Elucidation of Chemical Species and Reactivity at Methylammonium

Subscriber access provided by UNIV OF DURHAM

C: Surfaces, Interfaces, Porous Materials, and Catalysis

Elucidation of Chemical Species and Reactivity at Methylammonium Lead Iodide and Cesium Tin Bromide Perovskite Surfaces via Orthogonal Reaction Chemistry Weiran Gao, Kenneth Zielinski, Benjamin N. Drury, Alexander D Carl, and Ronald L. Grimm J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b05352 • Publication Date (Web): 19 Jul 2018 Downloaded from http://pubs.acs.org on July 19, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Elucidation of Chemical Species and Reactivity at Methylammonium Lead Iodide and Cesium Tin Bromide Perovskite Surfaces via Orthogonal Reaction Chemistry Weiran Gao,§ Kenneth Zielinski,§ Benjamin N. Drury, Alexander D. Carl, and Ronald L. Grimm*

Department of Chemistry and Biochemistry; Life Science and Bioengineering Center; Worcester Polytechnic Institute, 100 Institute Road, Worcester, Massachusetts 01609

§

These authors contributed equally to the work.

* Author to whom correspondence should be addressed [email protected] ORCID 0000-0003-0407-937X

1 ACS Paragon Plus Environment

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 45

Abstract We quantified the chemical species present at and reactivity of the (100) face of singlecrystal methylammonium lead iodide, MAPbI3(100), and polycrystalline cesium tin bromide, CsSnBr3. For these ABX3 perovskites, experiments utilized the orthogonal reactivity of the A+-site cation, the B2+-site cation, and the X–-site halide anion. Ambientpressure exposure to BF3 solutions probed the reactivity of interfacial halides. Reactions with p-trifluoromethylanilinium chloride probed the exchange reactivity of the A+-site cation. A complex-forming ligand, 4,4’-bis(trifluoromethyl)-2,2’-bipyridine, probed for interfacial B2+-site cations. Fluorine features in x-ray photoelectron spectroscopy (XPS) quantified reaction outcomes for each solution-phase species. XPS indicated adsorption of BF3 indicating surface-available halide anions on both MAPbI3(100) and on CsSnBr3. Temperature-programmed desorption (TPD) quantified a ~200 kJ mol–1 desorption activation energy from MAPbI3(100) and a ~215 kJ mol–1 desorption energy from CsSnBr3.

Adsorption of the fluorinated anilinium cation included no concomitant

adsorption of chlorine as revealed by the absence of Cl 2p features within the limits of XPS detection. We interpret the observation of the anilinium species as exchanging for interfacial methylammonium species on MAPbI3(100) surfaces and interfacial cesium on the polycrystalline CsSnBr3 surface.

Within detection limits, the bipyridine ligand

demonstrated no adsorption to MAPbI3(100) suggestive of a Pb2+ deficient surface, but adsorption to the polycrystalline CsSnBr3 that suggests surface-accessible Sn2+. The combination of results implies that methylammonium cations and iodide anions dominate tetragonal MAPbI3(100) surface that respectively enable cation exchange and Lewis adduct formation for surface derivatization. We discuss the present results in the context of interfacial stability, passivation, and reactivity for perovskite-based energy conversion.


Page 3 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1. Introduction High optical absorption, defect tolerance, and straightforward processing methods drive research and development efforts of perovskite-based solar-energy-conversion photovoltaics.1-4 Large bandgap tunability and ease of processing may enable scalable deployment of inexpensive tandem-junction solar cells that break single-junction Shockley–Queisser limits,5 as well as LEDs with wide spectral tunability.6,7 However, significant research remains before practical realization of devices based on such materials. Despite several desirable physical and electronic properties, solar-relevant perovskite materials undergo deleterious chemical reactions that degrade performance and long-term stability.

Critical challenges facing the lead halide perovskites are

decomposition and degradation associated with ambient environmental conditions including humidity, oxygen, and thermal stressors.8-14

Significant efforts address

passivation by surface functionalization with chemical methods including thiol functionalization;15 hydrophobic, fluorinated organic salts;16 and quaternary ammonium halides.17 However, a lack of understanding the fundamental surface composition and reactivity at the atom-and-bond level hinders the effective functionalization and passivation of the surfaces.

