-efficiency perovskite solar cells.

Experimental Section
Preparation of Acidified SnO 2 ETLs: F-doped tin oxide glass (FTO) was cleaned by ultrasonic vibration with acetone, ethanol, and deionized water for 30 min.The compared SnO 2 films were prepared by spin-coated SnO 2 (Alfa Aesar, 15% in H 2 O colloidal dispersion) solution that was diluted with 1:7 volume ratio with deionized water at 3000 rpm and 30 s.The modified solution was prepared by adding different acidic solution with different molar ratio of acidic to SnO 2 from 1:1 toward 1:8 in the long-time stored SnO 2 diluted solution.When the concentration of glacial acetic acid used was 23.21 mmol L −1 , the efficiency was the highest.The acidified SnO 2 films were prepared by spin-coating the modified SnO 2 solution at 3000 rpm and 30 s.All as-prepared structures were annealed at hotplate with 150 °C temperature for 10 min in air.
Fabrication of Perovskite Solar Cells: Perovskite solar cells were fabricated by a modified two steps process.PbI 2 was dissolved in DMF to obtain a clear PbI 2 solution with a concentration of 600 mg mL −1 .Next, 30 mL of the solution was taken and carefully dropped onto the surface of the conductive glass.Sufficient time was allowed for the PbI 2 solution to evenly spread and cover the glass surface before proceeding with the subsequent experiment.After spinning PbI 2 for 10 s when the spin speed was 2000 rpm, 30 mL isopropyl alcohol solution that contained 60 mg mL −1 FAI, 6 mg mL −1 MABr, and 6 mg mL −1 MACl was dropped into the substrate rapidly.The as-prepared substrates were transfer to a hot plate at 150 °C for 10 min in air.The HTL was prepared by spin-coating 72.3 mg mL −1 of Spiro-OMeTAD solution in acetonitrile at 2000 rpm for 40 s, and devices were oxidized in air for 36 h.Finally, a 50 nm thick Au electrode was deposited by thermal evaporation with a shadow mask (0.15 cm 2 active area).
Characterization and Measurement: The morphologies of the samples were characterized using field-emission scanning electron microscopy (Hitachi, SU8010).The films were investigated by X-ray diffractometer (Bruker, D8 Advance), X-ray photoelectron spectroscopy (Thermo, Escalab 250Xi), and Raman (Horiba, Labram Hr Evolution).The photoluminescence and time-resolved photoluminescence were tested with 530 nm laser (Edinburgh Instruments, LP320).The contact angle measurement was tested by DSA25E (KRÜSS).The absorbance measurement was tested by UV-2600 (Shimadzu).The photovoltaic parameters of solar cells were measured under Newport solar simulator AM 1.5G irradiation (100 mW cm −2 ) with a Keithley 2400 Source Meter, and IPCE curves were characterized by Zolix system.Electrochemical Impedance Spectroscopy was measured under AM 1.5G with an alternative signal amplitude of 10 mV and in a frequency range of 0.1 Hz-40 KHz (Autolab PGSTAT 302N) in glove box.M-S plots were obtained by the same workstation at an AC frequency of 10 3 Hz with an amplitude of 5 mV under light, the electrolyte was 1 m Na 2 SO 4 , and the reference electrode was Ag/AgCl.

Figure 1 .
Figure 1.SEM images of a) N-SnO 2 ETL, b) O-SnO 2 ETL, and c) A-SnO 2 ETL on FTO glass.AFM images of d) N-SnO 2 film, e) O-SnO 2 film on Si. f) Hydrodynamic measurement of N-SnO 2 and O-SnO 2 colloidal solution.
Figure 2a displays the contact angle test results for the A-SnO2 sample,

Figure 3 .
Figure 3. SEM images of perovskite layer on a) N-SnO 2 ETL, b) O-SnO 2 ETL, and c) A-SnO 2 ETL.d) Cross-sectional SEM image of perovskite layer on A-SnO 2 on glass.

Figure 4 .
Figure 4. a) J-V curves, b) Efficiency distribution plots, c) Hysteresis curves, and d) long-time stability curves of N-SnO 2 ETL, O-SnO 2 ETL, and A-SnO 2 ETL based perovskite solar cells.e) PCE variation of perovskite solar cells with different SnO 2 solution storage time based ETLs and acidified ETLs.f) J-V curves of different acid treated 30 days stored SnO 2 solution.(All devices are tested under AM 1.5G, reverse from 1.2 to 0 V, scan step of 0.04 V, and scan rate of 100 mV s −1 ).
Figure S7 (Supporting Information), where the AN-SnO 2 -based device has an efficiency of 19.45%.The incident photon-to-electron conversion efficiency (IPCE) spectrum is shown in Figure S8 (Supporting Information), and the integrated J SC of A-SnO 2 ETL based solar cell is 24.1 mA cm −2 .Figure S9 (Supporting Information) shows the IPCE curves of N-SnO 2 and O-SnO 2 -based devices and the integrated J SC of N-SnO 2 and O-SnO 2 is 21.3 and 22.1 mA cm −2 .
ct indicates better carrier transport ability of ETL, while a larger value of R rec means it is more difficult for the carrier to recombine in the solar cell.The R ct value of N-SnO 2 , O-SnO 2 , and A-SnO 2 -based devices is 260, 119, and 153 Ω, respectively, and the R rec value of N-SnO 2 , O-SnO 2 , and A-SnO 2based devices is 990, 535, and 667 Ω, The low R ct value and large R rec value of the A-SnO 2 sample demonstrate its excellent electron transmission ability and low recombination probability during the charge transport process.In order to further explore the role of acidification treatment, we also conducted test under different light intensities (Figure S16, Supporting Information).From the test results, theA-SnO 2 -based device has a smaller slope, indicating a lower trip density of ETL.

Table 1 .
Summary of photovoltaic parameters of solar cells with differentETLs with best PCE.SnO 2 , and A-SnO 2 ETLs.The device based on N-SnO 2 ETL has a V OC of 1.09 V, J SC of 22.6 mA cm −1 , FF of 69.9%, and PCE of 17.35%, which is comparable to reported literature results.The perovskite solar cell based on O-SnO 2 ETL has the lowest efficiency, mainly due to its low V OC , J SC , and FF.This low efficiency is attributed to using the solution stored for a long time to prepare the ETL, which is not conducive to high efficiency device fabrication.In contrast, the ETL-based device has the highest efficiency of 20.9%, with a V OC of 1.10 V, J SC of 25.22 mA cm −1 , and FF of 75.03%.The results demonstrate that acidification can effectively enhance ETL performance by producing a high-quality perovskite layer and facilitating charge transfer, leading to improved device efficiency.We also conducted acidification of a new SnO 2 (named as AN-SnO 2 ) solution and found that it can also improve device efficiency, as shown in