Defect Engineering in Earth‐Abundant Cu2ZnSnSe4 Absorber Using Efficient Alkali Doping for Flexible and Tandem Solar Cell Applications

To demonstrate flexible and tandem device applications, a low‐temperature Cu2ZnSnSe4 (CZTSe) deposition process, combined with efficient alkali doping, was developed. First, high‐quality CZTSe films were grown at 480 °C by a single co‐evaporation, which is applicable to polyimide (PI) substrate. Because of the alkali‐free substrate, Na and K alkali doping were systematically studied and optimized to precisely control the alkali distribution in CZTSe. The bulk defect density was significantly reduced by suppression of deep acceptor states after the (NaF + KF) PDTs. Through the low‐temperature deposition with (NaF + KF) PDTs, the CZTSe device on glass yields the best efficiency of 8.1% with an improved VOC deficit of 646 mV. The developed deposition technologies have been applied to PI. For the first time, we report the highest efficiency of 6.92% for flexible CZTSe solar cells on PI. Additionally, CZTSe devices were utilized as bottom cells to fabricate four‐terminal CZTSe/perovskite tandem cells because of a low bandgap of CZTSe (~1.0 eV) so that the tandem cell yielded an efficiency of 20%. The obtained results show that CZTSe solar cells prepared by a low‐temperature process with in‐situ alkali doping can be utilized for flexible thin‐film solar cells as well as tandem device applications.

To demonstrate flexible and tandem device applications, a low-temperature Cu 2 ZnSnSe 4 (CZTSe) deposition process, combined with efficient alkali doping, was developed.First, high-quality CZTSe films were grown at 480 °C by a single co-evaporation, which is applicable to polyimide (PI) substrate.Because of the alkali-free substrate, Na and K alkali doping were systematically studied and optimized to precisely control the alkali distribution in CZTSe.The bulk defect density was significantly reduced by suppression of deep acceptor states after the (NaF + KF) PDTs.Through the low-temperature deposition with (NaF + KF) PDTs, the CZTSe device on glass yields the best efficiency of 8.1% with an improved V OC deficit of 646 mV.The developed deposition technologies have been applied to PI.For the first time, we report the highest efficiency of 6.92% for flexible CZTSe solar cells on PI.Additionally, CZTSe devices were utilized as bottom cells to fabricate four-terminal CZTSe/perovskite tandem cells because of a low bandgap of CZTSe (~1.0 eV) so that the tandem cell yielded an efficiency of 20%.The obtained results show that CZTSe solar cells prepared by a low-temperature process with in-situ alkali doping can be utilized for flexible thin-film solar cells as well as tandem device applications.
[18][19] Among these researchers, Becerril-Romero et al. obtained the highest efficiency of 4.9% for a CZTSe device on 0.09 cm 2 of PI prepared by a sequential process; specifically, a metallic precursor stack was deposited by direct current (DC)-magnetron sputtering, followed by selenization. [19]A suitable process for PI-based CZTSSe solar cells such as low-temperature processes and/or alkali PDT needs to be established for high-performance devices.
This study focuses on the development of a PI-friendly and highly efficient CZTSe deposition process.Given that the PI decomposition temperature is <500 °C, CZTSe films should be deposited at a lower temperature, enabling these films to remain a microstructure with large grains and without deep holes.In an earlier study, we found that CZTSe films grown at 480 °C have densely packed and large-grained morphologies. [12]Herein, optimal alkali incorporation with an in-situ process was explored for application to flexible and alkali-free substrates.Basically, the Na alkali distribution can be controlled through bifacial treatments (i.e., before and after the CZTSe film growth).In addition to Na doping, heavier alkali (potassium) K doping has been investigated to enhance the defect properties of CZTSe by KF PDT.The correlation between alkali doping and defect properties was systemically studied using capacitance-voltage (C-V) and drive-level capacitance profiling (DLCP) measurements.When NaF and KF dual PDTs (denoted as (NaF + KF) PDTs) were applied, the defect density in the CZTSe bulk decreased to approximately 5 9 10 15 cm À3 and the best efficiency of the CZTSe device on glass showed 8.1%.To obtain flexible solar cells, CZTSe solar cells on PI-coated glass were fabricated and lifted off from the glass substrate.For the first time, we report the highest efficiency of 6.92% (designated area: 0.25495 cm 2 ) for flexible CZTSe solar cells on PI.Moreover, the large area CZTSe solar cell on 64 cm 2 of PI exhibited an efficiency of 3.3%.
In addition to flexible solar cells, a tandem configuration using a perovskite top cell is a highly effective design to maximize device performance by utilizing the full spectrum of light.A CZTSe film with a low bandgap (1.0 eV) is suitable for a bottom cell to sufficiently absorb long-wavelength regions.[22] Nevertheless, very few tandem cells combined with kesterite-based bottom cells have been reported.Todorov et al. [23] obtained an efficiency of 4.6% for a monolithic CZTSSe/perovskite tandem solar cell.Most recently, an efficiency of 22.27% was achieved for a four-terminal (4-T) CZTSSe/perovskite tandem solar cell. [24]Therefore, it indicates that the 4-T tandem structure is a highly effective strategy to maximize the performance of CZTSe solar cells.In this respect, we fabricated 4-T CZTSe/perovskite tandem solar cells.The wide-band gap perovskite top cell exhibited an efficiency of 17.1%.When the top and bottom cell were mechanically stacked for 4-T CZTSe/perovskite tandem solar cell, the tandem cell yielded an efficiency of 20%.The obtained results suggest that alkali doping and low-temperature processes enable the fabrication of high-quality CZTSe films with a low bandgap that provide a promising route for achieving flexible thin-film solar cells, as well as tandem device applications.Furthermore, highefficiency flexible CZTSe/perovskite tandem solar cells can be readily demonstrated using the flexible CZTSe bottom cells on PI.

