# Fileset

[s41467-024-45228-9.pdf](https://mdr.nims.go.jp/filesets/2a4b1e98-c7cb-469e-95da-1650b9248242/download)

## Creator

[Dhruba B. Khadka](https://orcid.org/0000-0001-9134-3890), [Yasuhiro Shirai](https://orcid.org/0000-0003-2164-5468), [Masatoshi Yanagida](https://orcid.org/0000-0002-8065-7875), [Hitoshi Ota](https://orcid.org/0000-0002-8339-9592), [Andrey Lyalin](https://orcid.org/0000-0001-6589-0006), Tetsuya Taketsugu, [Kenjiro Miyano](https://orcid.org/0000-0002-5869-3087)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Defect passivation in methylammonium/bromine free inverted perovskite solar cells using charge-modulated molecular bonding](https://mdr.nims.go.jp/datasets/4051e175-caf9-4d80-9db7-99ccdab82c9d)

## Fulltext

Defect passivation in methylammonium/bromine free inverted perovskite solar cells using charge-modulated molecular bondingArticle https://doi.org/10.1038/s41467-024-45228-9Defect passivation in methylammonium/bromine free inverted perovskite solar cellsusing charge-modulated molecular bondingDhruba B. Khadka 1 , Yasuhiro Shirai 1 , Masatoshi Yanagida 1,HitoshiOta 2,AndreyLyalin 3,4 , TetsuyaTaketsugu 4,5&KenjiroMiyano 1Molecular passivation is a prominent approach for improving theperformanceand operation stability of halide perovskite solar cells (HPSCs). Herein, wereveal discernible effects of diammoniummoleculeswith either an aryl or alkylcore ontoMethylammonium-free perovskites. Piperazine dihydriodide (PZDI),characterized by an alkyl core-electron cloud-rich-NH terminal, proves effec-tive inmitigating surface and bulk defects andmodifying surface chemistry orinterfacial energy band, ultimately leading to improved carrier extraction.Benefiting from superior PZDI passivation, the device achieves an impressiveefficiency of 23.17% (area ~1 cm2) (low open circuit voltage deficit ~0.327 V)along with superior operational stability. We achieve a certified efficiency of~21.47% (area ~1.024 cm2) for inverted HPSC. PZDI strengthens adhesion to theperovskite via -NH2I and Mulliken charge distribution. Device analysis corro-borates that stronger bonding interaction attenuates the defect densities andsuppresses ion migration. This work underscores the crucial role of bifunc-tional molecules with stronger surface adsorption in defect mitigation, settingthe stage for the design of charge-regulatedmolecular passivation to enhancethe performance and stability of HPSC.Exceptional optoelectronic properties of halide perovskite (HP) havehiked thepower conversion efficiency (PCE) of halideperovskite-basedsolar cells (HPSCs) over 26.1%, approaching to Shockley–Queisserlimit1,2. A deluge of experimental efforts on stoichiometric engineer-ing, crystallinity improvement, interface passivation, and carriertransport engineering has been used in the course of rapid progress3,4.However,HP films are still prone to degradationunder external factors(such as thermal/humidity stress, oxygen, and light) and intrinsicphenomena5,6. These deleterious characteristics have exerted bigchallenges to practicality. Indeed, the HPSC degradation stems fromdeleterious surface chemistries on the surface of HP film or at thedevice interface7. To address these detrimental defect chemistries,molecular passivation has been of great interest in improving the PCEand device stability8,9.Methylammonium (MA)-based HPs with a combination of variouscations and halides have been extensively used in the state-of-the-artHPSCs10–13. However, MA is prone to be released and decomposeswhen exposed to elevated temperatures and humid conditions, whichposes a persistent concern for device stability14–16. In recent years, MA-free formamidinium (FA)-based HP has garnered significant attentiondue to its higher thermal stability, enhanced moisture resistance,better absorbance in the near-infrared region, and tolerance to varyingReceived: 5 June 2023Accepted: 17 January 2024Check for updates1PhotovoltaicMaterials Group, Center for GREEN Research on Energy and Environmental Materials, National Institute for Materials Science (NIMS), 1-1 Namiki,Tsukuba, Ibaraki 305-0044, Japan. 2Battery Research Platform, Research Center for Energy and Environmental Materials (GREEN), National Institute forMaterials Science (NIMS), Namiki 1-1, Tsukuba 305-0044, Japan. 3Research Center for Energy and Environmental Materials (GREEN), National Institute forMaterials Science, Namiki 1-1, Tsukuba 305-0044, Japan. 4Institute for Chemical Reaction Design and Discovery (WPI-ICReDD), Hokkaido University, Sapporo001-0021, Japan. 5Department of Chemistry, Faculty of Science, Hokkaido University, Sapporo 060-0810, Japan. e-mail: KHADKA.B.Dhruba@nims.go.jp;SHIRAI.Yasuhiro@nims.go.jp; lyalin@icredd.hokudai.ac.jpNature Communications |          (2024) 15:882 11234567890():,;1234567890():,;http://orcid.org/0000-0001-9134-3890http://orcid.org/0000-0001-9134-3890http://orcid.org/0000-0001-9134-3890http://orcid.org/0000-0001-9134-3890http://orcid.org/0000-0001-9134-3890http://orcid.org/0000-0003-2164-5468http://orcid.org/0000-0003-2164-5468http://orcid.org/0000-0003-2164-5468http://orcid.org/0000-0003-2164-5468http://orcid.org/0000-0003-2164-5468http://orcid.org/0000-0002-8065-7875http://orcid.org/0000-0002-8065-7875http://orcid.org/0000-0002-8065-7875http://orcid.org/0000-0002-8065-7875http://orcid.org/0000-0002-8065-7875http://orcid.org/0000-0002-8339-9592http://orcid.org/0000-0002-8339-9592http://orcid.org/0000-0002-8339-9592http://orcid.org/0000-0002-8339-9592http://orcid.org/0000-0002-8339-9592http://orcid.org/0000-0001-6589-0006http://orcid.org/0000-0001-6589-0006http://orcid.org/0000-0001-6589-0006http://orcid.org/0000-0001-6589-0006http://orcid.org/0000-0001-6589-0006http://orcid.org/0000-0002-1337-6694http://orcid.org/0000-0002-1337-6694http://orcid.org/0000-0002-1337-6694http://orcid.org/0000-0002-1337-6694http://orcid.org/0000-0002-1337-6694http://orcid.org/0000-0002-5869-3087http://orcid.org/0000-0002-5869-3087http://orcid.org/0000-0002-5869-3087http://orcid.org/0000-0002-5869-3087http://orcid.org/0000-0002-5869-3087http://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45228-9&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45228-9&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45228-9&domain=pdfhttp://crossmark.crossref.org/dialog/?doi=10.1038/s41467-024-45228-9&domain=pdfmailto:KHADKA.B.Dhruba@nims.go.jpmailto:SHIRAI.Yasuhiro@nims.go.jpmailto:lyalin@icredd.hokudai.ac.jpprocessing conditions17,18. However, it forms a photoinactive “yellowphase” at room temperature in pristine formamidinium (FA)- per-ovskite due to the relatively large size of FA. Alkali or organic cationshave been introduced into the FA-perovskite film to promote thegrowth of the photoactive black phase at low-temperature crystal-lization. For example, Saliba and colleagues have reported the for-mation of a remarkably crystalline photoactive phase of FA-HP,achieved by introducing Rb and Cs ions without using MA and Br andsubjecting the material to annealing temperatures not exceeding100 °C. They have achieved a competitive PCE of 20.35% withimproved stability19. Similarly, Seok and co-workers demonstrated animpressive PCE of 24.4% using alloyed FA-perovskite with organiccation in the normal device configuration20. It has documented theformation of a highly crystalline α- phase by incorporating methyle-nediammonium dichloride as a divalent organic cation with an ionicradius equivalent to FA which induced a stronger ionic interaction ofits divalent state in FA-HP.In recent scenarios, several studies have focused on the versa-tile molecular passivation strategy aimed at mitigating variousintrinsic point defects in the HP film4,21,22. Surface or bulk areas canlead to the formation of under-coordinated Pb2+ ions, A-site vacan-cies/interstitial defects, and halide vacancies, causing recombina-tion and performance loss22–25. The molecular passivator withmultifunctional derivatives consisting of amine26,27, Lewis acids/bases28,29, supramolecules30–32, ionic polymer9,33, etc. have been usedfor mitigation of defect chemistry in the HPSCs. The ammonium-containing functional additives with alkyl or aryl halide or pseudo-halide counterparts have demonstrated a significant enhancementin PCE and operational stability34–41. The additive having strongeradsorption on the HP surface modulates the grain nucleation andgrowth or alleviates defect chemistry at interface and bulk, which iscrucial for a highly efficient and stable device22,42. Additionally, somefunctional molecules eliminate molecular iodine present within theperovskite and suppress both iodine and Pb-related deep trapsites43,44. Importantly, the introduction of molecular functionalderivatives of diammonium halide into alkyl or aryl core, serving asadditives or passivation, has also exhibited noteworthy enhance-ments in both the PCE and stability of HPSCs45–49. For example,Wakamiya and co-workers have demonstrated the universality ofsurface treatment with ethylenediammonium diiodide by wet anddry deposition methods on perovskite surfaces, achieving a decentPCE and device stability in various perovskite systems45,50. Liu andco-workers introduced a highly electronegative fluorine molecule;Cobalt (II) hexafluoro-2,4-pentanedionat for interface passivation ofMA-free HPSCs, resulting in themitigation of defect chemistries andenhancing hole-transport kinetics51. This modification demon-strated a remarkable PCE of 24.64% (normal device structure) as aconsequence of strong molecular interaction with the surfacecharge modulated passivation. Typically, inverted devices exhibitlower efficiency levels than conventional structures. Inverted HPSCswith an inorganic hole transport layer have gained significantinterest due to their potential for superior stability and compat-ibility with tandem solar cells52,53. An inverted HPSC with MA-free HPmixing Tris(chloromethyl) ammonium iodide using NiOx nano-particle as HTL reported a decent certified PCE of 23.2% with a smallactive area (0.04 cm2)42. Despite achieving an impressive PCE ofrecord level in small-area HPSCs, there remains a substantial PCEdisparity between small and large-area PSC devices4,54. Therefore,the fabrication of highly efficient inverted HPSC with a large areacontinues to pose a challenge. Despite a variety of functionalmolecules being wielded for the optimization of crystal growth anddefect passivation in MA-free HP, the influence of adjusting surfacecharge via bonding interactions is often overlooked.In this work, we report on passivation strategy through bond/charge regulated defect passivation by introducing bifunctionalmolecules with an aryl core (1,4-phenylenediamine dihydrodide(PEDAI)) or alkyl core (piperazine dihydriodide (PZDI)) onto the MA/Br-free 3D-HP film (FA0.84Cs0.12Rb0.04PbI3) in inverted HPSCs. Wedemonstrate that the different molecular properties lead to distinctdifferences in the film growth, material distribution, and device char-acteristics. The bifunctional molecular passivation significantly affectsthe film morphology and surface chemistries