WO2023172645A2 - Modulating heterointerface energetics for operationally stable perovskite solar cells - Google Patents

Abstract

A photovoltaic device includes a first electrode; a second electrode spaced apart from the first electrode; a perovskite layer between the first and second electrodes, the perovskite layer including a semiconducting halide perovskite photoactive material; and a vacuum-level modulating layer formed on a surface of the perovskite layer. The vacuum-level modulating layer provides crystal defect passivation of the surface of the perovskite layer, and the vacuum-level modulating layer further modulates surface energetics of the surface of the perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein.

Description

MODULATING HETEROINTERFACE ENERGETICS FOR OPERATIONALLY STABLE PEROVSKITE SOLAR CELLS CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This patent application claims priority benefit from U.S. provisional patent application no.63/317,777, filed on March 8, 2022, the entire content of which is incorporated herein by reference. STATEMENT OF GOVERNMENT INTEREST [0002] This invention was made with government support under Grant Number DE-EE0008751, awarded by the Department of Energy. The government has certain rights in the invention. BACKGROUND 1. Technical Field [0003] The currently claimed embodiments of the present invention relate to photovoltaic devices and methods, and more particularly to operationally stable perovskite solar cells. 2. Discussion of Related Art [0004] The instability problems of perovskite solar cells (PSCs) are the most pressing problems hindering market commercialization of PSC technology. PSCs incorporate a surface treatment step during the fabrication process for defect passivation purposes. Most successful examples use alkyl- or aryl-ammonium iodide salts, which have been used in all the efficiency-record-breaking PSCs since 2019. To apply such surface treatments onto the perovskite, the target ammonium iodide salt is dissolved in isopropyl alcohol or chloroform (at 10-20 mM concentration), and then applied to the perovskite surface via dynamic spin-coating. Despite the beneficial defect passivation effects of surface treatment, surface treatments cause a negative vacuum level change (ΔE_vac) at the heterointerface. The negative ΔE_vac detrimentally accumulates charges in a potential well, which accelerates ion migration and thus PSC degradation during operation. [0005] Consequently, there remains a need for improved operationally stable perovskite solar cells and methods of production. SUMMARY [0006] A photovoltaic device according to an embodiment of the current invention includes a first electrode; a second electrode spaced apart from the first electrode; a perovskite layer between the first and second electrodes, the perovskite layer including a semiconducting halide perovskite photoactive material; and a vacuum-level modulating layer formed on a surface of the perovskite layer. The vacuum-level modulating layer provides crystal defect passivation of the surface of the perovskite layer, and the vacuum-level modulating layer further modulates surface energetics of the surface of the perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein. [0007] A method of producing a photovoltaic device according to an embodiment of the current invention includes forming a first electrode; forming a second electrode spaced apart from the first electrode; forming a perovskite layer between the first and second electrodes, the perovskite layer comprising a semiconducting halide perovskite photoactive material; and forming a vacuum-level modulating layer on a surface of the perovskite layer. The vacuum-level modulating layer provides crystal defect passivation of the surface of the perovskite layer, and the vacuum-level modulating layer further modulates the surface energetics of the surface of the perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein. [0008] A method of treating a surface of a perovskite layer for use in a photovoltaic device according to an embodiment of the current invention includes receiving a material for forming a vacuum-level modulating layer on a surface of the perovskite layer, and forming the vacuum-level modulating layer on the surface of the perovskite layer. The vacuum-level modulating layer provides crystal defect passivation of the surface of the perovskite layer, and the vacuum-level modulating layer further modulates the surface energetics of the surface of the perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein. [0009] A perovskite layer for use in a photovoltaic device according to an embodiment of the current invention is produced according to a method of treating a surface of a perovskite layer according to An embodiment of the current invention. BRIEF DESCRIPTION OF THE DRAWINGS [0010] The present disclosure, as well as the methods of operation and functions of the related elements of structure and the combination of parts and economies of manufacture, will become more apparent upon consideration of the following description and the appended claims with reference to the accompanying drawings, all of which form a part of this specification, wherein like reference numerals designate corresponding parts in the various figures. It is to be expressly understood, however, that the drawings are for the purpose of illustration and description only and are not intended as a definition of the limits of the invention. As used in the specification and in the claims, the singular form of “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise. [0011] FIG. 1 is a schematic illustration of a photovoltaic cell according to an embodiment of the current invention. [0012] FIGS. 2A-2I show perovskite surface and heterointerface dynamics. FIGS. 2A, 2B (a, b) show UPS secondary electron cut-offs of various surface-treated perovskite films. Labels are BA: butylammonium, OA: octylammonium, DA: dodecylammonium. KPFM surface potential maps of the FIG. 2C (c), Reference, FIG. 2D (d), OAI-treated, and FIG. 2E (e), OATsO-treated films. Insets include the corresponding AFM topography images. All scale bars are 2 μm. FIG. 2F (f), Work function distributions and FIG. 2G (g), RMS surface roughness (R_q) of the films measured by KPFM and AFM, respectively. Charge density distribution profiles of the complete FIG.2H (h), OATsO-treated and FIG.2I (i), OAI-treated device cross-sections measured by cross-sectional KPFM. The devices were illuminated under open-circuit condition. Error bars demarcate the estimated spatial resolution of ~30 nm.^8,28 [0013] FIGS. 3A-3H show charge carrier dynamics, performance, and photostability. FIG. 3A (a), Steady-state and FIG. 3B (b), time-resolved PL spectra of the glass/perovskite films. FIG.3C (c), Steady-state and FIG.3D (d), time-resolved PL spectra of the glass/perovskite/spiro-MeOTAD films. Included are the extracted lifetimes fitted with a mono-exponential decay function for (b) and bi-exponential decay function for (d). FIG. 3E (e), Histogram showing the PCE distributions of the devices. FIG. 3F (f), Photostability evolution with time of encapsulated devices aged under continuous illumination at maximum power point. P₀ denotes the initial PCE. FIG. 3G (g), Photostability evolution with time of encapsulated devices aged under continuous illumination in open-circuit condition. Error bars represent the standard deviation of four devices for each condition. FIG.3H (h), Photostability PCE evolution of the most stable OATsO-treated device aged under open-circuit condition. Included are the PCE retentions in approximately 500 h intervals. [0014] FIGS. 4A-4J show STEM and EDX analyses of the aged devices. STEM bright field images of the aged FIG. 4A (a), OAI-treated and FIG. 4B (b), OATsO- treated device cross-sections. The OAI-treated device is seen to have a rougher heterointerface contacting spiro-MeOTAD. EDX elemental maps of bromine for the FIG. 4C (c), OAI-treated and FIG. 4D (d), OATsO-treated device cross-sections, and iodine for the FIG.4E (e), OAI-treated and FIG.4F (f), OATsO-treated device cross-sections. All scale bars in the STEM and EDX images represent 200 nm. Elemental distributions of bromine for the FIG.4G (g), OAI-treated and FIG.4H (h), OATsO-treated devices, and iodine for the FIG.4I (i), OAI-treated and FIG.4J (j), OATsO-treated devices. [0015] FIGS. 5A-5E show physical origins of the experimental observations. Activation energy for the FIG. 5A (a), intra-lattice and FIG. 5B (b), extra-lattice migrations of iodine or bromine calculated using first-principles NEB simulations. The forward migration pathways are indicated by the arrows in the example simulated supercells. Charge displacement with FIG. 5C (c), [OA]⁺ or FIG.5D (d), [TsO]- on the surface, superimposed on the defect-free slab models, and FIG. 5E (e), corresponding charge displacement profiles. Blue and yellow volumes on the slab models correspond to electron-depleted or electron-enriched regions, respectively. [0016] FIGS. 6A-6F show morphology of the perovskite films. Surface morphology of the: FIG. 6A (a), reference; FIG. 6B (b), OAI-treated; FIG. 6C (c), OATsO-treated; FIG.6D (d), OATFA-treated; FIG.6E (e), OABr-treated; and FIG.6F (f), OABF₄-treated perovskite films measured by SEM. All scale bars represent 2 μm. No obvious difference can be seen between the reference and treated perovskite films. [0017] FIGS. 7A-7G show topography of the passivated perovskite films. Representative 3D topography of the: FIG. 7A (a), OAI-treated; FIG. 7B (b), OABF₄- treated; and FIG. 7C (c), OATsO-treated perovskite films measured by AFM. All scale bars represent 2 μm. FIG. 7D (d), Comparison of the height depth distribution of the films. Depth distribution histograms for the: FIG. 7E (e), OAI-treated; FIG. 7F (f), OABF₄-treated; and g OATsO-treated perovskite films. Insets include the fitted statistical parameters. SD, standard deviation. [0018] FIGS. 8A-8B show heterointerface energy band diagrams. Schematic interpretation of the heterointerface band alignments of the FIG. 8A (a), OAI-treated and FIG. 8B (b), OATsO-treated devices under illumination in open-circuit condition. The band alignments are constructed based on the UPS and KPFM results. CBM, conduction band minimum; VBM, valence band maximum; E_fn, electron quasi-fermi level; E_fp, hole quasi-fermi level. The dashed gray lines in FIG. 8B (b) indicate E_fn and E_fp of the OAI-treated device from FIG.8A (a). The diagrams are not drawn to scale. Both surface treatments create a type I energy alignment at the heterointerface, but the vacuum level upshift of the OATsO-treated device minimized the potential well to mitigate the electron accumulation. [0019] FIGS. 9A-9H show device cross-sectional KPFM profiling. FIG. 9A (a), CPD profile; FIG. 9B (b), KPFM spatial mapping; FIG. 9C (c), corresponding AFM spatial mapping; and FIG. 9D (d), electric field distribution of the OAI-treated device. FIG. 9E (e), CPD profile; FIG. 9F (f), KPFM spatial mapping; FIG. 9G (g), corresponding AFM spatial mapping; and FIG. 9H (h), electric field distribution of the OATsO-treated device. Measurements were performed under illumination in open-circuit condition. All scale bars represent 300 nm. The CPD offsets were adjusted such that the CPD value of the buffer layer becomes zero. Note that this does not affect the electric field and charge displacement profiles, which calculate the derivatives of the CPD profiles. Although we do not expect the rough morphology seen in FIG. 9C (c) to affect the KPFM signal, we cannot completely rule this out at this stage. [0020] FIGS. 10A-10D show device photovoltaic parameters. Box plots showing the distribution of the: FIG. 10A (a), V_OC; FIG. 10B (b), J_SC; FIG. 10C (c), FF; and FIG. 10D (d), PCE of the devices. Centre line, median; box limits, 25^th and 75^th percentiles; whiskers, outliers. [0021] FIGS. 11A-11C show characterization of the OATsO-treated devices. FIG. 11A (a), Current density-voltage curves of the best-performing OATsO-treated device, in reverse scan (blue line) and forward scan (red line). Inset includes the measured photovoltaic parameters. FIG.11B (b), EQE spectrum and integrated J_SC of an OATsO- treated device. The integrated J_SC is 24.6 mA cm⁻², and therefore well matched (less than 3% discrepancy) with the measured value. FIG. 11C (c), Absorbance profile of an OATsO-treated film on glass measured by UV-Vis spectroscopy. Inset includes a Tauc plot and linear fits to estimate the optical bandgap. [0022] FIGS.12A-12E show third-party device performance measurements. FIG. 12A (a), Current density-voltage curve, and FIG. 12B (b), box plot showing the PCE distribution of the encapsulated OATsO-treated devices. Measurements were performed at the Molecular Foundry, Berkeley, CA, USA. As the measurements were fully done in ambient air (RH approximately 50%), all devices had to be encapsulated, which resulted in a drop in performance. FIG. 12C (c), PCE evolution with time under lightsoaking of an encapsulated OATsO-treated device. Current density-voltage curves of the same device FIG. 12D (d), before and FIG. 12E (e), after the encapsulation procedure, measured in-house. [0023] FIGS. 13A-13H show universality verification on a FAPbI₃ composition. FIG. 13A (a), UPS secondary electron cut-offs of the perovskite films. FIG. 13B (b), Steady-state and FIG. 13C (c), time-resolved PL spectra of the glass/perovskite films. FIG. 13D (d), Steady-state and FIG. 13E (e), time-resolved PL spectra of the glass/perovskite/spiro-MeOTAD films. The carrier lifetimes are fitted with a mono- exponential decay function. FIG. 13F (f), Box plots of the distribution of the device photovoltaic parameters. Centre line, median; box limits, 25^th and 75^th percentiles; whiskers, outliers. FIG.13G (g), Current density-voltage curves and FIG.13H (h), EQE spectrum and integrated J_SC of the best-performing device treated with OATsO. The integrated J_SC is 25.4 mA cm⁻², well matched (approximately 3% discrepancy) with the measured scan value. [0024] FIGS. 14A-14C show open-circuit stability test device performance. Evolution with time of the normalized average FIG. 14A (a), J_SC, FIG. 14B (b), V_OC, and FIG.14C (c), FF of the devices under continuous illumination with a metal halogen lamp. The encapsulated devices were aged in ambient atmosphere at RH ~ 40% and T ~ 40 °C in open‐circuit condition. Error bars represent the standard deviation of four devices for each condition. [0025] FIGS. 15A-15F show activation energy for halide migration. Energy profiles for the extra-lattice migration of FIG.15A (a), bromine and FIG.15B (b), iodine in a neutral uncharged or negatively charged environment, and FIG. 15C (c), corresponding activation energy barriers. Energy profiles for the intra-lattice migration of FIG. 15D (d), bromine and FIG. 15E (e), iodine in a neutral uncharged or negatively charged environment, and FIG.15F (f), corresponding activation energy barriers. [0026] FIG. 16 shows work function change with ammonium iodide surface treatments. UPS secondary electron cut-offs of perovskite films treated with various ammonium iodide salts. Labels are PEA: phenylethylammonium, BA: butylammonium, OA: octylammonium, DA: dodecylammonium. [0027] FIG. 17 shows XPS characterization of the films. High-resolution Pb 4f spectra of the films. Dashed vertical lines demarcate the peak positions for the Reference film. The Pb 4f peaks of the OAI-treated film shifted to a higher binding energy relative to the Reference film, consistent with the interaction of alkylammonium iodides with Pb.¹⁹ For the OATsO-treated film, the shift to lower binding energy is also consistent with the interaction of the -SO₃- group of [TsO]- with Pb.