On Free Boundary Problems Shaped by Varying Singularities

Basic Information

• Paper ID: 2401.08071
• Title: On free boundary problems shaped by varying singularities
• Authors: Damião J. Araújo, Aelson Sobral, Eduardo V. Teixeira, José Miguel Urbano
• Classification: math.AP (Analysis of PDEs)
• Publication Date: November 12, 2025 (v2)
• Paper Link: https://arxiv.org/abs/2401.08071

Abstract

This paper pioneering studies variational free boundary models with varying singularities. The theory strongly depends on the properties of the singular exponent γ(x) and how it varies. Under mild continuity assumptions on γ(x), the authors establish optimal regularity for minimizers. These estimates vary pointwise, leading to a continuum of free boundary geometries shaped by the singularities. The paper provides extensive analysis of free boundaries shaped by singularities. Using novel monotonicity formulas, the authors prove that if the singular exponent γ(x) varies in a W^{1,n+} manner, then the free boundary is locally a C^{1,δ} surface, except for a negligible singular set of Hausdorff codimension at least 2.

Research Background and Motivation

Problem Definition

The core problem studied in this paper concerns analyzing free boundary problems containing varying singularities, specifically involving the minimization of energy functionals:

$E(v,O) = \int_O F(Dv, v, x) dx$

where the Lagrangian function F(p⃗, v, x) is non-differentiable in the parameter v, and the degree of singularity varies with the spatial variable x. The associated Euler-Lagrange equation is:

$\Delta u = s(x,u)\chi_{\{u>0\}}$

where the singular term satisfies $\lim_{v\to 0} s(x,v) = \infty$, and the singularity strength varies with location.

Research Significance

1. Theoretical Value: Traditional free boundary theory assumes uniform singularities (such as the classical Alt-Phillips problem $\Delta u \sim u^{-p_0}$, where p₀ is constant). This paper is the first to systematically study the case where the singular exponent varies spatially, filling a theoretical gap.
2. Applied Value:
   • Non-uniform Catalytic Surfaces: In heterogeneous catalytic surfaces, the effective singularity strength varies due to fluctuations in local adsorption, microstructure, or wettability
   • Thin Film Flow: Thin films on patterned substrates
   • Porous Composite Materials: Material heterogeneity leads to local reaction rate differences
3. Mathematical Challenges: The variation in singularity introduces fundamental difficulties:
   • Free boundary geometry is no longer uniform but varies pointwise
   • Cannot rely on uniformity assumptions
   • Requires development of new analytical tools

Limitations of Existing Methods

Existing methods for handling singular PDE models depend to varying degrees on uniformity assumptions for the blow-up rates. When the singular exponent γ(x) varies:

• Regularity estimates and non-degeneracy properties have different homogeneities
• Cannot expect the free boundary to have uniform regularity
• Classical monotonicity formulas no longer apply

Core Contributions

1. Establishing a Variational Framework for Varying Singularities: First systematic study of minimizers of the energy functional $J^{\gamma(x)}_{\delta(x)}(v) := \int \frac{1}{2}|Dv|^2 + \delta(x)(v_+)^{\gamma(x)}dx$ where γ(x) and δ(x) are functions of spatial variables.
2. Optimal Regularity Estimates:
   • Under mild continuity assumptions, prove C^{1,α*} regularity of minimizers (Theorem 2.1)
   • Under weak Dini continuity conditions, obtain pointwise optimal growth estimates (Theorem 4.1): $u(y) \leq C'|y-z_0|^{\frac{2}{2-\gamma(z_0)}}$
3. Novel Monotonicity Formula: Develop a Weiss-type monotonicity formula applicable to varying singularities (Theorem 6.1) for classifying blow-up limits.
4. Free Boundary Regularity: Prove that under γ, δ ∈ W^{1,n+} conditions, the free boundary is locally a C^{1,δ} surface, except for a singular set of Hausdorff dimension at most n-2 (Theorem 7.1).
5. Hausdorff Measure Estimates: Under W^{2,∞} regularity assumptions, prove that the H^{n-1} measure of the free boundary is finite (Theorem 5.1).

