Impact of interdigitated electrodes design on the low frequency and random telegraph noise of single-layer graphene micro ribbons

High performance devices consisting of interdigitated electrodes (IDEs) on top of single-layer graphene (SLG) are candidates with favorable prospects for sensing applications. Graphene micro ribbons (GMRs) of various widths and IDE design geometries were fabricated and experimentally examined regarding their low-frequency noise (LFN) behavior. Measurements revealed a 1/f behavior and different kinds of trap activity behind it, which were studied through the analysis of random telegraph noise (RTN) signals. Our investigation suggests that adjusting the geometrical characteristics of either the GMR width or the IDE topology can significantly influence the signal-to-noise ratio (SNR) of SLG-based electronics. On the bright side, the results of our study can provide useful guidelines for fabrication decisions to maximize the SNR.

💡 Research Summary

This paper investigates how the geometry of interdigitated electrodes (IDEs) placed on single‑layer graphene (SLG) micro‑ribbons influences low‑frequency noise (LFN) and random telegraph noise (RTN), both of which are critical for the performance of graphene‑based sensors. Single‑layer graphene was grown by chemical vapor deposition (CVD) on copper foil, transferred onto a 300 nm SiO₂/Si substrate using a PMMA‑assisted wet‑transfer, and subsequently patterned into graphene micro‑ribbons (GMRs) with widths (W) of 50 µm, 100 µm, and 200 µm via electron‑beam lithography and oxygen plasma etching. Aluminum interdigitated electrodes were then fabricated on top of the ribbons with three different inter‑electrode spacings (G) of 8 µm, 15 µm, and 25 µm, yielding nine distinct device configurations (3 × 3 matrix).

Electrical characterization using an HP4155A parameter analyzer showed that the I‑V characteristics are essentially Ohmic for currents up to ~20 µA (ln V = ln R + n ln I with n≈1). Above this current level, a deviation from linearity appears, which the authors attribute to self‑heating of the graphene ribbon and possible grain‑boundary conduction. The resistance values extracted from the linear region follow the expected sheet‑resistance scaling R = R_sheet·