Comparison of Methods for Characterization of Magnetocaloric Materials

Abstract

I compare three different computational methods for calculating the magnetocaloric effect (MCE) in Fe, Gd, MnBi, and FeRh, using the software package UppASD.The studied Monte Carlo (MC) simulation has a classical approach while the two studied magnon spectra methods have a quantum mechanical approach. The goal was to study the performances of the methods at different temperatures and for different magnetic field changes. The studied quantity was ∆S(T, ∆H), with the Monte Carlo method calculating it with the specific heat capacity obtained with the Metropolis algorithm. The two magnon spectra methods obtain the entropy difference with the magnon density of states, with the adiabatic method (AMS) not including temperature effects in the magnon density of states, in contrast to the dynamical (DMS) method.

The MCE is explained both as an isothermal and as an adiabatic process. The dispersion relation of spin waves is derived and proceeded by magnon theory. The simulation processes are carefully explained and visualized. All the data obtained from UppASD was post-processed with a Python library coded by myself. With the Monte Carlo method it was possible to observe TC and, therefore, the obtained value of TC was used to determine the reliability of the used interaction energies.

The Monte Carlo method showed a good performance at TC for gadolinium and the best of all methods, however, its performance at lower temperatures were poor. The values obtained at T >> TC do not seem physically possible, but the reason behind this is unknown. With MnBi, it was evident that by using poor interaction parameters, both the MC and the AMS method performed poorly. The value for iron at TC for ∆H = 5 T were not in good agreement with the experimental value, however, the extracted experimental value from magnetization data was assumed to be wrong, considering that the obtained TC was close to TCexp. The AMS method worked poorly for materials with a ferromagnetic-to-paramagnetic phase transition, with the absence of temperature effects the reason why the Zeeman energy was not suppressed according to Hm << kBT. The method could observe the antiferromagnetic-to-ferromagnetic phase transition in FeRh though, which the MC method could not. The DMS method were only tested for Gd and showed the best values of all methods for low temperatures. The temperature effects seem to break the shape of the magnon density of states at higher temperatures. The method takes into consideration the suppression of the Zeeman energy due to the thermal energy, however, for ∆H = 4.9 T at 0.16 T /TC it was evident that the method underestimated the contribution of the Zeeman energy