In less than a decade, the organic-inorganic PSCs with the general structure of ABX 3
), cesium (Cs þ ), or rubidium (Rb þ ) [1]; B¼ Pb 2þ ,S n 2þ or Ge 2þ ; and X ¼ Cl À ,B r À ,I À halides or mixture thereof, as well as a non-halide compounds such as acetates [2]) have undergone a dramatic efficiency improvement from 3.8% [3] to over 23% [4]. Such staggering achievements have been credited to the perovskite tunable bandgap [5], ultraviolet (UV) to infrared (IR) light absorption range [6], high charge mobility [7], long carrier diffusion length [8], exciton low binding energy [9], acute optical band edge [10], and bipolar semi-conductivity [11 ]. Rapid enhancements within these domains have made PSCs a viable opportunity toward bulk manufacturing using simple deposition and processing methods with the potential to rival silicon photovoltaics. So far, most of the investigations have been focused on four major fields such as ink engineering [12,13]; thin film deposition methods such as spray coating [14e17], slot-die coating [18e21], blade coating [22], gravure coating [23,24], and inkjet printing [25e27]; annealing methods such as hotplate [28], hot air blowing [29], and radiative annealing [30]; and fabrication techniques such as two-step coating [31], anti-solvent dripping [32], gas quenching [33], vacuum-flash assisted [34], and vapor assisted solution processing [35]. Despite dramatic progress within these areas, crucial processes are impeding full and rapid roll to roll manufacturing of PSCs. Aside from moisture and thermal stability concerns [36e40], rapid cost-effective annealing of the perovskite and ETL has remained a challenge for high throughput roll-to-roll manufacturing of PSCs.
To date, PSCs have been fabricated by taking the advantage of high temperature annealed metal-oxide electron transport layers (ETLs), such as ZnO, TiO 2 or SnO 2 , in a furnace (up to 500 C) for approximately an hour [41,42], or have been annealed through medium temperature ranges (~150 C) within minutes time frame [43e45]. Prolonged annealing of PSCs not only doesn't meet the energy payback time [46], but also necessitates the application of long ovens, batched ovens, or lowered web speeds within the rollto-roll setup. In addition, the prolonged heat treatment using temperatures greater than 150 C is a hindrance to the application of roll-to-roll favorable PET flexible substrates. Therefore, taking advantage of cost effective and rapid annealing processes for the charge carrier and perovskite films is a crucial factor for roll-to-roll manufacturing of PSCs.
Intense pulsed light (IPL) is an economical roll-to-roll applicable method which can be used for rapid thermal processing of various thin films [47e52]. IPL anneals PSC thin films through energetic flashes of light, typically from a xenon lamp, within millisecond time frame to crystallize the material without damaging the underneath layers. [20]. However, they annealed the SnO 2 ETL at 220 C through two ovens, each having one meter length, and created the perovskite films by heating the substrates for 10 min at 130 C. Despite such improvements, none of these investigations managed high throughput processing of both the ETL and perovskite films which would make them feasible for rapid and continuous automated manufacturing. Successful fabrication of efficient PSCs necessitates production of thin films which possess superior morphology and optimal material crystallization based on optimum thermal annealing and tuned ink engineering [56,57].
In this paper, we have rapidly processed PSCs from a Cs 0.05 (MA 0.85 FA 0.15 ) 0.95 PbI 3 mixed triple cation perovskite ink with CH 2 I 2 alkyl-halide additive, in which, the ETL and the perovskite photoactive films were made through a single step spin-coating, both on rigid glass-FTO and flexible PET-ITO substrates, and were subsequently annealed using millisecond duration pulses from a Xenon lamp. The ink engineering and coating parameters were tuned in a manner which could deliver adequate wet film quality before the annealing process and mirror-like films with superior morphology and high crystallinity after the annealing process in the ambient with relative humidity about 60%. This study was conducted by initially finding the suitable pulse energy and counts for the optimum annealing of SnO 2 ETL on rigid glass-FTO slides, studying the impact of CH 2 I 2 on the perovskite film quality, and, finally, applying the optimized conditions to make PSCs on both the rigid FTO coated glass and roll-to-roll applicable PET-ITO flexible substrates. The morphological inspection of thin films was carried out using SEM, the structural properties were investigated using XPS and XRD techniques, PL and IS were used to examine the interfacial charge transport capability, and current-voltage curves were obtained to measure the device performance. Our results provided rapid cost-effective fabrication of efficient PSCs without any further dopants or treatments which can be directly implemented for a low cost and scalable fabrication of PSCs through roll to roll.
