ScalingStacks

2.1. Generalized Gibbons-Hawking ansatz [05CA]

Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.

Complete original source context · Original author HTML

2.1. Generalized Gibbons-Hawking ansatz

Suppose TnT^{n}, a nn-dimensional (real) torus, acts freely on an NN-(complex) dimensional Kähler manifold MM by Hamiltonian holomorphic isometries. Then MM can be considered as a principal TnT^{n}-bundle over a real manifold of dimension 2​N−n2N-n. The generalized Gibbons-Hawking ansatz expresses the Kähler and Ricci-flat conditions as differential equations in the nn moment map coordinates and N−nN-n holomorphic coordinates on the Kähler quotient.

Let 𝔱\mathfrak{t} denote the Lie algebra of the torus Lie group TnT^{n}, and let 𝔱ℤ\mathfrak{t}_{\mathbb{Z}} be the natural integral lattice in 𝔱\mathfrak{t}. We will fix a basis in 𝔱ℤ\mathfrak{t}_{\mathbb{Z}}. This defines affine coordinates uiu_{i} on the dual space 𝔱∗≅ℝn\mathfrak{t}^{*}\cong\mathbb{R}^{n}. Let YY be either ℂN−n\mathbb{C}^{N-n} or (ℂ∗)N−n≅ℝN−n×(S1)N−n(\mathbb{C}^{*})^{N-n}\cong\mathbb{R}^{N-n}\times(S^{1})^{N-n}, with the affine complex coordinates ηp=xp+i​yp\eta_{p}=x_{p}+iy_{p}, where ypy_{p} are the phase coordinates on the torus (S1)N−n(S^{1})^{N-n} in the latter case.

Consider a principal TnT^{n}-bundle π:M→B∘\pi:M\to B^{\circ} over an open set B∘B^{\circ} in 𝔱∗×Y\mathfrak{t}^{*}\times Y, with coordinates (ui,ηp,η¯q)(u_{i},\eta_{p},\bar{\eta}_{q}). Denote by [ν][\nu] its integral Chern class as an element in H2​(B,𝔱ℤ)H^{2}(B,\mathfrak{t}_{\mathbb{Z}}).

Theorem 2.1 (cf. [PP91]).

Let Vi​jV^{ij}, respectively Wp​qW^{pq}, be real symmetric, respectively hermitian, positive definite matrices of smooth functions on B∘B^{\circ}, locally given by some potential function Φ\Phi:

(1) Vi​j=∂2Φ∂uj​∂uj,Wp​q=−4​∂2Φ∂ηp​∂η¯q,1≤i,j≤nn+1≤p,q≤N.V^{ij}=\frac{\partial^{2}\Phi}{\partial u_{j}\partial u_{j}},\quad W^{pq}=-4\frac{\partial^{2}\Phi}{\partial\eta_{p}\partial\bar{\eta}_{q}},\qquad 1\leq i,j\leq n\quad n+1\leq p,q\leq N.

Then the following 𝔱\mathfrak{t}-valued 2-form is closed:

(2) Fj=−1​(12​∂Wp​q∂uj​d​ηp∧d​η¯q+∂Vi​j∂ηp​d​ui∧d​ηp−∂Vi​j∂η¯q​d​ui∧d​η¯q).F_{j}=\sqrt{-1}\left(\frac{1}{2}\frac{\partial W^{pq}}{\partial u_{j}}d\eta_{p}\wedge d\bar{\eta}_{q}+\frac{\partial V^{ij}}{\partial\eta_{p}}du_{i}\wedge d\eta_{p}-\frac{\partial V^{ij}}{\partial\bar{\eta}_{q}}du_{i}\wedge d\bar{\eta}_{q}\right).

