Bohr radii of elliptic regions (Q820307)

About ninety years ago, Harald Bohr proved a remarkable theorem concerning complex power series. Usually, the following result is referred to as Bohr's Theorem: If \(f(z)=\sum^\infty_{k=0}a_kz^k\) defines an analytic function mapping the complex open unit disc \(\mathbb{D}\) to itself, then \(\sum^\infty_{k=0}|a_kz^k|<1\) for all \(|z|<1/3\), and the constant \(1/3\), often called the Bohr radius, cannot be improved. Actually, H. Bohr proved a weaker version of this result, namely the statement of the theorem for a disc of radius \(R\leq 1/6\), whereas the sharp bound 1/3, the precise Bohr radius, was established a little later by F. Wiener, M. Riesz and I. Schur, independently. In the late 1990s, the interest in the phenomenon of the Bohr radius underwent an amazing revival, due to various generalizations of Bohr's theorem to several complex variables by \textit{H. P. Boas} and \textit{D. Khavinson} [Proc. Am. Math. Soc. 125, No. 10, 2975--2979 (1997; Zbl 0888.32001)], \textit{L. Aizenberg} [Proc. Am. Math. Soc. 128, 1147--1155 (2000; Zbl 0948.32001)], and many others thereafter. In general, it is very difficult to determine the exact values of the Bohr radii for specific domains in \(\mathbb{C}\) or in \(\mathbb{C}^n\). In the paper under review, the authors return to the classical case of one complex variable and study the Bohr phenomenon for simply connected plane domains \(E\) with analytic Jordan boundaries. In order to establish a meaningful concept of Bohr-type radius for analytic series in such a domain \(E\), and that in a way that makes it suitable for precise estimations, power series are replaced by other series that converge globally on the domain \(E\). To this end, the authors show that, under certain conditions on the domain \(E\), the space of holomorphic functions \(H(E)\) on \(E\) can be described by series of Faber polynomials [cf. \textit{P. K. Suetin}, Series of Faber Polynomials, Moskva: Nauka, 336 p. (1984; Zbl 0936.30026); Transl. Gordon and Breach, Amsterdam (1998; Zbl 0936.30027)], and that some specific properties of such Faber series can be utilized to define the Bohr radius for the domain \(E\) in a quite natural way. This is carried out in the first part of the paper, whereas the main part is then devoted to the finer analysis of this newly established concept of Bohr radius for a domain \(E\) bounded by an ellipse. In this concrete case, the Faber polynomials are well known and sufficiently manageable, as they are essentially Chebyshev polynomials, which enables the authors to derive explicit lower and upper bounds for the (modified) Bohr radius of the elliptic domain \(E\) under the condition that the eccentricity of the ellipse is small enough. The given bounds turn out to be roots of certain quadratic polynomials. On the other hand, it is shown that the Bohr phenomenon does not occur if the eccentricity of the elliptic boundary is larger than the value 0.41, which is quite remarkable. Finally, the classical Bohr radius 1/3 is obtained as the common limit of the author's bounds when the eccentricity of elliptic boundary tends to 0, meaning that \(E\) becomes the closed unit disc. In this vein, the author's approach transpires an interesting new proof of Bohr's classical theorem, and a certain generalization of it to elliptic domains in the bargain.