The heat trace for domains with curved corners
Pith reviewed 2026-05-21 18:57 UTC · model grok-4.3
The pith
For curvilinear polygons the heat trace expansion includes a new term at order t to the 1/2 that mixes corner angles with boundary curvatures.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
We derive the local heat trace expansion through order t^{1/2} for both Dirichlet and Neumann boundary conditions. The new coefficient decomposes into the usual smooth-boundary contribution and a sum of local curved-corner terms, each depending only on the interior angle α and the one-sided limiting curvatures κ± of the adjacent arcs. In the Dirichlet case, the curved-corner contribution has the form C_{1/2}(α,κ+,κ−)=c_{1/2}(α)(κ++κ−)/(4sin(α/2)), with c_{1/2}(α) given by an explicit sector heat kernel integral. We determine its sign for every 0<α<2π.
What carries the argument
The local curved-corner term C_{1/2}(α, κ+, κ−) that encodes the leading interaction between the corner opening angle and the limiting curvatures on either side.
If this is right
- The expansion separates contributions from smooth boundary parts and from each curved corner at order t^{1/2}.
- The Dirichlet curved-corner term has a sign that depends only on the convexity or reflex nature of the corner.
- Convex curvilinear polygons cannot share the full heat trace with a straight-sided polygon unless they are themselves straight-sided.
- This removes the straight corners assumption in prior results on isospectrality.
Where Pith is reading between the lines
- The sign law could provide similar obstructions for other boundary conditions or in higher dimensions.
- Numerical verification could be done by approximating the heat trace for a domain like a stadium with one curved corner.
- This interaction suggests that higher order terms in the expansion may involve more complicated combinations of curvature and angle derivatives.
Load-bearing premise
The asymptotic expansion of the heat kernel separates additively into smooth-boundary contributions along the arcs and local contributions from each isolated corner.
What would settle it
Computing the coefficient of t^{1/2} in the heat trace for a concrete example such as a quarter-circle with an additional curved corner and checking if it matches the sum of the smooth curvature term and the predicted C_{1/2} term.
read the original abstract
The heat trace of a planar polygon contains corner terms depending only on the opening angles, while the heat trace of a smooth planar domain contains curvature terms along the boundary. We show that, for curvilinear polygons, these two phenomena first interact at order $t^{1/2}$. We compute this first corner-curvature heat invariant and prove a sharp sign law for its Dirichlet angular factor: its sign is determined solely by whether the corner is convex or reflex. More precisely, we derive the local heat trace expansion through order $t^{1/2}$, for both Dirichlet and Neumann boundary conditions. The new coefficient decomposes into the usual smooth-boundary contribution and a sum of local curved-corner terms, each depending only on the interior angle $\alpha$ and the one-sided limiting curvatures $\kappa_{\pm}$ of the adjacent arcs. In the Dirichlet case, the curved-corner contribution has the form $C_{1/2}(\alpha,\kappa_+,\kappa_-) = c_{1/2}(\alpha)\frac{\kappa_+ + \kappa_-}{4\sin(\alpha/2)}$, with $c_{1/2}(\alpha)$ given by an explicit sector heat kernel integral. We determine its sign for every $0<\alpha<2\pi$. The sign law has a spectral consequence: it gives a new obstruction to a curvilinear polygon being Dirichlet-isospectral to a straight-sided polygon. In particular, every convex curvilinear polygon Dirichlet-isospectral to a straight-sided polygon must itself be straight-sided, removing the assumption of straight corners from the theorem of Enciso and G\'omez-Serrano.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. The paper derives the local heat trace expansion through order t^{1/2} for planar curvilinear polygons (C^∞ arcs meeting at isolated corners with one-sided limiting curvatures κ±) under Dirichlet and Neumann conditions. The t^{1/2} coefficient decomposes into the standard smooth-boundary curvature integral plus a sum of local curved-corner contributions; for Dirichlet, each such term is C_{1/2}(α,κ+,κ−)=c_{1/2}(α)(κ++κ−)/(4 sin(α/2)) where c_{1/2}(α) is given by an explicit sector heat kernel integral. The sign of c_{1/2}(α) is determined for all 0<α<2π, yielding a sign law and a new obstruction to Dirichlet isospectrality between curvilinear and straight-sided polygons.