Unless noted otherwise, all materials were purchased from Sigma Aldrich without further purification. PbI 2 (99.9985%), N,N-Dimethylformamide (99.8%), di-iodomethane (99%), and SnO 2 solution (15% in H 2 O colloidal dispersion) were purchased from Alfa Aesar. Methylammonium iodide (MAI), Formamidinium iodide (FAI), and cobalt dopant FK209 were purchased from Greatcell solar. The 2,20,7,70-Tetrakis(N,N-di-p-methoxyphenylamine)-9,90spirobifluorene (Spiro-MeOTAD) was acquired from Merck. FTO glass substrates (<20 U/sq., 2 cm  2cm 3 mm) and gold palettes (99.999%) were purchased from Hartford glass and Kurt J. Lesker Co., respectively. ITO coated PET sheets (60 U/sq.) were purchased from Sigma Aldrich.
FTO/ITO coated slides (2 cm  2 cm) were initially etched using zinc powder and 2 M HCl and were subsequently cleaned through sequential sonication in a 1:10 V/V diluted solution of Hellmanex detergent in DI water, DI water, ethanol, and DI water for 10 minutes each. After blowing with nitrogen to remove the remainder water, 20 minutes plasma treatment was applied to the cleaned substrates to improve the surface energy and thus the hydrophilicity. The SnO 2 ETL was then spin coated from a 1:4 V/V diluted solution of the SnO 2 colloidal dispersion in fresh DI water using the spinning rate of 2000 rpm for 30 seconds. Before annealing, the coated electrode places were wiped off with cotton swabs dipped in DI water. The substrates were immediately transferred to the IPL machine to perform the annealing process using different flash counts having different energy quantities and were later put under a UV lamp to provide surface treatment and remove the organic contaminants for 15 min. The perovskite ink was then passed through a 0.45 mm PTFE syringe filter and spin-coated on the SnO 2 coated substrates at two consecutive spinning rates of 1000 rpm for 10 s and 3000 rpm for 30 s. During the last 12 s of the spinning process, 200 mL of chlorobenzene solution was pipetted on the rotating substrate to take the solution to the supersaturation mode and form a transparent yellow wet film. The Cs 0.05 (MA 0.85- FA 0.15 ) 0.95 PbI 3 triple cation perovskite ink was made in a nitrogen filled glovebox in compliance to our previous ink engineering stated elsewhere [58]. In brief, 0.013 gr CsI, 0.6454 gr PbI 2 , 0.0343 gr FAI, and 0.18 gr MAI was initially dissolved in 1 mL DMF and 0.125 mL DMSO. After complete dissolution and the formation of a transparent yellow solution, 0.25 mL di-iodomethane was pipetted to the solution and stirred for 2 hours to provide the complete dissolution. During the spinning process, a dry air stream was put in the spin coater chamber to keep the relative humidity less than 10%. After heating the wet perovskite film for 20 seconds on a hotplate at 110 C, the slides were immediately taken to the IPL machine to perform rapid perovskite annealing through 5 pulses, each with 1.4 kJ energy. For flexible substrates, the wet film coated sheets were placed on a hotplate at 75 C for 10 s to form a dark brown film and were subsequently sintered through a single IPL shot with 900 J energy, an optimal condition where perovskite film damage was observed at higher energy quantities. Both hotplate and IPL annealing steps were carried out in the ambient with relative humidity of 60%. Right after perovskite annealing, the substrates were transferred to a nitrogen filled glovebox for the Spiro-MeOTAD hole-transport layer deposition. The Spiro-MeOTAD solution was made by initially dissolving a 72.3 mg of Spiro-MeOTAD in 1mL chlorobenzene. 28.8 mL4 -tert-butyl-pyridine was then pipetted to the solution and mixed. Later, a 17.5 mL of a stock solution of 520 mg/mL lithium bis (trifluoromethylsulphonyl) imide in anhydrous acetonitrile was pipetted to the solution and stirred until full dissolution. Finally, 29 mL of the cobalt dopant FK209 TFSI salt (300 mg/mL in anhydrous acetonitrile) was pipetted to the mixture to prepare the final solution. The Spiro-MeOTAD film was made by spin-coating 70 mL of the prepared solution at 1700 rpm for 30 seconds. Ultimately, an 80 nm gold electrode was deposited through a thermal evaporator.
