Technical Summary: Finite-Time Singularities of Lagrangian Mean Curvature Flow with Quantitatively Precise Dynamics

1. Problem Statement and Context

The central problem addressed is the formation of finite-time singularities in the Lagrangian Mean Curvature Flow (LMCF) within the ambient space $\mathbb{C}^n$. While the Thomas-Yau conjectures suggest that stable Lagrangian submanifolds should converge to special Lagrangians (SLs) under LMCF, it is well-established that finite-time singularities are generally unavoidable.

Prior to this work, the understanding of these singularities relied heavily on qualitative arguments using maximum principles, blow-up arguments, and compactness. A significant gap existed in the quantitative description of singularity formation. Specifically, while it was known that under the zero-Maslov assumption, finite-time singularities must be of Type II (where the curvature blows up faster than $(T-t)^{-1/2}$), the precise blow-up rates and the detailed dynamics leading to these singularities were unknown. It was unclear whether the mean curvature could remain uniformly bounded or what specific curvature blow-up rates might occur.

This paper aims to construct explicit solutions to the LMCF that develop finite-time singularities with quantitatively precise dynamics, providing the first rigorous description of the curvature blow-up rates and the geometric structure of the flow near the singularity.

2. Methodology

The authors employ a sophisticated analytical framework combining modulation analysis, spectral theory, and topological arguments to construct these solutions. The methodology is distinct from previous approaches that relied solely on maximum principles.

2.1. Reduction to a Fully Nonlinear Scalar PDE

By restricting the flow to cohomogeneity-one Lagrangian submanifolds invariant under a compact, connected Lie group $G \leq SU(n)$ (where generic orbits are $(n-1)$-dimensional), the LMCF reduces to a fully nonlinear, scalar-valued parabolic PDE for a profile function $h(s, \tau)$. This reduction transforms a quasilinear system into a scalar equation, allowing for the application of maximum principle techniques, though the full nonlinearity introduces significant analytic challenges near the singularity.

2.2. Modulation Analysis and Spectral Decomposition

The core of the construction involves analyzing the flow near a shrinking family of special Lagrangian desingularizations ($aL$). The profile function is decomposed as:
$$h = f_a + \partial_s u$$
where $f_a$ is the profile of a special Lagrangian cone desingularization scaled by a time-dependent parameter $a(\tau)$. The perturbation $u$ is further decomposed into a finite-dimensional component $\phi$ (spanned by eigenfunctions of the linearized operator) and a remainder term $v$:
$$u = \sum_{k=1}^K b_k(\tau)(\phi_{k,a} - \phi_{0,a}) + v$$
Here, $\phi_{k,a}$ are eigenfunctions of the linearized operator $H_a$ constructed in the companion paper [SS26]. The parameters $a(\tau)$ and $b_k(\tau)$ are modulation parameters controlled by a system of ordinary differential equations (modulation equations).

2.3. Overcoming Non-Integrability

A major technical hurdle is that the tangent flow (a transverse pair of SL cones) admits non-integrable deformations. Unlike previous constructions (e.g., Velázquez's work on mean curvature flow) where $L^2$ methods were available due to integrable deformations, the decay rate of the desingularization profile here ($O(|x|^{1-n})$) places it outside $L^2_{loc}$.
To address this, the authors replace standard $L^2$ spaces with a collection of weighted spaces $\{L^2(L_a)\}_{a>0}$ and utilize perturbative spectral theory developed in [SS26]. This allows for the control of the remainder term $v$ despite the lack of standard $L^2$ integrability.

2.4. The Ważewski Box Argument

The existence of a solution with the desired asymptotic behavior is proven using a Ważewski topological argument (often called a "box argument").

1. A time-dependent set (the "box") $B(\tau)$ is defined in the space of parameters $(u, a)$.
2. The authors derive a series of integral estimates (Section 3) and weighted Hölder estimates (Section 4) to show that if a solution exits this box, it must do so through a specific $(K-1)$-dimensional sphere on the boundary.
3. By assuming for contradiction that all solutions exit the box, one constructs a continuous retraction of a $K$-dimensional ball to its boundary, which is topologically impossible.
4. This contradiction implies the existence of at least one solution that remains in the box for all time, thereby exhibiting the desired asymptotic behavior.

2.5. Regularity and Estimates

The proof relies on establishing:

• Modulation Equations: Controlling the evolution of $a(\tau)$ and $b_k(\tau)$.
• Energy Estimates: $L^2_a$ and $H^k_a$ estimates for the remainder $v$.
• Weighted Hölder Estimates: Establishing $C^{2,\alpha}$ bounds in three distinct regions:
  • Parabolic Region: $|s| \sim 1$.
  • Outer Region: $|s| \gtrsim 1$.
  • Inner Region: $|s| \lesssim 1$ (near the singularity).
    Crucially, the inner region estimates utilize a blow-up argument combined with Liouville-type theorems for ancient solutions of the heat equation on the tangent cones and the desingularized SLs.

3. Key Results

The main theorem (Theorem 1.1) establishes the existence of a smooth, non-compact, $G$-invariant Lagrangian mean curvature flow $L_K(t)$ in $\mathbb{C}^n$ for specific dimensions and integers $K$:

• Conditions: $n \geq 4, K \geq 2$; or $n=3, K \in \{2, 3, 4\}$.
• Singularity Formation: The flow develops a finite-time singularity at the origin at time $T < \infty$.
• Curvature Blow-up Rate: The second fundamental form blows up with an explicit rate:
  $$\sup_{L_K(t)} |A_{L_K(t)}| \sim (T-t)^{-K/2}$$
  This provides the first quantitative confirmation of specific Type II blow-up rates for fully nonlinear parabolic PDEs arising from LMCF.
• Geometric Structure:
  • The tangent flow at the singularity is a unique transverse pair of $G$-invariant special Lagrangian cones $C_0 \cup C_1$.
  • The Type II blow-up limit is a smooth, $G$-invariant special Lagrangian desingularization $aL$.
• Initial Data: The initial Lagrangian angle oscillation can be taken to be arbitrarily small, meaning the initial data is arbitrarily close to being special Lagrangian.

4. Significance and Claims

The paper claims to be the first construction of Type II blow-up for a fully nonlinear parabolic PDE using either the Hadžić-Raphaël spectral strategy or the Dávila-del Pino-Wei gluing strategy.

• Quantitative Precision: It remedies the lack of quantitatively precise descriptions in prior singularity formation literature, providing explicit curvature blow-up rates and detailed asymptotic profiles.
• Analytic Innovation: The work demonstrates how to handle non-integrable deformations in the context of fully nonlinear PDEs by replacing $L^2$ methods with spectral analysis on a family of weighted spaces.
• Completion of a Research Program: The result is presented as a completion of a problem posed by Joyce [Joy15, Problem 3.12(a)] regarding the dynamics of special Lagrangian cones and their desingularizations in dimensions $n \geq 3$.
• Methodological Extension: It extends techniques previously applied to semilinear heat equations, the Keller-Segel system, and Yang-Mills heat flow to the more challenging setting of fully nonlinear geometric flows.

The authors explicitly note that their construction is based on a "modulation analysis around a shrinking family of cohomogeneity-one special Lagrangian desingularizations" and relies heavily on the perturbative spectral theory developed in their companion paper [SS26]. They do not claim to solve the general Thomas-Yau conjecture but rather provide a rigorous, quantitative model of how singularities can form and be resolved (in the sense of the blow-up limit) within this specific symmetric setting.