Using Visible Laser-Based Raman Spectroscopy to Identify the Surface Polarity of Silicon Carbide

详细信息    查看全文

• 作者：Yi-Chuan Tseng ; Yu-Chia Cheng ; Yang-Chun Lee ; Dai-Liang Ma ; Bang-Ying Yu ; Bo-Cheng Lin ; Hsuen-Li Chen
• 刊名：Journal of Physical Chemistry C
• 出版年：2016
• 出版时间：August 18, 2016
• 年：2016
• 卷：120
• 期：32
• 页码：18228-18234
• 全文大小：473K
• 年卷期：0
• ISSN：1932-7455

文摘

In this study we developed an approach to identify the surface polarity of silicon carbide (SiC) by using an excitation laser possessing a photon energy (2.33 eV) much lower than the band gap of 4H-SiC (3.30 eV). By gradually attenuating the intensity of the excitation laser, the effective depth that the laser could generate Raman signals could eventually be limited to within the ultrashallow region of the SiC wafer. Through three-dimensional finite-difference-time-domain (3D-FDTD) simulations, we found that the depth of the high electric field region could be limited from several micrometers below the surface to the near-surface region of 4H-SiC, merely by attenuating the power of the incident laser. Experimentally, we observed a clear trend in the Raman peak intensity ratio of the signals at 210 and 203 cm^–1 in the FTA mode: the intensity ratio of the Si face was always higher than that of the C face regardless of the measurement position on the 4H-SiC wafer. Even through the carrier concentrations in the 4H-SiC wafer were nonuniform, the resulting variability in peak intensity did not influence the trend in the intensity ratio, which could, therefore, be used to identify the surface polarity. This approach might also allow characterization of different polytypes of SiC, for example, 6H-SiC and 3C-SiC, the optical band gaps of which are lower than that of 4H-SiC. Because this optical approach using low-photon-energy laser-based Raman spectroscopy is nondestructive, simple, and rapid and employs excitation light of low photon energy, it should be very applicable for characterizing the surface properties of various other crystalline materials.