Search

The diffusion of Hg into CdTe has been studied, under saturated vapour pressure conditions, as a function of time at 300 and 350°C. The Hg concentration profiles showed two components at short and intermediate anneal times reverting to single component at long times, indicating that the diffusion of Hg in CdTe is rate limiting involving two stream diffusion.

The self-diffusion of Cd into Cu doped CdTe is reported as a function of Cu concentration at 523 and 820°C. Some of the diffusion profiles possessed two components and were fitted by a function composed of the sum of two complementary error functions (erfc) whereas others possessed a single component and were fitted by a single erfc. This implied that the diffusion was rate limiting. In addition D Cd increased linearly with [Cu] up to a concentration of 3 × 10 18 atoms cm −3 at both temperatures but at 820°C, D Cd showed a sudden increase at concentrations above this value.

The diffusion of Hg into CdTe under saturated vapour pressure conditions has been studied in the temperature range 160–403°C. Two component profiles were obtained and the diffusion was rate limited giving two values of the diffusivity. All Arrhenius graphs gave two straight lines with an abrupt change of gradient at 275°C, which was probably due to a change in the diffusion mechanism.

In this paper the use of the three standard defect etches for CdTe: Nakagawa's reagent, ferric chloride and Inoue's EAg-1 reagent have been compared and it has been demonstrated, by measuring the etch pit densities (EPD), that Inoue's EAg-1 reagent can be used successfully for making comparisons in CdTe slices which have been given various physical treatments. In addition a graph of the diffusivity of mercury in CdTe, D(Hg), obtained using radiotracer sectioning techniques, versus EPD shows that a tenfold increase (from 4×10 4 to 3×10 5 cm -2 ) in the surface EPD produces an increase of a factor of four in D(Hg)

Measurement of the self-diffusion of Cd in CdTe carried out in an atmosphere of excess Cd and in the temperature range 350–650°C is reported. The diffusion profiles, which were measured using an anodic oxidation sectioning technique, showed two components, a shallow Fickian one and a second, much deeper component which was non-Fickian. When the results for the shallow diffusion component were combined with those reported earlier by Jones and Stewart, an Arrhenius graph was obtained where the diffusion coefficient extended over nearly nine decades in the temperature range 350–920°C. The graph can be represented by the sum of two exponential functions given by the Arrhenius parameters D 01 = 125 cm 2 s −1 , E 1 = 2.55 eV, D 02 = 2.48×10 −8 cm 2 s −1 and E 2 = 1.08 eV,which indicates that at least two diffusion mechanisms were operatin this temperature range. No significant difference was observed in the diffusion coefficients for bulk grown and epitaxially grown material and no significant pressure dependency was observed at 489°C. For the shallow diffusion component, agreement with previously published data is good at temperatures above 625°C, but below this temperature, agreement with the limited amount of data available is poor. This may be due to the higher purity of CdTe used in this study.

The diffusion and solubility of Cu in CdTe single crystal material has been measured in the temperature range 200–400°C using a radiotracer sectioning technique. The diffusivity in this temperature range is given by the Arrhenius relationship: D = 6.65 x 10 -5 (cm 2 s -1 ) exp[-0.57 eV / kT ] and the corresponding surface Cu concentration in the diffused CdTe slices is given by the Arrhenius relationship C 0 = 1.56 x 10 23 (cm -3 ) exp[-0.55 eV / kT ]. The addition of Cd or Te to the capsule did not result in profiles that were significantly different to those given by the above expressions and diffusion probably occurred via a defect of the form (Cu i Cd v )'. The major portion of the activation energy for the diffusion was taken up in the formation of the defect rather than in its migration through the lattice. A much faster diffusion mechanism (≈100 times faster), which did not show any consistent behaviour with temperature, was also observed in the diffusion profiles. Because of this, CdTe will not act as an effective barrier, in the fabrication of Hg x Cd 1− x Te (MCT) infra-red detectors, to prevent Cu diffusing from the substrate through the CdTe buffer layer into the epitaxially grown MCT layer.

Measurements of the self diffusion of Cd into the ternary compound In0.67xCd1-xS, where 0

Results of measurements on the self-diffusion of Cd into CdTe carried out in an atmosphere of excess Cd and in the temperature range 570–920°C are given in this report. CdTe produced by Mining and Chemical Products Limited (MCP) and the University of Durham were used for the measurements. The diffusion profiles obtained using a radiotracer sectioning technique agreed closely with the solution to the diffusion equation which assumes a constant source term for the diffusion. The diffusion coefficients gave a straight line when plotted on an Arrhenius graph and did not show any flattening out at lower temperatures and the best fit through the points gave a value of E = (2.47±0.10) eV and D 0 = (50 × / ÷ 3.0) cm 2 s -1 . When compared with similar measurements reported in the sixties they agree closely with those due to Borsenberger and Stevenson but the agreement is poor when compared with the results due to Whelan and Shaw.

Additional measurements are reported here on the self-diffusion of Cd in CdS which were carried out in an atmosphere of excess Cd vapour in the temperature range 600–1000°C and measured using a radiotracer sectioning method. These results were analysed along with similar results reported by Jones much earlier and the values of D (Cd) reported here are similar to those reported by Jones. When all the values of D (Cd) over the whole temperature range are taken together, they are found to be consistently lower than similar results reported by Shaw and by Woodbury, but the difference with those reported by Shaw is not significant. When plotted on an Arrhenius graph, the results can be divided into two separate straight line sections with different slopes which change at an approximate temperature of 780°C. This behaviour is also similar to that reported by Shaw. In addition, no source dilution occurred during the diffusion anneal and the values of D (Cd) for CdS obtained from two different manufacturers gave similar results.