²⁰ [0028] FIG. 18 shows Cross-sectional KPFM schematic. Schematic illustrating the KPFM measurement setup and planar device structure of ITO/SnO₂/perovskite/spiro- MeOTAD/Au. The MgF₂ buffer layer prevents slippage of the cantilever tip at a “cliff”. Often, the device cross-sections are not parallel, due to the uncontrollable nature of the mechanical cleaving process. Therefore, the measured distance may not represent the layer thickness. Note that we purposely avoided using FIB or mechanical polishing for the KPFM samples, as such processing may modify the properties of the exposed cross- sectional surface.²¹ [0029] FIGS. 19A-19C show Cross-sectional KPFM profiling of another OAI- treated device. a, AFM spatial mapping of an OAI-treated device. Yellow arrows in (a) indicate damaged regions due to the mechanical cleaving process. b, KPFM spatial mapping, and c, CPD profile of the OAI-treated device. The magnified KPFM image in (b) spatially corresponds to the blue square region in (a). All scale bars represent 500 nm. The ΔE_vac value from the CPD profile matches those previously measured as shown in Supplementary FIG.20. [0030] FIG. 20 shows different measurements of the vacuum level change. Comparison of the ΔE_vac or ΔW of the perovskite devices or films measured from KPFM or UPS. [0031] FIGS. 21A-21F show PL mapping of the passivated films. Representative confocal PL maps of the a, OAI-treated, b, OABF₄-treated, c, OATsO-treated perovskite films on glass. All scale bars represent 5 μm. PL intensity distribution histograms of the d, OAI-treated, e, OABF₄-treated, f, OATsO-treated perovskite films on glass. Insets include the fitted statistical parameters. SD: standard deviation. [0032] FIGS. 22A and 22B show photo-transient measurements of the complete devices. Normalized a, transient photovoltage decay and b, transient photocurrent decay of the complete devices. [0033] FIGS. 23A-23F show GIXRD characterization of the films. Grazing incidence X-ray diffraction (GIXRD) of the a, Reference, b, OAI-treated, c, OATsO- treated, d, OATFA-treated, e, OABr-treated, and f, OABF₄-treated perovskite films. Measurements were performed at an ω = 0.2°, with an estimated penetration depth of <60 nm. [0034] FIGS. 24A-24F show device hysteresis behavior. a, Current density- voltage curve of the a, Reference, b, OAI-treated, c, OATsO-treated, d, OATFA-treated, e, OABr-treated, and f, OABF₄-treated devices, in reverse bias (black solid lines) and forward bias (red dashed lines). Insets include the measured photovoltaic parameters. We note that these devices did not have the MgF₂ anti-reflection coating. The hysteresis index (HI) is included in the insets. [0035] FIGS. 25A and 25B show STEM images of the aged devices. STEM bright field images of the aged a, OAI-treated and b, OATsO-treated device cross- sections. All scale bars represent 200 nm. [0036] FIGS. 26A and 26B show EDX analyses of the aged devices. EDX elemental maps of lead for the a, OAI-treated and b, OATsO-treated device cross- sections. All scale bars represent 200 nm. [0037] FIGS. 27A-27E show Halide distribution in the aged devices. Average atomic ratio distributions across the a, OAI-treated and b, OATsO-treated device cross- sections, and iodine for the c, OAI-treated and d, OATsO-treated device cross-sections. e, Peak values extracted from (a)-(d) at the electrode region. The average atomic ratios were calculated by simply multiplying the normalized signal intensity with the nominal perovskite stoichiometry. [0038] FIGS. 28A and 28B show XRD analyses of the aged devices. XRD diffraction patterns measured with an ω-2θ setup of the aged a, OAI-treated and b, OATsO-treated devices. These were the same devices used for the STEM analyses, and were recovered after FIB milling. X-rays from the ω-2θ diffraction setup were able to penetrate through the top electrode and spiro-MeOTAD layers to probe the active layer itself. Concurrently, given the large penetration depth of ω-2θ XRD (>700 nm), it is not sensitive to the ultrathin 2D interlayer, which is only detectable by the GIXRD measurements. [0039] FIG. 29 shows Half-device photostability test. Stability evolution with time of half-devices aged under continuous illumination in a nitrogen atmosphere. The time elapsed refers only to the time under illumination. Error bars represent the standard deviation of four devices for each condition. [0040] FIGS. 30A-30C show Electrostatic potentials of various cations. Electrostatic potential maps of a, [BA]⁺, b, [OA]⁺, and c, [DA]⁺. Included are their respective dipole moments (µ). [0041] FIGS. 31A-31D show Electrostatic potentials of various anions. Electrostatic potential maps of a, [I]-, b, [Br]-, c, [TFA]-, and d, [TsO]-. Included are their respective dipole moments (µ). [0042] FIGS. 32A-32F show Surface states and charge displacement. Charge displacement with surface iodine vacancy defect (V_I) with a, [OA]⁺ or b, [TsO]-, and c, corresponding charge displacement profiles. Charge displacement with surface lead- iodine antisite defect (Pb_I) with d, [OA]⁺ or e, [TsO]-, and f, corresponding charge displacement profiles. Blue and yellow volumes on the slab models correspond to electron-depleted or electron-enriched regions, respectively. The point defects are marked by pink squares. [0043] FIGS. 33A-33F show Charge displacement profiles with different anions and concentrations. Charge displacement profiles on defect-free surfaces with a, [I]-, b, [Br]-, and c, [TFA]-. Charge displacement profiles on defect-free surfaces with increasing [OA]⁺ or [TsO]- surface concentration from d, 25 %, to e, 50 %, and to f, 100 %. [0044] FIGS. 34A-34D show Device photovoltaic parameters with different tosylate-based salts. Box plots showing the distribution of the a, V_OC, b, J_SC, c, FF, and d, PCE of the devices. Center line: median, box limits: 25^th and 75^th percentile, whiskers: outliers. [0045] FIG. 35 shows ¹H NMR spectra of p-Toluenesulfonic acid (red), octylamine (blue) and OATsO salt (black) in DMSO-d₆. The chemical structures and reaction scheme are shown above the spectra. All atoms are labelled for identifying their corresponding peaks in the spectra. DETAILED DESCRIPTION [0046] Some embodiments of the current invention are discussed in detail below. In describing embodiments, specific terminology is employed for the sake of clarity. However, the invention is not intended to be limited to the specific terminology so selected. A person skilled in the relevant art will recognize that other equivalent components can be employed and other methods developed without departing from the broad concepts of the current invention. All references cited anywhere in this specification, including the Background and detailed description sections, are incorporated by reference as if each had been individually incorporated. [0047] The term “substrate” is intended to have a broad meaning that can include a single layer of a material, multiple layers, laminated layers, composite layers, flexible structures, rigid structures, organic materials, inorganic materials, or combinations thereof. For example, and without limitation, substrates can be glass, plastic, or crystalline materials (e.g., silicon or another semiconductor). [0048] Some embodiments of the current invention are directed to a method to control the surface energetics (i.e. n-type, p-type) in halide perovskite materials to control the heterointerface charge distribution in perovskite solar cells (PSCs). The terms “solar cells” and “photovoltaic cells” are used interchangeably herein. This method. according to some embodiments, is used to achieve long-term stable PSCs under illumination operational conditions. [0049] Therefore, some embodiments of the current invention are directed to a method to control the perovskite surface energetics, to solve the negative ΔE_vac and charge accumulation problems. We show that substituting the surface treatment iodide [I]- counter-anion with any of tetrafluoroborate [BF₄]-, trifluoroacetate [TFA]-, or bromide [Br]- partially negated (to varying extents) the negative ΔE_vac. With counter-anions that show even stronger surface electron-displacement character, such as tosylate [TsO]- and triflate [TfO]-, the negative ΔE_vac can be fully neutralized and even lead to positive ΔE_vac. We show that controlling the surface energetics can mitigate the potential well and avoid charge accumulation at the PSC heterointerface. [0050] In our studies using both experimental and computational methods, we show that controlling the perovskite surface energetics played a significant role in suppressing ion migration. As a result, switching the counter-anion from [I]- into [TsO]- was found to significantly improve the long-term stability of PSCs under illumination operational conditions according to some embodiments of the current invention. Under accelerated maximum power point photostability testing, the [TsO]--treated device sustained its performance with negligible degradation after ~800 h. Under accelerated open-circuit condition photostability testing, the [TsO]--treated devices sustained 87.0 % of their initial average performance after ~2,092 h. To further show that the suppressed ion migration in the [TsO]--treated devices, the original devices from the 2,096 h open- circuit condition photostability testing were sent for scanning transmission electron microscopy (STEM) analysis. Compared with the [I]--treated devices, significantly suppressed ion migration and heterointerface degradation was directly observed by the STEM analyses. [0051] Accordingly, FIG. 1 is a schematic illustration of a photovoltaic device 100 according to an embodiment of the current invention. This photovoltaic device can be a PSC, for example. This photovoltaic device includes a first electrode 102; a second electrode 104 spaced apart from the first electrode 102; a perovskite layer 106 between the first and second electrodes (102, 104) and a vacuum-level modulating layer formed 108 on a surface of said perovskite layer 106. The perovskite layer 106 includes a semiconducting halide perovskite photoactive material. In FIG. 1, the substrate 110 can be or can include either the first or second electrode (102, 104). The current example shows the second electrode 104 being the same as or integrated with the substrate 110. However, the broad concepts of the current invention are not limited to only this example of a particular configuration. In some embodiments, the substrate can be a transparent substrate, such as glass, with a layer of transparent electrically conductive material such as, but not limited to, indium tin oxide (ITO). [0052] The vacuum-level modulating layer 108 provides crystal defect passivation of the surface of the perovskite layer 106. The vacuum-level modulating layer 108 further modulates surface energetics of the surface of the perovskite layer 106 to mitigate formation of a potential well so as to avoid charge accumulation therein. [0053] In some embodiments, the photovoltaic device 100 of FIG. 1 can further include an electron transport layer 112 between the second electrode 104 and the perovskite layer 106. In some embodiments, the photovoltaic device of FIG. 1 can further include a hole transport layer 114 between the first electrode 102 and the perovskite layer 106. [0054] In some embodiments, the vacuum-level modulating layer can include a material comprising at least one of a tetrafluoroborate [BF4]-, a trifluoroacetate [TFA]-, a bromide [Br]-, a tosylate [TsO]-, or a triflate [TfO]- counter-anion. However, the general concepts of the current invention are not limited to only these materials. Methods of selecting materials according to further embodiments of the current invention that may be suitable for use as a vacuum-level modulating layer are described in detail below. Additional embodiments of the current invention are directed to methods of producing devices that include a vacuum-level modulating layer formed on a surface of said perovskite layer. EXAMPLES [0055] The following provides some examples to help explain further concepts of the current invention. However, the general concepts of the current invention are not intended to be limited t only the examples. [0056] Optoelectronic devices have heterointerfaces formed between dissimilar semiconducting materials. The relative energy level alignment between contacting semiconductors determinatively affects the heterointerface charge injection and extraction dynamics. For perovskite solar cells (PSCs), the heterointerface between the top perovskite surface and a charge-transporting material (CTM) is often treated for defect passivation^1–4 to improve PSC stability and performance. However, such surface treatments could also affect the heterointerface energetics.¹ Here we show that surface treatments may induce a negative work function shift (i.e. more n-type), which activates halide migration to aggravate PSC instability. Therefore, despite the beneficial effects of surface passivation, this detrimental side effect limits the maximum stability improvement attainable for PSCs treated in these ways. This trade-off between the beneficial and detrimental effects should guide further work on improving PSC stability via surface treatments. [0057] Progress in compositional and crystal growth engineering have made possible the fabrication of halide perovskite thin films with minimized bulk trap density, such that defects are predominantly located at the surface.^5,6 This has motivated the development of defect passivation treatments applied onto the top perovskite surface.^1–4 However, such treatments could also change the heterointerface energetics and thus charge carrier dynamics between the perovskite and top CTM.¹ In this work, we study the consequences of the altered heterointerface energetics on carrier extraction, trap passivation, charge accumulation, and ion migration. We show that a negative work function change (ΔW) accumulates charges in a potential well, which lowers the halide migration activation energy to limit PSC stability. A negative ΔW is equivalently described as a negative vacuum level change (ΔE_vac) at a heterointerface, and hereafter, ΔW and ΔE_vac are used interchangeably. Perovskite surface and heterointerface dynamics [0058] The genesis of this study began with our investigations on the surface energetics of perovskite films based on a (FAPbI₃)_0.95(MAPbBr₃)_0.05 composition. Ultraviolet photoelectron spectroscopy (UPS) measurements show that ubiquitously used iodide-based ammonium surface treatments generally result in a negative ΔW (FIG.2A, FIG. 16, Table 1), successively increasing in magnitude with longer alkylammonium chain length. Given that a negative ΔW is associated with a relatively more electron- enriched surface, we speculated initially that the negative ΔW can be modulated by increasing the electron-withdrawing ability of the counter-anion. Further UPS measurements (FIG. 2B) showed that substitution of iodide [I]- with bromide [Br]-, tetrafluoroborate [BF₄]-, or trifluoroacetate [TFA]- progressively negated the negative ΔW of OAI treatment towards that of the Reference film, but only substitution with tosylate [TsO]- fully neutralized the negative ΔW. Particularly, among the counter-anions, [TsO]- has the strongest electron-withdrawing character,⁷ which also justifies its ubiquitous use as the leaving group in synthetic heterolytic fission chemistry. The interaction of [TsO]- with the surface is further verified by XPS analysis (FIG.17). [0059] Table 1. Energy levels of the perovskite films. The work function (WF) and (ionization energy (IE) – WF) values are based on the UPS measurements. The vacuum level shift (ΔE_vac) and valence band maximum (VBM) values are calculated based on an aligned Fermi level.

