On perturbations of the DN map of a disk causing changes of surface topology

Basic Information

• Paper ID: 2511.18179
• Title: On perturbations of the DN map of a disk causing changes of surface topology
• Author: D.V. Korikov (PDMI RAS, ITMO University)
• Classification: math-ph (Mathematical Physics), math.CV (Complex Variables), math.MP (Mathematical Physics)
• Publication Date: November 25, 2025 (arXiv preprint)
• Paper Link: https://arxiv.org/abs/2511.18179

Abstract

This paper establishes the degeneracy properties of Schottky hyperbolic surfaces on genus 1 Riemann surfaces with boundary: when the Dirichlet-to-Neumann (DN) map approaches the DN map of the unit disk, the hyperbolic surface degenerates.

Research Background and Motivation

Core Problem

This paper investigates a fundamental mathematical problem in Electric Impedance Tomography (EIT): the stability of surface topological structure when the DN map of a Riemann surface with boundary undergoes perturbation.

Problem Significance

1. Uniqueness of DN maps: It is known that the DN map uniquely determines the conformal class of a surface (up to conformal diffeomorphisms fixing boundary points)
2. Continuity Paradox: Although the conformal class depends continuously on the DN map under fixed topology (measured in the Teichmüller metric), the surface topology itself is unstable under small perturbations of the DN map
3. Physical Significance: This corresponds to the question of whether internal topological structure can be stably inferred from boundary data in actual physical measurements

Limitations of Existing Research

1. Known examples: disks with small handles, where the DN map approaches that of the standard disk as handle size tends to zero
2. Lack of general theory: for general cases, the mechanism by which topological defects (such as additional handles) separate from the near-boundary region remains unclear
3. Key Difficulty: Any curve can be "inflated" by changing the metric via a conformal factor, making the characterization of geometric degeneracy dependent on choosing a canonical metric representative

Research Motivation

The author proposes a hypothesis: when the DN map approaches that of the disk, topological defects should be separated from the near-boundary region by short closed geodesics. To verify this hypothesis, one must choose hyperbolic metrics of constant curvature -1 as canonical representatives, where the boundary is geodesic.

Core Contributions

The main contributions of this paper include:

1. Main Theorem (Proposition 0.1): For a genus 1 surface with boundary M, when its DN map Λ approaches the DN map |∂φ| of the unit disk in operator space B, the length l₁(X) of the shortest closed geodesic on the Schottky hyperbolic surface X satisfies l₁(X) → 0
2. Technical Innovation: Establishes quantitative connections between spectral data of the DN map (via Hilbert transform) and the period matrix of the hyperbolic surface
3. Classification of Degeneracy Mechanisms: Through asymptotic behavior of the period matrix, identifies two types of degeneracy:
   • Contraction along homologically trivial closed curves (e.g., the boundary itself contracting)
   • Contraction along homologically non-trivial closed curves
4. Theoretical Refinement: Provides rigorous proof of the geometric mechanism underlying the instability of surface topology under DN map perturbations

Detailed Methods

Task Definition

Input: Smooth orientable surface (M, g) with boundary T (unit circle) and genus gen(M) = 1

Output: Prove that when DN map Λ → |∂φ|, the shortest closed geodesic length l₁(X) → 0 on the Schottky hyperbolic surface X

Constraints:

• M equipped with smooth metric g
• DN map defined as Λ: f ↦ ∂ᵥu_f, where u_f is the harmonic extension of f
• Schottky hyperbolic surface X formed by gluing two copies of M along the boundary

Core Method Architecture

The proof consists of 5 key steps:

Step 1: Spectral Analysis of Hilbert Transform

Define the Hilbert transform: $H = -Λ^{-1}∂_φ$

Key Properties:

• H is a bounded operator on L²(T;ℂ)
• iH is self-adjoint on H^{1/2}(T;ℂ)/ℂ with respect to the inner product (·,·)_Λ
• Spectral Structure:
  • Essential spectrum: {-i, +i} (infinite multiplicity)
  • Discrete spectrum: 0 (KerH = ℂ) and simple eigenvalues ±iμ, where μ ∈ (0,1)

Convergence: As Λ → |∂φ|, we have H → H₀ (Hilbert transform of the disk), thus μ → 1

Step 2: Harmonic Differentials and Period Matrices

Introduce the space of harmonic differentials: $D = \{υ ∈ Ω^1(M;ℝ) | dυ = 0, δυ = 0 \text{ in } M, υ(∂_φ) = 0 \text{ on } ∂M\}$