XRD measurements were carried out within the 2q range between 10 and 60 using a Bruker AXS D8 X-ray Diffractometer equipped with a position sensitive detector (PCD) and an X-ray source of CuKa (l ¼ 0.1548 nm) with the scanning speed of 1 Sec/
Step and the step size of 0.02 . IPL annealing was done through a Xenon S-2000 rack-mounted sintering system with linear Xenon flash lamp placed within 5.7 cm from the substrates. Surface morphology inspections were done through a FEI Nova NanoSEM 600 SEM machine with an accelerating voltage of 10 kV and a working distance of 5 mm. The impedance spectroscopy (IS) and the current-voltage curves were obtained using an Autolab PGSTAT128-N potentiostat with the scanning rate of 0.1 V/Sec. Each cell was illuminated from the back side with an active area of 0.12 cm 2 using an AM 1.5 simulated light from a Newport LCS-100 solar simulator. The transmission and absorption spectra were obtained using PerkinElmer Lambda 950 UVeVis spectrometer between 250 and 800 nm wavelengths. PL analysis was carried out using a Renishaw inVia Raman microscope with a CCD detector and a 632 nm HeÀNe laser source. X-ray photoelectron spectroscopy (XPS) data was obtained from VG scientific MultiLab 3000 under ultrahigh vacuum pressure range within 10 À8 Torr using an Mgka radiation X-ray source (hn > 1253.6 eV) with respect to the carbon C1S peak position.
In this study we have successfully fabricated planar n-i-p PSCs with ITO-PET and FTO-glass/SnO 2 /Perovskite/Spiro-MeOTAD/gold thin film structure through a single step spin coating with rapid flash annealing from a xenon lamp on both SnO 2 ETL and the perovskite photoactive film under ambient conditions. We have previously shown that di-iodomethane (CH 2 I 2 ) could improve the power conversion efficiency of CH 3 NH 3 PbI 3 solar cells by improving surface coverage, film morphology, and crystallinity of the perovskite material [58]. In this work, we, first, optimized the IPL annealing parameters for the diluted solution of the commercially available SnO 2 colloidal dispersion solution through four different IPL conditions, including 5 and 10 pulse counts, each with different light energies of 1.4 kJ and 2.1 kJ. Later, we incorporated di-iodomethane into the triple cation (
) perovskite precursor to determine how CH 2 I 2 would interact with a different perovskite chemistry. Finally, the optimum conditions were picked to fabricate PSCs on both rigid and flexible substrates.
To determine the surface coverage and film quality of SnO 2 , SEM images of the FTO coated glass and SnO 2 coated thin films on glass-FTO substrates were taken and are shown in Fig. 1. The image of FTO, shown in Fig. 1(A), displays distinct grain boundaries under 200 nm magnification. In comparison, images of the SnO 2 film coated on glass-FTO, Fig. 1(B), appear granulated and the distinct grain boundaries shown in Fig. 1(A) are less apparent. This indicates that glass-FTO surface is covered with SnO 2 , and the disappearance of grain boundaries determines successful uniform deposition of the SnO 2 film which should increase shunt resistances.