Significance. If the separation and explicit integral hold, the result supplies the first mixed corner-curvature heat invariant at order t^{1/2} and removes the straight-corner hypothesis from the Enciso–Gómez-Serrano isospectral rigidity theorem. The explicit sector-integral representation and sharp sign law are concrete strengths that make the coefficient falsifiable and potentially useful for further spectral geometry.
major comments (2)
- [Main expansion theorem (likely §3)] The central decomposition of the t^{1/2} coefficient into smooth-boundary curvature integral plus purely local corner terms C_{1/2}(α,κ+,κ−) requires that the local sector model near each corner produces an error o(t^{1/2}) uniformly in the curvature parameters. The manuscript must supply explicit remainder estimates controlling the contribution from curvature variation along the adjacent C^∞ arcs; without them the claimed separation at precisely this order is not justified (see the statement of the main expansion theorem and the construction of the local model).
- [Sign determination section (likely §4 or appendix)] The sign law for c_{1/2}(α) rests on the explicit sector heat kernel integral. The manuscript should include a self-contained verification or numerical check of this integral that confirms the sign is determined solely by convexity/reflexivity for every 0<α<2π; post-hoc analysis alone leaves a verification gap at the level of the central coefficient.
minor comments (2)
- [Introduction] Clarify the precise meaning of 'one-sided limiting curvatures' κ± in the introduction and ensure they are consistently denoted when the arcs are parametrized.
- [Abstract and §2] The abstract states an 'explicit sector heat kernel integral' for c_{1/2}(α); if this integral is left in unevaluated form, a brief remark on its computability or reduction would aid readability.
Simulated Author's Rebuttal
We thank the referee for the careful reading, positive assessment of the significance, and constructive major comments. We address each point below and will incorporate revisions to strengthen the justification and verification in the manuscript.
read point-by-point responses
-
Referee: [Main expansion theorem (likely §3)] The central decomposition of the t^{1/2} coefficient into smooth-boundary curvature integral plus purely local corner terms C_{1/2}(α,κ+,κ−) requires that the local sector model near each corner produces an error o(t^{1/2}) uniformly in the curvature parameters. The manuscript must supply explicit remainder estimates controlling the contribution from curvature variation along the adjacent C^∞ arcs; without them the claimed separation at precisely this order is not justified (see the statement of the main expansion theorem and the construction of the local model).
Authors: We agree that explicit remainder estimates are required to rigorously justify the o(t^{1/2}) error uniformly in the curvature parameters. The proof of the main expansion (Theorem 3.1) controls the error through the C^∞ regularity of the arcs and the construction of the local sector model, but these controls are not stated as standalone estimates. In the revised manuscript we will add a new appendix (or subsection of §3) that supplies the missing explicit remainder bounds, showing that the contribution from curvature variation along each adjacent arc is indeed o(t^{1/2}) uniformly for bounded κ±. This will make the separation of the local curved-corner terms fully justified at the claimed order. revision: yes
-
Referee: [Sign determination section (likely §4 or appendix)] The sign law for c_{1/2}(α) rests on the explicit sector heat kernel integral. The manuscript should include a self-contained verification or numerical check of this integral that confirms the sign is determined solely by convexity/reflexivity for every 0<α<2π; post-hoc analysis alone leaves a verification gap at the level of the central coefficient.
