Realization of fluoride sensor using two dimensional photonic structures and machine learning model

Abstract

Present research computes the concentration of fluoride in water with the help of all type of two dimensional photonic structures and machine learning model. The current work has a certain primacy relating to the several types of losses in the photonic structures. Further this paper considers silicon based square, triangular and honeycomb based two dimensional photonic crystal structures. Aside this, mathematical models are developed using the linear regression algorithm to prophesy the concentration of fluoride in the water by knowing the power emerging from the proposed structure. Moreover the above said machine learning model discloses the relationship between concentration of fluoride and output power as $$ {\text{concentration}} = 17979 \times {\text{output power}} - 11548,{\text{ concentration}} = 7902.8 \times {\text{output power}} - 4438.2{\text{ and concentration}} = 446.47 \times {\text{output power}} - 274 $$ with the error of 5.405%, 6.6762%, and 15.806% corresponding to the square, triangular and honeycomb structure respectively. Finally, the present model indicates that both square and triangular photonic structures are paramount candidates to estimate the fluoride concentration rather honeycomb as it suffers glut error.