Results and Discussion
Alkali doping has been widely adopted in chalcogenide solar cells to improve device performance.In the case of CIGSe solar cells, the world record cell was fabricated with not only Na doping, but also heavier alkali element doping.[27] As a result, many studies on CZTSSe solar cells have applied alkali doping to improve device performance.However, only Na doping of CZTSSe films by NaF evaporation is performed usually, instead of heavier alkali doping combined with Na doping.In this study, the effect of heavier alkali K doping on the defect behavior and device performance was investigated.
As shown in Figure 1, large-grained CZTSe surfaces were grown without pinhole defects.The CZTSe grains extended throughout the films, which is desirable in terms of the transport of photo-generated carriers.After alkali doping, white needle-shaped NaF and KF were coated on the surface of the CZTSe films (Figure 1b,c).When the NaF and KF dual post-deposition treatments (PDTs) were performed consecutively, the shape of the shown particles changed to a smaller circle type with a higher density (Figure 1d), compared with that of a single alkali PDT.In general, the particles formed by PDT are eliminated after deposition of the CdS buffer layer because they are soluble in waterbased solution. [26,28,29]According to the energy dispersive X-ray spectroscopy (EDS) analysis, all the CZTSe films had approximately the following range of chemical compositions: In a previous study, we found that Na bifacial treatment is a suitable approach to precisely control the Na content in the bulk and at the surface of CZTSe films. [26]To assess the impact of K doping on photovoltaic (PV) device performance, CZTSe solar cells were fabricated using the alkali doping methods, as shown in Figure 2 (see Figure 9 in the experimental section).All PV parameters obtained from the illuminated current density-voltage (J-V) data are summarized in Table S1, Supporting Information.As either Na or K alkali element was introduced after the CZTSe film deposition, the device efficiency increased by more than 50% (Figure 2).Basically, there was no noticeable change in the bandgap of the CZTSe films after alkali doping.According to the EQE spectra shown in Figure S1, Supporting Information, the spectral responses in the visible range significantly increased, which means that the bulk properties of the CZTSe films are enhanced by alkali doping.Specifically, when the KF PDT was applied instead of the NaF PDT, the current density was slightly lower than that of the PV device using only the NaF PDT.However, the NaF and KF dual PDTs (denoted as (NaF + KF) PDTs) improved the current density.In addition, the highest V OC was achieved by the (NaF + KF) PDTs, with the lowest deficit of 685 mV.Similar to CIGS solar cells, the NaF PDT combined with the heavier alkali KF PDT seems to be the optimal pathway to improve device performance. [30,31]o understand the changes in the chemical properties of the CZTSe film due to the (NaF + KF) PDTs, the SIMS depth profiles of the CZTSe films were investigated (Figure S2, Supporting Information).Figure 3a indicates that Na doping in the CZTSe films induces an increase in the Na concentration compared with that of the control film.However, after the KF PDT, the Na concentration decreased slightly, whereas the K concentration increased significantly, as shown in Figure 3b.Previous studies indicate that K atoms replace Na atoms in CZTSe films subjected to KF PDT; in addition, K is typically observed at the CZTSe grain boundaries. [32,33]It could affect the defect behavior and resulting device performance.
To clarify the correlation between device performance and defect properties because of alkali doping, C-V and DLCP measurements were performed at 300 and 140 K, as shown in Figure S3, Supporting Information.The estimated free carrier (N A ), bulk defect (N BULK ), and interface defect (N IF ) values were extracted from the C-V and DLCP data according to the method reported in the literature [34] and are summarized in Table 1.In terms of defect passivation in the CZTSe films, the (NaF + KF) PDTs seem to be more effective than the only NaF PDTterminated method because there were decreases in the N BULK and N IF values, meaning that the defect properties are enhanced by the (NaF + KF) PDTs.Overall, the N A value of the CZTSe film increased slightly to an optimal value of approximately 3 9 10 16 cm À3 .The results are consistent with temperature-dependent open-circuit voltage (V OC -T) measurements.For the (NaF + KF) PDTs, the difference of the energy gap and defect activation energy, D(E g -E A ), shows the lowest value, 0.05 eV among them.These results suggest that the (NaF + KF) PDTs are more desirable for fabricating high-efficiency CZTSe solar cells.
Admittance spectroscopy (AS) was performed to determine the defect activation energy (E A ), as shown in Figure 4. Owing to the slow response of deep traps in the depletion region, the measured capacitances are dependent on both the oscillation frequency and measurement temperature.The E A value decreased from 197 (control) to 137 (NaF PDT) and 69 meV ((NaF + KF) PDTs) above the valence band edge.When E A is approximately 200 meV, the potential point defects in the band structure are V Zn and Zn Sn defects. [24,34,35]Considering the Zn-rich film growth and higher formation energy of the V Zn point defect, the contribution of V Zn in the control device can be ruled out, but deep acceptor Zn Sn antisite defects in the bandgap are more likely to be formed, which is approximately 200 meV above the valence band edge. [36]For the defect transitional level in the range of 100-200 meV, Cu Zn isovalent defects can be formed as shallow deep acceptors. [37]herefore, the E A value of 197 meV in the control device arises from the Cn Zn and Zn Sn antisite defects.After alkali doping, the E A value shifted to a lower shallow level near the valence band maximum   (VBM).A significant reduction in E A (69 meV) was observed after the (NaF + KF) PDTs, indicating that the defect properties were improved by suppressing the formation of detrimental antisite defects.All Naand K-related defects were harmless, except for the Na Sn and K Sn antisites. [38]In the case of the control device, the dominant peak