Shifted convolution sums of Fourier coefficients with divisor functions (Q343213)

From the text: Let \(f(z)\) be a primitive holomorphic cusp form of even integral weight \(k\) for the full modular group. Denote its \(n\)th normalized Fourier coefficient (Hecke eigenvalue) by \(\lambda_f(n)\).'' ``Since Selberg's seminal paper [\textit{A. Selberg}, Proc. Sympos. Pure Math. 8, 1--15 (1965; Zbl 0142.33903)], shifted convolution sums with \(\mathrm{GL}(2)\) Fourier coefficients have been investigated extensively by many authors, which led to many important results, such as subconvexity and equidistribution quantum unique ergodicity, etc.'' ``In this paper we are interested in some mixed shifted convolution sums \[ \sum_{n\leq x}| \lambda_f(n^j)| d(n-1),\quad \sum_{n\leq x}| \lambda_f(n^i)\lambda_f(n^j)| d(n-1),\quad \sum_{n\leq x}| \lambda_f(n)^{2j}| d(n-1), \] where \(d(n)\) is the Dirichlet divisor function and \(i,j\in\mathbb Z^+\).'' Fix a positive integer \(j\). The author proves that there exist suitable positive constants \(c(f,j)\) and \( 0< \delta_j < 1\) (which can be explicitly determined) such that \[ \sum_{n\leq x}| \lambda_f(n^j)| d(n-1) = c(f,j) x (\log x)^{\delta_j} (1+o(1)) \] (Theorem 1.1). Similar results are obtained for the remaining two convolution sums (Theorems 1.2 and 1.3).