Examples of smooth compact toric varieties that are not quasitoric manifolds (Q470713): Difference between revisions

An \(n\)-dimensional normal variety \(X\) over \(\mathbb C\) on which \((\mathbb C^*)^n\) algebraically such that \(X\) contains a dense open orbit \(U\) which is equivariantly isomorphic to \((\mathbb{C}^*)^n\) where the group \((\mathbb C^*)^n\) acts on itself by translation is called a toric variety. A smooth complete toric variety is called a toric manifold. It is known that there are toric manifolds which are not projective manifolds when \(n\geq 3\), although when \(n=2\) any toric manifold is projective. Generalizing the notion of a projective toric manifold, \textit{M. W. Davis} and \textit{T. Januszkiewicz} [Duke Math. J. 62, No. 2, 417--451 (1991; Zbl 0733.52006)] introduced the notion of what they called toric manifolds, which are now referred to as quasitoric manifolds. A real \(2n\)-dimensional smooth compact manifold \(M\) which is acted on by the \(n\)-dimensional (compact) torus \(T=(\mathbb S^1)^n\) is a quasitoric manifold if the action is `locally standard' and if the quotient space \(M/T\) is a simple polytope. Every projective toric manifold is a quasitoric manifold and it is known that there are quasitoric manifold which are not projective toric manifolds. The restricted action of \(T\subset (\mathbb{C}^*)^n\) on a toric manifold \(X\) satisfies the local standardness condition. The author shows, in Theorem I, that there are infinitely many toric manifolds \(X\) of dimension \(n\) (over \(\mathbb C\)) such that \(X/T\) is not a simple polytope, for every \(n\geq 4\). Such toric manifolds cannot be quasitoric. This answers a question raised by \textit{V. M. Buchstaber} and \textit{T. E. Panov} in their book [Torus actions and their applications in topology and combinatorics. Providence, RI: American Mathematical Society (AMS) (2002; Zbl 1012.52021)]. From the introduction: ``In a 2003 preprint [``Some examples in toric geometry'', \url{arXiv:math/0306029}] \textit{Y. Civan} claimed the existence of a toric manifold as in Theorem I, but it has been recognized that his explanation of non-polytopality was insufficient. Our construction is based on ideas found in (his preprint) but is more explicit.'' Recall that a simplicial sphere (i.e., a simplicial complex homeomorphic to a sphere) is called \textit{polytopal} if it is (simplicially isomorphic to) the boundary complex of a simplicial polytope. The dual (or polar) of a simplicial polytope is a simple polytope and conversely. An example of a simplicial sphere which is not polytopal is the Barnette sphere \(\mathcal B\), which is \(3\)-dimensional---the smallest possible dimension in which a non-polytopal sphere can exist. It has been shown by Ishida, Fukukawa, and Masuda [\textit{H. Ishida} et al., Mosc. Math. J. 13, No. 1, 57--98 (2013; Zbl 1302.53091)] that the Barnette sphere cannot arise as the underlying simplicial sphere of a regular fan. The author first finds a simplicial fan in dimension \(4\) inducing the Barnette sphere \(\mathcal B\) and finds a suitable subdivision of the fan which is regular (and corresponds to a subdivision \(\mathcal B'\) of the Barnette sphere). The corresponding complete toric variety \(X\) is smooth but the quotient \(X/T=P\) cannot be a simple polytope. (If \(P\) were a simple polytope, its dual \(Q:=P^\circ\) would be simplicial and so \(\mathcal B'=\partial Q\) would then be polytopal.) To obtain infinitely infinitely many examples, one performs repeated blow ups. To obtain examples in higher dimensions, one takes repeated suspensions.