New AI tool discovers realistic 'metamaterials' with unusual properties

The properties of normal materials, such as stiffness and flexibility, are determined by the molecular composition of the material. But the properties of metamaterials are determined by the geometry of the structure they are composed of. Researchers digitally design these structures and then have them 3D printed. The resulting metamaterials can exhibit unnatural and extreme properties. For example, researchers have designed metamaterials that, despite being solid, behave like a liquid.

"Traditionally, designers use the materials at their disposal to design a new device or machine. The problem with this is that the available material properties are limited. Some properties we would like to have just don't exist in nature. Our approach is: tell us what you want as properties and we will develop a suitable material with those properties. What you then get is not really a material, but something halfway between a structure and a material, a metamaterial." Speaking is professor Amir Zadpoor from the Department of Biomechanical Engineering.

Such a materials discovery process requires solving a so-called inverse problem. The problem of finding the geometry that yields the desired properties. These inverse problems are notoriously difficult, which is where AI comes in. Researchers at TU Delft developed deep learning models that solve these inverse problems.

"Even when inverse problems were solved in the past, they were limited. This was due to the simplifying assumption that small-scale geometry can be made from an infinite number of building blocks. The problem with that assumption is that metamaterials are usually made by 3D printing. And real 3D printers have limited resolution. This limits the number of building blocks that can fit into a device, says first author Dr Helda Pahlavani.

TU Delft researchers' AI models break new ground by bypassing these simplifying assumptions. "So we can now simply ask: how many building blocks can you incorporate into the device with your manufacturing technique? The model then finds the geometry that gives the desired properties for the number of building blocks you can actually make."

A major practical problem under-researched in previous research was the durability of metamaterials. Most existing designs break down once they have been used a few times. "Until now, it has only been about the properties that can be achieved. Our study takes durability into account and selects the most durable designs from a large pool of design candidates. This makes our designs truly practical and not just theoretical adventures," says Zadpoor.

possibilities of metamaterials seem endless. But the full potential is far from being realised, says associate professor Mohammad J. Mirzaali. This is because finding the optimal design of metamaterials is currently still largely based on intuition. This involves trial and error and is therefore labour-intensive. Using a reverse design process, where the desired properties are the starting point of the design, is still rare within the metamaterials field. "But we think the step we have taken is revolutionary for metamaterials. It can lead to all kinds of new applications." The researchers see potential applications in orthopaedic implants, surgical instruments, soft robots, adaptive mirrors and exo-skeletons.

Source and image: TU Delft

Read also: Government sets out its vision on generative AI