Acoustic emission-based monitoring for size-controlled robotic powder grinding.

Abstract

Controlling particle size is essential in solid-state synthesis, yet powder grinding remains a critical and largely unresolved challenge in self-driving laboratories. Here, we present a robotic mortar-based grinding system integrated with structure-borne acoustic emission (AE) sensing for autonomous powder processing. By measuring high-frequency elastic waves transmitted through the mortar, the proposed approach enables robust AE acquisition in standard laboratory environments without specialized acoustic facilities and provides a practical basis for particle-size-dependent feedback and autonomous stopping. We model AE-derived total spectral power as a monotonic function of particle size using monotone Bernstein polynomial regression, enabling size-dependent threshold design for grinding termination. Using this threshold-based stopping strategy, autonomous grinding experiments on NaCl and citric acid demonstrate reliable size-controlled processing. In contrast, monosodium glutamate exhibits degraded threshold reliability at intermediate target sizes, where AE fluctuations approach the noise floor. These results reveal material-dependent regimes in which AE-based feedback is effective only when sufficient signal-to-noise ratio is maintained, thereby clarifying the intrinsic limitations of AE-driven autonomous powder processing in solid-state self-driving laboratories.