Emergence, persistence, and disappearance of logarithmic spirals in the spectra of singular integral operators (Q1924281)

The spectrum of the Cauchy singular integral operator in \(L_p\) space on piecewise smooth curves and an exponential (power-like) weight is well-known and consists of pieces of arcs. The situation changes considerably when the weights are more general e.g. for a general Muckenhoupt weight we get ``thick'' arcs as found by I. Spitkovsky with the help of the factorization method. The present paper systematically investigates what is happening when the weight is general (Muckenhoupt weight) and the curve is Carleson's. From the authors' introduction: ``In the absence of a weight the spectrum may contain certain circular arcs if the curve is sufficiently nice; these arcs metamorphose into logarithmic double spirals for more complicated curves and, in the case of a general composed Carleson curve, these double spirals may blow up to heavy sets whose boundaries are nevertheless composed of pieces of logarithmic spirals. Muckenhoupt weights may further thicken the spectrum; until some point the weights are unable to destroy the circular arcs and logarithmic spirals, but beyond this point some kind of interference between the curve and the weight result in a complete disappearance of spirality \dots''. Applying the modified method of Mellin transforms, due to A. Dynin, the authors start with the simplest (model) cases of ``star-like'' curves (straight and logarithmic spirals). They fix the emergence of double spirals in the spectra and, continuing the investigation, show the persistence of a double spiral under ``non-pathological'' deformation of the curve. In Subsection 3.3 they apply the Wiener-Hopf method and consider a Carleson-Jordan curve. They discover the emergence of massive horn-like spectra, which might disappear under certain interference between non spiral-like curves and non power-like weights (see Section 4). Many graphical illustrations of the spectra of the Cauchy singular integral operator for different curves and different weights serve for a better understanding of the exposed results.