of the initial PCE after 700 hours of operation under continuous light soaking at 1 sun illumination at 55± 5 ºC. This work represents a viable route to the sustainable implementation of the solar energy.

Photovoltaics represent a significant advancement in the field of renewable energy, particularly in the pursuit of sustainable and environmentally friendly energy solutions. Metal halide perovskite solar cells (PSCs) have demonstrated significant advancements, achieving certified power conversion efficiency (PCE) exceeding 26% 1,2 . Despite the significant progress made in enhancing efficiency, the practical deployment of PSCs necessitates addressing critical issues concerning longevity and scalability [3][4][5][6] . Notably, research efforts have primarily concentrated on perovskite absorbers with bandgap within the range of 1.49-1.56 eV [7][8][9][10][11][12][13][14][15][16] . However, to surpass the theoretical efficiency limit of single-junction solar cells, the integration of widebandgap (WBG) perovskite (with a bandgap of 1.65-1.70 eV) with crystalline silicon in tandem solar cells has proven to be an exceptionally promising strategy to boost the efficiencies to a certified value of 33.9% [17][18][19][20] . Nevertheless, WBG PSCs have not achieved the same level of PCE compared to those with narrower bandgaps, with their efficiencies constrained by larger loss of open-circuit voltage (VOC) 21 . Furthermore, scalability remains a crucial concern for WBG PSCs, especially since they are designed to be the top cells for wafer-sized silicon solar cells 22,23 .

Nearly all the reported high efficiency WBG perovskite and tandem cells were fabricated by spin-coating method. The reliance on glovebox-dependent solution-processes for upscaling is not commercially viable, making it imperative to produce WBG PSCs in ambient air for commercial production 24 . Unfortunately, there are only a few reports on scalable deposition of WBG perovskites in ambient condition, and none on efficient methylammonium (MA)-free WBG PSCs 21,25 . The reaction of methylammonium (MA + ) with formamidinium (FA + ), which has been frequently observed, could change the stoichiometry of perovskites and thus limit the stability at elevated temperatures 26,27 . Moreover, mixed-halide WBG PSCs have particularly struggled with poor operational stability, primarily due to halide-demixing induced phase segregation 28,29 . Xu et al. demonstrated using triple-halide WBG perovskites (bandgap of 1.67 eV) to minimize phase segregation by decreasing the amount of Br -to less than 20% 30 . In their study, chloride (Cl -) was incorporated to stabilize WBG perovskites by shrinking the lattice parameter and reducing Br content. Halide oxidation reaction has been proposed to explain halide segregation under operation conditions 28 . In this scenario, halide segregation is initiated by the oxidation of the most easily oxidized halide species, providing a gradient driving force for these species to migrate across the film. Therefore, minimizing or suppressing iodide oxidation may mitigate the halide-demixing induced phase segregation.

In this work, we report that the partial substitution of Cs + or FA + cation by a reductive methylhydrazinium (MHy + ) cation in WBG perovskites (Cs0.25FA0.75 Pb(I0.82Br0.15Cl0.03)3) with a bandgap of 1.65 eV can effectively inhibit the iodide oxidation and increase migration barrier for iodide ions, which mitigates phase segregation and improves the stability of WBG PSCs. Furthermore, hydrazinium group provides good defect passivation through Pb-N bonding, resulting in a record low VOC deficit of 0.37 V for WBG PSCs and minimodules. Consequently, we obtain a champion efficiency of 23.3% for small-area WBG PSCs with a VOC of 1.28 V. The uniformity of the films made by scalable blading process is also confirmed by the demonstration of efficient minimodules with a stabilized efficiency of 19.8%.