[0060] Kelvin probe force microscopy (KPFM) measurements were performed to verify the work function distributions (FIGS. 2C-2F). The mean work function of the OAI-treated film was decreased to 4.49 ± 0.09 eV, from 4.67 ± 0.08 eV of the Reference film, and in contrast with the 4.77 ± 0.11 eV of the OATsO-treated film. From the topographical atomic force microscopy (AFM) maps, the surface morphology and root- mean-square roughness of the treated films were negligibly different (FIG.2G), expected given the dilute (but common) solution concentration used for surface treatment. Further comparisons of the height depth distributions and scanning electron microscopy (SEM) images also suggest that the surface uniformity is relatively similar for the treated films (FIGS.6A-7G). Consequences of heterointerface energetics on charge carrier dynamics [0061] Band alignments constructed from the UPS results predict that a negative ΔE_vac may create a potential well to trap charges at the heterointerface (FIGS. 8A-8B, Note 1). Cross-sectional KPFM measurements under illumination in open-circuit condition were followed to investigate the real-time charge carrier distributions in complete devices of planar architecture ITO/SnO₂/perovskite/spiro-MeOTAD/Au (FIGS. 9A-9H, Note 2, FIGS. 18, and 19A-19C). The measured device ΔE_vac values at the perovskite/spiro-MeOTAD heterointerface are consistent with those obtained from the films (FIG. 20). Charge carriers are unextractable in open-circuit condition (split quasi Fermi level), and might accumulate at a contacting selective heterointerface of the opposite polarity.^8,9 This is observed as an accumulation of holes for both devices at the perovskite/SnO₂ heterointerface (FIGS. 2H, 2I). In contrast, a pronounced electron accumulation exists at the perovskite/spiro-MeOTAD heterointerface for only the OAI- treated device. The electron accumulation is noted to be significantly more severe than the counterpart hole accumulation. In principle, the ideal photovoltaic device would have a homogenous electric field distribution with no charge accumulation across its heterojunctions,¹⁰ which is seen for the OATsO-treated device at its perovskite/spiro- MeOTAD heterointerface. Both devices were relatively field-free (i.e. flat potential) along the active layer, indicative of a high-quality perovskite bulk, suggesting that the different behaviors were a consequence of the surface treatments. The results are consistent with those predicted based on the band diagrams (FIGS. 8A-8B), where the non-negative ΔE_vac of the OATsO-treated device avoided the potential well and charge accumulation. [0062] Photoluminescence (PL) spectroscopy indicates an effective suppression of charge-trapping defect states in the surface-treated films, as evidenced by their enhanced PL intensities and carrier lifetimes with a glass/perovskite architecture (FIGS. 3A, 3B). The defect passivation efficacies of the surface-treated films are noted to be relatively comparable. Comparing their PL intensity distributions indicate again that the surface uniformity is similar between the films (FIGS. 21A-21F). Despite the beneficial passivation effects, charge extraction into spiro-MeOTAD is sacrificially impeded in all surface-treated films (FIGS. 3C, 3D, Table 2), but to different extents, with the trend correlated with the magnitude of ΔE_vac. Compared to the Reference/spiro-MeOTAD film, the average carrier lifetime (τ_ave) more than doubled to 8.1 ns (from 3.0 ns) for the OAI- treated film and the steady-state PL intensity was 223% higher. In contrast, OATsO treatment simultaneously suppressed trap states while charge extraction is barely impeded, avoiding the trade-off seen for OABF₄ and OAI treatments. Further photo- transient measurements on complete devices complement the film PL results (FIGS.22A, 22B). We postulate that the charge obstruction and accumulation cannot be explained by the surface 2D phase, since the insulating large organic cation (OA⁺) is kept unchanged (Note 3, FIGS.23A-23F). [0063] Table 2. Time-resolved photoluminescence decay parameters of the glass/perovskite/spiro-MeOTAD films, fitted with a bi-exponential decay model.