Detailed Methodology

Task Definition

Input:

• Domain Ω ⊂ ℝⁿ
• Singular exponent function γ: Ω → (0,1]
• Weight function δ: Ω → ℝ₊
• Boundary data φ ∈ H¹(Ω) ∩ L∞(Ω)

Objective: Find a minimizer u of the energy functional in the admissible class $\mathcal{A} = \{v \in H^1(\Omega): v - \phi \in H^1_0(\Omega)\}$ and study:

1. Regularity of the minimizer
2. Geometric properties and regularity of the free boundary ∂{u > 0}

Core Technical Roadmap

1. Existence and Preliminary Regularity (Section 2)

Scaling Properties: For x₀ ∈ Ω and parameters A, B ∈ (0,1], define $w(x) := \frac{u(x_0 + Ax)}{B}$ Then w is a minimizer of the scaled functional, where $\tilde{\delta}(x) := B^{\gamma(x_0+Ax)}\left(\frac{A}{B}\right)^2\delta(x_0+Ax)$

Choosing $B = r^{\beta}$, $\beta = \frac{2}{2-\gamma_*(x_0,A)}$ maintains boundedness of $\|\tilde{\delta}\|_{L^\infty}$.

Local Regularity (Theorem 2.1): Using harmonic replacement techniques, prove $\|u\|_{C^{1,\alpha}(\Omega')} \leq C, \quad \alpha = \frac{\gamma_*(\Omega')}{2-\gamma_*(\Omega')}$

Key steps:

• For minimizer u and its harmonic replacement h, energy inequality holds: $\int_{B_R} |Du|^2 - |Dh|^2 dx \leq 2\int_{B_R} \delta(x)(h^{\gamma(x)} - u^{\gamma(x)})dx$
• Using $h^{\gamma(x)} - u^{\gamma(x)} \leq |u-h|^{\gamma_*(0,R)}$ and Sobolev inequality obtain $\int_{B_R} |Du-Dh|^2 dx \leq CR^{n+2\frac{\gamma_*(0,R)}{2-\gamma_*(0,R)}}$
• Apply Campanato embedding theorem

Non-degeneracy (Theorem 2.2): Prove that in {u > 0}, $\sup_{\partial B_r(y)} u \geq c r^{\frac{2}{2-\gamma_*(y,r)}}$

Using auxiliary function $\phi(x) = u(x)^{2-\frac{3}{2}\gamma_*(y,r)} - c|x-y|^2$ and maximum principle.

2. Weakly Dini Continuous Exponents and Optimal Estimates (Section 4)

Key Assumption: γ is continuous at free boundary point z₀, with modulus of continuity ω satisfying $\omega(1) + \limsup_{t\to 0^+} \omega(t)\ln\left(\frac{1}{t}\right) \leq \tilde{C}$

This is a weaker condition than classical Dini continuity.

Optimal Growth Estimate (Theorem 4.1): There exist constants r₀, C' such that $u(y) \leq C'|y-z_0|^{\frac{2}{2-\gamma(z_0)}}$

Proof strategy:

• From Theorem 2.1 obtain $u(y) \leq Cr^{\frac{2}{2-\gamma_*(z_0,r)}}$
• Estimate exponent difference: $\frac{2}{2-\gamma_*(z_0,r)} - \frac{2}{2-\gamma(z_0)} \leq 2\omega(r)$
• Use assumption condition: $r^{2\omega(r)} \geq e^{-2C_*}$

Strong Non-degeneracy (Theorem 4.2): $\sup_{\partial B_r(z_0)} u \geq c_* r^{\frac{2}{2-\gamma(z_0)}}$

Positive Density and Porosity (Theorem 4.3): Prove $\frac{|B_r(z_0) \cap \Omega(u)|}{|B_r(z_0)|} \geq \mu_0$ and the free boundary is porous, hence $\mathcal{H}^{n-\epsilon}(\mathcal{F}(u) \cap B_{1/2}) = 0$.

