Specific Heat of the Ferromagnets Tb(OH)3, Dy(OH)3 and Ho(OH)3

Publisher Website
Google Scholar
Abstract
Continuing a study of the magnetic rare‐earth hydroxides,1 we have measured the specific heats of hydrothermally crystallized powder samples of Tb(OH)3, Dy(OH)3 and Ho(OH)3 between 0.6° and 18°K. Sharp anomalies associated with paramagnetic‐ferromagnetic phase transitions were found at 3.71° Tb(OH)3; 3.48° Dy(OH)3; and 2.54°K Ho(OH)3, in good agreement with the Curie temperatures deduced from magnetic measurements.1 Analysis of the specific heats for Dy(OH)3 and Ho(OH)3 is complicated by Schottky contributions from low‐lying crystal‐field levels, and in the latter compound also by a large hyperfine contribution, Chf. However, for Tb(OH)3 the ground state doublet is far below the first excited state which is at 170°K,2 and the magnetic specific heat Cm could therefore be isolated by a simple analysis of the ``high'' temperature measurements to determine the lattice contribution, and a small calculated correction for Chf at the lowest temperatures. The corresponding total entropy change is estimated to be (0.69±0.05) R, in good agreement with the expected value R In2. The highly anisotropic ground state1−3 makes Tb(OH)3 an unusually good approximation to an Ising ferromagnet, and this was confirmed by an analysis of the low‐temperature tail of Cm. A value for the energy of a single‐spin reversal at T=0°K is deduced: Δ0=8.4±0.3°K, in reasonable agreement with the value 9.2±0.8°K estimated from the integrated specific heat and with the value 9.4±0.2°K obtained spectroscopically.2 The reason for the apparent small discrepancy is not clear. Comparison of the experimental Δ0 with the calculated magnetic dipole contribution (Δ0)dip=11.8°K shows that the ordering in Tb(OH)3 is predominantly, but not exclusively, due to dipolar interactions. A similar situation was found for Dy(OH)3, where Δ0=8.8±0.3°K was determined from the low‐temperature specific heat while the calculated (Δ0)dip=13.3°K. For Ho(OH)3 an accurate value for Δ0 could not be estimated, but it seems likely that the interactions are similar to those in Tb‐ and Dy(OH)3. Full details of this work are being published elsewhere.

This publication has 3 references indexed in Scilit: