Discovering novel halide perovskite alloys using multi-fidelity machine learning and genetic algorithm.

Expanding the pool of stable halide perovskites with attractive optoelectronic properties is crucial to addressing current limitations in their performance as photovoltaic (PV) absorbers. In this article, we demonstrate how a high-throughput density functional theory (DFT) dataset of halide perovskite alloys can be used to train accurate surrogate models for property prediction and subsequently perform inverse design using genetic algorithm (GA). Our dataset consists of decomposition energies, bandgaps, and photovoltaic efficiencies of nearly 800 pure and mixed composition ABX3 compounds from both the GGA-PBE and HSE06 functionals, and are combined with ∼100 experimental data points collected from the literature. Multi-fidelity random forest regression models are trained on the DFT + experimental dataset for each property using descriptors that one-hot encode composition, phase, and fidelity, and additionally include well-known elemental or molecular properties of species at the A, B, and X sites. Rigorously optimized models are deployed for experiment-level prediction over >150 000 hypothetical compounds, leading to thousands of promising materials with low decomposition energy, band gap between 1 and 2 eV, and efficiency of >15%. Surrogate models are further combined with GA using an objective function to maintain chemical feasibility, minimize decomposition energy, maximize PV efficiency, and keep bandgap between 1 and 2 eV; thus, hundreds more optimal compositions and phases are discovered. We present an analysis of the screened and inverse-designed materials, visualize ternary phase diagrams generated for many systems of interest using machine learning predictions, and suggest strategies for further improvement and expansion in the future.

Two-Dimensional Halide Pb-Perovskite-Double Perovskite Epitaxial Heterostructures.

Epitaxial heterostructures of two-dimensional (2D) halide perovskites offer a new platform for studying intriguing structural, optical, and electronic properties. However, difficulties with the stability of Pb- and Sn-based heterostructures have repeatedly slowed the progress. Recently, Pb-free halide double perovskites are gaining a lot of attention due to their superior stability and greater chemical diversity, but they have not been successfully incorporated into epitaxial heterostructures for further investigation. Here, we report epitaxial core-shell heterostructures via growing Pb-free double perovskites (involving combinations of Ag(I)-Bi(III), Ag-Sb, Ag-In, Na-Bi, Na-Sb, and Na-In) around Pb perovskite 2D crystals. Distinct from Pb-Pb and Pb-Sn perovskite heterostructures, growths of the Pb-free shell at 45° on the (100) surface of the lead perovskite core are observed in all Pb-free cases. The in-depth structural analysis carried out with electron diffraction unequivocally demonstrates the growth of the Pb-free shell along the [110] direction of the Pb perovskite, which is likely due to the relatively lower surface energy of the (110) surface. Furthermore, an investigation of anionic interdiffusion across heterostructure interfaces under the influence of heat was carried out. Interestingly, halide anion diffusion in the Pb-free 2D perovskites is found to be significantly suppressed as compared to Pb-based 2D perovskites. The great structural tunability and excellent stability of Pb-free perovskite heterostructures may find uses in electronic and optoelectronic devices in the near future.

A framework for materials informatics education through workshops.

Abstract: The burgeoning field of materials informatics necessitates a focus on educating the next generation of materials scientists in the concepts of data science, artificial intelligence (AI), and machine learning (ML). In addition to incorporating these topics in undergraduate and graduate curricula, regular hands-on workshops present the most effective medium to initiate researchers to informatics and have them start applying the best AI/ML tools to their own research. With the help of the Materials Research Society (MRS), members of the MRS AI Staging Committee, and a dedicated team of instructors, we successfully conducted workshops covering the essential concepts of AI/ML as applied to materials data, at both the Spring and Fall Meetings in 2022, with plans to make this a regular feature in future meetings. In this article, we discuss the importance of materials informatics education via the lens of these workshops, including details such as learning and implementing specific algorithms, the crucial nuts and bolts of ML, and using competitions to increase interest and participation.

Photocatalytic Ketyl Radical Initiated Cketyl -Csp2 /Csp3 Coupling on ZnIn2 S4.