was observed at 197 meV with a density of states (DOS) of 9.6 9 10 16 cm À3 ÁeV À3 .In contrast, the device subjected to the (NaF + KF) PDTs had the lowest defect energy level at 69 meV with a DOS of 7.2 9 10 17 cm À3 eV À3 , as shown in Figure 4b.From the perspective of defect chemistry, the decrease in E A after Na doping can be related to the relatively low formation energy of Na Zn . [39]The Na Zn defect has a transitional energy level below 100 meV, which is slightly smaller than that of the Cu Zn antisite. [40]The substitution energy of Krelated defects (e.g., K Zn , K Cu ) is much higher than 2 eV, and K + ion has a larger ionic radius (1.37 A) than Cu + (0.6 A) and Na + (0.99 A), thus the defect density should be negligible.According to atom probe tomography (APT) analysis, after the KF PDT, K atoms are usually segregated at the grain boundaries. [41]If K-containing compounds are formed in the grain boundaries, these compounds can passivate the grain boundaries and improve the defect properties, as shown in Table 1.Thus, the (NaF + KF) PDTs appear to be effective at suppressing detrimental deep defects and enhancing the acceptor doping level to achieve better device performance.
By device process optimization and MgF 2 antireflection coating, we achieved the highest efficiency of 8.1% along with improvement in all the PV parameters (V OC = 0.354 V, J SC = 39.0 mAcm À2 , and FF = 58.3),as shown in Figure 5a.The spectral responses in the visible region (500-800 nm) exceeded 90% as shown in Figure 5b.The EQE at the wavelengths above 800 nm increased compared with that in Figure S1, Supporting Information, although there was still a drop in the spectral responses owing to bulk recombination.If there is an additional improvement in CZTSe film quality, the efficiency of CZTSe PV devices prepared by a coevaporation method can exceed more than 10% in the near future.
To realize flexible CZTSe PV devices, PI was selected as a flexible substrate because CIGSe solar cells on PI have achieved an efficiency >20%. [14,42]Several research groups have attempted to fabricate kesterite-based CZTSSe solar cells on PI.However, most of these studies reported very low device efficiencies.Recently, Becerril-Romero et al. achieved an efficiency of 4.9% for a CZTSe device on PI (0.09 cm 2 ) using a two-step method with NaF + Ge doping. [19]In this study, CZTSe films were grown on PI filmcoated glass using optimized deposition conditions, specifically a low-temperature process and the aforementioned alkali doping.As shown in Figure S5, Supporting Information, the CZTSe films on PI also had large grains without pinhole defects, which were almost identical to the films on glass (Figure 1).This result confirms that the developed process is suitable for the deposition of high-quality CZTSe films on PI.After CZTSe device fabrication, PI was lifted off.Through process optimization, for the first time, we report the highest certified efficiency of 6.92% (designated area: 0.25495 cm 2 ) for flexible CZTSe solar cells on PI.Moreover, the large area CZTSe solar cell on 64 cm 2 of PI yielded an efficiency of 3.3% as seen in Figure 6.The PV parameters are presented in Table 2.
For the flexibility test of the CZTSe devices on PI, the bending properties were measured as shown in Figure 7.When the bending cycle of 100 with a 4 cm radius was applied, there was negligible changes in the device efficiency.As the bending radius decreased from 4 to 1 cm, the device performance slightly decreased.Particularly, after 1000 bending cycles, the bending radius of 1 cm induced 15% degradation relative to the initial device performance.Such a reduced efficiency seems to be related to some damage of the Al metal grid by either bending or multiple measurements of device performance.Nevertheless, the obtained results indicate that the flexible CZTSe devices are mechanically stable and can be utilized in various applications such as building-and vehicle-integrated photovoltaics.
To maximize the performance of CZTSe PV devices, a recent effective strategy is to fabricate a tandem structure consisting of a wide-bandgap perovskite-based cell on top of the CZTSe bottom cell because this approach effectively utilizes the full spectrum of incident light.As  shown in Figure 8a, the bottom and top solar cells were composed of CZTSe and semitransparent p-i-n structured perovskite, respectively.The 4-T configuration permits the separate optimization of the CZTSe and perovskite cells.The perovskite top cell and CZTSe bottom cell were optically coupled and electrically decoupled.Therefore, the top and bottom cells were individually processed and then mechanically stacked on top of each other.In this configuration, the perovskite top and CZTSe bottom cells can operate at their respective maximum power points, and the sum of each constituent cell corresponds to the efficiency of the 4-T tandem device.
The J-V curve and EQE spectra of the tandem solar cell are shown in Figure 8b,c.The 8.1%-efficient CZTSe stand-alone cell had an efficiency of 2.9% when operating as the bottom cell in the tandem configuration with the 17.1% efficient transparent perovskite solar cell.The J SC and V OC values of the stand-alone CZTSe solar cell decreased from 0.35 to 0.32 V and 39.0 to 16.3 mAcm À2 , respectively, when incorporated into the tandem configuration.The reduction in the PV parameters was consistent with the EQE measurements.The EQE spectra under the perovskite filter (see the transmittance of the perovskite filter in Figure S7, Supporting Information) clearly indicate no spectral response in the visible region of 300-750 nm, as shown in Figure 8c.The 4-T CZTSe/perovskite tandem solar cell yielded 20%.Table 3 summarizes the PV parameters of each constituent solar cell in the 4-T CZTSe/ perovskite tandem solar cell.
With regard to future opportunities, high-efficiency flexible CZTSe/ perovskite tandem solar cells can be realized by mechanically stacking CZTSe bottom and perovskite top cells prepared on a flexible PI substrate.Moreover, for a monolithic CZTSe/perovskite tandem solar cell, a perovskite top cell can be directly fabricated on the top of the bottom CZTSe cell by introducing a recombination layer and improving surface morphologies such as smooth and pinhole-free CZTSe surface.

