\(p\)-adic deformation of motivic Chow groups (Q1757202): Difference between revisions

Let \(Y_{\bullet}\) be a smooth projective formal scheme over \(W(k)\), \(k\) being a perfect field of characteristic \(p > 0\). Let \(m\) be a positive integer, and \(X = Y_{\bullet} \otimes_{W(k)} W_{m}(k)\). Assume (for simplicity) that \(\dim X < p\) and there is a good smooth projective lifting \(X_{\bullet}\) of \(X\) to \(W(W_m(k))\). The author explains how to define a motivic complex \(\mathbb{Z}_{X_{\bullet}}(r)\) and its finite level \(\mathbb{Z}_{X}(r)\) as complexes on \(X\) with respect to the Nisnevich topology, and characterizes the image of the map \[ \mathrm{CH}^{r}_{\mathrm{cont}}(Y_{\bullet}) := {H}_{\mathrm{cont}}^{2r}(X,\mathbb{Z}_{X_{\bullet}}(r)) \to {H}^{2r}(X,\mathbb{Z}_{X}(r)). \] The main theorem states that if \(r > p\), then an element \(\xi\) lies in the said image if and only if its cycle class falls in the image of \(H^{2r}(X_{\bullet},F^{r}\Omega^{\ast}_{X_{\bullet}}) \to H^{2r}(X_{\bullet},\Omega^{\ast}_{X_{\bullet}})\), \(F^{\bullet}\) being the Hodge filtration. When \(m=1\), \({H}^{2r}(X,\mathbb{Z}_{X}(r))\) is the usual Chow group of the algebraic variety \(X\) over \(k\), and this result due to \textit{S. Bloch} et al. [Invent. Math. 195, No. 3, 673--722 (2014; Zbl 1301.19005)]. Since the base ring is \(W_{m}(k)\) instead of \(k\), the author needs the ``relative'' version of the de Rham-Witt complex developed by the author and \textit{T. Zink} [J. Inst. Math. Jussieu 3, No. 2, 231--314 (2004; Zbl 1100.14506)]. It was conjectured by the author and \textit{T. Zink} [Doc. Math. 12, 147--191 (2007; Zbl 1122.14016)] that the Nygaard filtration on the relative de Rham-Witt complex ``equals'' the ``relative version'' of the filtration considered by \textit{P. Berthelot} and \textit{A. Ogus} [Notes on crystalline cohomology. Princeton, NJ: Princeton University Press (1978; Zbl 0383.14010)] in the filtered derived category. This is conjecture is proved in the article under review, under an extra hypothesis that \(p\) is bigger than \(\dim X\) (or more accurately, bigger than the cohomological degree one is interested in). Just like the ``absolute case'', both Nygaard and Ogus filtrations can be identified using a divided power filtration on a suitable divided power de Rham complex, when \(p\) is large. The author then explains how to construct syntomic complexes and motivic complexes in the relative context. The theorem is proved using the interrelations among the various definitions.