3. Hausdorff Measure Estimates (Section 5)

Under γ, δ ∈ W^{2,∞} assumptions, prove $\mathcal{H}^{n-1}(\mathcal{F}(u) \cap B_{1/2}) < C$.

Key Lemma (Lemma 5.1): Pointwise gradient estimate $|Du(x)|^2 \leq 2\delta(x)[u(x)]^{\gamma(x)} + c_1u(x)$

Proof uses auxiliary function $w(y) = |Du(y)|^2 - 2\delta(y)[u(y)]^{\gamma(y)} - K_1u(y) - \zeta(|y-x_0|)[u(y)]^{\gamma(y)}$ and proves its maximum point cannot be in the positive set interior or on the free boundary.

Integrability (Lemma 5.2): $u(x)^{-\gamma(x)/2} \in L^1(\Omega(u) \cap B_{1/2})$

Hausdorff Estimate: Using covering arguments and the above integrability, estimate $\mathcal{H}^{n-1}(\mathcal{F}(u) \cap B_r) \leq Cr^{n-1}$

4. Monotonicity Formula and Blow-up Classification (Section 6)

Monotonicity Formula (Theorem 6.1): For z₀ ∈ F(u), define $W_{u,z_0}(r) = r^{-(n+2(\beta_0-1))}J^\gamma_\delta(u,B_r(z_0)) - \frac{1}{2}\beta_0 r^{-((n-1)+2\beta_0)}\int_{\partial B_r(z_0)} u^2 d\mathcal{H}^{n-1} - \text{error terms}$

where $\beta_0 = \frac{2}{2-\gamma(z_0)}$, prove $\frac{d}{dr}W_{u,z_0}(r) \geq 0$.

Blow-up Classification (Corollary 6.1): If u₀ is a blow-up limit of u at z₀, i.e., $u_r(x) := \frac{u(z_0+rx)}{r^{\beta(z_0)}} \to u_0$ then u₀ is β(z₀)-homogeneous.

Proof key points:

• Scale the monotonicity formula
• Prove $W^\infty_{u_0,0}(r)$ is constant
• Use $\int_{\partial B_r}(\partial_\nu u_0 - \beta_0 r^{-1}u_0)^2 = 0$ to obtain homogeneity

5. Free Boundary Regularity (Section 7)

Linearized Equation: Define $w = \varrho(z_0)^{-1/\beta(z_0)}u^{1/\beta(z_0)}$, then in {w > 0} $\Delta w = \frac{h(x,w,Dw)}{w}$

where $h(x,s,\xi) = \delta(x)\frac{\gamma(x)}{\beta(z_0)}\varrho(z_0)^{\gamma(x)-2}s^{\beta(z_0)(\gamma(x)-\gamma(z_0))} - (\beta(z_0)-1)|\xi|^2$

Key Observation: As r → 0, $h_r(x,s,\xi) \to h(z_0,\xi) = (\beta(z_0)-1)(1-|\xi|^2)$ with convergence uniform at free boundary points z₀ ∈ F(u).

Flatness Improvement (Proposition 7.1): Under control conditions $\sup([γ]_{C^{0,\mu}(0)}, [δ]_{C^{0,\mu}(0)}) \leq \epsilon^2$ if $(x_n - \epsilon)_+ \leq w^-(x) \leq w(x) \leq w^+(x) \leq (x_n + \epsilon)_+$ then there exists ν ∈ ∂B₁ such that $(\nu \cdot x - \frac{\epsilon}{2}r)_+ \leq w(x) \leq (\nu \cdot x + \frac{\epsilon}{2}r)_+$

Dimension Reduction: Using Federer dimension reduction arguments, starting from classification of two-dimensional minimal cones, recursively prove Hausdorff dimension estimates for higher-dimensional singular sets.