To understand the feasibility of IPL annealing on SnO 2 ETL, FTO glass slides carrying the wet spin coated SnO 2 film were immediately transferred to the IPL machine and different flash counts carrying various energy quantities were applied on them. During the pulse annealing process, the water content evaporates from the wet film and the SnO 2 nanoparticles are crystallized and become capable for charge carrier transportation. It is expected for the charge carrier transportation capability to be depended on the degree of SnO 2 crystallization which, in turn, is depended on the heat flux per flash, as well as the duration and count of the flashes. Due to the limitation of our IPL machine being capable to provide 2 ms duration flashes, we only studied the crystallization through variation of the flash energy and counts. The applied pulse energies were 1.4 kJ and 2.1 kJ, and four different pulse counts of 1, 3, 5, 10, and 15 were selected for each pulse energy to anneal the SnO 2 film.
To study the chemical state, XPS was employed to probe the surface of our IPL annealed SnO 2 thin films and is highlighted in Fig. 2. Fig. 2(A) indicates the full XPS spectrum of the SnO 2 thin film fabricated through the optimal annealing condition of 5 pulses at 2.1 kJ energy, and the XPS patterns for 5 pulses at 1.4 kJ, 10 pulses at 1.4 kJ, and 10 pulses at 2.1 kJ conditions are included in Fig. S1 of the supplementary data. It can be clearly observed that all XPS graphs indicated a neat pattern demonstrating intense peaks of Sn and O without any other chemical species and functional groups which can suggest the formation of uncontaminated samples containing high concentration of Sn and O elements. The two small peaks at 293 and 295.5 eV are attributed to K2P 3/2 and K2P 1/2 orbitals of potassium ions used as the stabilizer in making the commercial product of the SnO 2 colloidal dispersion solution [59]. We assume that the bonding nature of surface level SneO observed in XPS spectra will be indicative of the bulk material given that the film is about 25e40 nm thick. The O1s peak can be deconstructed into two components indicative of OeSn 4þ and OeSn 2þ , as shown in Fig. 2(B) [60]. The OeSn 4þ peak is higher in binding energy than the OeSn 2þ peak; hence, the IPL processing condition that maximizes the O1s binding energy will maximize the surface concentration of SnO 2 . The O1S peak intensity for all the IPL annealing conditions remained similar which can be an indicator of unchanged oxygen concentration change. However, the binding energy of O1s peak was maximized when processed through 5 pulses at 2.1 kJ IPL condition, which can indicate enhanced SneO interaction and, thus, strengthened bonding that can better facilitate charge carrier transportation. Similar to O1s peak, the 3d 5/2 peak is a combination of the Sn 4þ and Sn 2þ oxidation states [60]. The Sn 2þ oxidation state arises from oxygen vacancies formed during the spin-coating process and is lower in binding energy than the Sn 4þ oxidation state [61]. The presence of Sn 4þ oxidation state indicates SnO 2 formation; therefore, the IPL processing condition that maximizes the binding energy of Sn 3d 5/2 peak is the optimal SnO 2 processing condition. Fig. 2(C) shows that the binding energy of Sn 3d 5/2 peak is maximized for SnO 2 thin films processed through 5 pulses at 2.1 kJ with respect to other processing conditions, which, in addition to the obtained results from O1S peak, suggests that this condition would yield the optimal crystallization of the deposited SnO 2 films and is consistent with other reports [62]. The peak intensity for all IPL conditions were almost identical, suggesting that the relative change in Sn concentration was insignificant. Table 1 contains Sn and O core level binding energies from the XPS surveys for each of the SnO 2 films annealed through different IPL conditions.