Visible-light-driven C-C bond formation utilizing ketyl radical (Cketyl ) species has attracted increasing attention recently, as it provides a direct route for the synthesis of complex molecules. However, the most-developed homogeneous photocatalytic systems for the generation and utilization of ketyl radicals usually entail noble metal-based (e. g., Ru and Ir) photosensitizers, which suffer from not only high cost but also potential degradation and hence pose challenges in product separation and purification. In contrast, readily accessible, inexpensive, and recyclable semiconductors represent a class of attractive and alternative photocatalysts but remain much less explored for photocatalytic ketyl radical initiated C-C bond formation. This work demonstrates that a wide range of industrially important chemicals, including substituted chromanes and tertiary alcohols, can be produced on ZnIn2 S4 under visible light irradiation through intramolecular cyclization (Cketyl -Csp2 ) and intermolecular cross-coupling (Cketyl -Csp3 ) reactions, respectively, using ketyl radicals. A suite of experimental studies aided by computational investigation were carried out to shed light on the mechanistic insights of these two types of ketyl radical initiated C-C coupling reactions on ZnIn2 S4 .

Universal machine learning framework for defect predictions in zinc blende semiconductors.

We develop a framework powered by machine learning (ML) and high-throughput density functional theory (DFT) computations for the prediction and screening of functional impurities in groups IV, III-V, and II-VI zinc blende semiconductors. Elements spanning the length and breadth of the periodic table are considered as impurity atoms at the cation, anion, or interstitial sites in supercells of 34 candidate semiconductors, leading to a chemical space of approximately 12,000 points, 10% of which are used to generate a DFT dataset of charge dependent defect formation energies. Descriptors based on tabulated elemental properties, defect coordination environment, and relevant semiconductor properties are used to train ML regression models for the DFT computed neutral state formation energies and charge transition levels of impurities. Optimized kernel ridge, Gaussian process, random forest, and neural network regression models are applied to screen impurities with lower formation energy than dominant native defects in all compounds.

Machine learning for impurity charge-state transition levels in semiconductors from elemental properties using multi-fidelity datasets.

Quantifying charge-state transition energy levels of impurities in semiconductors is critical to understanding and engineering their optoelectronic properties for applications ranging from solar photovoltaics to infrared lasers. While these transition levels can be measured and calculated accurately, such efforts are time-consuming and more rapid prediction methods would be beneficial. Here, we significantly reduce the time typically required to predict impurity transition levels using multi-fidelity datasets and a machine learning approach employing features based on elemental properties and impurity positions. We use transition levels obtained from low-fidelity (i.e., local-density approximation or generalized gradient approximation) density functional theory (DFT) calculations, corrected using a recently proposed modified band alignment scheme, which well-approximates transition levels from high-fidelity DFT (i.e., hybrid HSE06). The model fit to the large multi-fidelity database shows improved accuracy compared to the models trained on the more limited high-fidelity values. Crucially, in our approach, when using the multi-fidelity data, high-fidelity values are not required for model training, significantly reducing the computational cost required for training the model. Our machine learning model of transition levels has a root mean squared (mean absolute) error of 0.36 (0.27) eV vs high-fidelity hybrid functional values when averaged over 14 semiconductor systems from the II-VI and III-V families. As a guide for use on other systems, we assessed the model on simulated data to show the expected accuracy level as a function of bandgap for new materials of interest. Finally, we use the model to predict a complete space of impurity charge-state transition levels in all zinc blende III-V and II-VI systems.

Defective Ultrathin ZnIn2 S4 for Photoreductive Deuteration of Carbonyls Using D2 O as the Deuterium Source.

Deuterium (D) labeling is of great value in organic synthesis, pharmaceutical industry, and materials science. However, the state-of-the-art deuteration methods generally require noble metal catalysts, expensive deuterium sources, or harsh reaction conditions. Herein, noble metal-free and ultrathin ZnIn2 S4 (ZIS) is reported as an effective photocatalyst for visible light-driven reductive deuteration of carbonyls to produce deuterated alcohols using heavy water (D2 O) as the sole deuterium source. Defective two-dimensional ZIS nanosheets (D-ZIS) are prepared in a surfactant assisted bottom-up route exhibited much enhanced performance than the pristine ZIS counterpart. A systematic study is carried out to elucidate the contributing factors and it is found that the in situ surfactant modification enabled D-ZIS to expose more defect sites for charge carrier separation and active D-species generation, as well as high specific surface area, all of which are beneficial for the desirable deuteration reaction. This work highlights the great potential in developing low-cost semiconductor-based photocatalysts for organic deuteration in D2 O, circumventing expensive deuterium reagents and harsh conditions.