Functionalization via Atomic Layer Deposition (ALD)

enables passivation with carrier selectivity or blocking, and demonstrates promise.18 ALD-deposited, low-defect, electron-transport layers of TiO2 yielded increased solar energy conversion efficiency as compared to TiO2 deposited via other methods for methylammonium lead iodide-based devices.19 Notably, ALD film quality strongly influences device performance, and variations between deposition techniques for titania-based electron-transport layers yield drastic differences in subsequent device efficiency.19-20

Together, the ambient instability at perovskite surfaces and the

challenges associated with their passivation drive fundamental studies into the 3 ACS Paragon Plus Environment


Page 4 of 45

interfacial chemical species, their chemistry, and how their chemistry may be exploited to effect stable, long-term passivation with tunable electronic properties as desired. Knowledge of the interfacial chemical states and their reactivity should enable chemical routes to forming strong covalent or ionic bonds to species that prevent oxygen or water contact with the surface over the timescales required of commercial photovoltaics. Further, the wide energy tunability afforded by organic semiconductors and their chemical derivatization should yield efficient, carrier selective transfer across interfaces to a wide range of inorganic and hybrid perovskites. Researchers have studied the surfaces of several lead-containing and lead-free solar-and-LED-relevant perovskites. For methylammonium (MA) lead iodide, MAPbI3, surfaces relevant to those under study, individual, tetragonal, (100)-parallel layers carry alternating MA2Pb2I2 stoichiometry with a 4+ net charge and I4 with a 4– net charge, which implies the need for interfacial reconstruction.21

In contrast to tetragonal

MAPbI3(100), cubic CsSnBr3 layers that are (100)-parallel carry alternating CsBr and SnBr2 stoichiometry that are charge neutral,22 which suggests that charge neutralization need not drive any rearrangement to minimize surface energy. Considering tetragonal MAPbI3(100) as would exist at room temperature from a synthesis in γ-butylrolactone,23 early computational studies conclude that Pb-deficient and PbI2-rich surfaces may coexist.24

Computational studies of tetragonal MAPbI3(001) suggest lower surface

energy for methylammonium iodide surface termination.25-26

Importantly slices of

tetrahedral MAPbI3(001) demonstrates identical layer stoichiometry and nearly identical structure as layers of {100} family of cubic MAPbI3,27 and we refer to the tetrahedral MAPbI3(001) as having a pseudo-cubic surface. Similarly DFT studies starting from cubic CsPbBr3 note a relaxation into an orthorhombic but still pseudo-cubic structure that terminates in CsBr.28

High-resolution STM images and DFT calculations for 4 ACS Paragon Plus Environment

Page 5 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


MAPbI3 thin films deposited on Au(111) suggest the adoption of a tetragonal (001) structure with terminal MAI stoichiometry. Importantly, that work predicts highly oriented interfacial methylammonium groups with the ammonium group directed towards the bulk and the methyl group oriented away from the bulk in a surfactant-like geometry.29 This insight does not include orienting effects of an applied electric field,30 nor the factors that influence bulk orientation. Among lead-free perovskite materials, tin-based perovskites are particularly attractive for their range of band gap energies and low exciton-binding energies,31-32 however the interfacial Sn2+ in CsSnX3 (X= Cl, Br, I) is prone to ambient oxidation to Sn4+.33

Researchers addressed tin-perovskite-based

instabilities with the addition SnF2 to fill the cation vacancies,34 and by employing hydrazine as reducing agent during fabrication.35 Recent insight demonstrates that MAPbI3 surfaces can limit carrier lifetimes relative to bulk values,36 but the chemical species that create the energetic states that limit performance remain unclear. Thus, fundamental studies regarding the surface reactivity, including grain reformation chemistry,37 remain underexplored including the exact stoichiometry of tetragonal MAPbI3(100) surfaces as commonly prepared by near-ambient-temperature methods.. Elucidation of the interfacial chemical species and their reactivity for lead-and-tin-based perovskites should enable optimal passivation, carrier selectivity/collection, band-edge alignment, and band gap tuning.36,38-40 Such studies motivate the present investigation. Herein, we synthesize tetragonal, single-crystal MAPbI3 and polycrystalline CsSnBr3, and utilize surface-specific orthogonal reaction chemistry,41-42 to elucidate the chemical species present at each surface, and the reactivity of those species. We chose MAPbI3 for the ability to synthesize large-faceted single crystals that are ideal for surface analyses, and for its relevance in photovoltaics research. We chose CsSnBr3 to compare and contrast the reactivity on the hybrid inorganic-organic MAPbI3 with the 5 ACS Paragon Plus Environment


Page 6 of 45

purely inorganic CsSnBr3 for which little surface insight exists. For oxide-free surfaces, the possible interfacial species for a perovskite of stoichiometry ABX3 include the A+ cation, i.e. methylammonium or Cs+, the B2+ cation, i.e. Pb2+ or Sn2+, or the X– halide, i.e. iodide or bromide. Specifically, reactions with the Lewis-acid BF3 should yield Lewis adducts with halides (Lewis bases) and serve as indicator of interfacial halides. Reactions with p-trifluoromethylanilinium chloride probe the availability and reactivity of A-position cations and their sites.

