The Rabinowitz continuum of subcritical Gelfand problems and free boundary-type equations arising in plasma physics

1. Problem Statement and Motivation

The paper addresses the qualitative behavior of the Rabinowitz unbounded continuum of classical solutions to the subcritical Gelfand problem:
$$\begin{cases} -\Delta v = \mu(1 + v)^p & \text{in } \Omega \\ v > 0 & \text{in } \Omega \\ v = 0 & \text{on } \partial\Omega \end{cases} \quad (Q)_\mu$$
where $\Omega \subset \mathbb{R}^N$ ($N \ge 2$) is a bounded domain, $1 < p < p_N$ (with $p_N = \infty$ for $N=2$ and $p_N = \frac{N+2}{N-2}$ for $N \ge 3$), and $\mu > 0$.

Key Challenges:

• Connectivity and Monotonicity: While the existence of a continuum containing both minimal and non-minimal solutions is known (Rabinowitz, 1971), a sharp, detailed description of the global branch—specifically regarding the monotonicity of the $L^\infty$ norm or energy along the branch for general domains—is lacking.
• Limitations of Existing Methods: Classical results establish that on balls, solutions are radial and the branch has a "bell-shaped" profile in the $(\mu, v(0))$ plane. However, for general domains (even convex ones), the monotonicity of the norm along the branch containing non-minimal solutions is not well understood.
• Non-Minimal Solutions: Standard variational methods often focus on minimal solutions. Describing the global structure including non-minimal (mountain-pass) solutions requires new tools.

2. Methodology: The Constrained Approach

The authors introduce a novel approach by transforming the Gelfand problem $(Q)_\mu$ into a constrained free boundary problem arising in plasma physics (specifically related to the Grad-Shafranov equation for Tokamaks).

The Constrained Problem $(P)_\lambda$:
Instead of varying $\mu$, they study the following problem with a fixed integral constraint:
$$\begin{cases} -\Delta \psi = [\alpha + \lambda \psi]_+^p & \text{in } \Omega \\ \int_\Omega [\alpha + \lambda \psi]_+^p \, dx = 1 \\ \psi = 0 & \text{on } \partial\Omega \end{cases} \quad (P)_\lambda$$
where $\lambda \ge 0$ is a parameter, $\alpha \in \mathbb{R}$ is an unknown Lagrange multiplier, and $[\cdot]_+$ denotes the positive part.

The Connection:
If $(\alpha_\lambda, \psi_\lambda)$ solves $(P)_\lambda$ with $\alpha_\lambda > 0$, then $v(\lambda) = \frac{\lambda}{\alpha_\lambda}\psi_\lambda$ solves $(Q)_\mu$ with $\mu = \mu_\lambda = \lambda \alpha_\lambda^{p-1}$.

Core Strategy:

1. Uniqueness of $(P)_\lambda$: Establish the uniqueness of solutions to the constrained problem for a range of $\lambda$.
2. Stability Analysis: Define a weighted first eigenvalue $\sigma_1(\alpha_\lambda, \psi_\lambda)$ for the linearized operator of $(P)_\lambda$. The authors prove that solutions are stable (i.e., $\sigma_1 > 0$) along a specific branch.
3. Energy Parametrization: Replace the classical $L^\infty$ norm monotonicity with the monotonicity of the plasma energy $E_\lambda = \frac{1}{2} \int_\Omega |\nabla \psi_\lambda|^2$.
4. Global Continuation: Follow the curve of stable solutions from $\lambda=0$ until the critical point where $\alpha_\lambda \to 0^+$, thereby covering the entire Rabinowitz continuum.

3. Key Contributions and Theoretical Results

A. Uniqueness of Grad-Shafranov Type Equations (Theorem 1.1)

The authors solve a long-standing open problem regarding the uniqueness of solutions to equations of Grad-Shafranov type.

• Result: They define $\lambda^*(\Omega, p)$ as the supremum of $\lambda$ for which the weighted first eigenvalue $\sigma_1$ is positive. They prove $\lambda^*(\Omega, p) > \lambda_0(\Omega, p)$ (where $\lambda_0$ is a known lower bound).
• Significance: For any $\lambda \in [0, \lambda^*)$, there exists a unique solution to $(P)_\lambda$, forming a $C^1$ curve $G^*(\Omega)$. Along this curve, $\alpha_\lambda$ is strictly decreasing and the energy $E_\lambda$ is strictly increasing. This is the first unconditional uniqueness result for these equations in general dimensions.

