Cited by 0Geometric Analysis and Curvature FlowsAdvanced Differential Geometry ResearchNonlinear Partial Differential Equations

Generating shrinkers by mean curvature flow: Dedicated to Tom Ilmanen (1961–2025)

Open access
Hoffman, D., Martín, F., White, B.·2026·Mathematische Annalen

Abstract

We prove existence for many examples of shrinkers by producing compact, smoothly embedded surfaces that, under mean curvature flow, develop singularities at which the shrinkers occur as blowups.

Verified 2026-11-06 · 34 references not checked yet

AI snapshot

Written from the paper’s full text.

TL;DR

The paper proves existence of many shrinkers by constructing compact, smoothly embedded initial surfaces whose mean curvature flow develops singularities with those shrinkers as blowups. It recovers known examples and produces new families, including Platonic/polygonal shrinkers and high-genus shrinkers in \(S^2\times\mathbb R\).

Key findings

  • For a finite set \(Q\) of rays from the origin intersecting \(S^2\) in either the face centers of a Platonic solid or three or more equally spaced points in \(S^2\cap\{z=0\}\), there is a smoothly embedded shrinker \(\Sigma\) with the symmetries of \(Q\), disjoint from \(Q\), of genus \(g=\#Q-1\), occurring as a blowup of a compact embedded mean curvature flow (Theorem 12).
  • For each \(g\ge1\) and \(1+d_2<\delta<1+d_1\), there is a compact smoothly embedded surface in \(S^2\times\mathbb R\) whose flow has a singularity where every shrinker \(\Sigma_g\) is \(G_g\)-invariant, has genus \(g\), satisfies \(\Sigma_g\cap\{z=0\}=Q_g\), has \(\Sigma_g\setminus\{z=0\}\) of genus 0, and has entropy \(\le\delta\); for large \(g\), \(\Sigma_g\) has exactly one end and converges to \(\{z=0\}\cup\partial B(0,2)\) (Theorem 40).
  • For all sufficiently large \(g\), there is a compact smoothly embedded surface in \(S^2\times\mathbb R\) whose flow has a singularity where every shrinker \(\Sigma_g\) has genus \(2g\), three ends, is \(G_g\)-invariant, satisfies \(\Sigma_g\cap\{z=0\}=Q_g\), has \(\Sigma_g\setminus\{z=0\}\) of genus 0, and has entropy bounded independently of \(g\); as \(g\to\infty\), \(\Sigma_g\) converges to \(\{z=0\}\) with multiplicity 3 (Theorem 43).
  • All shrinkers at a given spacetime point have the same genus (Theorem 68).

Methods

  • The paper is theoretical: it constructs compact, smoothly embedded initial surfaces in \(\mathbb R^3\) or \(S^2\times\mathbb R\), then studies their level-set/Brakke mean curvature flow up to the first positive-genus singularity time \(T_{\mathrm{pos}}\).
  • Main tools include elliptic regularization, unit-regular Brakke flows, local and initial regularity theorems, maximum principles, Huisken monotonicity, topological arguments, and symmetry/barrier arguments.
  • The desired shrinkers are obtained as blowups at singularities, with genus and topology controlled by the constructed initial surfaces and by topological lemmas such as Lemma 25.

Limitations

  • The paper states it does not know whether the shrinkers in Theorem 12 are compact; in the regular \(n\)-gon case, for sufficiently large \(n\) the shrinker must be noncompact, and as \(n\to\infty\) it converges to a multiplicity 2 plane.
  • Several results are only for large genus: Theorem 40’s one-end behavior and Theorem 43’s genus \(2g\), three ends, and multiplicity-3 limit hold for large \(g\); Conjecture 56 that \(\Sigma\) has genus \(2g\) is true for sufficiently large \(g\) by Theorem 59.
  • The text makes evident that the constructions rely on substantial prior regularity and compactness results, and the paper has no associated data.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s00208-026-03546-x

Cites

0 of 34 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 20 cited sources checked · 0 supported · 0 flagged

  • Findingreasoning

    If T_pos < ∞, then there is a shrinker at time T_pos of genus ≥ 1.

    “Theorem 7. If T pos < ∞, then there is a shrinker at time T pos of genus ≥ 1.”
    Quote in the paperSubstantial claim
  • Findingnone given

    For a set Q of rays as specified in Theorem 12, there exists a smoothly embedded shrinker Σ that has the symmetries of Q and is disjoint from Q.

    “(1) Σ has the symmetries of Q and is disjoint from Q.”
    Quote in the paperSubstantial claim
  • Findingown data

    For large g, there are shrinkers that have the defining properties of those non-existent minimal surfaces.

    “By contrast, we show that (for large g) there are shrinkers that have the defining properties of those non-existent minimal surfaces.”
    Quote in the paperSubstantial claim
  • Findingnone given

    If T_fat < ∞, then there is a shrinker at time T_fat with genus ≥ 1.

    “Theorem 5.If T fat < ∞, then there is shrinker at time T fat with genus ≥ 1.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [CHH22]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [BK23]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Findingreasoning

    If (p,t) is a singular point with t ≤ T_fat, and if there is an ε > 0 such that genus(M(τ) ∩ B(p,ε)) ≤ g for every regular time τ ∈ (t-ε,t), then (p,t) has genus ≤ g.

    “Theorem 3. If (p, t) is a singular point with t ≤ T fat , and if there is an ϵ > 0 such that genus(M (τ ) ∩ B(p, ϵ)) ≤ g for every regular time τ ∈ (t -ϵ, t), then (p, t) has genus ≤ g.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim. 4
  • May combine two claims. 15
  • The quote only partly states this claim. 1
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Section 4 produces shrinkers that resemble skeletons of the Platonic solids.

