
2D MATERIAL DETECTOR
2D material Detector for mechanical exfoliated flakes
Automated detection of mechanically exfoliated 2D materials:
-
AI based Upgrade: to distinguish between contaminant and 2D materials based on a CNN neural network.
-
Eliminates the time-consuming manual search and documentation under optical microscopes.
-
Compatible with all 2D materials — even unknown van der Waals crystals.
-
Accelerates research in novel 2D material discovery.
Specification:
-
scan range: 5cm x 5cm maximum
-
Objective lens: 10x, 20x, 50x, all compatible
-
materials: graphene, hBN, TMDC, MXene, etc
-
including all 2d materials
-
substrate: SiO2/Si, glass,
-
detection speed: 1.3 sec/frame
-
scan speed: 0.5 um^2/sec.
-
accuracy: single layer (for graphene)
-
image format: jpg file
-
coordinate: recorded in file name




Mechanically exfoliated MoS₂: 6-hour scan on a 5 × 5 cm² SiO₂(300 nm)/Si substrate:
With our newly upgraded algorithm, we achieve over 90% detection accuracy for single-layer graphene. Built-in auto-focusing enables drift correction and stable, long-duration scans (>10 hours) across large areas. For example, scanning a 5 × 5 cm² wafer divided into 400 µm × 400 µm frames results in 125 × 125 = 15,625 frames. At 1.38 seconds per frame, a full scan takes approximately 6 hours.From over 15,000 images, the system successfully identifies just a handful (e.g., ~10) of flakes — a task impossible for humans, but made effortless by the machine.