Suppose, in addition, that detVi​j=detWp​q\det V^{ij}=\det W^{pq} and (F1,…,Fn)(F_{1},\dots,F_{n}) is in the cohomology class 2​π​[ν]2\pi[\nu]. Then there exist a connection on the bundle M→BM\to B with associated 1-forms AiA_{i} and the curvature (F1,…,Fn)(F_{1},\dots,F_{n}) such that MM is a Kähler manifold with Ricci-flat metric given by

(3) h=(V−1)i​j​d​zi⊗d​z¯j+Wp​q​d​ηp⊗d​η¯q,h=(V^{-1})^{ij}dz_{i}\otimes d\bar{z}_{j}+W^{pq}d\eta_{p}\otimes d\bar{\eta}_{q},

where d​zj=Vi​j​d​ui+−1⋅Ajdz_{j}=V^{ij}du_{i}+\sqrt{-1}\cdot A_{j} and d​ηpd\eta_{p} form a basis of holomorphic 1-forms. The holomorphic NN-form and the Kähler form:

(4) Ω=∧j=1kdzj⋀∧p=1ldηp,ω=duj∧Aj+−12⋅Wp​qdηp∧dη¯q\Omega=\wedge_{j=1}^{k}dz_{j}\bigwedge\wedge_{p=1}^{l}d\eta_{p},\qquad\omega=du_{j}\wedge A_{j}+\frac{\sqrt{-1}}{2}\cdot W^{pq}d\eta_{p}\wedge d\bar{\eta}_{q}

are compatible in the sense that Ω∧Ω¯=c​o​n​s​t⋅ωN\Omega\wedge\bar{\Omega}=const\cdot\omega^{N}.

Proof.

First we note that the local potential description of VV and WW by (1) insures that

(5) d​Fj=−12⋅d⁡(∂Wp​q∂uj)∧d​ηp∧d​η¯q+−1⋅d(∂Vi​j∂ηp)∧dui∧dηp−−1⋅d(∂Vi​j∂η¯q)∧dui∧dη¯q=−12​(∂2Wp​q∂ui​∂uj+4​∂2Vi​j∂ηp​∂η¯q)​d​ui∧d​ηp∧d​η¯q=0.dF_{j}=\frac{\sqrt{-1}}{2}\cdot d\left(\frac{\partial W^{pq}}{\partial u_{j}}\right)\wedge d\eta_{p}\wedge d\bar{\eta}_{q}\\ +\sqrt{-1}\cdot d\left(\frac{\partial V^{ij}}{\partial\eta_{p}}\right)\wedge du_{i}\wedge d\eta_{p}-\sqrt{-1}\cdot d\left(\frac{\partial V^{ij}}{\partial\bar{\eta}_{q}}\right)\wedge du_{i}\wedge d\bar{\eta}_{q}\\ =\frac{\sqrt{-1}}{2}\left(\frac{\partial^{2}W^{pq}}{\partial u_{i}\partial u_{j}}+4\frac{\partial^{2}V^{ij}}{\partial\eta_{p}\partial\bar{\eta}_{q}}\right)du_{i}\wedge d\eta_{p}\wedge d\bar{\eta}_{q}=0.

Moreover, if θj\theta_{j} denote the coordinates on the torus fiber such that ∂/∂θj\partial/\partial\theta_{j} are the Hamiltonian vector fields, then the connection 1-forms can be written up to exact forms on B∘B^{\circ} in terms of the local potential Φ\Phi:

(6) Aj=d​θj+−1​(∂2Φ∂uj​∂ηp​d​ηp−∂2Φ∂uj​∂η¯q​d​η¯q).A_{j}=d\theta_{j}+\sqrt{-1}\left(\frac{\partial^{2}\Phi}{\partial u_{j}\partial\eta_{p}}d\eta_{p}-\frac{\partial^{2}\Phi}{\partial u_{j}\partial\bar{\eta}_{q}}d\bar{\eta}_{q}\right).

And one can see explicitly that Fj=d​AjF_{j}=dA_{j}.