A reductive A-site cation, methylhydrazinium (MHy + ), was incorporated to partially substitute either formamidinium (FA + ) or Cs + to form a triple-cation, triple-halide WBG perovskites, denoted as MHy1-x-yCsxFAyPb(I0.82Br0.15Cl0.03)3, where x (from 0.05 to 0.25) and y (from 0.55 to 0.75) are the percentages of Cs and FA, respectively. MHy + has a larger ionic radius (264 pm) compared to the normally used A-site cations including FA + , MA + , and Cs + . Previous studies have reported that MHy + can form both three-dimensional (3D) perovskites (MHyPbBr3) or two-dimensional (2D) perovskites (MHy2PbBr4 or MHy2PbCl4) depending on synthesis conditions [31][32][33][34][35] . However, it remains uncertain whether MHy + can be partially incorporated into 3D WBG perovskites, despite the successful substitution of MA + with MHy + in MAPbI3. 36 In addition, it is unclear whether MHy + ions would exist in the formation of 2D perovskites in our WBG perovskite films or leave the films during thermal annealing. Therefore, we first investigated whether MHy + can remain in the annealed WBG perovskite films using nuclear magnetic resonance (NMR) measurements. The 1H NMR spectra of the annealed WBG perovskite with MHy + , or MHy-WBG perovskite, showed a signal peak belonging to MHy + (2.6 ppm), which is identical to the peak from MHyI (Supplementary Fig. 1). This proves that MHy + can stay in the final WBG perovskite films. The content of MHy in the annealed MHy-perovskite film was estimated by calculating the integrated area of the characteristic peak of 1 H in MHy + to that of FA + .

The calculated MHy + /FA + ratio was 6.8%, which is close to the ratio in the precursor solution (7.1%). Scanning electron microscopy (SEM) images show that the introduction of 5% MHy + did not notably change the surface morphology of WBG perovskite films (Supplementary Fig. 2). To check whether MHy + can enter the lattice and assess the dimensionalities of MHy + species, we performed the X-ray diffraction (XRD) measurements on WBG perovskite films with varying MHy + substitution ratios from 0 to 20%. As shown in Supplementary Fig. 3, the substitution of FA + by MHy + resulted in a shift of XRD peaks towards smaller diffraction angles, leading to an expansion of the lattice constant in all crystallographic directions. Furthermore, we did not detect any diffraction peaks at small angles that belong to MHy + -based 2D perovskite even when MHy + was increased to 20%.

We performed density functional theory (DFT) calculations to explore the role of MHy + on structural and electronic properties of WBG perovskites. Since halide migration has been often reported to have a smaller activation energy than A-cation migration, 37 we focused on halide migration in this study. We first calculated the migration energy barrier of iodide interstitial (Ii) in Cs0.25FA0.75PbI3, under two scenarios. In both scenarios, MHy + replacing Cs + and FA + were examined (Supplementary Fig. 4). The simulations were based on the nudged elastic band (NEB) method to calculate the migration barriers of Ii by considering Ii migrating from initial position to the adjacent lattice site along the shortest path in the (110) direction (see the migration pathways in Fig. 1). The DFT results reveal that MHy + -substituted perovskite exhibits a higher migration barrier (0.76 eV) for iodide ions traveling via an interstitial site when 5% of either Cs + or FA + is replaced by MHy + (0.58 eV and 0.68 eV, Fig. 1c-d and Supplementary Fig. 5). This suggests that the inclusion of MHy + can suppress Ii migration, likely due to lattice distortion, as compared to Cs + /FA + . The local lattice distortion, whether due to the size mismatch of A-site cations or the formation of Pb-N coordination bonds, may be responsible for the increased migration energy barrier 36,38,39 . It is worth noting that, while our DFT calculations offer some insights into the increased migration energy barrier associated with lattice distortions, they may not fully capture the complex, dynamic processes involved, such as the continuous formation of new interstitials under illumination and the diverse redox reactions involving iodide species. Furthermore, MHy + contains a strong reducing function group, which has been demonstrated to effectively reduce the detrimental I2 to I -. 40 Iodine generation can occur in both precursor and perovskite over aging. A transparent FAI solution turned into light yellow when heated at 60 ºC for two hours in ambient air conditions, indicating the formation of I2 by oxidization. To test the reducing capability of MHy + , we introduced MHyI into the oxidized FAI solution at a 1% weight ratio. Upon the addition of MHyI, the solution became transparent, demonstrating the effectiveness of MHy + in reducing I2 to I -(Supplementary Fig. 6). 41 Similar strategies, such as using dimethylammonium formate (DMAFo) as an additive, have also been adopted to inhibit iodide oxidation in air-processed perovskite solar cells. 41 But the primary contribution of DMAFo is the enhancement of perovskite precursor stability.