B. Global Parametrization of the Gelfand Continuum (Theorem 1.2)

Using the unique stable branch of $(P)_\lambda$, the authors construct a global parametrization of the Rabinowitz continuum for $(Q)_\mu$.

• Result: The set $R_+ = \{(\mu_\lambda, v(\lambda)) : \lambda \in [0, \min\{\lambda^*, \lambda_+\})\}$ is a real analytic branch of solutions.
• Monotonicity: The energy $E_\lambda$ is strictly increasing along the branch. This provides a meaningful generalization of the monotonicity of minimal solutions to non-minimal solutions as well.
• Asymptotics: As $\lambda \to \lambda_+$ (where $\alpha_\lambda \to 0^+$), the parameter $\mu_\lambda \to 0^+$ and $\|v(\lambda)\|_\infty \to +\infty$, confirming the unbounded nature of the continuum.

C. The Case of the Unit Ball (Theorems 1.4 and 1.5)

For the specific case where $\Omega = D_N$ (a ball of unit volume), the authors provide a complete characterization.

• Inequality Proof (Theorem 1.4): They prove that $\lambda^*(D_N, p) \ge \lambda_+(D_N, p)$. This implies that the stable branch covers the entire range of $\lambda$ up to the point where $\alpha_\lambda$ vanishes.
• Bell-Shaped Profile (Theorem 1.5): They explicitly demonstrate the "bell-shaped" profile of the solution branch in the $(\mu, E)$ plane:
  • $\mu_\lambda$ increases initially, reaches a maximum $\mu_t$ at some $\lambda_t$, and then decreases to 0.
  • $E_\lambda$ is strictly increasing throughout.
  • This recovers the classical result for radial solutions but via a new method that does not rely a priori on radial symmetry assumptions for the general theory.

4. Technical Innovations

1. Weighted Spectral Theory: The authors develop a refined spectral analysis for the linearized operator $L_\lambda = -\Delta - \tau_\lambda (\rho_\lambda)^{1/q} [\cdot]_\lambda$. Unlike standard operators, the weight $(\rho_\lambda)^{1/q}$ can vanish on a large set (when $\alpha_\lambda < 0$). They define a weighted first eigenvalue $\sigma_1$ that remains well-defined and positive, allowing for stability analysis even when solutions are not strictly positive everywhere.
2. Bending Lemma (Crandall-Rabinowitz Type): To handle the point where $\sigma_1 = 0$ (the turning point of the branch), they employ a refined bifurcation analysis. They prove that on a ball, the zero eigenvalue is simple and the transversality condition holds, ensuring the branch bends back smoothly without bifurcation.
3. Energy vs. Norm: The shift from tracking $\|v\|_\infty$ to tracking the energy $E_\lambda$ is crucial. While $\|v\|_\infty$ monotonicity is hard to prove for non-minimal solutions on general domains, the energy monotonicity is naturally derived from the stability condition ($\sigma_1 > 0$).

5. Significance and Impact

• Resolution of Open Problems: The paper resolves the open problem of the uniqueness of solutions for Grad-Shafranov type equations with integral constraints, a problem dating back to foundational work in plasma physics.
• New Global Description: It provides the first detailed description of the Rabinowitz continuum that includes non-minimal solutions for general domains, linking the existence of such solutions to the stability of the associated constrained problem.
• Methodological Shift: The approach of using constrained free boundary problems (plasma physics models) to study semilinear elliptic equations (Gelfand problems) offers a powerful new toolkit. It bypasses the difficulties of direct bifurcation analysis on the original Gelfand problem.
• Generalization: While the explicit "bell shape" is proven for balls, the framework establishes that the monotonicity of energy is a robust property of the continuum, suggesting that similar qualitative behaviors (like the existence of a turning point for $\mu$) may hold for broader classes of domains (e.g., convex domains), provided the inequality $\lambda^* \ge \lambda_+$ holds.

In summary, the paper successfully bridges plasma physics and nonlinear PDE theory to provide a rigorous, global, and monotonic parametrization of the solution set for subcritical Gelfand problems, extending classical results on minimal solutions to the full continuum of solutions.