Szegő kernel, regular quantizations and spherical CR-structures (Q2435112)

Let \((L, h)\) be a geometric quantization of a compact Kähler manifold \((M, g)\) (that is, \((L, h)\) is a positive Hermitian line bundle over \(M\) such that the Ricci curvature of \(h\) is just the Kähler form \(\omega_g\) given by \(g\)). If \((L^*, h^*)\) denotes the negative dual bundle, we can consider the unit disk bundle \(D\) over \(M\): \[ D = \{v \in L | \rho(v) = 1- h^*(v,v) > 0 \} \] Let \(S\) denote the Szegő kernel of \(D\): \[ S(v) = \sum_{j=1}^{+\infty} \, f_j(v) \overline{f_j(v)},\quad v\in D, \] where \(\{f_j\}\) is an orthonormal basis of the Hardy space \({\mathcal H}^2(X)\) and \(X = \partial D\). As we know (after the result by \textit{C. Fefferman} [Invent. Math. 26, 1--65 (1974; Zbl 0289.32012)] and \textit{L. Boutet de Monvel} and \textit{J. Sjöstrand} [Astérisque 34--35, 123--164 (1976; Zbl 0344.32010)]) we have the following decomposition \[ S(v) = a(v) \rho(v)^{-n-1} + b(v) \log \rho(v) \] where \(a(v)\) and \(b(v)\) are smooth functions; \(b(v)\) is just called the log-term of \(S\). Two questions naturally arise: { Question 1}: If the log-term vanishes, then is it \(X\) locally CR-equivalent to the sphere? { Question 2}: Assume that \(X\) is locally CR-equivalent to the sphere. Is it true that \(M\) is biholomorphic to \(\mathbb C P^n\). \textit{M. Engliš} and \textit{G. Zhang} [Math. Z. 264, No. 4, 901--912 (2010; Zbl 1197.32008)] gave a negative answer to Question 1 and a positive answer to question 2 when \(L\) is a positive line bundle over an Hermitian symmetric space of compact type. The authors show that the log-term of the Szegő kernel vanishes in many other cases. Concerning Question 2, they prove that, among all homogeneous spaces, \(\mathbb C P^n\) is the only one for which \(X\) is locally CR-equivalent to the sphere \(S^{2n+1}\).