First principles modeling of polymer encapsulant degradation in Si photovoltaic modules.

An outstanding issue in the longevity of photovoltaic (PV) modules is the accelerated degradation caused by the presence of moisture. Moisture leads to interfacial instability, de-adhesion, encapsulant decomposition, and contact corrosion. However, experimental characterization of moisture in PV modules is not trivial and its impacts can take years or decades to establish in the field, presenting a major obstacle to designing high-reliability modules. First principles calculations provide an alternative way to study the ingress of water and its detrimental effect on the structure and decomposition of the polymer encapsulant and interfaces between the encapsulant and the semiconductor, the metal contacts, or the dielectric layer. In this work, we use density functional theory (DFT) computations to model single chain, crystalline and cross-linked structures, infrared (IR) signatures, and degradation mechanisms of ethylene vinyl acetate (EVA), the most common polymer encapsulant used in Si PV modules. IR-active modes computed for low energy EVA structures and possible decomposition products match well with reported experiments. The EVA decomposition energy barriers computed using the Nudged Elastic Band (NEB) method show a preference for acetic acid formation as compared to acetaldehyde, are lowered in the presence of a water solvent or hydroxyl ion catalyst, and match well with reported experimental activation energies. This systematic study leads to a clear picture of the hydrolysis-driven decomposition of EVA in terms of energetically favorable mechanisms, possible intermediate structures, and IR signatures of reactants and products.

Spectroscopic Comparison of Thermal Transport at Organic-Inorganic and Organic-Hybrid Interfaces Using CsPbBr3 and FAPbBr3 (FA = Formamidinium) Perovskite Nanocrystals.

Thermal transport across interfaces depends on the matching of vibrational structure at the interface. This work examines the transfer of thermal excitation from an organic ligand coating to either all-inorganic cesium lead tribromide (CsPbBr3) nanocrystals or hybrid organic-inorganic formamidinium lead tribromide (FAPbBr3) nanocrystals using selective infrared optical excitation. These two semiconductors are directly compared because they (or similar semiconductors) are currently envisioned as strong candidates in many optoelectronic technologies and they differ due to the presence of an organic or inorganic cation, which introduces substantial differences in the phonon density of states in otherwise quite similar semiconductors. Infrared excitation of C-H vibrations of surface-bound ligands generates a temperature gradient between the organic ligand shell and nanocrystal core, which results in heat flow, measured by probing changes of the semiconductor band gap. Heat transfer to both perovskite compositions of comparable sizes is similar (25-30 ps), due to fast intramolecular vibrational relaxation and similar matching of low-energy phonons with the organic ligand, but FAPbBr3 samples show a slow bleaching kinetic on the order of 1 ns. This slow, heat-induced change in the semiconductor band gap is attributed not to interfacial heat transfer but instead to thermal equilibration between the organic and inorganic sublattices of FAPbBr3. Ab initio molecular dynamics simulations support the hypothesis that low-energy inorganic sublattice phonon modes are populated initially in the heat transfer process, with a slow thermal population of the higher-energy phonon modes associated primarily with the organic cation. Slow thermal equilibration of FAPbBr3 is likely to substantially impact the time-dependent response of optoelectronic devices that heat the semiconductor active layer and provide further evidence that the poor bulk thermal transport of hybrid perovskite materials extends to microscopic thermal processes.

Charge Transfer Dynamics of Phase-Segregated Halide Perovskites: CH3NH3PbCl3 and CH3NH3PbI3 or (C4H9NH3)2(CH3NH3) n-1Pb nI3 n+1 Mixtures.