Conclusions
To realize flexible CZTSe solar cells, alkali doping and low-temperature processes were systemically studied to obtain high-quality CZTSe films on a flexible PI substrate.The low-temperature-processed CZTSe films exhibited a large-grained surface morphology without pinhole defects.In-situ alkali doping during the CZTSe film growth was applied via bifacial treatments using NaF and KF sources.Through the lowtemperature deposition with (NaF + KF) PDTs, the device efficiency increased by more than 50% up to the best efficiency of 8.1%.According to SIMS depth profiling, Na doping in the CZTSe films led to an increase in the Na concentration; however, after the KF PDT, the Na concentration decreased slightly, whereas the K content in the film increased significantly.The capacitance-based analysis exhibited that the reduction in defect density in the CZTSe bulk subjected to the (NaF + KF) PDTs by one order of magnitude and the defect activation energy decreased from 197 (control) to 69 meV ((NaF + KF) PDTs) above the valence band edge.It indicates that detrimental defects are effectively passivated by the suppression of deep antisite defects in bulk and grain boundaries by alkali doping.For the first time, we report that the flexible CZTSe PV device on the PI exhibited the highest certified efficiency of 6.92%.After 1000 bending cycles with a bending radius of 1 cm, 15% degradation relative to the initial efficiency was shown due to damage of the Al metal grid.Furthermore, when the CZTSe solar cell was used as a bottom cell for 4-T CZTSe/perovskite tandem solar cells, the best 4-T tandem solar cell yielded 20%.The obtained results suggest that CZTSe solar cells prepared by a low-temperature process with in-situ alkali doping can be utilized for flexible thin-film solar cells as well as tandem device applications.

