Complete Calabi-Yau metrics from smoothing Calabi-Yau complete intersections (Q6190402)

The author extends the work of \textit{G. Székelyhidi} [Duke Math. J. 168, No. 14, 2651--2700 (2019; Zbl 1432.32034)] to construct complete Calabi-Yau metrics on non-compact manifolds which are smoothings of an initial complete intersection Calabi-Yau cone \(V_0\). The constructed Calabi-Yau manifold has tangent cone at infinity given by \(\mathbb{C} \times V_0\). This construction produces Calabi-Yau structures on new topological spaces which are not \(\mathbb{C}^*\)-equivariant and admit fibers having different complex structures including singularities. Let \(V_0 = \{f_1(x) = \cdots = f_k(x) = 0\} \subset \mathbb{C}^{n+k}\) be a conical complete intersection of dimension \(n\) with an isolated singularity at 0. Assume that \(f_1,\ldots,f_k\) are quasi-homogeneous under some action \(\xi = (w_1,\ldots,w_{n+k})\) where \(f_i\) has degree \(d_i\) and assume that \(1 < d_1 \leq \cdots \leq d_k\): \[ t\cdot x = (t^{w_1}x_1,\ldots,t^{w_{n+k}}x_{n+k}),\quad f_i(t\cdot x) = t^{d_i}f_i(x). \] Consider a smoothing of \(V_0\) by choosing some point \(p = (p_1,\ldots,p_k)\) defined as \[ \mathcal{X}= \{f_1(x) - zp_1 = \cdots = f_k(x) - zp_k = 0\} \] such that the total space is smooth and at least one fiber is smooth, which can be rescaled such that \(z=1\). We can label these fibers for any \(t \in\mathbb{C}\) as \[ V_t =\mathcal{X}\cap \{z = t\} \cong \{f_1(x) - tp_1 = \cdots = f_k(x) - tp_k =0\}. \] The singular fibers form an algebraic subvariety of \(\mathbb{C}\), so there are only finitely many of them, and they are bounded in the \(z\) direction. We define the index \(\ell\) to be the maximal index such that \(d_\ell = d_1\), that is, \(1 < d_1 = \cdots = d_\ell < d_{\ell + 1} \leq \cdots \leq d_k\). The main difference between this construction and that of \textit{G. Székelyhidi} [Duke Math. J. 168, No. 14, 2651--2700 (2019; Zbl 1432.32034)] is that when \(\ell < k\), the rescaling action does not take the fibers above \(V_{z_0}\) to a rescaling \(V_{tz_0}\), that is, it does not preserve \(\mathcal{X}\) as considered in \(\mathbb{C}^{n+k+1}\). In \(\mathcal{X}\) the fibers have potentially different complex structures including singularities, unlike in [\textit{G. Székelyhidi} [Duke Math. J. 168, No. 14, 2651--2700 (2019; Zbl 1432.32034)] or if \(\ell = k\) where all the fibers are the same. Here, the author allows \(\ell = k\), and this would reduce all the computations and geometry to the case of \textit{G. Székelyhidi} [Duke Math. J. 168, No. 14, 2651--2700 (2019; Zbl 1432.32034)], as the rescaling action would be \(\mathbb{C}^*\)-equivariant and therefore preserve \(\mathcal{X}\), taking fibers to rescaled fibers appropriately. Since the scaling action does not preserve the fiber structure, we need to construct a separate Calabi-Yau metric on each fiber that vary smoothly in the \(z\)-direction. The metrics on the fibers will vary smoothly because the computations only involve working near infinity, outside the algebraic set of singular fibers and therefore all the fibers are smooth. Another difference from \textit{G. Székelyhidi} [Duke Math. J. 168, No. 14, 2651--2700 (2019; Zbl 1432.32034)] is that the limiting geometry of the fibers differs from any of the fibers in a region near infinity. The scaling action \[ G_t(z,x_1,\ldots,x_{n+k}) = (tz,t^{w_1}x_1,\ldots,t^{w_{n+k}}x_{n+k}) \] applied to \(\mathcal{X}\) is expressed as \[ G^{-1}_t(\mathcal{X}) = \{f_1(x) - t^{1-d_1}zp_1 = \cdots = f_k(x) - t^{1-d_k}zp_k = 0\}.\tag{1} \] As \(t \to \infty\) this approaches \(X_0 = V_0 \times\mathbb{C}\), its tangent cone at infinity. The first \(\ell\) coordinates scale the most slowly, and under specific rescalings used in the analysis in later sections, the model geometry becomes the fiber \[ V_{p'} = \{f_1(x) -p_1= \cdots = f_{\ell}(x) - p_\ell = f_{\ell+1}(x) = \cdots = f_k(x) = 0\} \] over the point \(p' = (p_1,\ldots,p_{\ell},0,\ldots,0)\), the projection onto the first \(\ell\) coordinates, in a certain region near infinity. We therefore require this fiber to also be smooth. Not all complete intersections admit such a smooth fiber, or even perturbations that admit a smooth fiber of this form. For example, for \(f_1 = z_1^2 + z_2z_3 + z_4^3, f_2 = z_1z_4^2 + z_2^2 + z_3^2z_4 + z_5^3\) with weights \(\xi = (27,63/2, 45/2,18,21)\) and of degrees \(d_1 = 54\) and \(d_2 = 63\), no fiber \(V_{(t_1,0)} = \{f_1 - t_1 = f_2 = 0\}\subset\mathbb{C}^5\) is smooth and no perturbation of the polynomials fixes this singularity either by smoothing it, or perturbing it off the vanishing locus. In the case when no smooth fiber exists, it is possible the result can still hold as the singularities will be isolated and Gorenstein. In these circumstances, it may still be possible to invert the Laplacian and create a singular Calabi-Yau metric with which the methods presented here could be utilized, or model Calabi-Yau metrics around the singularities could be glued to asymptotically Calabi-Yau metrics to create a global singular Calabi-Yau metric on the fiber. This is a question for future study. Given a conical, singular, Calabi-Yau metric \(\omega_{X_0}\) on \(X_0\) whose homothetic transformations are the maps \(G_t\) given above, we produce a Calabi-Yau metric on \(\mathcal{X}\) that has tangent cone at infinity given by \((X_0,\omega_{X_0})\). The construction and methodology follow similarly the work of \textit{G. Székelyhidi} [Duke Math. J. 168, No. 14, 2651--2700 (2019; Zbl 1432.32034)] with care taken as the fibers are no longer all biholomorphic rescalings of the same fiber, nor is the limiting fiber geometry as discussed. The strategy is to construct model Calabi-Yau metrics near infinity in the \(z\)-direction and in the cone direction and glue them together. This asymptotical Calabi-Yau metric is then perturbed to be Ricci-flat outside a compact set, which allows to apply the result of \textit{H.-J. Hein}'s PhD thesis [On gravitational instantons. Princeton: Princeton University (PhD Thesis) 2010] to further perturb it to a globally Ricci-flat metric, giving the main theorem. Theorem. If \(V_0\) admits a Calabi-Yau cone metric \(\omega_0\), then there exists a complete Calabi-Yau metric on \(\mathcal{X}\) with tangent cone at infinity given by \((\mathbb{C} \times V_0,\sqrt{-1}\partial \overline{\partial} (|z|^2) + \omega_0)\).