Defect quantification in metal halide perovskites

Abstract

Devices and electrochemical and spectroelectrochemical methods of characterizing a perovskite layer of a semiconductor device are described. One example method includes contacting the perovskite layer with a first surface of an electrolyte layer, where a second surface of the electrolyte layer is in contact with a first counter electrode. The method further includes embedding a first reference electrode in the electrolyte layer, applying a voltage to the first reference electrode, measuring a current between the first reference electrode and the first counter electrode, and characterizing the perovskite layer based on the applied voltage and the measured current between the first reference electrode the first counter electrode. Characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

Description

DEFECT QUANTIFICATION IN METAL HALIDE PEROVSKITESCROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] This application claims priority to the provisional application with serial number 63 / 584,557 titled “Defect Quantification in Metal Halide Perovskites Projecting Photoluminescence and Photovoltaic Performance,” filed September 22, 2023. The entire contents of the above noted provisional application are incorporated by reference as part of the disclosure of this document.STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH

[0002] This invention was made with government support under Grant numbers N00014-20-1 -2440 awarded by Office of Naval Research, and DE-EE0010238 awarded by the Department of Energy. The government has certain rights in the invention.TECHNICAL FIELD

[0003] Embodiments of the present invention relate to methods and systems for analyzing electronic materials and semiconductor devices.BACKGROUND

[0004] Perovskite refers to a class of materials with a specific crystal structure having the chemical formula CaTiOs. Perovskites have unique properties that make them promising materials for use in electronic devices, such as solar cells, light-emitting diodes (LEDs), and detectors. These properties include high absorption coefficient, long charge-carrier diffusion lengths, and high defect tolerance. As a result, there has been a significant increase in interest in metal halide perovskites in recent years.

[0005] However, the properties and operational stability of metal halide perovskites for use in optoelectronic devices can be influenced by interfaces and defects. Interface and defect engineering can play a role in controlling the behavior of the charge carriers andgrowing high quality, substantially defect-free perovskite crystals. While there have been developments in methods for analyzing various defects in metal halide perovskites, most of the conventional methods show low sensitivity.

[0006] To address these issues and promote stronger ties between basic and applied research, in situ and operando characterization during manufacturing and / or operation of electronic devices and / or semiconductor materials is needed. These techniques may be useful analytical tools that can follow chemical reactions, physical processes, microstructural changes, and interfacial phenomena in real-world conditions and enhance rates of commercialization of viable technologies based on these optoelectronic materials. The techniques may be also essential in predicting lifetimes of optoelectronic devices, including but not limited to solar cells. For the high-throughput production of (opto)electronic devices, state-of-the-art multi-model characterization capabilities, as well as novel inline sensors, are needed to follow (electro)chemical and / or physical processes during manufacturing operation across length and time scales, or to understand device-transport layer interactions.BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 shows (A) XRD, (B) UV-vis, (C) SSPL, (D) TRPL spectra for stoichiometric Cs0.05FA0.79MA0.i6 / NiOx / ITO (black), excess Pbh Cs0.05FA083MA0.i6 / NiOx / ITO (blue) and excess FAI Cs0.05FA0.75MA0.i6 / NiOx / ITO (red) half-stacks and (E) device performance of a stoichiometric Cs0.05FA0.79MA0 i6 PAL using a hole transport layer of NiOx nanoparticles and C60 / BCP electron transport layer.

[0008] Figure 2 shows (A) Schematic view of the spectroelectrochemical cell configuration exploiting the semi-transparent nature of the SE / IL top contact; (B) Schematic view of the consequences of electron injection into the conduction band of the PAL, leading to a lowered probability for absorbance at and above EBG; (C) Schematic of the energetics and density of states at VBM and below, or at CBM and above, including shallow donor defects just below CBM distributed in an Urbach tail. The image shows the injection of electrons into these donor effects, accompanied by thermalization to populate states atCBM, changing the effective band gap that causes "bleaching" of the initially observed band edge absorbance. (D) Absorbance A at 730 nm (versus the baseline absorbance A at 820 nm) in a potential range of -0.6 V to 1.1 V of stoichiometric CsFAMA and 5A / 5E (first derivative) of this data, fitted to a Gaussian distribution with a FWHM = ca. 0.06 eV; (E-F) Comparable plots of A versus E and 5A / 5E for the excess Pbk and excess FA CsFAMA perovskite showing the additional absorbance bleaching events below CBM due to electron injection into cation defect states. The raw data is shown in Figure 13. It should be noted that in Figure 13, the "baseline" absorbance in the near-infrared region is not the same for all three perovskite compositions, which also means the background shift will be different for all perovskite compositions. The baseline absorbance shifts to positive absorbances due to scattering effects within the gel electrolyte that change due to an increase in transmittance of the perovskite film.

[0009] Figure 3 shows (A) Schematic view of the electrochemical cell configuration where the reference electrode (Ag / AgCI) is embedded within the SE / IL top contact and stacked between a perovskite-coated ITO and gold electrode, in which 3 mM of the redox probe molecules are added to this SE / IL layer so that reduction (electron injection) processes reduce the probe (electron acceptor); (B) Schematic view of the redox potentials for the two redox probes used here, benzoquinone (BQ) and chloranil (Ch) relative to the CBM of the stoichiometric perovskite; Polarographic CVs of the (C) stoichiometric Cs.05FA.79MA.i6 and (D) excess FAI Cs.05FA.83MA.i6 as a function of cathodic potential endpoints approaching and surpassing CBM combined with (right schematic) the redox reactions of the Ch and BQ, showing the two one-electron reduction processes, well separated in potential. The Ch probe was used to assess the stoichiometric film while the BQ probe was necessary for the film with excess FAI. CVs characterizations were taken at 0.05 V s'1were carried out at room temperature in inert atmosphere (H2O < 1 ppm; O2 < 1 PPm).

[0010] Figure 4 shows a schematic view of the band edge energetics and distributions of defect energies for the stoichiometric, excess FAI and excess Pbh CsFAMA perovskites. The percentages indicate the PCE measured for the devices using the three compositionsin combination with an ITO / NiOx hole transport layer and C60 / BCP / Ag electron transport layer.

[0011] Figure 5 shows a schematic view a dual-probe device with two separate electrochemical probes, affixed to the perovskite surface, with independent control of the potential of the perovskite film with respect to2B the internal reference electrode(s), and independent measurement capability of the currents induced by changes in potential at the perovskite / sol id-electrolyte (SE)-ionic liquid (IL) interface.

[0012] Figure 6 shows a schematic view of the spectroelectrochemical determination of band edges of perovskite films. The valence band is characterized using an oxidative process by extracting holes and the conduction band is characterized using a reductive process by injection electrons.

[0013] Figure 7 shows (A) CV scans in which the end potential is shifted closer and past the valence band offset of the perovskite; (B) expanded oxidation peaks including a black line that makes the distinction between Faradaic contributions due to defect states and the non-Faradaic background; (C) a peak charge vs probe concentration log-log calibration curve; and (D) predicted defect densities.

[0014] Figure 8 shows UV-vis spectra (A), steady-state (B) and time-resolved (C) photoluminescence data, and XRD spectra (D) for stoichiometric Cs.05FA.79MA.i6, excess Pbh Cs.05FA.75MA.i6, and excess FAI Cs.05FA83MA.i6 perovskite films on an ITO / glass substrate measured in ambient conditions. The average lifetimes <t> in (C) are calculated using the photoluminescence decay fitting described in literature.

[0015] Figure 9 shows a boxplot of the short circuit currents (A), open current voltages (B), fill factors (C) and power conversion efficiencies (D) for the forward and reverse J-V characteristics of five stoichiometric Cs.05FA.79MA.i6 PSCs under AM1.5 light conditions.

[0016] Figure 10 shows (A) Forward and reverse J-V characteristics of stoichiometric Cs.05FA.79MA.i6 under AM1.5 light conditions and (B) its dark J-V characteristics on a logarithmic scale. Forward and reverse J-V characteristics of (C) excess FAI Cs.05FA.83MA.i6and (D) excess Pbh Cs.05FA.75MA.i6 PSCs under AM1.5 light conditions. The relevant performance and reproducibility parameters for all PSCs are given in the inset.

[0017] Figure 11 shows a boxplot of the short circuit currents (A-E), open current voltages (B-F), fill factors (C-G) and power conversion efficiencies (D-H) for the forward and reverse J-V characteristics of five excess Pbh Cs 05FA75MA i6 and excess FAI Cs.05FA.83MA.i6 PSCs under AM1.5 light conditions.

[0018] Figure 12 shows cyclic voltammograms from -0.5 V vs. Ag / AgCI to anodic potential endpoints of 0.3 to 0.8 V vs. Ag / AgCI followed by cathodic sweep in presence of the SE / IL electrolyte at 0.05 V s-1on stoichiometric Cs.05FA.79MA.i6 in inert atmosphere (H2O < 1 ppm; O2 < 1 ppm).

[0019] Figure 13 shows all raw UV vis spectra of (A) stoichiometric Cs.05FA.79MA.i6, (B) excess Pbh Cs 05FA75MAi6, and (C) excess FAI Cs 05FA83MA i6 perovskite films as a function of potential in the range of -0.5 V to -0.9 V and -1.1 to -1.3 V with 0.1 V increments and -0.9 to -1.1 V with 0.02 V increments taken using an electrolyte-electrode stack as shown in the inset of Figure 2, without redox probe, in ambient conditions. The green lines indicate the absorbance at 730 and 820 nm above and below the band-edge absorbance of the perovskite, respectively.

[0020] Figure 14 shows a schematic of the redox potentials of the (A) Ch / Ch'' and (B) BQ / BQ'" probe measured in the SE / IL. CVs are taken at 0.05 V s-1.

[0021] Figure 15 shows scan rate dependent CVs of the one-electron reduction of (A) Ch / Ch*" and (B) BQ / BQ*" probe in the Au / SE / IL / ITO stack structure from 0.005 to 0.2 V s-1in a potential window between 0.5 V to -0.5 V vs. Ag / AgCI. Log maximum peak current density vs. log scan rate plot to check for Randles-Sevcik behavior and calculate the diffusion coefficient of (C) Ch / Ch*" and (D) BQ / BQ*" in the SE / IL.

