Ask Onix
AI uncovers materials that defy conventional physics
Researchers in the Netherlands have leveraged artificial intelligence to engineer metamaterials that thicken when stretched, a property once considered nearly impossible. These innovations could significantly enhance the durability of hip replacements and improve fracture healing.
The challenge of hip replacements
Hip implants, among the most common orthopedic procedures globally, face a critical limitation: wear and tear. Patients with artificial hips take approximately two million steps annually, subjecting implants to relentless forces. Over a decade or more, this stress often necessitates replacement surgeries.
Amir Zadpoor, a professor of orthopedics at Leiden University Medical Center, sought a solution by combining two opposing materials. One would thicken under compression, while the other would expand when stretched, cushioning the femur and securing the implant against the bone.
"That would reinforce the connection between the bone and the implant," Zadpoor explained.
The hunt for the 'holy grail' material
Traditional materials thin when stretched, like an elastic band. Zadpoor's team needed the opposite-an auxetic material that thickens under tension. However, known auxetic materials, such as those used in crash helmets, are typically too soft for load-bearing implants.
"We were trying to find this holy grail of auxeticity and also high stiffness," Zadpoor said. "That becomes a formidable hunt."
AI accelerates material discovery
To overcome this challenge, the team turned to artificial intelligence. By training an AI system to predict material behaviors, they input their desired properties. The machine generated designs for metamaterials-engineered structures with unconventional properties based on their microscopic architecture.
Sid Kumar, an associate professor of materials science at TU Delft, noted that while developing and training an AI model can take up to a year, once operational, it can produce viable designs in minutes or seconds.
"With machine learning, you can make the process orders of magnitude faster and explore thousands to millions of more structures," Zadpoor said.
Mimicking bone for better healing
Kumar and his colleagues applied AI to design metamaterials for soft bone implants, targeting complex fractures common in elderly patients. Current titanium or steel implants often fail to integrate properly with bone, leading to weak healing.
The team aimed to create a softer, lattice-like material that mimics the early stages of fracture healing. This metamaterial, resembling a thin, perforated bandage, would allow living cells to colonize and integrate with the bone.
"The early stage of fracture healing is decisive for success," said Xiao-Hua Qin, an assistant professor of biomaterial engineering at ETH Zurich.
Tailoring implants to human anatomy
Metal implants are more resilient than bone, which can be problematic as they absorb external forces, leading to bone deterioration. Kumar's team sought a metamaterial with properties matching trabecular bone-the porous, honeycomb-like structure at the ends of long bones that provides strength and shock absorption.
Using a machine learning model, they generated spinodoid designs that closely replicate human bone's curvature, porosity, and mechanical behavior. These designs could be 3D-printed and tailored to specific regions of an implant, offering varying stiffness, porosity, and tissue integration.
"This is important because you may want one region of the implant to be stiffer, another region to be more porous," said Mohammad Mirzaali, an associate professor of biomedical engineering at TU Delft.
Future prospects: deployable and personalized implants
Zadpoor's team advanced their research by combining three AI models to identify auxetic metamaterials suitable for bone implants. These materials could be durable, stress-resistant, and customizable to fit a patient's anatomy, potentially extending implant longevity.
Looking ahead, AI may enable implants that expand inside the body, reducing the need for invasive surgery. Kumar's team recently unveiled an AI-designed metamaterial that expands in all directions and can change shape in response to electrical currents.
"I think deployable implants are very exciting," Mirzaali said.
