Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Smooth edge rounding and dividing-curve connection in the standard Legendrian collar

Proved
ContactTopology.smooth_legendrian_edge_rounding

by Johan Mercedes · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

contact-topologyconvex-surfacesdifferential-topology

For every positive integer nnn and scale r>0r>0r>0, consider the standard Legendrian-collar contact form on its real-coordinate cover,

αn=sin⁡(2πnz) dx+cos⁡(2πnz) dy.\alpha_n=\sin(2\pi n z)\,dx+\cos(2\pi n z)\,dy.αn​=sin(2πnz)dx+cos(2πnz)dy.

Its coefficients are periodic in the longitudinal coordinate zzz with period one. There exist a number c>0c>0c>0 and a globally smooth embedded strip S:R2→R3S:\mathbb R^2\to\mathbb R^3S:R2→R3, with injective derivative everywhere, which agrees exactly with the two flat faces outside the bend:

S(s,z)=(0,r(c−s),z)(s≤0),S(s,z)=(r(c+s−1),0,z)(s≥1).S(s,z)=(0,r(c-s),z)\quad(s\le0),\qquad S(s,z)=(r(c+s-1),0,z)\quad(s\ge1).S(s,z)=(0,r(c−s),z)(s≤0),S(s,z)=(r(c+s−1),0,z)(s≥1).

Thus the joins to the flat pieces are smooth to every order. The longitudinal rulings are Legendrian.

The affine radial field about (rc,rc)(rc,rc)(rc,rc) is transverse to the whole strip. Its complete smooth flow is an actual family of diffeomorphisms and multiplies the contact form by ete^tet. The contact-volume coefficient of the model is exactly 2πn>02\pi n>02πn>0.

There is a smooth phase function with ψ(0)=0\psi(0)=0ψ(0)=0 and ψ(1)=π/2\psi(1)=\pi/2ψ(1)=π/2 such that the dividing locus on the bent part is exactly

z=k2n−ψ(s)2πn,k∈Z,0≤s≤1.z=\frac{k}{2n}-\frac{\psi(s)}{2\pi n},\qquad k\in\mathbb Z,\quad 0\le s\le1.z=2nk​−2πnψ(s)​,k∈Z,0≤s≤1.

Each branch has a globally smooth regular parametrization

γk(s)=S ⁣(s,k2n−ψ(s)2πn).\gamma_k(s)=S\!\left(s,\frac{k}{2n}-\frac{\psi(s)}{2\pi n}\right).γk​(s)=S(s,2nk​−2πnψ(s)​).

It joins the point (0,rc,k/(2n))(0,rc,k/(2n))(0,rc,k/(2n)) on the first face to (rc,0,k/(2n)−1/(4n))(rc,0,k/(2n)-1/(4n))(rc,0,k/(2n)−1/(4n)) on the second. This proves the quarter-turn connection of the dividing curves with a genuinely smooth rounding, rather than only tangent circular and flat pieces.

This is the explicit standard local model for edge rounding. It does not itself provide a compatible collar for arbitrary surfaces, descend the construction to a quotient manifold, construct a bypass disk, or prove a generic convex movie or the Legendrian-core classification.