Experimental Section
Fabrication of flexible Cu 2 ZnSnSe 4 solar cells: CZTSe films were grown on Mocoated glass substrates using a thermal co-evaporation method at 480 °C.The temperature profiles of the CZTSe film growth are shown in Figure 9a.The thickness of the CZTSe film was approximately 1 lm.For in-situ alkali doping, NaF and KF were evaporated at 700 and 550 °C, respectively.The flux of NaF and KF was fixed at 2.0 A/s.Specifically, Na doping was applied before and after the CZTSe film deposition to control the Na distribution in the film (Figure 9c).In   this study, the effects of K doping on device performance were explored by KF PDT, as depicted in Figure 9b,d.
For solar cell fabrication, CZTSe films on Mo-coated glass substrates were incorporated into the following well-known device structure for kesterite-based solar cells: indium tin oxide (ITO)/intrinsic ZnO (i-ZnO)/CdS/CZTSe/Mo/ glass. [43]A 60 nm-thick CdS buffer layer was deposited via chemical bath deposition.50 nm-thick i-ZnO and 120 nm-thick ITO layers were deposited using radio-frequency sputtering.An Al top electrode was deposited on top of the device using a thermal evaporator.The device area of 0.44 cm 2 was defined by mechanical scribing.For flexible solar cells, CZTSe films were deposited on PIcoated glass and lifted off from the glass. [44]abrication of Four-Terminal (4-T) CZTSe/Perovskite Tandem Solar Cells: Materials for perovskite top cells-MeO-2PACz was purchased from Tokyo Chemical Industry (TCI) Co., LTD.All other materials for the perovskite precursor and for evaporation were used without further purification unless otherwise stated.Lead iodide and lead bromide were purchased from TCI (99.99%, trace metal basis) or Sigma Aldrich (99.999%, trace metal basis).Formamidinium iodide (FAI) and methylammonium bromide (MABr) were purchased from GreatCell Solar Materials (grade > 99.88%), cesium iodide (CsI) was purchased from Alfa Aesar (99.999%, trace metal basis).All anhydrous solvents such as DMF, DMSO, chlorobenzene, and ethanol were obtained from Sigma Aldrich.C60 (99.99%) and BCP (>99.0%) were purchased from Nano-C or TCI.Cu (99.99%) was purchased from Itasco.
Semitransparent perovskite top cells for 4-T CZTSe/perovskite tandem cells-The washed substrate was treated under UV-ozone for 15 min and immediately moved into the glovebox.For hole transport layer deposition, 1 mM solution of MeO-2PACz in ethanol was spin-coated at 178 g for 30 s and annealed at 100 °C for 10 min.The perovskite precursor solution was prepared according to the chemical formula Cs 0.05 (FA 0.83 MA 0.17 )Pb(I 0.83 Br 0.17     semitransparent perovskite top cells: ITO/SnO 2 /C60/Perovskite/MeO-2PACz/ FTO.A 25 nm-thick C60 electron transport layer was deposited by using a thermal evaporator.A 10 nm-thick SnO 2 interlayer was introduced via atomic layer deposition to avoid sputter damage induced by ITO deposition.Then, a 130nm thick ITO layer was deposited using radio-frequency sputtering.Finally, a 100 nm-thick Ag grid was deposited on top of the device using a thermal evaporator.In order to make a 4-T tandem structure consisting of a semitransparent p-i-n structured perovskite solar cell on the top of the CZTSe bottom cell, the top and bottom cells were mechanically stacked together as shown in Figure 8a.Device characterization-Morphological and elemental analyses of the CZTSe films were carried out using field emission scanning electron microscopy (FE-SEM, Hitachi S-4700).The compositional depth profiles of the CZTSe films were investigated using secondary ion mass spectroscopy (SIMS) with a Cs + ion gun.
Temperature-dependent open-circuit voltage (V OC ) measurements of the CZTSe solar cells were carried out using a Keithley 2450 source meter and Linkam LTS 350 probe station under air mass (AM) 1.5G illumination.In addition, capacitance-based analysis was performed for the representative cells.Currentvoltage (C-V) and drive-level capacitance profiling (DLCP) analyses were performed at 300 and 140 K using an LCR meter (Agilent 4284A).Admittance spectroscopy (AS) was performed in the temperature range of 300-140 K, and the frequency varied from 10 2 to 10 6 Hz.The current density-voltage (J-V) characteristics of the CZTSe solar cells were measured using a solar simulator (MacScience, K201-LAB 50).The external quantum efficiency (EQE) spectra of the CZTSe solar cells were obtained using a MacScience K3100 instrument.The light intensity of the J-V and EQE measurements was calibrated using a Si reference cell.

