Improving the Photovoltaic Properties and Stability of Inverted Perovskite Solar Cells with Hydroxylamine‐Based Additives

Organic–inorganic halide perovskite solar cells (PVSCs) are considered a promising emerging photovoltaic technology that offer exceptional optoelectronic properties and the potential for economic solar energy conversion. Additive engineering‐based fabrication processes can achieve highly efficient and stable PVSCs that feature well‐controlled perovskite layers with a dense, uniform, “black” α‐phase crystal structure, as well as large grains and few defects. In this study, several hydroxylamine derivatives are introduced as additives to FAPbI3 precursor solutions to investigate their effects on the performance of PVSCs. The addition of hydroxylamine derivatives suppresses the formation of the unwanted δ‐phase and lead iodide, while the α‐phase cubic structure is preferentially formed without changing the bandgap of FAPbI3. Additionally, the additive‐treated perovskite films show improved stability compared with those without additives. Moreover, using X‐ray diffraction and X‐ray photoelectron spectroscopy analyses, it is discovered that the hydroxylamine‐based additives are not incorporated in the crystal lattices but rather resided on the surface or grain boundaries. Notably, the inverted PVSCs added with N‐methylhydroxylamine exhibit an improved power conversion efficiency, higher stability, and minimal hysteresis.


Introduction
Perovskite solar cells (PVSCs) can be fabricated using a simple wet-coating process, [1] and high power conversion efficiencies (PCE) [2,3] exceeding those of conventional solar cells have been reported for both experimental single cells and tandem cells. [4,5]he perovskite layers play an important role as light absorbers in the visible-light region.However, it is difficult to increase their reliability because of issues such as ion desorption from the perovskite structure and crystal phase shifts caused by heat and light. [6]egarding the composition of the perovskite layer, various combinations have been investigated for the basic ABX 3 composition, such as the methylammonium cation (MA + ), formamidinium cation (FA + ), and potassium cation (K + ) for the A site; the lead (II) ion (Pb 2+ ) and tin (II) ion (Sn 2+ ) for the B site; and the iodide ion (I − ) and bromide ion (Br − ) for the X site.Among these, MAPbI 3 is the most widely reported perovskite material, with an absorption edge at approximately 780 nm corresponding to an energy gap (Eg) of 1.59 eV, which allows it to achieve high PCE values.However, because MA + has a low molecular weight and is easily released from MAPbI 3 , FAPbI 3 has been introduced, in which MA + is replaced by the higher molecular weight FA + .Because the Eg of FAPbI 3 is close to the desired value in terms of the Shockley-Queisser limit, it is a promising composition for increasing the PCE; however, perovskites with the FAPbI 3 composition are still hindered by poor crystal phase stability. [7]Although the black cubic crystal phase (-phase) with a small bandgap is ideal for light absorbers, the -phase with a larger bandgap, which tends to appear at room temperature as a phase transition, has negative effects on device performance. [8]nother problem is the easy formation of lead iodide (PbI 2 ), which stems from the desorption and migration of ions from the perovskite surface. [9]PbI 2 is a yellow compound with a larger bandgap and smaller light absorption coefficient than the phase black perovskites; therefore, its generation negatively impacts the PCE of PVSCs.
The introduction of MABr has been reported as an effective method for stabilizing FAPbI 3 . [10]However, the introduction of bromine anions increases the bandgap, causing it to deviate from the ideal values.Other methods, including the introduction of organic [11] and inorganic cations, [12] have been used to stabilize perovskite films by introducing mixed cation compositions.Various amine compounds have been reported to act as organic cations, [13] with typical examples including imidazolines, dimethylammonium, guanidine, [14] and phenethylamine (PEA). [15]These compounds can be used as additives or passivation materials to replace amines on the perovskite surface or fill defects.This enhances the charge transfer properties of perovskite-based devices while suppressing existing instability issues, leading to higher-quality films and superior device performance.Furthermore, the introduction of PEA positively affects the crystal structure by producing 2D perovskite layers on the surface that serve as passivation layers.Additionally, the use of amines with relatively large molecular weights has been proposed to improve the long-term durability of perovskite structures; however, such molecules may induce other drawbacks, such as an increase in the electrical insulation properties of the material and detrimental effects on its crystal lattices.