It is noteworthy that determining the crystallinity using XRD was attempted; however, the results did not reveal any peaks due to the film being extremely thin. Therefore, XPS was used to demonstrate a response of the material at different IPL processing Fig. 3(A) displays the Tauc plot, photon energy versus (ahn) 2 ,of SnO 2 which is used to measure the bandgap of the material. As it is apparent, SnO 2 thin films annealed through 5 pulses at 1.4 kJ and 2.1 kJ IPL conditions had the lowest and highest bandgap of 3.72 eV and 3.74 eV, respectively. Fig. 3(B) shows the PL spectra for SnO 2 thin films annealed through different IPL conditions. The peak intensity was minimized when SnO 2 films were fabricated through 5 pulses at 2.1 kJ IPL condition. A reduction in PL intensity indicates less charge recombination at the SnO 2 /perovskite interface [63]. Therefore, it can be suggested that the annealing of film through 5 pulses at 2.1 kJ condition would yield the maximum charge transport in comparison to the other annealing conditions. In addition, it can be observed that the PL peaks were slightly shifted for the SnO 2 films annealed through different IPL conditions, demonstrating a blue shift behavior for the films annealed through flashes carrying higher heat fluxes. This blueshift is in accordance with ~0.02 eV bandgap difference shown in Fig. 3(A) which can indicate diminished trap emissions and, thus, stronger exciton emissions at elevated flash heat fluxes and conforms with other reports [64,65].
We further studied the interfacial and in-layer charge transportation of IPL processed PSCs through impedance spectroscopy (IS) as shown in Fig. 4. Operating at open-circuit voltage under low light intensity has been shown to isolate the hole accumulation zone built up between perovskite absorber and ETL [66,67]. A reduction in overall impedance of the PSCs at open-circuit voltage indicates that charge carriers were more efficiently extracted from the perovskite absorber layer. The impedance was minimized for PSCs, in which, the SnO 2 layer was annealed through 5 pulses at 2.1 kJ condition. These results are consistent with the PL data shown in Fig. 3(B) and suggest that annealing through 5 pulses at 2.1 kJ would be the optimum IPL condition. Further study on material characteristics of the low temperature annealed SnO 2 films obtained from the commercial colloidal dispersion is provided in our other work [68].
Alkyl-halide additives improve solvent-precursor interactions, allowing for higher perovskite film quality and better surface coverage. In our previous study, we added CH 2 I 2 to CH 3 NH 3 PbI 3 perovskite precursor solution and observed improvements in power conversion efficiency, surface coverage, and film morphology [58]. Herein, we incorporated CH 2 I 2 into the triple cation perovskite precursor solution to investigate the possible performance improvement of PSCs after introducing CH 2 I 2 into a new perovskite chemistry. To discern whether CH 2 I 2 improved the perovskite film quality and surface coverage, top-down SEM images of the annealed perovskite film, with and without CH 2 I 2 after IPL annealing, is shown in Fig. 5. Fig. 5(A) suggests that a fine perovskite film without observable pinholes was formed when CH 2 I 2 was incorporated into the precursor solution. In comparison, Fig. 5(B) contains a top-down SEM image of IPL annealed perovskite without the diiodomethane additive. The image indicates the formation of smaller grain sizes with dendrite structures within the perovskite film, indicating natural crystallization within the wet film prior to the annealing process. Formation of dendrite structures can reduce the quality of thin film, while smaller grain sizes can enhance charge recombination at the grain boundaries due to the formation of more grain boundary density within the film, thus, decreasing the photovoltaic performance. To find out the average grain sizes, ImageJ was used to obtain the size and area of individual grains within the captured top surface SEM image of the perovskite films after IPL annealing. The results indicated average crystal size and area of 201 nm, and 0.031 mm 2 for the CH 2 I 2 containing perovskite films, and 138 nm, and 0.015 mm 2 for the films made from pristine solution, delineating almost 1.45 times crystal growth for the devices made with aid of di-iodomethane additive.
