Toda brackets and congruences of modular forms (Q545858): Difference between revisions

One of the most classical invariants of homotopy theory is the degree of a map in \(\pi_n(S^n) = \pi_0^{st}(S^0)\). Adams defined his \(e\)-invariant \[ F_1\pi_*^{st} \to \mathbb{Q}/\mathbb{Z} \] on stable homotopy classes with vanishing degree, or equivalently, of Adams-Novikov filtration 1. The \(e\)-invariant detects the image of \(J\), which agrees with the \(1\)-line of the Adams-Novikov spectral sequence. In [Topology 38, No. 2, 387--425 (1999; Zbl 0924.55004)], the author went one step further to define the \(f\)-invariant \[ \pi_{2n}^{st} \to D/(M_0+M_{n-1})\otimes \mathbb{Q}/\mathbb{Z} \] on all even stable classes for \(n>0\). On these the \(e\)-invariant vanishes and they have Adams-Novikov filtration 2. In the target, \(M_i\) denotes the ring of integral modular forms of weight \(i\) and \(D\) the ring of linear combinations of rational modular forms such that the resulting \(q\)-expansion of the sum is integral, all with respect to an implicitly chosen congruence subgroup \(\Gamma_1(N) \subset SL_2(\mathbb{Z})\) for \(N\geq 2\). Here, integral refers to the ring \(\mathbb{Z}[\frac1N, \zeta_n]\) and rational to \(\mathbb{Q}(\zeta_n)\). The ring \(D\) was introduced in [\textit{N. M. Katz}, Ann. Math. (2) 101, 332--367 (1975; Zbl 0356.10020)], motivated by the study of congruences of modular forms. Like the \(e\)-invariant detects the \(1\)-line, the \(f\)-invariant detects the \(2\)-line of the Adams-Novikov spectral sequence away from \(N\). Such a complete invariant is a good tool to show whether two different descriptions of homotopy classes actually define the same class. For example, \textit{H. von Bodecker} gave in his 2009-preprint [``On the \(f\)-invariant of products'', \url{arXiv:0909.3968}] a formula for the \(f\)-invariant of products of elements of filtration 1, which implies (in principle) a complete determination of these products. Laures gives in the present paper formulae for the \(f\)-invariants of 3- and 4-fold Toda brackets in terms of \(e\)- and \(f\)-invariants. He mainly concentrates on the cases where one, in the case of 3-fold Toda brackets, or two, in the case of 4-fold Toda brackets, of the entries are equal to the prime \(p\) one is working at. The author gives several sample computations of \(f\)-invariants of Toda brackets. For example, he shows thus that \(\langle \sigma^2, 2,\sigma^2, 2\rangle\) contains the Kervaire class \(\beta_{8/8} \in \pi_{30}^{st}(S^0)\).