A deep machine-learning technique based on a convolutional neural network (CNN) is introduced. It has been used for the classification of powder X-ray diffraction (XRD) patterns in terms of crystal system, extinction group and space group. About 150 000 powder XRD patterns were collected and used as input for the CNN with no handcrafted engineering involved, and thereby an appropriate CNN architecture was obtained that allowed determination of the crystal system, extinction group and space group. In sharp contrast with the traditional use of powder XRD pattern analysis, the CNN never treats powder XRD patterns as a deconvoluted and discrete peak position or as intensity data, but instead the XRD patterns are regarded as nothing but a pattern similar to a picture. The CNN interprets features that humans cannot recognize in a powder XRD pattern. As a result, accuracy levels of 81.14, 83.83 and 94.99% were achieved for the space-group, extinction-group and crystal-system classifications, respectively. The well trained CNN was then used for symmetry identification of unknown novel inorganic compounds.

Most of the novel phosphors that appear in the literature are either a variant of well-known materials or a hybrid material consisting of well-known materials. This situation has actually led to intellectual property (IP) complications in industry and several lawsuits have been the result. Therefore, the definition of a novel phosphor for use in light-emitting diodes should be clarified. A recent trend in phosphor-related IP applications has been to focus on the novel crystallographic structure, so that a slight composition variance and/or the hybrid of a well-known material would not qualify from either a scientific or an industrial point of view. In our previous studies, we employed a systematic materials discovery strategy combining heuristics optimization and a high-throughput process to secure the discovery of genuinely novel and brilliant phosphors that would be immediately ready for use in light emitting diodes. Despite such an achievement, this strategy requires further refinement to prove its versatility under any circumstance. To accomplish such demands, we improved our discovery strategy by incorporating an elitism-involved nondominated sorting genetic algorithm (NSGA-II) that would guarantee the discovery of truly novel phosphors in the present investigation. Using the improved discovery strategy, we discovered an Eu(2+)-doped AB5X8 (A = Sr or Ba, B = Si and Al, X = O and N) phosphor in an orthorhombic structure (A21am) with lattice parameters a = 9.48461(3) Å, b = 13.47194(6) Å, c = 5.77323(2) Å, α = β = γ = 90°, which cannot be found in any of the existing inorganic compound databases.

The combinatorial chemistry (combi‐chem) of inorganic functional materials has not yet led to the discovery of commercially interesting materials, in contrast to the many successful discoveries of heterogeneous catalysts leading to commercialization. Novel materials for practical use are likely hidden in the multicompositional search space that contains an infinite number of possible stoichiometries, as well as a large number of well‐known materials. To discover new, inorganic luminescent materials (phosphors) from the SrO‐CaO‐BaO‐La2O3‐Y2O3‐Si3N4‐Eu2O3 search space, heuristics optimization strategies, such as the non‐dominated‐sorting genetic algorithm (NSGA) and particle swarm optimization (PSO) are coupled with high‐throughput experimentation (HTE) in such a manner that the experimental evaluation of fitness functions for the NSGA and PSO is accomplished by the HTE. The proposed strategy also involves the parameterization of the material novelty to avoid systematically a futile convergence on well‐known, already‐established materials. Although the process starts with random compositions, we finally converge on a novel, single‐phase, yellow‐green‐emitting luminescent material, La4–xCaxSi12O3+xN18−x:Eu2+, that has strong potential for practical use in white light‐emitting diodes (WLEDs).

Candidates for high‐energy cathodes in potassium‐ion batteries (KIBs) are selected by fully screening the inorganic compound structure database. The compounds that satisfy the specific conditions for plausible KIB cathodes are further subjected to theoretical and electrochemical verification, and KVP2O7 is finally pinpointed. KVP2O7 can reversibly desert/insert ≈60% of K+ (60 mA h g−1) during either chemical or electrochemical oxidation/reduction. KVP2O7 shows an average discharge potential of ≈4.2 V versus K/K+, which corresponds to an energy density of 253 W h kg−1 at 0.25 C. This high energy density characteristic of KVP2O7 is maintained both during fast charge/discharge (C/D) and prolonged redox cycles. The C/D of KVP2O7 is also accompanied by a phase transition between a monoclinic KVP2O7 (P21/c) and a triclinic K1−xVP2O7(P1true¯). The structure interpretation of a new K1−xVP2O7 phase indicates that K+‐extraction induces a conformational change of two tetrahedral PO4 units in pyrophosphates. The P1true¯ phase of K1−xVP2O7 (x ≈0.6) remains stable during the C/D process, although it returns to the inborn P21/c phase after thermal treatment. It is believed that the data‐mining protocol designed for this study will provide a new strategy for materials discovery and that the pinpointed KVP2O7 can be utilized as a reliable KIB cathode.

Here we report a facile, prompt protocol based on deep-learning techniques to sort out intricate phase identification and quantification problems in complex multiphase inorganic compounds. We simulate plausible powder X-ray powder diffraction (XRD) patterns for 170 inorganic compounds in the Sr-Li-Al-O quaternary compositional pool, wherein promising LED phosphors have been recently discovered. Finally, 1,785,405 synthetic XRD patterns are prepared by combinatorically mixing the simulated powder XRD patterns of 170 inorganic compounds. Convolutional neural network (CNN) models are built and eventually trained using this large prepared dataset. The fully trained CNN model promptly and accurately identifies the constituent phases in complex multiphase inorganic compounds. Although the CNN is trained using the simulated XRD data, a test with real experimental XRD data returns an accuracy of nearly 100% for phase identification and 86% for three-step-phase-fraction quantification.