The histograms of grain size and area are included in Fig. S2 of the supplementary data. The SEM images suggest that CH 2 I 2 additive improved surface coverage, grain growth, and film quality of the triple cation perovskite layer by implementing two significant roles. First, the addition of diiodomethane could potentially increase the solution boiling point and delay the unfavorable natural crystallization of the as-spinning solution, thus, hindering the inadequate crystallization of the film and insufficient grain growth upon annealing. Second, CH 2 I 2 could improve the crystallinity of perovskite layer due to its role as an iodide source. When exposed to IPL, the CeI bond is broken, yielding equal parts CH 2 I þ and I À . The I À ions are then incorporated into the perovskite material by filling iodine vacancies [58]. The released iodine from alkyl halide could replace the lost iodine through evaporation during the annealing processes, thus avoiding the formation of defected perovskite black phase upon the completion of annealing. Therefore, it is expected that di-iodomethane addition would improve the device performance through enhanced charge carrier production and extraction to the transport films, thus decreasing and increasing the series and shunt resistances, respectively. Crystallization kinetics, surface coverage, and improved crystallinity are all examples of how the CH 2 I 2 additive, in cooperation with IPL, can improve the quality of the perovskite absorber layer. Fig. 5(C) shows the absorbance spectra for the FTO/SnO 2 / perovskite and FTO/SnO 2 structures after optimum IPL annealing. Similar to our previous work for CH 3 NH 3 PbI 3 photoactive films, the triple cation perovskite films were brown in color after short-term hotplate annealing and became darker after IPL annealing, thus, becoming capable of efficiently absorbing the light energy. The FTO/ SnO 2 /perovskite structure demonstrated high absorption within 400e750 nm wavelengths range; whereas, the FTO/SnO 2 structure possessed little to no absorption within 400e800 nm range. This indicates that the full AM1.5 spectrum reaches the perovskite absorber without being consumed by the ETL. The final device will be constructed in a manner where light travels through glass, FTO, SnO 2 , then strikes perovskite where it can be absorbed.
XRD patterns shown in Fig. 2(D) reveal information about crystallinity of the perovskite layer after IPL annealing. The intense 2q peaks around 26.5 , 33.75 ,37 .5 ,5 1 .2 , and 54.5 are indicative of FTO/SnO 2 , and the strong peaks around 14.06 ,20 , 24.4 , 28.4 , 31.8 , 34.9 , 40.5 ,4 3 , and 50.5 are associated to (110), ( 112), ( 202), ( 220), (310), ( 312), ( 224), (314), and (404) crystal planes which indicates the formation of perovskite tetragonal black phase. The perovskite tetragonal black phase is the preferred conformation for PSCs. Fortunately, the patterns did not reveal the PbI 2 peak around 12.4 which indicated the complete formation of a pure perovskite black phase. Formation of a pure perovskite black phase would promote higher PSC efficiency due to fewer series and higher recombination resistances within the perovskite absorber layer. Therefore, the applied IPL annealing condition for the perovskite film could sufficiently crystallize the layer without causing post degradation of the film nor being insufficient to leave the perovskite yellow phase unconverted to black phase.
To verify the observed characterization results, we finally fabricated PSCs to measure the device functionality. The overall performance summary of PSCs fabricated using four different IPL conditions on SnO 2 ETL is shown in Table 2, and the current density-voltage (J-V) curve of champion cell for each IPL condition is shown in Fig. 6. The results indicate that exposing the SnO 2 thin films to higher pulse energy flux of 2.1 kJ can provide better device performance than the lower pulse energy of 1.4 kJ for all the range of applied pulse counts, which implies the significance of heat flux intensity on material crystallization. On the other hand, lower pulse counts for the higher pulse energy were found to provide the maximum device performance, determining the 5 pulse at 2.1 kJ as the optimal IPL annealing condition which could result in the highest device performance. As shown by the results, varying the IPL annealing conditions for SnO 2 ETL could considerably affect the current flow throughout the fabricated cells while maintaining the voltage and fill factor. This observation can indicate the significance of IPL annealing conditions on the crystallization of the metal oxide material which play a vital role on facilitating charge extraction at the thin film interface and throughout the film. It was found after several experiments that pulse counts higher than 10 could not enhance the current flow, while pulse counts lower than five resulted in dead cells, demonstrating a resistor behavior. This is consistent with the XRD pattern of unannealed drop casted SnO 2 shown in Fig. S3 of the supplementary data which demonstrated broad peaks with low intensity and, hence, its amorphous nature. Therefore, thermal annealing is required to crystallize the material and there is an existing optimal annealing condition, in which, exceeding the limit would not result in an operative or economical fabrication of device. It is noteworthy that providing pulse energies higher than 2.1 kJ were not possible due to the limitation of our current annealing machine. Perhaps higher pulses than the studied conditions followed by lower pulse counts can result in further enhanced device performance. Further study over IPL annealing conditions on PSCs thin film morphology and device efficiency will be discussed in the future works. In this interim, we have provided the photovoltaic parameters and conducted the study over the impact of IPL annealing on SnO 2 structure only for the IPL annealing conditions within the range of functional devices.