Device performance and hysteresis behavior [0064] For the surface-treated devices, the power conversion efficiency (PCE) trends as OATsO-treated > OABF₄-treated > OAI-treated devices, due primarily to an increasing fill factor (FF) (FIG. 3E, FIGS. 10A-10D). The device open-circuit voltages (V_OC) are marginally different, reflective of the comparable PL results of the glass/perovskite films. The best OATsO-treated device reached a PCE of 24.41% (FIGS. 11A-11C). We further verified the performance of encapsulated devices at an independent third-party laboratory, noting that the performance slightly decreased after the encapsulation procedure (FIGS. 12A-12E). Contrasting the surface-treated devices, the PCE trend can be explained as follows (further discussion in Note 4); firstly, a more negative ΔE_vac deepened the valence band offset with spiro-MeOTAD, which increased hole extraction resistance to sacrifice FF. Additionally, the heterointerface barrier observed from the KPFM profiling may also contribute to impede charge extraction.^9,11 Regardless, a negative ΔE_vac remains the cause that gave rise to both effects. Separate investigations on a FAPbI₃ composition further verified our observations (FIGS. 13A- 13H). [0065] Comparing the surface-treated devices, the hysteresis behavior generally improved as the negative ΔE_vac decreases in magnitude (FIGS.24A-24F). This provided the first hint at a correlation between charge accumulation and ion migration, given that ion migration is responsible for PSC hysteresis.^12,13 On the other hand, the Reference devices exhibited the lowest performance and worst hysteresis, due to the abundant, unpassivated heterointerface traps. Overall, the device performance results provide evidence for the sacrificial trade-off of a negative ΔE_vac to limit PSC performance. Device stability under continuous illumination [0066] We assessed the photostability of encapsulated devices under continuous illumination without an ultraviolet filter. All devices were aged in ambient atmosphere at ~40 °C. We preserved the original device architecture with spiro-MeOTAD as the HTM, without applying further modifications (e.g. copper phthalocyanine (CuPC) or poly(triarylamine) (PTAA) were not used). Under maximum power point (MPP) testing (FIG. 3F), the OATsO-treated device sustained its performance with negligible degradation after ~800 h. The OABF₄-treated device was also relatively stable, retaining 91.5% of their performance after ~500 h. Among the surface-treated devices, the OAI- treated device degraded the most rapidly to 84.8% of their performance after ~500 h. [0067] Contrasting the device degradation under MPP versus open-circuit condition (OC) testing, all devices were generally less stable under OC testing (FIG.3G, FIGS. 14A-14C, Note 5). However, the device degradation trends are identical, and correlated with the magnitude of ΔE_vac. Particularly, the OABF₄-treated devices with an intermediate negative ΔE_vac exhibited stability between that of the OAI-treated and OATsO-treated devices. Near the approximate halfway point of 1,014 h, the devices retained 94.3% (OATsO-treated), 86.2% (OABF₄-treated), and 74.8% (OAI-treated) of their average PCEs. Ending after 2,092 h, the OATsO-treated devices retained 87.0% of their initial PCE on average. On the other hand, the average PCE (65.1% of initial) of the OAI-treated devices decreased dramatically over the 2,092 h aging duration. The most stable OATsO-treated device (FIG.3H) retained 94.9% and 88.5% of its initial PCE after 1,014 h and 2,092 h, respectively. [0068] Analyzing the degradation trends, we postulated that the potential well and charge accumulation may have accelerated ion migration, considering that the KPFM profiling was also performed under illumination in open-circuit condition. Particularly, the more rapid “burn-in” decay of the OAI-treated and OABF₄-treated devices, observed in both the MPP (Table 4) and OC tests, strongly hints that ion migration has been aggravated by a negative ΔE_vac, given that transient ion migration underlies the “burn-in” regime.^12,14,15 The altered ion migration energetics was also hinted by the device hysteresis behavior. Moreover, the 2D interlayer of conventional OAI treatment likely contributed to impeding ion migration,¹⁶ yet the OABF₄ and OATsO-treated devices still had superior photostability, indicating that the ion migration energetics have been dominantly affected. [0069] Table 3. Energy levels of the FAPbI₃ perovskite films. The work function (WF) and (ionization energy (IE) – WF) values are based on the UPS measurements. The vacuum level shift (ΔE_vac) and valence band maximum (VBM) values are calculated based on an aligned Fermi level.

Table 4. Comparison of the device degradation at different time periods during the MPP stability test.

Analyses of the degraded devices [0070] To directly investigate the extent of ion migration, upon completion of the OC stability testing, the original OAI-treated and OATsO-treated devices (after 2,092 h illumination) were sent for scanning transmission electron microscopy (STEM) analysis. The encapsulation cover glasses were detached immediately prior to focused ion beam (FIB) milling to extract sample cross-sections. STEM bright field images of the planar device stacks reveal that the micron-scale grain sizes are still visibly intact even after the extended aging (FIGS. 4A, 4B; FIGS. 25A, 25B). The first significant difference between the two samples can be seen at the perovskite/spiro-MeOTAD heterointerface. The OAI-treated device had a rougher heterointerface morphology, which is in contrast with the negligibly different topography for the fresh films observed by AFM and SEM. This suggests that the roughening is inherent to the aging process, possibly associated with heterointerface degradation by ion migration-induced compositional loss. The elemental distributions were compared by X-ray energy-dispersive (EDX) mapping in STEM (FIGS. 4C-4F, FIGS. 26A-26B). On close inspection, large accumulations of both bromine (green arrows) and iodine (yellow arrows) can be seen for the OAI-treated device along the top gold/platinum region. This is verified by quantitative analysis of the elemental distributions (FIGS. 4G-4I). In consideration of the nominal stoichiometry of the perovskite, both bromine and iodine accumulation were approximately similar in order magnitude (FIGS. 27A-27E). We note that despite progress to completely rid the need for bromine in FAPbI₃-based compositions,^17,18 it was shown that bromine might remain necessary to stabilize the α-FAPbI₃ phase in state-of-the-art devices.¹⁹ The bulk devices that underwent FIB milling were further recovered and characterized by ω-2θ XRD (FIGS. 28A, 28B). The δ-FAPbI₃ peak supposedly at ~11.8° was not detected for both devices, and the α-FAPbI₃ phase remained dominant, ruling out possible aggravation of the perovskite phase metastability by iodine interstitial generation.²⁰ Aggravated ion migration and device instability [0071] We further simulated the ion migration pathways using the first principles nudged elastic band (NEB) methodology (Note 6). To model the charge accumulation at the perovskite/spiro-MeOTAD heterointerface, we calculated the activation energy for halide migration in either a neutral uncharged, or negatively charged environment. Two independent pathways were explored; 1) migration within a unit cell of a supercell (intra- lattice), and 2) migration escaping a supercell (extra-lattice) (FIGS. 15A-15F, Table 5). The intra-lattice migration (vacancy mediated) investigates possible coulombic screening effects by a charged environment to alter the bonding affinities and thus migration energetics,^21,22 while the extra-lattice migration is defect-independent. For intra-lattice migration (FIG. 5A), the energy barriers for both iodine (-38.7%) and bromine (-29.4%) decreased substantially in the negatively charged environment. Similarly observed for the extra-lattice scenario (FIG. 5B), the activation energy for iodine migration (-13.3%) and bromine migration (-17.8%) were both lower in the charged environment. Per the Arrhenius relationship, the rate constant has exponential dependance on the activation energy barrier. [0072] Table 5. Activation energy for iodine or bromine migration in a neutral uncharged or negatively charged environment calculated from first-principles nudged elastic band (NEB) simulations.

[0073] Ion penetration can result in irreversible chemical reactions with spiro- MeOTAD to degrade its hole-transporting functionality,²³ and also chemical corrosion of the top electrode layer,²⁴ which potentially contributed to the V_OC and FF decays of the devices. The robust encapsulation procedure ruled out environmental degradation factors, indicating that intrinsic mechanisms (i.e. ion migration) were responsible, and also excluded trapped charges catalyzing extrinsic degradation pathways by moisture and oxygen.²⁵ This is evidenced by the J_SC retention for the devices, demonstrating the excellent phase stability of the active layers. By excluding ion migration, half-device tests further indicate that the photostability of the surface-treated films are negligibly different (Note 7, FIG. 29). Tying together all results, we propose that the potential well and charge accumulation created by a negative ΔE_vac aggravated device instability, by detrimentally accelerating halide migration at the spiro-MeOTAD/perovskite heterointerface. More generally, the beneficial improvements of surface treatments are sacrificially limited by a negative ΔE_vac, but modulating the counter-anion presents a simple method to further improve PSC stability and performance. [0074] Backtracking, we also preliminarily explored the possible mechanistic origins of ΔE_vac using first principles (full discussion in Note 8), summarized as follows: Ligand-induced ΔW originate from two independent contributions;^26,27 1) the ligand intrinsic dipole moment

and 2) a charge density displacement by ligand-surface interactions and bond formation The di ^{+ -}

pole moments of [OA] and [TsO] were +20.3 D and -10.5 D, respectively (FIGS.30A-30C, FIG.31A-31D). On the other hand, [TsO]- induces a striking negative surface charge displacement, , of -0.022 e, three-fold larger in magnitude than the positive of [OA]⁺ (+0.006 e) (FIGS.5C-5E, Table 6)_. The negative sign of [TsO]- indicates a depletion of electrons from the surface, and is equivalent with a positive ΔW. This trend is reproducible even with defect states on the surface, or different surface concentrations and species (FIGS. 32A-32F, 33A-33F, Table 7). Therefore, we speculate that simultaneous contributions from both and for [TsO]- synergistically counterbalanced [OA]⁺, to result in a net positive ΔE_vac for OATsO treatment. More broadly, this proposes design principles for the ideal perovskite/CTM heterointerface. We preliminarily tested pairing [TsO]- with alkylammoniums with different chain lengths, but OATsO treatment yielded the best performance (FIGS.34A-34D, Note 9). [0075] Table 6. Surface charge displacement, , with different species attached to defect-free surfaces.