Figure 1 .
Figure 1.Scanning electron microscopy (SEM) top-view of the Cu 2 ZnSnSe 4 (CZTSe) films a) without and b-d) with various Na and K alkali doping treatments: b) NaF pre-absorber + KF post-deposition, c) NaF bifacial deposition, and d) NaF pre-absorber + NaF and KF dual post-deposition.

Figure 2 .
Figure 2. Summary of the photovoltaic parameters a) efficiency, b) fill factor, c) short-circuit current, and d) open-circuit voltage) measured from the illuminated current density-voltage (J-V) curves of Cu 2 ZnSnSe 4 (CZTSe) solar cells (Five cells fabricated in the same batch) with various alkali postdeposition treatments (PDTs).Basically, the NaF on top of the Mo film, except for the control sample was supplied.

Figure 7 .
Figure 7. Bending test of flexible CZTSe solar cells on PI as the number of bending cycles with various bending radii from 1 to 4 cm.
) 3 .First, PbI 2 and PbBr 2 were dissolved in a mixture of anhydrous dimethylformamide (DMF) and dimethyl sulfoxide (DMSO) (4:1 volume ratio) to nominal concentration of 1.5 M.Then, PbI 2 and PbBr 2 stock solutions were added to FAI and MABr, respectively, to obtain FAPbI 3 and MAPbBr 3 solutions with final concentration of 1.24 M. The molar ratio between lead and the respective cations was 1.09:1.00(9% lead excess) for both solutions.Next, the FAPbI 3 and MAPbBr 3 solutions were mixed in 83:17 volume ratio to obtain the desired composition of the mixed cation FA 0.83 MA 0.17 Pb(I 0.83 Br 0.17

) 3 .
Finally, 5 vol% of 1.5 M CsI stock solution in DMSO was added in the mixed cation solution to prepare the triple cation Cs 0.05 (FA 0.83 MA 0.17 )Pb(I 0.83 Br 0.17 ) 3 precursor solution.For perovskite film, 80 lL of the prepared perovskite precursor solution was spin-coated onto the fluorine-doped tin oxide substrate at 194 g for 35 s.The film was treated with 200 lL of chlorobenzene for 5 s and annealed at 100 °C for 30 min.For top cell fabrication, perovskite films were incorporated into the following device structure of

Figure 8 .
Figure 8. a) Schematic diagram of the 2 ZnSnSe 4 (CZTSe)/perovskite tandem cell in the four-terminal (4-T) configuration; b) light-J-V curves under standard AM 1.5G illumination of the CZTSe stand-alone cell (black color), CZTSe bottom cell (blue color), and perovskite top cell (red color); and c) external quantum efficiency (EQE) spectra for the corresponding devices.

Figure 9 .
Figure 9. Temperature profiles of the Cu 2 ZnSnSe 4 (CZTSe) film growth a) without and b-d) with various Na and K alkali doping treatments: b) NaF pre-absorber and KF post-deposition, c) NaF bifacial deposition, and d) NaF pre-absorber and NaF + KF dual post-deposition.

Table 2 .
Summary of the photovoltaic parameters measured from the illuminated current density-voltage (J-V) curves of Cu 2 ZnSnSe 4 (CZTSe) solar cells on polyimide (PI).