To further investigate the scalability of rapid thermal annealing through IPL on both SnO 2 and perovskite thin films, we also fabricated PSCs on roll-to-roll favorable flexible PET-ITO substrates. Fig. 7 shows the J-V characteristics of the forward and reverse scan of champion rigid and flexible substrates annealed through the optimum SnO 2 IPL annealing condition of 5 pulses at 2.1 kJ. The J-V characteristics of rigid and flexible PSCs indicated the maximum reverse scan efficiency of 12.56% and 7.6%, open circuit voltage (V oc ) of 1.02 and 0.993 V, short-circuit current density (J SC ) of 15.78 and 11.87 mA/cm 2 , and the fill factor of 78.3% and 64.75%, as well as the maximum forward scan efficiency of 11.33% and 6.16%, V oc of 0.992 and 0.979 V, short-circuit current density of 14.92 and 11.06 mA/ cm 2 , and the fill factor of 76.7% and 56.95%, demonstrating relatively low J-V hysteresis for both rigid and flexible PSCs. It can be observed from the experimental results that the flexible PSCs had a lower performance compared to rigid cells. This lack of performance can be attributed to the higher resistance of ITO coated sheets compared to FTO slides, handling of the flexible substrates during fabrication processes, as well as unsuitable nature of spincoating method to form highly uniform films with superior quality, particularly for coating water based SnO 2 films on smooth ITO coated sheets compared to rough FTO substrates, which require the application of actual roll-to-roll fabrication setup and will be addressed in future works. Our results delineated that CH 2 I 2 additive is also applicable for another perovskite chemistry and its synergy with sequential step IPL annealing is a potential candidate for rapid and economical industrial scale manufacturing of PSCs on both the rigid and roll-to-roll favorable flexible substrates.
In this study, fabrication of PSCs through intense flashed annealing was demonstrated for both the SnO 2 electron transport layer (ETL) and di-iodomethane (CH 2 I 2 ) assisted Cs 0.05 (MA 0.85- FA 0.15 ) 0.95 PbI 3 mixed triple cation perovskite thin films both on rigid glass-FTO and flexible PET-ITO substrates. It was shown that the addition of CH 2 I 2 could eliminate natural perovskite crystallization by increasing the precursor ink boiling point and form a superior dry film morphology with bound grains as a result of iodine release during the annealing processes. This resulted in PSCs with efficiency and fill factors as high as 12.56% and 78.3% for rigid FTO-glass substrates, and 7.6% and 64.75% for flexible ITO-PET sheets, when the SnO 2 ETL was optimally annealed through 5 pulses of IPL, each carrying 2.1 kJ energy. It was shown that sequential step IPL annealing can be used for rapid fabrication of PSCs and eliminate the need for prolonged annealing of the charge carrier and photoactive thin films. Our results demonstrated the possibility of obtaining efficient cells with high fill factor through IPL annealing from a single-step spin-coating for various PSC thin films without the application of any dopants or surface treatments.
To our knowledge, this work provided the fastest route of fabricating PSCs so far, and is the first reported manuscript on rapid thermal annealing of the SnO 2 ETL and di-iodomethane assisted triple cation perovskite absorber films within seconds of time frame cumulatively, which can provide the possibility of a swift and cost-effective fabrication of PSCs through roll to roll.