[0076] Table 7. Surface charge displacement, , with different species attached to defected surfaces. V_I and Pb_I represent iodine vacancy and lead-iodine antisite defects, respectively, which may readily form on the perovskite surface due to their low formation energies.^17,18

[0077] The references for the following notes follow immediately after the notes. [0078] Note 1: Further discussion on the heterointerface energy band diagrams. In FIGS.8A-8B, the perovskite bulk work function should also be affected by ΔE_vac. This is because the less negative ΔE_vac by replacing OAI treatment with OATsO treatment resultantly delocalizes the accumulated electrons in the potential well, which enhances the electron density and thus electron quasi-fermi level

work function of the perovskite bulk. Moreover, the change in bulk work function by surface treatment would also affect the band alignment and thus charge accumulation at the perovskite/SnO₂ heterointerface. Indeed, a slightly increased hole accumulation was observed for the OATsO-treated device (FIGS. 2H-2I). However, although the hole quasi-fermi level for the OATsO-treated device slightly up-shifted, its increase is not as significant as . The overall effect on the quasi-fermi level splitting is consistent with the slight V_OC enhancement by replacing OAI treatment with OATsO treatment (FIGS.10A-10D). [0079] Note 2: Derivation of the charge density distribution from KPFM. The electric field distribution across the device cross section, E( ), was calculated from the contact potential difference, V( ), according to the equation:¹

where is the work function of the probe and is the elementary charge. is constant. The charge density distribution in a device, , is then given by:¹

where are the vacuum and relative permittivity, respectively. In accordance

with the previous report,¹ the data were smoothened with the Savitzky–Golay processing by a second order polynomial regression and the same ratio of data points. [0080] Note 3: Further discussion on the low-dimensional phase formation. Surface-sensitive GIXRD (ω = 0.2°, penetration depth <60 nm) revealed that only OAI and OABr treatments form 2D perovskites. OABF₄, OATFA, and OATsO all did not form 2D perovskites (FIGS. 23A-23F), likely due to steric hindrance and the inability of asymmetrical counter-anions to be shaped into the ordered PbX₆ ^4- octahedrons. We note that the observed negative ΔE_vac and consequent charge obstruction/accumulation is separate and likely unrelated from 2D formation: (1) OABF₄ and OATFA do not form 2D perovskites, but their ΔE_vac is still negative; (2) 2D perovskites are known to have low conductivity. Fundamentally, the low conductivity is because of the large organic cation (e.g. insulating alkyl group of alkylammonium), not the iodide or bromide anion. In this work, the organic cation is fixed as octylammonium; (3) Relatedly, halide orbitals contribute density of states to the perovskite band edge structure. Therefore, iodide or bromide actually participate in charge conduction; (4) It is actually possible for 2D perovskites to form a type II band alignment, which does not obstruct charge extraction. Therefore, 2D perovskite, by itself, cannot be used to explain the charge obstruction and accumulation observations; (5) Directly comparing OABF₄ and OATsO allows us to exclude any possible effects of 2D formation. [0081] Note 4: Further discussion on the correlation between ΔE_vac, charge extraction, and device performance. Based on FIG. 3E and FIGS. 10A-10D, we observed that the varying performance of the surface-passivated devices is mainly contributed by their differing FFs. A reduced FF is indicative of less efficient charge extraction. We postulate that this is related to ΔE_vac as follows: [0082] Firstly, the negative ΔE_vac deepened the valence band offset (Table 1) to worsen the energy level mismatch with spiro-MeOTAD (HOMO level: -5.2 eV). This sacrifices FF by reducing hole extraction into spiro-MeOTAD, which is observed as an increased charge extraction lifetime from the PL results (FIGS. 3C, 3D). Similar phenomenon has been observed in published works that studied the charge extraction time after surface treatment by conventional alkylammonium salts paired with iodide or bromide as the counteranion.^2,3 Secondly, the negative ΔE_vac created the heterointerface energy barrier seen from the KPFM device profiling, which also contributed to impede charge extraction to reduce FF.^1,4,5 ΔE_vac progressively became less negative going from OAI treatment (-0.37 eV), to OABF₄ treatment (-0.22 eV), and to OATsO treatment (+0.07 eV). This minimizes the inefficient charge extraction trade-off, therefore maximizing the FF and PCE gains enabled by surface treatment. Consequently, the OATsO-treated devices have the best performance. [0083] On the other hand, we observed that the V_OC of the surface-passivated devices, while still different, are relatively less affected by a negative ΔE_vac and inefficient charge extraction. Contrary to popular belief, from the perspective of the fundamental physics of solar cell operation, series resistance is actually not directly correlated with V_OC.^6,7 Fundamentally, this is because at V=V_OC, the current flow across a solar cell, in other words the current flow through the series resistance, is exactly equal to zero (I=0).⁷ Published perovskite literature have also reported that the device FF can be more sensitive than V_OC to inefficient charge extraction and potential energy barriers,^4,5 which supports our device results. [0084] Finally, we further validated the generality of our observations and universality of our proposed strategy by applying it to a different, pure FAPbI₃ composition with DMSO replaced with NMP (fabrication details included in Methods). The results are presented in FIGS.13A-13H and Table 3. [0085] Note 5: Contrasting the device degradation under MPP aging versus OC aging. In general, we observed that all devices were less stable under OC stability test, compared to their corresponding degradation rate under MPP stability test. This is consistent with previous reports that studied PSC degradation behavior under different aging conditions.⁸ That previous study showed that PSC degradation rate increases in the trend MPP aging < short-circuit aging < OC aging (i.e. PSCs have the worst stability under OC aging). Therefore, the OC stability test can be used to represent a conservative estimate of the PSC lifetime under maximum accelerated degradation, although we note that MPP aging is more truly reflective of real working PSCs. [0086] Relatedly, a rapid “burn-in” decay stage is often observed during the first ~100 h of aging during PSC photostability tests.^8–11 Importantly, the “burn-in” decay has been shown to be more severe for OC aging, and less severe for MPP aging.⁸ This is also consistent with our own observations (FIG. 3F). Although the “burn-in” decay is less noticeable under MPP stability tests, it nevertheless still exists, where the devices decayed at a more rapid rate in the first 100 h compared to the following ~400 h (FIG.3F, Table 4). Specifically, the OABF₄-treated device performance decreased by -0.029 %/h in the first 100 h, compared to -0.014 %/h in the next 400 h. The OAI-treated device performance decreased by -0.064 %/h in the first 100 h, compared to -0.022 %/h for the next 400 h. Remarkably, the OATsO-treated devices displayed no obvious “burn-in” decay in both MPP aging and OC aging. General n-i-p devices are notoriously plagued by crippling initial “burn-in” decays,^8,12 so this improvement was already noteworthy. [0087] Note 6: Simulation of halide migration pathways by first principles. The energy barriers of intra- and extra-lattice migration of halides were predicted by the associated reaction pathway energy profiles based on nudged elastic band (NEB) and constrained energy minimization methods in VASP. For intra-lattice migration, initial and final structures consisting of one vacancy in neighboring halide sites in the 2x2x2 supercell were first generated and the structures were relaxed. We then performed a linear interpolation to generate the intermediate structures within defective initial and final structures. For extra-lattice migration, one selected halide atom on the relaxed PbX₂ terminated surface was manually moved along the surface normal direction up to ~7-8Å above the surface. The structures of this reaction pathway were also constructed by linear interpolation. For both intra- and extra-lattice migration energy profiles we used sufficiently large grid points. Energy barrier of halide migration are computed from the total energy difference between the initial (or final) state and saddle point. For the negatively charged environment, we found that energy corrections due to spurious charged defect-defect interactions in finite-size periodic cells in barrier calculations are around 0.02eV, i.e., they are negligibly small and may be excluded. [0088] Note 7: Half-device photostability tests. For the half-device stability testing, half-devices of structure ITO/SnO₂/perovskite, with or without surface treatment, were exposed to AM 1.5G illumination in a nitrogen glovebox (<100 ppm of O₂/H₂O). Periodically, a few half-devices were removed to deposit spiro-MeOTAD and electrode, to complete a full solar cell for J-V measurement. The remaining, unremoved half-devices were left untouched under AM 1.5G illumination. The recorded “time elapsed” in the figure plot refers only to the time exposed to illumination. [0089] Although the half-devices are not real and complete PSCs, this experiment allows us to study the perovskite photostability by itself when ion migration (to penetrate and chemically corrode the charge-transport material/electrode) is excluded. FIG. 29 plots the average over 4 half-devices for each condition. We found that the OAI-treated, OABF₄-treated, and OATsO-treated half-devices all had approximately similar stability. Importantly, this result suggests that the photostability of the perovskite active layer, by itself, is not altered by the surface treatments. In other words, if the negative consequences induced by ΔE_vac and ion migration are largely excluded, all surface-treated devices have comparable stability. In another perspective, this result also suggests that ΔE_vac and its effect on ion migration are the dominant and primary factors that influence the degradation of complete, real working PSCs under illumination. [0090] Note 8: Possible mechanistic origins of ΔE_vac. Ligand-induced work function changes, , is given by the relation:^13–15

where is the elementary charge, is the vacuum permittivity, A is the surface area, and 1^3–15

are two independent contributions, where respectively, is the ligand intrinsic dipole moment normal to the surface, while is related to a charge density displacement by interaction and bond formation of the ligand during chemisorption to the surface. The former is intensively reported in the perovskite community,¹⁶ while the latter has been relatively less explored.

