Anion‐Vacancy‐Defect Passivation of a 2D‐Layered Tin‐Based Perovskite Thin‐Film Transistor with Sulfur Doping

Metal halide perovskites have attracted a considerable amount of research attention with significant progress made in the field of optoelectronics. Despite their outstanding electrical characteristics, structural defects impede their potential performance due to the polycrystalline nature of solution‐processed perovskite films. Herein, the effective p‐type doping and defect passivation of phenethylammonium tin iodide ((PEA)2SnI4) perovskite films using xanthate additives as a sulfur source is reported. Sulfur can be introduced to the iodine vacancies mainly at the grain boundaries of the perovskite film, passivating the electrical defects originating from the iodine vacancy and increasing the hole concentration. The Fermi‐level shift toward the valence band maximum of the sulfur‐doped perovskite film is confirmed using ultraviolet photoemission spectroscopy, resulting in p‐type doping. Finally, the electrical performance improvement for the 0.2% sulfur‐doped (PEA)2SnI4 thin‐film transistor with a mobility of 1.45 cm2 V−1 s−1, an on/off ratio of 2.9 × 105 is demonstrated, and hysteresis of 10 V is reduced.


Introduction
Metal halide perovskites (MHPs) are rapidly emerging as promising semiconductor materials due to their superior electrical properties and ease of fabrication. [1][2][3] Due to their outstanding optoelectronic properties, MHPs show great potential, particularly in the field of photovoltaics. Consequently, the certified power conversion efficiency of perovskite solar cells (PSCs) has achieved a value of 25.5%, which surpasses the efficiency www.advelectronicmat.de multicomponent perovskites containing mixed cations and anions have been used to overcome the disadvantages of pure perovskite compounds. [27,28] Nevertheless, these approaches may be of concern in regard to the complexity, reproducibility, and limitation of configuration due to their structural instability. Additive engineering is one of the promising strategies used for defect passivation and enhancement of the electrical properties of perovskites. Most of the additive engineering approaches have been aimed at modifying the optoelectronic properties related to the bandgap or improving the air/moisture instability. [29,30] Furthermore, it has been typically designed for cation doping using alkali metal components to suppress the formation of halide defects, whereas anion doping has been rarely reported. [31] Alternatively, anion doping can provide potential ways to improve the electrical properties due to the possibility of the generation of excess hole carriers in halide perovskites.
Herein, we propose an anion doping strategy in (PEA) 2 SnI 4 perovskite by simply introducing xanthate additives as a sulfur source into the precursor solution. Sulfur doping can generate additional hole carriers by the passivation of iodine defects (V I + ) at the GBs, allowing effective charge-carrier transport throughout the perovskite film. The hole doping effect has been confirmed using ultraviolet photoemission spectroscopy (UPS), which shows a shift in the Fermi energy level toward the valence band maximum (VBM). Upon 0.2% sulfur doping in perovskite, the optimized TFT exhibited enhanced electrical characteristics with a mobility of 1.45 cm 2 V −1 s −1 , an on/off ratio of 2.9 × 10 5 , and reduced hysteresis of 10 V.

Results and Discussion
The possible mechanism for doping in (PEA) 2 SnI 4 perovskite using xanthate (Xth) additives is illustrated in Figure 1a.
In general, the Sn-I chemical bond is relatively weak with a low V I + formation energy of 0.08 eV due to the volatile nature of halide species in the (PEA) 2 SnI 4 perovskite. [32,33] Since the vacancies of the iodide anions (I − ) lead to the formation of undercoordinated central metal cations (Sn 2+ ), the formation of iodine vacancy (V I + ) defects cannot be avoided. Moreover, these defect sites are prone to being located at the GBs with a high density of trap states. For the passivation of these defect sites, the Xth additives act as a sulfur source and are directly added to the perovskite precursor in subpercentage in the molar ratio. Sulfur atoms can be introduced substitutionally at the positively charged iodine vacancies due to their similar ionic radius to that of iodine atoms. [34,35] The sulfur atoms can coordinate to the undercoordinated Sn 2+ ions in the perovskite lattice upon the decomposition of the Xth additive. [36] Similar to previous report, the thermogravimetric analysis (TGA) of potassium ethyl xanthate (PEX) and SnI 2 mixture (PEX:SnI 2 = 2:1) shows that the decomposition of the xanthates to metal sulfides occurs in a single step around 80-120 °C (Figure 1b). [37] The mixture was completely decomposed at ≈120 °C, allowing the injection of S 2− into the perovskite precursor during annealing. Accordingly, divalent sulfides (S 2− ) will partially fill the iodine vacancies, leading to hole doping to balance the charge neutrality. [38] As shown in Figure 1c, X-ray photoelectron spectroscopy (XPS) was performed to confirm the presence of S atoms in the sulfur-doped (PEA) 2 SnI 4 film. Figure 1. a) Schematic of the possible mechanism for the substitutional doping of (PEA) 2 SnI 4 perovskite upon adding the Xth additive. b) Thermogravimetric analysis (TGA) of the decomposition of potassium ethyl xanthate (PEX) and SnI 2 mixture (PEX:SnI 2 = 2:1). c) X-ray photoelectron spectroscopy (XPS) peak of S 2p for the pristine and sulfur-doped (PEA) 2 SnI 4 film. d) X-ray diffraction (XRD) patterns obtained for (PEA) 2 SnI 4 thin films prepared with different concentrations of the Xth additive from 0% to 0.5%.

