ORNL demonstrates autonomous synthesis of customized materials with AI

How to apply this

Using an AI-based autonomous laboratory approach, one can try atom-scale assembly and synthesis of customized materials on a small scale.

1. Prepare synthesis equipment with autonomous control and real-time measurement channels, then narrowly define the target material and the permissible process range.
2. Collect initial experimental data and literature or existing experimental records, and determine the input variables and output metrics for AI to explore.
3. Perform small-scale repeated synthesis under conditions proposed by AI, and immediately record the signal and output of each experiment.
4. Use independent analysis to confirm whether the product has the target structure and composition, and repeat at least once under the same conditions to check reproducibility.
5. Adjust the exploration range based on the results and update the next experimental conditions in a closed-loop manner.

Data needed

The material composition to be attempted, precursors/reagents, the initial range of process conditions, real-time measurement signals such as growth, structure, and composition data, and safely controllable synthesis equipment are needed.

Checking the result

Check whether repeated experiments under the same conditions produce similar results, confirm the structure, composition, and morphology of the products with independent analyses such as XRD, microscopy, and composition analysis, and compare by reproducing once more without automatic control.

What can go wrong

The most common failure is that, even when AI appears to fit well, performance drops sharply for compositions or equipment conditions outside the learned range, or measurement error and process variation are mistaken for materials effects.

이 가이드는 원문의 방법을 바탕으로 AI가 제안한 적용 방향입니다. 원문에 적힌 내용이 아니며 같은 결과가 재현된다는 보장도 아닙니다.