by their basicity andadsorption energy. Thus, the PZDI additive effectively quenches thesurface or bulk defects in the HP film resulting in a longer carrierlifetime and better interface quality. Consequently, the HPSCs withPZDI treatment resulted in an enhanced large area (>1cm2) deviceperformance from 19.68 to 23.17% (inverted configuration) with anincrease in the device parameters and reduced VOC deficit of 0.327 Vwith superior device stability under thermal and moisture stress.Density functional theory (DFT) calculations show that the alkyl coreamine has stronger bonding interaction with uncoordinated Pb2+ andiodine traps. This method is equally effective in improving deviceperformance in narrow and wide bandgap HPSC systems. The workgets insight into the device characteristics with synergetic effect in thesurface chemistry, elemental distribution, photophysics, and defectprofile.Results and discussionSurface passivation and film growth characterizationFigure 1a illustrates the chemical structures of diammonium iodidemolecules (DIMs) with aryl and alkyl cores, respectively, employed asbonding/charge-regulated molecular passivation. The optimizedstructures of the freemolecules, namely PEDAI (C6H8N2*2HI) and PZDI(C4H10N2*2HI), were obtained using Gaussian software at the B3LYP/def2TZVP level of theory. The bond lengths and Mulliken charges forcorresponding molecules are shown in the adjoining figure. Thesemolecules exhibit distinct characteristics in terms of bond length andMulliken charge distribution, which are crucial for effectively sup-pressing charge defects in the perovskite film. Notably, consideringthe charge cloud induced in iodine and nitrogen atoms within thePEDAI and PZDI molecules, the PZDI induces stronger adsorptionresulting in effective molecular interactions with uncoordinated Pbatoms and charge defects within the perovskite film. The DFT calcu-lation section provides a comprehensive discussion of the theoreticalaspects.To test the effectiveness of DIMs in a perovskite device, thesemolecules were introduced on the 3D-HP film as an interfacial pas-sivation layer (IPL) as shown in schematics (Fig. 1b, c). The surfacemorphology of MA/Br-free HP film (treated without and withPEDAI or PZDI) was studied with scanning electron microscope(SEM)measurement (Fig. 1d–f). One can notice a slight change in thegrain size with an overlayer surface grown on the film. The HP filmwith PEDAI passivation grows with unevenly distributed small crys-tallite. It suggests that the DIM with an alkyl core has beneficialsurface coverage compared to the aryl core. The film with PZDItreatment forms well-covered surface and grain boundarieswhich is propitious for the elimination of localized defects in per-ovskite film8,22.We collected X-ray diffraction (XRD) results of control andsurface-treated perovskite films to study the crystal growth. XRDresults of the PEDAI or PZDI-treated HP films with reference to thecontrol film are displayed in Fig. 1g and Fig. S1a, b. The PZDI-treatedHPfilms show a dominant (110) characteristic diffraction peak of the α-phase of FA- HP. It suppresses the δ-Cs/RbPbI3 phase and residual PbI2.Moreover, additionalpeaks at 2θ < 10o aredominantly grownon theHPfilms with a higher content of DIM indicating the evolution of a 2Dphase of PEDAI and PZDI interacting with PbI248. The characteristicXRD peak at lower 2θ on the HP film with PEDAI treatment grows withhigher intensity indicating higher tendency for the formation of 2Dphase compared to alkyl counterpart. This result is parallel to theArticle https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 2formation of low-dimensional perovskite on Sn-Pb 3D perovskitesthrough surface treatment with piperazine-based amine49.Figure 1h depicts the absorption spectra of the control and DIM-treated HP films, with varying concentrations (as shown in Fig. S1c, d).Notably, there is no significant differenceobserved at the band edgeofthe HP film as evidenced by the insets in Fig. 1h and Fig. S1c, d. The PLspectra (Fig. 1i) ofHPwith PZDIdemonstrated an intensifiedpeakwhilePEDAI suppressed it. The variation in PL intensity among the HP filmstreated with different concentrations of DIM (Fig. S2) aligns with animprovement in opto-physical quality. One can see a minimal shiftingof the PL characteristic peak (~817–818 nm). This suggests that theseadditives did not incorporate into the 3D-HP lattice which is alsoconsistent with XRD patterns.To gain deeper insights into the formation of the 2D phase, weprepared HP film by mixing DIM in the perovskite precursor solution,as depicted in Fig. S3a1-b1, and powder crystal by mixing (PbI2 andPEDAI or PEDAI). XRD results revealed that the HP films with DIMmixing and powder crystal grow without any remanence of the PbI2peak suggesting a strong tendency for the formation of a new crystalphase. Besides that, XRD patterns (Fig. S3a2-b2) demonstrated a moreintensified XRD peak at 2θ<10o for the HP film prepared with mixedDIMcompared to the surface treatment alone. Importantly, theHPfilmwith mixed PEDAI was found to be grown with much higher XRD peakintensity at ~4.8° indicating a more preference towards 2D phase for-mation. This observation was further supported by SEM images asdisplayed in Fig. S3a3-b3 which depicted nano-sheet-like features in theHP mixed containing PEDAI, underscoring the presence of PEDAI-based 2D phase. This observation is consistent with the small crystal-lite observed in the PEDAI passivated film. Interestingly, in contrast tothe PEDAI-mixed scenario, the HP film with mixed PZDI exhibitedgrowth as an overlayer on crystal grains or at grain boundaries, ratherthan forming a flake structure that closely resembled that of the pas-sivated film. To confirm the crystal structure of the 2D phasewith PZDIin HP film, single crystals were grown by mixing 2:1 and 1:1 molecularFig. 1 | Schematic of surface treatment and film growth. a Chemical structure,bond length, and Mulliken charges with summed H (iodine and nitrogen atoms)(PEDAI and PZDI). b Surface passivation of MA-free (FA0.84Rb0.04Cs0.12PbI3) 3D-HPfilm. c Illustration of interfacial interaction of diammonium iodide molecularpassivator and HP surface. d–f SEM images of HP films; without passivation, withPEDAI and PZDI passivation. g XRD patterns. h Absorption spectra. i PL spectrawithout and with IPL treatment on the HP films.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 3ratios of PZDI and PbI2 using the reported antisolvent vapor-assistedcrystallization method as described in Fig. S4. We obtained singlecrystal (2D phase) of composition (PZDI)2(PbI4)2 (orthorhombic)(Supplementary data 1, CCDC 2311444) and (PZDI)3Pb2I7 (monoclinic)(Supplementary data 2, CCDC 2311446) with DMSO complex con-firmed by single crystal XRD analysis. The results (SupplementaryFig. S5 and Table S1) show the simulated XRD patterns and corre-sponding crystal unit and packing fraction. Although the growthcondition is completely different from the film preparation, the XRDdata ofPZDI-containedHPfilmare found tobe close to the synthesizedsingle crystal data. This confirmed the formation of small crystallite onthe surface of HP film with PZDI treatment.Furthermore, the PL spectra of respective HP films (Fig. S3c) alsoshowcased distinct PL peaks at higher energy regimes suggesting theexistence of DIM-based 2D phases. Notably, the PL characteristic fea-tures of respective films are parallel to XRD patterns. These resultscollectively suggest the HP films with PEDAI tend to favor 2D phaseformation rather than forming a fully covered 3D-HP surface. It isbelieved that these characteristic differences could have a significanteffect on the device’s performance.Photovoltaics performance and photophysicsTo investigate the effect of surface treatment using DIM with aryl oralkyl core on the photovoltaic properties, we fabricated HPSCs withinverted device configuration as depicted in Fig. 2a. A typical cross-sectional image of a complete device is displayed in Fig. 2b. The cur-rent density-voltage (J-V) curves for the best HPSCs without and withPEDAI or PZDI treatment are shown in Fig. 2cwith a large device area of≈1 cm2 (Fig. S6) The J-V characteristics with varying concentrations ofDIM are given in Fig. S7. The deviceparameters have been summarizedin Table 1 (Tables S2–S3). The HPSCs with PZDI-treated HP (1 mg/ml)demonstrated champion PCE of ~23.17% with negligible J-V hysteresis.However, contrary to our expectation, the PCE of the PEDAI passivateddevice rolls off with reduced JSC and FF with reference to the controldevice as given in Table 1. A decrease of Jsc and an increase of Voc (bothsmall) seem to be typical behavior of slight passivation of interfacedefects. It suggests that PEDAI has an interface passivation effect tosome extent. These results corroborate that the amine inDIMwith aryland alkyl core greatly affects the device results. It is a consequence ofthe chemical interaction of the HPwith DIM that drives the film qualityimpacting morphology, surface chemistry, and defect profile. It isknown that PEDAI has a delocalized lone pair of electrons of nitrogenatomswhile that in PZDI is localized. The nitrogen site inDIMenhancesthe surface adhesion on the HP film which is favorable for defectmitigation. This could result in a stark characteristic difference inHPSCs. We will discuss more insight in the succeeding paragraphaccounting for these aspects.Figure 2d presents the PCE statistics of the control device and thatwith PEDAI or PZDI treatments (Figs. S8−10 and Tables S2–S3). ForPZDI, the device performance is found to be improvedwith an increasein all device parameters. However, the device with a higher con-centration of PZDI rolls off with a lower VOC and FF. In contrast, theHPSCs with PEDAI treatment are inferior to the control device and thedevice with a higher concentration of PEDAI further deteriorates bydropping VOC and FF like the PZDI case. These inferior device para-meters with higher PEDAI or PZDI content could be due to the accu-mulation of 2D HP phase on the surface unevenly. This observation isin line with other reports32,48.We validated HPSCs with MA/Br-free HP film with PZDI treatmentof PCE ≈21.47% (area ≈1.024 cm2) under standard conditions (accre-dited independent photovoltaic test laboratory, AIST PV Lab, Japan).The official certified data is given in Fig. S11. Our certified PCE of thechampion device has a record-level device efficiency for inverted p-i-nconfiguration of HPSCs with MA/Br-free HP for a large area of >1 cm2.For comparative evaluation, the few certified device reports with anarea >1 cm2 are tabulated (Table S4), where our champion HPSCcompares favorably amongst the reports.The external quantum efficiency (EQE) data for HPSCs withoutand with DIM passivation are shown in Fig. 2e. The EQE response forthe device with PZDI presents a better spectral response in theabsorption energy band range of the HP layer (λ > 450 nm) and inter-facial regime (450> λ > 330nm)55. It is attributed to the betterment inthe bulk and interface quality