Work by NOMAD CoE researchers published in npj Computational Materials
Scientists from the NOMAD Laboratory at the Fritz Haber Institute of the Max Planck Society have recently proposed a workflow that can dramatically accelerate the search for novel materials with improved properties. They demonstrated the power of the approach by identifying more than 50 strongly thermally insulating materials. These can help alleviate the ongoing energy crisis by allowing for more efficient thermoelectric elements, i.e., devices that can convert otherwise wasted heat into useful electrical voltage.
Discovering new and reliable thermoelectric materials is paramount for harnessing the more than 40% of energy given off as waste heat globally and for mitigating the growing challenges of climate change. One way to increase the thermoelectric efficiency of a material is to reduce its thermal conductivity, κ, thereby maintaining the temperature gradient needed to generate electricity. However, the cost associated with studying these properties limited the computational and experimental investigations of κ to only a minute subset of all possible materials. A team at the NOMAD Laboratory has recently sought to reduce these costs by creating an AI-guided workflow that hierarchically screens out materials to efficiently find new and better thermal insulators.
The work, recently published in npj Computational Materials, proposes a new way of using Artificial Intelligence (AI) to guide the high-throughput search for new materials. Instead of using physical/chemical intuition to screen out materials based on general, known or suspected trends, the new procedure uses advanced AI methods to learn the conditions that lead to the desired outcome. This work has the potential to quantify and increase the efficiency of the search for new energy materials.
Figure 1. The Sobol indexes Si and SiT as well as the kernel SHAP values for each feature in the model. These values are all feature importance metrics where 0 means the feature has no effect, and a larger value means that feature is more important.
The first step in designing these workflows is to use advanced statistical and AI methods to approximate the target property of interest, in this case κ. To this end, the Sure-Independence Screening and Sparsifying Operator (SISSO) approach is used. SISSO is a machine learning method that reveals the fundamental dependencies between different material properties from a set of billions of possible expressions. Compared to other “black-box” AI models, this approach is similarly accurate, but additionally yields the analytical relationships between different material properties. This allows us to apply modern feature importance metrics to find out which material properties are most important. In the case of κ, these are the molar volume, Vm; the high-temperature limit Debye Temperature, θD,∞; and the anharmonicity metricfactor, σA, as illustrated in Figure 1.
Figure 2. a) Schematic of the high-throughput workflow used to screen for new thermal insulators. b) A scatter plot showing the predicted thermal conductivity for 227 thermodynamically stable electrical insulators from both a SISSO and kernel-ridge regression (KRR) model. The color corresponds to which of the tests outlined in part (a) failed.
Furthermore, the described statistical analysis allows rules-of-thumb to be identified for the individual features that enable the a priori estimation of the potential of material as a thermal insulator. Working with the three most important features therefore enabled the creation of AI-guided computational workflows for discovering new thermal insulators, as shown in Figure 2. These workflows use state-of-the-art electronic structure programs to calculate each of the selected features. During each step, materials were screened out that are unlikely to be good insulators based on their values of Vm, θD,∞, and σA.
This makes it possible to reduce the number of calculations needed to find thermally insulating materials by over two orders of magnitude. In this work, this is demonstrated by identifying 96 thermal insulators (κ < 10 Wm-1K-1) in an initial set of 732 materials. The reliability of this approach was further verified by calculating κ for four of these predictions with the highest possible accuracy.
Besides facilitating the active search for new thermoelectric materials, the formalisms proposed by the NOMAD team can be also applied to solve other urgent material science problems.
Read the full publication here:
T. Purcell, M. Scheffler, L. M. Ghiringhelli, and C. Carbogno
Accelerating Materials-Space Exploration by Mapping Materials Properties via Artificial Intelligence: The Case of the Lattice Thermal Conductivity, npj Computational Materials 9, 112 (2023).
https://www.nature.com/articles/s41524-023-01063-y