Nonlinear Coefficient Optimization in Emerging Silicon Nitride Waveguides for High Conversion Efficiency of Four Wave Mixing Process Dharma P. Permana (a), Yvan Klaver (b), David A. Marpaung (a,b), Alexander A. Iskandar (a*)
(a) Physics of Magnetism and Photonics Research Division, Faculty of Mathematics and Natural Sciences, Institut Teknologi Bandung, Indonesia
*a.a.iskandar[at]itb.ac.id
(b) Nonlinear Nanophotonics, MESA+ Institute of Nanotechnology, University of Twente, Enschede, the Netherlands
Abstract
Efficient single photon source is one of the important components needed to operate chip sized photonic quantum circuits and quantum processor for practical applications of quantum computer. A promising candidate for efficient single photon source integration to the existing quantum processor and quantum circuits can be achieved by utilizing four wave mixing (FWM) process in nonlinear optical media. Strong FWM can be achieved by enhancement of nonlinear coefficient which depends on the intrinsic properties of materials and optical properties of the waveguide. Taking into consideration the value of nonlinear index, refractive index, broad transparency window, and negligible effect of two photon absorption, silicon nitride (SiN) is chosen as the material platform, while double stripe waveguide is chosen as the platform geometry due to its low propagation loss, tight optical confinement, and can be fabricated with COMS-compatible technologies with high yield. However, the major problem of employing emerging waveguide designs (such as the double stripe waveguide) is the inaccuracy of the commonly used nonlinear coefficient calculation model which known as scalar model due to its weakly guiding approximation. Another approach to calculate the nonlinear coefficient in the strong guidance regime has been developed in the form of vectorial model, and perturbative model. In this work, we used scalar, vectorial, and perturbative models to numerically calculate the nonlinear coefficients of several structure with experimental demonstrations of FWM process in order the determine the accuracy of each model. The results of our calculations shows that the each of the models accuracy changing depends on the fabrication methods of the SiN waveguides- thus, no models can consistently gives accurate results of nonlinear coefficients. Despite, none of the models can consistently give accurate value of the nonlinear coefficients, these models can predict the optimum thickness value which posses the largest nonlinear coefficient. By doing thickness variation of single stripe SiN waveguides, all the models shows that the nonlinear coefficient peaked around 400 nm and 500 nm thickness simultaneously. Further, we^ve done optimization of double stripe waveguides by changing the stripes thickness and varying the gap between each stripe to find the optimum gap and thickness value with largest nonlinear coefficient. From this optimization we found that the optimum stripe thicknesses needed to achieve maximum value of the nonlinear coefficient also increases with the increment of gap size, while the peak value of the nonlinear coefficient simultaneously decreasing for all models with the increment of gap size.
Keywords: nonlinear coefficient- inhomogeneous cross section- high index contrast- four wave mixing- double stripe waveguide- silicon nitride
