Speciation of Lanthanide Metal Ion Dopants in Microcrystalline All-Inorganic Halide Perovskite CsPbCl3.

Lanthanides are versatile modulators of optoelectronic properties owing to their narrow optical emission spectra across the visible and near-infrared range. Their use in metal halide perovskites (MHPs) has recently gained prominence, although their fate in these materials has not yet been established at the atomic level. We use cesium-133 solid-state NMR to establish the speciation of all nonradioactive lanthanide ions (La3+, Ce3+, Pr3+, Nd3+, Sm3+, Sm2+, Eu3+, Eu2+, Gd3+, Tb3+, Dy3+, Ho3+, Er3+, Tm3+, Yb3+, Lu3+) in microcrystalline CsPbCl3. Our results show that all lanthanides incorporate into the perovskite structure of CsPbCl3 regardless of their oxidation state (+2, +3).

De Novo Crystal Structure Determination from Machine Learned Chemical Shifts.

Determination of the three-dimensional atomic-level structure of powdered solids is one of the key goals in current chemistry. Solid-state NMR chemical shifts can be used to solve this problem, but they are limited by the high computational cost associated with crystal structure prediction methods and density functional theory chemical shift calculations. Here, we successfully determine the crystal structures of ampicillin, piroxicam, cocaine, and two polymorphs of the drug molecule AZD8329 using on-the-fly generated machine-learned isotropic chemical shifts to directly guide a Monte Carlo-based structure determination process starting from a random gas-phase conformation.

Graph-convolutional neural networks for (QM)ML/MM molecular dynamics simulations.

To accurately study the chemical reactions in the condensed phase or within enzymes, both quantum-mechanical description and sufficient configurational sampling are required to reach converged estimates. Here, quantum mechanics/molecular mechanics (QM/MM) molecular dynamics (MD) simulations play an important role, providing QM accuracy for the region of interest at a decreased computational cost. However, QM/MM simulations are still too expensive to study large systems on longer time scales. Recently, machine learning (ML) models have been proposed to replace the QM description. The main limitation of these models lies in the accurate description of long-range interactions present in condensed-phase systems. To overcome this issue, a recent workflow has been introduced combining a semi-empirical method (i.e. density functional tight binding (DFTB)) and a high-dimensional neural network potential (HDNNP) in a Δ-learning scheme. This approach has been shown to be capable of correctly incorporating long-range interactions within a cutoff of 1.4 nm. One of the promising alternative approaches to efficiently take long-range effects into account is the development of graph-convolutional neural networks (GCNNs) for the prediction of the potential-energy surface. In this work, we investigate the use of GCNN models - with and without a Δ-learning scheme - for (QM)ML/MM MD simulations. We show that the Δ-learning approach using a GCNN and DFTB as a baseline achieves competitive performance on our benchmarking set of solutes and chemical reactions in water. This method is additionally validated by performing prospective (QM)ML/MM MD simulations of retinoic acid in water and S-adenoslymethionine interacting with cytosine in water. The results indicate that the Δ-learning GCNN model is a valuable alternative for the (QM)ML/MM MD simulations of condensed-phase systems.

Supramolecular Modulation of Hybrid Perovskite Solar Cells via Bifunctional Halogen Bonding Revealed by Two-Dimensional 19F Solid-State NMR Spectroscopy.

There has been an ongoing effort to overcome the limitations associated with the stability of hybrid organic-inorganic perovskite solar cells by using different organic agents as additives to the perovskite formulations. The functionality of organic additives has been predominantly limited to exploiting hydrogen-bonding interactions, while the relevant atomic-level binding modes remain elusive. Herein, we introduce a bifunctional supramolecular modulator, 1,2,4,5-tetrafluoro-3,6-diiodobenzene, which interacts with the surface of the triple-cation double-halide perovskite material via halogen bonding. We elucidate its binding mode using two-dimensional solid-state 19F NMR spectroscopy in conjunction with density functional theory calculations. As a result, we demonstrate a stability enhancement of the perovskite solar cells upon supramolecular modulation, without compromising the photovoltaic performances.

Rapid Structure Determination of Molecular Solids Using Chemical Shifts Directed by Unambiguous Prior Constraints.

