Computer Fourier‐transform techniques for precise spectrum measurements of oscillatory data with application to the de Haas–van Alphen effect

摘要: Methods are developed for the high‐accuracy analysis of oscillatory data using discrete fast Fourier transform. A wide dynamic range and linearity response together with good separation individual lines in a spectrum achieved by digital filtering to reduce sidelobes less than −120 dB relative central peak. Periodic errors associated nature transform reduced interpolation fitting. To make use high accuracy inherent data, least‐squares method is which fits line shapes accurately matched filter Fourier‐transform spectrum. Frequencies measured an (computational) better 0.00001% amplitudes 0.001%. Although more generally applicable problems requiring highly precise spectral analysis, techniques here directed de Haas–van Alphen oscillations where many frequencies present strong function magnetic field. An interpretation made ’’amplitude’’ from it possible calculate (i.e., Fermi surface areas), Dingle temperatures electron scattering times), effective masses 0.00005% or better. This realized record having only modest number points.