The integrability of the complex structure follows from the fact that the differential ideal generated by (1,0)(1,0)-forms is closed:

(7) d⁡(d​zj)=d​Vi​j∧d​ui+−1⋅d​Aj=−∂Vi​j∂uk​d​ui∧d​uk−∂Vi​j∂ηp​d​ui∧d​ηp−∂Vi​j∂η¯q​d​ui∧d​η¯q−(12​∂Wp​q∂uj​d​ηp∧d​η¯q+∂Vi​j∂ηp​d​ui∧d​ηp−∂Vi​j∂η¯q​d​ui∧d​η¯q)=(12​∂Wp​q∂uj​d​η¯q−2​∂Vi​j∂ηp​d​ui)∧d​ηp,d(dz_{j})=dV^{ij}\wedge du_{i}+\sqrt{-1}\cdot dA_{j}\\ =-\frac{\partial V^{ij}}{\partial u_{k}}du_{i}\wedge du_{k}-\frac{\partial V^{ij}}{\partial\eta_{p}}du_{i}\wedge d\eta_{p}-\frac{\partial V^{ij}}{\partial\bar{\eta}_{q}}du_{i}\wedge d\bar{\eta}_{q}\\ -\left(\frac{1}{2}\frac{\partial W^{pq}}{\partial u_{j}}d\eta_{p}\wedge d\bar{\eta}_{q}+\frac{\partial V^{ij}}{\partial\eta_{p}}du_{i}\wedge d\eta_{p}-\frac{\partial V^{ij}}{\partial\bar{\eta}_{q}}du_{i}\wedge d\bar{\eta}_{q}\right)\\ =\left(\frac{1}{2}\frac{\partial W^{pq}}{\partial u_{j}}d\bar{\eta}_{q}-2\frac{\partial V^{ij}}{\partial\eta_{p}}du_{i}\right)\wedge d\eta_{p},

where we have only used ∂Vi​j∂uk=∂Vk​j∂ui\frac{\partial V^{ij}}{\partial u_{k}}=\frac{\partial V^{kj}}{\partial u_{i}}.

It is equally easy to verify the Kähler condition:

(8) d​ω=−d​uj∧d​Aj+−12⋅d​Wp​q∧d​ηp∧d​η¯q=−−1​(12​∂Wp​q∂uj​d​uj∧d​ηp∧d​η¯q)+−12​∂Wp​q∂uj​d​uj∧d​ηp∧d​η¯q=0.d\omega=-du_{j}\wedge dA_{j}+\frac{\sqrt{-1}}{2}\cdot dW^{pq}\wedge d\eta_{p}\wedge d\bar{\eta}_{q}\\ =-\sqrt{-1}\left(\frac{1}{2}\frac{\partial W^{pq}}{\partial u_{j}}du_{j}\wedge d\eta_{p}\wedge d\bar{\eta}_{q}\right)+\frac{\sqrt{-1}}{2}\frac{\partial W^{pq}}{\partial u_{j}}du_{j}\wedge d\eta_{p}\wedge d\bar{\eta}_{q}=0.

Finally, the Ricci-flatness is manifest since det(h)=detV−1⋅detW=1\det(h)=\det V^{-1}\cdot\det W=1 in the complex coordinates d​zj,d​ηpdz_{j},d\eta_{p}. ∎

We are interested in applying the Gibbons-Hawking ansatz to description of the metrics on the toric Calabi-Yau hypersurfaces. Given a torus fibration of such hypersurface near the large complex structure point one may approximate the true Calabi-Yau metric by non-compact solutions of GH equations over different regions of the base. Away from the discriminant locus a semi-flat metric gives a good approximation. We try to argue that as the tori shrink to zero size the metric behavior near singular fibers can also be approximated by certain yy-periodic solutions of GH ansatz. This local description is the main subject of the paper.

Original mathematics by the credited authors. Source-backed reader collection; mathematical self-containment is not assessed.