Target-Driven Design of Deep-UV Nonlinear Optical Materials via Interpretable Machine Learning

A Machine-Learning-Assisted Crystalline Structure Prediction Framework To Accelerate Materials Discovery

Modern crystal structure prediction methods based on structure generation algorithms and first-principles calculations play important roles in the design of new materials. However, the cost of these methods is very expensive because their success mostly relies on the efficient sampling of structures and the accurate evaluation of energies for those sampled structures. Herein, we develop a Machine-learning-Assisted CRYStalline Materials sAmpling sysTem (MAXMAT) aiming to accelerate the prediction of new crystal structures. For a given chemical composition, MAXMAT can generate efficient crystal structures with the help of a Python package for crystal structure generation (PyXtal) and can quickly evaluate the energies of these generated structures using a well-developed machine learning interaction potential model (M3GNET). We have used MAXMAT to perform crystal structure searches for three different chemical systems (TiO2, MgAl2O4, and BaBOF3) to test its accuracy and efficiency. Furthermore, we apply MAXMAT to predict new nonlinear optical materials, suggesting several thermodynamically synthesizable structures with high performance in LiZnGaS3 and CaBOF3 systems.

Recent advances in rational structure design for nonlinear optical crystals: leveraging advantageous templates

Nonlinear optical (NLO) crystals that can expand the spectral range of laser outputs have attracted significant attention for their optoelectronic applications. The research progress from the discovery of new single crystal structures to the realization of final device applications involves many key steps and is very time consuming and challenging. Consequently, exploring efficient design strategies to shorten the research period and accelerate the rational design of novel NLO materials has become imperative to address the pressing demand for advanced materials. The recent shift in paradigm toward exploring new NLO crystals involves significant progress from extensive "trial and error" methodologies to strategic approaches. This review proposes the concept of rational structure design for nonlinear optical crystals leveraging advantageous templates. It further discusses their optical characteristics, promising applications as second-order NLO materials, and the relationship between their structure and performance, and highlights urgent issues that need to be addressed in the field of NLO crystals in the future. The review aims to provide ideas and driving impetus to encourage researchers to achieve new breakthroughs in the next generation of NLO materials.

Two Hydroxyfluorooxoborates Achieving Deep-Ultraviolet Cutoff Edges and Moderate Birefringence by Assembling Multi-Anionic Groups

Assembling multi-anionic groups is conducive to utilizing respective advantage to achieve the enhancement of optical performance. Two new hydroxyfluorooxoborates, Ama2-Rb2B3O3F4(OH) and K8Cs2B15O14(OH)7F20 ⋅ H2O with [B3O3F4(OH)] six-membered rings were synthesized for the first time. The title compounds exhibit short ultraviolet cutoff edges (<200 nm) and K8Cs2B15O14(OH)7F20 ⋅ H2O possesses a moderate experimental refractive index difference of 0.051@546 nm.

Multi-fidelity machine learning for predicting bandgaps of nonlinear optical crystals

Nonlinear optical (NLO) materials are of great importance in modern optics and industry because of their intrinsic capability of wavelength conversion. Bandgap is a key property of NLO crystals. In recent years, machine learning (ML) has become a powerful tool to predict the bandgaps of compounds before synthesis. However, the shortage of available experimental data of NLO crystals poses a significant challenge for the exploration of new NLO materials using ML. In this work, we proposed a new multi-fidelity ML approach based on the multilevel descriptors developed by us (Z.-Y. Zhang, X. Liu, L. Shen, L. Chen and W.-H. Fang, J. Phys. Chem. C, 2021, 125, 25175-25188) and the gradient boosting regression tree algorithm. The calculated and experimental bandgaps of NLO crystals were collected as the low- and high-fidelity labels, respectively. The experimental values were predicted based on chemical compositions of crystals without prior knowledge about crystal structures. The multi-fidelity ML model overcame the performance of single-fidelity predictor. Furthermore, it was observed that less accurate predictions on the low-fidelity label may result in more accurate prediction on the high-fidelity label, at least in the present case. Using the multi-fidelity ML model with the best performance in this work, the mean absolute error on the test set of experimental bandgaps was 0.293 eV, which is smaller than that using the single-fidelity model (0.355 eV). It is far from perfect but accurate enough as an effective computational tool in the first step to discover novel NLO materials.