Preamble
import Mathlib.Analysis.Calculus.ContDiff.Deriv
import Mathlib.Analysis.Calculus.ContDiff.Operations
import Mathlib.Analysis.Calculus.Deriv.Comp
import Mathlib.Analysis.Calculus.Deriv.MeanValue
import Mathlib.Analysis.Calculus.Deriv.Pi
import Mathlib.Analysis.Calculus.Deriv.Prod
import Mathlib.Analysis.Calculus.FDeriv.Prod
import Mathlib.Analysis.SpecialFunctions.ExpDeriv
import Mathlib.Analysis.SpecialFunctions.SmoothTransition
import Mathlib.Analysis.SpecialFunctions.Trigonometric.ArctanDeriv
import Mathlib.Analysis.SpecialFunctions.Trigonometric.Basic
import Mathlib.Analysis.SpecialFunctions.Trigonometric.Deriv
import Mathlib.MeasureTheory.Integral.IntervalIntegral.FundThmCalculus
import Mathlib.Tactic.FieldSimp
import Mathlib.Tactic.FinCases
import Mathlib.Tactic.FunProp
import Mathlib.Tactic.Linarith
import Mathlib.Tactic.LinearCombination
import Mathlib.Tactic.NormNum
import Mathlib.Tactic.Ring
import Mathlib.Topology.Constructions.SumProd
import Mathlib.Topology.Order.IntermediateValue
open Set Function
open scoped ContDiff
Formal statement
theorem ContactTopology.smooth_legendrian_edge_rounding (n : ℕ) (hn : 0 < n)
    (r : ℝ) (hr : 0 < r) :
    let a : (Fin 3 → ℝ) → (Fin 3 → ℝ) := fun y =>
      ![Real.sin (2*Real.pi*n*y 2),Real.cos (2*Real.pi*n*y 2),0]
    let α : (Fin 3 → ℝ) → (Fin 3 → ℝ) → ℝ := fun y v =>
      Real.sin (2*Real.pi*n*y 2)*v 0+Real.cos (2*Real.pi*n*y 2)*v 1
    ∃ c : ℝ, ∃ S : ℝ × ℝ → (Fin 3 → ℝ), ∃ X : (Fin 3 → ℝ) → (Fin 3 → ℝ),
    ∃ Φ : ℝ → (Fin 3 → ℝ) → (Fin 3 → ℝ), ∃ phase : ℝ → ℝ,
    ∃ γ : ℤ → ℝ → (Fin 3 → ℝ),
      0 < c ∧ ContDiff ℝ ∞ S ∧ Topology.IsEmbedding S ∧
      (∀ p, Injective (fderiv ℝ S p)) ∧ ContDiff ℝ ∞ X ∧
      ContDiff ℝ ∞ (fun p : ℝ × (Fin 3 → ℝ) => Φ p.1 p.2) ∧
      ContDiff ℝ ∞ phase ∧ (∀ k, ContDiff ℝ ∞ (γ k)) ∧
      (∀ y, X y=![y 0-r*c,y 1-r*c,0]) ∧
      (∀ t y, Φ t y=![r*c+Real.exp t*(y 0-r*c),r*c+Real.exp t*(y 1-r*c),y 2]) ∧
      (∀ y, Φ 0 y=y) ∧ (∀ t y, Φ (-t) (Φ t y)=y) ∧
      (∀ t y, HasDerivAt (fun s => Φ s y) (X (Φ t y)) t) ∧
      (∀ t y v, α (Φ t y) (fderiv ℝ (Φ t) y v)=Real.exp t*α y v) ∧
      (∀ y,
        a y 0*(fderiv ℝ a y (Pi.single 1 1) 2-fderiv ℝ a y (Pi.single 2 1) 1)+
        a y 1*(fderiv ℝ a y (Pi.single 2 1) 0-fderiv ℝ a y (Pi.single 0 1) 2)+
        a y 2*(fderiv ℝ a y (Pi.single 0 1) 1-fderiv ℝ a y (Pi.single 1 1) 0)=2*Real.pi*n) ∧
      (∀ y v : Fin 3 → ℝ, α ![y 0,y 1,y 2+1] v=α y v) ∧
      (∀ s z, s ≤ 0 → S (s,z)=![0,r*(c-s),z]) ∧
      (∀ s z, 1 ≤ s → S (s,z)=![r*(c+s-1),0,z]) ∧
      (∀ p v : ℝ × ℝ, X (S p) ≠ fderiv ℝ S p v) ∧
      (∀ p : ℝ × ℝ, α (S p) (fderiv ℝ S p (0,1))=0) ∧
      phase 0=0 ∧ phase 1=Real.pi/2 ∧
      (∀ k s, γ k s=S (s,(k:ℝ)/(2*n)-phase s/(2*Real.pi*n))) ∧
      (∀ k s, deriv (γ k) s ≠ 0) ∧
      (∀ k, γ k 0=![0,r*c,(k:ℝ)/(2*n)] ∧
        γ k 1=![r*c,0,(k:ℝ)/(2*n)-1/(4*n)]) ∧
      ∀ s ∈ Icc (0:ℝ) 1, ∀ z : ℝ,
        (α (S (s,z)) (X (S (s,z)))=0 ↔
          ∃ k : ℤ, z=(k:ℝ)/(2*n)-phase s/(2*Real.pi*n)) := by sorry
Source
Ko Honda, Gluing tight contact structures, https://arxiv.org/abs/math/0102029, Section 1.5, edge-rounding discussion, arXiv p.11: standard collar form sin(2*pi*n*z) dx + cos(2*pi*n*z) dy and the dividing-curve connection shift -1/(4n). This is a constructed smooth model on the real-coordinate cover; the all-orders straight joins, complete affine contact flow, and exhaustive smooth dividing branches are derived details, not a verbatim general-manifold theorem.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me