From intuition to optimization: a review of inverse design applied to optical nanotweezers

Optical nanotweezers, which integrate nanophotonic structures such as plasmonic antennas, dielectric resonators, and metasurfaces, have emerged as powerful tools for manipulating nanoparticles, biomolecules, and viruses with high precision. However, traditional designs rely heavily on heuristic approaches and simple geometries, limiting performance. Inverse design offers a paradigm shift by using computational optimization methods to discover non-intuitive, high-performance structures for optical trapping. This review provides an overview of key inverse design methodologies, including adjoint-based optimization, topology optimization, and machine learning approaches, and highlights their broader impact across nanophotonics. We then focus on their application to optical nanotweezers, where recent advances demonstrate enhanced trapping stiffness, reduced optical heating, and tailored trapping landscapes. Representative examples include algorithmic design of plasmonic nanotweezers, topology-optimized plasmonic nanoapertures, inverse-designed dielectric nanocavities, and force-centric designs based on Maxwell stress tensor engineering. Collectively, these studies illustrate how inverse design enables stable, efficient, and versatile optical trapping at the nanoscale. Despite these advances, challenges remain in fabrication, computational cost, and the interpretability of optimized structures. We conclude with a perspective on future directions, emphasizing opportunities in reconfigurable tweezers, lab-on-a-chip integration, and quantum nanophotonics. By bridging photonic optimization with nanoscale manipulation, inverse-designed optical nanotweezers offer promising opportunities for future applications in biophysics, materials science, and quantum technologies.