primary mechanism by which low polarity (or non-polar) species cause a . We first computed the respective dipole moments of various species in this work (Supplementary Fig. 15, 16). The negative (i.e. more n-type) with increasing alkylammonium chain length is correlated with their successively increasing positive dipole moments. On the other hand, while both [I]- and [Br]- are non-polar, the experimentally measured for the films treated with OAI or OABr were different (FIG. 2B), hinting of a contribution

Additionally, the dipole moment of [TsO]- has an opposite sign but only half that of [OA]⁺, and thus the net dipole remains positive, which should have resulted in a negative , and thus cannot fully explain the experimental observations. Therefore, the contribution

must be taken into consideration. [0091]

is related to the charge density differential, , as a function of position ( , , ) by:

[0092] The averaged charge density displacement as a function of , , and charge displacement, , in terms of , is given by:

where is calculated from:

[0093] Here, , and are the charge densities of the surface-ligand complex, free-standing surface, and free ligand, respectively. Subsequently, the charge displacement, , near the surface positioned at is then calculated by:

[0094] We computed the with the different ligands to explore the contribution (FIGS. 5C-5E). Strikingly, [TsO]- induces a pronounced negative surface charge displacement, , of -0.022 e, three-fold larger in magnitude than the positive of [OA]⁺ of +0.006 e_. Calculations with varying ligand surface density corroborate the trends (FIGS. 33A-33F). We did further calculations by introducing iodine vacancy (V_I) or lead-iodine antisite (Pb_I) defects to the surfaces of the slab models. Both defects have been reported to form readily on the perovskite surface.^17,18 We present our results in FIGS. 32A-32F and Table 7, which show that the charge displacement trends remained the same even on defected surfaces with V_I or Pb_I. Specifically, [OA]⁺ still induces a positive charge displacement (+ even with either V_I or Pb_I present on the surface. On the other hand, [TsO]- continues to induce a negative charge displacement (- that is larger in magnitude than the + of [OA]⁺. Considering the results altogether, we therefore speculate that the contributions from both -

