Physics-informed machine learning (PIML) for wave control integrates physical principles with data-driven models to enhance the manipulation of acoustic, elastic, and water waves. By embedding governing equations and boundary conditions into machine learning algorithms, we use PIML for more efficient, accurate, and adaptive wave control.

Precise acoustic beam shaping is vital for applications like ultrasound imaging and manipulation, yet controlling acoustic pressure profiles during propagation remains challenging. Acoustic Diffraction-Resistant Adaptive Profile Technology (ADAPT) addresses this by creating propagation-invariant beams with customizable profiles. Utilizing wave number modulation and beam multiplexing, it enables flexible acoustic beams using linear array ultrasonic transducers. ADAPT can maintain beam profiles in lossy materials by compensating for attenuation, particularly benefiting shear wave elasticity imaging for tissue mechanics assessment. ADAPT overcomes acoustic beam shaping limitations and finds applications in medicine, biology, and material science. Science Advances 9.44 (2023): eadi6129.

Acoustic methods enable label-free, non-contact, and biocompatible particle manipulation. Traditional approaches are limited by axisymmetric pressure patterns. Emerging acoustic holography offers greater freedom in pressure distribution but struggles with micro/nanoscale particles due to decreasing acoustic forces. We introduce acoustofluidic holography (AFH), utilizing both arbitrary acoustic fields and controlled fluid motion for effective micro/nano particle manipulation. AFH works across various materials, from cells to metals, with sizes spanning hundreds of micrometers to tens of nanometers. ACS nano 14.11 (2020): 14635-14645.

We introduce an acoustofluidic centrifugation method that combines acoustic wave actuation and droplet spinning for rapid (<1 min) nanoparticle enrichment and size-based separation. This technique efficiently processes biological samples, such as DNA segments and exosome subpopulations, overcoming limitations in nanoscale (<100 nm) bioparticle manipulation. The method has applications in biology, chemistry, engineering, material science, and medicine, addressing physical phenomena across different size scales from cells to black holes. Science advances 7.1 (2021): eabc0467.