Fig. 2: Dynamic mechanical characterization of four-ribbon-cross devices by measuring the amplitude of the thermomechanical noise in air using LDV.

a–g Thermomechanical noise peak of devices 5–11 using LDV, where the red solid lines in (a–g) were based on Lorentz fitting and extracted resonance frequencies and quality factors. The seven four-ribbon-cross devices have identical trench width (2 µm) and ribbon width (5 µm) but different proof mass dimensions (10 µm × 10 µm × 16.4 µm in (a); 15 µm × 15 µm × 16.4 µm in (b); 20 µm × 20 µm × 16.4 µm in (c); 25 µm × 25 µm × 16.4 µm in (d); 30 µm × 30 µm × 16.4 µm in (e); 40 µm × 40 µm × 16.4 µm in (f); 50 µm × 50 µm × 16.4 µm in (g)). h Resonance frequencies of devices 5–11 (157.3 kHz, 104.3 kHz, 53.1 kHz, 49.4 kHz, 25.2 kHz, 15.6 kHz and 13.6 kHz) versus the side length of the proof mass of devices 5–11. i Quality factor of devices 5–11 (70, 112.5, 38.6, 14.2, 23.7, 14.3 and 14.6) versus the side length of the proof mass of devices 5–11. j–p SEM images of devices 5–11 in (a–g), respectively. One graphene ribbon of device 5 (j) and one graphene ribbon of device 7 (l) had defects