Comparison of the Bergman and Szegö kernels (Q643439)

The authors use \(L^{2}\) estimates for the \(\overline{\partial}\)-operator on bounded, pseudoconvex domains in \(\mathbb{C}^{n}\) with class \(C^{2}\) smooth boundary to control the quantity \[ \frac{S(z,z)}{\delta(z)K(z,z)}, \] where \(S\) is the Szegő kernel function, \(K\) is the Bergman kernel function, and \(\delta\) is the distance to the boundary of the domain. Their upper bound implies that if \(\varepsilon\) is an arbitrary positive number, then the indicated fraction is bounded above by \(|\log \delta(z)|^{n+\varepsilon}\) when \(z\) is sufficiently close to the boundary of the domain. If, additionally, the domain is of finite type or admits a defining function that is plurisubharmonic on the boundary, then the indicated fraction is bounded below by \(|\log\delta(z)|^{-(1+\varepsilon)}\) when \(z\) is sufficiently close to the boundary. On a more general class of domains, there is a logarithmic lower bound that involves the Diederich-Fornæss exponent. In the special case of convex domains, no logarithmic factors are needed, and the indicated fraction is bounded above and below by constants.