The aim of this study was to stabilize the perovskite layer by suppressing the desorption and migration of ions from the surface using relatively small organic cations without affecting the bandgap or crystal structure of the perovskite.Therefore, the introduction of hydroxylamine derivatives has been attempted.Hydroxylamine compounds are expected to interact more strongly with the surrounding molecules than amino groups alone and are less likely to desorb because of the expected hydrogenbonding nature of the hydroxyl groups.
A previous study showed that the incorporation of hydroxylamine hydrochloride into MAPbI 3 facilitated charge carrier mobility and improved PCE. [16] However, hydroxylamine (NH 2 OH, molecular weight 33 g mol −1 ) is hygroscopic and unstable, raising concerns regarding its corrosiveness to metals and possible hazardous effects on aquatic organisms. [17]Therefore, materials with low reactivity and environmental risk are required.
In this study, several new hydroxylamine derivatives with increased molecular size and reduced reactivity (due to substituent groups) were investigated.We evaluated the properties and stability of perovskite films prepared from precursor solutions with each of these hydroxylamine-based additives, fabricated the corresponding inverted PVSCs, [18] and assessed their photovoltaic characteristics and reliability to verify the effects of the additives.The addition of a small amount of hydroxylamine derivatives was revealed to suppress the formation of unwanted PbI 2 and preferentially generate the desired cubic -phase, while also suppressing the degradation of the perovskite films and cells without affecting the band gap or perovskite crystal structure.We also observed that some of the added hydroxylamine derivatives effectively increased the perovskite crystal grain size.As a result, high PCEs exceeding over 17% were achieved, highlighting the potential of our approach for improving the performance of PVSCs.
Various analyses, including UV-vis absorption spectroscopy, X-ray diffraction (XRD), scanning electron microscopy (SEM), Xray photoelectron spectroscopy (XPS), Fourier transform infrared (FT-IR) spectroscopy, and current-voltage (J-V) measurements, were conducted to determine the effects of the additives on the power conversion efficiency and stability of PVSCs under ambient conditions and continuous illumination.

Effects of Hydroxylamine
To investigate the impact of hydroxylamine-based additives on perovskite films and inverted PVSCs, we first examined the influence of two additives, hydroxylamine (HoA) and OH-free MoA, as depicted in Figure 1.
The UV-vis absorption spectra of the perovskite films, shown in Figure 2a,b, were analyzed to investigate the effects of the MoA and HoA additives.Interestingly, no change in the absorption edge was observed regardless of the type or amount of additive used (MoA or HoA), indicating that the bandgap of the perovskite was marginally affected by the presence of these additives or their corresponding amounts.In addition, the absorbance of the perovskite films decreased at higher additive concentrations, hindering their solar light absorbance.This may imply that the PCE of PVSCs will decrease as the additive amount increases, as more incident light is transmitted through the film than absorbed.
