Artificial Intelligence (AI) Expands the Frontiers of Materials Science

by Chema Arcos

A new review article led by ICN2 scientists analyses the impact that the application of AI and machine learning is having on key areas of materials science, including optical and electron spectroscopy and microscopy, synchrotron experiments, material design and atomistic modelling.

Over the past decade, artificial intelligence (AI) has transformed numerous scientific fields. In astronomy, for example, AI is helping to efficiently detect and classify celestial objects, while in medicine, AI-driven imaging techniques have improved the accuracy of diagnosing a wide range of diseases. Materials science is no exception. Tools such as machine learning are accelerating the discovery, design, and characterisation of new materials, as well as developing novel strategies to enhance the efficiency and sustainability of processes.

A recent review article led by Dr Emigdio Chávez, under the supervision of ICREA Prof. Aitor Mugarza (head of the ICN2 Atomic Manipulation and Spectroscopy Group), has analysed the recent advances achieved through the implementation of AI in materials science, as well as its potential impact on society and the main challenges that remain to be addressed. The study, published in the journal Advanced Intelligent Systems, also involved researchers from the Advanced Electron Nanoscopy Group and the Theoretical and Computational Nanoscience Group, as well as from institutions including IFAE, Universidad de La Frontera (Chile) and Ohio University (USA).

The scientists evaluated the use of AI and machine learning tools in several areas related to materials science, including Raman and infrared spectroscopy, synchrotron experiments, scanning probe microscopy, electron microscopy, materials design, metamaterials, atomistic modelling and drug design. The future of AI in materials science promises rapid advances in the discovery, design and characterisation of new materials, with applications in a wide range of fields. However, in a context where energy availability is a global concern, the energy implications of AI still pose many challenges. In this context, the adoption of sustainable practices and the development of more efficient algorithms are crucial.

Another challenge for AI in fields such as spectroscopy and electron and scanning probe microscopy is to train models with robust data sets from different conditions. Collaboration between research centres is an excellent strategy for this. In this context, initiatives such as the In-situ Correlative Facility for Advanced Energy Materials (InCAEM) are leading this effort and can be very useful in advancing materials science.

About InCAEM

The InCAEM project, led by ICREA Prof. Jordi Arbiol and coordinated by the ALBA Synchrotron, is a collaborative initiative within the Spanish Advanced Materials Programme. It is developing a unique infrastructure to enable the correlative use of a variety of techniques, such as (scanning) transmission electron microscopy (TEM/STEM), scanning probe microscopy (SPM), atomic force microscopy (AFM), scanning tunneling microscopy (STM) and synchrotron radiation techniques to address key scientific challenges. These objectives are aligned with the European Green Deal’s goal of fostering a sustainable economy.

In addition, InCAEM will develop a robust computational infrastructure for automated AI-based data analysis, essential for managing the large datasets generated by in situ experiments. This integration will streamline experimental workflows, enhance data interpretation, and accelerate scientific discovery, advancing our understanding of materials for energy and environmental sustainability.

Reference article:

Chávez-Angel, E; Eriksen, MB; Castro-Alvarez, A; Garcia, JH; Botifoll, M; Avalos-Ovando, O; Arbiol, J; Mugarza, A. Applied Artificial Intelligence in Materials Science and Material Design. Advanced Intelligent Systems (2025). 2400986. https://doi.org/10.1002/aisy.202400986.