No abstract

The magnetization of three Pb 1Ϫx Eu x Se samples, with xϭ1.3,3.0, and 4.1 %, was measured at 30 and 50 mK in magnetic fields H up to 50 kOe, and at 0.6 K in fields up to 180 kOe. For xϭ1.3% and with Hʈ͓100͔, a magnetization step ͑MST͒ due to an energy-level crossing for isolated Eu 2ϩ ions was observed at 30 and 50 mK. The magnetic field at this MST, 1.76Ϯ0.2 kOe, was close to the predicted value H c ϭ1.98 kOe. At the same low temperatures ͑30 and 50 mK͒ but at higher fields, a magnetization ''ramp'' due to pairs was observed in all samples. For xϭ3.0 and 4.1 % this ramp consisted of well-resolved MST's arising from pairs. A ramp due to open triplets was also observed in these two samples. The MST's due to pairs were used to obtain the value J/k B ϭϪ0.24Ϯ0.03 K for the dominant antiferromagnetic exchange constant. Comparisons between the measured magnetization curves at 30 or 50 mK and theoretical simulations indicates that this J is the nearest-neighbor ͑NN͒ exchange constant J 1 . At 0.62 K the magnetization of each of the three samples rose rapidly with H in fields below several kOe. At higher fields a rounded ramp due to pairs and triplets was present. This ramp ended near 40 kOe, and complete saturation was achieved near 50 kOe. A model which includes only the NN exchange constant J 1 gave a reasonably good account for all the data at 0.62 K. Calculated magnetization curves for pairs, and for open and closed triplets, at various values of k B T/͉J͉ are presented. The effects of the single-ion and dipole-dipole anisotropies on the MST's due to pairs are also discussed.

The magnetization of Pb1−xEuxTe samples with x = 1.9, 2.6 and 6.0% was measured at 20 mK in fields up to 50 kOe, and at 0.6 K in fields up to 180 kOe. The 20 mK data show the magnetization steps (MSTs) arising from pairs and from triplets. The pair MSTs are used to obtain the dominant Eu-Eu antiferromagnetic exchange constant, J/kB = −0.264 ± 0.018 K. The exchange constant for triplets is the same. Comparison of the magnetization curves with theoretical simulations indicates that the Eu ions are not randomly distributed over all the cation sites. The deviation from a random distribution is much smaller if J is assumed to be the nearest-neighbor exchange constant J1 rather than the next-nearest-neighbor exchange constant J2. On this basis, J is tentatively identified as J1. However, the possibility that J = J2 cannot be excluded completely. To obtain agreement with the data, it must be assumed that the Eu ions tend to bunch together. Comparision with microprobe data indicates that the length scale for these concentration variations is smaller than a few µm. The theoretical simulations in the present work improve on those performed earlier by including clusters larger than three spins.75.50. Pp, 75.30.Et, 75.60.Ej

No abstract

No abstract