Chromatic splitting for the \(K(2)\)-local sphere at \(p=2\) (Q2138596)

The main result of the paper is the calculation of the homotopy type of \(L_1 L_{K(2)} S^0\) and \(L_{K(1)} L_{K(2)} S^0\) at the prime 2, where \(L_{K(n)}\) is localization with respect to Morava \(K\)-theory and \(L_1\) is localization with respect to 2-local \(K\)-theory. Specifically, the authors construct weak equivalences \[ L_1( S^0_2 \vee S^{-1}_2 ) \vee L_0( S^{-3}_2 \vee S^{-4}_2 ) \vee L_1(\Sigma^{-2} V(0) \vee \Sigma^{-3} V(0) ) \longrightarrow L_1 L_{K(2)} S^0 \] and \[ L_{K(1)}( S^0 \vee S^{-1} \vee \Sigma^{-2} V(0) \vee \Sigma^{-3} V(0) ) \longrightarrow L_{K(1)} L_{K(2)} S^0 \] where \(S^n_2\) is the 2-completed \(n\)-sphere and \(V(0)\) is the mod-2 Moore spectrum. In particular, \(L_1 S^0_2 \to L_1 L_{K(2)} S^0\) is a split inclusion. Both weak equivalences are induced from a particular map from wedges of spheres and \(V(0)\) to \(L_{K(2)} S^0\). The first weak equivalence shows that the sphere summands predicted by the Chromatic Splitting Conjecture (see [\textit{M. Hovey}, Contemp. Math. 181, 225--250 (1995; Zbl 0830.55004)]) are correct. However, the Moore spectra summands are new. This result proves the Strong Chromatic Splitting Conjecture false, as was already known by work of [\textit{A. Beaudry}, Geom. Topol. 21, No. 6, 3213--3230 (2017; Zbl 1421.55010)]. However, the result is compatible with the weaker form of the conjecture, see [\textit{A. Beaudry} and \textit{T. Barthel}, in: Handbook of homotopy theory. Boca Raton, FL: CRC Press. 163--220 (2020; Zbl 1476.55026)]. A major component to the result is a radical simplification of group cohomology. Let \(E\) be Morava \(E\)-theory at height 2 at the prime 2. Let \(\mathbb{W}\) be the Witt vectors on \(\mathbb{F}_4\). There is an isomorphism \[ E_* \cong \mathbb{W} [\![ u_1 ]\!] [u^{\pm 1}] \] where the power series is in degree 0 and \(u\) has degree 2, hence \(E_0 = \mathbb{W} [\![ u_1 ]\!]\). The inclusion of constants into the power series ring induce isomorphisms in group cohomology \[ H_c^*(\mathbb{G}_2, \mathbb{F}_4) \longrightarrow H_c^*(\mathbb{G}_2, E_0/2) \quad \text{ and } \quad H_c^*(\mathbb{G}_2, \mathbb{W}) \longrightarrow H_c^*(\mathbb{G}_2, E_0) \] The authors make the conjecture that this result is true at all heights and all primes, referred to as the Chromatic Vanishing Conjecture.