www.advelectronicmat.de
The S 2p signal was observed at ≈163 eV in the case of the 10% sulfur-doped film, whereas no distinct peak was revealed for the pristine film. Since sulfur has low relative sensitivity factor (RSF) of 1.67 for S 2p, (cf. RSF of Sn 3d 5/2 = 15.03), the reliable quantitative detection of S with XPS analysis is challenging for the S doped (PEA) 2 SnI 4 . The 10% sulfur-doped (PEA) 2 SnI 4 shows clear S 2p peak with ratio of S/Sn = 0.10, which is well matched with the initial PEX content. X-ray diffraction (XRD) analysis was used to confirm whether an undesirable phase transformation occurred in the layered structure of (PEA) 2 SnI 4 upon doping with S. There was no significant shift observed in the peak positions in the XRD patterns, as shown in Figure 1d, still exhibiting the (00l) diffraction peaks that are in good agreement with those previously reported. [39] For the pristine film, the full width at half maximum (FWHM) of the most intense (002) peak at 5.47° was slightly smaller than the 0.5% sulfur-doped film, which implies the larger crystallite size in the direction perpendicular to the substrate according to the Scherrer equation. Along c-axis of (PEA) 2 SnI 4, the average crystallite size of the thin film was decreased from 28.2 nm for the pristine film to 24.0, 19.7, and 16.9 nm for the 0.1%, 0.2%, and 0.5% sulfur-doped films, respectively. In addition, no new secondary phases were detected, suggesting that the dopants were incorporated into the perovskite lattice without segregation. It is found that the Xth dopant had no effect on the distortion of the layered perovskite structure at even higher doping regimes up to a doping level of 10% ( Figure S1, see the Supporting Information). The bright images indicate the regions where charges were accumulated, [40] which appear to be mainly located at the GBs. In addition, the overall grain size decreases upon increasing the amount of dopant used, which is consistent with the trend of decreasing horizontal and vertical crystallite sizes from microscopy (AFM and SEM) and XRD measurement. Although the addition of Xth did not form a noticeable amount of secondary phase in XRD, the reduced grain size of the sulfur-doped film was attributed to the Xth additives, which may act as impurities at the GBs and hinder grain growth in the perovskite film.
Nevertheless, the corresponding conductive atomic force microscopy (C-AFM) images demonstrate that the doped films were more conductive than the pristine film. Figure 2e-h shows the C-AFM current-mapping images and topography images of the pristine and 0.2% sulfur-doped perovskite films, respectively. The data were obtained at five random locations in each film to acquire the average current on the surface. The average current of the pristine film was 1.74 nA, whereas that of the doped film was increased to 9.54 nA. In the case of the pristine film, some areas in the grain interior exhibit a low current signal with a dark-brown color in the image. By contrast, www.advelectronicmat.de the doped film displays a relatively high current signal in the grain interior with a homogeneous distribution of the current signal in the areas. We propose that the dopants contribute to the overall conductivity not only at the GBs but also at the grain interior of the film. In particular, the current measured at the GBs was much higher than that observed at some of the grain regions, demonstrating the enhanced electrical performance of the film at the GBs. The increased current suggests that the electrical conductivity was improved, which is beneficial for charge carrier transportation. [41] Additionally, the root-mean-square (R q ) of the pristine film obtained from the AFM topography images was 6.49 nm, whereas that of the doped film was reduced to 5.08 nm, demonstrating its moderately smooth surface with relatively low surface roughness. Moreover, the doped film clearly shows better surface coverage, which has a more homogenous distribution with a low density of pinholes and voids. These results are expected due to the presence of many nucleation sites in the doped film formed due to the addition of small amounts of the Xth dopant. The nucleation sites induced by the dopants can initiate crystal growth during the formation of the perovskite film, promoting the continuous perovskite film with low surface roughness. Therefore, the incorporation of the Xth additives can improve charge carrier transport with an improved surface coverage of the perovskite film. Upon defect passivation, it was evident that the charge carriers in the doped film can be efficiently transported via the more conducting GBs as well as the grain interior.
