New algorithm can discover materials with unusual characteristics—including invisibility

Metamaterials are artificially engineered materials. Scientists create them by combining multiple elements from composite materials such as a metal and an electrical insulator. The result is an entirely new material with properties not found in nature. Engineers can then use these materials to create new devices or improve existing ones.

Let's say you want to build a real-life invisibility cloak. To achieve invisibility, a metamaterial needs to possess certain optical properties. Specifically, scientists would have to design the material so that they could control how light moves around an object without being reflected or absorbed. This design is possible, but it would take just the right material with just the right structure.

There are hundreds of thousands of potential material structures with optical responses that fall somewhere along the optical spectrum. Sifting through them to find a new material design has traditionally taken hours or even days.

Now, Northeastern professor Yongmin Liu has developed a new method for quickly discovering materials that have desirable qualities. In a paper published recently in ACS Nano, Liu and his co-authors describe a machine learning algorithm they developed and trained to identify new metamaterial structures. The new method is much faster and more accurate than previous approaches, paving the way for engineers to design next-generation materials.

The algorithm Liu and his team built was trained with a data set of 30,000 different samples, each representing a specific relationship between a metamaterial structure and corresponding optical property. Once the algorithm learned those relationships, it was able to predict new ones.