Efficient modeling and optimization approach for grating-based flow cytometers

This work presents a computationally efficient transmission matrix model and optimization scheme for the design of silicon nitride grating couplers in integrated scattering-based flow cytometry systems. The proposed model accurately simulates the optical power flow through the grating coupler and microfluidic channel system, enabling precise evaluation of the transient associated with a polystyrene bead's passage through the channel. The transmission matrix model yields a four to five orders-of-magnitude improvement in computational efficiency compared to a finite difference time domain solver, making it suitable for optimization loops consisting of many iterations and objective function evaluations. The model's computational speed is leveraged to quickly simulate the effect of variations in bead dimensions or position in the channel. The model is incorporated into a Bayesian optimization scheme that maximizes the peak to baseline height of the transient by tuning the parameters of both uniform and linearly apodized grating configurations. Results demonstrate a linearly apodized grating configuration, optimized for a dynamic system, yields a 2.49 dB improvement in peak to baseline transmission on the best uniform grating configuration, optimized for a static system.