of PZDI-treated HPSC. Note that theintegrated current values from EQE spectra are 23.04, 22.68, and24.08mA/cm2 for the control and PEDAI or PZDI-treated HPSCs, whichare in the range of the JSC of respective devices. We also calculated thebandgap (Eg) of the HP absorber layer from EQE analysis (Eag~1.513,1.516, and 1.515 eV for the control, PEDAI, and PZDI; Fig. S12a–c). Theyare in close agreement with the Eg obtained from the absorptionspectra (Fig. S9d–f) and PL spectra (Fig. 1i).To explore the characteristic insight, we analyzed the photoresponse of the HPSCs. Figure 2f presents the VOC variation withlogarithmic of light intensity (ln(I)). The slopes are estimated to be1.40, 1.34, and 1.16 kBTq−1 for the control, PEDA, and PZDI-passivateddevices, respectively. A device with a higher slope signifies morecharge recombination at open circuit conditions. It suggests that theHPSCs with PZDI experience reduced trap-assisted recombinationwhich ameliorates the device performance. We recorded the TPVresponse as displayed in Fig. 2g by triggering VOCwith transient photoillumination. The TPV decay signal analysis reveals a longer carrierlifetime for the PZDI-treated device (12.34 µs). While the device withPEDAI (7.62 µs) shows a slight increase in carrier lifetime compared tothe control device (6.22 µs). It suggests that the PZDI treatment pas-sivates the defect in the HP film with propitious surface chemistry andmitigating defect.To understand the carrier lifetime, we measured the time-resolved photoluminescence (TRPL) spectra (Fig. 2h) and fitted themwith a bi-exponential decay equation48; I(t) = A0 +A1e�ðt�t0 Þτ1 + A2e�ðt�t0 Þτ2 ,whereA0 is a constant for the baseline offset, A1 and A2 are the relativeamplitude. The decay time, τ1 and τ2 accounts for the nonradiativerecombination at the interface and radiative recombination at the bulklayer56. The HP film with PZDI treatment shows a significantly longerlifetime (τ1 ≈331 ns and τ2 ≈2285ns) compared to that of the HP withPEDAI (τ1 ≈85 ns and τ2 ≈783 ns) or control (τ1 ≈98 ns and τ2 ≈678 ns).Interestingly, the PEDAI-treated film shows only a small difference incarrier lifetime compared to the control film. It corroborates thatPEDAI is not as effective as PZDI for the attenuation of a deleteriousdefect in the HP film. These results indicate the surface treatment withPZDI is propitious for defect passivation due to stronger localizednitrogen bonding in HP film and hence leads to the superiority ofdevice performance.Modulation of surface chemistry and interfaceTo understand surface energy, we measured ultraviolet photoelec-tron spectroscopy (UPS). The cutoff energy corresponding to thework function (ϕ) (Fig. 3a) and the onset energy (Ei) (Fig. 3b) cal-culated from the UPS results. The band structure has been con-structed by combining with optical bandgap and UPS result (Fig.S13). The values of ϕ and Ei are found to be slightly increased withDIM treatment. The results demonstrate a downshift of EV (by 0.287or 0.278 eV) and EC (0.283 or 0.276 eV) levels for PEDAI or PZDI-treated film. It modulates the band alignment at the HP/C60interface27,36. Indeed, the interfacial band alignment induced by DIMtreatment is beneficial for effective carrier transport resulting inbetter device performance. Although the PZDI or PEDAI shows asimilar effect in surface energy, there is a significant improvement indevice parameters for the device with PZDI treatment. It suggeststhat the surface energy modification by PEDAI has only a minimaleffect on device performance.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 4To explore surface chemistry, we carried out X-ray photoelectronspectroscopy (XPS) measurements. In the C 1s XPS core (Fig. 3c), thebinding energies centered at ~284.4, 286.2, and 287.4 eV are assignedto C-C/C=C, C-N-C, and N =C-N, respectively. The C-N-C characteristicpeak for PZDI-treated film indicates its dominant interaction with theHP surface. TheN 1sXPS cores for the corresponding film (Fig. 3d) alsoindicate the respective chemical binding characteristics. The surfacepassivated film shows a small shift of the XPS characteristic core of Pb4f and I 3d (~0.11 and 0.18 eV for PEDAI and 0.16 and 0.29 eV for PZDI;respectively, Fig. S14) towards higher binding energy. It indicates aFig. 2 | Photovoltaic characterization of the HPSCs with surface treatment.a Schematics of the device structure. b STEM-cross-sectional images of devices.c J-V characteristics of the control and IPL treatment (best IPL content; 1mg/ml);(open/filled symbols (forward/reverse) scan direction). d Statistical box for PCE oftheHPSCswithout andwith surface passivation (PEDAI or PZDI). These data consistof 50devices from6batches. e EQE spectra of devices. f Light intensity vs. VOC plot.g Transient photovoltage (TPV) decay curves of the respective devices. h TRPLdecay spectra for corresponding films.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 5Fig. 3 | Characterizations of surface chemistry and elemental distribution.a, b UPS spectra of the HP films; control, PEDAI, and PZDI treatment. c, d XPS-spectra analysis; C-1s core and N-1s core. e–g ToF-SIMS depth profiles of control,PEDAI, and PZDI passivated HP films. h–j Reconstructed 3Dmaps; the distributionsof passivated molecules in HP film. There are selected ionic species; ITO (In+), NiOx(Ni+), HP control (FAI+, PbI+, CsI+, RbI+) and with PEDAI (PEDA2+) or PZDI (PZD2+).Table 1 | Solar cell data and VOC deficit of the HPSCs with MA/Br-free perovskite (without and with surface treatment (PEDAIor PZDI))Condition Eag (eV) Scan JSC (mAcm−2) VOC (V) FF PCE (%) VOC deficit (Eag/q -VOC) (V)Control 1.513 F 23.62 1.104 0.717 18.69 0.402R 23.56 1.112 0.750 19.65PEDAI 1.517 F 22.98 1.142 0.703 18.46 0.381R 22.79 1.145 0.736 19.21PZDI 1.515 F 24.76 1.185 0.772 22.65 0.327R 24.78 1.188 0.787 23.17VOC deficit calculated with reference to Eag extracted from EQE data. The letters: F and R refer to the scan directions.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 6stronger ionic bonding induced on the film surface with PZDI treat-ment. The Cs 3d and Rb 3d core levels demonstrate almost similarspectral features indicating only a weak interaction with the passivat-ing molecule. This surface analysis implicates that the bifunctionalsurface passivator establishes stronger interaction with nitrogenbonding to uncoordinated Pb2+ or Iodine antistites48,57. The function-ality of the PZDI molecule is superior due to higher electron density inthe vicinity of the N-atom compared to PEDAI with an aryl core.Furthermore, we conducted a cross-sectional transmission elec-tron microscopy (TEM) measurement to investigate the interfacialstructure formed with the 3D HP host. A set of Fig. S15 illustrates thescanning transmission electronmicroscopy (STEM) images of both thecontrol and PEDAI or PZDI-treated HPSCs, revealing a noticeable fea-ture at the HP/C60 interface (Fig. S15b3-4 and S15c3-4). High-resolutionTEM (HR-TEM) images at the interface of the PEDAI or PZDI-treateddevice display a wide interplanar d-spacing (Fig. S15b4-c4), providingevidence for the presence of a 2D phase either at the interface orburied within the 3D HP host HP. One can see an unevenly distributed2D phase not only at the HP/C60 interface but also in the subsurfacebulk for PEDAI-treated devices which could be detrimental to devicequality. We found that both molecules, PEDAI and PZDI, modify thesurface or grain boundary chemistries and the interfacial band struc-ture byvirtueof theirmolecular functionalities. This finding alignswithprevious studies on interface passivation39,58.To explore the spatial distributions of molecular passivator, time-of-flight secondary ion mass spectrometry (ToF-SIMS)32,45,59 was usedto track the ionic distribution of PEDA2+ and PZD2+ in the HP. Thecharacteristic ionic species are shown in Fig. 3e–g. It shows an identicalionic distribution in the perovskite bulk. The characteristic signalsfrom PEDAI or PZDI are found to be significantly higher on the surfacewith a deep gradient to the bulk (Fig. S16). The 3D maps (Fig. 3h–j)demonstrate that the PEDA2+ and PZD2+ cations introduced by surfacetreatment are mainly distributed on the HP top surface. This obser-vation is analogous to the report by the Wakamiya group in ethyle-nediammonium iodide-treated Sn-PbmixedHP film45. Importantly, thePZD2+ additive shows uniform surface coverage with negligible bulkdiffusion. In contrast, the PEDA2+ ions are found to have uneven dis-tribution on the top surface with enriched bulk diffusion compared tothe PZDI case. These results are analogous to the surface feature ofSEM images of respective films. This observation corroborates that thePZDI or PEDAI mainly passivates the surface defect density by itsfunctional characteristics to modify the surface chemistry of thepristine perovskite layer.Moreover, we also delved into the analysis of residual stress in therespected devices following surface treatment by examination of XRDspectra using the 2θ− sin2ψ method60, as detailed in Fig. S17. Weselected the (012) plane at 31.6° as the focal point for analysis due to itsability to provide grain information and its diluting orientation effectin the linear relationship of 2θ- sin2ψ. The data illustrate how thescattering peaks progressively shift to the left as the ψ angle variesfrom 10 to 60° at different levels. The stress induced in perovskite canbe calculated by fitting the plots (2θ- sin2ψ). Notably, a negative slopeindicates the presence of tensile stress within the perovskite film. Alower negative slope was observed for PZDI treatment compared toPEDAI and control device which aligns with the other reports60,61.Moreover, theGIXRDanalysis confined to the (012)plane (Fig. S18)wasalso analyzed accounting for the penetration depth corresponding tothe angle of the grazing incident. It revealed a distinct effect with anincreasing depth profile suggesting the modification of the surface,grain boundaries, and the bulk of HP film through the surface treat-ment. This could have a crucial role in strain attenuation. It corrobo-rates that the surface treatment passivates the defect on the surface orat grain boundaries to some extent which results in lower strain in thefilm. Thus, it is suggested that the surface treatment using PZDI withstronger -NH functionality regulates the HP film by mitigating theresidual strain which is crucial in influencing carrier dynamics, deviceresults, and device stability as observed in this work.Effect of surface treatment on defectsTo get insight into the defect densities, we investigate the admittancespectroscopy of HPSCs with surface treatment. Mott-Schottky (M-S)plot and carrier profile (NCV) were extracted from capacitance-voltage(C-V) data62–65. M-S plots (Fig. 4a) exhibit fully depleted curves for V >diffusion potential (VD) suggesting intrinsic characteristic junction.One can see a slight hysteresis in theM-S curve near VD for the controldevicewhich almost disappeared for the surface-passivateddevice. It isattributed to the reduction in ionic