Relieving the Ion Migration and Increasing Superoxide Resistance with Glutathione Incorporation for Efficient and Stable Perovskite Solar Cells

Perovskite solar cells (PSCs) have the great potential for the next‐generation energy application, while the serious issue about longevity has still been a long‐standing concern. The rapid crystallization process would inevitably introduce defects, and the perovskite film decomposed rapidly when exposed to water/oxygen. Herein, the reduced l‐glutathione (GSH) is introduced into the perovskite precursor to tailor the crystallization process and passivate defects, thus relief the detrimental ion migration. Furthermore, GSH can effectively inhibit the formation of superoxide. As a result, the optimized PSCs demonstrate the champion power conversion efficiency of 22.89% and maintained 91% of its initial efficiency under ambient environment for 1000 h.

perovskite to tailor the crystallization process and inhibit the generation of O 2 − . The multi-functional groups in GSH such as sulfhydryl (-SH), carbonyl (-COOH) could bind with Pb 2+ to slow down the crystallization process by enhancing the energy barrier. While the amino (-NH 2 ) would interact with I − to form hydrogen bond to inhibit ion migration. Furthermore, GSH could react with the O 2 − to restrict the decomposition of the organic compounds and the oxidation of I − ion. Attributed to the synergistic effects of GSH, the power conversion efficiency (PCE) increased from 21.53% to 22.89% accompanied with the enhanced stability. The unencapsulated PSCs devices could remain 91% of its initial efficiency after storage in ambient environment for 1000 h.

Results and Discussion
The perovskite film was fabricated through the two-step deposition process, and GSH shown in Figure S1a, Supporting Information, was dissolved in the PbI 2 precursor solution with various concentration. We assumed that the Lewis base functional groups would chela with Pb 2+ , and the -NH 2 would bind with I − through the hydrogen bond. [12] First, we conducted Fourier transform infrared spectroscopy (FTIR) analysis to confirm this assumption. The full FTIR spectrum was plotted in Figure S1b, Supporting Information. The detailed FTIR spectrum was shown in Figure 1a,b. The wavenumber at around at 3500 cm −1 could be attributed to the NH vibrations, and the characteristic peak at 1643 cm −1 could be assigned to CO, respectively. [13] After the addition of PbI 2 , the NH characteristic peak moved to higher wavenumber (Figure 1a), while the CO characteristic peak was down shifted to 1626 cm −1 (Figure 1b). The interaction between GSH and PbI 2 would influence the electronic cloud, thus the corresponding characteristic peaks were shifted. While we did not observe the -SH characteristic peak at ≈2600 cm −1 , since it was easily oxidized under ambient environment. We then conducted X-ray photoelectron spectrum (XPS) to further investigate the interaction between PbI 2 and GSH. The full XPS spectrum was shown in Figure S2, Supporting Information, and the S 2p peak at binding energy (BE) of 164 eV could be identified for the GSH-PbI 2 film in Figure S3, Supporting Information, which confirmed that the GSH was successfully incorporated into the PbI 2 film. [14] We focused on the Pb 4f and I 3d spectrum to explore the possible interaction mechanism. As shown in Figure 1c for the pristine PbI 2 film, the BE of Pb 4f was located at 143.3 and 138.5 eV, which could be assigned to 4f 5/2 and 4f 7/2 of divalent Pb 2+ , respectively. After the addition of GSH, the BE of Pb 4f core level shifted to lower energy. At the same time, the two obvious shoulder peaks could be identified, which may be attributed to the formation of the GSH-PbI 2 complex. [15] Furthermore, compared with the pristine PbI 2 film, the I 3d were all shifted to higher BE due to the formation of NH…I hydrogen bond in the GSH-PbI 2 film as shown in Figure 1d. [12] The X-ray diffraction

