Almost global solutions of 1D nonlinear Klein-Gordon equations with small weakly decaying initial data

Technical Summary: Almost Global Solutions of 1D Nonlinear Klein-Gordon Equations with Small Weakly Decaying Initial Data

Problem Statement
The paper investigates the Cauchy problem for the one-dimensional semilinear Klein-Gordon equation with quadratic nonlinearity:
$$\begin{cases} \partial_t^2 u - \partial_x^2 u + u = F(u, \partial u), & (t, x) \in [0, \infty) \times \mathbb{R}, \\ (u, \partial_t u)(0, x) = (u_0, u_1)(x), \end{cases}$$
where $F$ is a smooth quadratic function of $(u, \partial u)$. The primary focus is on the lifespan of classical solutions when the initial data $(u_0, u_1)$ are small but possess only weak decay at spatial infinity.

Existing literature establishes that if initial data decay sufficiently fast (e.g., in weighted Sobolev spaces with high-order weights), global or almost global solutions exist. Specifically, for 1D equations without the null condition, the lifespan $T_\varepsilon$ is known to be bounded by $e^{C/\varepsilon^2}$ for rapidly decaying data. However, standard vector field methods, which typically require high-order spatial weights (e.g., $x \partial_x$) to close energy estimates, fail when the initial data decay slowly (e.g., with weights like $\langle x \rangle^{1/2}$), as the resulting Sobolev norms may become infinite. This paper addresses the gap between rapidly decaying data and the specific case of weakly decaying data in 1D.

Methodology
The authors employ a combination of Fourier analysis, normal form transformations, and a refined dispersive estimate approach tailored to weak decay. The methodology proceeds as follows:

1. Normal Form Transformations:

   • First Transformation: The quadratic nonlinearity is transformed into a cubic nonlinearity. This is achieved by introducing the profile $U = (\partial_t + i\Lambda)u$ (where $\Lambda = (1-\partial_x^2)^{1/2}$) and applying a normal form transformation to eliminate the quadratic terms. This relies on the fact that the phase function $\Phi_{\mu_1 \mu_2}$ associated with quadratic interactions is non-vanishing in 1D.
   • Partial Second Transformation: To handle the "bad" cubic nonlinearity (specifically the term involving frequencies where the phase $\Phi_{++-}$ could be small), a second partial normal form transformation is performed. This transforms specific cubic terms into quartic terms, provided the frequencies satisfy certain separation conditions (e.g., high-low-low interactions).

2. Z-norm and Dispersive Estimates:

   • Instead of relying on the $L^1$ norm (which requires fast decay) or high-order vector fields, the authors introduce a specialized $Z_\alpha$-norm (defined via Littlewood-Paley decomposition and dyadic spatial localization). This norm is designed to capture the decay properties of solutions with initial data weighted by $\langle x \rangle^{1/2+}$.
   • The proof utilizes a linear dispersive estimate of the form $\|e^{\pm it\Lambda} f\|_{L^\infty} \lesssim (1+t)^{-\alpha} \|\Lambda^{3/2+} f\|_{Z_\alpha}$ for $\alpha \in (0, 1/2]$. This replaces the standard $(1+t)^{-1/2}$ decay associated with $L^1$ data.

3. Phase Function Analysis and Kernel Estimates:

   • A significant portion of the technical work involves the precise analysis of the Schwartz kernels associated with the cubic and quartic nonlinearities.
   • The authors perform a delicate classification of space-time and frequency interactions (e.g., Low-Low-High, High-Low-Low, etc.).
   • For "bad" phases where stationary points might exist, the authors analyze the critical points of the phase function $\Psi$ in physical space. They distinguish between regions near critical points (handled by bootstrap assumptions and dispersive estimates) and regions away from critical points (handled by the stationary phase method).
   • Crucially, the authors note that the 2D analysis techniques from previous works (e.g., [9]) cannot be directly applied because the 1D decay rate $(1+t)^{-1/2}$ is slower than the 2D rate $(1+t)^{-1}$. This slower decay necessitates a more refined analysis of the phase function to avoid losing too much regularity or decay in the estimates.

4. Bootstrap Argument:

   • The proof closes via a bootstrap argument. Assuming the solution exists and satisfies a bound on the $Z_\alpha$-norm and high-order Sobolev norm, the authors derive improved estimates that allow the continuation of the solution up to the claimed lifespan.

Key Contributions and Results

• Theorem 1.1 (Main Result): The paper proves the existence of almost global solutions for initial data with small $H^s$-Sobolev norms and a weight of $\langle x \rangle^{1/2+}$.

  • If the initial data satisfies a smallness condition involving the $Z_{1/2}$-norm (corresponding to weight $\langle x \rangle^{1/2+}$), the solution exists on the time interval $[0, e^{\kappa_0/\varepsilon^2} - 1]$.
  • If the weight is slightly weaker, $\langle x \rangle^{1/2}$ (corresponding to $Z_\alpha$ with $\alpha \in (0, 1/2)$), the solution exists on $[0, \kappa_0 \varepsilon^{-2/(1-2\alpha)}]$.
  • The decay rate of the solution is shown to be $\|(\Lambda u, \partial_t u)(t)\|_{L^\infty} \leq C\varepsilon (1+t)^{-\alpha}$.

• Corollary 1.2: For initial data with weight $\langle x \rangle^\beta$ where $\beta > 1/2$, the solution exists globally in the sense of almost global time $e^{\kappa_1/\varepsilon^2}$. This recovers the optimal lifespan bound known for rapidly decaying data but under a weaker decay assumption on the initial data.

• Corollary 1.3: For initial data with weight $\langle x \rangle^{1/2}$, the solution exists on $[0, \varepsilon^{-M}]$ for any $M > 0$. While not exponential, this represents a significant extension of existence time compared to polynomial bounds, demonstrating that even with minimal decay, the solution persists for a very long time.

• Optimality: Remark 1.1 notes that the exponential lifespan bound $e^{C/\varepsilon^2}$ is optimal for the case $\alpha = 1/2$, citing a counterexample from [7] where the lifespan cannot exceed this bound for certain cubic nonlinearities.

Significance and Claims
The paper claims to extend the theory of almost global existence for 1D Klein-Gordon equations to the regime of weakly decaying initial data.

• Overcoming Vector Field Limitations: The authors explicitly state that their method avoids the vector field method, which fails for data with low-order weights (like $\langle x \rangle^{1/2}$) because the weighted Sobolev norms of the derivatives would diverge.
• Refined Phase Analysis: The paper highlights that the slower time decay in 1D ($t^{-1/2}$) compared to 2D ($t^{-1}$) makes the analysis of the "bad" phase terms more difficult. The authors claim their detailed analysis of the Schwartz kernel and the critical points of the phase function allows them to close the bootstrap argument despite this slower decay.
• Comparison to Prior Work: The results generalize previous findings for rapidly decaying data (e.g., [12]) and provide a bridge to understanding the behavior of solutions with minimal decay, suggesting that global existence might still be possible for Euler-Poisson systems with similar low-order weights (Remark 1.7).

The paper does not propose new applications or experimental setups; its contribution is strictly within the mathematical analysis of partial differential equations, specifically establishing the lifespan and decay properties of solutions under relaxed initial data conditions.