NMR-based crystallography approaches involving the combination of crystal structure prediction methods, ab initio calculated chemical shifts and solid-state NMR experiments are powerful methods for crystal structure determination of microcrystalline powders. However, currently structural information obtained from solid-state NMR is usually included only after a set of candidate crystal structures has already been independently generated, starting from a set of single-molecule conformations. Here, we show with the case of ampicillin that this can lead to failure of structure determination. We propose a crystal structure determination method that includes experimental constraints during conformer selection. In order to overcome the problem that experimental measurements on the crystalline samples are not obviously translatable to restrict the single-molecule conformational space, we propose constraints based on the analysis of absent cross-peaks in solid-state NMR correlation experiments. We show that these absences provide unambiguous structural constraints on both the crystal structure and the gas-phase conformations, and therefore can be used for unambiguous selection. The approach is parametrized on the crystal structure determination of flutamide, flufenamic acid, and cocaine, where we reduce the computational cost by around 50%. Most importantly, the method is then shown to correctly determine the crystal structure of ampicillin, which would have failed using current methods because it adopts a high-energy conformer in its crystal structure. The average positional RMSE on the NMR powder structure is ⟨rav⟩ = 0.176 Å, which corresponds to an average equivalent displacement parameter Ueq = 0.0103 Å2.

A Bayesian approach to NMR crystal structure determination.

Nuclear Magnetic Resonance (NMR) spectroscopy is particularly well suited to determine the structure of molecules and materials in powdered form. Structure determination usually proceeds by finding the best match between experimentally observed NMR chemical shifts and those of candidate structures. Chemical shifts for the candidate configurations have traditionally been computed by electronic-structure methods, and more recently predicted by machine learning. However, the reliability of the determination depends on the errors in the predicted shifts. Here we propose a Bayesian framework for determining the confidence in the identification of the experimental crystal structure, based on knowledge of the typical errors in the electronic structure methods. We demonstrate the approach on the determination of the structures of six organic molecular crystals. We critically assess the reliability of the structure determinations, facilitated by the introduction of a visualization of the similarity between candidate configurations in terms of their chemical shifts and their structures. We also show that the commonly used values for the errors in calculated 13C shifts are underestimated, and that more accurate, self-consistently determined uncertainties make it possible to use 13C shifts to improve the accuracy of structure determinations. Finally, we extend the recently-developed ShiftML model to render it more efficient, accurate, and, most importantly, to evaluate the uncertainties in its predictions. By quantifying the confidence in structure determinations based on ShiftML predictions we further substantiate that it provides a valid replacement for first-principles calculations in NMR crystallography.

Phase Segregation in Potassium-Doped Lead Halide Perovskites from 39K Solid-State NMR at 21.1 T.

Organic-inorganic lead halide perovskites are a promising family of light absorbers for a new generation of solar cells, with reported efficiencies currently exceeding 22%. A common problem of solar cells fabricated using these materials is that their efficiency depends on their cycling history, an effect known as current-voltage ( J- V) hysteresis. Potassium doping has recently emerged as a universal way to overcome this adverse phenomenon. While the atomistic origins of J- V hysteresis are still not fully understood, it is essential to rationalize the atomic-level effect of protocols that lead to its suppression. Here, using 39K MAS NMR at 21.1 T we provide for the first time atomic-level characterization of the potassium-containing phases that are formed upon KI doping of multication and multianion lead halide perovskites. We find no evidence of potassium incorporation into 3D perovskite lattices of the recently reported materials. Instead, we observe formation of a mixture of potassium-rich phases and unreacted KI. In the case of Br-containing lead halide perovskites doped with KI, a mixture of KI and KBr ensues, leading to a change in the Br/I ratio in the perovskite phase with respect to the undoped perovskite. Simultaneous Cs and K doping leads to the formation of nonperovskite Cs/K lead iodide phases.

Formation of Stable Mixed Guanidinium-Methylammonium Phases with Exceptionally Long Carrier Lifetimes for High-Efficiency Lead Iodide-Based Perovskite Photovoltaics.

