Fig. 1: Shortcomings of the Fourier transform analysis. | Nature Communications

Fig. 1: Shortcomings of the Fourier transform analysis.

From: Dictionary learning in Fourier-transform scanning tunneling spectroscopy

Fig. 1

a The experimentally obtained LDoS of NaFeAs over a 100 × 100 nm2 area of the sample at energy ω = 60 mV from ref. 18 with junction conditions V = − 100 mV and I = 300 pA measured at T = 26 K. Color bar units are nS. b, c Real part and magnitude, respectively, of the FT of a showing large phase noise. Color bar units are nS−1. d Simulated image of scattering from a single point defect. e, f Real part and magnitude, respectively, of the FT of d. The LDoS scale in d was chosen such that the FT magnitude for the simulated single defect had a maximum of 1. gi Corresponding simulated LDoS and FTs for scattering from 70 point defects that are randomly distributed in space. The color bars for di are in arbitrary units. All FT spectra are shown for  −3π/5a ≤ qxqy≤ 3π/5a. Comparison with the single defect case shows that noise in the FTs arises from the random placement of defects. Furthermore, the maximum amplitude of the FT scales approximately as \(\sqrt{N}\), resulting in a loss of signal fidelity in the FT analysis.

Back to article page