Discovery and Optimization of Materials Using Evolutionary Approaches Tu C. Le † and David A. Winkler* ,†,‡,§,∥ † CSIRO Manufacturing, Bag 10, Clayton South MDC, Victoria 3169, Australia ‡ Monash Institute of Pharmaceutical Sciences, 381 Royal Parade, Parkville 3052, Australia § Latrobe Institute for Molecular Science, La Trobe University, Bundoora 3046, Australia ∥ School of Chemical and Physical Sciences, Flinders University, Bedford Park 5042, Australia ABSTRACT: Materials science is undergoing a revolution, generating valuable new materials such as flexible solar panels, biomaterials and printable tissues, new catalysts, polymers, and porous materials with unprecedented properties. However, the number of potentially accessible materials is immense. Artificial evolutionary methods such as genetic algorithms, which explore large, complex search spaces very efficiently, can be applied to the identification and optimization of novel materials more rapidly than by physical experiments alone. Machine learning models can augment experimental measurements of materials fitness to accelerate identification of useful and novel materials in vast materials composition or property spaces. This review discusses the problems of large materials spaces, the types of evolutionary algorithms employed to identify or optimize materials, and how materials can be represented mathematically as genomes, describes fitness landscapes and mutation operators commonly employed in materials evolution, and provides a comprehensive summary of published research on the use of evolutionary methods to generate new catalysts, phosphors, and a range of other materials. The review identifies the potential for evolutionary methods to revolutionize a wide range of manufacturing, medical, and materials based industries. CONTENTS 1. Introduction A 2. Evolutionary Algorithms B 2.1. Genetic Algorithms C 2.1.1. Overview of the Process C 2.1.2. Materials Genomes C 2.1.3. Fitness Functions C 2.1.4. Fitness Landscapes D 2.1.5. Genetic Operators D 2.2. Structure−Property Model Methods as in Silico Fitness Functions D 2.3. Genetic Algorithm Software and Optimiza- tion Parameter Choices F 3. Examples of Materials Discovery and Optimiza- tion Using Evolutionary Algorithms G 3.1. Catalytic Materials G 3.1.1. Catalyst Evolution Using Experimental Fitness Functions G 3.1.2. Catalyst Evolution Using Computational Models as Fitness Functions L 3.2. Phosphors O 3.3. Other Materials Q 4. Optimization Using Structure−Property Models of Materials Evolutionary Landscapes T 5. Conclusions and Perspective W Author Information X Corresponding Author X Notes X Biographies X References X 1. INTRODUCTION The vastness of materials structure space is still not widely recognized in the materials science community. Estimating how many new materials could be made from the elements in the periodic table using the laws of chemical valence and reactivity is very difficult. However, most studies have agreed on a number of close to 10 100 , considerably larger than the estimated number of particles of matter in the Universe. For all intents and purposes, the number of possible materials we could synthesize is infinite. This realization immediately raises a threat and an opportunity. The threat revolves around the impossibility of exhaustively exploring such vast space using even the most optimistic projections of the capabilities of robotics and automation. Although these high throughput materials synthesis and characterization capabilities are being developed at a rapid pace, 1,2 and they will undoubtedly improve the efficiency of discovery of new and useful materials, they do not provide an answer to the materials space problem. The Received: November 26, 2015 Review pubs.acs.org/CR Published XXXX by the American Chemical Society A DOI: 10.1021/acs.chemrev.5b00691 Chem. Rev. XXXX, XXX, XXX−XXX