The effect of the hydroxylamine-based additives on the perovskite crystal structure was verified via XRD.The XRD results, shown in Figure 2c,d, revealed that the use of MoA (2%, 5%) and HoA (1%, 2%, 5%) as additives led to a reduction in the peak originating from the unfavorable PbI 2 .In contrast, the -phase peak became prominent when the amount of additive increased to more than 2%, suggesting that an excess amount of additives may interfere with the formation of the -phase crystal structure.These results suggest that mixing small amounts of HoA (1%) is effective in obtaining an ideal perovskite crystal structure.Furthermore, SEM analysis (Figure 2e) was performed to investigate the influence of the additives on the morphology of the perovskite films.On the surface of the film with 1% addition of HoA, the amount of PbI 2 was obviously less than that of the control film, and it was observed that a dense and uniform film was formed. [19]hese XRD results also correlate with the UV-vis absorption results described above.Light yellow -phase crystals have a larger band gap than the -phase and are known to have less optical absorption after 600 nm than the -phase crystal structures. [20]The -phase increases with increasing additive concentration, which is consistent with a decrease in optical absorption from 600 to 800 nm.Since PbI 2 is also a yellow wide-bandgap material, the improvement in optical absorption with 1% HoA addition is due to the suppression of PbI 2 formation and the preferential formation of the photoactive -phase in the visible light wavelength range.These results clearly demonstrate that the introduction of hydroxylamine additives in the PVSCs fabrication process has an advantageous effect on the development of the perovskite surface, possibly resulting in enhanced device performance. [21]igure 3a,b shows the J-V characteristics of the inverted PVSCs fabricated with 1%, 2%, and 5% MoA and HoA as additives.The device structure composed of ITO / MeO-2PACz / perovskite / C 60 / BCP / Ag is depicted schematically in Figure S1 (Supporting Information).For both MoA and HoA additives showed higher PCE than the control device when added in small amounts but showed a decreasing trend with increasing additive concentration.This result is in accordance with the XRD and SEM results, indicating that the optimal concentration of the additive is 1% under these experimental conditions.In addition to efficiency, the stability of PVSCs is a crucial factor affecting their commercial applications. [22]Therefore, we also fabricated inverted PVSCs, including a control device (FAPbI 3 ) and devices with 1% MoA and HoA additives, to investigate their stability by conducting a continuous pseudo-solar irradiation test with a solar simulator (1 sun) in air (temperature 25 °C, relative humidity 30-40%).As shown in Figure 3c, HoA 1%, MoA 1% and the control devices reached approximately 78%, 68%, and 30% of their initial power conversion efficiency after 100 h, respectively.Although the HoA and MoA devices degraded at similar rates, devices containing 1% HoA exhibited better durability than devices containing 1% MoA, indicating that the addition of HoA significantly improved PVSC stability by maintaining a higher PCE over time.
The results obtained from the UV-vis absorption spectra, SEM, and XRD measurements provide strong evidence that the addition of HoA to the perovskite precursor solutions effectively modified the perovskite, resulting in an increase in PCE from 14.81% to 16.50% (Table S1, Supporting Information) and a longer lifetime of the perovskite cells.

Comparison of Hydroxylamine-Based Additives
To confirm whether other hydroxylamine-based additives have the same favorable effects on PVSCs as the hydroxylamine molecules studied above, similar tests were conducted on the three hydroxylamine derivatives with the substituents shown in Figure 4.The derivatives are N-MHoA, N-IpHoA, and N-tBHoA.
Since the results described above clearly indicated that a 1% additive concentration was optimal, subsequent experiments were performed with a fixed additive concentration of 1%.
