On stable summands of Thom spectra of \(B(\mathbb{Z}/2)^n\) associated to Steinberg modules (Q1970357)

This is a generalization of a result of Mitchell and Priddy which shows that \(e_n B(\mathbb{Z}/2)= M(n)\) and there exists a stable splitting \(M(n) \simeq L(n)\vee L(n-1)\). Here \(e_n\in \widehat\mathbb{Z}_2 [\text{GL}_n (\mathbb{F}_2)]\) denotes the Steinberg idempotent. After replacing a \(\text{GL}_n\)-action on \(B(\mathbb{Z}/2)^n\), the author defines a stable summand \(M(n)_k\) of the Thom spectrum \((B(\mathbb{Z}/2)^n)^{k \widetilde\alpha_n}\) by using a homotopy colimit such that \(M(n)_0=M(n)\) and \(M(n)_1= L(n)\). Here \(\widetilde\alpha\) is the vector bundle associated to the reduced real regular representation of \((\mathbb{Z}/2)^n\). The main theorem in this paper is the existence of a cofiber sequence \(\Sigma^k M(n-1)_{2k+1}\to M(n)_k \to M(n)_{k+1} \to\Sigma^{k+1} M(n-1)_{2k+1}\), which respects the above splitting if \(k=0\). This sequence is also related to Kuhn's sequence. In Corollary B, it is said that \(K(m)^* (M(n)_{2k+ 1}) \cong 0\) if \(n>m\), and the connecting map of the cofiber sequence \(M(n-1)_{4 k+1} \to M(n-1)_{4k-1}\) is a \(K(n-1)_*\)-equivalence. This follows immediately from a result of Kuhn that \(K(m)^* (L(n))\cong 0\) if \(n>m\). The main theorem is proved by showing the exactness of the sequence \(0\to H^*(M(n)_{k+1}) \to H^*(M (n)_k) \to H^*(M(n-1)_{2k+1}) \to 0\) using a colimit spectral sequence of Bousfield-Kan.