The second Hilbert coefficients and the homological torsions of parameters (Q2254327)

Let \((A,m)\) be a Noetherian local ring of dimension \(d >0\), and denote by \(E = E_A(A/m)\) the injective envelope of \(A/m\). Let \(M\) be a finitely generated \(A\)-module of dimension \(r\) with \(M_j = \Hom_A(H^j_m(M), E)\) for all \(j \in \mathbb Z\). Let \(I\) be an \(m\)-primary ideal in \(A\), there exist integers \(\left\{e^i_I(M) \right\}_ {0 \leq i \leq r}\) such that \[ l_A(M/I^{n+1}M)= e^0_I(M) {n+r \choose r} -e_I^1(M){n+r-1 \choose r-1}+ \dots +(-1)^re_I^r(M) \] for all \(n \gg 0\). The coefficient \(e^i_I(M)\) is called the \(i^{th}\) Hilbert coefficient of \(M\) with respect to \(I\), and the leading coefficient \(e^0_I(M)\) is called the multiplicity. The homological degree of \(M\) with respect to \(I\) is defined inductively by \(\mathrm{hdeg}_I(M)=l_A(M)\) when \(r \leq 0\) and \(\mathrm{hdeg}_I(M)=e_I^0(M)+ \sum_{j=0}^{r-1}{r-1 \choose j} \mathrm{hdeg}_I(M_j)\) when \(r>0\). The homological torsions of modules are defined by \(T_I^i(M)= \sum_{j=1}^{r-i} {r-i-1\choose j-1}\mathrm{hdeg}_I(M_j)\) for any \(1 \leq i \leq r-1\). Let \(Q\) be a parameter ideal for \(M\). The first coefficient \(e^1_Q(M)\) has been previously studied (see for instance [\textit{L. Ghezzi} et al., J. Lond. Math. Soc., II. Ser. 81, No. 3, 679--695 (2010; Zbl 1200.13037); ``The Chern and Euler coefficients of modules'', to appear in Acta Math. Vietnam., \url{arXiv:1109.5628}; \textit{S. Goto} and \textit{K. Ozeki}, Bull. Braz. Math. Soc. (N.S.) 45, No. 4, 679--709 (2014; Zbl 1320.13030); \textit{S. Goto} and \textit{K. Ozeki}, Contemp. Math. 555, 97--118 (2011; Zbl 1264.13017)]) and is sufficiently understood. In the paper under review, the authors study the second coefficient \(e^2_Q(M)\) in terms of the homological degrees and torsions of modules. They show in theorems 1.3 and 1.4 that if \(M\) is a finitely generated \(A\)-module with \(\dim_AM= \dim A \geq 3\), then for every parameter ideal \(Q\) of \(A\): {\parindent=6mm \begin{itemize}\item[(1)] \(e^2_Q(M) \leq T^2_Q(M)\). \item[(2)] \(- \sum_{j=2}^{d-1} {d-3 \choose j-2}\mathrm{hdeg}_Q(M_j) \leq e^2_Q(M)\), when \(M\) is unmixed. \item[(3)] \(e^2_Q(M)=T^2_Q(M)\) is equivalent to \(l_A(M/QM)-e^0_Q(M)+e^1_Q(M)=\mathrm{hdeg}_Q(M) -e^0_Q(M)-T^1_Q(M)\) when \(M\) is unmixed . \item[(4)] Further, in the case when \(e^2_Q(M)=T^2_Q(M)\), they obtain: {\parindent=12mm \begin{itemize}\item[{\(\bullet\)}] \((-1)^ie^i_Q(M) = T^i_Q(M)\) for \(3 \leq i \leq d - 1\) and \(e^d_Q(M) = 0\), \item[{\(\bullet\)}] \(l_A(M/Q^{n+1}M) = \sum^d_{i=0} (-1)^ie^i_Q(M) {n+d-i \choose d-i}\) for all \(n \geq 0\), \item[{\(\bullet\)}] \(Q = \left(a_1, a_2, \dots, a_d \right)\) for some \(a_1, a_2, \dots, a_d \in A\) that form a \(d\)-sequence on \(M\), \item[{\(\bullet\)}] \(QH^i_m(M) = (0)\) for all \(1 \leq i \leq d-3\). \end{itemize}}\end{itemize}}