Methylammonium (MA)- and formamidinium (FA)-based organic-inorganic lead halide perovskites provide outstanding performance as photovoltaic materials, due to their versatility of fabrication and their power conversion efficiencies reaching over 22%. The proposition of guanidinium (GUA)-doped perovskite materials generated considerable interest due to their potential to increase carrier lifetimes and open-circuit voltages as compared to pure MAPbI3. However, simple size considerations based on the Goldschmidt tolerance factor suggest that guanidinium is too big to completely replace methylammonium as an A cation in the APbI3 perovskite lattice, and its effect was thus ascribed to passivation of surface trap states at grain boundaries. As guanidinium was not thought to incorporate into the MAPbI3 lattice, interest waned since it appeared unlikely that it could be used to modify the intrinsic perovskite properties. Here, using solid-state NMR, we provide for the first time atomic-level evidence that GUA is directly incorporated into the MAPbI3 and FAPbI3 lattices, forming pure GUA xMA1- xPbI3 or GUA xFA1- xPbI3 phases, and that it reorients on the picosecond time scale within the perovskite lattice, which explains its superior charge carrier stabilization capacity. Our findings establish a fundamental link between charge carrier lifetimes observed in photovoltaic perovskites and the A cation structure in ABX3-type metal halide perovskites.

Positional Variance in NMR Crystallography.

We propose a method to quantify positional uncertainties in crystal structures determined by chemical-shift-based NMR crystallography. The method combines molecular dynamics simulations and density functional theory calculations with experimental and computational chemical shift uncertainties. In this manner we find the average positional accuracy as well as the isotropic and anisotropic positional accuracy associated with each atom in a crystal structure determined by NMR crystallography. The approach is demonstrated on the crystal structures of cocaine, flutamide, flufenamic acid, the K salt of penicillin G, and form 4 of the drug 4-[4-(2-adamantylcarbamoyl)-5-tert-butylpyrazol-1-yl]benzoic acid (AZD8329). We find that, for the crystal structure of cocaine, the uncertainty corresponds to a positional RMSD of 0.17 Å. This is a factor of 2.5 less than for single-crystal X-ray-diffraction-based structure determination.

Phase Segregation in Cs-, Rb- and K-Doped Mixed-Cation (MA)x(FA)1-xPbI3 Hybrid Perovskites from Solid-State NMR.

Hybrid (organic-inorganic) multication lead halide perovskites hold promise for a new generation of easily processable solar cells. Best performing compositions to date are multiple-cation solid alloys of formamidinium (FA), methylammonium (MA), cesium, and rubidium lead halides which provide power conversion efficiencies up to around 22%. Here, we elucidate the atomic-level nature of Cs and Rb incorporation into the perovskite lattice of FA-based materials. We use 133Cs, 87Rb, 39K, 13C, and 14N solid-state MAS NMR to probe microscopic composition of Cs-, Rb-, K-, MA-, and FA-containing phases in double-, triple-, and quadruple-cation lead halides in bulk and in a thin film. Contrary to previous reports, we have found no proof of Rb or K incorporation into the 3D perovskite lattice in these systems. We also show that the structure of bulk mechanochemical perovskites bears close resemblance to that of thin films, making them a good benchmark for structural studies. These findings provide fundamental understanding of previously reported excellent photovoltaic parameters in these systems and their superior stability.

Cation Dynamics in Mixed-Cation (MA)x(FA)1-xPbI3 Hybrid Perovskites from Solid-State NMR.

Mixed-cation organic lead halide perovskites attract unfaltering attention owing to their excellent photovoltaic properties. Currently, the best performing perovskite materials contain multiple cations and provide power conversion efficiencies up to around 22%. Here, we report the first quantitative, cation-specific data on cation reorientation dynamics in hybrid mixed-cation formamidinium (FA)/methylammonium (MA) lead halide perovskites. We use 14N, 2H, 13C, and 1H solid-state MAS NMR to elucidate cation reorientation dynamics, microscopic phase composition, and the MA/FA ratio, in (MA)x(FA)1-xPbI3 between 100 and 330 K. The reorientation rates correlate in a striking manner with the carrier lifetimes previously reported for these materials and provide evidence of the polaronic nature of charge carriers in PV perovskites.

Artificial intelligence for natural product drug discovery.

Developments in computational omics technologies have provided new means to access the hidden diversity of natural products, unearthing new potential for drug discovery. In parallel, artificial intelligence approaches such as machine learning have led to exciting developments in the computational drug design field, facilitating biological activity prediction and de novo drug design for molecular targets of interest. Here, we describe current and future synergies between these developments to effectively identify drug candidates from the plethora of molecules produced by nature. We also discuss how to address key challenges in realizing the potential of these synergies, such as the need for high-quality datasets to train deep learning algorithms and appropriate strategies for algorithm validation.

A Machine Learning Model of Chemical Shifts for Chemically and Structurally Diverse Molecular Solids.