Exposure to 4,4’-bis(trifluoromethyl)-2,2’-

bipyridine that can ligate to d-block and p-block metal cations probes for interfacial Pb2+ or Sn2+.43-47

For each reaction, x-ray photoelectron spectroscopy (XPS) probes for

fluorine as a proxy for the extent of surface reactivity. Substrate overlayer models yield coverage values for each adsorbed interfacial species from the XPS data.48-50 Temperature-programmed desorption (TPD) elucidates the Lewis adduct bond strength between BF3 and perovskite surfaces. With these studies, we aim to provide the insight for chemical pathways towards passivation from atmospheric oxidation. We further aim to elucidate chemical hooks by which direct covalent or ionic contacts could be made to organic semiconductors, carrier-selective contacts, or buffers between perovskite light absorbers and a metal oxide contact.

2. Experimental Section The supporting information details the preparation of all chemicals, syntheses, and reaction procedures, as well as the collection and quantification of spectra. 2.1. Perovskite syntheses Inverse-temperature crystallization from methylammonium iodide and lead iodide in γ-butylrolactone (GBL) yielded large, dodecahedral methylammonium lead iodide crystals with 0.1–1 cm2 facets of which the largest was previously ascribed to the (100) 6 ACS Paragon Plus Environment

Page 7 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


face of the tetragonal phase.23,51 Crystals were removed from the GBL solution at the synthesis temperature followed by cooling to room temperature that notably yields the tetragonal or β phase.21 Synthesis via a high-temperature melt in quartz ampoules yielded CsSnBr3 polycrystalline samples with ~0.2–1.5 cm2 faces but no visible facets to an unaided eye.52 Sanding and polishing removed interfacial SnOx as detected in the photoelectron spectra of as-synthesized CsSnBr3. Following synthesis, dicing, and polishing, all samples were used directly or stored in a recirculating glovebox until further use. 2.2. Surface Basicity Study Samples were transferred from their inert environments to the Schlenk line in atmosphere-isolated flasks. Cannula transfer of an ethereal BF3 solution submerged individual wafer samples under an argon ambient at either 25 or 40 °C. Following either 5 min or 60 min exposures, the reagent solution was removed and wafers were rinsed via sequential submersions in diethyl ether. Samples were dried via evacuation. Under an air ambient the MAPbI3 wafers were directly mounted on XPS sample pucks, while the CsSnBr3 samples were anaerobically transferred to a recirculating glovebox for sample mounting and subsequent transfer to the XPS via an air-free sample suitcase (Transfer Engineering and Manufacturing, Inc., Freemont, CA). 2.3. Ammonium Exchange Study Samples were transferred to reagent flasks as in the basicity study and submerged in the p-trifluoromethylanilinium chloride solution for one hour at 25 °C followed by rinsing in DCM, drying under vacuum, and sample mounting/transfer as described above.



Page 8 of 45

2.4. Metal Ligation Study Coupling of 4-(trifluoromethyl)-2-bromopyridine yielded 4,4’-bis(trifluoromethyl)-2,2’bipyridine via literature procedures.53-54 Wafer transfer, exposure to the fluorinated-bpy compound, rinsing, and drying proceeded as above for 60 min exposures of MAPbI3(100) and CsSnBr3 at room temperature.