Lead halide perovskites present a versatile class of solution-processable semiconductors with highly tunable bandgaps that span ultraviolet, visible, and near-infrared portions of the spectrum. We explore phase-separated chloride and iodide lead perovskite mixtures as candidate materials for intermediate band applications in future photovoltaics. X-ray diffraction and scanning electron microscopy reveal that deposition of precursor solutions across the MAPbCl3/MAPbI3 composition space affords quasi-epitaxial cocrystallized films, in which the two perovskites do not alloy but instead remain phase-segregated. First-principle calculations further support the formation of an epitaxial interface and predict energy offsets in the valence band and conduction band edges that could result in intermediate energy absorption. The charge dynamics of variable mixtures of the relatively narrow bandgap (1.57 eV) MAPbI3 perovskite and wide bandgap (3.02 eV) MAPbCl3 are probed to map charge and energy flow direction and kinetics. Time-resolved photoluminescence and transient absorption measurements reveal charge transfer of photoexcited carriers in MAPbCl3 to MAPbI3 in tens of picoseconds. The rate of quenching can be further tuned by replacing MAPbI3 with two-dimensional Ruddlesden-Popper (BA)2(MA) n-1Pb nI3 n+1 ( n = 3, 2, and 1) perovskites, which also remain phase-separated.

Infrared-pump electronic-probe of methylammonium lead iodide reveals electronically decoupled organic and inorganic sublattices.

Organic-inorganic hybrid perovskites such as methylammonium lead iodide (CH3NH3PbI3) are game-changing semiconductors for solar cells and light-emitting devices owing to their defect tolerance and exceptionally long carrier lifetimes and diffusion lengths. Determining whether the dynamically disordered organic cations with large dipole moment benefit the optoelectronic properties of CH3NH3PbI3 has been an outstanding challenge. Herein, via transient absorption measurements employing an infrared pump pulse tuned to a methylammonium vibration, we observe slow, nanosecond-long thermal dissipation from the selectively excited organic mode to the inorganic sublattice. The resulting transient electronic signatures, during the period of thermal-nonequilibrium when the induced thermal motions are mostly concentrated on the organic sublattice, reveal that the induced atomic motions of the organic cations do not alter the absorption or the photoluminescence response of CH3NH3PbI3, beyond thermal effects. Our results suggest that the attractive optoelectronic properties of CH3NH3PbI3 mainly derive from the inorganic lead-halide framework.

Heterogeneous mesoporous manganese/cobalt oxide catalysts for selective oxidation of 5-hydroxymethylfurfural to 2,5-diformylfuran.

We report a heterogeneous catalytic protocol for the oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-diformylfuran (DFF) using a mesoporous manganese doped cobalt oxide material. The absence of precious metals and additives, use of air as the sole oxidant, and easy isolation of products, along with proper catalyst reusability, make our catalytic protocol attractive for the selective oxidation of HMF to DFF.

Optimization of Organotin Polymers for Dielectric Applications.

Recently, there has been a growing interest in developing wide band gap dielectric materials as the next generation insulators for capacitors, photovoltaic devices, and transistors. Organotin polyesters have shown promise as high dielectric constant, low loss, and high band gap materials. Guided by first-principles calculations from density functional theory (DFT), in line with the emerging codesign concept, the polymer poly(dimethyltin 3,3-dimethylglutarate), p(DMTDMG), was identified as a promising candidate for dielectric applications. Blends and copolymers of poly(dimethyltin suberate), p(DMTSub), and p(DMTDMG) were compared using increasing amounts of p(DMTSub) from 10% to 50% to find a balance between electronic properties and film morphology. DFT calculations were used to gain further insight into the structural and electronic differences between p(DMTSub) and p(DMTDMG). Both blend and copolymer systems showed improved results over the homopolymers with the films having dielectric constants of 6.8 and 6.7 at 10 kHz with losses of 1% and 2% for the blend and copolymer systems, respectively. The energy density of the film measured as a D-E hysteresis loop was 6 J/cc for the copolymer, showing an improvement compared to 4 J/cc for the blend. This improvement is hypothesized to come from a more uniform distribution of diacid repeat units in the copolymer compared to the blend, leading toward improved film quality and subsequently higher energy density.