Key Connection:

• dimℝ D = 2 (since gen(X) = 2)
• Via Hodge decomposition, establish correspondence between boundary data and harmonic differentials: $υ(ν) = -Λ(H² + I)f, \quad f = cη + \bar{c}\bar{η}$ where η is an eigenfunction of H

Inner Product Formula (Equation 4): $\|(υ, υ')\|_{L²(M,T^*M)} = 2(1-μ²)ℜ(c\bar{c}')$$\|(\star υ, υ')\|_{L²(M,T^*M)} = 2(μ²-1)μℑ(c\bar{c}')$

Step 3: Computation of Period Matrices

Choose symmetric homology basis (Equation 5): $a_+, a_- := τ ∘ a_+, b_+, b_- := -τ ∘ b_+$

Introduce dual basis υ_a, υ_b ∈ D satisfying: $\int_{q_+} υ_e = δ_{qe}, \quad (q, e = a, b)$

Auxiliary Period Matrix B (Equation 6): $B_{qe} = \int_{q_+} \star υ_e$

Via Riemann bilinear relations, obtain: $\det B = μ^{-2}$

b-Period Matrix (Equation 10): $B = -\frac{(μ²+1)i}{2μ²B_{ab}}\begin{pmatrix} 1 & \frac{μ²-1}{μ²+1} \\ \frac{μ²-1}{μ²+1} & 1 \end{pmatrix} + \frac{B_{bb}}{μ²B_{ab}}\begin{pmatrix} 1 & 0 \\ 0 & -1 \end{pmatrix}$

Step 4: Hyperelliptic Curve Representation and Degeneracy Analysis

Hyperelliptic Representation: A genus 2 Riemann surface X can be represented as: $y² = P(x) = x(x-1)(x-λ)(x-μ)(x-ν)$

By analyzing the relationship between theta constants (Equation 13) and parameters λ, μ, ν, examine two degeneracy cases:

Case i: {B_r}_r is bounded

• Can assume B_r → B_∞ = diag(B₁, B₂) (diagonal matrix)
• From μ → 1 and Equation (10), deduce that B_∞ diagonalizes
• Decomposition properties of theta functions lead to μ, ν → ∞
• Conclusion: Hyperelliptic surface X_r degenerates

Case ii: (ℑB)^{-1} → 0

• Theta constants e_{2a₁,2a₂,2b₁,2b₂} → δ_{a₁,0}δ_{a₂,0}
• Leads to λ → 1
• Conclusion: Surface contracts along homologically non-trivial closed curve

Step 5: Application of Mumford Compactness Theorem

Proof by Contradiction:

• Assume l₁(X) ↛ 0, then there exists a subsequence {X_r}_r with l₁(X_r) ≥ ε > 0
• By Mumford compactness theorem, the set K_ε = {X ∈ M₂ | l₁(X) ≥ ε} is compact in moduli space M₂
• But Step 4 shows that {X_r}_r cannot be contained in any compact set
• Contradiction, therefore l₁(X) → 0

Technical Innovations

1. Bridge Between Boundary Data and Geometry: Via spectral parameter μ of the Hilbert transform, directly control the period matrix, establishing quantitative connections between boundary observables and internal geometry
2. Application of Moduli Space Geometry: Cleverly utilize Giavedoni's results on fundamental domains of real Riemann surfaces (Equation 11), ensuring boundedness of period matrices in appropriate homology bases
3. Synthesis of Degeneracy Theory: Combine classical results of Fay on asymptotic expansions of period matrices of degenerating surfaces to identify degeneracy types
4. Unification of Topology and Analysis: Transform topological instability (genus change) into geometric degeneracy (geodesic length tending to zero), then quantify via analytical tools (operator spectrum)

Experimental Setup

Theoretical Nature

This is a pure mathematical theory paper with no numerical experiments or empirical research. All results are rigorous mathematical proofs.

Geometric Examples

The paper provides a concrete geometric example at the end (Figure 1) to illustrate the theorem:

Construction: Consider the case where the boundary ∂M itself is a short closed geodesic

• When l(∂M) → 0, by the collar lemma, there exists a metric neighborhood $\{x ∈ X | \text{dist}_h(x, ∂M) ≤ L\}$ biholomorphic to an annulus {z ∈ ℂ | log|z| ∈ (-r, r)}, where $L ∼ -\log(l(∂M)), \quad r = \frac{π²}{l(∂M)}$
• M is biholomorphic to the unit disk D with |z| ≤ e^{-π²/l(∂M)} removed, with a surface with handles attached at boundary |z| = e^{-r}

Literature Support: Such degenerating surface families have been constructed in reference 6

Experimental Results

Main Theoretical Results

Theorem (Proposition 0.1): If gen(M) = 1, then as Λ → |∂φ| in B, we have l₁(X) → 0.

