ScalingStacks

4.3 More background on the Tian-Yau metric [023H]

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

4.3 More background on the Tian-Yau metric

We recall some details of the Tian-Yau metric (cf. [11, section 3] for more expositions).

In our setup, D1D_{1} is a Fano manifold in its own right with anticanonical bundle d2​L0|D1d_{2}L_{0}|_{D_{1}}, and Y=D2∩D1Y=D_{2}\cap D_{1} is an anticanonical divisor defined by some section S∈H0​(D1,d2​L0)S\in H^{0}(D_{1},d_{2}L_{0}), which in our context is the restriction of S2∈H0​(X¯,d2​L0)S_{2}\in H^{0}(\bar{X},d_{2}L_{0}) to D1D_{1}. Up to a multiplicative constant S−1S^{-1} can be viewed as a holomorphic volume form ΩD1\Omega_{D_{1}} on D1∖D2D_{1}\setminus D_{2} with a simple pole along YY, which we normalize to have residue ΩY\Omega_{Y} along YY. In local coordinates,

ΩD1≈d​log⁡ξ2∧ΩY,\Omega_{D_{1}}\approx d\log\xi_{2}\wedge\Omega_{Y},

where ξ2\xi_{2} is a local defining function of D2∩D1⊂D1D_{2}\cap D_{1}\subset D_{1}. Recall hL0⊗d2h_{L_{0}}^{\otimes d_{2}} is the Hermitian metric on L0|YL_{0}|_{Y} whose curvature form is the Calabi-Yau metric on YY in the class d2​c1​(L0)d_{2}c_{1}(L_{0}). We extend hL0h_{L_{0}} to a smooth, positively curved metric on D1D_{1}, so that

ωC​a​l′=n−1n​d​dc​(−log⁡|S|)nn−1\omega_{Cal^{\prime}}=\frac{n-1}{n}dd^{c}(-\log|S|)^{\frac{n}{n-1}}

defines a Kähler form on a neighbourhood of infinity in D1∖D2D_{1}\setminus D_{2}. Up to an approximately holomorphic diffeomorphism, the metric ωC​a​l′\omega_{Cal^{\prime}} on (D1∖D2)(D_{1}\setminus D_{2}) outside a compact set, agrees with the Calabi ansatz on the total space of the line bundle L0→YL_{0}\to Y, up to exponentially small errors. The slightly unusual exponent nn−1\frac{n}{n-1} is because dimY=n−2\dim Y=n-2. Asymptotically, the distance to a fixed basepoint in D1∖D2D_{1}\setminus D_{2} is uniformly equivalent to rD1∼(−log⁡|S|)n2​(n−1).r_{D_{1}}\sim(-\log|S|)^{\frac{n}{2(n-1)}}. We can arrange ωC​a​l′\omega_{Cal^{\prime}} to extend to an exact global Kähler metric on D1∖D2D_{1}\setminus D_{2}, agreeing with the previous formula outside a compact set, still denoted ωC​a​l′\omega_{Cal^{\prime}}.

A technical subtlety is that SS is only defined up to a multiplicative constant, and correspondingly log⁡|S|\log|S| has an additive constant ambiguity, which is fixed by the integral normalization condition

∫D1∖D2(ωC​a​l′n−1−d2n−2​∫Yc1​(L0)n−24​π​−1(n−1)2​ΩD1∧Ω¯D1)=0.\int_{D_{1}\setminus D_{2}}\left(\omega_{Cal^{\prime}}^{n-1}-\frac{d_{2}^{n-2}\int_{Y}c_{1}(L_{0})^{n-2}}{4\pi}\sqrt{-1}^{(n-1)^{2}}\Omega_{D_{1}}\wedge\overline{\Omega}_{D_{1}}\right)=0. (30)

This makes sense because the integrand has exponential decay near infinity, and morever Stokes theorem on large compact sets shows that the integral only depends on the asymptotic information of ωc​a​l′\omega_{cal^{\prime}} near infinity. The preferred normalization of log⁡|S|\log|S| resolves the additive ambiguity of x2=−log⁡|S2|x_{2}=-\log|S_{2}|; a completely analogous normalization on the Tian-Yau space D2∖D1D_{2}\setminus D_{1} fixes the additive ambiguity of x1=−log⁡|S1|x_{1}=-\log|S_{1}|.

