ProofComplete

Moduli Spaces - Key Proof

We outline the proof of the Bogomolov-Tian-Todorov theorem, which establishes the unobstructedness of complex structure deformations for Calabi-Yau manifolds. This fundamental result distinguishes Calabi-Yau geometry from general algebraic geometry.

ProofProof of Bogomolov-Tian-Todorov Theorem

Let XX be a Calabi-Yau nn-fold with KX≅OXK_X \cong \mathcal{O}_X.

Step 1: Deformation theory setup

Complex structure deformations are governed by the Kodaira-Spencer-Kuranishi theory. Infinitesimal deformations lie in H1(X,TX)H^1(X,T_X), and obstructions to extending them lie in H2(X,TX)H^2(X,T_X). We must show H2(X,TX)=0H^2(X,T_X) = 0.

Step 2: Serre duality for tangent sheaf

By Serre duality on a Calabi-Yau manifold: H2(X,TX)βˆ—β‰…Hnβˆ’2(X,TXβˆ—βŠ—KX)=Hnβˆ’2(X,Ξ©X1βŠ—KX)H^2(X,T_X)^* \cong H^{n-2}(X, T_X^* \otimes K_X) = H^{n-2}(X, \Omega_X^1 \otimes K_X)

Since KXβ‰…OXK_X \cong \mathcal{O}_X, this equals: Hnβˆ’2(X,Ξ©X1)H^{n-2}(X, \Omega_X^1)

Step 3: Hodge decomposition

On a compact KΓ€hler manifold, Hnβˆ’2(X,Ξ©X1)H^{n-2}(X, \Omega_X^1) embeds into the Dolbeault cohomology: Hnβˆ’2(X,Ξ©X1)βŠ‚Hnβˆ’1,1(X)βŠ‚Hnβˆ’1(X,C)H^{n-2}(X, \Omega_X^1) \subset H^{n-1,1}(X) \subset H^{n-1}(X,\mathbb{C})

We analyze this using the Hodge decomposition: Hnβˆ’1(X,C)=Hnβˆ’1,0βŠ•Hnβˆ’2,1βŠ•β‹―βŠ•H0,nβˆ’1H^{n-1}(X,\mathbb{C}) = H^{n-1,0} \oplus H^{n-2,1} \oplus \cdots \oplus H^{0,n-1}

Step 4: Vanishing from Calabi-Yau condition

For a Calabi-Yau manifold, Hnβˆ’1,0(X)=0H^{n-1,0}(X) = 0 unless n=1n=1. This follows from the fact that Hnβˆ’1,0(X)=H0(X,KXβŠ—Ξ©Xnβˆ’1)H^{n-1,0}(X) = H^0(X, K_X \otimes \Omega_X^{n-1}) and KXβ‰…OXK_X \cong \mathcal{O}_X, giving: Hnβˆ’1,0(X)=H0(X,Ξ©Xnβˆ’1)H^{n-1,0}(X) = H^0(X, \Omega_X^{n-1})

A non-zero element would be a holomorphic (nβˆ’1)(n-1)-form, but on a Calabi-Yau manifold with hp,0=0h^{p,0} = 0 for 0<p<n0 < p < n, this vanishes.

Actually, we need to show Hnβˆ’2(X,Ξ©X1)=0H^{n-2}(X, \Omega_X^1) = 0. Consider the exact sequence: 0β†’OXβ†’Ξ©X1β†’Ξ©X[1]β†’00 \to \mathcal{O}_X \to \Omega_X^1 \to \Omega_X^{[1]} \to 0

where Ξ©X[1]=Ξ©X1/OX\Omega_X^{[1]} = \Omega_X^1/\mathcal{O}_X fits into: Hnβˆ’2(X,OX)β†’Hnβˆ’2(X,Ξ©X1)β†’Hnβˆ’2(X,Ξ©X[1])H^{n-2}(X, \mathcal{O}_X) \to H^{n-2}(X, \Omega_X^1) \to H^{n-2}(X, \Omega_X^{[1]})

Step 5: Explicit vanishing

For a Calabi-Yau manifold of dimension nβ‰₯3n \geq 3, we have Hnβˆ’2(X,OX)=Hnβˆ’2,0(X)=0H^{n-2}(X, \mathcal{O}_X) = H^{n-2,0}(X) = 0 since hp,0=0h^{p,0} = 0 for 0<p<n0 < p < n.

The key observation is that: Hnβˆ’2(X,Ξ©X1)βŠ‚Hnβˆ’1,1(X)∩Im(c1:Hnβˆ’2(X,Ξ©X1)β†’Hnβˆ’1(X,C))H^{n-2}(X, \Omega_X^1) \subset H^{n-1,1}(X) \cap \text{Im}(c_1: H^{n-2}(X,\Omega_X^1) \to H^{n-1}(X,\mathbb{C}))

But the Calabi-Yau condition c1(X)=0c_1(X) = 0 implies constraints. The detailed argument uses the trace map: tr:Ξ©X1β†’OX\text{tr}: \Omega_X^1 \to \mathcal{O}_X

and shows that elements of Hnβˆ’2(X,Ξ©X1)H^{n-2}(X,\Omega_X^1) must vanish.

Step 6: Conclusion

Having established H2(X,TX)=0H^2(X,T_X) = 0, Kuranishi's theorem guarantees that the moduli space Mc\mathcal{M}_c is smooth at [X][X] with tangent space H1(X,TX)β‰…Hnβˆ’1,1(X)H^1(X,T_X) \cong H^{n-1,1}(X) of dimension hnβˆ’1,1h^{n-1,1}.

β– 
Remark

The proof crucially uses the Calabi-Yau condition in two ways: the triviality of KXK_X for Serre duality, and the vanishing of Hp,0H^{p,0} for 0<p<n0 < p < n. For non-Calabi-Yau varieties, obstructions generally exist, making moduli spaces singular.

This unobstructedness result makes Calabi-Yau moduli spaces particularly tractable and is fundamental to constructing mirror pairs and studying their geometry.