Faster calculations of optical trapping using neural networks trained by T-matrix data: an application to micro and nanoplastics

Abstract

We employ neural networks to improve and speed up optical force calculations for dielectric particles. The network is first trained on a limited set of data obtained through accurate light scattering calculations, based on the Transition matrix method, and then used to explore a wider range of particle dimensions, refractive indices, and excitation wavelengths. This computational approach is very general and flexible. Here, we focus on its application in the context of micro and nanoplastics, a topic of growing interest in the last decade due to their widespread presence in the environment and potential impact on human health and the ecosystem.

arXiv comment25 pages, 5 figures

References in corpus (9)

• Deep Learning Microscopy
• Momentum of an electromagnetic wave in dielectric media
• Quantitative Digital Microscopy with Deep Learning
• Theory of Optical Tweezers
• Electromagnetic forces for an arbitrary optical trapping of a spherical dielectric
• Machine learning reveals complex behaviours in optically trapped particles
• Deep Learning for Optical Tweezers
• Optical tweezers in a dusty universe
• Investigation of dust grains by optical tweezers for space applications