Sharp spectral estimates for free boundary problems arising in plasma physics

1. Problem Statement and Context

The paper investigates a constrained semilinear free boundary problem arising in the modeling of plasma physics (specifically Tokamak devices). The problem, denoted as $(P)_\lambda$, is defined on a bounded domain $\Omega \subset \mathbb{R}^N$ ($N \ge 2$) with $C^3$ boundary:

$$\begin{cases} -\Delta \psi = [\alpha + \lambda \psi]_+^p & \text{in } \Omega \\ \int_\Omega [\alpha + \lambda \psi]_+^p = 1 \\ \psi = 0 & \text{on } \partial\Omega \end{cases}$$

Key Parameters and Constraints:

• $1 < p < p_N$, where $p_N = \infty$ for $N=2$ and $p_N = \frac{N}{N-2}$ for $N \ge 3$.
• $\lambda \ge 0$ is a parameter related to the inverse statistical temperature.
• $\alpha \in \mathbb{R}$ is an unknown Lagrange multiplier enforcing the integral constraint.
• $\psi \in C^{2,r}_0(\Omega)$ is the unknown potential.
• The term $[\cdot]_+$ denotes the positive part, creating a free boundary $\partial \{x \in \Omega : \alpha + \lambda \psi > 0\}$.

The Core Challenge:
Unlike standard Gelfand problems, the constraint $\int_\Omega [\alpha + \lambda \psi]_+^p = 1$ prevents the use of classical "minimal solution" theory. Consequently, the standard properties of the linearized operator (such as the simplicity of the first eigenvalue or the validity of the maximum principle) do not automatically hold. The authors focus on the first eigenvalue $\sigma_1$ of the associated non-local linearized operator:
$$L_\lambda[\phi] = -\Delta \phi - \tau_\lambda V_\lambda [\phi]_\lambda$$
where $\tau_\lambda = \lambda^p$, $V_\lambda = [\alpha_\lambda + \lambda \psi_\lambda]_+^{p-1}$, and $[\phi]_\lambda = \phi - \langle \phi \rangle_\lambda$ represents the deviation from the weighted mean.

The central question is: For which values of $\lambda$ is the first eigenvalue $\sigma_1$ strictly positive? This positivity is crucial for establishing the uniqueness of solutions and the monotonicity of the solution branch.

2. Methodology

The authors employ a combination of variational analysis, spectral theory for non-local operators, and bifurcation theory.

• Variational Formulation: They utilize the fact that solutions correspond to critical points of a free energy functional $F_\lambda(\rho)$ involving non-extensive entropy. This allows them to establish that $\sigma_1 \ge 0$ for variational solutions.
• Non-Local Eigenvalue Analysis: The linearized operator $L_\lambda$ is non-local due to the term $[\phi]_\lambda$. The authors analyze the spectrum of the associated compact operator $T_\lambda$ on $H^1_0(\Omega)$, relating the eigenvalues $\mu_j$ of $T_\lambda$ to $\sigma_j$ via $\mu_j = \frac{\tau_\lambda}{\tau_\lambda + \sigma_j}$.
• Bifurcation and Implicit Function Theorem: To study the global behavior of the solution curve, they use the Implicit Function Theorem. A critical step involves analyzing the behavior of the solution branch when $\sigma_1$ approaches zero.
• Bending Lemma (Crandall-Rabinowitz Type): A key technical tool is Proposition 4.1, a refined "bending lemma." It characterizes the local behavior of the solution curve $(\lambda(s), \alpha(s), \psi(s))$ near a point where $\sigma_1 = 0$. It establishes that if $\sigma_1=0$ is simple, the curve bends in a specific direction determined by the sign of $\langle [\phi_1]_\lambda, \psi \rangle_\lambda$.
• Symmetry and Convexity: For the main result, they restrict the domain to a ball $D_N$ (unit volume), leveraging the uniqueness and regularity of solutions on balls (Theorem B) and the symmetry of the domain to rule out bifurcation points.

3. Key Contributions and Results

Main Theorem (Theorem 1.1)

The paper's primary contribution is the proof that for the domain being a ball $D_N$ in any dimension $N \ge 2$:
$$\lambda^*(D_N, p) = +\infty$$
Interpretation: For any $\lambda \ge 0$, the first eigenvalue $\sigma_1(\alpha_\lambda, \psi_\lambda)$ of the linearized operator is strictly positive.

• This implies that the solution branch on a ball never encounters a bifurcation point where the first eigenvalue vanishes.
• Consequently, the solution curve is globally defined, simple, and monotonic in terms of energy and the parameter $\alpha$.

Implications for Uniqueness (Theorem 5.1)

As a direct corollary of Theorem 1.1 and Theorem B, the authors provide a new proof of the uniqueness of solutions for the Emden equation on a ball:
$$\begin{cases} -\Delta u = u^p & \text{in } B_1 \\ u > 0 & \text{in } B_1 \\ u = 0 & \text{on } \partial B_1 \end{cases}$$
The proof relies on the fact that if $\sigma_1 > 0$ everywhere, the mapping from the parameter $\lambda$ to the solution is injective, preventing multiple solutions for the same physical parameters.

Failure of Isoperimetric Properties

The authors demonstrate that the first eigenvalue $\sigma_1$ does not satisfy a Faber-Krahn type isoperimetric inequality.

• In standard eigenvalue problems, the ball minimizes the first eigenvalue.
• Here, if $\lambda^*(\Omega, p) = +\infty$ were true for all domains, it would imply the uniqueness of the Emden problem on all domains. Since the Emden problem is known to have non-unique solutions on certain non-convex or non-symmetric domains, $\lambda^*(\Omega, p)$ must be finite for those domains.
• This highlights a fundamental difference between this constrained free boundary problem and standard mean-field equations where domain topology does not affect the positivity of $\sigma_1$.

Open Problem and Conjectures

The paper leaves open the question of whether $\lambda^*(\Omega, p) = +\infty$ holds for convex domains in general dimensions.

• They conjecture that for 2D domains that are symmetric and convex with respect to coordinate axes, the solution branch is global and unique ($\lambda^* = +\infty$).
• They note that proving this would yield a complete description of the global solution branch for Gelfand problems, extending previous results by Korman and others.

4. Significance

1. Resolution of a Long-standing Issue: The paper resolves the question of spectral stability for plasma physics free boundary problems on spherical domains (balls), proving that the first eigenvalue remains positive for all $\lambda$. This confirms the stability of the physical solutions in these geometries.
2. New Proof of Uniqueness: It offers a novel, spectral-theoretic approach to the classical uniqueness of the Emden equation on balls, bypassing some traditional ODE techniques by using global bifurcation analysis.
3. Methodological Advancement: The development of the "bending lemma" for non-local operators with vanishing weights provides a robust tool for analyzing constrained problems where standard maximum principles fail.
4. Physical Relevance: By establishing the monotonicity of the energy and the parameter $\alpha$ along the solution branch, the results support the physical interpretation of the plasma equilibrium states, ensuring that the transition from low to high temperature regimes (increasing $\lambda$) follows a unique, stable path on spherical domains.
5. Limitations and Future Directions: The work explicitly identifies the failure of general isoperimetric properties, suggesting that the geometry of the domain is critical for the existence of multiple solutions in this class of problems. This sets a clear agenda for future research on non-spherical and non-convex domains.

In summary, the paper establishes that on balls, the constrained plasma free boundary problem is well-behaved globally: solutions are unique, the first eigenvalue is always positive, and the solution branch is a smooth, monotonic curve without bifurcations.