To further confirm the location of the sulfur dopants in the perovskite film, we investigated the distribution of the surface components in the doped film using time-of-flight secondaryion mass spectroscopy (TOF-SIMS). In the obtained 2D TOF-SIMS images, Figure 2i shows the total signal clearly indicating the interface between the grains and grain boundaries. The distribution of the I ions was mainly located inside the grain interiors consisting of a SnI 6 octahedral network, as shown in Figure 2j. Figure 2k shows that the sulfur ions were abundant at the GBs in the doped film. Although the signal was also collected inside the grain interior, the intensity was significantly lower than that observed at the GBs. This result indicates that the sulfur components were mainly located at the iodine vacancies formed at the GBs in the perovskite film. Consequently, it was confirmed that the location of the surface passivation upon sulfur doping mainly occurs at the GBs.
We performed UPS measurements to identify the electrical doping effect in (PEA) 2 SnI 4 prepared using the Xth additive, which was directly associated with the work function and valence band energy. [42] The intersection point of the intensity at high binding energy represents the secondary electron cutoff energy (E cut-off ), as shown in Figure 3a. The work function was calculated from the difference between the incident photon energy and E cut-off . Figure 3b shows the intersection point at low binding energy, which indicates the VBM of the Fermi energy level, i.e., E v − E F . By combining the work function and E v − E F values, we determined the energy band diagram of the pristine and sulfur-doped (PEA) 2 SnI 4 films, as shown in Figure 3d. The conduction band minimum was deduced from the optical bandgap derived from Tauc plots (Figure 3c). There were no significant differences in the optical bandgap, yielding a constant value of 1.97 eV, whereas the Fermi energy level was shifted toward the VBM upon sulfur doping. The E F downshift indicates the increase in the hole concentration due to the addition of holes to the perovskite structure. [43] The E F shift can be attributed to the substitutional doping of sulfur atoms at the iodine vacancies in the perovskite lattice. Moreover, the negatively charged sulfur dopant can contribute to the generation of additional mobile carriers in the semiconductor layer, increasing the hole carrier concentration. [35] Furthermore, we observed an  Figure S2, see the Supporting Information). From the results, it can be concluded that effective p-type doping is achieved using the sulfur dopant, which simultaneously passivates the iodine vacancies.
Based on the p-type doping effect, we investigated the TFT performance via (PEA) 2 SnI 4 perovskite prepared using different sulfur dopant amounts. An illustration of the TFT architecture is shown in Figure 4a, wherein the channel layer is composed of the perovskite thin film. Figure 4b and Table 1 show the transfer characteristics curves obtained for the TFT and the extracted parameters of the 0 (pristine), 0.1%, 0.2%, and 0.5% sulfur-doped devices, respectively. The pristine device exhibits a relatively poor electrical performance with an average mobility of 0.45 cm 2 V −1 s −1 and large hysteresis. In the doped channels, an increase in the on-state current by at least an order of magnitude was observed, including a shift in the threshold voltage (V TH ) to the positive gate voltage bias. This trend in the V TH shift was attributed to an increase in the hole carrier concentration, inducing p-type doping. The optimized device was the 0.2% sulfur-doped TFT, which showed the best performance with a mobility of 1.45 cm 2 V −1 s −1 , an on/off ratio of 2.9 × 10 5 , and slightly reduced hysteresis of 10 V. Figure 4c,d shows the output characteristics of the pristine and the optimized devices, demonstrating an increase in the current upon doping, which was in agreement with the transfer characteristics (all related output curves can be found in Figure S3, see the Supporting Information). All of the samples exhibit changes in current between the forward-sweep direction and backwardsweep direction in the transfer characteristics. As shown in Figure S4 (Supporting Information), the negative bias stability tests of pristine and sulfur-doped (PEA) 2 SnI 4 TFTs showed the improved bias stability of the 0.2% sulfur-doped device while stressing at V GS = V DS = −20 V.
The current observed for the forward direction was larger than the current for the backward direction, displaying clockwise hysteresis behavior. The observed hysteresis can be attributed to hole carrier trapping due to the migration of ions in the semiconductor layer. We conducted temperature-dependent measurements from 100 to 300 K to separate the ion migration effect in the channel. Figure 4e shows the transfer characteristics curve obtained at 100 K. As shown in Figure 4f-h, all of the devices showed improved electrical performance as the temperature decreased. The negligible hysteresis window (ΔV) was observed due to the suppression of ion conductivity at low temperatures. In the case of the optimized device with 0.2% Xth, the field-effect mobility (µ FE ) linearly increases from 1.45 to 2.1 cm 2 V −1 s −1 . The number of trapped hole carriers (ΔN T ) was significantly reduced from 1.5 × 10 12 to 2.4 × 10 11 cm −2 , corresponding to the reduced hysteresis window (ΔV). The subthreshold swing (SS) also decreased from 2.6 to 0.68 V dec −1 , suggesting a decrease in the trap density. Overall, the S-doped (PEA) 2 SnI 4 TFT device exhibits an enhanced hole mobility with reduced hysteresis when compared to those prepared without S-doping.

