Low-temperature specific heat and low-field magnetic susceptibility were measured in ${\mathrm{Cd}}_{1-x}{\mathrm{Mn}}_x\mathrm{Te}$ mixed crystals for $0.002<~x<~0.70$. Three regions of composition can be distinguished in both experiments. For $x<\mathrm{}0.17\mathrm{}$ the crystal is paramagnetic at all temperatures. For $0.17<x<\mathrm{}0.7\mathrm{}$, a spin-glass phase is observed, as evidenced by the characteristic cusp in the susceptibility and a linear temperature dependence of the specific heat in the low-temperature regions. Because the material is an insulator at low temperatures, and the Mn interactions are only antiferromagnetic, we believe that the observed spin-glass phase is produced by the frustration of the lattice. For $x=0.7\mathrm{}$ an antiferromagnetic phase is observed. To understand the experimental behavior of the specific heat and susceptibility as a function of temperature and magnetic field for $0.002<x<\mathrm{}0.17\mathrm{}$, we must assume that the distribution of Mn ions deviates strongly from a random distribution. The number of pairs is more than doubled for $x=0.05\mathrm{}$, and 30% higher for $x=0.1\mathrm{}$, than statistically predicted. The number of larger clusters, like triplets, is also significantly higher. Analysis of the data yielded the value of the exchange integral for the nearest-neighbor interaction to be $\frac{J_{\mathrm{NN}}}{k}=-0.55\mathrm{}\pm 0.05$ K. The interaction is stronger within the larger clusters, and is described by a different exchange constant $\frac{J_{\mathrm{NN}}^{\prime }}{k}=-4.3\mathrm{}$ K.

Low-dc-field magnetic-susceptibility measurements are reported for $\mathrm{Cu}\mathrm{Mn}$ alloys containing 1.08- and 2.02-at.% Mn using a superconducting quantum-interference device. The zero-field cooled initial susceptibility shows a cusplike peak at $T_g$, the transition temperature to a spin-glass state. After field cooling from $T>\mathrm{}{T}_g$ (even for $H=2\mathrm{}$ G), the susceptibility below $T_g$ is nearly independent of temperature with the constant value of $\sim 0.97\chi \left(T_g\right)$. The high-temperature susceptibility follows a Curie-Weiss law until just above $T_g$; the value of the Curie temperature is very small. Using the theory of Sherrington and Kirkpatrick, the spin-glass order parameter $q\left(T\right)\mathrm{}$ is extracted from the susceptibility data. A critical index $\beta \simeq 1$ is obtained assuming a power law for $q\left(T\right)\mathrm{}\propto {(1-\frac{T}{T_g})}^{\beta }$. The results are qualitatively in agreement with the prediction of an Edwards-Anderson-type mean-field theory.

A systematic study of the low-dc-field magnetic susceptibility and the specific heat has been carried out on mixed ${\mathrm{Hg}}_{1-x}{\mathrm{Mn}}_x\mathrm{Te}$ crystals, in the composition range $0<~x<~0.35$. The alloy with $x=0.35\mathrm{}$ showed spin-glass behavior below $T=10.9\mathrm{}$ K. At this Mn concentration the sample is a very poor conductor at low temperatures, so that the Ruderman-Kittel-Kasuya-Yosida (RKKY) mechanism cannot be responsible for the spin-glass transition. Also, since the Mn ions interact only antiferromagnetically, the observed spin-glass phase does not result from competition between ferro- and antiferromagnetic interaction. It must therefore be ascribed to the frustration of the antiferromagnetic interactions inherent in an fcc sublattice over which the Mn ions are distributed. For $x<~0.25$, the ${\mathrm{Hg}}_{1-x}{\mathrm{Mn}}_x\mathrm{Te}$ samples remain paramagnetic down to 1 K. Experimental results for the specific heat and the susceptibility for $x<\mathrm{}0.1\mathrm{}$ are discussed in terms of a cluster model which leads to an estimated value of the antiferromagnetic exchange constant $\frac{J}{k}\approx -0.7\mathrm{}\pm 0.3$ K. When a random distribution of Mn ions over the fcc sublattice is assumed, calculated values for the specific heat and the susceptibility differ substantially from the experimental results for the low Mn concentration. This therefore indicates that in real crystals the distribution of Mn ions is very different from random. To obtain agreement between calculated and experimental results, the number of single ions has to be reduced to less than 30% of the number corresponding to a purely random distribution, leading to the conclusion that the magnetic ions prefer to cluster rather than to remain isolated in ${\mathrm{Hg}}_{1-x}{\mathrm{Mn}}_x\mathrm{Te}$. For HgTe, the Debye temperature ${\Theta }_D$ is 141 K at $T=0\mathrm{}$ K, and goes through an anomalously low minimum value of ${\Theta }_D=77\mathrm{}$ K at 7 K. A very small linear term in the low-temperature specific heat of HgTe gives an estimate of the electric charge carrier density in the ${10}^{18}$ ${\mathrm{cm}}^{-3\mathrm{}}$ range.

Crystal structure and magnetic susceptibility were investigated for a thiospinel compound CuIr 2 S 4 , which exhibits a transition from metallic to insulating on cooling at the transition temperature T c of 230 K. It was found that the transition is accompanied by a first-order structural transition. As the compound is cooled through T c , the structure transforms from cubic to tetragonal with a volume contraction of 0.7%. The tetragonal phase has an axial ratio of about 1.03. Magnetic susceptibility, which is paramagnetic in the metallic region, decreases sharply and becomes diamagnetic at T c on cooling. The susceptibility is mostly temperature independent in each region above and below T c . Magnetic measurements also revealed that high pressure stabilizes the insulating phase, in contrast with most of the compounds that exhibit metal-insulator transitions. Possible mechanisms of the metal-insulator transition are discussed.