On the Hurewicz image of the Steinberg summand \(M(n)\) (Q1957135): Difference between revisions

Let \(p\) be a prime, and cohomology \(H^\ast(-)\) and homology with coefficients in \(\mathbb F_p\). There has been a lot of research on the classifying spaces \(B(\mathbb Z/p)^n\). The following results about the Hurewicz image of \(B(\mathbb Z/p)^n\) are known: \textit{I. Hansen} [Math. Z. 157, 43--52 (1977; Zbl 0387.57019)] has determined the image of the stable Hurewicz map \(\pi^S_\ast(B\mathbb Z/p)\rightarrow H_\ast(B\mathbb Z/p;\mathbb Z)\) completely. \textit{N. Minami} [Trans. Am. Math. Soc. 351, No.~6, 2325--2351 (1999; Zbl 0932.55012)] has shown that the mod \(p\) stable Hurewicz map \(\pi^S_\ast(B(\mathbb Z/p)^n)\rightarrow H_\ast(B(\mathbb Z/p)^n)\) has trivial image in some dimensions, but it is far from complete determination. A lot of work has been done on the \(\mathcal{A}_\ast\)-primitive elements \(PH_\ast(B(\mathbb Z/p)^n)\), or equivalently the indecomposable quotient module \(\mathcal{Q}H^\ast(B(\mathbb Z/p)^n)=\mathbb F_p\otimes_AH^\ast(B(\mathbb Z/p)^n)\), which lifts to a minimal \(\mathcal{A}\)-generating set for \(H^\ast(B(\mathbb{Z}/p)^n)\), where \(\mathcal{A}\) and \(\mathcal{A}_\ast\) are the Steenrod algebra and its dual, respectively. One of the reasons is the existence of \textit{W. M. Singer}'s transfer [Math. Z. 202, No.~4, 493--523 (1989; Zbl 0687.55014)] \(\varphi:F_p\otimes_{GL_n}PH_\ast(B(\mathbb Z/p)^n)\rightarrow Ext^{n,\ast+n}_\mathcal{A}(\mathbb F_p,\mathbb F_p)\), where \(Ext^{n,\ast+n}_\mathcal{A}(\mathbb F_p,\mathbb F_p)\) is the \(E_2\)-term of the Adams spectral sequence converging to \(\pi^S_\ast(S^0)\). For each idempotent \(e\) in the group ring \(\mathbb{F}_p[GL_n(\mathbb{Z}/p)]\) is obtained a stable summand \(eB(\mathbb{Z}/p)^n\) with \(H^\ast(eB(\mathbb{Z}/p)^n)\cong H^\ast(B(\mathbb{Z}/p)^n)e\) [cf. \textit{S. A. Mitchell, S. B. Priddy}, Topology 22, 285--298 (1983; Zbl 0526.55010), Am. J. Math. 106, 219--232 (1984; Zbl 0551.55005)]. Each idempotent corresponds to a modular representation of \(GL_n(\mathbb{Z}/p)\) over \(\mathbb{F}_p\). \textit{S. A. Mitchell} [Math. Z. 189, 1--9 (1985; Zbl 0547.55017)] has shown that the transfer \(tr_n:B(\mathbb Z/p)^n_+\rightarrow S^0\) factors trough the summand corresponding to the trivial representation. Singer's transfer \(\varphi_n\) is the algebraic analog of \(tr_n\). In the present paper the author considers a spectrum \(M(n)\) defined by \(e_nB(\mathbb{Z}/p)^n\), where \(e_n\) is the Steinberg idempotent corresponding to the Steinberg representation (for details of the Steinberg idempotent see \textit{S. A. Mitchell} and \textit{S. B. Priddy} [op. cit.] and \textit{R. Steinberg} [Can. J. Math. 8, 580--591 (1956; Zbl 0073.01502)]). \textit{N. J. Kuhn} [Math. Proc. Camb. Philos. Soc. 92, 467--483 (1982; Zbl 0515.55005)], \textit{N. J. Kuhn, S. A. Mitchell} and \textit{S. B. Priddy} [Bull. Am. Math. Soc., New