The stable Adams conjecture and higher associative structures on Moore spectra (Q2078809)

The Adams conjecture, proven in different ways by \textit{D. Quillen} [Topology 10, 67--80 (1971; Zbl 0219.55013)] and \textit{D. Sullivan} [Ann. Math. (2) 100, 1--79 (1974; Zbl 0355.57007)], states that for \(p\) odd and \(q\) a topological generator of \(\hat{\mathbb{Z}}_p^\times\), the composite \[\mathrm{J} \circ (\psi^q - 1) : \mathrm{BU} \to \mathrm{BU} \to \mathrm{BSL}_1(\mathbb{S})\] of the relevant Adams operation and the \(\mathrm{J}\)-homomorphism is \(p\)-locally nullhomotopic. The stable Adams conjecture is the stronger statement that this composite ought to be null as an infinite loop map. This paper provides two new proofs of the stable Adams conjecture, and shows that the previously accepted proof [\textit{E. M. Friedlander}, Math. Proc. Camb. Philos. Soc. 87, 109--150 (1980; Zbl 0426.55010)] is incorrect. The first new proof proceeds by applying the \(\mathrm{K}\)-theory functor to a certain multinatural transformation. The second new proof is essentially a correction to Friedlander's work. The authors give applications to the existence of \(\mathbb{A}_n\)-structures on Moore spectra \(\mathrm{M}_p(i)\). Following recent work of the first author [Adv. Math. 402, Article ID 108319, 26 p. (2022; Zbl 1501.55010)], they observe that \(\mathrm{M}_p(i)\) is the Thom spectrum associated to a map \(f_{p,i}: S^1 \to \mathrm{BGL}_1(\hat{\mathbb{S}}_p)\) representing a class \(1+p^i u \in \pi_1(\hat{\mathbb{S}}_p) \cong \hat{\mathbb{Z}}_p\), for \(u \in \hat{\mathbb{Z}}_p^\times\), and that if \(f_{p,i}\) is an \(\mathbb{A}_n\)-map then \(\mathrm{M}_p(i)\) inherits an \(\mathbb{A}_n\)-structure. They define a Thomified \(\mathbb{A}_n\)-structure to be one induced in this manner. At odd primes \(p\), the authors show that \(\mathrm{M}_p(i)\) admits a Thomified \(\mathbb{A}_n\)-structure if and only if \(n < p^i\) (Theorem 1.13). When \(p=2\), the result is slightly weaker: they show that for \(i \geq 3\), \(\mathrm{M}_2(i)\) admits a Thomified \(\mathbb{A}_{2^{i-1}-1}\)-structure but does not admit a Thomified \(\mathbb{A}_{2^i}\)-structure. The authors point out (Remark 7.7) that there may exist `exotic' (i.e. not Thomified) \(\mathbb{A}_n\)-structures on \(\mathrm{M}_p(i)\), and suggest this as a direction for future work.