Comment on 'Interpretation of Fermi level pinning on 4H-SiC using synchrotron photoemission spectroscopy' [Appl. Phys. Lett. 84, 538 (2004)]

摘要

In a recent letter in this journal, Han and Lee reported on an investigation of pinning mechanisms of the Fermi level on 4H-SiC using synchrotron radiation photoemission spectroscopy (SRPES) and deep level transient spectroscopy (DLTS). The authors suggested that the Fermi level (E_(F)) pinning at the surface of p-type SiC (p-SiC) originated from two kinds of hole traps in p-SiC, observed in DLTS measurements. However, the authors neglected important features of the problem. The description is given later. First, Han and Lee showed that the Fermi level (E_(F)) pinning on the surface of p-SiC could be interpreted using schematic energy band diagrams, deduced from the SRPES results, as shown in Fig. 4 of Ref. 1. For the measurement of SRPES spectra, the thickness of the Ni layer (deposed using a thermal evaporator) of Ni/p-SiC was calculated to be 0.4 nm by the authors. In order to determine the Schottky barrier height (φ) of Ni/p-SiC Schottky diodes and characterize the E_(F) pinning on the surface of p-SiC using DLTS, a Ni (100 nm) was deposited on the p-SiC surface by the authors, using an electron beam evaporator. Wu and Khan have shown the movement of the surface E_(F)) on n- and p-GaN as a function of metal coverage and suggested that the final stabilized position gives the φ. Monch has also indicated that a sufficiently thick metal coverage is required before the surface band bending reaches its final barrier height which can be correlated with electrically measured Schottky barrier heights. The insufficiently thick Ni layer used in the photoemission measurements in Ref. 1 might lead to an error in the value of the Schottky barrier of 0.95 eV deduced from the analysis of their Fig. 4. This may explain the discrepancy with the value of φ=1.15 eV given in Fig. 1 of Ref. 1. Further investigations are require about the SRPES results; as shown in Fig. 4 of Ref. 1.