2024 AIChE Annual Meeting

(271b) Building a “Defective” Materials Genome: High-Throughput Computational Materials Design Including Point Defects

Author

First principles or ab initio approaches have become central to materials science as they offer powerful ways to solve the fundamental laws of physics at the atomistic level. Essential materials properties can now be assessed through these methods and they are routinely used to understand and design materials. The predictive power of ab initio techniques in assessing materials properties provides an opportunity for large-scale, accelerated, computational searches for new materials. In this high-throughput approach, we can screen thousands of materials by their computed properties even before any experiment has been performed. High-throughput computing had many successes in orienting experimental research in promising directions in various fields from batteries to catalysis. Large and free databases of computed materials properties such as the Materials Project have even emerged.

However, high-throughput screening and large data in materials science have mainly focused so far on bulk properties (e.g., band gap, elastic constants, ...) while defects can be sometimes essential in driving materials performance. In this talk, I will discuss the opportunities and challenges in bringing point defects in computational screening and large databases, setting up the stage to build a materials genome for defects.

More concretely, I will discuss the results of two recent projects bringing defects to the center of computational materials design in different fields: photovoltaics and quantum information science. In photovoltaics, point defects are central to the performance of the solar absorber as they act as recombination centers, limiting carrier lifetime and efficiency. We will show how “defect-tolerant” materials (i.e., with favorable intrinsic defects) can be identified computationally among thousands of inorganic compounds and discuss some of our findings including the recent discovery of a promising new ternary phosphide: BaCd2P2. This material was not only predicted computationally but follow-up experimental synthesis and characterization confirmed its interest as a high-performance solar absorber with long carrier lifetime.

While defects are detrimental in photovoltaics, the field of quantum information science consider them as important building blocks for quantum sensors and networks. More specifically, color centers in semiconductors can be used as single photon emitters or spin-photon interface. There are a handful of known “quantum defects” and the NV center in diamond is by far the most studied. While many of the required properties of quantum defects are known (emission, spin multiplicity and coherence, ...) and computable through first principles, all the current color centers considered for quantum technologies have been found serendipitously. I will highlight how a quantum defect by design approach can be set-up where defects with appropriate properties are searched for computationally and targeted experimentally. I will illustrate this approach discussing our work on designing quantum defects in silicon and WS2 including recent experimental confirmation of our theoretical prediction.