www.advmatinterfaces.de
(XRD) measurement of PbI 2 film with or without GSH was also conducted and plotted in Figure 1e. The diffraction peaks intensity of GSH-PbI 2 film was strongly decreased under the same annealing time, the slowly crystallization process also confirmed the strong interaction between GSH and PbI 2 .
We then conducted in situ X-ray diffraction (XRD) measurement to investigate the crystallization process after the addition of GSH as shown in Figure 2a,b. Before the annealing process (0 s), the diffraction peaks at 6.6°, 7.2°, and 8.7° could be identified for the both two samples, which could be assigned to the MAI/FAI-PbI 2 -DMSO intermediate phase, and the peak at 11.8° could be attributed to the formation of δ-phase FAPbI 3 (δ-FAPbI 3 ). While the diffraction peaks of intermediate phase and δ-FAPbI 3 were effectively inhibited after the addition of GSH (Figure 2b at 0 s), which could be ascribed to the strong interaction between GSH and PbI 2 as we have discussed above. [16] With the extension of annealing process to 5 s, the intermediate phase vanished, accompanied the formation of perovskite phase (α-phase FAPbI 3 , α-FAPbI 3 ). It should be noted the detrimental δ-FAPbI 3 fully transformed to α-FAPbI 3 for the GSH-incorporated perovskite film when annealing process to 10 s. As for the pristine sample, the δ-phase FAPbI 3 fully transformed to perovskite phase and the PbI 2 peak emerged at 12.8° when the annealing time was prolonged to 30 s. With the annealing process going on, the peak intensity of the (100) plane increased continuously for the GSHincorporated perovskite film, and the PbI 2 peak appeared after 3 min annealing time, which confirmed that the incorporation of GSH could effectively retard the crystallization process. The complete perovskite film could be obtained after 10 min annealing time and the XRD patterns of the two sample were shown in Figure S4, Supporting Information. It can be seen that, after introducing GSH, the peak intensity of (100) plane was much stronger than that of the pristine film, which indicated a better crystallographic orientation. [17] Scanning electron microscopy (SEM) was conducted to investigate the variation of the perovskite film morphology after the addition of GSH. As shown in Figure 2c,d, the perovskite film with or without the addition of GSH all demonstrated the dense and smooth morphology. While the GSH incorporated film exhibited a larger grain size (800 nm) compared with the pristine film (300 nm). The cross-sectional SEM images shown in Figure S5, Supporting Information, also confirmed the superiority with the addition of GSH. The pristine sample demonstrated pinhole and creak among the film, while the large grain across the complete perovskite film could be observed in the GSH incorporated perovskite film, which would be beneficial for decreasing the defects and promoting carrier transport. [18] The surface roughness was also detected by the atomic force microscopy (AFM) and shown in Figure 2e,f and Figure S6, Supporting Information, the surface root-mean-square (RMS) decreased from 29.5 to 20.3 nm by introducing GSH as addictive, which agreed well with the SEM images. The slightly enhancement of absorption intensity shown in Figure S7a, Supporting Information, also confirmed the improvement of perovskite film quality without changing the bandgap ( Figure S7b, Supporting Information). [16] The space charge limited current (SCLC) measurement was conducted to evaluate then defects density of the perovskite films with or without GSH incorporation. We fabricated electron-only (E-only) and hole-only (H-only) devices then characterized the J-V curves in the dark. The trap-state density (N trap ) could be calculated by the equation: (1)