UV-vis absorption spectroscopy was conducted to examine the effects of the three hydroxylamine-based additives on the optical properties of the perovskite films.Figure 5a shows that the absorption onset wavelength remained unchanged regardless of the hydroxylamine-based additive used, suggesting that the addition of hydroxylamine-based additives during the fabrication process did not affect the bandgap of the perovskite material.XRD measurements were performed to investigate the effect of hydroxylamine-based additives on the perovskite crystal structure.Figure 5b confirms that PbI 2 -derived peaks were generated in the control film, whereas the peaks derived from both the phase and PbI 2 were significantly suppressed in the films with each of the three hydroxylamine additives, indicating that the perovskite films with these additives exhibited a more desirable perovskite crystal structure than the control film. [23]The SEM images in Figure 5c show the influence of hydroxylamine-based additives on the perovskite film surface.On the surface of the film with hydroxylamine-based additives, the amount of PbI 2 was obviously lower than that of the control film.It was also found that the grain size of perovskite with N-MHoA was larger than the control film.The observed increase in the grain size of the perovskite is expected to improve the photovoltaic properties owing to the reduction in nonradiative recombination caused by the minimized grain boundaries. [24]igure 6a and Table 1 show the photovoltaic properties of the inverted PVSCs with the three new hydroxylamine-based additives.The energy diagrams of the fabricated devices are shown in Figure S2 (Supporting Information).The control device exhibited a PCE of 14.81%, whereas devices with hydroxylamine additives exhibited PCEs higher than 16%.In particular, the use of the N-MHoA additive resulted in a remarkable improvement in PCE, with the device exhibiting a J SC of 21.62 mA cm −2 , V OC of 1.05 V, fill factor of 0.76, and PCE of 17.18%.The J SC values calculated from the integration of the external quantum efficiency (EQE) spectra agree well with the J SC values obtained from the J-V curves.EQE spectrum of N-MHoA cell is provided in Figure S3 (Supporting Information).It was also observed that the hysteresis was very small for the N-MHoA device, suggesting that the presence of N-MHoA can effectively passivate the defects on the surface of the perovskite films.Table S1 (Supporting Information) lists the characteristics of devices with the five additives investigated in this study.Figure 6b shows the results of continuous light exposure lifetime measurements using a solar simulator.Compared with the control device, the use of these hydroxylamine additives tended to enhance the long-term durability, whereas the HoA and N-MHoA devices showed substantially improved stability, retaining more than 70% of the initial PCE even after 100 h.The N-MHoA and N-IpHoA devices showed initial degradation in the first few hours but thereafter showed relatively stable characteristics.In contrast, the N-tBHoA-based device underwent rapid degradation within the first few hours of operation.These results highlight the potential of hydroxylamine-based additives, particularly HoA and N-MHoA, for further improving the efficiency and stability of inverted lead halide PVSCs without MA + .In addition, Figure 6c shows the maximum steadystate photocurrent and the verified output at the maximum output point for the control device and the device with 1% N-MHoA, which showed the best performance.The control device yielded a PCE of 14.31% and a photocurrent of 19.17 mA cm −2 .In contrast, the device incorporating N-MHoA exhibited a PCE of 16.52% and a photocurrent of 20.19 mA cm −2 .These results suggest that N-MHoA is effective in improving operational stability under practical conditions.Furthermore, Figure 6d shows the results of space charge limited current (SCLC) measurements of a glass/ITO/MeO-2PACz/perovskite/PTAA/Ag device fabricated to investigate the effect of additives on perovskite defect den-sity.The defect density was calculated according to the following equation:    N defects is the trap density,  is the relative permittivity,  0 is the vacuum permittivity, V TFL is the trap filling limiting voltage, which is the intersection of the ohmic curve and the TFL fitting curve, q is the elementary charge, and L is the thickness of the perovskite film, respectively. [25,26]The VTFL values for the control, HoA 1%, and N-MHoA 1% devices were 0.73, 0.31, and 0.29 V, respectively.Correspondingly, the calculated defect state densities were 2.36 × 10 16 , 1.00 × 10 16 , and 9.39 × 10 15 cm −3 for the control, HoA 1%, and N-MHoA 1% devices.These results indicate that hydroxylamine additives are effective in improving the crystallinity of perovskites and decreasing the density of defect states.
To elucidate the mechanisms responsible for the observed enhancements in terms of PCE and lifetime, the results were analyzed with XRD, XPS, and FT-IR.As shown in Figure 7a, the XRD results show that there was no shift in peak position even when hydroxylamine-based additives were added, suggesting that the additives were not incorporated into the perovskite crystal lattice.The XPS analysis shown in Figure 7b revealed the presence of an O1s-derived peaks from hydroxylamine-based additives on the perovskite surface.These results also suggest that the additives may not incorporated into the perovskite crystal lattice but rather exist on the perovskite surface or grain boundaries.Furthermore, as shown in Figure 7c, the characteristic peaks of Pb 4f 5/2 and Pb 4f 7/2 for the film with added N-MHoA, which showed the best performance, and the control film shifted to the lower energy side, from 143.47 to 143.32 eV and 138.56 to 138.41 eV, respectively.This trend was also observed for the I 3d 3/2 and I 3d 5/2 peaks.This shift in peak position indicates that there is interaction between the additive molecules and perovskite through hydrogen and coordination bonds, suggesting that the hydroxylamine additive can effectively deactivate crystal defects.