[0022] Figure 16 shows (A) Focused image of the oxidation process (from 0 V to potentials between -0.7 and -0.78 V vs. Ag / AgCI in Figure 3C) of stoichiometric Cs.05FA.79MA.i6, where the black lines represent the baseline of the peak considering thebackground; (B) Charge integration of donor defects near the CBM of stoichiometric Cs.05FA.79MA.i6 as a function of Eend, redox which follows a Gaussian distribution (y = A exp(- 1 / 2((x-xc) / w)2) + yo fit is added as a black line); (C) Calibration curve of the charge density of the Ch’’ oxidation process using CV for known concentrations of the Ch / CK probe in the SE / IL electrolyte on an ITO electrode. (D) Complete energy band diagram of stoichiometric Cs.05FA.79MA.i6 measured using a CV and LSV / DPV approach with ferrocene as a hole capture probe in a SE / IL a p-chloranil as an electron capture redox probe based on the calibration in (C).

[0023] Figure 17 shows a schematic of the CBM of excess FAI Cs.05FA.83MA.i6 combined with the standard reduction potentials for the redox reactions of the Ch probe and the measured CVs as a function of cathodic potential endpoints (A) approaching CBM and (B) surpassing CBM in the presence of 3 mM Ch in the SE / IL, at 0.05 V s-1followed by anodic sweeps to the initial potential. All these characterizations were carried out in inert atmosphere (H2O < 1 ppm; O2 < 1 ppm). The CVs are shifted by -0.1 V on the potential axis to compensate for an Ag / AgCI reference potential of 0.2 V (versus the Fc / Fc+redox couple).

[0024] Figure 18 shows (A) The charge integration of donor defects in excess FAI Cs.05FA.83MA.i6 perovskite films as a function of Eend, redox which follows a Gaussian distribution (y = A exp(-1 / 2((x-xc) / w)2) + yo fit is added as a black line); (B) Calibration curve of charge density of the BQ’’ oxidation process using cyclic voltammetry for known concentrations of BQ / BQ’’ probe in the SE / IL electrolyte on an ITO electrode; (C) Complete energy band diagram of excess FAI Cs.05FA.83MA.i6 in an inert atmosphere measured using a CV and LSV / DPV approach with ferrocene as a hole and 1 ,4-benzoquinone as an electron capture redox probe in a SE / IL electrolyte.

[0025] Figure 19 shows (A) CVs and (B) DPVs of the CBM of excess Pbh Cs.05FA.75MA.i6 as a function of cathodic potential endpoints from -0.4 V to -0.8 V vs. Ag / AgCI and from -0.8 V to -1.1 V vs. Ag / AgCI in the presence of 5 mM BQ in the SE / IL. CVs are taken at 0.05 V s-1followed by anodic sweeps to the initial potential. The DPV parameters are 0.001 V increment, 0.05 mV amplitude, 0.05 s pulse width, 0.005 s samplewidth, and 0.3 s pulse period. These characterizations were carried out at room temperature and in an inert atmosphere (H2O < 1 ppm; O2 < 1 ppm). (C) Charge integration of donor defects in excess Pbh Cs.05FA.75MA.i6 as a function of Eend, redox which follows a Gaussian distribution (y = A exp(-1 / 2((x-xc) / w)2) + yo fit is added as a black line) in combination with its complete energy band diagram. (D) Calibration curve of the peak area of the BQ'" oxidation process taken using an LSV / DPV approach for known concentrations of BQ / BQ-' probe in the SE / IL electrolyte on ITO.

[0026] Figure 20 illustrates a set of operations for using an electrochemical method for characterizing a perovskite layer of a semiconductor device in accordance with an example embodiment.

[0027] Figure 21 illustrates a set of operations for using a spectroelectrochemical method for characterizing a perovskite layer of a semiconductor device in accordance with another example embodiment.

[0028] Figure 22 illustrates a set of operations for using an electrochemical method for characterizing a perovskite layer of a semiconductor device in accordance with another example embodiment.DETAILED DESCRIPTION

[0029] Near-band edge defects in metal halide perovskite active layers (PALs) contribute to performance loss and instabilities in photovoltaic devices. However, many conventional metrologies are unable to quantify both near-valence maximum and nearconduction minimum (CBM) defects with energy resolution necessary to optimize semiconductor materials. Herein we demonstrate an innovative (spectro)electrochemical approach for determining the concentration and energy distribution of both donor and acceptor defects in high performing, wide-gap (Cs.o5FA.79MA.i6)Pb(l.87Br.i3)3-based photovoltaics, benchmarked against time-resolved photoluminescence and device performance. The present disclosure presents limits-of-detection of defects near CBM and VBM of 1015cm-3and 1014cm-3, respectively, with an energy resolution of ca. 5-10 meV,well below that of conventional spectroscopic and electrical characterization methods. The approach uses semi-transparent solid electrolyte / ionic liquid (SE / IL) (a combination of solid electrolyte and ionic liquid) top contacts to characterize both halide migration and undercoordinated lead defects under device relevant bias. Addition of electron acceptors or electron donors (i.e., redox probes) to the SE / IL enables coulometric quantification with energy resolution.

[0030] As disclosed herein, the sensitivity of the approach to detect defects arising from small modifications to the A-site cation (formamidimium (FA+)) concentration (<2%), mimicking compositional imperfections at nanometer length scales arising from nonoptimized processing and / or because operational / environmental stress. Overall, the disclosed methodology provides the first complete near-band edge energetic diagram of PALs as a function of small changes in composition. The present technique is nondestructive, compatible with characterization during processing, and will enable the semiconductor community to link molecular origins of defects with limitations in device performance for a wide array of opto-electronic platforms.

[0031] The methods, systems, and materials for electrochemical and spectroelectrochemical characterization of perovskite in semiconductor devices (e.g., in solar cells), are described in further detail below.

[0032] Solution processed, double and triple cation / mixed halide perovskite (single junction and tandem) solar cells (PSCs) consistently demonstrate competitive power conversion efficiencies (PCEs) with conventional semiconductor materials, but commercially competitive stabilities have not yet been released to make perovskites a truly disruptive solar market technology. The ease of (solution) processing of perovskite materials is directly coupled to the formation of defects originating at grain boundaries, interfaces with electrical contacts, and / or in regions where nucleation / growth leaves nanometer-scale regions with small departures from stoichiometry. Many of these energetically shallow defect states - distributed close to the conduction band minimum (CBM) and the valence band maximum (VBM) - were initially believed to be electronically benign. Yet there is an emergingunderstanding that many defect states are (electro)chemically reactive, with redox activity manifesting when the perovskite material is under "stress," including device relevant (operando) electric fields. The impact of defects is exacerbated by additional external stressors include intense illumination, high temperatures, and oxygen and / or water-rich environments.

[0033] As the reactivity of defects has been identified as the primary reason for longterm instabilities of perovskite semiconductors, promoting long-term device stability requires a routine but selective characterization of defects, including their type, energetic distribution, and concentration. The chemical speciation of defects in triple cation perovskites has been challenging, with defect states most often characterized using non-selective electrical, optical, and magnetic techniques. Defect densities have been estimated to be in the range of 1015- 1016cm-3for (initially) high performing polycrystalline thin film (400-500 nm) active layers, and as low as 1011cm-3for single crystal perovskites. The ideal approach would be amenable to in-line processing integration of the active and transport layers of the device stack as well as compatible with accelerated degradation testing.

[0034] Here, we demonstrate the combined use of spectroscopy approaches and electrochemistry to characterize the energetic distribution of donor defect states in devicerelevant, "stoichiometric" triple cation, mixed halide perovskite thin films (Cs0.05FA0.79 MAo.i6)Pb(lo.87Bro.i3)3) or CsFAMA at high energy resolution (< 10 meV) near CBM. The stoichiometric perovskites demonstrate spectroscopic properties, including photoluminescence (PL) lifetimes, which make them competitive with active layers known to produce excellent photovoltaic (PV) performance in freshly prepared CsFAMA thin films, and we consistently create PV platforms with power-conversion-efficiencies (PCE) exceeding 17-18%.

[0035] In some embodiments, our (spectro)electrochemical characterization platform uses a solid electrolyte / ionic liquid (SE / IL), with embedded reference and counter electrodes as the top contact. This semi-transparent, stick and peel (non-destructive) electrolyte enables device-relevant electric fields (i.e., electrochemical biasing) across the perovskiteactive layer (PAL) so that reactive defects are detected as a function of their energetic position with respect to the band maxima. To demonstrate the further sensitivity of our approach, we benchmark the characterization to include PALs with small (<2%) deviations from stoichiometry (enrichment or depletion of the major A-site cation) that might occur because of non-optimized processing at nanometer length scales and / or environmental or electrical stress on the PAL. Only small changes in EBG are seen with changes in stoichiometry, but the energetic distribution of cation and anion defect states is altered significantly so that PL lifetimes and PV performance are impacted. For cation defect quantification near the CBM, we use two different quinone-based electron acceptor redox probes (with different first one-electron reduction potentials) in the SE / IL top contact. For anion defect quantification, an electron donor probe (ferrocene) has been used. For the CBM analysis, voltametric responses using the quinone probes provide energetic distributions of mid-gap defects, well below CBM for excess FAI perovskites, and shallow cation defects lying within 100-200 meV of CBM for both stoichiometric and excess Pbk PALs. Defect concentrations are quantified presently down to ca. 1015cm’3, with an anticipated limit of detection of ca. 1014crrr3. Such detection limits that are better than spectroscopic, electronic and photoemission protocols, with speciation (anion versus cation defects) not available in those other approaches. This characterization platform provides a powerful general approach to do sensitive and selective in-line (operando) characterization of reactive defects in PALs and is one of the few ways to quantify low defect concentrations in top performing PSCs which ultimately control their long-term stability.Example Results and DiscussionBasic Properties of Three Perovskite Film Compositions and Their PV Performance

[0036] We start with the understanding that the final ABXs perovskite is created by insertion of A-site cations into a solvent adduct of a Pbk nanocrystal, forming a colloidal dispersion, followed by nucleation and growth of the final crystalline film. The final product state represents a free energy minimum that can be difficult to reach if the stoichiometries of the reaction intermediates are not locally sustained. To show the consequences of suchdepartures, we create a prototypical wide-gap material (Cs0.05FA0.79 MAo.i6)Pb(lo.87Bro.i3)3) or CsFAMA, accompanied by a "Pbh-rich" active layer, created by slightly lowering the formamidimium iodide (FAI) content in the precursor solution and an "FAI-rich" active layer, created by raising the precursor FAI content.

