Tetrahedral curves via graphs and Alexander duality (Q2457281)

Let \(R = k[a,b,c,d]\). A tetrahedral curve is a one-dimensional subscheme of \(\mathbb P^3\) defined by an ideal of the form \((a,b)^{p_1} \cap \dots \cap (c,d)^{p_6}\). The name, introduced by the reviewer and \textit{U. Nagel} [Int.\ Math.\ Res.\ Not.\ 2005, No. 15, 899--939 (2005; Zbl 1093.14044)], comes from the fact that one can view the support of such curves as being the union of 6 lines forming the edges of a tetrahedron. In the case where \(p_2 = p_5 = 0\), these curves were studied by Schwartau in his 1982 Ph.D.\ thesis. Schwartau's motivating problem was to determine when these curves are arithmetically Cohen-Macaulay (ACM), and he gave an explicit answer. This motivated the work of Nagel and the reviewer mentioned above, to examine when the more general curves (i.e.\ \(p_2\) and \(p_5\) are not necessarily 0) are ACM, and more generally to examine the even liaison class of these curves. An algorithmic answer was given in the former case, using liaison theory. The current paper instead gives an explicit answer to this question, more in the lines of Schwartau's result. The methods avoid liaison theory entirely, relying instead on combinatorial methods. The basic idea is to view the ideal \(I\) of a tetrahedral curve as a monomial ideal, polarize it (to obtain a squarefree monomial ideal in more variables), take the Alexander dual to obtain a new squarefree monomial ideal, view this new ideal as the edge ideal of a graph \(G\), and then make a careful study of the combinatorial properties of \(G\) and its complementary graph \(\bar G\). Finally, the problem is reduced to a numerical question about the partitioning of two positive integers, subject to four inequalities. The first part of this paper is a very clearly written exposition of the tools that will be used to derive the main results.