[Paper Review] Eliminate obstructions: curves on a 3-fold

Sen Yang|arXiv (Cornell University)|Nov 22, 2016

Algebraic Geometry and Number Theory6 references3 citations

TL;DR

This paper resolves an open question by TingFai Ng on deforming curves on a 3-fold by introducing a K-theoretic framework using Milnor K-theoretic cycles. It shows that obstructed first-order deformations of curves can be lifted to second order by embedding them in a nodal curve with a fixed component, leveraging Koszul complexes and deformation theory in perfect complexes, thereby generalizing Ng's divisor result to codimension-two cycles.

ABSTRACT

By using higher K-theory, we reinterpret and generalize an idea on eliminating obstructions to deforming cycles, which is known to Mark Green, Phillip Griffiths and TingFai Ng(for the divisor case). As an application, we show how to eliminate obstructions to deforming curves on a 3-fold. This answers affirmatively an open question by TingFai Ng.

Motivation & Objective

• To resolve an open question posed by TingFai Ng regarding the existence of second-order deformations for curves on a 3-fold when classical subscheme deformations are obstructed.
• To generalize Green-Griffiths' program on eliminating obstructions in cycle deformation theory beyond the divisor case.
• To provide a K-theoretic reinterpretation of Ng's method using Milnor K-theoretic cycles that detect nilpotent structures.
• To establish a framework where deformations of curves as cycles can be lifted to higher order by fixing auxiliary components.
• To demonstrate that the obstruction to deforming a curve as a cycle can be eliminated via a nodal extension with a fixed residual curve.

Proposed method

• The paper uses Milnor K-theoretic cycles to detect nilpotent structures in deformations, enabling cycle-level lifting where subscheme deformations fail.
• It constructs a nodal curve $\tilde{C} = C \cup D$ in the 3-fold $X$, where $D$ is fixed (i.e., $v' = 0$) to avoid introducing new obstructions.
• The deformation of the curve $C$ is realized as a difference of cycles: $(C, v) = (\tilde{C}, \tilde{v}) - (D, 0)$, with $\tilde{v}$ extending to second order.
• Koszul complexes are used to model the deformation of the structure sheaf of the curve; these are decomposed into components corresponding to $C$ and $D$.
• The deformation of the Koszul complex $L$ is lifted to $X_1 = X[\varepsilon]/(\varepsilon^2)$ and $X_2 = X[\varepsilon]/(\varepsilon^3)$, with $L'$ and $L''$ in $Z^M_2(D^{\mathrm{Perf}}(X_1))$ and $Z^M_2(D^{\mathrm{Perf}}(X_2))$ respectively.
• The key is showing that $\mu(Y^\prime) = (L' - \mu(Z'))$ deforms to second order, even when $\mu(Y^\prime)$ alone may not be a Milnor K-theoretic cycle.

Experimental results

Research questions

• RQ1 Can every first-order deformation of a curve on a 3-fold be extended to a second-order deformation as a cycle, even when the subscheme deformation is obstructed?
• RQ2 Is there a way to eliminate obstructions to cycle deformation by embedding the curve in a nodal curve with a fixed component?
• RQ3 Can the method used for divisors by Ng be generalized to curves in 3-folds using K-theoretic cycles?
• RQ4 Under what conditions is a first-order deformation of a curve as a cycle liftable to second order via a nodal extension?
• RQ5 Does the use of Milnor K-theoretic cycles allow for a consistent deformation theory of cycles that detects nilpotent structures?

Key findings

• The paper affirms that every first-order deformation of a curve on a 3-fold as a cycle can be lifted to second order, answering Ng’s question in the affirmative.
• When $b_1 \notin (f,g,h)$, the deformation $\mu(Y^\prime)$ is a Milnor K-theoretic cycle and deforms to second order.
• When $b_1 \in (f,g,h)$, $\mu(Y^\prime)$ is not a Milnor K-theoretic cycle, but the combination $\mu(Y^\prime) + \mu(Z^\prime) - \mu(Z)$ still deforms to second order.
• The deformation $L'$ of the Koszul complex $L$ decomposes as $\mu(Y^\prime) + \mu(Z^\prime)$, with $Z^\prime$ a first-order deformation of $Z$, and this decomposition is preserved under lifting.
• The construction ensures $((\mu(Y^\prime) + \mu(Z^\prime)) - \mu(Z)) \mid_Y = \mu(Y^\prime)$, so the deformation of $C$ is correctly induced.
• The method successfully eliminates obstructions by fixing the residual curve $D$, ensuring no new obstructions are introduced, and enables cycle-level lifting via K-theoretic cycles.