for [TsO] synergistically counterbalanced the otherwise negative work function change induced by [OA]⁺. Consistent with this reasoning, for the other counter-anions, although the magnitudes of their negative are also larger than that of [OA]⁺ (FIGS.33A-33F, Table 6), their dipole moment contributions are however less than that of [TsO]-, and therefore insufficient to negate the net negative . [0095] Note 9: Pairing tosylate anion with alkylammoniums of different chain lengths. We compared the performances of devices surface treated with either BATsO (BA: butylammonium), HATsO (HA: hexylammonium), OATsO (OA: octylammonium), and DATsO (DA: dodecylammonium). The alkyl chain length increases in the order: BA (4 carbons) < HA (6 carbons) < OA (8 carbons) < DA (12 carbons). In general, the device V_OC increases with longer alkyl chain length, whereas the FF decreases. As a result, the OATsO-treated devices showed the best performance, by achieving a balance between high V_OC and FF (FIGS. 34A-34D). We speculate that the V_OC trend is related to the cation dipole moments: the positive dipole moment increases with longer alkyl chain length (as we also calculated in FIGS.30A-30C), which enhances the cation’s interaction with traps to further reduce charge recombination at the heterointerface.¹⁶ On the other hand, the FF seems to decrease with longer alkyl chain length, due to the insulating nature of the alkyl group to increase series resistance. A similar device performance trend with alkyl chain length was also observed for iodide-based salts in a previous work.³ [0096] Note 10: Chemical structure and purity of the ammonium salts. We used ¹H NMR (400 MHz, DMSO-d₆) to characterize the structure of our synthesized ammonium salt. We compared the chemical shift of the acid and amine reactants with the neutral salt product, shown in FIG. 35. The reaction can be simply comprehended as a generic acid-base neutralization for organic species. The peaks of the reactants, p- Toluenesulfonic acid and octylamine (red and blue, respectively), remain almost unchanged in the salt product (black), except the proton on the sulfonyl group. Moreover, all peaks in the product shifted to lower-field regions compared to their corresponding original peaks. The chemical shift is particularly evident on proton e and f. Considering that proton e and f are closest to the ionic bond between the ammonium cation and tosylate anion, they are most susceptible to the inductive effect, and therefore exhibited the most evident variation of chemical shift. Since we mixed the reactants stoichiometrically, and the reaction is rapid and complete, very few reactants remained in the formed salt, as observed in the ¹H NMR spectra. The purity of the product is thus very high. [0097] References for Notes 1. Bergmann, V. W. et al. Local Time-Dependent Charging in a Perovskite Solar Cell. ACS Appl. Mater. Interfaces 8, 19402–19409 (2016). 2. Yoo, J. J. et al. An interface stabilized perovskite solar cell with high stabilized efficiency and low voltage loss. Energy Environ. Sci.12, 2192–2199 (2019). 3. Kim, H. et al. Optimal Interfacial Engineering with Different Length of Alkylammonium Halide for Efficient and Stable Perovskite Solar Cells. Adv. Energy Mater.9, 1902740 (2019). 4. Cai, M. et al. Control of Electrical Potential Distribution for High-Performance Perovskite Solar Cells. Joule 2, 296–306 (2018). 5. Tsai, H. et al. Design principles for electronic charge transport in solution- processed vertically stacked 2D perovskite quantum wells. Nat. Commun.9, 2130 (2018). 6. Green, M. A. Solar Cells: Operating Principles, Technology, and System Applications. Prentice Hall Series in Solid State Physical Eelctronics (1981). 7. PVEducation. Solar Cell Operation: Series Resistance. https://www.pveducation.org/pvcdrom/solar-cell-operation/series-resistance. (accessed 01/20/2022). 8. Domanski, K., Alharbi, E. A., Hagfeldt, A., Grätzel, M. & Tress, W. Systematic investigation of the impact of operation conditions on the degradation behaviour of perovskite solar cells. Nat. Energy 3, 61–67 (2018). 9. Tan, S. et al. Steric Impediment of Ion Migration Contributes to Improved Operational Stability of Perovskite Solar Cells. Adv. Mater.32, 1906995 (2020). 10. Wang, Z. et al. Efficient ambient-air-stable solar cells with 2D–3D heterostructured butylammonium-caesium-formamidinium lead halide perovskites. Nat. Energy 2, 17135 (2017). 11. Bai, S. et al. Planar perovskite solar cells with long-term stability using ionic liquid additives. Nature 571, 245–250 (2019). 12. Domanski, K. et al. Migration of cations induces reversible performance losses over day/night cycling in perovskite solar cells. Energy Environ. Sci.10, 604–613 (2017). 13. Brown, P. R. et al. Energy Level Modification in Lead Sulfide Quantum Dot Thin Films through Ligand Exchange. ACS Nano 8, 5863–5872 (2014). 14. Heimel, G., Rissner, F. & Zojer, E. Modeling the Electronic Properties of π- Conjugated Self-Assembled Monolayers. Adv. Mater.22, 2494–2513 (2010). 15. de Boer, B., Hadipour, A., Mandoc, M. M., van Woudenbergh, T. & Blom, P. W. M. Tuning of Metal Work Functions with Self-Assembled Monolayers. Adv. Mater.17, 621–625 (2005). 16. Lu, J. et al. Interfacial benzenethiol modification facilitates charge transfer and improves stability of cm-sized metal halide perovskite solar cells with up to 20% efficiency. Energy Environ. Sci.11, 1880–1889 (2018). 17. Wu, W.-Q. et al. Reducing Surface Halide Deficiency for Efficient and Stable Iodide-Based Perovskite Solar Cells. J. Am. Chem. Soc.142, 3989–3996 (2020). 18. Wang, R. et al. Constructive molecular configurations for surface-defect passivation of perovskite photovoltaics. Science 366, 1509–1513 (2019). 19. Guo, Y. et al. Phenylalkylammonium passivation enables perovskite light emitting diodes with record high-radiance operational lifetime: the chain length matters. Nat. Commun.12, 644 (2021). 20. Liu, K. et al. Zwitterionic-Surfactant-Assisted Room-Temperature Coating of Efficient Perovskite Solar Cells. Joule 4, 2404–2425 (2020). 21. Jiang, C.-S. et al. Carrier separation and transport in perovskite solar cells studied by nanometre-scale profiling of electrical potential. Nat. Commun.6, 8397 (2015). Methods Perovskite film fabrication [0098] All materials were purchased from Sigma-Aldrich, unless otherwise stated. For the (FAPbI₃)_0.95(MAPbBr₃)_0.05 composition, 889 mg mL⁻¹ FAPbI₃ (FAI: GreatCell Solar, PbI₂: TCI America), 33 mg mL⁻¹ MAPbBr₃ (MABr: 1-Material, PbBr₂: Alfa Aesar Chemicals), and 33 mg ml⁻¹ MACl (GreatCell Solar) were dissolved in a dimethylsulfoxide (DMSO)/dimethylformamide (DMF) mixed solvent (1:8 v/v). The films were deposited at 4000 rpm for 20 s. After 10 s, 0.2 mL of diethyl ether was dropped on the film. The film was then annealed for 10 min at 150 °C. For the FAPbI₃ composition, 1266 mg FAPbI₃ and 34 mg MACl were dissolved in 192.8 μL N- methylpyrrolidone (NMP) and 1 mL DMF. The films were deposited at 4000 rpm for 20 s. After 10 s, 0.2 mL of diethyl ether was dropped on the film. The film was then annealed for 5 min at 100 °C, and subsequently 10 min at 150 °C. Subsequently, for post- fabrication surface treatment, 10 mM of the respective ammonium salts were dissolved in isopropyl alcohol, followed by deposition at 5000 rpm for 30s. Only OABr was dissolved in chloroform. All treatments, except with PEAI, were annealed at 100 °C for 5 min. No annealing was done for PEAI. The reference films without post-treatment were washed with pure isopropyl alcohol at 5000 rpm, followed by 100 °C annealing for 5 min. Device fabrication [0099] Substrates of indium tin oxide (ITO) on glass was cleaned by successive ultrasonication in detergent, deionized water, acetone, and isopropyl alcohol, for 20 min each. Cleaned substrates were treated by ultraviolet ozone for 20 min. For the electron- transporting material, SnO₂ colloidal solution (Alfa Aesar Chemicals) was diluted in water in 1:5 ratio. The solution was spun at 3000 rpm for 30s, and the film was subsequently annealed for 35 min at 165 °C. 10 mM potassium hydroxide or potassium chloride in water was spun at 3000 rpm for 30s, and subsequently annealed at 100 °C for 10 min. We note that potassium chloride treatment was slightly improved over potassium hydroxide. The perovskite film fabrication and post-treatment procedures were performed as described above. Subsequently, 60.0 mg of spiro-MeOTAD (p-OLED Corp), 25.5 μl of 4-tert-butylpyridine, 15.5 μl of Li-TFSI (520 mg mL⁻¹ in acetonitrile), and 12.5 μl of FK209 (p-OLED Corp) (375 mg mL⁻¹ in acetonitrile) were dissolved in 0.7 mL of chlorobenzene. The solution was spun at 3000 rpm for 30s. For the top electrode, 100 nm gold was thermally deposited at an evaporation rate of 0.5 A s^-1. The device active area is determined by a shadow mask to be 0.13 cm². For the anti-reflection coating, 150 nm MgF₂ was thermally deposited at an evaporation rate of 1 A s^-1. MgF₂ was deposited on all devices, unless otherwise stated. Atomic Force Microscopy and Kelvin Probe Force Microscopy [00100] To obtain the spatial variation of work function for each perovskite sample, we performed Kelvin Probe Force Microscopy (KPFM), which is an Atomic Force Microscopy (AFM) based technique to acquire work function information of the sample surface via measuring the contact potential difference for every pixel. For cross- sectional KPFM, the devices were cleaved simply to expose their cross-sections. A map of contact potential difference, or the relative surface potential of a sample to that of a biased AFM tip, is simultaneously obtained with a topographic image of the sample surface. The work function of the AFM tip is calibrated before and after every measurement with Highly Ordered Pyrolytic Graphite, whose work function is well- known to be 4.6 eV. The real work function value of each perovskite sample is calculated from the measured surface potential of each sample, with respect to the tip work function. The temperature and humidity near the AFM were recorded before every measurement. [00101] The AFM tips used for the KPFM measurement are Au-coated NSC36/Cr- Au tips (MikroMasch Co.) since the measurement of electric properties requires at least one conductive component. Also, each KPFM measurement involved the electrically grounded perovskite samples with conductive sample holders, through which the surface potential of any sample may have the same zero point. All AFM-based measurements were conducted with a commercial AFM NX-10 (Park Systems Corp.). Scanning Transmission Electron Microscopy and X-ray Energy Dispersive Spectroscopy [00102] The Scanning Transmission Electron Microscopy (STEM) images and X- ray Energy Dispersive Spectroscopy (EDX) maps were taken using a JEOL 2800 S/TEM equipped with dual 100 mm² Silicon Drift Detectors (SDD) at 200 kV with a probe size of 1 nm. To perform the STEM analyses on the devices, the device cross-section was lifted out and mounted on a Cu grid using a Tescan GAIA3 SEM/FIB microscope. The sample was protected by a 2 μm thick Pt layer to prevent Ga implantation during the milling procedure. The sample was successively polished, first using 30 kV 600pA, then subsequently at 15kV 150pA, and finally at 3kV 90pA, until reaching electron- transparent thickness. First-principles Density Functional Theory computation [00103] All bulk and slab first principles calculations were performed using density functional theory (DFT) in the plane-wave/pseudopotential approach implemented in the VASP package.²⁹ A revised Perdew-Burke-Ernzerhof generalized gradient approximation (PBEsol) was used for the exchange-correlation functional.³⁰ We included dispersion corrections to the total energy using Grimme's DFT-D3 scheme.³¹ Projector-augmented- wave (PAW) pseudopotentials were used to describe valence-core interactions.³² Plane- wave expansions with kinetic energies up to 300 eV were chosen. Both ionic positions and supercell dimensions were allowed to relax using conjugate gradient algorithm until all residual forces were smaller than 0.02 eV/Å. Bulk 4x4x4 and slab 4x4x1 Γ-center k- point mesh were adopted for Brillouin-zone sampling. For bulk calculations 2x2x2 supercells were used. For surface energy calculations, the 2x2xL surfaces were formed along (001) by periodic slabs including 9 to 11 atomic layers for a surface separated by 10-15 Å of vacuum. [00104] For the electrostatic potential maps, all structures were pre-optimized with very tight GFN2-xTB method using xTB program (version 6.2)^33,34 prior to conformational search. Conformational analysis of each structure was performed in water using metadynamic sampling in extended tight binding Conformer–Rotamer Ensemble Sampling Tool (xtb CREST) program package (version 6.2).³⁵ iMTD-GC workflow was used for conformational search algorithm with 6 kcal/mol energy and 0.5 Å RMSD thresholds at 298.15 K. [00105] All further density functional theory (DFT) calculations of lowest conformers were performed using Gaussian16. Geometry optimization and frequency calculations were completed at ωB97X-D/6-311++G(d,p) level in the gas phase. Optimized structures were verified by frequency calculations as minima (zero imaginary frequencies). ESP charges and maps were computed with the B3LYP method, GD3BJ³⁶ empirical dispersion and LANL2DZ basis set for iodide and aug-cc-pvtz basis set for all other atoms based on the Hirshfeld population analysis.³⁷ Dipoles were calculated with ωB97X-D/6-31+G(d) method and basis set. ESP maps were generated using GaussView 6.0.16. Preparation of ammonium salts [00106] p-Toluenesulfonic acid monohydrate was first dehydrated to remove the water of crystallization. Using 100 ml toluene dissolves 1 g p-Toluenesulfonic acid monohydrate in a 250 ml two-neck round-bottom flask. Vacuum distillation of toluene (~20 mmHg is a sufficient vacuum to lower the boiling point of toluene to a reasonable value) was followed to obtain solid p-Toluenesulfonic acid using a vacuum oven at 100 ^oC, 20 mmHg, for one week get solid p-Toluenesulfonic acid. For the anion substitution of alkylammonium salts, alkylamine (e.g. octylamine) was added stoichiometrically to the acid (e.g. p-Toluenesulfonic acid) in an isopropyl alcohol mixed solvent, followed by stirring at 50 °C overnight. The chemical structure and purity of the salt was further verified by ¹H NMR (Note 10, FIG. 35). Typically, a concentrated stock solution (>500 mM) is first prepared, followed by solution dilution to reach the final 10 mM concentration. Trifluoroacetic acid (>99.5%) was obtained from EMD Millipore Corp. Octylammonium iodide and octylammonium bromide were purchased commercially from GreatCell Solar. Device encapsulation and stability testing [00107] Device encapsulation was done inside a nitrogen-filled glovebox (<0.6 ppm of O₂/H₂O) by using a UV-curable adhesive (Nagase America LLC.) applied to a custom designed cover glass (AMG Korea). The glovebox is completely absent of any chemical solvents at all to ensure the most pristine atmosphere. The devices were kept in the glovebox for at least 2 h before and after encapsulation. The cover glass was superimposed on the active layer and fixed in position with the adhesive. This is then exposed to UV illumination for 2 min to cure the adhesive and seal the cover glass to the device. Care was taken to block the device active layer area from the UV illumination, by superimposing black electrical tape on the cover glass area that corresponded with the device active layer area. We note that in our experience, it was important to pay attention to the date of expiry of the UV-curable adhesive, with fresh adhesives preferred. [00108] For the maximum power point stability testing, encapsulated devices were exposed to continuous illumination by a simulated AM 1.5G spectrum illumination from a solar simulator. The devices were biased with a voltage corresponding to their maximum power point. The aging atmosphere is in open ambient air (RH 40 ± 10 %). For the open-circuit stability testing, encapsulated devices were exposed to continuous illumination by a metal halogen lamp source (90 ± 10 mW cm^-2). The devices were placed in an in-house built aging chamber under open-circuit condition. The chamber atmosphere is in open ambient air (RH 40 ± 10 %). Periodically, the devices are transferred to a simulated AM 1.5G spectrum illumination from a solar simulator, also in ambient air (RH 40 ± 10 %), to measure their performance. The devices are immediately returned to the aging chamber upon measurement completion. Material and device characterization [00109] SEM was done using a FEI Nova NanoLab 600 DualBeam (FIB/SEM) instrument in secondary electron mode. The films were coated with a ~1 nm-thick gold layer by sputtering to prevent the charging during the measurement. XPS measurements were carried out on an XPS AXIS Ultra DLD (Kratos Analytical). An Al Kα (1,486.6 eV) X-ray was used as the excitation source. UPS measurements were carried out using a He discharge lamp, emitting ultraviolet energy at 21.2 eV. All UPS measurements were performed using standard procedures with a −9 V bias applied between the samples and detectors. PL spectroscopy was performed with a Horiba Jobin Yvon system, with a 532 nm xenon lamp or a 640 nm monochromatic laser excitation source. TRPL spectra were measured using a Picoharp 300 with time-correlated single-photon counting capabilities (PLD 800B, PicoQuant) with a repetition frequency of 100 kHz. The confocal PL maps were measured using a Leica Confocal SP8-STED/FLIM/FCS confocal laser scanning microscope, using a HC PL APO oil objective (40 ×/1.40) and a 514 nm argon pulsed diode laser. XRD was performed by a X-ray PANalytical diffractometer at a scan rate of 4° min^-1 with Cu Kα radiation source. The simulated AM 1.5G 1-sun spectrum illumination (100 mW cm⁻²) was from an Oriel Sol3A class AAA solar simulator (Newport). The light intensity was first calibrated with a NREL-certified Si photodiode with a KG-5 filter. A Keithley 2401 source meter was used to perform the current density-voltage device measurements. A 0.100 cm² sized metal aperture was used to precisely define the device active area during measurement. EQE was measured with a custom designed Enlitech system under AC mode (frequency = 133 Hz) without light bias with a lock-in amplifier with a current preamplifier under short-circuit condition. The photo-transient measurements were done with a pulsed red dye laser (Rhodamine 6G, 590 nm) pumped by a nitrogen laser (LSI VSL-337ND-S) as the perturbation source. The pulse width was 4 ns at a repetition frequency of 10 Hz. The laser pulse intensity was monitored to maintain the amplitude of transient V_OC below 5 mV. Then, the voltage under open-circuit and current under short-circuit conditions were measured over a 1 MΩ and a 50 Ω resistor, and recorded on a digital S4 oscilloscope (Tektronix DPO 4104B). References 1. Lee, J.-W., Tan, S., Seok, S. Il, Yang, Y. & Park, N.-G. Rethinking the A cation in halide perovskites. Science 375, eabj1186 (2022). 2. Jung, E. H. et al. Efficient, stable and scalable perovskite solar cells using poly(3- hexylthiophene). Nature 567, 511–515 (2019). 3. Jiang, Q. et al. Surface passivation of perovskite film for efficient solar cells. Nat. Photonics 13, 460–466 (2019). 4. Yoo, J. J. et al. An interface stabilized perovskite solar cell with high stabilized efficiency and low voltage loss. Energy Environ. Sci.12, 2192–2199 (2019). 5. Ni, Z. et al. Resolving spatial and energetic distributions of trap states in metal halide perovskite solar cells. Science 367, 1352–1358 (2020). 6. Yang, Y. et al. Top and bottom surfaces limit carrier lifetime in lead iodide perovskite films. Nat. Energy 2, 16207 (2017). 7. Smith, M. B. & March, J. March’s Advanced Organic Chemistry: Reactions, Mechanisms, and Structure, 6th Edition. John Wiley & Sons, Inc. 8. Bergmann, V. W. et al. Local Time-Dependent Charging in a Perovskite Solar Cell. ACS Appl. Mater. Interfaces 8, 19402–19409 (2016). 9. Cai, M. et al. Control of Electrical Potential Distribution for High-Performance Perovskite Solar Cells. Joule 2, 296–306 (2018). 10. Söderström, T., Haug, F.-J., Terrazzoni-Daudrix, V. & Ballif, C. Optimization of amorphous silicon thin film solar cells for flexible photovoltaics. J. Appl. Phys. 103, 114509 (2008). 11. Tsai, H. et al. Design principles for electronic charge transport in solution- processed vertically stacked 2D perovskite quantum wells. Nat. Commun.9, 2130 (2018). 12. Tan, S. et al. Steric Impediment of Ion Migration Contributes to Improved Operational Stability of Perovskite Solar Cells. Adv. Mater.32, 1906995 (2020). 13. Azpiroz, J. M., Mosconi, E., Bisquert, J. & Angelis, F. De. Defect migration in methylammonium lead iodide and its role in perovskite solar cell operation. Energy Environ. Sci.8, 2118–2127 (2015). 14. Domanski, K. et al. Migration of cations induces reversible performance losses over day/night cycling in perovskite solar cells. Energy Environ. Sci.10, 604–613 (2017). 15. Bai, S. et al. Planar perovskite solar cells with long-term stability using ionic liquid additives. Nature 571, 245–250 (2019). 16. Wang, Z. et al. Efficient ambient-air-stable solar cells with 2D–3D heterostructured butylammonium-caesium-formamidinium lead halide perovskites. Nat. Energy 2, 17135 (2017). 17. Kim, M. et al. Methylammonium Chloride Induces Intermediate Phase Stabilization for Efficient Perovskite Solar Cells. Joule 3, 2179–2192 (2019). 18. Min, H. et al. Efficient, stable solar cells by using inherent bandgap of α-phase formamidinium lead iodide. Science 366, 749–753 (2019). 19. Yoo, J. J. et al. Efficient perovskite solar cells via improved carrier management. Nature 590, 587–593 (2021). 20. Tan, S. et al. Shallow Iodine Defects Accelerate the Degradation of α-Phase Formamidinium Perovskite. Joule 4, 2426–2442 (2020). 21. Zhao, J. et al. Strained hybrid perovskite thin films and their impact on the intrinsic stability of perovskite solar cells. Sci. Adv.3, eaao5616 (2017). 22. Lin, Y. et al. Excess charge-carrier induced instability of hybrid perovskites. Nat. Commun.9, 4981 (2018). 23. Carrillo, J. et al. Ionic Reactivity at Contacts and Aging of Methylammonium Lead Triiodide Perovskite Solar Cells. Adv. Energy Mater.6, 1502246 (2016). 24. Besleaga, C. et al. Iodine Migration and Degradation of Perovskite Solar Cells Enhanced by Metallic Electrodes. J. Phys. Chem. Lett.7, 5168–5175 (2016). 25. Ahn, N. et al. Trapped charge-driven degradation of perovskite solar cells. Nat. Commun.7, 13422 (2016). 26. Heimel, G., Rissner, F. & Zojer, E. Modeling the Electronic Properties of π- Conjugated Self-Assembled Monolayers. Adv. Mater.22, 2494–2513 (2010). 27. Brown, P. R. et al. Energy Level Modification in Lead Sulfide Quantum Dot Thin Films through Ligand Exchange. ACS Nano 8, 5863–5872 (2014). 28. Zerweck, U., Loppacher, C., Otto, T., Grafström, S. & Eng, L. M. Accuracy and resolution limits of Kelvin probe force microscopy. Phys. Rev. B 71, 125424 (2005). 29. Kresse, G. & Furthmuller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci.6, 15–50 (1996). 30. Perdew, J. P. et al. Restoring the Density-Gradient Expansion for Exchange in Solids and Surfaces. Phys. Rev. Lett.100, 1–4 (2008). 31. Grimme, S., Antony, J., Ehrlich, S. & Krieg, H. A consistent and accurate ab initio parametrization of density functional dispersion correction (DFT-D) for the 94 elements H-Pu. J. Chem. Phys.132, 154104 (2010). 32. Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953–17979 (1994). 33. Grimme, S., Bannwarth, C. & Shushkov, P. A Robust and Accurate Tight-Binding Quantum Chemical Method for Structures, Vibrational Frequencies, and Noncovalent Interactions of Large Molecular Systems Parametrized for All spd- Block Elements ( Z = 1–86). J. Chem. Theory Comput.13, 1989–2009 (2017). 34. Bannwarth, C., Ehlert, S. & Grimme, S. GFN2-xTB—An Accurate and Broadly Parametrized Self-Consistent Tight-Binding Quantum Chemical Method with Multipole Electrostatics and Density-Dependent Dispersion Contributions. J. Chem. Theory Comput.15, 1652–1671 (2019). 35. Grimme, S. et al. Fully Automated Quantum-Chemistry-Based Computation of Spin-Spin-Coupled Nuclear Magnetic Resonance Spectra. Angew. Chemie Int. Ed. 56, 14763–14769 (2017). 36. Grimme, S., Ehrlich, S. & Goerigk, L. Effect of the damping function in dispersion corrected density functional theory. J. Comput. Chem.32, 1456–1465 (2011). 37. Hirshfeld, F. L. Bonded-atom fragments for describing molecular charge densities. Theor. Chim. Acta 44, 129–138 (1977). [00110] The embodiments illustrated and discussed in this specification are intended only to teach those skilled in the art how to make and use the invention. In describing embodiments of the invention, specific terminology is employed for the sake of clarity. However, the invention is not intended to be limited to the specific terminology so selected. The above-described embodiments of the invention may be modified or varied, without departing from the invention, as appreciated by those skilled in the art in light of the above teachings. It is therefore to be understood that, within the scope of the claims and their equivalents, the invention may be practiced otherwise than as specifically described.

