Rational period functions in higher level cases (Q2517123)

Let \(p\in\{1,2,3\}, k\in\mathbb Z\) and let \(f\) be a meromorphic function on the complex upper half plane. For \(\gamma=\left(\begin{smallmatrix} a&b\\c&d\end{smallmatrix}\right)\in\text{SL}_2(\mathbb R)\), put \(f|_k(z)=(cz+d)^{-k}f(\gamma z)\). Let \(T=\left(\begin{smallmatrix} 1&1\\0&1\end{smallmatrix}\right)\), \(W_p=\left(\begin{smallmatrix} 0&-1/\sqrt p\\ \sqrt p&0\end{smallmatrix}\right)\) and \(U=TW_p\). Let \(\mathrm{RPF}_{2k}(\Gamma_0^+(p))\) be the set of rational functions \(q(z)\) satisfying \(q|_{2k}(W_p+1)=0\) and \(q|_{2k}(U^{n_p-1}+U^{n_p-2}+\cdots +1)=0\), where \(n_p\) is the order of \(U\) as an element of \(\text{PSL}_2(\mathbb R)\) and \(\Gamma_0^+(p)=\langle \Gamma_0(p),W_p\rangle=\langle T,W_p\rangle\). An element of \(\mathrm{RPF}_{2k}(\Gamma_0^+(p))\) is called a rational period function of weight \(2k\) for \(\Gamma_0^+(p)\). In this article, the authors extend the results for rational period functions for \(\Gamma_0^+(1)=\text{SL}_2(\mathbb Z)\) in [\textit{M. I. Knopp}, Duke Math. J. 45, 47--62 (1978; Zbl 0374.10014); Glasg. Math. J. 22, 185--197 (1981; Zbl 0459.10017)] to the cases that \(p=2,3\). Let \(p=2,3\). They show that for odd positive integers \(k\), \(q_{2k}^+(z)=(z^2-z-1/p)^{-k}+(z^2+z-1/p)^{-k}\in \mathrm{RPF}_{2k}(\Gamma_0^+(p))\), and that if \(z_0\) is a finite pole of \(q(z)\in \mathrm{RPF}_{2k}(\Gamma_0^+(p))\), then \(z_0\) is a real quadratic number and further if \(z_0\in\mathbb Q\), then \(z_0=0\). Furthermore, they show that \(q(z)\in \mathrm{RPF}_{2k}(\Gamma_0^+(p))\) if and only if there exists a modular integral \(F(z)\) for \(q(z)\), that is to say, \(F(z)\) is a meromorphic function satisfying \(F|_{2k}T=F\) and \(F|_{2k}(W_p-1)=q(z)\). For an positive integer \(n\) greater than \(1\) and prime to \(p\), they define a Hecke operator \(\hat{T}_{2k,n}\) on \(\mathrm{RPF}_{2k}(\Gamma_0^+(p))\) following Knopp [loc. cit.], by \(\hat{T}_{2k,n}(q)=(F|T_n)|_{2k}(W_p-1)\), where \(T_n\) is the Hecke operator on modular forms and \(F\) is any modular integral of \(q\). By showing that the set \(\{\hat{T}_{2k,n}(q_{2k}^+)\}\) is linearly independent, they obtain that \(\mathrm{RPF}_{2k}(\Gamma_0^+(p))\) is infinite dimensional in the case that \(k\) is an odd positive integer. The restriction to primes \(\{2,3\}\) is essential, because the arguments need that \(U\) has finite order and that \(\Gamma_0(p)^+=\langle \Gamma_0(p),W_p\rangle\) is generated by \(T\) and \(W_p\). These conditions are satisfied only for \(p\in\{1,2,3\}\).