Nuclear magnetic resonance (NMR) chemical shifts are a direct probe of local atomic environments and can be used to determine the structure of solid materials. However, the substantial computational cost required to predict accurate chemical shifts is a key bottleneck for NMR crystallography. We recently introduced ShiftML, a machine-learning model of chemical shifts in molecular solids, trained on minimum-energy geometries of materials composed of C, H, N, O, and S that provides rapid chemical shift predictions with density functional theory (DFT) accuracy. Here, we extend the capabilities of ShiftML to predict chemical shifts for both finite temperature structures and more chemically diverse compounds, while retaining the same speed and accuracy. For a benchmark set of 13 molecular solids, we find a root-mean-squared error of 0.47 ppm with respect to experiment for 1H shift predictions (compared to 0.35 ppm for explicit DFT calculations), while reducing the computational cost by over four orders of magnitude.

Structure determination of an amorphous drug through large-scale NMR predictions.

Knowledge of the structure of amorphous solids can direct, for example, the optimization of pharmaceutical formulations, but atomic-level structure determination in amorphous molecular solids has so far not been possible. Solid-state nuclear magnetic resonance (NMR) is among the most popular methods to characterize amorphous materials, and molecular dynamics (MD) simulations can help describe the structure of disordered materials. However, directly relating MD to NMR experiments in molecular solids has been out of reach until now because of the large size of these simulations. Here, using a machine learning model of chemical shifts, we determine the atomic-level structure of the hydrated amorphous drug AZD5718 by combining dynamic nuclear polarization-enhanced solid-state NMR experiments with predicted chemical shifts for MD simulations of large systems. From these amorphous structures we then identify H-bonding motifs and relate them to local intermolecular complex formation energies.

Ba-induced phase segregation and band gap reduction in mixed-halide inorganic perovskite solar cells.

All-inorganic metal halide perovskites are showing promising development towards efficient long-term stable materials and solar cells. Element doping, especially on the lead site, has been proved to be a useful strategy to obtain the desired film quality and material phase for high efficient and stable inorganic perovskite solar cells. Here we demonstrate a function by adding barium in CsPbI2Br. We find that barium is not incorporated into the perovskite lattice but induces phase segregation, resulting in a change in the iodide/bromide ratio compared with the precursor stoichiometry and consequently a reduction in the band gap energy of the perovskite phase. The device with 20 mol% barium shows a high power conversion efficiency of 14.0% and a great suppression of non-radiative recombination within the inorganic perovskite, yielding a high open-circuit voltage of 1.33 V and an external quantum efficiency of electroluminescence of 10-4.

Chemical shifts in molecular solids by machine learning.

Due to their strong dependence on local atonic environments, NMR chemical shifts are among the most powerful tools for strucutre elucidation of powdered solids or amorphous materials. Unfortunately, using them for structure determination depends on the ability to calculate them, which  comes at the cost of high accuracy first-principles calculations. Machine learning has recently emerged as a way to overcome the need for quantum chemical calculations, but for chemical shifts in solids it is hindered by the chemical and combinatorial space spanned by molecular solids, the strong dependency of chemical shifts on their environment, and the lack of an experimental database of shifts. We propose a machine learning method based on local environments to accurately predict chemical shifts of molecular solids and their polymorphs to within DFT accuracy. We also demonstrate that the trained model is able to determine, based on the match between experimentally measured and ML-predicted shifts, the structures of cocaine and the drug 4-[4-(2-adamantylcarbamoyl)-5-tert-butylpyrazol-1-yl]benzoic acid.

Probing Protein Dynamics Using Multifield Variable Temperature NMR Relaxation and Molecular Dynamics Simulation.

Understanding the interplay between protein function and dynamics is currently one of the fundamental challenges of physical biology. Recently, a method using variable temperature solid-state nuclear magnetic resonance relaxation measurements has been proposed for the simultaneous measurement of 12 different activation energies reporting on distinct dynamic modes in the protein GB1. Here, we extend this approach to measure relaxation at multiple magnetic field strengths, allowing us to better constrain the motional models and to simultaneously evaluate the robustness and physical basis of the method. The data reveal backbone and side-chain motions, exhibiting low- and high-energy modes with temperature coefficients around 5 and 25 kJ·mol-1. The results are compared to variable temperature molecular dynamics simulation of the crystal lattice, providing further support for the interpretation of the experimental data in terms of molecular motion.