Research Article

A Novel THz Differential Spectral Clustering Recognition Method Based on t-SNE

Table 1

The sample information of CMC and silicon.

The sample nameSpecificationDescription

Aluminum oxide (Al2O3) sheet with defectsLength × width × thickness (mm): 101.0 × 101.0 × 3.0 mm
Bulk density ≥ 3.65 g/cm3
Surface roughness Ra ≥ 0.5 μm
Principal component content ≥ 95%
Gray-white, defects are internal irregularities cracks

Beryllium oxide (BeO) sheet with defectsLength × width × thickness (mm): 50.8 × 50.8 × 1.5
Bulk density ≥ 2.85 g/cm3
Surface roughness Ra ≥ 0.1 μm
Principal component content ≥ 99%
Pure white, highly toxic powder, the defect is the shape of the regular set of holes 4
Outside diameter: 4.2–5.9 mm
Inner diameter: 2.6–4.2 mm
Hole depth: 0.8 mm

Beryllium oxide (BeO) sheet with zero defectsLength × width × thickness (mm): 50.8 × 50.8 × 1.5
Bulk density ≥ 2.85 g/cm3
Surface roughness Ra ≥ 0.1 μm
Principal component content ≥ 99%
Pure white, highly toxic powder, zero defects

Monocrystalline silicon with zero defectsThickness :   450 ± 10 μm
Diameter :  76.3 ± 0.2 mm
N dopant (phosphorus)
Crystal orientation:<111>
Crystal growth: CZ resistivity: 14–45 Ωcm zero defects