[0037] Figure 1 provides a summary of the characterization of perovskite / NiOx / ITO stacks for the three different compositions; complementary data for the perovskite / ITO stacks of three perovskites used for the electrochemical characterization is shown in Figure 8. Figure 1 A provides comparative X-ray diffraction (XRD) patterns for three formulations of CsFAMA based on 2% excess Pbh (Cso.o5FAo.75MAo i6)Pb(lo.87Bro.i3)3) and 2% excess FAI (Cso.o5FAo83MAo.i6)Pb(lo.87Bro.i3)3; procedures for their fabrication are included in the Experimental Procedures Section herein. In the XRD patterns, the diffraction peak at 12.7 degrees evidentiary shows the overall concentration of unreacted crystalline Pbl2 in the perovskite film is dependent on the initial precursor solution composition. Briefly, the stoichiometric Cs0.05FA0.79MA0.i6 film reveals small concentrations of unreacted Pbh, with domains sizes that meet or exceed the X-ray scattering coherence length requirement, the presence of which is known to decrease nonradiative recombination rates and increase carrier lifetimes. The off-stoichiometry perovskite films either present even higher Pbh concentrations (excess Pbh) or no detectable Pbh (excess FAI). For the latter, this suggests any residual Pbk has been incorporated into the ABX3 lattice, and thus we would posit that unreacted FA+and h may alter defect types, densities, and energetic distributions in the bandgap region and near the band maxima. Figures 1 B and 1 C respectively show the comparative band edge absorption (EBG) and steady-state photoluminescence (SSPL). In Figure 1 B, the band edge absorption (EBG = 1.62 eV) undergoes only small shifts between the three PAL compositions. However, the steady-state photoluminescence (SSPL) shows at least a factor of 10x larger PL yield for the stoichiometric Cs0.05FA079MA0.i6 versus the excess FAI Cs005FA0.83MA0.i6. Representative plots of the time-resolved photoluminescence (TRPL) are provided in Figure 1 D with average lifetimes <T> provided, for which the order of magnitude for the stoichiometric Cs0.05FA079MA0.i6 is consistent with previous reports, to assess spatial heterogeneous sites of recombination. We note that complementary datashown in Figure 8for ITO / perovskite samples show similar trends between perovskite formulations but with overall lower carrier lifetimes, presumed to be due to surface-assisted recombination effects of the ITO substrate. The stoichiometric Cs0.05FA0.79MA0.i6 exhibits carrier lifetimes in excess of 250 ns on NiOx and over 200 ns even on ITO, which is consistent with the low recombination probabilities, high fill factors and large Voc seen in top-performing photovoltaic devices. Both non-stoichiometric systems show significantly lower time constants, especially on ITO, indicating higher recombination rates and increased defect densities relative to the stoichiometric Cs0.05FA0.79MA0.i6 system.

[0038] Being able to create consistently high performance in PV platforms with PALs that (spectro)electrochemically demonstrate low defect densities is critical in validating our characterization protocols. Stoichiometric Cs0.05FA0.75MA0.i6 PALs show consistently good performance in Figure 1 E among multiple devices and experimental trials, with Voc > 1 V, Jsc > 23 mA cm-2, FF > 0.75 and PCE = ca. 17-19% under AM1.5 illumination. Device construction details and the reproducibility in values for Jsc, l / oc, FF and PCE are discussed in the Experimental Procedures Section herein and in Figure 9; the dark JA / characteristics are added for reference in Figures 10A and 10B. We readily acknowledge this is not state of the art for this PAL composition, which has demonstrated over 22%. Our PV platforms use non-optimized ITO / NiOx nanoparticle thin films as a hole-selective interlayer in an attempt to enhance the PV efficiency and reproducibility. (Spectro)electrochemical characterization of fully optimized p-i-n and n-i-p platforms will be published elsewhere. Additionally, we have not focused on modification of PALs to mitigate any specific type of effect (e.g., remove under-coordinated Pb ionic sites); PL lifetimes that can extend to greater than 1 ps with careful optimization of multiple interfaces have been reported. The illuminated JV curves for device stacks using the excess Pbl2 and excess FAI PALs are shown in Figures 10C and 10D, along with their reproducibility in Figure 11. Collectively these three examples clearly indicate that PL lifetime alone is not the only indicator for predicting PV device performance. Interestingly, the excess Pbh PALs, despite their considerably shorter PL lifetimes, show only a small decrease of overall PV efficiency, with comparable Voc, Jsc, FF and PCE. Alternatively, the excess FAI PV devices, despite PL lifetimes on NiOx over100 ns, exhibit significantly diminished PV performance, with losses in Voc, significantly lower Jsc, FF, and PCE = 4-5%. For these FAI PV devices, we also noted significant hysteresis in the JV behavior consistent with ion displacement and electrochemical reactions accompanying electronic conduction in the PAL.(Spectro)electrochemical Characterization of Energetic Distributions

[0039] Energetic distributions in semiconductor materials are characterized via JV (or voltametric) behavior based on the knowledge that Faradaic currents, resolved from background capacitive (ion displacement) currents, represent electron injection into available states. Cyclic voltametric scans are shown in Figure 12, with no redox couples in the SE / IL. For the top performing stoichiometric PAL, only capacitive (ion displacement) currents are observed in the potential region that lies between VBM and CBM, and currents exponentially increase as potentials are applied past the band maxima. This is analogous to a shift of the surface Fermi level into the conduction (CB) or valence (VB) band regions, associated with redox processes of the PAL, respectively.

[0040] Spectroelectrochemical approaches, including electroabsorption studies, monitor changes in spectroscopic signatures that accompany electrochemical processes, often with greater sensitivity towards Faradaic processes. For conjugated polymers, molecular (non-fullerene) electron acceptors, and perovskite thin films that electron injection into the CB and the VB results in loss of band edge absorbance. Ultimately, more cathodic (negative) or anodic (positive) excursions past the band edge onset produce currents and absorbance changes that identify these band edge energies. This electro-bleaching effect can be used as a sensitive indicator for measuring band energies and as a first order estimation of the energetic distribution of reactive defect states just below bands, assuming these defects are optically active. This approach also eliminates difficulty in differentiation between Faradaic currents and non-Faradaic currents, the latter being predominantly ionic motion in the electrolyte and / or perovskite.

[0041] Figure 2A shows schematic views of the transmission spectroelectrochemical characterization of PALs made possible by the semi-transparent nature of the SE / ILelectrolyte top contact. We note that current (spectro)electrochemical studies are performed on perovskite / ITO interfaces, so as to directly compare with the defect quantifications described below. Figure 2B shows schematically the origin of the absorbance bleaching process at increasingly negative potentials where electrons are injected into the CBM region lowering the absorptivity at and above EBG. Due to dynamic disorder induced by near-edge defects in the PAL, the density of states (DOS) extends into the band gap following an exponential behavior described as the Urbach tail in Figure 2C. The optical absorption associated with this tail, however, is proportional to the joint DOS that describes the allowed optical transitions between occupied and unoccupied states. When electrons are injected into defect states, unoccupied states are slowly filled with electrons so that the allowed optical transitions shift to higher photon energies as seen in Figure 2C so that the principal optical transition becomes less dominant and bleaches accordingly. Figures 2D-2F provide both the difference in the absorbance edge AA (at EBG, 765 nm, •, top axis) and the first derivative of that behavior, 6A / 6E (■, bottom axis) as a function of potential (left axis) and wrt the vacuum level (right axis) for each perovskite composition. Fits are included to guide the eye, where the Gaussian centroid of the 5A / 5E defines CBM. For the stoichiometric perovskite in Figure 2D, the potential-dependent absorbance bleach is centered around - 3.74 eV versus vacuum (-0.96 V vs. Ag / AgCI) with a FWHM of 0.061 ± 0.00(4) eV. For the excess FAI film (Ecenter = -3.69 eV; FWHM = 0.039 ± 0.00(4) eV) in Figure 2E, the 5A / 6E is significantly broader and shifted toward vacuum suggesting a significant difference in CBM, while the Gaussian fit to 5A / 5E for the excess Pbh PAL (Ecenter = -3.71 eV; FWHM = 0.053 ± 0.00(3) eV) in Figure 2F is comparable to that seen for the stochiometric,Cso.o5FAo.75MAo.i6 film. Non-Faradaic processes, associated with ion displacement, are assumed to be spectroelectrochemically silent.Example Defect Quantification Using Redox Probes in the SE / IL to Predict JA / Behavior

[0042] A major challenge in defect quantification in metal halide perovskite photoactive layers is that a multitude of defect states are possible. Examples include anion vacancies and the resultant mobile anions, metal cation sites that are "under-coordinated," and avariety of anti-site and microstructural defects. Anion (acceptor) defects above the VBM have been attributed to mobile iodide within CsFAMA perovskites. Quantification of cation (donor) defects, on the other hand, requires an electron-acceptor redox probe be added to the SE / IL top contact. Briefly, the potential of the ITO / perovskite contact in Figure 3A is cathodically swept or stepped to more cathodic (negative) potentials, approaching, and then surpassing CBM. If there are available states in the perovskite film, electron injection occurs from the back contact into reactive defect states within the perovskite bandgap. Electronaccepting probes with appropriate first one-electron reduction potentials are then reduced by these electron-rich states at the perovskite / SE / IL interface. Further sweeping towards more cathodic potentials results in electron injection into CB states and yields reversible oxidation reactions on the return anodic sweep.

[0043] We strategically selected 2,3,5,6-tetrachlorocyclohexa-2,5-diene-1 ,4-dione (p- chloranil - Ch) and 1 ,4 p-benzoquinone (BQ) as electron acceptor probes in the SE / IL as indicated in Figure 3B based on a validation study presented in Figures 14 and 15 and the explanations provided in connection with Figure 15. Figure 3B (based on Figure 15) shows that the one-electron reduction potential for Ch is around 200 meV of the spectroelectrochemically determined CBM for stoichiometric Cs.05FA.79MA.i6 which is critical to measure electron injection into defect states through interaction with Ch to form Ch’ , which is apparent from the CV response in Figure 3C. We can readily differentiate Faradaic peaks on top of the capacitive background as a function of increasingly cathodic potentials as seen in Figure 16. As the applied potential is made more negative, Ch’ is formed in the SE / IL, and upon scan reversal (anodic wave), the Ch’ product is oxidized back to Ch. The charge passed in that oxidation event is correlated with the number of charges injected into either the population density of the mid-gap states below CBM, or states within CB itself and can be coulometrically quantified. In separate calibration experiments, we use an ITO bottom contact with the same SE / IL electrolyte, and a range of specified Ch concentrations in order to provide a means to determine absolute charge densities for all of these states (see the discussion in connetino with Figure 16).