We then carried out several measurements to investigate the photostability of WBG perovskite films with and without MHy + . Fig. 2a-b show the photoluminescence (PL) spectra of encapsulated Cs0.25FA0.75Pb(I0.82Br0.15Cl0.03)3 and MHy0.05Cs0.25FA0.7Pb(I0.82Br0.15Cl0.03)3 perovskite films after aging under continuous 1-sun illumination for different durations up to 500 hours. The Cs0.25FA0.75Pb(I0.82Br0.15Cl0.03)3 perovskite film showed an obvious PL redshift, and full width at half maximum (FWHM) of PL peak became larger with increasing aging duration, suggesting the emerging of the photo-induced film degradation (Fig. 2c). In contrast, the MHy0.05Cs0.25FA0.7Pb(I0.82Br0.15Cl0.03)3 perovskite film showed no PL peak shifting or broadening.

A photoelectrochemical model has been developed by Kerner et al., which rationalizes that iodide oxidation is the first step of mixed halide segregation 28,42 . To elucidate the role of MHy + in preventing iodide oxidation, we tracked the formation of I2 in WBG perovskite films by immersing the films in toluene at 65 ºC when the films were illuminated for 500 h under continuous light at 1 sun intensity. We then measured the absorption spectra of the toluene solution to evaluate the released I2 during the aging. As shown in Fig. 2d, ultraviolet-visible (UV-vis) absorption measurements show that the toluene solution with aged control WBG perovskite film has a distinct absorption peak at 360 nm, which is assigned to I2. By contrast, the MHy + -incorporated perovskite film did not generate I2 after long-time illumination at high temperature, proving that MHy + can inhibit iodide oxidation. Therefore, employing the MHy⁺ cation can facilitate efficient reduction of iodine not only in the solution state but also in the film state since it can be incorporated into perovskite lattice.

To verify the increased iodide migration energy barrier due to MHy + substitution from DFT calculation, we measured the temperature-dependent conductivity to determine the activation energy (Ea) of ion migration. Based on the Arrhenius plot shown in Fig. 2e, we derived Ea values of 0.43 eV and 0.81 eV for control and MHy + WBG perovskite films, respectively, which are in alignment with the DFT calculation results. We further estimated the concentration of mobile ions using a mobile ion charging-discharging method. The typical transient currents in devices are shown in Fig. 2f. The calculated mobile ion concentrations decrease from 4.2 × 10 15 cm -3 for the control WBG device to 2.4 × 10 15 cm -3 for MHy + device 43 . The mobile ion concentration in the MHy + device is approximately two times lower than that in the control device, as the result of improved crystallinity after MHy + incorporation 44 . The reduced mobile ion concentration together with the increased ion migration activation energy are beneficial for suppressing phase segregation. 38 Overall, the mechanism of suppressed phase segregation in MHy + -incorporated WBG perovskites is attributed to two key factors. The first is the potential of the reductive MHy + cation to prevent halide oxidation, which in turn slows down the phase segregation process. The halide oxidation has suggested to initiate the phase segregation process. 28 The second factor is associated with the increased migration energy barrier of iodide species induced by the lattice distortion, thereby mitigating the phase segregation of WBG perovskite during operation.