Separately, the fluorinated bpy

compound was exposed to cesium bromide powder and to tin(II) bromide powder in order to verify that bpy would not ligate to interfacial Cs+ but rather to interfacial Sn2+ on CsSnBr3 surfaces. Photoelectron spectroscopic results detailed in the supporting information section confirms ligation exclusively to the tin(II) bromide surfaces and not to CsBr within XPS-detection limits. 2.5. Grain Reformation of CsSnBr3 Dropwise addition of water to a mixture of methylammonium chloride and sodium hydroxide yielded the methylamine reagent gas at 40 °C. The resulting gas passed through a CaO drying column to remove water vapor and was collected in a liquidnitrogen cooled trap. The CsSnBr3 wafers were transferred to the Schlenk line as above, exposed to ~500 Torr of methylamine for 10 min followed by evacuation and transfer to a powder x-ray diffractometer as detailed below. 2.6.X-ray Photoelectron Spectroscopy (XPS) A PHI5600 XPS system with a third-party data acquisition system (RBD Instruments, Bend, Oregon) acquired all photoelectron spectra as detailed previously.55 Acquisitions included wide-energy survey scans as well as high-resolution scans of the C 1s, O 1s, and F 1s regions for all samples and the B 1s, Br 3d, Cs 3d, Cl 2p, I 3d, N 1s, Pb 4f, and Sn 3d regions as necessary. A substrate overlayer model converted photoelectrondetermined peak area ratios to fractional monolayer coverage values for the reagent molecules that functionalized the MAPbI3 and CsSnBr3 surfaces.48-50 The supporting 8 ACS Paragon Plus Environment

Page 9 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


information details the model, key parameters, and methods for extracting coverage values from raw XPS data. 2.7.Temperature-Programmed Desorption Temperature-programmed desorption (TPD) quantified the strength of interaction between adsorbed BF3 and perovskite substrates.56

Detailed in the supporting

information, the TPD experiments utilize an in-house-modified chamber affixed onto the XPS analysis chamber. Throughout a 0.33 K s–1 heating ramp, a 70 eV electron impact quadrupole ion guide mass spectrometer monitors masses 48 and 49 m/z as the primary ionization products 10BF2+ and 11BF2+ of BF3, as well as 50 and 51 m/z as controls. Under the assumption of first-order desorption, an Arrhenius-style model quantifies the activation energy for desorption.56 2.8.Powder X-ray diffraction (XRD) A Bruker-AXS D8 focus powder X-ray diffractometer with Cu–Kα1 radiation collected xray diffraction data in the range of 10–60° (2θ) with a step size of 0.05° and 5 s. X-ray tube operating conditions were 40 kV and 40 mA. In the nitrogen-purged recirculating glovebox, quartz ampoules from the CsSnBr3 synthesis were broken and several crystals were added to an oven-dried mortar. Following grinding, the resulting black powder was transferred to a sample stage and covered using low-static Kapton tape (KaptonTape.com, Torrance, CA) to minimize contact with O2 during data acquisition. Powder XRD data of CsBr and SnBr2 powders were separately collected using similar acquisition and preparation conditions.



Page 10 of 45

3. Results 3.1. Adsorption of BF3 X-ray photoelectron spectroscopy quantifies the presence of fluorine on GBLsynthesized MAPbI3(100) surfaces following exposure to an ethereal BF3 solution. No attempt was made via heating or ion sputtering to minimize signals at 284.8–285.0 eV ascribed to adventitious carbon adsorption. Figure 1 presents the F 1s, I 3d5/2, N 1s, C 1s, B 1s, and Pb 4f regions of characteristic photoelectron spectra for a GBL-synthesized MAPbI3(100) face following a 60 min 40 °C exposure to the BF3 solution (frame A, top); spectra following a 5 min, 25 °C exposure (frame B), and a face with no BF3 exposure (frame C, bottom). For both the BF3-exposed and non-exposed samples in Fig. 1, the I 3d, N 1s, C 1s, and Pb 4f spectral regions resemble previously reported photoelectron spectra of methylammonium lead iodide surfaces. Scans quickly acquired data in the Pb 4f region to minimize beam-damage-induced observation of Pb0 features, however some spectra demonstrate a small quantity of low-binding-energy features as in the Pb 4f region of frame C. Importantly for the adsorption models discussed below, the Pb 4f region demonstrates no quantifiable feature towards higher binding energy that could be ascribable to highly oxidized lead cations. O 1s spectra corresponding to each Pb 4f region demonstrate no features ascribable to metal oxides (not shown). Unique to the BF3-exposed surface in frames A and B, we ascribe the feature at ~687 eV to BF3 adsorption that is absent in the F 1s region for frame C and well matches that from prior reports of BF3-containing compounds.57 The 1:3 boron-to-fluorine atom ratio, a 0.159:1 boron-to-fluorine sensitivity factor ratio, and the lower F 1s signal in frame B as compared to frame A contributes to no observation of B 1s in frame B over a signal background. For five samples exposed to BF3 for 60 min at 40 °C similarly to those in frame A, the average F 1s-to-I 3d peak area ratio was 0.12 ± 0.06. For five samples 10 ACS Paragon Plus Environment

Page 11 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


exposed to BF3 for 5 min at 25 °C similarly to those in frame B, the F 1s-to-I 3d peak area ratio was 0.032 ± 0.004.

