A colts study of the deep levels in n- and p- type cdte thin films deposited by hot wall evaporation

CdTe thin films, both undoped and with different dopants, have been deposited unto graphite and Corning 7059 glass substrates using a Three-Stage Hot Wall Vacuum Evaporator (TSHWVE) system. The dopants were incorporated into the CdTe
thin films using a "delta doping" technique. The conductivity type of the doped CdTe thin films was determined using the hot probe method, and the film stoichiometry was determined using X-ray and Auger electron spectroscopy measurements. Schottky diodes fabricated on the CdTe thin films that were deposited on graphite substrates
have been studied using Current-Voltage (I-V), Capacitance-Voltage (C-V), and Capacitance Deep Level Transient Spectroscopy (CDLTS). The conductivity type of CdTe films that were undoped and doped with Antimony (Sb), Phosphorus (P), Gold (Au), Silver (Ag), and Copper (Cu) were found to
be p-type, while Indium (In) doped CdTe thin films were found to be n-type. The highest carrier concentration of the CdTe films are 1 x 1016 cm-3, 1 x 1017 cm-3. and 7.5 x 1015 cm-3 for In-, Sb-, and P-doped CdTe, respectively.
From the I-V measurements for the Schottky diodes fabricated on n- and p-type CdTe thin films, it was found that the series resistance of the diodes is still large (60 Q for the best diode on p-type and 217 @ for the best diode on n-type), and its effect on the C-V data had to be corrected. The large value of diode ideality factor and diffusion voltage suggest that the diodes are more likely to be MIS
structures, rather than ideal Schottky barriers.For the In-doped CdTe films three majority carrier traps are found with
activation energies measured from the conduction band of 0.23 + 0.05 eV, 0.46 + 0.06 eV, and 0.78 ₺ 0.05 eV. The trap concentrations of the three traps are 0.10 x 1014 cm-3, 1.11x1014 cm-3, and 3.04x1014 cm-3, respectively, while the capture cross-section df the three traps are 0.56x10-16 cm2, 0.44x10-16 cm2, and 0.47x10-16 cm2 respectively. By comparison with the theoretical scheme of Kroger, one can
tentatively identify the level at 0.23 + 0.05 eV as being caused by Cd interstitials, the level at 0.46 + 0.06 eV as being caused by Te vacancies, and the level at 0.78 + d.os eV as being caused by Cd vacancies.For the Sb-doped CdTe films three majority carrier traps are found with activation energies measured from the valence band of 0.27 + 0.06 eV, 0.50 ₺ 0.06 dV, and 0.80 + 0.06 eV. The trap concentrations of the three traps are 0.39x1013 com -3. 1.02x1013 cm-3, and 2.70x1013 cm-3, respectively, while the capture cross-section of the traps are 0.62x10-16 cm2, 0.92x10-16 cm2, and 1.70x10-16 cm2, respectively. By comparison with the defect levels given by Zanio, one can tentatively identify the
level at 0.27 ₺ 0.06 eV as being caused by a defect due to Sb, and the levels at 0.50 0.06 eV and at 0.80 + 0.06 eV as being caused by a complex Te interstitial or a doubly charged Cd vacancy. For the P-doped CdTe films three majority carrier traps are found with activation energies measured from the valence band of 0.28 + 0.05 eV, 0.50 + 0.06 eV and 0.75 ₺ 0.05 eV. The trap concentrations of the three traps are 0.27x1013 cm-3 1:56x1013 cm-3, and 2.7x1013, respectively, while their capture cross-sections are 363x10-16 cm2, 4.21x10-16 cm2, and 4.48x10-16 cm2, respectively. By comparison with the defect levels given by Zanio, one can tentatively identify the level at 0.28 + 0.05 ev as being caused by a defect due to P, and the levels at 0.50 + 0.06 eV and 0.75 ₺ 005 eV as being caused by a complex Te interstitial or a doubly charged Cd vacancy. Our capture measurements on In-, Sb-, and P-doped CdTe showed non-expo-
dential transients, however they could be fitted very well by Pons theory, and aallowed us to determine values for the trap concentration (NT), the trap capture rate (En,p) and the trap capture cross-section (on,p). However, the capture cross-sections so derived are approximately two orders of magnitude smaller than estimates given in the literature earlier. Whether the discrepancy is due to errors in the earlier estimates or in problems with the applicability of Pons theory to CdTe is unknown it this time. However, errors in our capture measurements producing non-exponential transients which are misinterpreted by Pons theory do not seem likely.