[0044] For the case of excess FAI Cs.osFA.ssMA. , the Ch / Clr probe was unable to quantify the defects, because the voltammograms are affected by the redox activity of the iodide anion (see Figure 17 and related explanation provided in the Brief Description of the Drawings section). We hypothesized that due to the excess FAI, we had a number of undercoordinated Pb sites, which would lie deeper within the gap compared to the stoichiometric Cs.05FA.79MA.i6. Results using the BQ probe are presented in Figure 3D for the excess FAI system. With increasingly cathodic potentials, BQ is reduced at the perovskite / SE / IL interface from electron injection into states that are in the middle of the bandgap in addition to the states near CBM. We note that the deeper defects in the excess FAI Cs.05FA.83MA.i6 film overlap with the second one-electron reduction potential of BQ (to form BQ2-) ca. 430 meV farther from vacuum versus Ctr / Ch2’ as demonstrated in Figure 3E to shut off the redox activity of the iodide anions. As shown in Figure 2F for the spectroelectrochemical bleach of excess FAI Cs.05FA.83MA.i6, we notice that these deep donor defects cause the Urbach tail as described in Figure 2C to extend much deeper into the band gap. When electrons are now injected into these deeper donor defect states, unoccupied CB states within the tail are filled with electrons so that, even at higher cathodic end potentials, the allowed optical transitions shift to higher photon energies. Additionally, we confirm that the CBM of the excess FAI film is shifted closer to the vacuum level. Summaries of the redox activity and the calibration experiments are provided in connection with Figure 18 and associated description. Summaries of the redox activity and the calibration experiments for the excess Pbh CsFAMA film are also provided in Figure 19 and associated description.

[0045] To provide full energy descriptions, we have also characterized anionic defects for the CsFAMA, excess Pbh and excess FAI active layers, using the electron donor ferrocene (Fc) in the SE / IL layer. In these studies, anionic defects (k) are oxidized at potentials close to and beyond VBM, to create h'. This species then oxidizes Fc in the diffusion layer adjacent to the perovskite / SE / IL interface to Fc+, which is coulometrically quantified upon scan reversal. The potential dependence of Fc+formation helps to define both the actual VBM and the concentration of reactive defects which lie just above VBM, at concentrations down to ca. 1015cm-3, and potentially down to 1014cm-3. Results and datafor the quantification of the acceptor defects are summarized in connection with Figure 16 and associated description for stoichiometric Cs.05FA.79MA.i6, 4 for excess FAI Cs.05FA.83MA.i6 and 5 for excess Pbh Cs.05FA.75MA i6.

[0046] Figure 4 now summarizes the coulometric quantification of cation defects near CBM and anion defects near VBM, for the stoichiometric, excess FAI and excess Pbh perovskite, as a function of energy wrt the vacuum level, showing the near-Gaussian distribution of states just below CBM and just above VBM. The results combine the electrochemical approach for defect characterization and the band edge characterization with the spectroelectrochemical studies demonstrated in connection with Figures 16, 18 and 19 and associated description. The band edge maxima, defect concentrations and the FWHM for all fits are summarized in Table 1. For the stoichiometric triple cation perovskite in Figure 4, we note the low defect density for cation defects (ca. 1016cm-3, at ca. -3.85 eV) and anion defects at a 10x lower concentration (ca. 1015cm-3, at -5.4 eV). We anticipate that further optimization of the processing conditions for this active layer would even further lower the as-deposited defect densities by another 10x, and that our electrochemical approach to quantification would be able to track even those lower concentrations.Table 1. Example band edge maxima, defect concentrations and the FWHM of the Gaussian fit of the defects for stoichiometric Cs.05FA.79MA i6, excess FAI Cs05FA83MA.i6 and excess Pbh Cs.05FA.75MA.i6.*defect density behavior vs. energy follows an exponential behavior instead of Gaussian behavior

[0047] For the excess FAI PAL, the apparent EBG is closer to vacuum than for the stoichiometric film and reactive cation defects are no longer detected just below CBM, and are replaced by higher defect concentrations which where chemically reactive using redox probe Ch, but not using the BQ (ca. 1018cm-3) with a much broader energetic distribution (FWHM = ca. 0.069 eV) centered at ca. -4.21 eV, i.e. , 0.52 eV below CBM as demonstrated in Figure 4. Changes in reactive anion defect concentrations and energetic distributions are also especially notable, with total concentrations rising to ca. 1018cm-3, demonstrating a broad exponential tail of states extending from VBM to -5.24 eV into the gap of the active layer material. This explains the significantly diminished PV performance to 4% and the losses in Voc in FAI excess devices (Figures 10C and 10D). We interpret this change in distribution and concentration of anion defects as arising from the excess I’ accompanying the added FAI, for which there were no reaction pathways, allowing them to accumulate at the termini of the perovskite lattice, in grain boundary regions, and even just at the surface of the perovskite thin film to act as iodide interstitials. These new mid-gap cation defect states introduced in the excess FAI active layer suggests the presence of point defects where the overall coordination environment of the central heavy metal cations is changed so it is easier to reduce (or farther from vacuum), for example where Pb replaces an I site (i.e., Pbi) as discussed in DFT literature. The spectroelectrochemical data provides confirmation that both cation defects near CBM and anion defects near VBM show energetic distributions much like the "Urbach tails" that have been recently proposed for perovskite opto-electronic materials.

[0048] For the excess Pbh active layers, anion defects are essentially below our detection limit (below 1014cm-3), while cation defects are distributed energetically similar tothe stoichiometric active layers and with comparable defect densities as demonstrated in Figure 4. The (spectro)electrochemically determined EBG is nearly the same as for the stoichiometric material, but anion defect densities are now well below our LOD (ca. 1014cm-3). Electrochemically reactive cation defects appear just below CBM by ca. 100 meV, as for the stoichiometric active layers, and donor defect densities at close to the same level (ca. 1016cm-3). Clearly the excess Pbl2 in the precursor solution significantly lowers the reactive anion defect density, which explains the good PV performance (Figures 10E and 10F), but also increases the stability of these films due to slower photoinduced halide segregation. This statement is consistent with observations that state that a small excess of Pbh in the perovskite film enhances stability in perovskite energy conversion and light emitting devices.

[0049] In the area of hybrid organic-inorganic perovskites, advances are continuously being made via increased grain sizes and lower defect densities thru the use of coordinating solvents, additives and interface modifiers which slow the nucleation / growth process, and post processing of these thin films via thermal and / or solvent annealing, leading to impressive improvements in PSC performance and stabilities. Yet direct connections between chemical insights and defect quantification have remained minimal, in part due to the metrological challenge of quantifying both donor and acceptor defects with energy resolution. Herein we have demonstrated the first selective and direct quantification of reactive donor defects within a device-relevant triple cation PAL by spectroelectrochemically profiling the band edge energetics near CBM, noting the presence of a sub-population of electroactive states below CBM that appear when the active layer is under device-relevant stress, and where filling those states leads to changes in optical properties (both absorbance and photoluminescence) of the perovskite in ways that can be used to estimate their energetic distribution. Addition of a high electron affinity redox probe allows for coulometric quantification of these defect states down to concentrations which are much lower than those assessed using conventional spectroscopic or electronic probes. This approach allows for independent and selective characterization of defects near VBM and CBM, made possible by the use of the SE / IL top contact, and selection of probes which are energetically positioned, and with the right concentrations, to ensure complete capture of charge fromthese states in the perovskite / electrolyte interfacial region, as soon as formed. Additionally, we have confirmed that only slight compositional changes within the perovskite material have important consequences for its electronic structure and defect concentrations and thus long-term stability. It should be noted that this electrochemical probe provides the option to be used with complementary surface characterization techniques to promote operando analysis. In general, this approach will find use in the characterization of a much broader range of semiconductor materials, including thin films of organic and inorganic semiconductors, semiconductor and metal quantum dot thin films, and material blends and partial or full device stacks, where the SE / IL functions as a removable top contact which nevertheless rapidly electronically equilibrates with a diverse array of materials. Further embodiments of this characterization platform are predicted to be crucial for the in-line (scalable process) improvement of perovskite opto-electronic and energy conversion device stabilities and durabilities, which remain the primary impediment to commercialization and widespread market penetration of these technologies.Example Dual-Electrochemical-Probe Device for Characterization of Perovskite

[0050] Figure 5 shows a schematic view an example dual-probe device with two separate electrochemical probes, affixed to the perovskite surface, with independent control of the potential of the perovskite film with respect to the internal reference electrode(s), and independent measurement capability of the currents induced by changes in potential at the perovskite / solid-electrolyte (SE)-ionic liquid (IL) interface. The device of Figure 5 includes an ITO, a charge selective layer, and a perovskite layer. A charge selective layer can be applied to selectively allow the passage of one type of charge carrier (either electrons or holes) while blocking the other type to enhance the efficiency and performance of the device. The device of Figure 5 further includes two SE / IL probes in contact with the perovskite layer such that the two SE-IL are not in direct contact with each other.

[0051] The SE / IL probe at right (e.g., a first SE / IL probe) is configured for: a) characterization of band edge energies and energetics and distributions of cation defects near the conduction band minimum (CBM) and; b) quantification of cation defect densitiesvia redox reactions of the internal quinone (Q) probe molecule(s). The first SE-IL probe positioned on the right hand side of the device includes a first SE / IL layer positioned between a perovskite layer and a first counter electrode (CE). The first CE is coupled with the ITO. A first reference electrode (RE) on the right hand side is embedded in the first SE / IL and coupled with the ITO. The cation detect states of the perovskite layer are measured by applying a voltage to the first RE and measuring a current between the first RE and the first CE.

