On spaces of modular forms spanned by eta-quotients (Q488958)

The paper under review considers the problem when the graded rings of holomorphic modular forms for \(\Gamma_0(N)\) are generated by holomorophic or weakly holomorphic eta-quotients of level \(N\). An eta-quotient of level \(N\) is a function of the form \[ f(z)=\prod_{\delta|N}\eta(\delta z)^{r_{\delta}}\;, \] where \(\eta(z)=q^{1/24}\prod_{n=1}^{\infty}(1-q^n)\)~is the eta function defined for \(z\) in the upper half plane \( {\mathbb H}\) and \(r_{\delta}\in {\mathbb Z}\). In the 1950s, \textit{M. Newman} considered these eta-quotients systematically and found conditions for \(f(z)\) being a modular form for \(\Gamma_0(N)\) [Proc. Lond. Math. Soc. (3) 7, 334--350 (1957; Zbl 0097.28701); ibid. (3) 9, 373--387 (1959; Zbl 0178.43001)]. Let \(M_k(\Gamma_0(N))\) denote the space of modular forms of weight \(k\) for \(\Gamma_0(N)\). In this paper, 8 theorems and 3 corollaries are given, some of which are the following: { Theorem 1.} There are precisely 121 positive integers \(N\leq 500\) so that the graded ring of modular forms for \(\Gamma_0(N)\) is generated by eta-quotients. { Theorem 2.} Suppose that \(f(z)\) of the above form is in \(M_k(\Gamma_0(N))\). Then \[ \sum_{\delta|N}|r_{\delta}|\leq 2k\prod_{p|N}\left(\frac{p+1}{p-1} \right)^{\text{min}\{2,\text{ord}_p(N)\}}\;. \] { Corollary 3.} If \(p\) is prime, then \(M_2(\Gamma_0(4p))\) is spanned by eta-quotients if and only if \(p\leq 13\). { Theorem 7.} Suppose \(f(z)\in M_k(\Gamma_0(N))\cap {\mathbb Z}[[q]]\) has the property that \(f\) is non-zero on \( {\mathbb H}\). Then \(f(z)=cg(z)\) where \(c\in {\mathbb Z}\) and \(g(z)\) is an eta-quotient. Also when the compactification \(X_0(N)\) of the algebraic curve \( {\mathbb H}/\Gamma_0(N)\) has genus zero, the following is obtained : { Theorem 11.} Assume that \(N\) is a positive integer so that \(X_0(N)\) has genus zero, and there is a holomorphic eta-quotient in \(M_k(\Gamma_0(N)\). Then the number of eta-quotients in \(M_k(\Gamma_0(N))\) is equal to the number of tuples of non-negative integers \((c_d:\,d|N)\) with \[ \sum_{d|N}c_d\,\phi\left(\gcd\left(d,\,\frac{N}{d}\right)\right)=\frac{k}{12}\cdot [\text{SL}_2({\mathbb Z}):\Gamma_0(N)]\;. \] Here \(\phi(n)\) is the Euler totient function. The key to the proofs for most of the above theorems is the following existence lemma : { Lemma 14.} If \(N\) is a positive integer, then for each divisor \(d\) of \(N\), there is a holomorphic eta-quotient \(E_{d,N}(z)\in M_{k_d}(\Gamma_0(N))\) that vanishes only at the cusps \(c/d\). Moreover, if \(E_{d,N}(z)=\prod_{\delta|N}\eta(\delta z)^{r_\delta}\), \[ \frac{1}{2k_d}\sum_{\delta|N}|r_{de}|\leq \prod_{p|N}\left(\frac{p+1}{p-1}\right)^{\text{min}\{2,\text{ord}_p(N)\}}\;. \]