polarization at the interface. The VDvalue for the PSCwith PZDI (1.148 V) is greater than PEDAI (1.065 V) andcontrol device (0.991 V) which is parallel to the VOC of the respectivedevice. The carrier profile (NCV) (Fig. 4b) extracted from C-V data ana-lysis comprises the free carrier and defect density. It showed a carrierdistribution in a bulk (NBCV ) in the range of ∼3.46–6.94 × 1015 cm−3. TheNBCV with PZDI is slightly lower by some fraction. The carrier profile atthe edge accounts for the interface defect density profile (NIFCV∼15.26 × 1017, ~11.56 × 1017, and ∼2.43 × 1017 cm−3 for control, PEDAI, andPZDI treated devices, respectively). The interface defect density issuppressed by 6 times for the PZDI-passivated device. It corroboratesthat the PZDI effectively attenuates the recombination centers leadingto the improvement in the device parameter.For the quantitative analysis of the defect profile, we investigatedthermal admittance spectroscopy, an effective technique for estimat-ing optoelectronic properties; the defect level and defect density tothin film solar cells HPSCs66,67, chalcogenide solar cells68, and organicsolar cells69). Figure 4c shows the capacitance-frequency (C-f) spectrameasured at room temperature (under dark). All device reveals a pla-teau regime (1 to 100 kHz) with a slightly lower value for PZDI treateddevice that could stem from the HP accounting for defect density.Besides that, the lower frequency capacitance response is much stee-per for the control device which is attributed to the interfacial chargeaccumulation or ionic polarization. It implicates a suppression ofinterfacial charge accumulation for the HPSCs with PZDI treatment.Furthermore, we measured the temperature-dependent capaci-tance-frequency (C-f-T) spectra to analyze the defect density profile.The trap state (Et) is calculated from Arrhenius plot70 by analyzing theresonance frequency (ωo) obtained from the C-f-T analysis as given inFig. S19. The Arrhenius plots (Fig. 4d–f) revealed shallower defectstates in the PZDI-treated device (Et3, E0t3 ∼0:154, 0:374eV) comparedto the PEDAI-treated (Et2, E0t2 ∼0:212, 0:408eV) or control device(E 0t1, E0t2 ∼0:241, 0:423eV). We calculated the defect density profiles(Fig. 4g–i) using the equation70,71, NtðEωÞ =� VDqWωkBTdCdω� �, where, VD,W,q, andω denote the diffusion potential, the space charge regionwidth,elementary charge, and applied frequency, respectively.We found that the integrated trap densities for the control device(Nt1, N0t1 ∼ 1.08× 1017, 9.89× 1016 cm−3) are attenuated for the PEDAI(Nt2,N0t2~7.38× 1016, 8.49× 1016 cm−3) or PZDI (Nt3,N0t3~3.22× 1016,4.03× 1016 cm−3) treated devices. These results are in the range ofreported trap densities for the perovskite film (1016−1019 cm−3)27,72. Onother the hand, the defect densities in our devices are more than 106than a single crystal (1010 cm−3) which calls for more effort to lower thetrap densities to achieve superior film quality. From the comparativeanalysis, the trap densities (N0t1) primarily assigned for defects in thebulk are decreased by ~2.5 times in the device with PZDI treatment or1.65 times in the device with PEDAI treatment indicating the improvedbulk quality of HP film. The shallower trap state profile (Nt1) is assumedto be defects at the surface or GBs in the HP film. These shallowerdefect densities are found to be significantly lowered in the PZDI orPEDAI-treated devices. These results consolidate that the PEDAI orPZDI stays on the film surface or diffuses into the bulk through the GBsto passivate the defect states due to molecular interaction with theArticle https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 7characteristics of nitrogen terminals. We found that the PZDI passi-vation is rather efficient for mitigating the defect chemistries at thesurface, GBs, and bulk compared to PEDAI. Thus, the capacitancespectra analysis well agrees with the advantageous properties of HPfilm induced with surface treatment as discussed in previousparagraphs.Theoretical insights on surface passivationTo elucidate the effects of the DIM passivator on the HP film, weconducted the first-principles calculations based on density functionaltheory (DFT). A detailed description of our computational and theo-retical methods can be found in the Supplementary Information. Thepseudo-cubic structure of FAPbI3 was used as a model of the bulkstructure (Fig. S20). The perovskite’s surface was modeled by a 2 x 2slabofPbI2-terminated surface (001)withfive PbI2 layers and a vacuumregion of ~25 Å (Fig. S21a). The corresponding total density of elec-tronic states (DOS) calculated for the defect-free PbI2 terminatedsurface is shown in Fig. S21b. In the case of full surface coverage, weaccommodated two PEDAI (PZDI) molecules (Figs. S22 and S23), withthe I atoms of PEDAI (PZDI) adsorbing atop the Pb atoms of the top-most PbI2 surface layer, as illustrated in Fig. S24a, b. Bothmolecules areassumed a tilted orientation, with their N atoms forming a plane par-allel to the perovskite surface. Our calculations revealed that PEDAImolecules exhibit a preference for forming a chain along the (100)direction, while PZDImolecules forma chain along the (010) direction.The interaction of PEDAI and PZDI with the surface induces a slightdistortion of the surface PbI2 layers and causes a rotation of the FAmolecules within the first and second PbI2-FAI bilayers. This effect isparticularly prominent in the case of PEDAI@FAPbI3 (see Fig. S24a). Itcould also affect the distribution of passivating molecules during thefilm formation.As one can see, adsorption of PEDAI or PZDI molecules does notintroduce any defect states in the forbidden zone of defect-free per-ovskite (DOS for PEDAI and PZDI; Fig. S24c, d), slightly increasing thebandgap from 1.53 eV calculated for the pure FAPbI3, up to 1.60 eV(1.59 eV) for the perovskite covered by PEDAI (PZDI) which is analo-gous to slightly higher surface band energy obtained from UPS ana-lysis. Both DIM molecules are attached strongly to the defect-freesurface, which protects the unstable surface without introducing anyadverse electronic effects. In contrast, on the defective surface, thesemolecules show remarkable electronic functions.Figure 5 shows the role of PEDAI/PZDI passivation of the defected(IPb antisite) PbI2-terminated surface of FAPbI3 defined in the previousstudies25,32. Indeed, the IPb antisite defect results in the formation of anunoccupied defect state 0.1 eV above the Fermi level as well as somedefect states in themiddle of the forbidden zone (0.4–1.0 eV above theFermi level), as it is seen from the analysis of the total DOS of theperovskite surfacewith IPb defect, presented in Fig. 5d, e by black lines.The detrimental defect states can be effectively passivated with PZDItreatment. Thus, in the case of PZDI passivation, the density of thedefect states is considerably reduced, slightly shifting down the bot-tomof the conductionbandby0.06 eV toward the Fermi level (Fig. 5e).Fig. 4 | Capacitance characteristics of HPSCs. a Mott–Schottky plots (open/filled symbols forward/reverse scan direction). b Carrier distribution calculated from C-Vdata. c C-f response. d–f Arrhenius plots. g–i Defects profile (Nt).Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 8On the other hand, the passivation of the IPb defect with PEDAI doesnot eliminate the low-lying defect state at 0.1 eV above the Fermi level,as well as introduces the narrow defect state at 1.66 eV, with the edgeof the conductivity band shifted to 1.78 eV. Therefore, from the ana-lysis of the electronic structure of the passivated surface with IPbantisite defect, one can suggest that PZDI passivation should lead toconsiderably better solar-cell performance. This observation is in linewith the calculated defect profile (Fig. 4).Since PZDI or PEDAI, both contain cations and anions, thesecould also interact with other charge defects. From the analysis ofthe Mulliken charges (Figs. S22 and S23), it is found that in PZDI, thecharge distribution in the tail is strongly polarized, with both of I inthe -NH2I anchor possessing an excess of the negative charge of−0.68|e|, while the NH2 counterpart is positively charged, with thenet charge of +0.34|e|. Here, |e| is an elementary charge. On the otherhand, in the case of the more stable trans-isomer the PEDAI moleculeone of the -NH3I anchor is polarized with the Mulliken charge on Iequal to −0.64|e| and charge on the NH3 counterpart of +0.49|e|,while another -NH3I anchor is overall almost neutral with a littlepolarization. In the case of the cis-isomer of PEDAI both of the -NH3Ianchors are overall almost neutral with a little polarization of chargesbetween I and N. This feature explains the difference in the interac-tion of PEDAI and PZDI molecules with the perovskite surface andtheir ability to quench the defects. Theoretical analysis of the changein Gibbs free energy upon adsorption demonstrates that PZDImolecules bind considerably stronger to the surface with IPb antisitedefect in comparison with PEDAI, with the binding energy 1.54 eV permolecule (1.32 eV for PEDAI). Thus, our theoretical analysis corro-borates that PZDI passivates the defective surface with a strongerquenching tendency forming the stable film covering the surface.The observed differences in the distribution tendencies of PEDAIand PZDI in the HP film, as seen in the ToF-SIMS results (Fig. 3h–j), canbe correlated with their distinct characteristics such as stark chargedifference, their preference for forming a chain (PEDAI/PZDI along-(100)/(010) direction) and surface binding energy (1.32/1.54 eV permolecule for PEDAI/PZDI). These characteristics could play a crucialrole in the distribution of PEDAI or PZDI in the HP film. The PEDAImolecule having (100) preferential direction and less binding energypromotes the surface coverage aswell as the penetration into the bulk,whichexplains the distinct variations observed in theToF-SIMS results.Thus, penetration of PEDAI into the perovskite film tends to reduceconductivity, and therefore, one might expect a decrease in deviceperformance. However, since surface passivation is still effective(Fig. 5d), an improvement in the VOC is achieved. This explains theobserved characteristics of low JSC, low FF, but high Voc, along withenhanced stability compared to the control device. Therefore, theoverarching strategy here is to find a species that stays primarily on thesurface and penetrates into the bulk only through defects on grainboundaries, without significantly compromising crystallinity. Having asimilar molecular structure to PEDAI and PZDI, one might anticipate aFig. 5 | DFT calculations of defect passivation. a Optimized structures of IPbantisite defect. b, c PEDAI and PZDI adsorbed on the PbI2-terminated surface ofFAPbI3 with IPb antisite defect. d, e total DOS calculated for the PbI2-terminatedsurfaceof FAPbI3with IPb antisite defectpassivatedwith PEDAI and PZDI. Black linescorrespond to the total DOS calculated for the unpassivated perovskite surfacewith the IPb defect.