Main Theorem (Theorem 7.1): Under γ, δ ∈ W^{1,n+} conditions, the free boundary F(u) is locally a C^{1,δ} surface, except for a singular set of Hausdorff dimension ≤ n-2.

Technical Innovations

1. Handling Variable Exponents: By introducing w⁻ and w⁺ to sandwich the original function w, using equations they satisfy with definite sign to construct strict upper and lower solutions.
2. Generalization of Monotonicity Formula: Extend classical Weiss monotonicity formula to variable exponent case, introducing error terms to handle variation in γ(x).
3. Linearization Strategy: Prove that at small scales, variable exponent problems are well approximated by constant exponent problems with controllable error terms.
4. Unified Regularity Theory: Despite varying singularities, establish a unified regularity framework; the key is proving uniform convergence.

Experimental Setup

This is a pure theoretical mathematics paper with no numerical experiments. All results are obtained through rigorous mathematical proofs.

Theoretical Verification Structure

The paper employs a hierarchical theoretical construction:

1. Foundational Theory (Section 2): Results not depending on γ(x) continuity
2. Refined Theory (Sections 3-4): Optimal estimates under weak continuity assumptions
3. Measure Theory (Section 5): Hausdorff estimates under W^{2,∞} assumptions
4. Regularity Theory (Sections 6-7): C^{1,δ} regularity under W^{1,n+} assumptions

Hierarchy of Assumptions

• Weakest: 0 < γ₊(Ω) ≤ γ*(Ω) ≤ 1 (Theorems 2.1, 2.2)
• Intermediate: Weak Dini continuity (Theorems 4.1-4.3)
• Stronger: γ, δ ∈ W^{2,∞} (Theorem 5.1)
• Optimal: γ, δ ∈ W^{1,n+} (Theorem 7.1)

Experimental Results

Main Theoretical Results

1. C^{1,α} Regularity (Theorem 2.1):
   • Under only boundedness and measurability of γ
   • Obtain α = γ₊(Ω')/(2-γ₊(Ω'))
   • This regularity is suboptimal but requires no continuity assumption
2. Optimal Pointwise Estimates (Theorem 4.1):
   • Under weak Dini continuity conditions
   • Obtain precise growth rate $u(y) \sim |y-z_0|^{2/(2-\gamma(z_0))}$
   • Estimates are pointwise, reflecting local variation in singularity
3. C^{1,δ} Regularity of Free Boundary (Theorem 7.1):
   • Under W^{1,n+} assumptions
   • Hausdorff dimension of singular set ≤ n-2
   • First regularity result for variable exponent case

Theoretical Findings

1. Geometric Diversity: Unlike the uniform geometry in constant exponent case, variable exponent problems exhibit a continuum of free boundary geometries.
2. Critical Continuity: Weak Dini continuity is the critical condition for obtaining optimal pointwise estimates, weaker than classical Dini continuity.
3. Regularity Threshold: W^{1,n+} regularity is the natural threshold ensuring C^{1,δ} regularity of the free boundary, consistent with general variable exponent p(x)-Laplacian theory.

Example Analysis (Example 4.1)

Consider singular exponent $\gamma(x,v) = \frac{1}{2} - \frac{3}{(\ln(\min(v(x),e^{-3})))^2}$

• Globally only measurable and bounded, γ(x,v) ≥ 1/6
• On free boundary {u = 0}, γ ≡ 1/2
• Theorem 2.1 gives C^{12/11} regularity (based on γ₊ = 1/6)
• Theorem 4.1 gives C^{4/3} regularity (based on γ = 1/2 at free boundary)

This demonstrates significant improvement of pointwise estimates over global estimates.

Related Work

Classical Theory

1. Alt-Phillips Problem 3, 22, 23: Studies $\Delta u \sim u^{-p_0}$ (p₀ constant) free boundary problems, foundational for this paper.
2. Obstacle Problem 21: Corresponds to p₀ = 0 case.
3. Cavitation Problem: Corresponds to limit p₀ ↗ 1.