To further understand how the additive affects the perovskite, FT-IR spectra are shown in Figure 7d: peaks near 3340, 1600, and 1350 cm −1 correspond to N-H stretching vibrations and O-H stretching vibrations, N-H stretching vibration, and C-N stretching vibration, respectively.As the results, when N-MHoA is used as an additive, the N-H and O-H vibrations around 3340 cm −1 shift slightly downward and the peak intensity is enhanced.Similar changes are also observed for N-H vibration around 1600 cm −1 , and C-N vibration around 1350 cm −1 .These changes suggest the formation of hydrogen bond-mediated interactions between perovskite and N-MHoA. [27]From the above XPS and FT-IR results, it is considered that the interaction between the hydroxylamine additive and perovskite may be formed through I⋯N-H or I⋯O-H hydrogen bonding, which contributes to the passivation of the iodine vacancy and improves the stability and power conversion efficiency.

Conclusions
The use of a small amount (1%) of hydroxylamine-based additive in the precursor solution for the FAPbI 3 -based perovskite films suppressed the formation of unwanted PbI 2 and -phase crystal structures, resulting in the preferential formation of -phase cubic perovskite structures suitable for solar cells.Devices fabricated with compositions containing hydroxylamine-based additives exhibited obviously improved conversion efficiencies and smaller J-V hysteresis in the voltage sweep direction.In particular, a perovskite film with N-MHoA as an additive showed an increased grain size, improved photovoltaic characteristics such as short-circuit current density, open-circuit voltage, and a PCE of 17.18%.The stability under 1-sun light irradiation was also greatly improved compared to that of the cell without additives.
Preparation of Hole Transport Layer Solutions (MeO-2PACz): In an N 2filled glovebox, 4.02 mg MeO-2PACz and 6 mL of IPA solvent were added to a 6 mL glass vial, which was then stirred at 45 °C overnight.Right before being used, the solution was filtered through a 0.45 μm PTFE filter to make a 2 mм solution of MeO-2PACz.
Preparation of Perovskite Precursor Solutions: FAI and PbI 2 were weighed separately in 4 mL glass vials under a nitrogen atmosphere, and then mixed with DMF and DMSO at a volume ratio of 4:1 to create 1.48 м solutions for each compound.The solutions were stirred at 45 °C overnight and then filtered through 0.45 μm PTFE filters before use.To prepare the FAPbI 3 precursor solution, 200 μL of each FAI and PbI 2 solution were mixed.KI was weighed in another 4 mL glass vial under a nitrogen atmosphere and mixed with DMF and DMSO at a volume ratio of 4:1 to make a 1.48 м solution, which was also stirred at 45 °C overnight and filtered prior to its usage.A hydroxylamine-based additive was weighed into a 4 mL glass vial under a nitrogen atmosphere.Then, 400 μL of DMF and 100 μL of DMSO were added to make a 1.48 м L solution.This solution was stirred at 45 °C overnight and filtered before use.Precursor solutions with different amounts of additives were prepared by mixing the FAI, PbI 2 , KI, L solutions with 1%, 2%, and 5% L additive added based on mole ratio for comparison.
