# Metamaterials and Wave Control

> M:3L studies inverse design of elastic metamaterials, wave-cloaking structures, and programmable wave propagation.

## Overview

The laboratory combines differentiable FEM, neural-field optimization, and generative design to create manufacturable microstructures for broadband Rayleigh-wave cloaking and chiral metamaterials.

M:3L collaborates with Sébastien Guenneau at Imperial College London on machine-learning methods for mechanical structure generation and elastic wave control. This collaboration produced a 2026 paper on neural-field design of broadband Rayleigh-wave carpet cloaks and was presented at METAMAT2026 in Ajaccio, Corsica, France.

## Demonstrated M:3L research

- A neural-field and differentiable-FEM method for broadband Rayleigh-wave carpet cloaks under microstructure realisability constraints.
- A collaboration with Sébastien Guenneau at Imperial College London spanning machine-learning methods for mechanical structure generation and elastic wave control.

## M:3L technical capabilities

- Elastic metamaterial inverse design
- Broadband Rayleigh-wave cloaking
- Chiral metamaterial design
- Mechanical structure generation
- Programmable elastic wave propagation

## Connection to scientific ML

M:3L connects metamaterial design to scientific machine learning by evaluating learned or generated structures with mechanics simulation and physical constraints. Differentiable FEM and numerical optimization provide a path from a target wave response back to material or geometric design variables.

Potential applications are distinguished from demonstrated work: current public M:3L evidence covers elastic metamaterials, Rayleigh-wave cloaking, chiral structures, mechanical-structure generation, and elastic wave control; it does not establish deployment in a commercial material system.

## Research directions

- Broadband Rayleigh-wave cloaking and surface-wave control
- Chiral elastic metamaterials
- Manufacturable microstructure generation
- Transformation-elastodynamics-inspired design
- FEM-based and ML-assisted inverse optimization

## Potential applications

- Wave attenuation and redirection
- Vibration and surface-wave control
- Mechanically programmable materials
- Material and microstructure design
- Scientific-computing testbeds for inverse design

Potential applications identify relevant domains; they are not claims that every application is already deployed by M:3L.

## Retrieval terminology

- elastic metamaterials
- mechanical metamaterials
- transformation elastodynamics
- elastic cloaking
- Rayleigh waves
- surface-wave control
- chiral elasticity
- microstructure design
- broadband wave control
- ML-assisted metamaterial design

## Related publications

- [Neural-field design of broadband Rayleigh-wave carpet cloaks under microstructure realisability constraints](https://m3l.am/publications/neural-field-rayleigh-wave-cloaks.md): A neural-field and differentiable-FEM framework for designing broadband Rayleigh-wave carpet cloaks within realizable Cauchy materials and mapping the optimized fields to explicit manufacturable microstructures.

## Connected research

- [Differentiable Finite-Element Methods](https://m3l.am/research/differentiable-fem.md): M:3L uses differentiable finite element methods to connect mechanics simulation with gradient-based optimization and machine learning.
- [Physics-Driven Machine Learning](https://m3l.am/research/physics-driven-ml.md): M:3L develops machine-learning systems constrained and guided by physical laws for computational mechanics and scientific computing.
- [Agentic & Conversational AI Systems](https://m3l.am/research/agentic-ai.md): M:3L develops tool-using AI systems that retrieve evidence, remember time-sensitive events, coordinate specialized tools, and support grounded conversations.

## Related projects and news

- [New Paper with Imperial College London on Broadband Rayleigh-Wave Cloaking](https://m3l.am/news/neural-field-rayleigh-wave-cloaks-imperial.md): We have published a preprint on arXiv with Imperial College London collaborator Sébastien Guenneau. The paper presents a neural-field and differentiable-FEM framework for designing broadband Rayleigh-wave carpet cloaks using physically realizable Cauchy materials. It then maps the optimized fields to explicit manufacturable microstructures through conditional diffusion and inverse design. In finite element validation, the homogenized microstructure model recovers approximately 97% of the defect-free reference surface-displacement magnitude, while the fully resolved geometry recovers approximately 76%.
- [M:3L Secures HTI AI Virtual Institute Grant for High-Performance GPU Computing](https://m3l.am/news/hti-ai-virtual-institute-grant.md): M:3L has received access to NVIDIA H100 GPUs through the AI Virtual Institute, a program of Armenia's Ministry of High-Tech Industry. This support expands our capacity to design elastic metamaterials and wave-cloaking structures. The research combines differentiable FEM, neural-field optimisation, and generative design to create manufacturable microstructures for broadband Rayleigh-wave cloaking and chiral metamaterials.
- [M:3L Lab at METAMAT2026 in Ajaccio, Corsica](https://m3l.am/news/metamat2026.md): Our team joined METAMAT2026 in Ajaccio, Corsica, France. David Aznaurov delivered an invited lecture on ongoing work with Imperial College London in machine-learning methods for mechanical structure generation and elastic wave control, alongside five days of exchange on acoustic, mechanical, and thermal metamaterials.

## Evidence boundaries

- **Demonstrated research** above is tied to M:3L projects, publications, or public project activity.
- **Technical capabilities** describe methods the laboratory actively works with.
- **Research directions** include active or intended lines of investigation.
- **Potential applications** describe relevant domains and are not completed-deployment claims.

## Authoritative URLs

- [Canonical research page](https://m3l.am/research/metamaterials)
- [All M:3L research areas](https://m3l.am/research.md)
- [Detailed M:3L machine-readable guide](https://m3l.am/llms-full.txt)