Claims

We Claim: 1. A photovoltaic device, comprising: a first electrode; a second electrode spaced apart from said first electrode; a perovskite layer between said first and second electrodes, said perovskite layer comprising a semiconducting halide perovskite photoactive material; and a vacuum-level modulating layer formed on a surface of said perovskite layer, wherein said vacuum-level modulating layer provides crystal defect passivation of said surface of said perovskite layer, and wherein said vacuum-level modulating layer further modulates surface energetics of said surface of said perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein. 2. The photovoltaic device according to claim 1, wherein said first electrode is at least partially transparent to light within an operating band of wavelengths of said photovoltaic device. 3. The photovoltaic device according to claim 2, further comprising an electron transport layer between said first electrode and said perovskite layer. 4. The photovoltaic device according to any one of claims 1-3, further comprising a hole transport layer between said first electrode and said perovskite layer. 5. The photovoltaic device according to any one of claims 1-4, wherein said vacuum-level modulating layer comprise a material comprising at least one of a tetrafluoroborate [BF4]-, a trifluoroacetate [TFA]-, a bromide [Br]-, a tosylate [TsO]-, or a triflate [TfO]- counter-anion. 6. A method of producing a photovoltaic device, comprising: forming a first electrode; forming a second electrode spaced apart from said first electrode; forming a perovskite layer between said first and second electrodes, said perovskite layer comprising a semiconducting halide perovskite photoactive material; and forming a vacuum-level modulating layer on a surface of said perovskite layer, wherein said vacuum-level modulating layer provides crystal defect passivation of said surface of said perovskite layer, and wherein said vacuum-level modulating layer further modulates the surface energetics of said surface of said perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein. 7. The method according to claim 6, wherein said first electrode is at least partially transparent to light within an operating band of wavelengths of said photovoltaic device. 8. The method according to claim 7, further comprising forming an electron transport layer between said first electrode and said perovskite layer. 9. The method according to any one of claims 6-8, further comprising forming a hole transport layer between said first electrode and said perovskite layer. 10. The method according to any one of claims 6-9, wherein said vacuum-level modulating layer comprise a material comprising at least one of a tetrafluoroborate [BF4]-, a trifluoroacetate [TFA]-, a bromide [Br]-, a tosylate [TsO]-, or a triflate [TfO]- counter-anion. 11. The method according to claim 6 or 10, wherein said vacuum-level modulating layer is formed by spin coating. 12. A method of treating a surface of a perovskite layer for use in a photovoltaic device, comprising: receiving a material for forming a vacuum-level modulating layer on a surface of said perovskite layer; and forming said vacuum-level modulating layer on said surface of said perovskite layer, wherein said vacuum-level modulating layer provides crystal defect passivation of said surface of said perovskite layer, and wherein said vacuum-level modulating layer further modulates the surface energetics of said surface of said perovskite layer to mitigate formation of a potential well so as to avoid charge accumulation therein. 13. The method according to claim 12, further comprising, prior to said receiving, selecting said material for forming said vacuum-level modulating layer on said surface of said perovskite layer, wherein said selecting is based on at least one of an effect of a surface charge modification or an effect of a diploe moment of said material on said surface of said perovskite layer. 14. The method according to claim 12, wherein said vacuum-level modulating layer comprise a material comprising at least one of a tetrafluoroborate [BF4]-, a trifluoroacetate [TFA]-, a bromide [Br]-, a tosylate [TsO]-, or a triflate [TfO]- counter- anion. 15. A perovskite layer for use in a photovoltaic device produced according to the method of any one of claims 12-14.