Machine learning assisted hepta band THz metamaterial absorber for biomedical applications

Prince Jain; Himanshu Chhabra; Urvashi Chauhan; Krishna Prakash; Akash Gupta; Mohamed S Soliman; Md Shabiul Islam; Mohammad Tariqul Islam

doi:10.1038/s41598-023-29024-x

Machine learning assisted hepta band THz metamaterial absorber for biomedical applications

Sci Rep. 2023 Jan 31;13(1):1792. doi: 10.1038/s41598-023-29024-x.

Authors

Prince Jain¹, Himanshu Chhabra², Urvashi Chauhan³, Krishna Prakash⁴, Akash Gupta⁵, Mohamed S Soliman^{6

7}, Md Shabiul Islam⁸, Mohammad Tariqul Islam⁹

Affiliations

¹ Department of Mechatronics Engineering, Parul University, Vadodara, Gujarat, India. princeece48@gmail.com.
² Department of Mechatronics Engineering, Parul University, Vadodara, Gujarat, India.
³ Department of Electronics and Communication Engineering, Parul University, Vadodara, Gujarat, India.
⁴ Department of Electronics and Communication Engineering, SRM University, Amaravati, Andhra Pradesh, India.
⁵ Division of Electronics and Communication Engineering, The LNM Institute of Information Technology (LNMIIT), Jaipur, India.
⁶ Department of Electrical Engineering, College of Engineering, Taif University, P.O. Box 11099, Taif, 21944, Saudi Arabia.
⁷ Department of Electrical Engineering, Faculty of Energy Engineering, Aswan University, Aswan, 81528, Egypt.
⁸ Faculty of Engineering (FOE), Multimedia University, Persiaran Multimedia, Cyberjaya, 63100, Selangor, Malaysia.
⁹ Department of Electrical, Electronic and Systems Engineering, Faculty of Engineering and Built Environment, Universiti Kebangsaan Malaysia, Bangi, Malaysia. tariqul@ukm.edu.my.

Abstract

A hepta-band terahertz metamaterial absorber (MMA) with modified dual T-shaped resonators deposited on polyimide is presented for sensing applications. The proposed polarization sensitive MMA is ultra-thin (0.061 λ) and compact (0.21 λ) at its lowest operational frequency, with multiple absorption peaks at 1.89, 4.15, 5.32, 5.84, 7.04, 8.02, and 8.13 THz. The impedance matching theory and electric field distribution are investigated to understand the physical mechanism of hepta-band absorption. The sensing functionality is evaluated using a surrounding medium with a refractive index between 1 and 1.1, resulting in good Quality factor (Q) value of 117. The proposed sensor has the highest sensitivity of 4.72 THz/RIU for glucose detection. Extreme randomized tree (ERT) model is utilized to predict absorptivities for intermediate frequencies with unit cell dimensions, substrate thickness, angle variation, and refractive index values to reduce simulation time. The effectiveness of the ERT model in predicting absorption values is evaluated using the Adjusted R² score, which is close to 1.0 for n_min = 2, demonstrating the prediction efficiency in various test cases. The experimental results show that 60% of simulation time and resources can be saved by simulating absorber design using the ERT model. The proposed MMA sensor with an ERT model has potential applications in biomedical fields such as bacterial infections, malaria, and other diseases.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Computer Simulation
Electric Impedance
Electricity*
Machine Learning*
Physical Examination