We extended our DFT calculations to examine surface passivation by MHy + , focusing on two representative perovskite surfaces (i.e., FAI-and PbI2-rich). As illustrated in Supplementary Fig. 7, for the FAI-rich surface, the replacement of surface FA + with MHy + enhances the binding energy of A-site cation with perovskite from -4.93 eV to -5.14 eV; and for the PbI2-rich surface, the MHy + ions passivate the undercoordinated Pb 2+ ions at the surface through the formation of Pb-N coordination bonds. In the perovskite bulk, the MHy + replacement can suppress the formation of iodide interstitials, as evident by the increased formation energy of iodide interstitials compared to the control case under the moderate growth condition (1.47 eV vs 1.10 eV, Supplementary Fig. 8). Therefore, MHy + ions play a dual role in reducing defect densities within perovskites and enabling higher activation energy for iodide migration, underscoring its significance in enhancing WBG perovskite device stability and efficiency.

To have a better insight into the effects of MHy + on the optoelectronic properties of WBG perovskite films and devices, we performed various measurements. To evaluate the passivation effect of MHy + , we studied the charge carrier recombination kinetics of the 5% MHy + -substituted substitution can effectively suppress the nonradiative recombination. Confocal PL intensity and

PL lifetime mapping were used to further understand the charge carrier dynamics uniformity in WBG perovskite films (Fig. 3b-c). The MHy + perovskite film shows uniformly enhanced PL intensity and recombination lifetime.

We also investigated the charge carrier dynamics in WBG PSCs. As shown in Fig. 3d, electroluminescence (EL) spectra of the control and MHy + devices were recorded at an injection current of 21 mA/cm 2 (close to device's JSC value). The MHy + device shows a 3.8-fold higher EL intensity than the control device, which is expected to increase the VOC by ~35 mV. We then measured transient photovoltage (TPV), and trap density state (tDOS) of control and MHy + devices.

The TPV was conducted with a light bias of 1 sun light intensity which made the devices operate at VOC condition for both control and target devices. This was evidenced by the observed VOC of 1.27 V for the MHy + device and 1.23 V for the control device. A perturbation of carrier concentration was induced by a weak nanosecond pulse laser, and the decay of the perturbation carrier density has a lifetime of 0.76 µs in the MHy + device and 0.48 µs in the control device (Fig. 3e), directly confirmed the slower recombination in the device with MHy + . To find out the origin of longer carrier recombination lifetime, we characterized the defect density in the devices using thermal admittance spectroscopy. The MHy + device exhibits a lower trap density in the trap depth region from 0.25 to 0.40 eV characterized by thermal admittance spectroscopy (Fig. 3f). The observed trap bands are associated with positive and negative iodide interstitials 21,45 , agreeing with the computation results that big size MHy + incorporation can reduce iodide interstitial formation.

The densities of both negatively charged iodide interstitial (Ii -) and positively charged iodide interstitial (Ii + ) are reduced. Hole trapping at iodide sites, such as negative Ii -, has been proposed to induce lattice instability and halide phase segregation in mixed halide perovskites. 46 Our previous result shows that reducing the concentration of Ii + defect will suppress energy loss for PSCs. The primary enhancement in device efficiency can be attributed to the increased FF and VOC, originating from reduced charge recombination and prolonged carrier lifetime. We further investigated the long-term operational stability of encapsulated WBG PSCs under 1 sun illumination. No temperature controlled was applied, so the temperature of the devices was increased to 55±5 ºC by the illumination light. The encapsulated MHy + WBG PSCs demonstrated a remarkably low PCE reduction of less than 5% after 800 hours, representing one of the best reported stabilities for WBG PSCs (Table S1). In contrast, the control device exhibited approximately 20% loss within the same timeframe (Fig. 4e).