[0052] The SE / IL probe at left (e.g., a second SE / IL probe) is configured for: a) characterization of band edge energies and energetics and distributions of anion defects near the valence band maximum (VBM) and; b) quantification of anion defect densities via redox reactions of the internal ferrocene (Fc) probe molecule. The second SE-IL probe positioned on the left hand side of the device includes a second SE / IL layer positioned between the perovskite layer and a second counter electrode (CE). The second CE is coupled with the ITO. A second reference electrode (RE) on the left hand side is embedded in the second SE / IL and coupled with the ITO. The anion detect states of the perovskite layer are measured by applying a voltage to the second RE and measuring a current between the second RE and the second CE.

[0053] Figure 6 shows an example schematic view of the spectroelectrochemical determination of band edges of perovskite films. The conduction band is characterized using a reductive process (e.g., reduction of (p-chloranil - Ch) and / or 1 ,4 p-benzoquinone (BQ) by injection electrons). In contrast, the valence band is characterized using an oxidative process by extracting holes (e.g., oridation of ferrocine). The electrochemical probe characterization method using the dual-electrochemical probe device described with respect to Figure 5 can be combined with the spectroelectrochemical characterization of Figure 6. For example, both measurements can be performed with the same solid-electrolyte / ionic liquid “pads” (e.g., layers). Alternatively, electrochemical probe characterization and the spectroelectrochemical characterization could be done sequentially on different SE / IL pads.

[0054] Figure 7 shows (A) CV scans in which the end potential is shifted closer and past the valence band offset of the perovskite; (B) expanded oxidation peaks including a line that makes the distinction between Faradaic contributions due to defect states and the non- Faradaic background; (C) a peak charge vs probe concentration log-log calibration curve; and (D) predicted defect densities.

[0055] Figure 7A shows the consecutive cyclic voltammograms in which the end potential is shifted closer and past the valence band offset of the perovskite. When these CV scans are enlarged, Figure 7B demonstrates four expanded oxidation peaks including a black line that makes the distinction between Faradaic contributions due to defect states and the non-Faradaic background. Coulombic analysis of the Faradaic peaks for each potential, defines the peak charge as a function of end potential. By measuring the peak charge from the CV data of different concentrations of the used probe in the SE / IL electrolyte on an ITO working electrode, we construct a peak charge vs probe concentration log-log calibration curve shown in Figure 7C. This probe concentration is recalculated from the number of probe molecules in a volume of 1 cm3based on the mass, m, of added probe, the thickness of the gel, t, and the contact area, A, following:where cch is calculated in mol cm’3. By multiplying this number with Avogradro’s number, the concentration is represented as a density in cm-3. Defect densities in Figure 7D are predicted based on log^Qox, defects j~bionor defect—10a‘ -- ' where the Qox, detects is the maximum charge measured for the defect peak and the linear regression coefficients a and b from the log-log calibration plot in Figure 7C.Experimental ProceduresExample Sample Preparation

[0056] ITO coated glass slides (sheet resistance 9 - 15 O cr1(ohms per square), Colorado Concept Coatings LLC, 96041 ) were cut into squares (10 x 10 x 1.09 mm3). All these ITO substrates were pre- rinsed and sonicated in solutions of sodium dodecyl sulfate- water, acetone (ACS, 99.5%, Beantown Chemical), and finally isopropanol (ACS, 99.5%, Beantown Chemical), and allowed to dry. (FA.79MA.i6Cs.o5)Pb(l.87Br.i3)3 films (or later called Cs.05FA.79MA.i6) are deposited starting from a precursor solution containing ca. 110 mg of formamidinium iodide (FAI, >99.99%, Greatcell Solar), ca. 324 mg lead (Pbh, >98%, TCI America), ca. 14 mg of methylammonium bromide (MABr, >99.99%, Greatcell Solar), ca. 47 mg of lead bromide (PbBr2, >98%, TCI America), and ca. 11 mg of cesium iodide (Csl, 99.99% trace metals basis, Acros Organics). For the Pbh-rich (or Cs 05FA75MA i6) films, the precursor solution contains ca. 90 mg of FAI, ca. 324 mg Pbh, ca. 14 mg of MABr, ca. 47 mg of PbBr2, and ca. 11 mg of Csl. For the FAI-rich (or Cs.05FA.83MA.i6) films, the precursor solution contains ca. 119 mg of FAI, ca. 324 mg Pbk, ca. 14 mg of MABr, ca. 47 mg of PbBr2, and ca. 11 mg of Csl. All these salts were transferred to a hot 1-dram vial (stored at 150°C to minimize water content) in air with a micro stir bar and then immediately loaded into the nitrogen glovebox. Within the glovebox, 512 pL of N.N-dimethylformamide (DMF, anhydrous, 99.9%, Sigma Aldrich) and 128 pL of dimethylsulfoxide (DMSO, 99+%, Alfa Aesar) were added to the vial (generating solution concentrations of ca. 1 :1.1 :0.2:0.2:0.05 M for FAI:Pbl2:MABr:PbBr2:Csl, respectively) and then the precursor solution was stirred for 20 minutes. The solution was then filtered with 0.22 pm PTFE syringe filter (VWR). Before film processing, the circulation in the glovebox was turned off. The precursor solution was dispensed onto the cleaned ITO substrates (± 30 pL) at room temperature and then spincast at 1000 RPM for 10 seconds (1000 acceleration) followed immediately by 4000 RPM for 35 seconds (6000 acceleration). With 10 seconds remaining on the spin-coating cycle, 100 pL of chlorobenzene (anti-solvent, extra dry, 99.8%, Acros Organics) was added to the film in a clean dispensing motion as close to the center / top-surface of the spinning film aspossible. All films were immediately transferred to a hot plate for 40 minutes pre-heated at 100°C. The glovebox was purged slightly during film processing.Example Device ConstructionPreparation of the SE / IL Electrolyte

[0057] The electrolyte was made by adding 1 g poly(vinylidene fluoride-co- hexafluoropropylene) (or PVDF-HFP) pellets to 10 g acetone. The solution was stirred overnight in air on a hot plate set at 50 °C. When dissolved, 1 mg of the redox probe (in this case, ferrocene, p-chloranil, 1 ,4-benzoquinone) is weighed and then the cooled 10 wt% PVDF-HFP / acetone solution is added. To make the electrolyte solution, 160 mg 1 -ethyl-3- methylimidazolium bis(trifluoromethylsulfonyl)imide or [EMIM][TFSI] (stored in a glovebox) was dissolved in the redox probe spiked PVDF-HFP / acetone solution. 0.5 mL of the solution was then emptied in a rectangular watch glass (35 x 40 mm2) and was allowed to dry overnight.Example Electrochemical Cell Construction

[0058] First, a silver foil (5 x 13 mm2, 99.9% metals basis, Alfa Aesar) was polished with 4000 grit sandpaper, 1 pm silica and sonicated in deionized water. Using a two- electrode cell filled with 1 M HCI (ACS, 36-38%, EMD) and a Pt counter electrode, we chlorodized the silver foil at 0.4 V for 2 minutes. Three-electrode sandwich-type perovskite device structures were constructed as a 5-layered stack consisting of an ITO film transparent electrode on glass, a Cs.05FA.79MA.i6 film as the active layer, a SE / IL electrolyte film wrapped around an Ag / AgCI foil and a gold electrode. The gold electrode was thoroughly cleaned and sonicated in ethanol (anhydrous, >95%, Decon Laboratories Inc.) before use. The device stack was pressed and kept with a binder clip. All device structures were assembled and tested in inert atmosphere (O2 < 1 ppm; H2O < 1 ppm).Example Solar Cell Construction

[0059] ITO / glass substrates were sequentially cleaned with sodium dodecyl sulfate- water, deionized water, acetone and isopropanol under ultrasonication. The ITO / glass substrates were then dried with N2 blowing and treated with low pressure O2 plasma treatment for 10 min. A hole transport layer was deposited using an aqueous dispersion of 23 mg / mL of NiOx particles which were synthesized by precipitation route known in the art. This dispersion was tired and sonicated for 30 mins. Then, 100 pL were dropped on cleaned ITO and spin-coated at 3000 rpm for 30s. Finally, films were transferred to a hot plate and annealed at 80°C for 10 mins. The perovskite films were fabricated by the anti-solvent crystallization approach as discussed before in a glovebox. The devices with an area of 7.5 mm2were finished by thermally evaporating C60 (50 nm), bathocuproine (10 nm) and Al (100 nm) in sequential order by using a custom evaporation chamber system from Angstrom Engineering under a vacuum of 10‘7Torr.

[0060] Simulated AM1.5G irradiation was produced by a 300W Xenon arc lamp (Newport, USA) as a light source and an AM 1.5G optical filter (Newport, USA). An engineered diffuser (Newport, USA) is used to attenuate the output. The diffused light intensity and the distance to the device was calibrated to receive an incident power density of 100 mW / cm2using a calibrated reference cell from Photo Emission Tech Inc. JV data acquisition was carried out with a source meter (Keithley 2400 Sourcemeter, USA) and inhouse software (National Instruments Labview, version 8.2, USA). The devices were measured immediately after fabrication without any preconditioning. A voltage scan was measured from -0.2 V to 1 .1 V starting from the negative potential with a dwell time of 5 ms, a voltage increment of 10 mV and a current limit of 200 mA. The devices were measured in nitrogen atmosphere.Perovskite Film Characterization0 / 20 X-Ray Diffraction (XRD)

[0061] For 0 / 20-XRD experiments, MAPbh films on ITO were mounted on a clay holder in air and loaded into a Phillips X'PERT MPD system, with "PreFIX" source module and a "X'Celerator" detector module, using CuKaradiation (A max — 1 .541874 A) with electron gun cathode-anode power settings of 45 kV at 40 mA). Films were analyzed from 5° to 90° in 26 with 0.0167° step size, scanning symmetrically (i.e. , source and detector at same angle with respect to the surface normal - thus preferentially detecting scattering from crystal planes parallel to the substrate plane) with a total scan time of approximately 35 minutes. Optical hardware settings include a 0.04 rad soller slit (source and detector side), a 2° divergence slit at 140 mm from sample (source side), a 11.6 mm horizontal mask (source side), a 0.02 mm-thick Nickel filter (source side), and an overhead sample beam knife. Processing / Analysis of XRD patterns were performed with X'Pert HighScore (for background subtraction) and Mercury 2020.3.0 software (for identifying Bragg peaks from crystal structure).Ultraviolet-visible (UV-Vis) Spectrophotometry