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 9similar effect. However, we stress that their structural integrity whenthey are placed at the actual interface matters.Furthermore, to confirm experimentally the effectiveness ofmolecular passivation using PZDI, we employed the surface treatmentmethod in perovskite derivatives with varying optical Eg. The detailedfabrication method for the devices has been explained in the experi-mental section. Regarding the wide bandgap HP (Eg~1.72 eV, WB-HP;FA0.84Cs0.12Rb0.04Pb(I0.63Br0.37)3), we observed that the application ofPZDI treatment led to a significant improvement in the PCE, increasingit from 17.53 to 19.42% with a notable reduction in the VOC deficit from0.448 to 0.404 V (Table 2). The J-V curves, statistics, and EQE spectraare given in Fig. S25 and Table S5. We attribute the reduction in VOCdeficit to the mitigation of non-radiative recombination and anenhancement in the quality of the interface achieved through PZDIpassivation39.Similarly, for the narrow bandgap HP (Eg~1.26 eV, NB-HP;FA0.85MA0.1Cs0.05(Pb0.5Sn0.5)I3) -based HPSCs, the device results of con-trol and surface-treated NB-HPSCs are given in Fig. S26 and Table S6. TheNB-HPSCs with PZDI passivation demonstrated significant improvementin PCEof 20.32%with device parameters; JSC ~31.46mAcm−2,VOC ~0.860V,and FF ~0.751. While the control device has a PCE of 16.85% (JSC~30.72mAcm−2, VOC ~0.785V, and FF ~0.730). The device efficiency of NB-HPSCs using Sn/Pb binary HP materials with reduced MA is in the com-petitive range of another report73. The VOC deficit of NB-HPSCs is sig-nificantly lowered from 0.478 to 0.405V (Table 2) which is attributed toattenuation of surface or bulk recombination. This result further confirmsthe effectiveness of PZDI treatment for NB-HPSCs, consolidating bothexperimental results and theoretical observations. Importantly, thisreport corroborates the universality of surface treatment using bifunc-tional diammonium molecules for the enhancement of device PCE andstability by mitigating detrimental defects by modifying the surface andbulk defect chemistry of perovskite film. This observation is parallel toother reports of similar molecular derivatives45,48–50,74. While there is stillconsiderable room for improvement, it is noteworthy that, to the best ofour knowledge, the performance metrics of these devices fall within therange of the highest reported PCE for HPSCs using different perovskiteband derivatives.Operational stability and monitoring of HPSC degradationDespite continuous breakthroughs in device efficiency, the stability ofHPSCs is still a stumbling block to their competitive reliability inpractical applications. To evaluate the device stability, we tracked thedevice parameters of HPSCs (encapsulated) at the maximum powerpoint tracking (MPPT) conditions under 1 sun irradiation under heat,light, or humidity stress. The device stability data under different agingstress conditions (ISOS-L-2, ISOS-L-3, procedure)75 were recorded(Fig. 6 and Figs. S27–28 and Tables S7–8). As shown in Fig. 6a, b, thedevice with PZDI treatment demonstrated superior operational devicestability under respectivemonitoring conditions. Interestingly, despitethe ineffectiveness in improving device performance, the PEDAI-treated device showed better device stability compared to the controldevice suggesting the beneficial effect of surface passivation. At ele-vated temperature (~60± 5 °C; ~35–40% RH), the performance of thecontrol device dropped to ~57.82% of initial PCE in 1000h whichsignificantly lowered to ~37.80% in 200h under heat and moisturestress (T = 35± 5 °C; RH~ 60–65%). Similarly, HPSCs with PZDI treat-ment retained ~89.48% and 86.74% of original PCE under respectiveaging conditions. While the PEDAI treated device demonstrated com-paratively better device stability than the control device retaining74.32% and 72.67% of the original PCE under respective aging condi-tions. These data corroborate that the surface treatment with DIMmultifunctional molecules significantly improves the device stabilityunder thermal and humidity stress as a consequence of propitioussurface chemistry and interfacial surface modulation with strongadsorption energy31. This observation indicates that the superiordevice stability stems from better interface quality and moisture sta-bility in the surface-treated HPSCs.To consolidate the superior moisture stability data, the watercontact angles were taken to study the hydrophobicity of respectiveHP films (Fig. 6c–e). We noticed a significant drop in water contactangle from 66.42° (to~ 0 s) to 42.36° (t~ 1min) for the control film. ThePEDAI-treated HP film shows a higher contact angle of 84.50° (to~ 0 s)which retains at 78.40° after 1min. Similarly, the PZDI-treated HP filmdemonstrates a contact angle of 90.60° (to~ 0 s) to 86.20° (t~ 1min). Itcorroborates that the HP films with surface treatment result in excel-lent moisture tolerance. Themoisture resistivity of the passivated filmis attributed to its dense distribution on the film surface (Fig. 3h–j),which agrees with the trend of device stability under higher humiditystress.Moreover, to contemplate the interfacial deterioration underaging conditions, the capacitance-voltage curves of aged devices weremeasured. Amore pronounced C-V hysteresis was seen for the controlHPSC (Fig. 6f–h) suggesting a deteriorated interface compared to theHPSCs with surface passivation. This observation substantiates thatthe control HPSC degrades due to the corrosion of the interfacialjunction and increasing dominance of accumulated ions at theinterface76. A sharp transition of theM–S curve in the device with PZDItreatment indicates a smaller depletion layer capacitance (Cdl) that isattributed to low interfacial defect density. It retains a more stableinterfacial junction that stems from intact bulk capacitance (Cg)77,78One can see a plateau capacitance (Cs) region for V>VD which is cor-related to interfacial charge accumulation and electrode polarization.A suppressed C-V hysteresis in PZDI-treated HPSC signifies the sup-pression of ionic motion or interfacial charge accumulation inducedwith scan directions65,79,80. This result is in line with an earlier report oninterfacial degradation analysis7. Interestingly, the M-S characteristicfeatures for the aged PEDAI device are not as intact as the PZDI-treateddevice. These characteristic disparities indicate that alky amine israther meticulous for surface passivation as supported by theoreticalcalculations. Thus, this work corroborates that localized electrondensity in alkyl amine enhances the interfacial adhesion stabilizing theinterface and bulk that is benign for device efficiency and operationalstability.In summary, we demonstrated a very effective approach to miti-gate the defects in MA-free perovskites using charge-modulatedbifunctional molecules with amine terminal. It is found that the dia-mmonium iodide with aryl or alkyl core amine has a remarkable effectonboth the efficiency and stability of invertedHPSCs. The PEDAIwith aTable 2 | Photovoltaic parameters HPSCs (control and target; with PZDI treatment) using perovskite absorber with wide andnarrow bandgap (WB, NB)Perovskite system Device Eag (eV) JSC (mAcm−2) VOC (V) FF PCE (%) VOC deficit (V)MA-free- WG-HP Control 1.722 18.80 1.274 0.732 17.53 (16.76 ± 0.69) 0.448Target 1.725 19.32 1.321 0.761 19.42 (18.46 ±0.42) 0.404NG-HP Control 1.263 31.15 0.785 0.730 17.85 (16.96 ±0.77) 0.478Target 1.265 31.46 0.860 0.751 20.32 (19.17 ±0.56) 0.405The bandgap (Eag) is estimated from EQE analysis. VOC deficit calculated with reference to Eag . The device statistics are given in parentheses; average values of PCE and standard deviation.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 10delocalized lone pair of electrons with an aryl core exerts weak inter-action on the surface defects. The PZDI molecules with alkyl core arerather propitious for the improvement in film quality inducing a sur-face chemical environment for quenching the surface and bulk defectdue to stronger interaction with localized electron density. Conse-quently, the PCE of HPSCs with PZDI treatment has been boosted from19.68% to 23.17% with a large area of ≥ 1 cm2 and a reduced VOC deficitof 0.327 V. The molecular distribution of surface-treated film reveals auniform surface coverage of PZDI with scant diffusion through thebulk layer from ToF-SIMS mapping. Moreover, the PZDI treated-perovskite film exhibited superior operational device stability underheat and humidity stress compared to PEDAI or control devices.This report corroborates that the amine with an alkyl core is crucialfor enabling molecular interaction for defect passivation. Hence itpaves a new way for designing molecular passivators with charge-regulated molecular bonding for HPSCs and other optoelectronicapplications.MethodsMaterials and precursor solutionAll chemicals were purchased from commercial suppliers as men-tioned and unless otherwise specified, they were used as received.Formamidinium iodide (FAI, GreatCells), 1,4-phenylenediamine dihy-driodide (PEDAI, TCI, >98.5%), piperazine dihydriodide (PZDI, TCI,>98.5%). PbI2 (Wako Chemicals, >98.5%), [2-(3,6-Dimethoxy-9H-carba-zol-9-yl) ethyl] phosphonic Acid (MeO-2PACz) (TCI), C60 (TCI, 99%),and Bathocuproine (BCP) (Sigma Aldrich, 99% purity) were purchasedand used as received. We used the NiOx target (purity >99.9%) fromKojundo Chemical Laboratory Co. Ltd, Japan.Fabrication of MA-free HP; FA0.84Cs0.12Rb0.04PbI3MA-free halide perovskite: the precursor solution (1.05 M) was pre-pared by dissolving FAI (0.84M), CsI (0.12M), RbI (0.04M), PbI2 (1M),and 5-AVAI (1 mM) in the mixture of dimethylformamide (DMF), anddimethyl sulfoxide (DMSO) (4:1) solvent for 2 h at 60 °C temperature.Forfilmdeposition, the precursorwas spin-coated at 1000 rpm for 10 s(ramping slope 2 s) and 5000 rpm for 40 s followed by dripping 800 μlof CB at the 35th second of the second step. Then to promote thecrystallization, those as-grown films were simply placed on a hot plateat 60 °C for 1min with further annealing at 100 °C for 45min For thesurface treatment strategy, the PEDAI or PZDI precursor solutions ofvarious concentrations (0.5, 1, 2mg/ml) wereprepared by dissolving inisopropyl alcohol (IPA) at 60 °C for 2 h. For surface passivation, thePEDAI or PZDI solutions were spin-coated onto the HP film atFig. 6 | Operational stability of PSCs, hydrophobicity of film, and capacitanceresponse of devices. a, b Device stability monitoring under MPPT conditions:T = 60 ± 5 °C; 30–35% RH (ISOS-L-2, procedure) and T = 35 ± 5 °C; RH~ 60–65%(ISOS-L-3, procedure). c–e Images of the water contact angle on the surface ofcontrol, PEDAI, and PZDI perovskite films at different water loading times (initial(0min) and after 1min). f–hM–S plots (open/filled symbols: reverse scan direction)of aged HSPCs (T= 60 ± 5 °C; RH~ 30–35%; 1000h). The color-shaded regionrepresents the characteristic capacitance regime. Overlapped shaded regionsrepresent the lagging of the M-S curve induced by interfacial deterioration.