Device Fabrication and Characterization: Inverted PVSCs were fabricated using ITO-coated glass substrates with a sheet resistance of 10 Ω sq −1 .The patterned ITO-coated glass substrates were sequentially washed in an ultrasonic bath containing detergent, deionized water, acetone, and isopropanol for 30 min each.After washing, the substrates were oven dried at 65 °C for 12 h before use.The substrates were then treated with ultraviolet (UV) ozone for 20 min.Subsequently, the prepared MeO-2PACz solution (100 μL) was spin-coated onto the ITO substrates at 3000 rpm for 30 s and dried at 100 °C for 10 min.Next, the perovskite precursor solution (100 μL) was filtered through a 0.45 μm PTFE filter and spin-coated onto the MeO-2PACz layer at 1000 rpm for 10 s and 6000 rpm for 30 s, for a total of 40 s.Chlorobenzene (150 μL) was added 15 s after the spin-coating process started.The films were then annealed at 150 °C for 30 min.The thickness of the perovskite layer was approximately 370 nm, as measured using a Kosaka laboratory microfigure-measuring instrument (ET-200).
Once the perovskite layers were baked, the substrates were transferred to a high-vacuum chamber for deposition.C 60 (20 nm), BCP (8 nm), and Ag (70 nm) were thermally deposited on the perovskite layer using a shadow mask at a rate of 1 Ås −1 each and pressures below 9 × 10 −5 Pa.The effective area of the device was 9 mm 2 .Current density-voltage (J-V) characteristics were measured using a CEP-2000 integrated device manufactured by Bunko Giken Inc., simulating AM 1.5G.Before the fabricated devices were measured, J-V was calibrated using standard solar cells.Device measurements were performed in reverse scan (1.2 → 0 V, step 0.02 V, delay time 200 ms) and forward scan (0 → 1.2 V, step 0.02 V, delay time 200 ms).The maximum steady-state photocurrent and corresponding output at the maximum output point were calculated by scanning the voltage every 7 s in steady-state light conditions.The photovoltaic properties of the encapsulated PVSCs were evaluated in the laboratory under ambient conditions.The UV-vis absorption spectra of the perovskite films were analyzed using an ultraviolet-visible spectrophotometer (UV2600, Shimadzu Corporation).Fourier-transform infrared spectra were determined using a Shimadzu IRAffinity-1S spectrophotometer.The SEM was performed using a JEOL JSM-7600FA microscope.XRD measurements were performed using an X-ray diffractometer (Rigaku Corporation) for the XRD analysis.XPS was performed using an X-ray photoelectron spectrometer (Thermo Fisher Scientific Inc.).Continuous pseudo-sunlight irradiation tests were evaluated using a continuous irradiation life device W32-2400SOL3 (System House Sunrise, Inc.).

Figure 3 .
Figure 3. a) J-V curves of inverted PVSCs with varying amounts of HoA and b) MoA additives with both forward and reverse scans.c) Device photostability under 1-sun irradiation at 25 °C and in ambient air.

Figure 5 .
Figure 5. a) UV-vis absorption spectra of the perovskite films with hydroxylamine-based additives.b) The XRD patterns of the perovskite films with and without hydroxylamine-based additives.c) The top-view SEM images of the ITO / MeO-2PACz / perovskite.

Figure 6 .
Figure 6.a) J-V curves for inverted PVSCs with different additives in the forward (F) and reverse (R) scan directions.b) Device photostability under 1-sun irradiation at 25 °C and in ambient air.c) The steady-state current density and PCE output of the PSCs with and without N-MHoA.d) Dark J-V curves of device with and without hydroxylamine-based additives with hole-only structure.

Figure 7 .
Figure 7. a) XRD patterns of the perovskite films with and without hydroxylamine-based additives.b) XPS spectra of the perovskite films with and without hydroxylamine-based additives on the O 1s signal.c) XPS spectra of the perovskite films with and without N-MHoA on the Pb 4f signal and I 3d signal.d) FT-IR spectra with and without N-MHoA.

Table 1 .
Photovoltaic performance of inverted PVSCs with different additives in the forward (F) and reverse (R) scan directions.