Structural strategy for advancing nonlinear optical effects in 1D-[MX2]∞ chains: internal distortion and atomic types

This work aimed to alleviate the limitations of existing mid-infrared nonlinear-optics (MIR-NLO) crystals by conducting theoretical research on 1D-[MX2]∞ (1D = one-dimensional; M = metallic element; X = anionic element) structures in relation to NLO. An analysis was conducted on the electronic structure and optical properties of six selenides (BaZnGeSe4, KxBa1-x/2Ga2Se4, KxBa1-x/2GayIn2-ySe4, KxBa1-x/2In2Se4, BaZnSiSe4, SrZnSiSe4) and three arsenides (Cs2SnAs2, Rb2SnAs2, K2SnAs2) using first principles. Afterwards, the inherent characteristics of these 1D-[MX2]∞ chains were further examined. The calculation results suggest that the presence of internal distortion (non-centrosymmetric) and high electronegativity in M elements can greatly enhance the NLO capability of 1D-[MX2]∞ structures. Furthermore, the first-ever prediction of K2SnAs2's NLO ability has been made. Ultimately, the establishment of a theoretical structure (K2BaSn2As4) provides guidance for the subsequent creation of high-performance MIR-NLO crystals.

Room-Temperature Phase Transition Material with Switchable Second-Order Nonlinear Optical Properties

Phase transition materials with switchable second-order nonlinear optical (NLO) properties have attracted extensive attention because of their great application potential in photoelectric switches, sensors, and modulators, while metal-free organics with NLO switchability near room temperature remain scarce. Herein, we report a hydrogen-bonded metal-free organic crystal, 2-methylpropan-2-aminium 2,2-dimethylpropanoate (1), exhibiting a room-temperature phase transition and favorable NLO switchability. Through investigations on its thermal anomalies, dielectric properties, and crystal structures, we uncover that 1 holds a near-room-temperature phase transition at 303 K from noncentrosymmetric point group C2v to centrosymmetric one D2h, which is attributed to the order-disorder transformations of both tert-butylamine cations and dimethylpropionic acid anions. Accompanied by symmetry change during the phase transition, 1 exhibits reversible and repeatable NLO "on-off" switchability with a desirable switching contrast ratio of ca. 19 between high and low NLO states. This discovery demonstrates a metal-free organic crystal with NLO switching behavior near room temperature, serving as a promising candidate in smart and ecofriendly photoelectric functional materials and devices.

A UV non-hydrogen pure selenite nonlinear optical material for achieving balanced properties through framework-optimized structural transformation

For non-centrosymmetric (NCS) oxides intended for ultraviolet (UV) nonlinear optical (NLO) applications, achieving a wide band gap, large second harmonic generation (SHG) intensity, and sufficient birefringence to satisfy phase matching is a significant challenge due to their inherent incompatibility. To address this issue, this study proposes a strategy called framework-optimized structural transformation. Building upon centrosymmetric (CS) NaGa(SeO3)2 as a foundation, an original UV selenite NLO material, NaLu(SeO3)2, was successfully synthesized. The derived NaLu(SeO3)2 exhibits a balanced comprehensive performance, including a band gap (5.3 eV), an SHG response (2.7 × KDP), a UV cut-off edge (210 nm), a laser-induced damage threshold (LIDT) (151.69 MW cm-2), birefringence (Cal: 0.138@546 nm, Exp: 0.153@546 nm), thermal stability (∼575 °C) and environmental stability. Notably, its SHG effect, band gap, LIDT, and birefringence are all the largest among UV non-hydrogen pure selenite materials. Such progress can be attributed to the successful arrangement of the SeO3 groups by optimizing the cations on the framework of the parent compound.