Variable Exponent Theory

1. Variable Exponent p(x)-Growth Functionals 1: Studies $\int F(x,Du)$ where F has p(x)-growth in p.
2. Variable Exponent Non-variational Theory 8: Bronzi et al. study regularity of variable exponent fully nonlinear elliptic equations.
3. Degenerate Diffusion Equations 4: C¹ regularity for variable exponent degenerate equations.

Free Boundary Theory Progress

1. Single-phase Problems 10, 11: De Silva-Savin work on Lipschitz free boundaries and degenerate single-phase problems.
2. Negative Power Alt-Phillips Functionals 12: Studies negative power case.
3. Fractional Laplacian Related 27: Yang's work on fractional Laplacian free boundary problems.

Innovation of This Paper

This paper is the first to systematically study spatially varying singular exponents in free boundary problems, a fundamental generalization of classical theory:

• Classical theory: Uniform singularity → Uniform geometry
• This paper: Varying singularity → Continuum of geometries

Conclusions and Discussion

Main Conclusions

1. Varying Singularities are Tractable: Despite fundamental difficulties from singularity variation, complete theory can be established under appropriate regularity assumptions.
2. Hierarchical Optimal Regularity:
   • No continuity: C^{1,α*} (suboptimal)
   • Weak Dini continuity: Pointwise optimal estimates
   • W^{1,n+}: C^{1,δ} free boundary
3. Geometric Complexity: Free boundary geometry varies pointwise with singularity, exhibiting rich structure.

Limitations

1. Regularity Assumptions:
   • Hausdorff estimates require W^{2,∞} (strong)
   • Free boundary regularity requires W^{1,n+}
   • Whether results hold under weaker assumptions remains open
2. Two-dimensional Dependence: Dimension reduction arguments depend on complete classification of two-dimensional minimal cones, limiting method applicability.
3. Singular Set: While proving Hausdorff dimension ≤ n-2, no finer structural description is given.
4. Uniqueness: Paper does not discuss uniqueness of minimizers (Remark 2.1 indicates possible failure in general).

Future Directions

1. More General Variational Models (Example 4.1): Study $J(v) = \int \frac{1}{2}|Dv|^2 + \delta(x)(v_+)^{\gamma(x,v(x))}dx$ where γ depends on the solution itself.
2. Parabolic Case: Extend theory to evolution problems.
3. Weaker Regularity Assumptions: Explore what results hold under γ ∈ W^{1,p} (p ≤ n).
4. Fine Structure of Singular Set: Study whether singular set has better properties (e.g., rectifiability).
5. Applied Problems: Apply theory to concrete physical models (e.g., non-uniform catalysts, composite materials).
6. Two-phase Problems: Remark 2.2 indicates method applies to two-phase problems, deserving deeper study.

In-Depth Evaluation

Strengths

1. Pioneering Research:
   • First systematic study of free boundary problems with varying singularities
   • Fills important gap between classical Alt-Phillips theory and practical applications
   • Establishes complete theoretical framework
2. Technical Depth:
   • Develops novel monotonicity formula for variable exponent case
   • Clever use of w⁻ and w⁺ sandwich technique
   • Sophisticated linearization arguments proving variable exponent problems are well approximated by constant exponent problems
3. Result Optimality:
   • Theorem 4.1 pointwise estimates are optimal
   • Weak Dini continuity condition is nearly necessary
   • W^{1,n+} assumption consistent with general variable exponent theory
4. Theoretical Completeness:
   • Complete theoretical chain from existence to regularity
   • Hierarchical results under different regularity assumptions
   • Both local results and global measure estimates
5. Clear Exposition:
   • Logical structure and clear reasoning
   • Sufficient exposition of main ideas
   • Rigorous technical details

