Continuity and finiteness of the radius of convergence of a \(p\)-adic differential equation via potential theory (Q890755)

The paper under review studies the radius of convergence of a differential equation on a smooth Berkovich curve over a non-archimedean complete valued field of characteristic 0. In particular, under the additional assumption that the curve has no boundary or that the differential equation is overconvergent, the authors gave a new proof of the continuity of the radius function (previously proved by \textit{F. Baldassarri} [Invent. Math. 182, No. 3, 513--584 (2010; Zbl 1221.14027)]) and that the radius function factorizes by the retraction through a locally finite graph (previously proved by the two authors [Acta Math. 214, No. 2, 307--355 (2015; Zbl 1332.12013); Acta Math. 214, No. 2, 357--393 (2015; Zbl 1332.12012)]). Roughly speaking, this new proof relies only on two technical inputs: (1) the radius functions on an annulus or a disc is continuous, concave, piecewise log-linear whose slopes have bounded denominators (this is proved by \textit{G. Christol} and \textit{B. Dwork} [Ann. Inst. Fourier 44, No. 3, 663--701 (1994; Zbl 0859.12004)] and \textit{K. S. Kedlaya} [\(p\)-adic differential equations. Cambridge: Cambridge University Press (2010; Zbl 1213.12009)]), and (2) the potential theory on Berkovich curves, as developed by A. Thuillier in his thesis. The key step lies in proving the radius of convergence is a super-harmonic function, at least locally in affinoid neighborhoods of type \(2\) points. The main theorems follow from general properties of super-harmonic functions.