A Machine Learning Study on High Thermal Conductivity Assisted to Discover Chalcogenides with Balanced Infrared Nonlinear Optical Performance

Exploration of novel nonlinear optical (NLO) chalcogenides with high laser-induced damage thresholds (LIDT) is critical for mid-infrared (mid-IR) solid-state laser applications. High lattice thermal conductivity (κL ) is crucial to increasing LIDT yet often neglected in the search for NLO crystals due to lack of accurate κL data. A machine learning (ML) approach to predict κL for over 6000 chalcogenides is hereby proposed. Combining ML-generated κL data and first-principles calculation, a high-throughput screening route is initiated, and ten new potential mid-IR NLO chalcogenides with optimal bandgap, NLO coefficients, and thermal conductivity are discovered, in which Li2 SiS3 and AlZnGaS4 are highlighted. Big-data analysis on structural chemistry proves that the chalcogenides having dense and simple lattice structures with low anisotropy, light atoms, and strong covalent bonds are likely to possess higher κL . The four-coordinated motifs in which central cations show the bond valence sum of +2 to +3 and are from IIIA, IVA, VA, and IIB groups, such as those in diamond-like defect-chalcopyrite chalcogenides, are preferred to fulfill the desired structural chemistry conditions for balanced NLO and thermal properties. This work provides not only an efficient strategy but also interpretable research directions in the search for NLO crystals with high thermal conductivity.

Cd8(BO3)4SiO4: Metal Cation Inducing the Formation of Isolated [BO3] and [SiO4] Units in Borate Silicate

The first compound of cadmium-borate silicate Cd8(BO3)4SiO4, crystallizing in space group P42/n (no. 86), has been successfully synthesized by the conventional high-temperature solution method and melts congruently. The zero-dimensional anionic groups of Cd8(BO3)4SiO4 are isolated [BO3] triangles and isolated [SiO4] tetrahedra which are filled in the framework formed by [CdO6] polyhedra. It has a moderate birefringence (Δn = 0.053 at 546 nm), which is measured by experiment and evaluated by first-principles calculations; meanwhile, the source of birefringence is revealed through the response electronic distribution anisotropy method. The UV-vis-NIR diffuse reflectance spectrum indicates that Cd8(BO3)4SiO4 possesses a wide optical transparency range, with a UV cutoff edge at about 254 nm. This work enriches the structure chemistry of borate silicates, and we discussed the possible methods for the exploration and synthesis of novel optical crystals possessing zero-dimensional anionic groups in the borate silicate system.

Mg assists in modulating the dimensionalities of the anionic frameworks of borates

Three Mg-containing borates were obtained by high-temperature spontaneous crystallization. In the (A2O)- or (A2O-MO)-MgO-B2O3 system (A is alkali metal and M is alkaline-earth metal) reported in the ICSD, Li4Mg3SrB12O24 is the first compound that contains one-dimensional infinite anionic chains, and the two examples of the isostructural A2Mg3B16O28 (A = Rb, Cs) exhibit a two-dimensional infinite bilayer structure for the first time, which contributes to the enrichment of the structural chemistry of Mg-containing borates. Besides, the results of comparison and analysis in this system clearly show that Mg not only affects the anionic frameworks of borates to produce low-dimensional structures but, together with the ratio of Ncation/NB, is responsible for the dimensionalities of the anionic frameworks in borates. The optical properties of the three compounds also show that they all have short cutoff edges, and Cs2Mg3B16O28, in particular, could reach the deep-ultraviolet region (<200 nm).

Vanadate Crystals with an Enhanced Birefringence and a Broadened Transparency Spectrum through Controlled [VO3]∞ Chain Arrangements and Alkali Metal Cation Introduction

The vanadate (VO) polyhedron offers a compelling avenue for exploring birefringent materials within the infrared frequency range. Among many potential building blocks, the implementation of [VO3]∞ chains demonstrated great potential as effective birefringent functional units. In this article, we successfully synthesized the Li0.8Na0.2CsV2O6·H2O compound, which exhibits a remarkable birefringence of 0.134 at 546.1 nm, as confirmed by the experiment. Notably, the introduction of alkali metals in this compound led to a significantly shorter cutoff edge at 340 nm. Through a comprehensive investigation, Li0.8Na0.2CsV2O6·H2O has the shortest UV cutoff edge among all vanadates, whose birefringences are larger than 0.1, to the best of our best knowledge. This finding underscores the application potential of this novel material as a birefringent crystal.