Key Intermediate Results

1. Spectral Convergence (Step 1): $μ → 1 \quad \text{as} \quad Λ → |∂φ|$
2. Period Matrix Degeneracy (Steps 3-4):
   • Case i: B → diagonal matrix
   • Case ii: (ℑB)^{-1} → 0
3. Hyperelliptic Parameter Degeneracy (Step 4):
   • Case i: μ, ν → ∞
   • Case ii: λ → 1

Geometric Interpretation

The two types of degeneracy correspond to:

• Type a (Figure 1 left): The boundary itself contracts, M becomes an approximate disk with handles
• Type b (Figure 1 right): Internal closed curve contracts, hyperbolic surface splits into two lower-genus surfaces

Theoretical Significance

1. Geometric Mechanism of Topological Instability: Proves that small perturbations of the DN map necessarily accompany the appearance of short geodesics, explaining why topology is unstable
2. Consistency with Known Results: Completely consistent with examples constructed in reference 6
3. Escape from Moduli Space: The conformal class X_r escapes toward the boundary of moduli space M₂ (boundary of Deligne-Mumford compactification)

Related Work

Uniqueness Theory of DN Maps

1. Lassas-Uhlmann 9: Proved that DN map determines the conformal class of Riemannian manifolds (up to diffeomorphisms fixing the boundary)
2. Lee-Uhlmann 10: DN map is a first-order pseudodifferential operator, differing from ∂_φ by a smooth operator

Stability Theory

3. Belishev-Korikov 2, 7: Under fixed topology, conformal class depends continuously on DN map (Teichmüller metric)
4. Korikov 6: Constructed examples of topological instability (small handle disks)

Spectral Methods

5. Belishev 1: BC method for solving Calderon problem, introducing Hilbert transform
6. Korikov 8: Established relationship between DN map and period matrix (directly used in this paper)

Degeneracy Theory

7. Fay 15: Asymptotic expansion of period matrices for degenerating Riemann surface families
8. Giavedoni 5: Fundamental domains of period matrices for real Riemann surfaces
9. Mumford 11: Compactness theorem (set of X with l₁(X) ≥ ε is compact in moduli space)

Hyperbolic Geometry

10. Farb-Margalit 3: Mapping class groups and collar lemma
11. Wang-Xue 14: Applications of length spectrum

Hyperelliptic Curves

12. Gaudry 4, van Wamelen 13: Relationship between theta functions and hyperelliptic curve parameters

Conclusions and Discussion

Main Conclusions

1. Core Theorem: The Schottky hyperbolic surface of a genus 1 surface with boundary necessarily degenerates when the DN map approaches that of the disk (shortest geodesic length tends to zero)
2. Degeneracy Mechanisms: Degeneracy occurs in two ways:
   • The boundary itself contracts (homologically trivial)
   • Internal handle contracts (homologically non-trivial)
3. Theoretical Significance: Provides complete geometric explanation for the instability of surface topology under DN map perturbations

Limitations

1. Genus Restriction: Proof applies only to gen(M) = 1; higher genus cases remain unresolved
2. Qualitative Results: Although l₁(X) → 0 is proved, quantitative convergence rates are not provided
3. Single Boundary: Only considers simply connected boundary (unit circle)
4. Hyperelliptic Dependence: Proof essentially exploits hyperellipticity of genus 2 surfaces; generalization to higher genus requires new techniques

Future Directions

Research directions implicit in the paper:

1. Higher Genus Generalization: Extend results to gen(M) > 1
   • Must handle more complex period matrices (genus g hyperbolic surface has genus 2g)
   • Degeneracy analysis for non-hyperelliptic curves is more difficult
2. Quantitative Estimates: Establish quantitative relationship between l₁(X) and ‖Λ - |∂φ|‖_B
3. Multiply Connected Boundaries: Study cases where boundary has multiple connected components
4. Inverse Problems: Given short geodesics, can one construct corresponding degenerating DN map families?
5. Numerical Methods: Develop numerical algorithms for computing DN maps and period matrices