Figure 1. XP spectra of (A) a GBL-synthesized MAPbI3(100) face following an ethereal BF3 treatment for 60 min at 40 °C, (B) a MAPbI3(100) face following a 5-min, 25 °C-BF3 treatment, and (C) an untreated MAPbI3(100) sample. A correlation between F 1s feature areas and treatment conditions indicates a reaction between BF3 and the perovskite. The scale bar denotes 5000 cps for the I 3d5/2 region with others magnified as indicated.

A substrate overlayer model further interprets the XP spectra in Fig. 1. We consider BF3 adsorption to surface-available halides, however that requires particular assumptions regarding the halide surface density on each of the tetragonal MAPbI3(100) and polycrystalline CsSnBr3 surfaces. The supporting information section details the assumptions and analyses for halide coverage.

Briefly, a tetragonal (100)-parallel layer of MAPbI3 that carries MA2Pb2I2

stoichiometry and a 4+ elementary charge per unit cell, while the layer that carries I4 stoichiometry possesses a 4– charge per unit cell and twice the iodide surface density as the former. A cubic MAPbI3(100) layer with MAI stoichiometry that is charge neutral has an approximately average iodide surface density between the tetragonal MA2Pb2I2 and I4 layers that we take to approximate the surface density of iodide ions on an idealized MAPbI3 surface. Based on the overlayer model detailed in the supplementary information this model of a



“complete” monolayer coverage of BF3 would have a F 1s-to-I 3d5/2 peak area ratio of 0.089.

Based on this “ideal” fluorine-to-iodide monolayer ratio, we interpret the results for the 60 min exposures at 40 °C (similar to Fig. 1A) to be well in excess of monolayer adsorption. This interpretation considers the comparably smooth but not atomically flat GBL-synthesized MAPbI3(100) SEM included in the supporting information that versus the highly textured polycrystalline CsSnBr3 surfaces under study. We discount the possibility of multilayer formation due to weak intermolecular forces between adjacent BF3 molecules that would likely not survive evacuation in an XPS load lock. With such strong F 1s signals due to 60-min-40-°C exposures, we do not utilize the overlayer equation to assign a coverage but assume near-surface infiltration of BF3 as discussed below. In contrast to the 60-min-40-°C exposures, a 5 min exposure at room temperature (similar to Fig. 1B) yields a limited, sub-monolayer coverage of BF3. From the overlayer model, the average F 1s-to-I 3d peak area ratio of 0.032 ± 0.004 corresponds to a coverage of 40 ± 5% on a model MAPbI3(100) surface. Considering that “real” MAPbI3(100) surfaces such as those provided by inverse-temperature crystallization in GBL have step edges, pits, and defects that should lead to an increased amount of interfacial halide sites and opportunistic BF3 bonding opportunities, we assume that this 40% value represents an upper bound for “real” surface adsorption for the given reaction conditions. Figure 2 presents XP spectra of a representative melt-synthesized polycrystalline CsSnBr3 sample following a 60 min exposure to an ethereal BF3 solution at room temperature (frame A, top) and a sample with no BF3 exposure (frame B, bottom). Figure 2 includes the Cs 3d5/2, F 1s, Sn 3d5/2, C 1s, and Br 3d photoelectron regions without the B 1s region as Br 3p photoelectron signals obstruct and overwhelm photoelectrons ascribed to boron. Importantly, both the BF3-exposed and non-exposed spectra demonstrate a small feature at 487 eV that we attribute to nominal near12 ACS Paragon Plus Environment

Page 12 of 45

Page 13 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


interfacial oxidation of Sn2+ to Sn4+. Interestingly, a single Gaussian-Lorentzian d-style doublet poorly describes the Br 3d spectral region. Instead of one chemical feature, a doublet shown in red and a second, smaller doublet at higher binding energy shown in blue better fit the Br 3d region. Rather than the possibility of oxidized bromide species, we attribute the red feature to photoelectrons from bulk bromide, and the blue feature to interfacial bromide species that may be non-ideally terminated at the CsSnBr3 surface. Importantly, and unlike BF3 interactions with GBL-synthesized MAPbI3(100), reactions between BF3 and melt-synthesized polycrystalline CsSnBr3 demonstrate little dependence on temperature between 25 and 40 °C and little dependence on time between 5 and 60 min exposures. For four samples treated to a 60 min, 25 °C BF3 solution, spectra yielded an average F 1s-to-Br 3d ratio of 0.73 ± 0.13.