Closed-Loop Error-Correction Learning Accelerates Experimental Discovery of Thermoelectric Materials

The exploration of thermoelectric materials is challenging considering the large materials space, combined with added exponential degrees of freedom coming from doping and the diversity of synthetic pathways. Here, historical data is incorporated, and is updated using experimental feedback by employing error-correction learning (ECL). This is achieved by learning from prior datasets and then adapting the model to differences in synthesis and characterization that are otherwise difficult to parameterize. This strategy is thus applied to discovering thermoelectric materials, where synthesis is prioritized at temperatures <300 °C. A previously unexplored chemical family of thermoelectric materials, PbSe:SnSb, is documented, finding that the best candidate in this chemical family, 2 wt% SnSb doped PbSe, exhibits a power factor more than 2× that of PbSe. The investigations herein reveal that a closed-loop experimentation strategy reduces the required number of experiments to find an optimized material by a factor as high as 3× compared to high-throughput searches powered by state-of-the-art machine-learning (ML) models. It is also observed that this improvement is dependent on the accuracy of the ML model in a manner that exhibits diminishing returns: once a certain accuracy is reached, factors that are instead associated with experimental pathways begin to dominate trends.

Li2RbSO4Cl with a Short Ultraviolet Absorption Edge and an Acentric Structure through Assembling Heteroleptic [LiO3Cl] Tetrahedra

Currently, short-wavelength nonlinear optical materials are urgently needed. Through substituting homoleptic [LiO4] in centrosymmetric LiRbSO4 with heteroleptic [LiO3Cl] tetrahedra, an acentric sulfate chloride, Li2RbSO4Cl, was designed and synthesized by the high-temperature melting method. Li2RbSO4Cl shows a relatively short ultraviolet absorption edge (<200 nm) among newly reported sulfate chlorides. Millimeter-sized crystals were grown due to the congruent melting behavior and high thermal stability of the compound.

Target-Oriented Synthesis of Borate Derivatives Featuring Isolated [B3O3] Six-Membered Rings as Structural Features

Borates provide an excellent platform for investigating the optical nonlinearity and linearity of crystals as photoelectric functional materials. In our work, borate derivatives with isolated [B₃O₃] six-membered rings as structural features are the preferred system due to their simple functional units and excellent properties. Herein, by utilizing the target-oriented synthesis, a series of borate derivatives, A₂[B₃O₃F₄(OH)] (A= NH₄, Rb, Cs) (ABOFH), K_2.3Cs_0.7B₃O₃F₆ (KCsBOF), and Cs₃[B₃O₃(OH)₃]Cl₃ (CsBOHCl), with novel heteroanionic groups containing [BO_xF_4-x] (x = 0-3) and/or [BO₂(OH)] units were obtained. ABOFH, KCsBOF, and CsBOHCl construct different two-dimensional pesudolayers featuring [B₃O₃F₄(OH)], [B₃O₃F₆], and [B₃O₃(OH)₃] units, respectively. Also, the optical properties and the arrangement information of these anionic groups were studied. Among the total five compounds, (NH₄)₂[B₃O₃F₄(OH)] and Cs₃[B₃O₃(OH)₃]Cl₃ with enlarged birefringence and sufficient band gaps were screened out as promising birefringent crystals due to the optimally aligned configuration of birefringence-active heteroanionic units. The successful results of target-oriented synthesis indicate a more profound conclusion that the borate system now has more diversified structural chemistry, and an effective strategy was proposed to modify the arrangement and species of anionic units to optimize the performance of optical crystals.