[0062] Absorbance data of the perovskite films were measured using an Ocean Optics Balanced Deuterium Tungsten Source (210-2500 nm) and an OCEAN-FX-XR1 fiber optic spectrometer (200-1025 nm) with a 25 pm slit controlled with the Oceanview software package. The perovskite film spectra were collected relative to a bare ITO substrate.Photoluminescence (PL) Spectroscopy

[0063] Steady-state photoluminescence data were recorded using a spectrograph (Fluorolog-3 from Horiba Scientific) equipped with a Xenon lamp, single-grating excitation and emission monochromators in combination with a R928P photomultiplier tube detector controlled using the Fluoressence software package. The perovskite films were excited at 640 nm (5 nm bandpass) and the emission spectrum was collected from 660 to 900 nm with an increment of 1 nm (5 nm bandpass). Time-resolved photoluminescence data wererecorded using a pulsed laser-diode (DeltaPro Lifetime Fluorometer) that allows time correlated single photon counting that was added to the Fluorolog-3 spectrograph from Horiba Scientific. The temporal range measured goes from 25 ps to 5 ps for the perovskite films. The maximum intensity was set to 10,000 counts.Spectroelectrochemical Measurements

[0064] Potential-controlled (CH Instruments Electrochemical Analyzer Model 660C potentiostat) spectroelectrochemical measurements of perovskite films on ITO working electrodes are recorded using a three-electrode device configuration. The complete device, including the lead halide perovskite film, is placed in the path of the light beam of an UV-vis spectrophotometer (Ocean Optics system as described above) and a background spectrum is taken after a fixed waiting period (t = 5 min) at a potential of -0.5 V vs. Ag / AgCI where the perovskite is expected to show the optical and electrical properties of the as-deposited film. Steady-state absorption spectra are acquired after 5 min at defined potentials (20 mV intervals in the bleaching regime around the conduction band) between -0.5 V vs. Ag / AgCI to -1 .3 V vs. Ag / AgCI using the bulk electrolysis feature in the control software.Electrochemical Measurements

[0065] The perovskite film is evaluated using a three-electrode configuration using a gold electrode and an Ag / AgCI foil as a counter and reference electrode, respectively. All voltammograms were recorded using a CH Instruments Electrochemical Analyzer Model 660C potentiostat controlled with the corresponding software package. Prior to evaluating the perovskite films using the electrochemical approach, the redox probes were evaluated by cyclic voltammetry at different scan rates (10-500 mV s-1for 10 cycles) to determine their mass transport and kinetic behavior in the SE / IL. Those redox probe test structures consisted of a solid electrolyte film with redox probe wrapped around an Ag / AgCI foil pressed in between a gold counter and an ITO working electrode. All perovskite device structures were assembled and tested in inert atmosphere (O2 < 1 ppm; H2O < 1 ppm).Validation of Ch and BQ as Electron Acceptor Probe

[0066] To validate the addition of Ch and BQ to the SE / IL for defect characterization, the reversibility, diffusivity and stability of the Ch 7Ch and BQ 7BQ redox process are demonstrated at an ITO working electrode in Figures 15A and 15B. From the cyclic voltametric peak separation in Figure 15A, we determine a heterogeneous rate constant of ko ~ 2 x 10"5cm s-1using the Nicholson method, with a diffusion coefficient of 2.8 x 10"8cm2s-1at 50 mV s-1for the Ch probe in the SE / IL in Figure 15C, which is slightly slower compared to the Fc / Fc+probe used to characterize acceptor defects in lead halide perovskites. For the BQ probe in the SE / IL in Figure 15D, we determine a heterogeneous rate constant of ko « 4 x 1O-7cm s’1, with a diffusion coefficient of 2.1 x 10’9cm2s’1at 50 mV s-1, which is slightly slower compared to Ch 7Ch.Ch Probe Calibration for Stoichiometric Cs osFAygMA -ie

[0067] In Figure 16A, we demonstrate four expanded oxidation peaks from the CV scans in Figure 3C including a black line that makes the distinction between Faradaic contributions and the non-Faradaic background. When focusing on the donor defect contributions in stoichiometric CsFAMA as demonstrated in Figure 16B, the peak charges as a function of the cathodic end potential follow a Gaussian distribution (y = A exp(-1 / 2((x- xc) / w)2) + yo fit is added as a black line) at energies close to the measured the CBM. The defect quantification using the amount of charge from the CV data in Figure 16B is done based on a peak charge vs. Ch concentration calibration curve measured on an ITO electrode. First, the Ch concentration in the SE / IL electrolyte is recalculated from the number of Ch molecules in a volume of 1 cm3based on the mass, m, of added Ch, the thickness of the gel, t, and the contact area, A, using Equation (1 ), which was described earlier:where cch is calculated in mol cm-3. By multiplying this number with Avogradro's number, the concentration is represented as a density in cm-3. This type of calibration allows us toconsider diffusion and distribution of Ch within the SE / IL electrolyte and to demonstrate the linear dependence between the Ch concentration and the measured signal. This linear range for the calibration (= the signal is directly proportional to the concentration of the analyte) is expected to be between its detection limit (= the lowest concentration at which the probe can be detected) and the higher concentrations at which the signal starts to be limited by mass transport and distribution limitations, which would result in a saturation of the signal response. In Figure 16C, the measured charge densities are calibrated using known concentrations of the Ch probe in the SE / IL electrolyte in a log-log calibration plot. The logarithmic scale is used because the data we are displaying covers a very large range of Ch concentrations (1015cm-3to 1019cm-3), which skews the curve when plotted on a linear scale. The slope from the calibration determines how strongly the current increases with the amount of p-chloranil present in the solid electrolyte (comparable to the use of the extinction coefficient in Beer's law), while the intercept describes the charge density of the peak with the assumption the Ch concentration is zero. Subsequently, with the assumption that the Ch probe is homogeneously distributed in the SE / IL electrolyte, the integrated charge densities for the Ch'" oxidation measured on the ITO electrode can be used as a calibration for the integrated charge densities measured on the perovskite film. Defect densities are predicted based on Equation (2), which was described earlier: l0g(Qox,clefects)~^^donor defect 10a(2) where the Qox, defects is the maximum charge for the defect peak and the linear regression coefficients a and b from the log-log calibration plot in Figure 16C. By using this defect characterization method for the stoichiometric Cs.05FA.79MA.i6 perovskite film, we calculate a defect density of 1.62 x 1016cm-3under inert conditions as demonstrated in Figure 16D. The detection limit for this CV approach with the Ch probe on stoichiometric Cs.05FA.79MA.i6 films is determined using the standard deviation of the minimum resolvable peak charge density on three perovskite films for reproducibility. Using the calibration curve in Figure 16C, this standard deviation is associated with the detection limit of the defect density withinthe stoichiometric Cs.05FA.79MA.i6 for the CV method under the specified conditions based on log(LOD = 10

[0068] In the case of stoichiometric Cs 05FA79MA.i6, the minimum detectable charge density for the CV approach is 13.4 nC cm-2, which corresponds to a donor defect density detection limit of ± 2 x 1014cm-3using the calibration curve for the Ch probe.Ch Probe for Excess FAI Cs 05FA.83MA.i6 Films

[0069] Figure 17 illustrates Cs Probe results for Excess FAI Cs.05FA.83MA.16 Films. Schematic of the CBM of excess FAI Cs.05FA.83MA.16 combined with the standard reduction potentials for the redox reactions of the Ch probe and the measured CVs as a function of cathodic potential endpoints (A) approaching CBM and (B) surpassing CBM in the presence of 3 mM Ch in the SE / I L, at 0.05 V s-1 followed by anodic sweeps to the initial potential. All these characterizations were carried out in inert atmosphere (H2O < 1 ppm; 02 < 1 ppm). The CVs are shifted by -0.1 V on the potential axis to compensate for an Ag / AgCI reference potential of 0.2 V (versus the Fc / Fc+ redox couple).Calibration of BQ Probe for Excess FAI Cs.05FA.83MA.i6

[0070] By reducing the excess FAI Cs.05FA.83MA.i6 film in Figure 3D, the first CV (purple scan) illustrates the electrochemical activity caused by the charge transfer process through defect states within the perovskite. Once the cathodic end potential passes -0.7 V vs. Ag / AgCI (-4.0 eV vs. local vacuum; cyan scan in Figure 3D), the forward scan demonstrates a current decrease, while, in the reverse scan, the defect oxidation peak slowly disappears. Using these redox peaks, Figure 18A demonstrates that the Coulometric analysis of the oxidation process of the reversible donor defects as a function of cathodic end potential. This broad distribution of defect states demonstrates a maximum at -0.58 V vs. Ag / AgCI (E = -4.12 eV) deep within the band gap of the excess FAI Cs.05FA.83MA.i6 perovskite with a maximum defect density of 8.71 x 1017cm-3calculated using a calibration procedure for theBQ probe in Figure 18B. Combined with the LSV / DPV measurement of the acceptor states in the excess FAI perovskite in Figure 18C using the procedure described in literature, we are able to draw a complete band diagram of the excess FAI Cs.05FA.83MA.i6 perovskite in Figure 18C.Defects in Excess Pbh Cs.05FA.75MA.i6 Films

[0071] To check if defect states are responsible for this dA / dE change, we run cyclic voltametric scans on an excess Pbh Cs 05FA75MA.i6 film with a BQ probe as demonstrated in Figure 19A. The first evidence of an electron transfer reaction is the appearance of the BQ2-oxidation, which happens when the excess Pbh Cs.05FA.75MA.i6 film is reduced past - 0.94 V vs. Ag / AgCI (-3.76 eV vs. local vacuum level). Once we pass the CBM, the current density increases very fast, which means a stability decrease under high operating voltages. Based on the procedure described in literature, we switch to a LSV / DPV approach to improve sensitivity and the energy resolution as seen in Figure 19B. By reducing excess Pbh Cs.05FA.75MA.i6 in Figure 19B, the first reverse oxidation scan (purple scan) illustrates the capacitive behavior within the band gap. Once the cathodic end potential passes -0.8 V vs. Ag / AgCI (-4.0 eV vs. local vacuum; orange scan), the reverse DPV scan, demonstrates that the initially reduced electron capture probe, BQ2', formed during the charge transfer process at the donor defects of the perovskite, is oxidized back to BQ*’. Using these potential-dependent redox peaks, Figure 12C demonstrates that the Coulometric analysis of the oxidation process of the reversible donor defects as a function of cathodic end potential follows a Gaussian distribution. This distribution of defect states demonstrates a maximum at -0.96 V vs. Ag / AgCI (or -3.74 eV) close the CBM of the excess Pbh Cs 05FA75MA i6 perovskite with a maximum donor defect density of 1.76 x 1016cm-3calculated using a calibration procedure for the BQ probe in Figure 19D. The energy levels of these defects lie close to the dA / dE peak seen in the spectroelectrochemical bleach data in Figure 2E, where the defect peak seems to form around -3.86 eV vs. the local vacuum level. Combined with the LSV / DPV measurement of the acceptor states in the excess Pbhperovskite using the procedure described in literature, which does not show any acceptor defects, we draw a complete band diagram of excess Pbh Cs.05FA.75MA.i6 in Figure 19C.