We evaluated whether this composition can be upscaled by fabricating WBG perovskite modules using the blade coating process in ambient conditions. Fig. 5a shows a photograph of an encapsulated WBG perovskite minimodule with an aperture area of 25 cm 2 . The subcell width for WBG perovskite module was 6.5 mm. A high geometric FF of 96% was achieved with a narrow dead area width of 270 µm (Fig. 5a). We fabricated multiple MHy + WBG perovskite minimodules, which showed an average PCE of 18.4±0.8% (Supplementary Fig. 11). The best-performing MHy + WBG perovskite minimodule exhibited an aperture PCE of 19.9%, with a VOC of 8.89 V, short-circuit current (ISC) of 73.94 mA, and a FF of 0.757 (Fig. 5b). It is notable that each subcell has a VOC of 1.27 V from the minimodule, showing the good reproducibility of the composition.

The stabilized minimodule aperture PCE is 19.8% at a bias voltage of 6.7 V, as shown in Fig. 5c.

The operational stability of a MHy + WBG perovskite minimodule was also tested. Fig. 5d shows the evolution of PCE of an encapsulated WBG perovskite module under 1 sun illumination under VOC condition in ambient air. The encapsulated WBG perovskite minimodule maintained 94% of the initial PCE after 700 h of operation under continuous 1 sun illumination (55±5 ºC). Fig. 5e shows the I-V curves of WBG perovskite minimodules after operational testing for different durations. The PL spectra of the WBG perovskite minimodules were recorded before and after operational stability test (Supplementary Fig. 12), which showed no obvious change in terms of PL peak shift and broadening.

In summary, we have successfully demonstrated that the reductive A-site cation MHy + can be incorporated into WBG perovskites, improving both efficiency and stability of WBG perovskite devices. The reductive MHy + can inhibit the halide oxidation and provide high diffusion barrier for the migration of halide species, which significantly suppresses the halide segregation of WBG perovskites, thus improving the operational stability. We found that MHy + can provide uniform passivation to perovskite films, enabling a VOC of 1.28 V for WBG perovskite with a bandgap of 1.65 eV, which is 93.4% of the Shockley-Queisser limit VOC (1.37 V). In addition, we have demonstrated the scalable fabrication of high-performance of WBG perovskite modules in ambient air conditions, achieving a champion module efficiency of 19.9% with an aperture area of 25 cm 2 along with good long-term operational stability. By providing a viable pathway for the large-scale deployment of high-performance solar modules in ambient air, our work supports the transition to clean energy sources, helping to reduce greenhouse gas emissions and promote environmental sustainability.

Lead(II) iodide (PbI2, 99.99% trace metals basis, TCI), Formamidinium iodide (FAI, 99.99%, GreatCell Solar),, 4-fluoro-phenethylammonium iodide (4-F-PEAI, >99%, GreatCell Solar), L-αphosphatidylcholine (LP, ≥99%, Sigma-Aldrich), 5%v/v FAH2PO2 (synthesized in our lab), dodecylammonium iodide (GreatCell Solar), lead bromide (PbBr2), caesium iodide (CsI), BCP, dimethyl sulfoxide, N, N-Dimethylformamide (DMF), isopropanol (IPA), bathocuproine (BCP), 2-methoxyethanol (2-ME) and toluene were purchased from Sigma-Aldrich, N-methyl-2pyrrolidone (NMP, 99.9%, Sigma-Aldrich), TPABr3 (TCI), C60 was purchased from NANO-C company, Copper (Cu) pellets were purchased from Kurt J. Lesker Company.

Synthesis of Methylhydrazine iodide (MHyI): 2.63 mL of methylhydrazine solution was diluted with 5mL of anhydrous IPA and stirred with an ice bath. Then 100 mL of HI solution was added to the above solution. The reaction was stirred for a further 2 h and then worked up by evaporation of the solvent. The white precipitate was recrystallized and washed 2-3 times with ethanol and diethyl ether. The white MHyI power obtained was dried in a vacuum oven at 60 ºC. The MHyI power was dried in a vacuum oven at 60 ºC.