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 115000 rpm for 40 s (ramping slope 3 s) and annealed at 100 °Cfor 5min.All the solutions were filtered using 0.2 µm syringe filters justbefore the deposition to avoid the risk of unwanted particles in theprecursor solution.Perovskite solar cell FabricationSolar cell devices were fabricated on pre-cleaned patterned indium tinoxide (ITO) coated glass substrates (15Ω square–1). The ITO substrates(4.5 cm ×3.5 cm) were pre-cleaned in an ultrasonic bath with deter-gent, pure water, and 2-propanol, followed by an ultraviolet-ozonetreatment for 5min to remove theorganic residuals. TheNiOx (~20nm)film was deposited by sputtering as mentioned in our earlier reports7.In brief, the pre-cleaned ITO substrates were loaded in the depositionchamber and evacuated until <2× 10−3 Pa then pure argon gas wasintroduced at the rate of 20 sccm. The NiOx deposition was carried outin an argon gas pressure of 3.5 Pa and rf power supply of 150 W for7min at room temperature. Then substrates were transferred into anitrogen-filled glove box (<1.0ppm O2 and H2O) and the rest of thesteps were carried out inside the glove box. The sputtered NiOx thinfilm was treated with MeO-2PACz (0.1 weight% in ethanol) by spincoating at 5000 rpm for 50 s and subsequently dried at 100 °C for10min on a hot plate. For surface treatment, the PEDAI or PZDI pre-cursorsdissolved in IPAwere spin-coatedonto theHPfilm at 5000 rpmfor 40 s (ramping slope 3sec) and annealed at 100 °C for 5min The C60(26-28 nm at 0.1Å/s) layer was deposited by thermal evaporation as anelectron transport layer (ETL). After that, the electron selective layer(ESL), BCP (5–6 nm, at0.01 Å/s)was deposited by evaporation at a basepressure of ~2 × 10−4 Pa. Then, to complete the device structure, sam-ples were then transferred into the evaporation chamber connected tothe glove box for metal contact deposition. Finally, 140 nm of Ag wasthermally evaporated at a pressure <10−4 Pa. Four devices in ITO sub-strate (4.5 cm ×3.5 cm) with an active device area of ~1.26 × 1.26 cm2were sealed using UV-curable resins before the subsequent measure-ments in ambient conditions.Fabrication of other perovskites system devicesFabrication of wide-bandgap (WB) halide perovskite (WB-HPSCs)Preparation of MA-free WB-HP; FA0.84Cs0.12Rb0.04Pb(I0.63Br0.37)3.Wide bandgap halide perovskite (WB-HP): the precursor solution (1 M)was prepared by dissolving FAI (0.84M), CsI (0.12M), RbI (0.04M), PbI2(0.45M), PbBr2 (0.55M), and 5-AVAI (1mM) in 1 ml of mixture of DMFand DMSO (4:1) solvent for 2 h at room temperature. Note that weadopted cations composition as in regular MA-free HPSCs while thehalide composition was taken with the reference of MA-free WB-HPreported by the Snaith group81 and our previous report71. For filmdeposition, theprecursorwas spin-coated at 1000 rpm for 10 s (rampingslope 2 s) and 5000 rpm for 40 s followed by dripping 800 μl of CB atthe 42th second of the second step. Then to promote the crystallization,those as-grownfilmswere simplyplacedon ahotplate at 60 °C for 2minand at 100 °C for 45min The surface passivated devices were preparedby spin coating PZDI same as regular bandgapHP devices. the ETL layer;C60-fused N-methylpyrrolidine-meta-dodecyl phenyl (C60MC12) wasused as reported in our earlier report for WB-HHPSC71.Finally, we fabricated WB-HP-based HPSCs of inverted configura-tions; ITO/NiOx/MeO-2PACz/WB-HP/C60MC12/BCP/Ag.Fabrication of narrow bandgap (NB) halide perovskite (NB-HPSCs)Preparation of NW-HP; FA0.85MA0.1Cs0.05(Pb0.5, Sn0.5)I3. Narrowbandgap halide perovskite (NB-HP): the precursor solution (1.4M) wasprepared by dissolving FAI (0.85M), MAI (0.1M), CsI (0.05M), PbI2(0.5M), SnI2 (0.5M), SnF2 (0.05M), and MASCN (0.05M) in the mix-ture of DMF and DMSO (4:1) solvent for 1 h at 50 °C temperature. Notethat the NB-HP; Cs0.05MA0.1FA0.85(Pb0.5, Sn0.5)I3; composition wasinspired froma two-stepprocess reportbyYanfa Yanand co-workers73.We have used a single precursor solution for the preparation of NB-HPfilm. For film deposition, the precursor was spin-coated at 5000 rpmfor 50 s (ramping slope 3 s) followed by dripping 150 μl of CB at the20th second of the second step. Then, as deposited films were placedon a hot plate at 60 °C for 2min and at 100 °C for 10min For surfacetreatment with PZDI, the PZDI solution in IPAwas spin-coated onto theNW-HP film at 5000 rpm for 40 s (ramping slope 3 s) and annealed at100 °C for 5min.Finally, we completed NB-HP-based HPSCs of inverted config-urations; ITO/PEDOT:PSS/NB-HP/C60/BCP/Ag.Synthesis of PZDI-based single crystalsA single crystal was synthesized by the antisolvent vapor-assistedcrystallization method reported by Bakr and co-workers82. Two sets ofprecursors (5ml) were prepared by dissolving (1) PbI2 (1 M) + PZDI (1M) (1:1) and (2) PbI2 (0.5 M) + PZDI (1 M) (1:2) were dissolved in DMSOand stirred overnight. The precursor solution was filtered and trans-ferred into 10mLvials. The vialwith anopened lidwas thenplaced intoa sealed bottle filled with 3mL tetrahydrofuran (THF) as an anti-solvent. The single crystalsweregrownalongwith the slowdiffusion ofthe vapor of the anti-solvent THF into the precursor solution. Finally,single crystals were obtained after keeping the precursor withoutdisturbing it for 72 h. These single crystals with DMSO complex wereanalyzed by measuring single crystal XRD.Materials and device characterizationsIn NIMS Battery Research Platform facilities, X-ray diffraction (XRD)patterns of fabricated MA-free-HP films were collected using anadvanced X-ray diffractometer (Rigaku SmartLab, CuKα radiation,λ = 1.54050 Å). The stresses weremeasured according to the 2θ−sin2φmethod using Cu-Kα radiation in a RIGAKU SmartLab X-ray dif-fractometer and beam/parallel slit analyzer (PB/PSA) optics. The dif-fracted rays weremeasured at different angles (tilt angle,ψ = 10°–60°)by fixing a diffraction plane of (012) for e perovskite film. GrazingIncidence X-ray Diffraction (GIXRD) analysis was measured with theRigaku SmartLab diffractometer (Rigaku SmartLab, CuKα radiation,λ = 1.54050 Å) with grazing incident angle (ω−0.05°, 0.1°, 0.2°,……3°.).The single crystal XRD data was collected byXtaLAB Synergy-DW (Mo)and XtaLAB miniII (X-ray source-Mo Kα (λ =0.71073 Å, 50 kV–24mA,λ = 1.54056 Å) in Rigaku R & D facilities in Japan. X-ray photoelectronspectroscopy (XPS) spectra were obtained using a Versa Probe II(ULVAC-PHI, Japan). Perovskite film samples for XPS measurementswere prepared in an N2-filled glove box and transferred to theXHPSChamber through an N2-filled transfer vessel in order to avoidoxygen contamination. XPS with a nonmonochromatic source wasmeasured (Al Kα; 1486.6 eV, spot size 10–300μm) at a pass energy of187.85 eV (1.5 eV step size) for the survey scanandpass energy 46.95 eV(0.1 eV step size) for the fine scan with spot size 100μm. The XPSspectra were calibrated with the binding energy of 284.8 eV for C1s.In NIMS Namiki foundry research facilities and category-3 sharedfacilities, themorphologyoffilms and cross-sectional imageswere takenby a high-resolution scanning electron microscope (SEM) at 5 kV accel-erating voltage (Hitachi, S-4800). The photoluminescence (PL) spectrawere collected using the micro-PL spectrometer (HORIBA, LabRamHR-PL NF(UV-NIR)) ∼532 nm laser diode (10mWcm−2) as an excitationsource. The carrier lifetimes weremeasured with a fluorescence lifetimespectrometer (Quantaurus-τ from Hamamatsu-Photonics K.K., C11367)equipped with ∼405nm laser diode (typical peak power of 400mW) at200kHz repetition rate. The absorption spectra films were measuredusing the UV-Vis-NIR spectrometer (UV-2600i, Shimadzu). The absorp-tion spectra and photoluminescence (PL) spectra of various films weremeasured using the UV-Vis-NIR spectrometer (UV-2600i, Shimadzu).The band structure of the film was measured using Ultraviolet photo-electron spectroscopy (UPS, Thermo Fisher Scientific, Inc.) with a He Iline (21.22 eV) from a helium discharge lamp.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 12The samples for transmission electron microscopy (TEM) wereprepared by using a focused ion beam (FIB) technique (JEOL, JIB-4501)inside a glove box. Before the sample preparation with the FIB, wedeposited a thin layer of osmium and carbon on the top of the sample(Ag layer peeled off) to protect it from damage during milling. Sampleextraction was performed with an FIB accelerating voltage of 30 kVanda current of800pA.Once the lamellawas extracted andweldedontheMogrid, itwas thinned to electron transparency at 10 kVand 10 pA.The SEM accelerating voltage was kept at 5 kV for the entire process.The TEM samples were about 50–60nm thick. The finished TEM wasimmediately transferred for TEM analysis to limit overall exposure toair to <2min STEM/EDX was carried out at room temperature (~300K)using an analytical TEM (JEOL JEM-ARM200F for HR, a 200 kV accel-eration voltage) equipped with a cold-field emission gun and a JEOLEDX detector. To minimize the damage from the electron beam, welimited the exposure time to within 5 s in every TEM observation.Time-of-flight secondary ion mass spectrometry (ToF-SIMS)measurements were carried out using a ToF.SIMS 5 (ION-TOF GmbH)instrument equipped with a 60 keV and pA current Bi2+ beam foranalysis and a 10 keV and nA current Ar gas cluster ion beam (Ar-GCIB)for sputtering in non-interlaced mode to have minimal interfacialmixing. The sputtered area was 700× 700μm2 and the analysis areawas 100 × 100μm2.The current density–voltage (J–V) curves were measured at thescan rate of 0.05V/s under 1 sun with an AM1.5G spectral filter(100mW/cm2) coupled with an MPPT system (Systemhouse SunriseCorp.). Eachdevice’s efficiencywas confirmedunderMPPT for 2minoftracking. The light intensity was calibrated by a silicon (Si) diode (BS-520BK). For the stability test, the encapsulated devices weremeasuredatMPPTconditions. Thedeviceswerekept under 1-Sun intensity underMPPT conditions (elevated temperature, 60 ± 5 °C, relative humidity,30–35% RH and ~35± 5 °C; 60–65% RH, RT) during device stabilitymonitoring (Bunkoukeiki Corp. BIR-50 solar cell light resistance testsystem, incubator type 50× 50 mm irradiation, Systemhouse SunriseCorp.). The external quantum efficiency (EQE) spectra were obtainedusing a spectrometer (SM-250IQE, Bunkokeiki, Japan).The transient photovoltage was measured using a commercialPAIOS system (PAIOS V.4.3). A pulse intensity was used to induce aspike in photovoltage. The capacitance spectra (C-f) were taken fromPAIOS v. 4.3 software, which scans from 20Hz to 2MHz at 30mVAC inthe dark at a bias voltage of 0 V. The C–Vmeasurements were taken at10 kHz with a voltage amplitude of 30mV AC in the dark. The scanfrequency is determined from the plateau region (corresponds to Cg)of the capacitance-frequency spectra C-f scan at zero bias. The thermalcapacitance spectra (C-f-T) were measured using an LCR meter(IM3536, Hioki), with a voltage amplitude of 30mV under dark con-ditions in the temperature range of 253–353K. The temperature wasvaried by using the controlled chamber (SU-221) (±0.1 K).Solar cell certificationDevice certification was obtained from The National Institute ofAdvanced Industrial Science and Technology (AIST), Japan. It is regis-tered as ISO / IEC 17025 accreditation laboratory (IAJapan ASNITE 0021Calibration) according to international mutual recognition arrange-ments (MRA) for ILAC and APAC.Theoretical and computational methodsDensity functional theory (DFT) calculations have been performedusing the fully constrained meta-generalized-gradient approximation(meta-GGA) SCAN83 for the exchange−correlation functional and theprojector-augmented wave (PAW) formalism as implemented in theVienna ab initio simulationpackage (VASP)84,85. This approximation haspreviously demonstrated a remarkable accuracy for the description oflattice constants and weak interactions in a large diversity of bondedmolecules and materials83,86, which is especially important for theinvestigation of the complicated process of molecular adsorption andpassivation of perovskite surfaces. The pseudo-cubic structure of themodel FAPbI3 bulk was optimized using the regular Γ-centered6 × 6 × 6 k-point mesh for Brillouin zone sampling. The energy cutoffof 500 eV for the planewave basis set and the convergence criterion of10−6 eV for the self-consistent loop were employed. The optimizedlatticeparameters of the pseudo-cubic FAPbI3, a = 6.4782Å, b = 6.3080Å, c = 6.4012 Å are in goodagreementwith an experimentally observedcubic unit cell of dimension ∼6.36 Å87. The optimized lattice of bulkFAPbI3 was used to construct 2x2 slab of PbI2-terminated surface (001)with five PbI2 layers and a vacuum region of ~25 Å. Two top PbI2-FAIbilayers were fully relaxed, while the atoms in the bottom layers werefrozen. For the Brillouin zone sampling of the slab, the Γ-centered3 × 3 × 1 and 6 × 6 × 1 k-point meshes have been used for structuralrelaxations and DOS calculations, respectively.The adsorbed PEDAI and PZDImolecules have been optimized onthe perovskite surface, where all atoms in the molecules and two topPbI2-FAI bilayers were fully relaxed until forces were <0.01 eV Å−1. Thebinding energy of the molecule to the surface was evaluated from thechange in Gibbs free energy as Eb = � ðΔEel � TΔS+ΔEZPE Þ. HereΔEel = Epassivatedsurface � Esurface + Emol� �, where Epassivatedsurface is the totalenergy of the passivated surface, while Esurface and Emol are energies ofthe free surface and the isolated PEDAI or PZDI molecule, as followsfrom DFT calculations. T is the temperature of the system taken equalto 293 K, ΔS is the change of entropy upon molecular adsorption, andΔEZPE is the change in zero-point energy. To estimate the change inentropy, we have used the ideal gas approximation for the freemolecules as implemented inGaussian09 and considered S =0 for theadsorbed molecules, because they are immobilized on the surface,losing the translational and rotational degrees of freedom.Reporting summaryFurther information on research design is available in the NatureResearch Reporting Summary linked to this article.Data availabilityThe data that support the findings of this study are available from thecorresponding authors upon request.References1. Almora, O. et al. Device performance of emerging photovoltaicmaterials (Version 3). Adv. Energy Mater. 13, 2203313 (2023).2. Green, M. A. et al. Solar cell efficiency tables (version 62). Prog.Photovolt. Res. Appl. 31, 651–663 (2023).3. Kumar Jena, A., Kulkarni, A. & Miyasaka, T. Halide perovskite pho-tovoltaics: background, status, and future prospects. Chem. Rev.119, 3036–3103 (2019).4. Khadka, D. B. et al. Advancing efficiency and stability of lead, tin,and lead/tin perovskite solar cells: strategies and perspectives.Solar RRL 7, 2300535 (2023).5. Yang, T. et al. Achieving long‐term operational stability of per-ovskite solar cells with a stabilized efficiency exceeding 20% after1000 h. Adv. Sci. 6, 1900528 (2019).6. Khadka, D. B., Shirai, Y., Yanagida, M., Uto, K. & Miyano, K. Analysis ofdegradation kinetics of halide perovskite solar cells induced by lightand heat stress. Sol. Energy Mater. Sol. Cells 246, 111899 (2022).7. Khadka, D. B., Shirai, Y., Yanagida, M. & Miyano, K. Insights intoaccelerated degradation of perovskite solar cells under continuousillumination driven by thermal stress and interfacial junction. ACSAppl. Energy Mater. 4, 11121–11132 (2021).8. Liu, Y. et al. Ultrahydrophobic 3D/2D fluoroarene bilayer-basedwater-resistant perovskite solar cells with efficiencies exceeding22%. Sci. Adv. 5, eaaw2543 (2019).9. Bai, S. et al. Planar perovskite solar cells with long-term stabilityusing ionic liquid additives. Nature 571, 245–250 (2019).Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 1310. Saliba, M. et al. Cesium-containing triple cation perovskite solarcells: improved stability, reproducibility and high efficiency. EnergyEnviron. Sci. 9, 1989–1997 (2016).11. Peng, J. et al. Centimetre-scale perovskite solar cellswithfill factorsof more than 86 percent. Nature 601, 573–578 (2022).12. Lin, X. et al. In situgrowthof grapheneonboth sidesof aCu–Ni alloyelectrode for perovskite solar cells with improved stability. NatEnergy 7, 520–527 (2022).13. Peng, W. et al. Reducing nonradiative recombination in perovskitesolar cells with a porous insulator contact. Science 379, 683–690(2023).14. Bisquert, J. & Juarez-Perez, E. J. The causes of degradation of Per-ovskite solar cells. J. Phys. Chem. Lett. 10, 5889–5891 (2019).15. Wang, S., Jiang, Y., Juarez-Perez, E. J., Ono, L. K. &Qi, Y. Accelerateddegradation of methylammonium lead iodide perovskites inducedby exposure to iodine vapour. Nat. Energy 2, 16195 (2017).16. Conings, B. et al. Intrinsic thermal instability of methylammoniumlead trihalide perovskite. Adv. Energy Mater. 5, 1500477 (2015).17. Lee, J., Seol, D., Cho, A. & Park, N. High‐efficiency perovskite solarcells based on the black polymorph of HC(NH 2) 2 PbI 3. Adv. Mater.26, 4991–4998 (2014).18. Castriotta, L. A. et al. Air-processed infrared-annealed printedmethylammonium-free perovskite solar cells and modules incor-porating potassium-doped graphene oxide as an interlayer. ACSAppl. Mater. Interfaces 13, 11741–11754 (2021).19. Turren-Cruz, S.-H., Hagfeldt, A. & Saliba, M.Methylammonium-free,high-performance, and stable perovskite solar cells on a planararchitecture. Science 362, 449–453 (2018).20. Kim, G. et al. Impact of strain relaxation on performance ofa-formamidinium lead iodide perovskite solar cells. Science 370,108–112 (2020).21. Jin, H. et al. It’s a trap! On the nature of localised states and chargetrapping in lead halide perovskites.Mater. Horiz. 7, 397–410 (2020).22. Zhu, J. et al. Formamidine disulfide oxidant as a localised electronscavenger for >20% perovskite solar cell modules. Energy Environ.Sci. 14, 4903–4914 (2021).23. Motti, S. G. et al. Defect activity in lead halide perovskites. Adv.Mater. 31, 1901183 (2019).24. Moia, D. & Maier, J. Ion transport, defect chemistry, and the devicephysicsof hybridperovskite solar cells.ACSEnergyLett.6, 1566–1576(2021).25. Liu, N. & Yam, C. First-principles study of intrinsic defects in for-mamidinium lead triiodide perovskite solar cell absorbers. Phys.Chem. Chem. Phys. 20, 6800–6804 (2018).26. Wang, F. et al. Phenylalkylamine passivation of organolead halideperovskites enabling high-efficiency and air-stable photovoltaiccells. Adv. Mater. 28, 9986–9992 (2016).27. Zheng, X. et al. Defect passivation in hybrid perovskite solar cellsusing quaternary ammoniumhalide anions and cations.Nat. Energy2, 17102 (2017).28. Bi, D. et al. Multifunctional molecular modulators for perovskitesolar cells with over 20% efficiency and high operational stability.Nat. Commun 9, 4482 (2018).29. Li, T. et al. Multiple functional groups synergistically improve theperformance of inverted planar perovskite solar cells. Nano Energy82, 105742 (2021).30. Abate, A. et al. Supramolecular halogen bond passivation oforganic–inorganic halide perovskite solar cells. Nano Lett. 14,3247–3254 (2014).31. Ruiz-Preciado, M. A. et al. Supramolecular modulation of hybridperovskite solar cells via bifunctional halogen bonding revealed bytwo-dimensional 19 F solid-state NMR spectroscopy. J. Am. Chem.Soc. 142, 1645–1654 (2020).32. Khadka, D. B., Shirai, Y., Yanagida, M., Tadano, T. & Miyano, K.Interfacial embedding for high‐efficiency and stablemethylammonium‐free perovskite solar cells with fluoroarenehydrazine. Adv. Energy Mater. 12, 2202029 (2022).33. Zhu, X. et al. Ionic‐liquid‐perovskite capping layer for stable24.33%‐efficient solar cell. Adv. Energy Mater. 12, 2103491 (2022).34. Fu, C. et al. From structural design to functional construction:amine molecules in high‐performance formamidinium‐based per-ovskite solar cells. Angew. Chem. Int. Ed. 61, e2021170 (2022).35. Pham, N. D. et al. Tailoring crystal structure of FA 0.83 Cs 0.17 PbI 3perovskite through guanidiniumdoping for enhanced performanceand tunable hysteresis of planar perovskite solar cells. Adv. Funct.Mater. 29, 1806479 (2019).36. Gharibzadeh, S. et al. Two birds with one stone: dual grain-boundary and interface passivation enables >22% efficient invertedmethylammonium-free perovskite solar cells. Energy Environ. Sci.14, 5875–5893 (2021).37. Alanazi, A. Q. et al. Atomic-level microstructure of efficientformamidinium-based perovskite solar cells stabilized by5-ammonium valeric acid iodide revealed by multinuclear andtwo-dimensional solid-state NMR. J. Am. Chem. Soc. 141,17659–17669 (2019).38. Xu, W. et al. Rational molecular passivation for high-performanceperovskite light-emitting diodes.Nat. Photonics 13, 418–424 (2019).39. Yang,G. et al. Stableand low-photovoltage-lossperovskite solar cellsby multifunctional passivation. Nat. Photonics 15, 681–689 (2021).40. Lin, Y.-H. et al. A piperidinium salt stabilizes efficient metal-halideperovskite solar cells. Science 369, 96–102 (2020).41. Zhang, J.,Wu, S., Liu, T., Zhu, Z. & Jen, A. K. Y. Boosting photovoltaicperformance for leadhalide perovskites solar cellswith BF4 − anionsubstitutions. Adv. Funct. Mater. 29, 1808833 (2019).42. Pan, T. et al. Surface‐energy‐regulated growth of α‐PhaseCs0.03FA0.97PbI3 for highly efficient and stable inverted per-ovskite solar cells. Adv. Mater. 35, 2208522 (2023).43. Chen, S., Xiao, X., Gu, H. & Huang, J. Iodine reduction for repro-ducible and high-performance perovskite solar cells and modules.Sci. Adv. 7, eabe8130 (2021).44. Zhang, W. et al. Enhanced optoelectronic quality of perovskite thinfilms with hypophosphorous acid for planar heterojunction solarcells. Nat. Commun. 6, 10030 (2015).45. Hu, S. et al. Optimized carrier extraction at interfaces for 23.6%efficient tin–lead perovskite solar cells. Energy Environ. Sci. 15,2096–2107 (2022).46. Song, Q. et al. Bridging the buried interface with piperazine dihy-driodide layer for high performance inverted solar cells. Small 19,e2208260 (2023).47. Wang, H. et al. Skillfully deflecting the question: a small amount ofpiperazine-1,4-diium iodide radically enhances the thermal stabilityofCsPbI 3 perovskite. J.Mater. Chem.CMater. 