The main existence theorem and asymptotic information of the Tian-Yau metric is summarized as follows:

Theorem 4.2.

[16][10, Prop. 2.9] There is a complete Ricci-flat Kähler metric

ωT​Y=d​dc​ϕT​Y=ωC​a​l′+d​dc​ϕT​Y,r​e​l\omega_{TY}=dd^{c}\phi_{TY}=\omega_{Cal^{\prime}}+dd^{c}\phi_{TY,rel}

solving the complex Monge-Ampère equation

ωT​Yn−1=d2n−2​∫Yc1​(L0)n−24​π​−1(n−1)2​ΩD1∧Ω¯D1,\omega_{TY}^{n-1}=\frac{d_{2}^{n-2}\int_{Y}c_{1}(L_{0})^{n-2}}{4\pi}\sqrt{-1}^{(n-1)^{2}}\Omega_{D_{1}}\wedge\overline{\Omega}_{D_{1}},

with exponential decay estimate for some constant c>0c>0 depending on D1,D2D_{1},D_{2}:

|∇ωC​a​l′kϕT​Y,r​e​l|ωC​a​l′=O⁡(e−c​rD1n−1n),rD1→+∞,k≥0.|\nabla_{\omega_{Cal^{\prime}}}^{k}\phi_{TY,rel}|_{\omega_{Cal^{\prime}}}=O(e^{-cr_{D_{1}}^{\frac{n-1}{n}}}),\quad r_{D_{1}}\to+\infty,\quad k\geq 0.

The volume growth rate of geodesic balls on the Tian-Yau space is Vol​(B⁡(rD1))∼rD12​(n−1)n\text{Vol}(B(r_{D_{1}}))\sim r_{D_{1}}^{\frac{2(n-1)}{n}}, which is less than quadratic. As such, solving the Poisson equation with potential decaying at infinity would require an integral normalization on the forcing term, which is related to (30) as the Poisson equation is the linearization of the complex Monge-Ampère equation.

Proposition 4.3.

Let ff be a smooth function on the Tian-Yau space satisfying the integral normalization ∫f​ωT​Yn−1=0\int f\omega_{TY}^{n-1}=0 and the fast decay

|∇T​Ykf|=O⁡(e−c​rD1n−1n),∀k≥0,|\nabla_{TY}^{k}f|=O(e^{-cr_{D_{1}}^{\frac{n-1}{n}}}),\quad\forall k\geq 0,

then there is a smooth function uu with ΔT​Y​U=f\Delta_{TY}U=f with fast decay

|∇T​YkU|=O⁡(e−c​rD1n−1n),∀k≥0,|\nabla_{TY}^{k}U|=O(e^{-cr_{D_{1}}^{\frac{n-1}{n}}}),\quad\forall k\geq 0,

where the decay rate c>0c>0 may be shrinked.

Remark 4.4.

The existence of solution with L∞L^{\infty} bound and L2L^{2}-gradient bound follows from [9, Thm 1.5]. The exponential type decay estimate is because the Tian-Yau space is CYL​(1n)\text{CYL}(\frac{1}{n}) in the sense of Hein [10, Def. 2.5], and the proof of [10, Prop. 2.9] works almost verbatim for the Poisson equation.

To solve the Poisson equation without the integral normalization condition on the forcing term, we need to allow solutions growing at infinity. Consider the special function (−log⁡|S|)1n−1(-\log|S|)^{\frac{1}{n-1}} defined near infinity, which we extend smoothly to a global function u0u_{0} on the Tian-Yau space D1∖D2D_{1}\setminus D_{2}.

