Hydrogen sensors for sub-ppm levels are vital for various applications. We have presented in 2021 the Electrostatically Formed Nanowire (EFN), a multi-gate very large-scale integration (VLSI) compatible transistor for Ppm level Hydrogen sensing [1]. We show here that the EFN sensor response for Hydrogen is improved by employing a specially designed large area sensing antenna that enhances the EFN sensor response by several orders of magnitudes. Figure 1 below shows the sensor response of the large area (red) and the small area conventional (blue) EFN to H2 concentration in the range of 30 ppb to 100 ppm. For the concentration of 30 ppb, the large area antenna EFN shows a normalized average response of ~87%, whereas, which to the best of our knowledge is a world record for near room temperature sensing. 

Another significant advantage of the EFN is its ability to achieve very high selectivity to various gases and gas mixtures using a single sensor. This is achieved by creating a two-dimensional ‘response map’ representing the measured response as a function of the applied voltage on the EFN back and side gates. We then apply both classical machine learning algorithms and deep learning to differentiate (with over 90 % accuracy) between low concentrations of various gas molecules and their mixtures. Such single sensor selectivity together with the EFN CMOS compatibility paves the way to a very promising miniature, low power selective H2 sensing platform.

References

[1] A. Mukarjee, M. Gnaim, I. Shem Tov, L. Hargreaves, J. Hayon, A. Shluger, and Y. Rosenwaks, Sensors and Actuators B 2021, 248, 130509.

[2] I. Shem Tov, A. Mukherjee, Z. Mutsafi, E. Pikhay, D. Greental, Y. Roizin, Y. Mazor, and Y. Rosenwaks, ACS Applied Electronic Materials, 2024, 6, 3610.