7, 11757–11763 (2019).48. Hou, M. et al. Aryl diammonium iodide passivation for efficient andstable hybrid organ‐inorganic perovskite solar cells. Adv. Funct.Mater. 30, 2002366 (2020).49. Hu, S. et al. Synergistic surface modification of tin–lead perovskitesolar cells. Adv. Mater. 35, 2208320 (2023).50. Hu, S. et al. A universal surface treatment for p–i–n perovskite solarcells. ACS Appl. Mater. Interfaces 14, 56290–56297 (2022).51. Li, Z. et al. 24.64%‐efficiency ma‐free perovskite solar cell with Vocof 1.19 V enabled by a hinge‐typefluorine‐rich complex. Adv. Funct.Mater. 33, 2212606 (2023).52. Wang, Y. et al. Oxidation-resistant all-perovskite tandem solar cellsin substrate configuration. Nat. Commun. 14, 1819 (2023).53. Li, H. et al. 2D/3D heterojunction engineering at the buried interfacetowards high-performance inverted methylammonium-free per-ovskite solar cells. Nat. Energy 8, 946–955 (2023).54. Qiu, L., He, S.,Ono, L. K., Liu, S. &Qi, Y. Scalable fabrication ofmetalhalide perovskite solar cells and modules. ACS Energy Lett. 4,2147–2167 (2019).Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 1455. Nakane, A. et al. Quantitative determination of optical and recom-bination losses in thin-film photovoltaic devices based on externalquantum efficiency analysis. J. Appl. Phys. 120, 064505 (2016).56. Oliver, R. D. J. et al. Understanding and suppressing non-radiativelosses in methylammonium-free wide-bandgap perovskite solarcells. Energy Environ. Sci. 15, 714–726 (2022).57. Chen, J., Kim, S.-G., Ren, X., Jung, H. S. & Park, N.-G. Effect ofbidentate and tridentate additives on the photovoltaic performanceand stability of perovskite solar cells. J. Mater. Chem. A Mater. 7,4977–4987 (2019).58. Kim, D. et al. Efficient, stable silicon tandem cells enabled by anion-engineeredwide-bandgapperovskites.Science368, 155–160 (2020).59. Khadka, D. B., Shirai, Y., Yanagida, M., Tadano, T. & Miyano, K.Alleviating defect and oxidation in tin perovskite solar cells using abidentate ligand. Chem. Mater. 35, 4250–4258 (2023).60. Zhu, C. et al. Strain engineering in perovskite solar cells and itsimpacts on carrier dynamics. Nat. Commun. 10, 815 (2019).61. Li, F. et al. Regulating surface termination for efficient invertedperovskite solar cellswith greater than 23%efficiency. J. Am.Chem.Soc. 142, 20134–20142 (2020).62. Khadka, D. B., Shirai, Y., Yanagida, M., Masuda, T. & Miyano, K.Enhancement in efficiency and optoelectronic quality of perovskitethin films annealed in MACl vapor. Sustain Energy Fuels 1,755–766 (2017).63. Matsushita, A., Yanagida, M., Shirai, Y. & Miyano, K. Degradation ofperovskite solar cells by the doping level decrease of HTL revealedby capacitance spectroscopy. Sol. Energy Mater. Sol. Cells 220,110854 (2021).64. Khadka, D. B., Shirai, Y., Yanagida, M. & Miyano, K. Pseudohalidefunctional additives in tin halide perovskite for efficient andstable Pb-free perovskite solar cells. ACS Appl. Energy Mater. 4,12819–12826 (2021).65. Miyano, K., Yanagida, M. & Shirai, Y. Impedance spectroscopyrevisited. Adv. Energy Mater. 10, 1903097 (2020).66. Ye, S. et al. A breakthrough efficiency of 19.9% obtained in invertedperovskite solar cells by using an efficient trap state passivatorCu(thiourea)I. J. Am. Chem. Soc. 139, 7504–7512 (2017).67. Khadka, D. B., Shirai, Y., Yanagida, M. & Miyano, K. Attenuating thedefect activities with a rubidium additive for efficient and stable Sn-based halide perovskite solar cells. J. Mater Chem. C Mater. 8,2307–2313 (2020).68. Khadka, D. B., Kim, S. Y. & Kim, J. H. A nonvacuum approach forfabrication of Cu2ZnSnSe4/In2S3 thin film solar cell and optoe-lectronic characterization. J. Phys. Chem. C 119, 12226–12235(2015).69. Zhang, K. et al. Reducing limitations of aggregation‐induced pho-tocarrier trapping for photovoltaic stability via tailoring inter-molecular electron–phonon coupling in highly efficient quaternarypolymer solar cells. Adv. Energy Mater. 12, 2103371 (2022).70. Walter, T., Herberholz, R., Müller, C. & Schock, H. W. Determinationof defect distributions from admittance measurements and appli-cation to Cu(In,Ga)Se2 based heterojunctions. J. Appl. Phys. 80,4411–4420 (1996).71. Khadka, D. B., Shirai, Y., Yanagida,M., Noda, T. &Miyano, K. Tailoringthe open-circuit voltage deficit of wide-band-gap perovskite solarcells using alkyl chain-substituted fullerene derivatives. ACS Appl.Mater. Interfaces 10, 22074–22082 (2018).72. Saidaminov, M. I. et al. High-quality bulk hybrid perovskite singlecrystals within minutes by inverse temperature crystallization. Nat.Commun. 6, 7586 (2015).73. Li, C. et al. Low-bandgap mixed tin–lead iodide perovskites withreducedmethylammonium for simultaneous enhancement of solarcell efficiency and stability. Nat. Energy 5, 768–776 (2020).74. Hu, S. et al. Mixed lead–tin perovskite films with >7 μs chargecarrier lifetimes realized by maltol post-treatment. Chem. Sci. 12,13513–13519 (2021).75. Khenkin, M. V. et al. Consensus statement for stability assessmentand reporting for perovskite photovoltaics based on ISOS proce-dures. Nat. Energy 5, 35–49 (2020).76. Lopez-Varo, P. et al. Device physics of hybrid perovskite solar cells:theory and experiment. Adv. Energy Mater. 8, 1702772 (2018).77. Tsai, H. et al. High-efficiency two-dimensional Ruddlesden–Popperperovskite solar cells. Nature 536, 312–316 (2016).78. Zhang, X. et al. Inorganic CsPbI3 perovskite coating on PbS quan-tumdot for highly efficient and stable infrared light converting solarcells. Adv. Energy Mater. 8, 1702049 (2018).79. Chen, B., Yang, M., Priya, S. & Zhu, K. Origin of J-V hysteresis inperovskite solar cells. J. Phys. Chem. Lett. 7, 905–917 (2016).80. Yanagida, M., Shirai, Y., Khadka, D. B. & Miyano, K. Photoinducedion-redistribution in CH3NH3PbI3 perovskite solar cells. Phys.Chem. Chem. Phys. 22, 25118–25125 (2020).81. McMeekin, D. P. et al. A mixed-cation lead mixed-halide perovskiteabsorber for tandem solar cells. Science 351, 151–155 (2016).82. Shi, D. et al. Low trap-state density and long carrier diffusion inorganolead trihalide perovskite single crystals. Science 347,519–522 (2015).83. Sun, J., Ruzsinszky, A. & Perdew, J. P. Strongly constrained andappropriately normed semilocal density functional. Phys. Rev. Lett.115, 036402 (2015).84. Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initiototal-energy calculations using a plane-wave basis set. Phys. Rev. B54, 11169–11186 (1996).85. Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the pro-jector augmented-wave method. Phys. Rev. B 59, 1758–1775 (1999).86. Sun, J. et al. Accurate first-principles structures and energies ofdiversely bonded systems from an efficient density functional. Nat.Chem. 8, 831–836 (2016).87. Weller, M. T., Weber, O. J., Frost, J. M. & Walsh, A. Cubic perovskitestructure of black formamidinium lead iodide, α-[HC(NH 2) 2]PbI 3,at 298 K. J. Phys. Chem. Lett. 6, 3209–3212 (2015).AcknowledgementsThe authors are grateful to Yamaguchi Kazuo-San (XPS), Nariaki Sato-San (ToF-SIMS), andKeisukeShinoda-san/MakikoOshida-san (TEM) fromthe NIMS battery research platform for the technical support forrespective measurement and analysis. We are thankful to Rigaku R & Dfacilities in Japan for carrying out single crystal XRD measurements.Calculations were performed using computational resources of theInstitute for Solid State Physics, the University of Tokyo, Japan; and theResearchCenter forComputationalScience,Okazaki, Japan (Project: 23-IMS-C016).Author contributionsD.B.K. conceived the idea, designed, and performed the device fabri-cation work. Y.S. and M.Y. optimized the hole transport layer and vali-dated device results. M.Y. and D.B.K. performed UPS measurement anddata analysis. D.B.K. with technical staff collected STEM, ToF-SIMS, andXPS data and performed data analysis. H.O. with D.B.K. performed andanalyzed XRD measurements. D.B.K. carried out device characteriza-tions and analysis. A.L. and T.T. provided theoretical support. K.M.supervised the analysis of material characterizations and device analy-sis. D.B.K. wrote the original manuscript. All authors discussed theresults and reviewed the manuscript.Competing interestsThe authors declare no competing interests.Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 15Additional informationSupplementary information The online version containssupplementary material available athttps://doi.org/10.1038/s41467-024-45228-9.Correspondence and requests for materials should be addressed toDhruba B. Khadka, Yasuhiro Shirai or Andrey Lyalin.Peer review information Nature Communications thanks the anon-ymous reviewer(s) for their contribution to thepeer reviewof thiswork. Apeer review file is available.Reprints and permissions information is available athttp://www.nature.com/reprintsPublisher’s note Springer Nature remains neutral with regard to jur-isdictional claims in published maps and institutional affiliations.Open Access This article is licensed under a Creative CommonsAttribution 4.0 International License, which permits use, sharing,adaptation, distribution and reproduction in any medium or format, aslong as you give appropriate credit to the original author(s) and thesource, provide a link to the Creative Commons license, and indicate ifchanges were made. The images or other third party material in thisarticle are included in the article’s Creative Commons license, unlessindicated otherwise in a credit line to the material. If material is notincluded in the article’s Creative Commons license and your intendeduse is not permitted by statutory regulation or exceeds the permitteduse, you will need to obtain permission directly from the copyrightholder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.© The Author(s) 2024Article https://doi.org/10.1038/s41467-024-45228-9Nature Communications |          (2024) 15:882 16https://doi.org/10.1038/s41467-024-45228-9http://www.nature.com/reprintshttp://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/ Defect passivation in methylammonium/bromine free inverted perovskite solar cells using charge-modulated molecular bonding Results and discussion Surface passivation and film growth characterization Photovoltaics performance and photophysics Modulation of surface chemistry and interface Effect of surface treatment on defects Theoretical insights on surface passivation Operational stability and monitoring of HPSC degradation Methods Materials and precursor solution Fabrication of MA-free HP; FA0.84Cs0.12Rb0.04PbI3 Perovskite solar cell Fabrication Fabrication of other perovskites system devices Fabrication of wide-bandgap (WB) halide perovskite (WB-HPSCs) Preparation of MA-free WB-HP; FA0.84Cs0.12Rb0.04Pb(I0.63Br0.37)3 Fabrication of narrow bandgap (NB) halide perovskite (NB-HPSCs) Preparation of NW-HP; FA0.85MA0.1Cs0.05(Pb0.5, Sn0.5)I3 Synthesis of PZDI-based single crystals Materials and device characterizations Solar cell certification Theoretical and computational methods Reporting summary Data availability References Acknowledgements Author contributions Competing interests Additional information