Critical role of morphology on the dielectric constant of semicrystalline polyolefins.

A particularly attractive method to predict the dielectric properties of materials is density functional theory (DFT). While this method is very popular, its large computational requirements allow practical treatments of unit cells with just a small number of atoms in an ordered array, i.e., in a crystalline morphology. By comparing DFT and Molecular Dynamics (MD) simulations on the same ordered arrays of functional polyolefins, we confirm that both methodologies yield identical estimates for the dipole moments and hence the ionic component of the dielectric storage modulus. Additionally, MD simulations of more realistic semi-crystalline morphologies yield estimates for this polar contribution that are in good agreement with the limited experiments in this field. However, these predictions are up to 10 times larger than those for pure crystalline simulations. Here, we show that the constraints provided by the surrounding chains significantly impede dipolar relaxations in the crystalline regions, whereas amorphous chains must sample all configurations to attain their fully isotropic spatial distributions. These results, which suggest that the amorphous phase is the dominant player in the context, argue strongly that the proper polymer morphology needs to be modeled to ensure accurate estimates of the ionic component of the dielectric constant.

Rational Co-Design of Polymer Dielectrics for Energy Storage.

Although traditional materials discovery has historically benefited from intuition-driven experimental approaches and serendipity, computational strategies have risen in prominence and proven to be a powerful complement to experiments in the modern materials research environment. It is illustrated here how one may harness a rational co-design approach-involving synergies between high-throughput computational screening and experimental synthesis and testing-with the example of polymer dielectrics design for electrostatic energy storage applications. Recent co-design efforts that can potentially enable going beyond present-day "standard" polymer dielectrics (such as biaxially oriented polypropylene) are highlighted. These efforts have led to the identification of several new organic polymer dielectrics within known generic polymer subclasses (e.g., polyurea, polythiourea, polyimide), and the recognition of the untapped potential inherent in entirely new and unanticipated chemical subspaces offered by organometallic polymers. The challenges that remain and the need for additional methodological developments necessary to further strengthen the co-design concept are then presented.

A polymer dataset for accelerated property prediction and design.

Emerging computation- and data-driven approaches are particularly useful for rationally designing materials with targeted properties. Generally, these approaches rely on identifying structure-property relationships by learning from a dataset of sufficiently large number of relevant materials. The learned information can then be used to predict the properties of materials not already in the dataset, thus accelerating the materials design. Herein, we develop a dataset of 1,073 polymers and related materials and make it available at http://khazana.uconn.edu/. This dataset is uniformly prepared using first-principles calculations with structures obtained either from other sources or by using structure search methods. Because the immediate target of this work is to assist the design of high dielectric constant polymers, it is initially designed to include the optimized structures, atomization energies, band gaps, and dielectric constants. It will be progressively expanded by accumulating new materials and including additional properties calculated for the optimized structures provided.

Machine Learning Strategy for Accelerated Design of Polymer Dielectrics.

The ability to efficiently design new and advanced dielectric polymers is hampered by the lack of sufficient, reliable data on wide polymer chemical spaces, and the difficulty of generating such data given time and computational/experimental constraints. Here, we address the issue of accelerating polymer dielectrics design by extracting learning models from data generated by accurate state-of-the-art first principles computations for polymers occupying an important part of the chemical subspace. The polymers are 'fingerprinted' as simple, easily attainable numerical representations, which are mapped to the properties of interest using a machine learning algorithm to develop an on-demand property prediction model. Further, a genetic algorithm is utilised to optimise polymer constituent blocks in an evolutionary manner, thus directly leading to the design of polymers with given target properties. While this philosophy of learning to make instant predictions and design is demonstrated here for the example of polymer dielectrics, it is equally applicable to other classes of materials as well.

Poly(dimethyltin glutarate) as a prospective material for high dielectric applications.

Poly(dimethyltin glutarate) is presented as the first organometallic polymer, a high dielectric constant, and low dielectric loss material. Theoretical results correspond well in terms of the dielectric constant. More importantly, the dielectric constant can be tuned depending on the solvent a film of the polymer is cast from. The breakdown strength is increased through blending with a second organometallic polymer.