Lemma 4.5.

The function f0=Δ​u0f_{0}=\Delta u_{0} satisfies the fast decay near infinity

|∇T​Ykf0|=O⁡(e−c​rD1n−1n)|\nabla_{TY}^{k}f_{0}|=O(e^{-cr_{D_{1}}^{\frac{n-1}{n}}})

and the total integral

12​π​(n−1)​∫D1∖D2f0​ωT​Yn−1=∫D1∖D2d​dc​u0∧ωT​Yn−2=d2n−2n−1​∫Yc1​(L0)n−2≠0.\frac{1}{2\pi(n-1)}\int_{D_{1}\setminus D_{2}}f_{0}\omega_{TY}^{n-1}=\int_{D_{1}\setminus D_{2}}dd^{c}u_{0}\wedge\omega_{TY}^{n-2}=\frac{d_{2}^{n-2}}{n-1}\int_{Y}c_{1}(L_{0})^{n-2}\neq 0.
Proof.

The exponential type decay is because the deviation between the Tian-Yau metric and the Calabi ansatz is exponentially small, and on the Calabi ansatz model (−log⁡|S|)1n−1(-\log|S|)^{\frac{1}{n-1}} is precisely harmonic. This is the intimately connected to the freedom to add a constant to −log⁡|S|-\log|S| in the Calabi ansatz, without affecting the complex Monge-Ampère measure.

To evaluate ∫f0​ωT​Yn−1\int f_{0}\omega_{TY}^{n-1}, we take a very large compact set K={log|S|≤R}K=\{\log|S|\leq R\}, and consider the R→∞R\to\infty limit. We have up to exponentially suppressed errors

∫Kd​dc​u0∧ωT​Yn−2=∫∂Kdc​u0∧ωT​Yn−2≈∫∂Kdc​u0∧ωC​a​l′n−2.\int_{K}dd^{c}u_{0}\wedge\omega_{TY}^{n-2}=\int_{\partial K}d^{c}u_{0}\wedge\omega_{TY}^{n-2}\approx\int_{\partial K}d^{c}u_{0}\wedge\omega_{Cal^{\prime}}^{n-2}.

Computing in the Calabi ansatz model, this is

−1n−1∫∂Kdclog|S|∧(ddc(−log|S|))n−2-\frac{1}{n-1}\int_{\partial K}d^{c}\log|S|\wedge(dd^{c}(-\log|S|))^{n-2}

which is

1n−1​∫Y(d​dc​(−log⁡|S|))n−2=1n−1​∫Y(d2​c1​(L0))n−2=d2n−2n−1​∫Yc1​(L0)n−2.\frac{1}{n-1}\int_{Y}(dd^{c}(-\log|S|))^{n-2}=\frac{1}{n-1}\int_{Y}(d_{2}c_{1}(L_{0}))^{n-2}=\frac{d_{2}^{n-2}}{n-1}\int_{Y}c_{1}(L_{0})^{n-2}.

Taking the R→+∞R\to+\infty gives the total integral. ∎

Corollary 4.6.

Let ff be a smooth function on the Tian-Yau space satisfying the fast decay

|∇T​Ykf|=O⁡(e−c​rD1n−1n),∀k≥0,|\nabla_{TY}^{k}f|=O(e^{-cr_{D_{1}}^{\frac{n-1}{n}}}),\quad\forall k\geq 0,

then there is a smooth function UU with ΔT​Y​U=f\Delta_{TY}U=f in the form

U=const⋅u0+U~,|∇T​YkU|=O⁡(e−c​rD1n−1n),∀k≥0,U=\text{const}\cdot u_{0}+\tilde{U},\quad|\nabla_{TY}^{k}U|=O(e^{-cr_{D_{1}}^{\frac{n-1}{n}}}),\quad\forall k\geq 0,

for some possibly shrinked c>0c>0.

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