Ser. 7, 255--258 (1982; Zbl 0502.55011)], \textit{N. J. Kuhn} and \textit{S. B. Priddy} [Math. Proc. Camb. Philos. Soc. 98, 459--480 (1985; Zbl 0584.55007)] and \textit{S. A. Mitchell} and \textit{S. B. Priddy} [op. cit] have shown that \(M(n)\) has various good properties. Though it is difficult to determine \(H^\ast(eB(\mathbb{Z}/p)^n)\cong H^\ast(B(\mathbb{Z}/p)^n)e\) in terms of modular representation theory, Mitchell and Priddy have described \(H^\ast(M(n))\) in terms of the Steenrod algebra. Moreover \(B(\mathbb{Z}/p)^n\) has \(p^{\frac{n(n-1)}{2}}\) stable summands equivalent to \(M(n)\). The author studies the mod \(p\) Hurewicz image of \(M(n)\). This describes a certain part of the mod\,\(p\) stable Hurewicz image of \(B(\mathbb{Z}/p)^n\). In two other previous papers [Recent progress in homotopy theory. Proceedings of a conference, Baltimore, MD, USA, March 17-27, 2000. Providence, RI: American Mathematical Society (AMS). Contemp. Math. 293, 125--139 (2002; Zbl 1022.55007); J. Lond. Math. Soc., II. Ser. 75, No.~2, 317--329 (2007; Zbl 1122.55012)] the author has determined a basis of \(\mathcal{Q}H^ast(M(n))\). The results imply dim \(\mathcal{Q}H^k(M(n))=\) dim \(PH_k(M(n))\leq 1\). For a spectrum \(X\), let \(h(X)_\ast:\pi_\ast(X)\rightarrow H_\ast(X)\) be the mod \(p\) Hurewicz map. Since the image of \(h(X)_\ast\) is included in \(PH_\ast(X)\), the image of \(h(M(n))_k\) is trivial if \(\dim \mathcal{Q}H^k(M(n))=0\). If dim \(\mathcal{Q}H^k(M(n))=1\), it is isomorphic to \(0\) or \(\mathbb{F}_p\). The main result of the paper is that the image of \(h(M(n))_\ast\) is trivial for most dimensions. If \(L(n)=\sum^{-n}(Sp^{p^n}(S^0)/Sp^{p^{n-1}}(S^0))\), where \(Sp^mS^0\) is the \(m\)th symmetric product of \(S^0\), then \(M(n)\simeq L(n)\vee L(n-1)\) [\textit{S. A. Mitchell} and \textit{S. B. Priddy}, Topology 22, 285--298 (1983; Zbl 0526.55010)], and then \(H^\ast(M(n))\cong H^\ast(L(n))\oplus H^\ast(L(n-1))\). Because of this, the author actually studies \(L(n)\). The author gives a partial computation of the \(BP\)-primitive elements \(PBP_\ast(L(n))\) by using the \(BP\)-Adams operation \(\psi^{n+1}\). By refinement of his previous results the author describes a basis of \(PH_\ast(L(n))\). The Hurewicz map \(h(L(n))_\ast:\pi_\ast(L(n))\rightarrow PH_\ast(L(n))\) factors through the usual reduction \(PBP_\ast(L(n))\rightarrow PH_\ast(L(n))\). These induce triviality of \(h(L(n))_\ast\) for certain dimensions. And finally the author obtains the following theorem. Theorem C. The image of \(h_H(M(n))_k:\pi_k(M(n))\rightarrow H_k(M(n))\) is trivial if : \(k\neq 2(p-1)(p^{k_1}+...+p^{k_{n-1}}+1)-n, (k_1>...>k_{n-1}\geq 1), 2(p-1)(p^{k'_1}+...+p^{k'_{n-2}}+1)-n+1,(k'_1>...>k'_{n-2}\geq 1)\) for \(p>2\); \(k\neq 2(2^{k_1}+...+2^{k_n})-n,(k_1>...>k_n, k_n\in \{0,1,2\}), 2(2^{k'_1}+...+2^{k'_{n-1}})-n+1, (k'_1>...>k'_{n-1},k'_{n-1}\in \{0,1,2\})\) for \(p=2\).