Figure 2. XP spectra of (A) a melt-synthesized polycrystalline CsSnBr3 surface following an ethereal BF3 treatment, and (B) with no BF3 exposure. We attribute the F 1s spectral feature at 687 eV in frame A to adsorbed molecular BF3. Both BF3-treated and non-treated samples demonstrate a small feature in the Sn 3d5/2 region at 487 eV ascribed to Sn4+. The scale bar represents 1500 cps for Cs 3d5/2.

For adsorption to an idealized CsSnBr3(100) surface in which one monolayer of adsorption corresponds to one BF3 molecule per atomically flat (100) unit cell area, a 13 ACS Paragon Plus Environment


Page 14 of 45

substrate overlayer model yields a F 1s-to-Br 3d ratio of 0.079 for 100% BF3 coverage as detailed in the supporting information. The experimentally determined F 1s-to-Br 3d ratio yields a calculated monolayer coverage of 550 ± 80%.

For clarification, the

exponential terms in overlayer models yield non-linear relationships between overlayer coverages and the resulting intensity ratios.

As mentioned above, the supporting

information includes demonstrative SEM images from both the GBL-synthesized MAPbI3(100) and melt-synthesized polycrystalline CsSnBr3 surfaces under study. The CsSnBr3 samples demonstrate significant surface roughness that we qualitatively assume are in excess of a 5:1 geometric area ratio compared to an atomically flat CsSnBr3(100). Thus, we ascribe the calculated coverage as an artifact of the simplistic overlayer model and interpret the results in Fig. 2 as that of interfacial adsorption of BF3 to CsSnBr3.

3.2. Desorption and interaction energy of adsorbed BF3 Temperature-programmed desorption elucidates the interaction energy of BF3 with GBL-synthesized MAPbI3(100) and with melt-synthesized polycrystalline CsSnBr3 surfaces. Desorption experiments were conducted on at least three samples for each specific exposure/perovskite combination. Figure 3 presents representative desorption traces for (frame A) GBL-synthesized MAPbI3(100) following a 60-min, 40-°C exposure to ethereal BF3; (frame B) MAPbI3(100) following a 5-min, 25-°C exposure; and (frame C) polycrystalline CsSnBr3 following a 5-min, 25-°C exposure. In frames A–C, the larger trace corresponds to the desorption signal at 49 m/z and the smaller trace corresponds to the desorption signal at 48 m/z. Frame B magnifies the desorption traces 10× relative to the traces presented in frames A and C. For comparison to Redhead-style desorption,56 frame D includes model first-order desorption traces for the experimentally utilized 0.33 K s–1 heating rate and a 1013 s–1 preexponential factor. The red, black, green, and 14 ACS Paragon Plus Environment

Page 15 of 45 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


blue dashed traces in Fig. 3C respectively correspond to desorption with Ea values of 185, 200, 215, and 230 kJ mol–1. For clarity, Fig. 3 does not include the acquired traces at 50 and 51 m/z that serve as controls for the presented data which do not demonstrate the same shape as the traces at 48 and 49 m/z. The results in Fig. 3 support the desorption of BF3 from the perovskites. The ~4:1 signal intensity ratios for the 49 m/z vs 48 m/z traces correspond to terrestrial ratios of 11

B to 10B, and matches the expected intensities for BF2+ that is the primary ionization

product of BF3 electron impact.58 Comparing the desorption from GBL-synthesized MAPbI3 surfaces in frames A and B, the stronger intensity in frame A vs frame B corroborates the higher relative BF3 signals in the photoelectron spectra in Fig. 1A–B. The traces in frame A demonstrate maximum desorption intensity at ~350 °C and shapes that are significantly broader than the model traces in frame D. The trace in frame B only demonstrates the expected isotopic pattern over a narrow temperature range between 260 and 320 °C with a peak at ~310 °C as highlighted by the arrow. Importantly for frame B, control traces at 50 and 51 m/z demonstrate the similarly erratic behavior observed between 320 and 380 °C indicating that the spectral features in that temperature range do not correspond to BF3 detection. We interpret the arrowhighlighted spectral feature in frame B to result from BF3 desorption with other features resulting from the breakdown of the MAPbI3 sample.

Thermogravimetric studies

reported the decomposition of MAPbI3 and evolution of methylammonium iodide in a process that initiates at

Elucidation of Chemical Species and Reactivity at Methylammonium

Recommend Documents