The resolution of the bracket powers of the maximal ideal in a diagonal hypersurface ring (Q1946109)

Let \(k\) be a field and for each pair of integers \((n,N)\) consider the module \(Q = R/(x^N,y^N,z^N)\) over the ring \(R = k[x,y,z]/(x^n+y^n+z^n)\). The authors study for which \(k\) and for which \((n,N)\) the projective dimension of \(Q_N\) over \(R\) is finite. One of the main theorems of the paper is: let \(c\) be the characteristic of \(k\), then \(Q\) has finite projective dimension if and only if one of three conditions hold. Namely \(n\) divides \(N\), or \(c =2\) and \(n \leq N\), or \(c= p\) is an odd prime and there exist an odd integer \(J\) and a power \(q = p^e\), \(e \geq 1\), such that \(|Jq-N/n| < \{q/3\}\), where \(\{q/3\}\) is the integer which is closest to \(q/3\). More generally, the authors explicitly resolve \(Q\) over \(R\), and the main technique involves use of Pfaffians to get a free resolution over \(k[x,y,z]\) first. Using this technique, they are able to give an explicit presentation for the first and the second module of syzygies when the projective dimension of \(Q\) is infinite. The former has a very complicated presentation, while the latter has a remarkably uncomplicated one. The question ``Does \(Q\) have finite projective dimension?'' is closely related to a Weak Lefschetz Property problem. Let \(\overline{k}\) be the algebraic closure of \(k\), then the authors prove that \(Q\) has finite projective dimension if and only if \(\overline{k}[x,y,z]/(x^a,y^a,z^a)\) has the WLP for at least one of the integers \(a =\lceil N/n \rceil\) or \(a = \lfloor N/n \rfloor\). Finally, when \(k\) has positive characteristic \(p\) they look at applications to Frobenius powers, and in particular they ask three questions. Let \(Q_N\) denote \(R/(x^N,y^N,z^N)\), so that \(F^t(Q_N) = Q_{p^tN}\). The first question is: ``When is there an integer \(N\) such that \(Q_N\) has infinite projective dimension, but some Frobenius power of it has finite projective dimension?''. They answer it showing that this condition is equivalent to the ring \(R\) being not F-injective. The second question is: ``Is the tail of the resolution of \(Q_{p^tN}\) (up to shift) eventually a periodic function of \(t\)?''. They show that there is an integer \(t_0\) depending on \(N\) such that for \(t \geq t_0\) the answer to the question is positive. Finally, the third question is: ``Can one use socle degrees to predict that the tail of the resolution of \(Q_{p^tN}\) is a shift of the tail of the resolution of \(Q_N\)?'' They give a positive answer also in this case.