Coefficients of products of powers of eta functions (Q1601121)

Define sequences \(b(m)\) and \(c(m)\), \(m=0,1,2\dots\), by \[ \sum_{m=0}^\infty b(m)q^m= \prod_{n=1}^\infty \frac{(1-q^n)^3} {(1-q^{2n})} \quad\text{and}\quad \sum_{m=0}^\infty c(m) q^m= \prod_{n=1}^\infty \frac{(1-q^{2n})^3} {(1-q^n)}. \] Let \(p\) be a prime. Then for the following values of \(p\) and \(\varepsilon\) \[ \varepsilon= \begin{cases} \phantom{-}1, &\text{if }p\equiv 7,13\text{ or }23\pmod{24},\\ -1, &\text{if }p\equiv 11,17\text{ or }19\pmod{24}, \end{cases} \] the author proves that \[ b(p^2m+ (p^2-1)/24)= \varepsilon b(m) \quad\text{and}\quad c(p^2m+ 5(p^2-1)/24)= \varepsilon c(m). \] These results, in the cases \(p\equiv 7\) or \(11\pmod {24}\) were conjectured by the reviewer, \textit{M. Hirschhorn} and \textit{R. Lewis} [Ramanujan J. 4, 137-155 (2000; Zbl 0989.11053)]. The remaining conjectures in the above cited paper have since been settled by \textit{S. Ahlgren} [J. Number Theory 89, 222-233 (2001; Zbl 0981.11017)]. The methods used in each paper are different. The paper under review makes use of the imaginary quadratic field \(\mathbb{Q} (\sqrt{-6})\) and properties of \(L\)-functions. Cooper et al. used explicit formulas arising from Macdonald identities of ranks 1 and 2, and Ahlgren used modular forms.