[Paper Review] Obstructions to the deformations of curves to other hypersurfaces

Bin Wang|arXiv (Cornell University)|Oct 2, 2011

Advanced Numerical Analysis Techniques5 references3 citations

TL;DR

This paper establishes necessary obstructions for deforming a smooth curve $C_0$ on a smooth hypersurface $f_0 \subset \mathbb{P}^n$ to other hypersurfaces in first-order deformations. It proves that for rational curves, the vanishing of $H^1(N_{C_0}f_0(1))$ is a necessary condition for such deformations, offering a new approach to Clemens' conjecture on rational curves in quintic threefolds and deriving numerical bounds on curve invariants under weaker deformation assumptions than previous works.

ABSTRACT

We are interested in obstructions to the FIRST order deformation of a pair of a smooth hypersurface $f_0$ and a smooth curve $C_0$ contained in $f_0$. In the first half of the paper, we give necessary conditions for the pair to deform in the first order. In particular, for a rational curve $C_0$, this necessary condition is $$H^1(N_{C_0}f_0(1))=0.$$ In the second half, we apply the necessary conditions from the first half of the paper to study the geometry of smooth curves in hypersurfaces. The main application is for the case where $C_0$ is a rational curve.

Motivation & Objective

• To identify necessary conditions for the first-order deformation of a pair $(C_0 \subset f_0)$, where $C_0$ is a smooth curve on a smooth hypersurface $f_0 \subset \mathbb{P}^n$, to other hypersurfaces.
• To address the gap in understanding whether $H^1(N_{C_0}f_0) = 0$ is not only sufficient but also necessary for such deformations, particularly for rational curves.
• To apply these obstructions to the geometry of curves in hypersurfaces, especially in the context of Clemens' conjecture on rational curves in quintic threefolds.
• To derive numerical bounds on the geometric genus and degree of curves in hypersurfaces under weaker deformation assumptions than those in prior works.

Proposed method

• Formalizing the first-order deformation of the pair $(C_0 \subset f_0)$ via the map $P_S^s: H^0(\bar{c}_0^*TX_S) \to T_{f_0}S$, where $S$ is a parameter space of hypersurfaces.
• Analyzing the surjectivity of $P_A^s$ and $P_E^s$ for two specific parameter spaces: $S = A$ (a linear family of hypersurfaces) and $S = E = \mathbb{P}(H^0(\mathcal{O}_{\mathbb{P}^n}(h)))$.
• Using Riemann-Roch and Serre duality to compute dimensions of cohomology groups $h^i(c_0^*Tf_0(1))$, particularly focusing on $h^1$.
• Applying the condition $H^1(N_{C_0}f_0(1)) = 0$ as a necessary obstruction for rational curves to deform to other hypersurfaces.
• Deriving numerical inequalities involving genus $g$, degree $d$, and hypersurface degree $h$ via cohomological vanishing and dimension counting.
• Leveraging $GL(n+1)$-symmetry on the universal hypersurface $X_E$ to construct global sections of $Tf_0(1)|_{C_0}$, proving lower bounds on $h^0(N_{C_0}f_0(1))$.

Experimental results

Research questions

• RQ1 What are the necessary cohomological obstructions for a smooth curve $C_0$ on a smooth hypersurface $f_0 \subset \mathbb{P}^n$ to deform to other hypersurfaces in the first-order setting?
• RQ2 Is the vanishing of $H^1(N_{C_0}f_0(1))$ a necessary condition for the first-order deformation of a rational curve $C_0$ to other hypersurfaces?
• RQ3 How do the numerical bounds on the geometric genus $g$ and degree $d$ of curves in hypersurfaces change when the deformation assumption is weakened to first-order only, rather than full deformation to generic hypersurfaces?
• RQ4 Can the results provide new insight into Clemens' conjecture on the finiteness of rational curves in a general quintic threefold?
• RQ5 What is the role of the surjectivity of the map $P_S^s$ in constraining the geometry of the curve-hypersurface pair?

Key findings

• For a rational curve $C_0$ on a smooth hypersurface $f_0 \subset \mathbb{P}^n$, the condition $H^1(N_{C_0}f_0(1)) = 0$ is a necessary obstruction for the first-order deformation of $C_0$ to other hypersurfaces.
• The paper proves that if $P_A^s$ is surjective and $C_0$ is rational, then $H^1(N_{C_0}f_0(1)) = 0$ is required for such deformations, providing a new necessary condition toward Clemens' conjecture.
• Under the assumption that $P_A^s$ is surjective, the paper derives the inequality $g(h - n + 1) \geq (h - 2n)d - n + 1$ or $g \geq \frac{d}{2} + 1$ for the geometric genus $g$ and degree $d$ of the curve.
• The paper shows that $h^0(N_{C_0}f_0(1)) \geq n - 2$, and thus $h^0(N_{C_0}f_0) - (n - 4) \geq 2$, under the surjectivity of $P_E^s$, which supports the non-vanishing of certain cohomology groups.
• The derived bound $g(h - n + 1) \geq (h - 2n)d - n + 1$ is weaker than Clemens' sharper bound $g \geq \frac{1}{2}(h - 2n + 1)d + 1$ when full deformation is assumed, but is valid under the strictly weaker first-order deformation assumption.
• The paper establishes that the condition $H^1(N_{C_0}f_0(1)) = 0$ is necessary for rational curves to deform to other hypersurfaces, even though it is not yet proven sufficient, offering a new pathway toward resolving Clemens' conjecture.