Totally reflexive modules over connected sums with \(\mathfrak{m}^3 =0\) (Q2074408)

In this review any ring \(R\) is a Noetherian standard graded algebra over a field \(k\). Recall that by definition, an \(R\)-module \(M\) is totally reflexive if \(M = M^{**}\) and \[ \mathrm{Ext}_{R}^{i}(M,R) = \mathrm{Ext}_{R}^{i}(M^*,R) = 0, \qquad\forall i \geq 1. \] Note that totally reflexive modules, introduced in [\textit{M. Auslander} and \textit{M. Bridger}, Stable module theory. Providence, RI: American Mathematical Society (AMS) (1969; Zbl 0204.36402)], play an important role in the theory of Gorenstein dimension. Also, recall that by definition, a ring \(R\) is called \(G\)-regular if the only totally reflexive \(R\)-modules are the free \(R\)-modules. In Theorem 3.1 in [\textit{Y. Yoshino}, NATO Sci. Ser. II, Math. Phys. Chem. 115, 255--273 (2003; Zbl 1071.13505)], it is assumed that \((R,\mathfrak{m}_R)\) is a local ring with \(\mathfrak{m}^3 = 0\) which is not Gorenstein and also not \(G\)-regular. As the author of the paper under review reports, there is no known criterion for deciding if a given non-Gorenstein ring is \(G\)-regular or not. On the other hand, a result in [\textit{L. W. Christensen} et al., Adv. Math. 218, No. 4, 1012--1026 (2008; Zbl 1148.14004)] states that ``if \(R\) is not Gorenstein, then there are either infinitely many isomorphism classes of indecomposable totally reflexive modules, or the only totally reflexive modules are free''. These are the things that have motivated the author to investigate totally reflexive modules over connected sums with \(\mathfrak{m}^3 = 0\). The author gives a criterion for rings with \(\mathfrak{m}^3 = 0\) which are obtained as connected sums of two other rings to have nontrivial totally acyclic modules. As an example, we describe what author proves in Corollary 3.7: Let \(P_1 = k[x_1,\dots,x_n]\) and \(P_2 = [y_1,\dots,y_m]\). For the given \[ R_1 = \frac{P_1}{(a_1,\dots,a_t)},~S_1 = \frac{P_2}{(b_1,\dots,b_u)} \] with \(\mathfrak{m}^3_{R_1} = 0\) and \(\mathfrak{m}^3_{S_1} = 0\), set \[ I_1 = \mathfrak{m}_{P_1} a_1 + (a_2,\dots,a_t),~I_2 = \mathfrak{m}_{P_2} b_1 + (b_2,\dots,b_u). \] Let \(R\) be the following ring: \[ \frac{P}{I_1P + I_2 P + (f_1 - g_1) + (x_i y_j | 1 \leq i \leq n, 1 \leq j \leq m)}, \] where \(P = k[x_1,\dots,x_n,y_1,\dots,y_m]\), \(I_1\) is an ideal of \(P_1 = k[x_1,\dots,x_n]\), \(I_2\) is an ideal of \(P_2 = [y_1,\dots,y_m]\), and \(f_1 \in P_1\), and \(g_1 \in P_2\). Now, let \(R_1\) and \(S_1\) be non-Gorenstein rings. Also, let \(f\) and \(g\) be part of minimal systems of generators for the defining ideals of \(R_1\), respectively \(S_1\). If \(R\) is not Gorenstein, then \(R\) has minimal totally acyclic complexes if and only if both \(R_1\) and \(S_1\) have minimal totally acyclic complexes such that the following conditions are satisfied: \[ (f)R^{b_i}_0 \subseteq \mathrm{im}(\tilde{A}_{i+1}) \text{~and~} (g)S^{b_i}_0 \subseteq \mathrm{im}(\tilde{B}_{i+1}) \qquad\forall~i. \]