Comparative Study of Defects in Semiconductors by Electrolyte Electroreflectance and Spectroscopic Ellipsometry

Abstract

It is shown that, because of the electrostriction and polarization of defects by the modulating electric field, electroreflectance spectra are more affected by defects than are third-derivative spectra from ellipsometry data. The theory of electroreflectance is generalized to include these effects. Plasticity and long-range strains are shown to lead to a first-derivative line shape and polarizable defects to a second-derivative line shape. For defect-rich samples these new terms dominate the usual third-derivative line shape. PACS numbers: 78.20.Jq, 71.55.Fr Low-field electrolyte electrorefiectance (EER) has been used primarily to determine accurately the interband transition energies, E,, in semiconductors. According to the existing theory' the EER spectrum is proportional to a linear combination of the third derivatives of the real and imaginary parts, e) and e2, of the complex dielectric function, e(E), with respect to energy. This third-derivative expression leads to an approximate parameterized functional form, the third-derivative functional form (TDFF), which fits remarkably well the EER spectra obtained from defect-free semiconducting surfaces. The fits yield values for the critical energy E,, the linewidth I,, and the phase 0, associated with the Ith transition. Because e~(E) and e2(E) can be measured by automatic spectroscopic ellipsometry (ASE), it is also possible to compute numerically from the independently measured ASE data the third-derivative line shape (TDL) which, according to this theory, should be proportional to the EER line shape.However, our EER data for mercury cadmium telluride (Hgt "Cd"Te) cannot be well fitted by the TDFF and are in serious disagreement with the TDLs obtained from ASE data. In particular, we have studied the effects of different surface treatments on EER spectra4 5 and have reported4 that chemical or chemomechanical treatment of the surface of Hg~"Cd"Te samples substantially changes the linewidth and overall shape of their EER spectra. In contrast, Arwin, Aspnes, and Righer have recently presented ASE results which show that similar surface treatments performed on comparable samples do not measurably affect the linewidth of the TDL. This contrast between EER and ASE results led us to suspect5 the possibility of an interaction between the EER modulating electric field and the defects. This idea was further supported by the failure of the conventional TDFF to fit most Hg~"Cd"Te EER spectra except over a narrow range around each critical-point energy.These results have led us to formulate a new, more general theory of EER which incorporates the effect of defects on the electroreflectance line shape and thus allows us to obtain quantitatively inforrnation about defects in near-surface regions of defectuous semiconductors such as Hg~"Cd"Te. This treatment includes two new terms resulting primarily from the electrostriction and polarization of defects by the EER modulating electric field, and extends the applicability of the attractively simple EER technique to the stu...

Photoreflectance of sulfur-annealed copper indium disulfide

“…These discrepancies have been thought to have resulted from a Burstein-Moss shift due to impurities doping, 1 I or from the effect of nonstoichiometry due to the native defects. 8 In a previous work we studied the temperature-dependent energy gaps of CulnS 2 by the surface barrier electroreflectance (SBER) method. 9 The observation of positive temperature coefficients below 120 K indicates that the band-gap shift is related to the reduction of d-leve1 contributions in the upper valence band due to thermal expansion and the competition of thermal expansion with the electron-phonon interaction.…”

Overview of compositional measurement techniques for HgCdTe with emphasis on IR transmission, energy dispersive X-ray analysis and optical reflectance

“…= W~RZTBRSR&&,B (38) where RI, R,, R , and R , are the reflectivities from mirrors 1, 2, 3 and 4, respectively, R, is the reflectivity from the beam splitter, TB is the transmission of the beam splitter, RdFtA and RdelB are voltage response functions of detectors A and B, respectively, and R, is the reflectivity of the sample under test. R, measured over the E , transition energy range provides the desired reflectance data.…”

“…)/(38) a cancellation of I,, R I and R , yieldsTo obtain R,, mirror 4 is placed at the sample location and detector B is moved to collect the beam to obtain a mirror calibration, as shown in figure26(b). For the mirror calibration the ratio of J&lB/&efA becomes…”