Eddy current determination of the electrical conductivity-temperature relation of Cdl_xZn xTe alloys

A multifrequency eddy current sensor has been installed in a vertical Bridgman furnace and used to measure the electrical conductivity of Cd l_XZn,Te alloys (for x = 0, 0.045 and 0.08) as a function of temperature during heating and cooling through the melting transition. The conductivity of the x = 0.0 and 0.08 samples increased exponentially with temperature up to the melting point. A 4-6 fold increase of conductivity accompanied melting, sufficient for the proposed eddy current sensing of liquid-solid interfaces in this materials system. Above the melting point, the liquid phase conductivity again exponentially increased with temperature. The x = 0.045 sample exhibited similar behavior except in a ~ 30°C interval immediately below the melting/solidification transition on heating and cooling. In this temperature interval, an "anomalous" decrease in conductivity with an increase in temperature was repeatedly observed. Zn has been found to depress the liquid conductivity while that of the solid (near its melting point) exhibited a weak maximum in conductivity at x = 0.045. These observations raise the possibility of eddy current monitoring of melt composition and segregation/homogenization behaviors during post-solidification annealing. 1. I n t r o d u c t i o n Single crystal Cd 1_ ,Zn~Te ( x = 0.045) solid solution alloys are used as substrates for the epitaxial growth of Hg I ~Cd ~Te thin film infrared focal plane array ( IRFPA) detectors [ 1]. As detector manufacturers seek to increase the size and number of IRFPAs per substrate, a demand has been created for large area substrates with low defect densities, uniform distributions of Zn and high infrared transmission coefficients. Either a vertical or horizontal variant of the Bridgman method can be used for the growth of this substrate quality material [1-3]. Unfortunately, * Corresponding author. both the seeded and unseeded growth of vertical Bridgman grown material is usually multigrained with significant Zn segregation (0.02 < x < 0.07), Te precipitation and a sometimes high density of dislocations [3,4]. In spite of many experimental efforts to investigate the relationships between material purity, the controllable growth parameters and the resulting mater ia l charac te r i s t i c s [5,6], the y ie ld of C d l _ ~ Z n , Te of a quality suitable for large area substrates remains disappointingly low ( < 10%). Since much of the poor yield is directly associated with the growth process (e.g. melt stoichiometry, solidification velocity, interface shape, temperature gradients, cooling rate, etc.), intensive efforts are under way to improve this technology. 0022-0248/97/$17.00 Copyright © 1997 Elsevier Science B.V. All rights reserved PII S0022-0248(96)00497-6 324 H.N.G. Wadley, B.W. Choi / Journal qf CP3'stal Growth 172 (1997) 323-336 One approach involves the in situ monitoring of the growth process with multifrequency eddy current sensors [7-13]. The potentially large difference in electrical conductivity of many solid and liquid semiconductors [7,14,15] has led to the proposed use of eddy current sensors for monitoring the liquidsolid interface during crystal growth [7,8,12]. The relationships between an eddy current sensor's frequency response, the electrical conductivities of the solid/liquid and the position/curvature of the liquid-solid interface are complex. Recent electromagnetic finite element modelling has identified several concepts to recover the interface shape during vertical Bridgman growth [8,12], and a companion paper [16] explores their application to Cd0.%Zn0.0aTe. The companion study has also revealed potential applications of eddy current sensors for monitoring the composition of the liquid prior to solidification and observing segregation related phenomena in the solid during post-growth annealing. In order to realize the potential of eddy current sensing, reliable electrical conductivity data as a function of temperature and Zn concentration are needed for both the solid and liquid phases of Cdl_,Zn.,Te alloys over the range of Zn concentrations and temperatures likely to be encountered during vertical Bridgman growth. Comparatively little data relating the electrical conductivity of CdTe to temperature exist for the elevated temperatures encountered in crystal growth, and no data have been published concerning the role of Zn upon the conductivity of Cd~_ ~Zn,Te alloys at high temperature. This has arisen because of the difficulty of making ohmic contacts with semiconductors containing volatile elements like Cd. The temperature and composition dependence of the electrical conductivity of equiatomic CdTe has been investigated by Glazov and coworkers using an electrodeless induction method [ 14,15]. Glazov et al. observed the electrical conductivity of the solid to gradually increase with temperature up to a value of a few hundred S / m in the 10001040°C temperature range. An abrupt rise in conductivity commenced at about 1045°C. During this transition, the conductivity increased by about one order of magnitude and was then followed by a non-linear variation of conductivity with temperature between 1045 and l l00°C. In other semiconductors [15], the abrupt rise in conductivity normally accompanied melting, but in the Glazov et al. experiment it occurred more than 40°C below the accepted melting point of 1092 _+ I°C for equiatomic CdTe [16,17]. Several factors may have contributed to the anomalously low temperature where the conductivity discontinuity was observed. Glazov et al.'s thermometry may have been inaccurate because of the small sample size and the potentially large errors associated with contact thermocouple measurements. A second possibility is that the composition tested may not have been that of equiatomic CdTe due to the high vapor pressure of Cd near the melting point and the significant free volume present in their ampoules [18]. It is also unclear if the trends in conductivity reported by Glazov et al. [14,15] on heating would also occur during cooling. In this study, the electrical conductivities for both solid and liquid CdTe, Cd095sZn004sTe and Cd092Zn00sTe contained in quartz ampoules with very small free volumes have been deduced from multifrequency eddy current sensor data collected in a multi-zone furnace assembly. The temperature and Zn concentration dependence of the electrical conductivity has been obtained for the solid and liquid phases between ~ 650 and 1150°C. 2. Experimental procedures 2.1. Eddy current measurement technique Eddy current testing has become a widely used method for non-destructive materials evaluation and inspection [19-21]. |t enables quantitative measurements of material properties such as electrical conductivity or magnetic permeability, the dimensions of conducting samples via lift-off effects, as well as the detection of crack-like discontinuities in metals. The principle underlying the eddy current method is electromagnetic induction [20,21]. Fluctuating electromagnetic fields are created within the test object by passing an alternating current through a nearby primary (driving) coil. These fluctuating electromagnetic fields are used to induce eddy currents in the test object. The eddy currents in turn create a secondary electromagnetic field which perturbs that of the primary coil. This can be sensed either as a change in the impedance of the primary coil (the H.N.G. Wadley. B. W. Choi / Journal qf Co'stal Growth 172 (1997) 323-336 325