[0072] Figure 20 illustrates a set of operations for using an electrochemical method for characterizing a perovskite layer of a semiconductor device in accordance with an example embodiment. At 2002, the perovskite layer is contacted with a first surface of a first electrolyte layer. For example, the first electrolyte layer can be a solid electrolyte / ionic liquid (SE / IL) layer. A second surface of the first SE / IL layer is in contact with a first counter electrode. At 2004, a first reference electrode is embedded in the first electrolyte layer. At 2006, a voltage is applied to the first reference electrode. At 2008, a current between the first reference electrode and the first counter electrode is measured. At 2010, the perovskite layer is characterized based on the applied voltage and the measured current between the first reference electrode the first counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

[0073] Figure 21 illustrates a set of operations for using a spectroelectrochemical method for characterizing a perovskite layer of a semiconductor device in accordance with another example embodiment. At 2102, the perovskite layer is contacted with a first surface of an electrolyte layer. For example, the electrolyte layer can be a solid electrolyte / ionic liquid (SE / IL) layer. The electrolyte layer includes an electron-acceptor redox probe, the electrolyte layer is in contact with a counter electrode on a second surface of the electrolyte layer, and the second surface is opposite to the first surface of the electrolyte layer. At 2104, a reference electrode is embedded in the electrolyte layer. At 2106, a voltage to is applied to the reference electrode. At 2108, a current between the counter electrode and the reference electrode is measured. At 2110, concurrently with applying the voltage to the reference electrode, an absorbance measurement is performed on a region of the perovskite layer to produce an absorbance spectrum. At 2112, the region of the perovskite layer is characterized based on the applied voltage, the measured current, and the absorbance spectrum.

[0074] Figure 22 illustrates a set of operations for using an electrochemical method for characterizing a perovskite layer of a semiconductor device in accordance with an example embodiment. At 2202, a voltage is applied to a reference electrode. At 2204, a current between the reference electrode and a counter electrode is measured. The reference electrode is affixed within a probe layer; the probe layer includes an electrolyte layer. For example, the electrolyte layer can be a solid electrolyte / ionic liquid (SE / IL) layer. The electrolyte layer includes two quinone electron-acceptor redox probes; a first surface of the probe layer is in contact with the perovskite layer, and a second surface of the probe layer is in contact with a counter electrode. At 2204, the perovskite layer is characterized based on the applied voltage and the measured current between the reference electrode the counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

[0075] Various operations disclosed herein can be implemented using a processor / controller is configured to include, or be coupled to, a memory that stores processor executable code that causes the processor / controller carry out various computations and processing of information. The processor / controller can further generate and transmit / receive suitable information to / from the various system components, as well as suitable input / output (IO) capabilities (e.g., wired or wireless) to transmit and receive commands and / or data. The processor / controller may, for example, provide signals to control the operation of various components such as voltage light sources, and detectors that are disclosed herein. The processor / controller may be further configured to perform various method steps and computations that are disclosed in this patent document.

[0076] Various information and data processing operations described herein may be implemented in one embodiment by a computer program product, embodied in a computer- readable medium, including computer-executable instructions, such as program code, executed by computers in networked environments. A computer-readable medium may include removable and non-removable storage devices including, but not limited to, Read Only Memory (ROM), Random Access Memory (RAM), compact discs (CDs), digital versatile discs (DVD), etc. Therefore, the computer-readable media that is described in thepresent application comprises non-transitory storage media. Generally, program modules may include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Computer-executable instructions, associated data structures, and program modules represent examples of program code for executing steps of the methods disclosed herein. The particular sequence of such executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps or processes.

[0077] Only a few implementations and examples are described and other implementations, enhancements and variations can be made based on what is described and illustrated in this patent document.Exemple Clauses

[0078] Clause 1 . An electrochemical method of characterizing a perovskite layer of a semiconductor device, the method comprising: contacting the perovskite layer with a first surface of a first electrolyte layer, wherein a second surface of the first electrolyte layer is in contact with a first counter electrode; embedding a first reference electrode in the first electrolyte layer; applying a voltage to the first reference electrode; measuring a current between the first reference electrode and the first counter electrode; and characterizing the perovskite layer based on the applied voltage and the measured current between the first reference electrode the first counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer. The first electrolyte layer can, for example, be a solid electrolyte / ionic liquid layer.

[0079] For example, the device of Figure 5 includes an ITO, a charge selective layer and a perovskite layer. A first SE / IL layer positioned on the right hand side of the device is positioned between a perovskite layer and a first counter electrode (CE). The first CE is coupled with the ITO. A first reference electrode (RE) on the right hand side is embedded in the first SE / IL and coupled with the ITO. The cation detect states of the perovskite layer are measured by applying a voltage to the first RE and measuring a current between the first RE and the first CE.

[0080] Clause 2. The method of any of the clauses herein, wherein the first electrolyte layer includes two electron-acceptor redox probes.

[0081] Clause 3. The method of any of the clauses herein, wherein the two electronacceptor redox probes include a 2,3,5,6-tetrachlorocyclohexa-2,5-diene-1 ,4-dione probe and a 1 ,4 p-benzoquinone probe.

[0082] Clause 4. The method of any of the clauses herein, wherein: applying the voltage to the first reference electrode causes a reduction of the two electron-acceptor redox probes in the electrolyte layer, the reduction causing an electron injection into cation defect states of the perovskite layer. The reduction can thereby be used to characterize cation defect states of the perovskite layer. Cation and anion defect characterization results is described, for example, with respect to Figure 4.

[0083] Clause 5. The method of any of the clauses herein, further comprising: contacting the perovskite layer with a first surface of a second electrolyte layer, wherein a second surface of the second electrolyte layer is in contact with a second counter electrode; embedding a first reference electrode in the first electrolyte layer; applying a voltage to the second reference electrode; measuring a current between the second reference electrode and the second counter electrode; and characterizing the perovskite layer based on the applied voltage and the measured current between the second reference electrode the second counter electrode, wherein characterizing the perovskite layer further includes characterizing anion defect states of the perovskite layer. The second electrolyte layer can, for example, be a solid electrolyte / ionic liquid layer.

[0084] For example, in Figure 5, a second SE / IL layer positioned on the left hand side of the device is positioned between the perovskite layer and a second counter electrode (CE). The second CE is coupled with the ITO. A second reference electrode (RE) on the left hand side is embedded in the second SE / IL and coupled with the ITO. The anion detect states of the perovskite layer are measured by applying a voltage to the second RE and measuring a current between the second RE and the second CE.

[0085] Clause 6. The method of any of the clauses herein, wherein the second electrolyte layer includes a ferrocene probe.

[0086] Clause 7. The method of any of the clauses herein, wherein: applying the voltage to the second reference electrode causes an oxidation of ferrocene probes in the electrolyte layer, the oxidation causing extraction of holes from anion defect states of the perovskite layer. Cation and anion defect characterization results is described, for example, with respect to Figure 4.

[0087] Clause 8. The method of any of the clauses herein, wherein characterizing the perovskite layer includes: characterization of band edge energies and energetics and distributions of the cation defects near a conduction band minimum and quantification of cation defect densities via redox reactions of quinone probe molecules (e.g., see right hand side of Figure 6); and characterization of band edge energies and energetics and distributions of anion defects near valence band maximum and quantification of anion defect densities via redox reactions of ferrocene probe molecules (e.g., see left hand side of Figure 6). For example, using the dual-probe device of Figure 5, band edge energies and energetics of cation and anion defects can be measured using a single device.

[0088] Clause 9. The method of any of the clauses herein, wherein: the characterization of the perovskite layer includes quantifying donor defects and / or acceptor defects in the perovskite layer by comparing the measured currents versus at a particular applied voltage against a respective calibration curve (e.g., as described with respect to the calibration curve of Figure 7C), and the respective calibration curve is constructed based on cyclic voltammogram measurements (e.g., as described with respect to Figure 7A) of calibration samples comprising different probe concentrations.

[0089] Clause 10. The method of any of the clauses herein, wherein the characterization includes quantifying cation defects in the perovskite layer with a density as low as 1016cm-3 (e.g., as described with respect to Figure 4).

[0090] Clause 11. The method of clause 1 , wherein the characterization includes quantifying anion defects in the perovskite layer with a density as low as 1015cm-3 (e.g., as described with respect to Figure 4).

[0091] Clause 12. The method of any of the clauses herein, further comprising: concurrently to applying the voltage to the first reference electrode or the second reference electrode, perform an absorbance measurement on a region of the perovskite layer to produce an absorbance spectrum; and characterizing the region of the perovskite layer based on the applied voltage, the measured current, and the absorbance spectrum. For example, Figure 2A-2F illustrate spectroelectrochemical characterization of PALs. Similar measurement can be performed on the oxidative process (ferrocine) and the reductive process (quinone). The spectroelectrochemical characterization can be combined with the dual-probe electrochemical measurement by the device of Figure 5. As shown in Figure 6, the valence band is characterized using an oxidative process by extracting holes and the conduction band is characterized using a reductive process by injection electrons holes and the conduction band is characterized using a reductive process by injection electrons.

[0092] Clause 13. The method of any of the clauses herein, wherein the first counter electrodes is a gold electrode and first reference electrode is an Ag / AgCI electrodes.