Weaknesses

1. Strength of Assumptions:
   • W^{2,∞} assumption (Theorem 5.1) is too strong, difficult to verify in applications
   • W^{1,n+} assumption, while natural, still excludes many interesting cases
2. Two-dimensional Dependence:
   • Free boundary regularity proof strongly depends on classification of two-dimensional minimal cones
   • If classification fails (e.g., other singularity types), method may not apply
3. Lack of Concrete Examples:
   • Besides Example 4.1, few concrete application examples
   • No numerical simulations verifying theoretical predictions
4. Uniqueness Problem:
   • Does not discuss uniqueness of minimizers
   • Remark 2.1 mentions possible non-uniqueness but lacks deep analysis
5. Singular Set Structure:
   • Only provides Hausdorff dimension upper bound
   • Does not study whether singular set is rectifiable or has other geometric structure
6. Computational Complexity:
   • Results are highly abstract
   • Unclear how to actually compute or estimate free boundary for given γ(x)

Impact

1. Theoretical Contribution:
   • Establishes theoretical foundation for free boundary problems with varying singularities
   • Novel monotonicity formula may have applications elsewhere
   • Sandwich technique and linearization strategy have methodological value
2. Subsequent Research:
   • Will inspire more research on variable exponent free boundary problems
   • Parabolic case, obstacle problems, two-phase problems all deserve study
   • May advance development of variable exponent PDE theory
3. Application Potential:
   • Provides theoretical tools for free boundary problems in non-uniform materials
   • May apply to catalytic reactions, thin film flow, biological invasion models
   • Requires further numerical method development for practical application
4. Reproducibility:
   • As pure theoretical work, proofs are verifiable
   • Lacks numerical implementation, limiting practical application
   • Recommend subsequent work develop numerical algorithms

Applicable Scenarios

1. Non-uniform Catalytic Surfaces: Heterogeneity in catalytic surface leads to spatial variation in local reaction rates.
2. Composite Materials: Spatial variation in diffusion coefficients and reaction rates in porous composites.
3. Thin Film Problems: Thin film flow on patterned substrates with spatially varying wettability.
4. Biological Models: Biological population invasion models where environmental heterogeneity affects diffusion and growth rates.
5. Image Processing: Variable exponent image segmentation models adapting to local image features.
6. Shape Optimization: Shape optimization problems with spatially varying constraints.

Recommendations

1. Numerical Methods: Develop efficient numerical algorithms for solving variable exponent free boundary problems.
2. Concrete Applications: Collaborate with applied scientists to apply theory to practical problems.
3. Relaxed Assumptions: Research whether meaningful results hold under weaker regularity assumptions.
4. Uniqueness: Investigate conditions for uniqueness of minimizers.
5. Singular Set: Further characterize geometric structure of singular set.

Selected References

3 H.W. Alt and D. Phillips, A free boundary problem for semilinear elliptic equations, J. Reine Angew. Math. 368 (1986), 63–107. (Classical Alt-Phillips problem)

10 D. De Silva and O. Savin, Regularity of Lipschitz free boundaries for the thin one-phase problem, J. Eur. Math. Soc. 17 (2015), 1293–1326.

11 D. De Silva and O. Savin, On certain degenerate one-phase free boundary problems, SIAM J. Math. Anal. 53 (2021), 649–680. (Degenerate single-phase problems, important reference for this paper's methods)

12 D. De Silva and O. Savin, The Alt–Philips functional for negative powers, Bull. London Math. Soc. 55 (2023), 2749–2777.

21 A. Petrosyan, H. Shahgholian and N. Uraltseva, Regularity of free boundaries in obstacle-type problems, AMS Graduate Studies in Mathematics 136 (2012). (Classical textbook on free boundary problems)

Overall Assessment: This is a high-quality theoretical mathematics paper that pioneering studies free boundary problems with varying singularities. The theory is profound, methods are innovative, and results are optimal. While assumptions are relatively strong, it establishes a solid theoretical foundation for this emerging field. The paper has significant theoretical value for PDEs and free boundary problems, and provides theoretical support for practical applications. Recommend subsequent research focus on numerical method development and concrete applications to fully realize the theory's value.