www.advmatinterfaces.de
where ε 0 was the vacuum permittivity, ε was the relative dielectric constant of perovskite, V TFL was the trap-filled limit voltage, e is the electron charge and L was the thickness of perovskite. [19] As shown in Figure 3a,b, the V TFL for the pristine and GSH incorporated E-only devices was located at 0.371 and 0.207 V, and the calculated N trap was 6.68 and 3.73 × 10 15 cm −3 , respectively. As for the H-only devices, the V TFL was estimated at 0.324 V for the pristine devices and 0.237 V for the GSH incorporated devices. The value of N trap was calculated to be 5.82 and 4.26 × 10 15 cm −3 , respectively. The decreased N trap could be attributed to the passivation effect and the improved crystal quality after the addition of GSH. Photoluminescence (PL) mapping and time-resolved PL (TRPL) measurement was conducted with the device structure of glass/perovskite to investigate the charge recombination behavior. As shown in Figure 3c,d, the PL intensity of the perovskite film was significantly enhanced with the addition of GSH. [20] The TRPL spectrum was plotted in Figure 3e, the results were fitted via a bi-exponential decay model and the results were summarized in Table S1, Supporting Information. As expected, the average charge lifetime (τ ave ) was greatly prolonged from 359.4 ns for the pristine perovskite film to 524.5 ns for GSH-incorporated perovskite film. The enhanced PL intensity and charge lifetime of the perovskite film with the addition of GSH was attributed to the suppression of nonradiative recombination caused by trap states. We also did PL measurements on the samples with SnO 2 layer. As shown in Figure S8, Supporting Information, the PL intensity of the GSH-incorporated sample was quenched more significantly, which confirmed that the charge transport between the perovskite and SnO 2 layer were also improved. [21] Electrochemical impedance spectroscopy (EIS) measurement was conducted to further investigate the electrical properties of the PSCs devices. The related Nyquist plot was shown in Figure 3f. In general, the semicircle at low frequency corresponded to the recombination resistance (R rec ). Compared with the pristine devices, the GSH-incorporated devices demonstrated higher R rec , indicating a decreased recombination rate, which was in accordance with the TRPL results. [22] Mott-Schottky analysis was carried out to evaluate the variations of the build-in potential (V bi ) in the PSCs devices. As shown in Figure 3g, The V bi values were improved from 0.79 V for the pristine devices to 0.91 V for the GSH-incorporated devices. The increased V bi could not only facilitate the charge separation but also contribute to the enhancement of open voltage (V oc ). [23] The dependency of V oc on light intensity was examined to assess the ideality factor. As shown in Figure S9, Supporting Information, the ideality factor decreased from 1.47 to 1.22 with the incorporation of GSH, which implied that the Shockley-Read-Hall recombination was effectively suppressed. [24] The trap density of state (tDOS) was further investigated by thermal admittance spectroscopy as shown in Figure 3h, the reduction in the trap density confirmed that GSH incorporation could effectively improve the quality of perovskite film. [25] To explore the mechanism for the reduction in tDOS shown in Figure 3h, we then conducted density functional theory (DFT) calculations to evaluate the variation of defects formation energy before and after GSH-incorporation, the unit cell of FA 0.75 MA 0.25 PbI 3 was simplified to FAPbI 3 to predigest the calculation. [26] We calculated the formation energy of various defects on the surface of perovskite films and the results were shown in Figure 4a,b and Figure S10, Supporting Information. It could be identified that the iodine vacancy (V I ) formation energy increased from 0.55 to 0.69 eV after GSH-incorporation, which would be beneficial for suppressing defect formation. The formation energy of lead vacancy (V Pb ) and Pb-I anti-site (I Pb ) defects were also enhanced from 3.19 and 3.21 eV to 3.27 and 4.53 eV, respectively. The detailed parameters could be found in Table S2, Supporting Information. [27] The ion migration behavior was investigated by the temperature-dependent conductivity measurement under dark conditions. [28] The activation energy (E a ) could be extracted by the Nernst-Einstein relation:

www.advmatinterfaces.de
Where σ(T) is the conductivity as a function of temperature T, k b is the Boltzmann constant, and σ 0 is a constant. The E a could be calculated by a linear fitting from Figure 4c at the high temperature region. The E a for the perovskite film with GSH incorporation was 0.482 eV, which was substantially higher than that of the pristine sample (0.258 eV), indicating a strong interaction between GSH and defects which would inhibit the ion migration. It has been previously reported that iodide ion was of similar size to the O 2 − species, thus the V I sites were the preferred location for the formation O 2 − by direct electron transfer from the perovskite to oxygen. [29] As demonstrated in Figure 4a-c, the defects formation energy and E a was significantly enhanced with the incorporation of GSH which would inhibit the formation of O 2 − species. [4,30] Hydroethidine (HE), which demonstrated a characteristic emission peak at 610 nm when exposed to the O 2 − species, was employed as a molecular fluorescent probe to evaluate the yield of O 2 − . [31] The emission data under various aging time was collected and the rate of increase in emission at 610 nm was shown in Figure 4d. It was evident that the perovskite samples with GSH-incorporation demonstrated a relatively lower yield of O 2 − formation. To confirm the generality of the GSH strategy, we also fabricated the FACsPbI 3 perovskite and minored the yield of O 2 − by the molecular fluorescent probe. The result shown in Figure S13, Supporting Information, further verified the universality of GSH strategy. This could be attributed to the following two factors: i) the improved crystal quality and decreased defect sites which could prevent the absorption of oxygen molecular; ii) the reaction between GSH and O 2 − species which lead to the decreased concentration of O 2 − in the perovskite film. According to the previous report, the oxygen and light induced decomposition process could be described by the following equation: The I 0 could not only serve as the non-radiative recombination center but also initiate the cascade decomposition process of the perovskite film. Since the decomposition product I 0 was volatile, we thus conducted XPS analysis of the perovskite film after aging for 200 h to examine the ratio of I/Pb which could demonstrate the iodine evolution indirectly. [32] The full XPS spectrum and S 2p characteristic peak were shown in Figures S11 and S12, Supporting Information, respectively. The Pb 4f and I 3d spectrum for the aged perovskite film was plotted in Figure 4e,f. The calculated I/Pb ratio was 2.92 of the perovskite film with GSH-incorporation, which was higher than that of the pristine sample (2.78). At the meantime, the distinct Pb 0 peak could be indicated for the pristine perovskite film which was absent in the sample with GSH-incorporation, indicating that the GSH could passivate the no-coordinated Pb 2+ thereby www.advmatinterfaces.de reducing the formation of metallic lead. [33] It has been proved that the main product between GSH and O 2 − was oxidized glutathione (GSSG) and the schematic diagram was shown in Figure S14, Supporting Information. It could be distinguished that the GSSG could still locate at grain boundary and maintain the ability to passivate the no-coordinated Pb 2+ defects. [29] To evaluate the effect of GSH incorporation on the photovoltaic performance, the complete PSCs device was fabricated based on the perovskite film with various GSH concentration. 15 devices were fabricated to verify the reproducibility and the detailed parameters was shown in Figure S15, Supporting Information. The champion device with the V oc of 1.157 V, short current (J sc ) of 24.98 mA cm −2 , fill factor (FF) of 79.2% and PCE of 22.89% was obtained at a concentration of 1 wt%, which was significantly higher than that of the pristine sample with a V oc of 1.124 V, J sc of 24.63 mA cm −2 , FF of 77.8% and PCE of 21.53% as shown in Figure 5a and Table S3, Supporting Information. Meanwhile, the negligible hysteresis (reverse scan: 22.89%, forward scan: 22.55%) was observed in the GSH incorporated PSCs devices, while the pristine devices exhibited stronger hysteresis (reverse scan: 21.53%, forward scan: 20.29%). Steadystate output (SOP) of the devices at the maximum power point (MPP) was conducted to verify the reliability of the device in the operating condition. As shown in Figure 5b, the stable J SC and PCE could be obtained for the GSH incorporated samples at MPP of 0.98 V for over 500 s, which was significantly higher than that of the pristine samples. The effect of the GSH incorporation on the stability of the perovskite films was investigated. It should be noted that the perovskite film was storage in ambient environment without encapsulation. As shown in Figure 5c, the PbI 2 peak intensity of the pristine perovskite evidently enhanced after aging, and the color blenched from minor black to yellow. While for the perovskite film with GSHincorporation, we could still observe the dominant perovskite diffraction peak, accompanied by a minor enhancement of the PbI 2 peak, which indicated the integrity of the perovskite www.advmatinterfaces.de crystal structure. The irradiance and thermal stability of the PSCs devices were investigated to prove the superiority of the GSH incorporation. As shown in Figure S16a, Supporting Information, the device incorporated with GSH could maintain 88% of its initial PCE after continuous illumination for 400 h, while the pristine device only retained 47% of the initial PCE. For the thermal stability measurement, the devices were kept at 85 °C in the glove box. The device incorporated with GSH could retain 83% of its primary PCE, while the PCE of the reference device dropped to 39%. It could be concluded that with the GSH-incorporation, the irradiance and thermal stability were enhanced simultaneously due to the defects passivation and suppressed ion migration. [34] Finally, the ambient stability of the unencapsulated devices with or without GSH incorporation was compared. As shown in Figure 5d, the PCE of the pristine PSCs devices rapidly dropped to 52% of its initial efficiency, while the PCSs with GSH-incorporation could maintain 91% of its original efficiency after storage for 1000 h under ambient environment.

Conclusion
In this work, we introduced the reduced GSH into the planar PSCs devices. The incorporation of GSH could not only tailor the crystallization process, passivate the defects, inhibit the ion migration, but also enhance the superoxide resistance. As a result, the champion PCE of 22.89% was achieved, which was higher than that of the pristine devices. Importantly, the device stability was enhanced simultaneously. The unencapsulated PSCs with GSH-incorporation could maintain 91% of its initial efficiency after storage in ambient environment for 1000 h. This facial and effective strategy would speed up the commercialization of high-performance PSCs.

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