In-Depth Evaluation

Strengths

1. Innovation of Methods

• Cross-disciplinary Synthesis: Cleverly combines inverse problem theory (DN maps), complex analysis (Riemann surfaces), differential geometry (hyperbolic metrics), and algebraic geometry (moduli space theory)
• Technical Bridge: Establishes quantitative connections between spectral data of boundary operators and surface geometry, having methodological significance in inverse problem theory

2. Mathematical Rigor

• Clear logical structure with solid theoretical foundation at each step
• Fully utilizes modern mathematical tools (Mumford compactness theorem, Teichmüller theory, theta functions)
• Complete technical details (homology basis selection, fundamental domain application)

3. Theoretical Depth

• Reveals the geometric essence of DN map topological instability: not arbitrary topological change, but specific mechanisms through geodesic contraction
• Connects two seemingly contradictory phenomena: stability of conformal class (fixed topology) vs. instability of topology

4. Clarity of Exposition

• Introduction well articulates research motivation
• Proof structure is clear (5 steps), easy to follow
• Provides geometric intuition (Figure 1) and concrete examples

Weaknesses

1. Limited Applicability

• Genus 1 Restriction: Serious limitation, as applications may encounter more complex topologies
• Dependence on hyperellipticity makes generalization difficult

2. Lack of Quantitative Estimates

• Although l₁(X) → 0 is proved, convergence rates are not provided
• For practical applications (determining if DN map perturbation is small enough to preserve topology), quantitative information is needed

3. Pure Existence Results

• Theorem only states "short geodesic exists" without specifying its location or properties
• No discrimination criteria provided for two degeneracy types

4. Absence of Algorithms and Numerical Verification

• As a theory paper, no computational methods provided
• No numerical examples verify theoretical predictions

Impact

Contributions to the Field

1. Inverse Problem Theory: Provides theoretical foundation for topological stability problems in electrical impedance tomography
2. Geometric Analysis: Demonstrates how to study internal geometry through boundary data
3. Riemann Surface Theory: Offers new perspective for studying degenerating surface families (via DN maps)

Practical Value

• Direct Application Limited: Due to genus 1 restriction and lack of quantitative estimates, direct practical value is limited
• Theoretical Guidance: Provides theoretical framework for understanding topological uncertainty in actual measurements
• Methodological Value: Developed techniques (Hilbert transform-period matrix connection) may apply to other problems

Reproducibility

• Theoretical Results: Proofs are fully verifiable with rigorous logic
• Numerical Implementation: Principles can be implemented in practice, though paper provides no code or algorithms
• Literature Dependence: Heavily references author's previous work 8; full understanding requires consulting these references

Applicable Scenarios

Theoretical Research

• Mathematicians studying relationships between DN maps and surface geometry
• Scholars researching degeneracy theory of Riemann surfaces
• Researchers in inverse problem theory

Potential Applications

1. Medical Imaging: Topological identification in electrical impedance tomography
2. Geophysics: Inferring underground structure from surface measurements
3. Materials Science: Detecting internal defects (holes, cracks) in materials

Constraints

• Requires 2D models
• Requires low-genus topology
• Requires precise boundary measurements

References

The paper cites 15 key references covering:

Inverse Problem Theory:

• 1 Belishev (2003) - BC method
• 9 Lassas-Uhlmann (2001) - DN map uniqueness
• 10 Lee-Uhlmann (1989) - Pseudodifferential operator properties

Author's Previous Work:

• 2 Belishev-Korikov (2023) - Teichmüller stability
• 6 Korikov (2021) - Topological instability examples
• 7 Korikov (2024) - Non-orientable surfaces
• 8 Korikov (2025) - Period matrix determination (core dependency of this paper)

Riemann Surface Theory:

• 5 Giavedoni (2013) - Real surface fundamental domains
• 12 Silhol (1989) - Real algebraic surfaces
• 15 Fay (1973) - Degeneracy theory

Hyperbolic Geometry:

• 3 Farb-Margalit (2012) - Mapping class groups
• 11 Mumford (1971) - Compactness theorem
• 14 Wang-Xue (2025) - Length spectrum

Algebraic Geometry:

• 4 Gaudry (2007) - Theta function computation
• 13 van Wamelen (1998) - Hyperelliptic curve equations

Overall Assessment

This is a high-quality pure mathematics theory paper making important contributions at the intersection of inverse problem theory and Riemann surface geometry. Although results are limited to genus 1 surfaces, the proof techniques are sophisticated, revealing the deep geometric mechanism underlying DN map topological instability. The paper provides important theoretical insights into understanding the relationship between boundary measurements and internal topology, laying foundations for future research. Main directions for improvement are generalization to higher genus and establishment of quantitative estimates.