[0093] Clause 14. The method of any of the clauses herein, wherein the perovskite layer is further in contact with a charge selective layer.

[0094] Clause 15. The method of any of the clauses herein, wherein the semiconductor device is a solar cell.

[0095] Clause 16. A spectroelectrochemical method of characterizing a perovskite layerof a semiconductor device, the method comprising: contacting the perovskite layer with a first surface of an electrolyte layer, wherein: the electrolyte layer includes an electronacceptor redox probe, the electrolyte layer is in contact with a counter electrode on a second surface of the electrolyte layer, and the second surface is opposite to the first surface of the electrolyte layer, embedding a reference electrode to the electrolyte layer; applying a voltage to the reference electrode; measuring a current between the counter electrode and thereference electrode; and concurrently with applying the voltage to the reference electrode, perform an absorbance measurement on a region of the perovskite layer to produce an absorbance spectrum; characterizing the region of the perovskite layer based on the applied voltage, the measured current, and the absorbance spectrum. The electrolyte layer can, for example, be a solid electrolyte / ionic liquid layer.

[0096] For example, an absorbance measurement, described with respect to Figures 2A-2F and Figure 6, can be combined with the described electrochemical probe measurement using a single SE / IL or dual Se / IL probe device (e g., described with respect to Figure 7). The measurement can be used to determine conduction band and valence band energy of the perovskite layer.

[0097] Clause 17. The method of any of the clauses herein, wherein the characterizing the region of the perovskite layer includes determining conduction band and valence band energy of the perovskite layer.

[0098] Clause 18. The method of any of the clauses herein, wherein the SE / IL layer includes a ferrocene probe.

[0099] Clause 19. The method of any of the clauses herein, wherein the SE / IL layer includes two electron-acceptor redox probes including a 2,3,5,6-tetrachlorocyclohexa-2,5- diene-1 , 4-dione probe and a 1 ,4 p-benzoquinone probe.

[0100] Clause 20. The method of any of the clauses herein, wherein the counter electrode is made of optically transparent conductive material (e.g., ITO).

[0101] Clause 21 . An electrochemical method of characterizing a perovskite layer of a semiconductor device, the method comprising: applying a voltage to a reference electrode; and measuring a current between the reference electrode and a counter electrode, wherein: the reference electrode is affixed within a probe layer, the probe layer including an electrolyte layer and two quinone electron-acceptor redox probes, a first surface of the probe layer is in contact with the perovskite layer, a second surface of the probe layer is in contact with a counter electrode, and characterizing the perovskite layer based on the applied voltage andthe measured current between the reference electrode the counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

[0102] Clause 22. An device (e.g., Figure 5) for characterizing a perovskite layer of a semiconductor device by electrochemistry, the device comprising: a probe layer including an electrolyte layer, such as a solid electrolyte / ionic liquid (SE / IL) layer, and one or more electron-acceptor redox probes, a reference electrode embedded in the probe layer; and a counter electrode in contact with a surface of the probe layer, wherein: the reference electrode is configured to receive a voltage when a second surface of the probe layer is positioned in contact with the perovskite layer, the device is configured to carry a measurable current between the reference electrode and the counter electrode, and the device is configured for characterizing the perovskite layer based on the applied voltage and the measured current between the reference electrode and the counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

[0103] Clause 23. The device of any of the clauses herein, wherein: the device is part of a measurement system, the measurement system further comprising: a voltage supply configured to provide the voltage to the reference electrode; and a current measurement device configured to measure the current between the reference electrode and the counter electrode.

[0104] Clause 24. The device of any of the clauses herein, the system further comprising: at least one processor; at least one non-transitory memory coupled to the at least one processor, and storing instructions that, when executed by the at least one processor, cause the measurement system to perform the methods of any of the clauses disclosed herein.

Claims

CLAIMSI / We claim:

1. An electrochemical method of characterizing a perovskite layer of a semiconductor device, the method comprising: contacting the perovskite layer with a first surface of a first electrolyte layer, wherein a second surface of the first electrolyte layer is in contact with a first counter electrode; embedding a first reference electrode in the first electrolyte layer; applying a voltage to the first reference electrode; measuring a current between the first reference electrode and the first counter electrode; and characterizing the perovskite layer based on the applied voltage and the measured current between the first reference electrode the first counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

2. The method of claim 1 , wherein the first electrolyte layer is a solid electrolyte / ionic liquid(SE / IL) layer.

3. The method of claim 1 or 2, wherein the first electrolyte layer includes two electron-acceptor redox probes.

4. The method of claim 3, wherein the two electron-acceptor redox probes include a 2,3,5,6-tetrachlorocyclohexa-2,5-diene-1 ,4-dione probe and a 1 ,4 p- benzoquinone probe.

5. The method of claim 3, wherein:applying the voltage to the first reference electrode causes a reduction of the two electron-acceptor redox probes in the electrolyte layer, the reduction causing an electron injection into cation defect states of the perovskite layer.

6. The method of claim 1 , further comprising: contacting the perovskite layer with a first surface of a second solid electrolyte layer, wherein a second surface of the second electrolyte layer is in contact with a second counter electrode; embedding a first reference electrode in the first electrolyte layer; applying a voltage to the second reference electrode; measuring a current between the second reference electrode and the second counter electrode; and characterizing the perovskite layer based on the applied voltage and the measured current between the second reference electrode the second counter electrode, wherein characterizing the perovskite layer further includes characterizing anion defect states of the perovskite layer.

7. The method of claim 6, wherein the second electrolyte layer is a solid electrolyte / ionic liquid(SEZIL) layer.

8. The method of claim 6 or 7, wherein the second electrolyte layer includes a ferrocene probe.

9. The method of claim 8, wherein: applying the voltage to the second reference electrode causes an oxidation of ferrocene probes in the electrolyte layer, the oxidation causing extraction of holes from anion defect states of the perovskite layer.

10. The method of claim 1 , wherein characterizing the perovskite layer includes:characterization of band edge energies and energetics and distributions of the cation defects near a conduction band minimum and quantification of cation defect densities via redox reactions of quinone probe molecules; and characterization of band edge energies and energetics and distributions of anion defects near valence band maximum and quantification of anion defect densities via redox reactions of ferrocene probe molecules.11 . The method of claim 1 , wherein: the characterization of the perovskite layer includes quantifying donor defects and / or acceptor defects in the perovskite layer by comparing the measured currents versus at a particular applied voltage against a respective calibration curve, and the respective calibration curve is constructed based on cyclic voltammogram measurements of calibration samples comprising different probe concentrations.

12. The method of claim 1 , wherein the characterization includes quantifying cation defects in the perovskite layer with a density as low as 1016cm-3.

13. The method of claim 1 , wherein the characterization includes quantifying anion defects in the perovskite layer with a density as low as 1015cm-3.

14. The method of claim 6, further comprising: concurrently to applying the voltage to the first reference electrode or the second reference electrode, perform an absorbance measurement on a region of the perovskite layer to produce an absorbance spectrum; and characterizing the region of the perovskite layer based on the applied voltage, the measured current, and the absorbance spectrum.

15. The method of claim 1 , wherein the first counter electrodes is a gold electrode and first reference electrode is an Ag / AgCI electrodes.

16. The method of claim 1 , wherein the perovskite layer is further in contact with a charge selective layer.

17. The method of claim 1 , wherein the semiconductor device is a solar cell.

18. A spectroelectrochemical method of characterizing a perovskite layer of a semiconductor device, the method comprising: contacting the perovskite layer with a first surface of an electrolyte layer, wherein: the electrolyte layer includes an electron-acceptor redox probe, the electrolyte layer is in contact with a counter electrode on a second surface of the electrolyte layer, and the second surface is opposite to the first surface of the electrolyte layer, embedding a reference electrode in the electrolyte layer; applying a voltage to the reference electrode; measuring a current between the counter electrode and the reference electrode; concurrently with applying the voltage to the reference electrode, performing an absorbance measurement on a region of the perovskite layer to produce an absorbance spectrum; and characterizing the region of the perovskite layer based on the applied voltage, the measured current, and the absorbance spectrum.

19. The method of claim 18, wherein the electrolyte layer comprises a solid electrolyte / ionic liquid (SE / IL) material.

20. The method of claim 18, wherein the characterizing the region of the perovskite layer includes determining conduction band and valence band energy of the perovskite layer.

21. The method of claim 18, wherein the electrolyte layer includes a ferrocene probe.

22. The method of claim 18, wherein the electrolyte layer includes two electronacceptor redox probes including a 2,3,5,6-tetrachlorocyclohexa-2,5-diene-1 ,4- dione probe and a 1 ,4 p-benzoquinone probe.

23. The method of claim 18, wherein the counter electrode is made of optically transparent conductive material.

24. An electrochemical method of characterizing a perovskite layer of a semiconductor device, the method comprising: applying a voltage to a reference electrode; measuring a current between the reference electrode and a counter electrode, wherein: the reference electrode is affixed within a probe layer, the probe layer including a electrolyte layer and two quinone electron-acceptor redox probes, a first surface of the probe layer is in contact with the perovskite layer, and a second surface of the probe layer is in contact with a counter electrode; and characterizing the perovskite layer based on the applied voltage and the measured current between the reference electrode the counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

25. A device for characterizing a perovskite layer of a semiconductor device by electrochemistry, the device comprising: a probe layer including an electrolyte layer and one or more electron-acceptor redox probes, a reference electrode embedded in the probe layer; and a counter electrode in contact with a surface of the probe layer, wherein: the reference electrode is configured to receive a voltage when a second surface of the probe layer is positioned in contact with the perovskite layer, the device is configured to carry a measurable current between the reference electrode and the counter electrode, and the device is configured for characterizing the perovskite layer based on the applied voltage and the measured current between the reference electrode and the counter electrode, wherein characterizing the perovskite layer includes characterizing cation defect states of the perovskite layer.

26. The device of claim 25, wherein: the device is part of a measurement system, the measurement system further comprising: a voltage supply configured to provide the voltage to the reference electrode; and a current measurement device configured to measure the current between the reference electrode and the counter electrode.

27. The device of claim 26, the system further comprising: at least one processor;at least one non-transitory memory coupled to the at least one processor, and storing instructions that, when executed by the at least one processor, cause the measurement system to perform any of the methods of claims 1 -26.

Images