Authors: We appreciate the referee’s request for a self-contained verification. The sign of c_{1/2}(α) follows from the explicit integral representation of the sector heat kernel; our analytic argument already shows that the sign is governed by convexity (α<π) versus reflexivity (α>π) for all 0<α<2π. To close the verification gap we will add, in the revised version, a short appendix containing both a brief analytic summary of the sign determination and a numerical quadrature check of the integral over a fine grid of angles in (0,2π). The numerical results will confirm that the sign changes exactly at α=π and depends only on convexity/reflexivity, thereby providing concrete, reproducible evidence for the sign law. revision: yes
Circularity Check
Derivation of t^{1/2} curved-corner heat trace coefficient is self-contained via sector integral
full rationale
The paper derives the local heat trace expansion through order t^{1/2} by separating the contribution into the standard smooth-boundary curvature integral plus purely local corner terms C_{1/2}(α, κ+, κ−) whose angular factor c_{1/2}(α) is obtained from an explicit sector heat kernel integral. No step reduces a claimed prediction to a fitted parameter by construction, no load-bearing uniqueness theorem is imported from the authors' prior work, and the separation hypothesis is stated as an assumption on the domain class rather than derived from the result itself. The derivation therefore remains independent of its own outputs.
Axiom & Free-Parameter Ledger
axioms (1)
- standard math The heat kernel on a planar domain with piecewise C^∞ boundary admits a short-time asymptotic expansion that separates into interior, boundary, and corner contributions at each order.
Reference graph
Works this paper leans on
-
[1]
T. Branson and P. Gilkey. The asymptotics of the Laplacian on a manifold with boundary.Communications in partial differential equations, 15(2):245–272, 1990
work page 1990
-
[2]
J. Cheeger. Spectral geometry of singular Riemannian spaces.J. Differential Geom., 18(4):575–657, 1983. [3]NIST Digital Library of Mathematical Functions.https://dlmf.nist.gov/, Release 1.2.4 of 2025-03-15. F. W. J. Olver, A. B. Olde Daalhuis, D. W. Lozier, B. I. Schneider, R. F. Boisvert, C. W. Clark, B. R. Miller, B. V. Saunders, H. S. Cohl, and M. A. M...
work page 1983
-
[3]
A. Enciso and J. G´ omez-Serrano. Spectral determination of semi-regular polygons.Journal of Differential Geometry, 122(3):399–419, November 2022
work page 2022
-
[4]
D. Grieser. Basics of the b-calculus. InApproaches to Singular Analysis: A Volume of Advances in Partial Differential Equations, pages 30–84. Springer, 2001
work page 2001
-
[5]
D. Grieser. Notes on heat kernel asymptotics. Lecture notes, 2004. Available athttps://web.math.ku.dk/grubb/notes/ heat.pdf
work page 2004
-
[6]
H. McKean Jr and I. Singer. Curvature and the eigenvalues of the laplacian.Journal of Differential Geometry, 1(1-2):43–69, 1967
work page 1967
-
[7]
M. Kac. Can one hear the shape of a drum?The american mathematical monthly, 73(4P2):1–23, 1966
work page 1966
-
[8]
Melrose.The Atiyah-Patodi-Singer Index Theorem
R. Melrose.The Atiyah-Patodi-Singer Index Theorem. A K Peters/CRC Press, 1st edition, 1993
work page 1993
-
[9]
M. Nursultanov, J. Rowlett, and D. A. Sher. The heat kernel on curvilinear polygonal domains in surfaces.Annales math´ ematiques du Qu´ ebec, 49(1):1–61, 2025
work page 2025
-
[10]
M. van den Berg and S. Srisatkunarajah. Heat equation for a region inR 2 with a polygonal boundary.J. London Math. Soc. (2), 37(1):119–127, 1988. 22 S. LOOI AND D. SHER
work page 1988
-
[11]
M. van den Berg and S. Srisatkunarajah. Heat flow and brownian motion for a region inR 2 with a polygonal boundary. Prob. Theory and Related Fields, 86:41–52, 1990. California Institute of Technology, The Division of Physics, Mathematics and Astronomy, 1200 E California Blvd, Pasadena CA 91125 Email address:looi@caltech.edu DePaul University, Department o...
work page 1990
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.