Conclusions
We have demonstrated the effective p-type doping and defect passivation of the (PEA) 2 SnI 4 perovskite films upon adding Xth as a hole dopant. It was confirmed that the Xth additives can contribute to the conductivity at the GBs, facilitating the charge transport throughout the film. TOF-SIMS tomography proved that the sulfur atoms from the Xth additives were mainly localized at the GBs rather than the grain interior. The shift in the Fermi energy (E F ) was attributed to doping and verified using UPS with respect to the work function and valence band edge. Consequently, (PEA) 2 SnI 4 perovskite with sulfur doping exhibited improved electrical performance with a mobility of 1.45 cm 2 V −1 s −1 , an on/off ratio of 2.9 × 10 5 , and reduced hysteresis of 10 V. Moreover, the hysteresis behavior observed in the transfer curves can be significantly reduced using the low-temperature measurements at 100 K. We believe this work presents an effective method for passivating defects and increasing the electrical properties in halide perovskite-based electronic and optoelectronic devices.

Experimental Section
Solution Preparation: The perovskite solution was prepared using 0.2 m phenethylammonium iodide and 0.1 m of tin (II) iodide in 3:1 mixed solvent comprised of N,N-dimethylformamide (DMF) and dimethyl sulfoxide (DMSO), which was stirred at 60 °C for 3 h. The 0.01 m potassium ethyl xanthate (PEX) in DMF was added to the precursor solution at a molar ratio of 0.1%, 0.2%, and 0.5%, and stirred at 60 °C for 30 min.
Device Fabrication: For the fabrication of the perovskite TFTs, p ++ Si wafer with thermally oxidized 200 nm SiO 2 was used as the substrate and cleaned in acetone and isopropyl alcohol using an ultrasonic bath for 10 min. The cleaned Si/SiO 2 substrate was treated with oxygen plasma for 20 min. The precursor solution was spin-coated onto the substrate at 4000 rpm for 30 s and annealed at 120 °C for 10 min. For the source and drain electrodes, 40 nm thick gold patterns were thermally evaporated on the perovskite film using a shadow mask. The channel width and length were 1000 and 50 µm, respectively. All of the fabrication steps were conducted in an Ar-filled glove box.
Characterization: The morphology of the perovskite films was obtained using field-emission scanning electron microscopy (FE-SEM, JSM-6700 F). A commercial AFM system (Park Systems, NX-10) was used to perform the C-AFM measurements with a Pt-coated AFM tip (Multi75E-G, BudgetSensors) at a bias voltage of 2 V and scan rate of 0.5 Hz. A TOF-SIMS V instrument (ION-TOF GmbH) was utilized for TOF-SIMS imaging using a 25 keV Bi 3 2+ primary ion beam source. X-ray diffraction (XRD) was performed using a X'pert Pro (PANalytical) with Cu Kα radiation at 40 kV/30 mA. UPS was carried out using a He-I phonon source (21.2 eV) to probe the work function and VBM. Optical transmittance spectra were obtained using a Cary5000 UV-VIS-NIR spectrophotometer (Agilent Technologies, USA) to determine the bandgap of the perovskite films. Thermogravimetric analysis (TGA) was conducted in the temperature range of room temperature to 300 °C at a heating rate of 10 °C min −1 under a N 2 atmosphere. X-ray photoelectron spectroscopy (XPS) was performed using K-alpha+ (ThermoFisher Scientific, USA) with Al K-α radiation.
Electrical Characterization: The TFT device was characterized on a Keithley 4200-SCS instrument (Tektronix, USA) and MST-4000A probe station (MS TECH, Korea) in a vacuum system. For the thermaldependent studies on the perovskite TFT devices, the samples were cooled using liquid argon in the range of 100-300 K. The field-effect saturation mobility of the TFT device was calculated as follows where L and W are the channel length and width, respectively and C i is the areal capacitance of SiO 2 . The number of trapped carriers (ΔN T ) was obtained from the hysteresis window (ΔV) in the transfer curve, which can be calculated as follows where the ΔV is the difference in the gate voltage between the forward and backward sweep direction at a drain current of 10 −7 A.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.