AI Technology Predicts Precursor Materials for Inorganic Compounds with 80% Accuracy

As the world grapples with the complexities of materials discovery, a breakthrough in artificial intelligence (AI) technology has emerged from a joint research team in Korea, revolutionizing the synthesis process of target materials by automatically identifying necessary precursor materials solely based on their chemical formulas.

This innovation, characterized as “Retrieval-Retro: Retrieval-based Inorganic Retrosynthesis with Expert Knowledge,” signifies a crucial leap forward in efficiently utilizing AI to predict the required intermediate materials for synthesis, a task traditionally marred by costly and repetitive experiments. The implications are profound, with potential applications spanning across various industries such as batteries and semiconductors, where the discovery of new materials has become a pressing challenge.

By leveraging a deep neural network trained on an extensive dataset of published research papers detailing material synthesis processes, this AI technology overcomes the hurdles posed by complex structures of inorganic materials, offering a universal prediction capability regardless of the target material’s application. With plans to establish a web-based public service for AI-based materials discovery by 2026 and aiming for 90% prediction accuracy, this development is poised to significantly enhance the efficiency of new material development worldwide, underscoring the transformative potential of AI in chemical technologies and materials science.

Key themes that define this issue include the integration of AI in materials discovery, the challenges and opportunities in predicting precursor materials for inorganic compounds, and the broader implications for industries reliant on advanced materials.

Introduction: Revolutionizing Material Synthesis with AI

The quest for efficient material synthesis has become a cornerstone in various industries, including batteries and semiconductors. Traditionally, identifying the right precursor materials for synthesis has been a costly and time-consuming process, relying on repetitive experiments. However, with the advent of Artificial Intelligence (AI), researchers have been exploring ways to utilize AI to streamline this process. A recent breakthrough by a joint research team from the Korea Research Institute of Chemical Technology (KRICT) and the Korea Advanced Institute of Science and Technology (KAIST) has developed an innovative AI-based retrosynthesis methodology that predicts the necessary precursor materials for synthesizing specific target materials, solely based on their chemical formulas.

Unlike organic materials, such as drug compounds, inorganic materials like metals pose significant challenges due to their complex structures and diverse elemental compositions. This complexity makes it difficult to determine synthesis pathways, a crucial step in material discovery. Existing AI technologies have primarily focused on organic materials, leaving a gap in research for inorganic compounds.

The KRICT-KAIST research team has overcome this challenge by developing an AI technology that learns the inverse process of predicting precursor materials using only the target material’s chemical formula. This approach analyzes the types and ratios of elements present in the target material, calculating thermodynamic formation energy differences to identify precursors that facilitate easier synthesis reactions.

The team employed a deep neural network specialized in chemical data to enhance prediction accuracy. The AI model was trained on approximately 20,000 published research papers detailing material synthesis processes and precursor materials. Testing on around 2,800 synthesis experiments not included in the training dataset showed an impressive success rate of over 80% in predicting necessary precursor materials within just 0.01 seconds, utilizing GPU acceleration.

Looking ahead, the research team aims to expand the training dataset to achieve a 90% prediction accuracy by 2026 and establish a web-based public service for AI-based materials discovery. Future research will focus on fully automated materials discovery that predicts both precursor materials and synthesis pathways based solely on the target material’s chemical formula.

The development of this AI-based retrosynthesis methodology marks a significant leap forward in material synthesis, offering the potential to revolutionize industries reliant on efficient material discovery. Automating the prediction of precursor materials can significantly reduce the time and cost associated with traditional methods, paving the way for more rapid advancement in fields such as battery and semiconductor development.

As KRICT President Young-Kuk Lee emphasized, “This research is expected to enhance the efficiency of new material development across various industries.” The global implications are profound, suggesting a future where material discovery is no longer a bottleneck but a catalyst for innovation. With ongoing developments and the planned expansion of this technology, the world may soon see a significant shift in how materials are discovered and synthesized, ushering in a new era of efficiency and discovery.