Application of the maximum-entropy technique to the analysis of de Haas-van Alphen data.

Abstract

In this work it is shown how the analysis of de Haas-van Alphen (dHvA) magnetic oscillations, traditionally performed with the use of fast Fourier transform (FFT), can be greatly improved by employing maximum entropy spectral analysis (MESA). In electronic systems where, for example, magnetic breakdown leads to close-lying dHvA frequencies of different amplitudes, such analysis is far superior to the conventional FFT, where large bandwidth constantly leads to problems of interpretation. In the paper the applicability of the MESA technique is discussed and its advantages are demonstrated by analyzing test data obtained in ${\mathrm{Ni}}_{3}$Al and Zr${\mathrm{Zn}}_{2}$. Both compounds are weakly ferromagnetic itinerant electron systems showing rich interplay of exchange splitting, spin-orbit effects, and magnetic breakdown. The amplitude of the dHvA oscillations is relatively weak, damped by impurity scattering and the effect of high effective masses. By showing examples of analysis in these systems it is argued that the MESA technique may form a basis for the next generation of high-resolution dHvA analysis of the Fermi surface in solids.