A two-dimensional limit theorem for Mellin transforms of the Riemann zeta-function (Q1041394)

The author continues his investigation [Acta Arith. 226, 173--184 (2006; Zbl 1098.11044)] concerning the (modified) Mellin transforms \[ {\mathcal Z}_k(s) := \int_1^\infty |\zeta(\tfrac 12+it)|^{2k}x^{-s}\,dx \qquad(k\in\mathbb N, \Re s > \sigma_0(k) >1). \] This function plays an important rôle in the theory of \(\zeta(s)\), the Riemann zeta-function. In the present work he proves the two-dimensional limit theorem for the Mellin transforms \({\mathcal Z}_1(s)\) and \({\mathcal Z}_2(s)\): let \(\sigma_1> 1/2\) and \(\sigma > 5/6\). Then on \((\mathbb C^2, {\mathcal B}(\mathbb C^2))\) (\({\mathcal B}(S)\) is the class of Borel sets on \(S\)) there exists a probability measure \(P_{\sigma_1,\sigma_2}\) such that the probability measure \(P_T\) converges weakly to \(P_{\sigma_1,\sigma_2}\) as \(T\to\infty\). We have \[ P_T(A) := \nu_T\Bigl(({\mathcal Z}_1(\sigma+it), {\mathcal Z}_1(\sigma+it))\in A\Bigr),\quad \nu_T(\dots) = {1\over T}\mu\Bigl(\{t\in [0,T];\dots\}\Bigr), \] where \(\mu(\cdot)\) denotes measure. The proof is based on intricate techniques from probabilistic number theory.