Patterned ITO glass was cleaned with acetone and isopropanol using ultrasonic cleaning for 15 minutes, followed by 15 minutes of ultraviolet ozone treatment. The double hole transport layers were sequentially blade coated on ITO substrates by using 0.5 mgmL -1 Meo-2PACz in methanol and 2.2 mgmL -1 PTAA in toluene with a 200 μm blade gap (same coating speed of 20 mm/s). In detail, Meo-2PACz was first blade-coated on ITO substrate and then annealed at 100 °C for 5 min and washed by methanol to remove the residue molecule. PTAA layer was then blade-coated onto the Meo-2PACz-coated ITO substrate. To blade-coat MHyxCs0.25FA0.7-xPb(I0.82Br0.15Cl0.3)3 WBG perovskite films, 1.4 M MHyxCs0.25FA0.7-xPb(I0.82Br0.15Cl0.3)3 ink solution was dissolved in DMF with FAH2PO2 (0.15 wt%), LP (0.05 wt%, LP was dissolved in 2methoxyethanol), n-octylammonium iodide (0.1 wt%), phenylethylammonium chloride (0.05 wt%), TPABr3 (0.1 mol%) and 20 mol% of NMP (compared with lead), and then blade coated onto a PTAA/Meo-2PACz-coated ITO substrate using an N2 knife (20 psi) at a coating speed of 20 mms -1 and coating gap of 250 μm. The as-coated solid perovskite films were then annealed at 70 °C for 10 min and followed by annealing at 100 °C for another 3 min in air. Then 30 nm of C60, 6 nm of BCP and 80 nm of copper were sequentially deposited by thermal evaporation. The minimodules were fabricated on pre-patterned large ITO glass substrates with P1 width of 30 μm followed by the same procedure as the small solar cells. The laser scribing was performed twice with a Keyence laser marker (MD-U1000C, 355 nm). The width of each subcell is 6.5 mm, while the final P2 and P3 widths were measured to be ~80 and ~60 μm, respectively.

The total width of the non-working area was measured to be ~250 μm, giving a GFF of 94.7%. A polydimethylsiloxane (PDMS) layer was attached to the front side of the perovskite solar cells as an anti-reflection coating.

The J-V characteristics of solar cells and modules were measured using a Keithley 2400 source meter under the simulated AM 1.5 G 1 sun illumination (100 mW cm -2 ) using a solar simulator The energy barriers for iodine ion migration were calculated by taking the differences between the total energy of the ground state for the perovskite supercells under consideration and the energies at the saddle points along the diffusion pathway. Mhy0.05Cs0.2FA0.75PbI3 with and without an I interstitial (Ii). The red arrow represents the Ii migration pathways. c,d, Calculated relative activation energies for the Ii migration in Cs0.25FA0.75PbI3 and MHy0.05Cs0.2FA0.75PbI3. The MHy-substituted perovskite exhibits a higher energy barrier for Ii migration (0.76 eV) compared to Cs0.25FA0.75PbI3 (0.58 eV). The DFT calculations were performed at generalized gradient approximation (GGA)/Perdew-Burke-Ernzerh (PBE) theory level of theory. under continuous light illumination. c, Time-dependent PL peak and full width at half maximum (FWHM) of PL. d, Absorptance spectra of toluene solutions in which WBG perovskite films were soaked at 65 °C for 500 hours. e, The temperature-dependent conductivity of the control and MHy + WBG perovskite films. f, Transient ion migration currents. batches. In the boxplots, whiskers: maxima and minima; bounds of box: 25th and 75th percentile; center: mean) are shown. e, Operational stability test of small-area control and MHy + PSCs. WBG perovskite module (left) and a microscope image of all lines P1, P2 and P3. b, The I-V curves of best-performing WBG perovskite module with an aperture area of 25 cm 2 . c, The stabilized efficiency of champion WBG perovskite module. d, Evolution of efficiency of an encapsulated WBG perovskite module. e, I-V curves of the WBG perovskite module measured after light soaking for different durations.