On the Cohen-Macaulayness of the conormal module of an ideal (Q2376684)

The authors contribute to the study related to the question: Let \(R\) be a regular local ring and \(I\) a perfect \(R\)-ideal that is generically a complete intersection, for which ideals \(I\) does the Cohen-Macaulayness of \( I/I^{2}\) imply that \(R/I\) is Gorenstein? A bit of history of the progress and the authors' contribution are given below in the authors' own words. ``Earlier work by several authors showed that the following ideals have the property conjectured in the question: \(\bullet\) perfect prime ideals of height \(2\) (see work of \textit{J. Herzog} [Math. Z. 163, 149--162 (1978; Zbl 0372.13014)]); \(\bullet\) licci ideals (proved by \textit{C. Huneke} and \textit{B. Ulrich} in [Commutative algebra, Proc. Microprogram, Berkeley/CA (USA) 1989, Publ., Math. Sci. Res. Inst. 15, 339--346 (1989; Zbl 0731.13008)]); \(\bullet\) squarefree monomial ideals in a polynomial ring whose square is Cohen-Macaulay over any field (proved recently by \textit{G. Rinaldo, N. Terai} and \textit{K.-I. Yoshida} [J. Commut. Algebra 3, No. 3, 405--430 (2011; Zbl 1237.13045)]). Also, work of Minh-Trung and Trung-Tuan gives a complete description of all \(2\)- or \(3\)-dimensional Stanley-Reisner ideals whose second power is Cohen-Macaulay whereas the third power is not Cohen-Macaulay and it is shown that in all these cases, the ideal is Gorenstein (independent of the base field) [\textit{Nguyen Cong Minh} and \textit{Ngo Viet Trung}, J. Algebra 322, No. 12, 4219--4227 (2009; Zbl 1206.13028), J. Algebra 328, No. 1, 77--93 (2011; Zbl 1213.13023)]. However, in general, the question is wide open. In the present paper we prove that the question has a positive answer if \(I\) is any monomial ideal (extending a result of Rinaldo-Terai-Yoshida) and for ideals defining stretched algebras, short algebras or algebras with low multiplicity (precise definitions will be given later in the paper). We use several different techniques and ad hoc methods, including careful estimates (both in the local and homogeneous settings) of the Hilbert function of the square of the ideal \(I\) by means of the Hilbert function of \(I\) . We also provide examples for which the question has a negative answer (to our best knowledge these are the first negative examples in the literature) and employ them to prove the sharpness of our main result''.