    “Section 4 produces shrinkers that resemble skeletons of the Platonic solids.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Findingnone given

    Assertion (3) in the Platonic solid case is new.

    “Assertion (3) in the Platonic solid case is new.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingcitation

    The Platonic shrinkers in Section 4 were first produced by Ketover using minimax, but it was not known until now that they arise as blowups.

    “In particular, the Platonic shrinkers in Section 4 were first produced by Ketover [Ket16] using minimax, but it was not known until now that they arise as blowups.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [Ket16]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • The quote only partly states this claim.

    Findingnone given

    The shrinker Σ in Theorem 12 has genus g, where g is the number of rays minus 1.

    “(2) Σ has genus g.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingcitation

    The genus-g examples in Section 7 were produced for high genus by desingularization and for every genus by minimax, but, except in the case of high genus, it was not known until now that they arise as blowups.

    “Similarly, the genus-g examples in Section 7 were produced for high genus by desingularization [KKM18,Ngu14] and for every genus by minimax [BNS24], but, except in the case of high genus, it was not known until now that they arise as blowups.”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites [KKM18,Ngu14]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [BNS24]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Findingnone given

    The surface M in this proof is disjoint from Q and intersects Z in 4 points, whereas the corresponding M in [IW24] intersects every horizontal ray from the origin, and it intersects Z in a pair of points.

    “However, the surface M in that paper is quite different than the surface M in our proof of Theorem 12: the M in this proof is disjoint from Q and intersects Z in 4 points, whereas the corresponding M in [IW24] intersects every horizontal ray from the origin, and it intersects Z in a pair of points.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [IW24]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Findingreasoning

    Suppose for each g that Σ_g ∈ S_g. Then Σ_g converges as g → ∞ to the plane {z = 0} with multiplicity 3, and the convergence is smooth except along a circle C(r) in the plane {z = 0}.

    “Moreover, Σ g converges as g → ∞ to the plane {z = 0} with multiplicity 3, and the convergence is smooth except along a circle C(r) in the plane {z = 0}.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingnone given

    In the singularity of Theorem 40, every shrinker Σ_g is G_g-invariant and has genus g.

    “(1) Σ g is G g -invariant and has genus g.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    Under the hypotheses of Lemma 60, E_g := U_g ∪ Γ_g lies in {z > 0} and is properly embedded in R^3.

    “Then E g := U g ∪ Γ g lies in {z > 0} and is properly embedded in R 3 .”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    In Theorem 61, one of the following holds: (1) Σ is the plane {z = 0} with multiplicity 3, and Σ_g has genus 2g for all sufficiently large g; (2) Σ is the plane {z = 0} together with the sphere ∂B(0, 2), each with multiplicity 1; (3) Σ is the plane {z = 0} together with the cylinder {x^2 + y^2 = 2}, each with multiplicity 1.

    “Furthermore, one of the following holds: (1) Σ is the plane {z = 0} with multiplicity 3, and Σ g has genus 2g for all sufficiently large g.(2) Σ is the plane {z = 0} together with the sphere ∂B(0, 2), each with multiplicity 1. (3) Σ is the plane {z = 0} together with the cylinder {x 2 + y 2 = 2}, each with multiplicity 1.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    In the proof of Theorem 61, one of the following holds: (1) Σ+ is empty; (2) Σ+ is a disk bounded by a circle C(r) in the plane {z = 0}; (3) Σ+ is an annulus whose boundary is a circle C(r) in the plane {z = 0}.

    “Claim 1.One of the following holds: (1) Σ + is empty. (2) Σ + is a disk bounded by a circle C(r) in the plane {z = 0}. (3) Σ + is an annulus whose boundary is a circle C(r) in the plane {z = 0}.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    After passing to a subsequence, the Σ_g converge to a shrinker Σ, and the convergence is smooth away from a circle C(r) := {(x, y, 0) : x^2 + y^2 = r^2}.

    “Theorem 61.After passing to a subsequence, the Σ g converge to a shrinker Σ, and the convergence is smooth away from a circle C(r) := {(x, y, 0) : x 2 + y 2 = r 2 }.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    Under the hypotheses of Theorem 23, if Σ is a shrinker at the origin at time T_Q to the flow M(•), then Σ has genus g and is disjoint from Q.

    “If Σ is a shrinker at the origin at time T Q to the flow M (•), then Σ has genus g and is disjoint from Q.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    In the singularity of Theorem 43, every shrinker Σ_g is G_g-invariant and satisfies Σ_g ∩ {z = 0} = Q_g.

    “(2) Σ is G g -invariant and Σ g ∩ {z = 0} = Q g .”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingreasoning

    If M ∈ M′′_g and t < T_pos(M) is a regular time, then M(t) ∈ M_g and the number of points in M(t) ∩ Z_+ is even.

    “Suppose that M ∈ M ′′ g .If t < T pos (M ) is a regular time, then M (t) ∈ M g , and the number of points in M (t) ∩ Z + is even.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S1 (asserted).

    Premisecitation

    For 2-dimensional shrinkers in R3, the Lee-Zhao hypothesis is that no end of the shrinker is asymptotically cylindrical.

    “(For 2-dimensional shrinkers in R 3 , the only case considered in this paper, the hypothesis is that no end of the shrinker is asymptotically cylindrical.)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [LZ24]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    In the regular n-gon case, Theorem 12 was proved by Ilmanen and White by similar methods to those in this paper.

    “In the regular n-gon case (Case [ii]), Theorem 12 was proved, by similar methods to those in this paper, by Ilmanen and White [IW24].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [IW24]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    In the cases of the cube and the octahedron, shrinkers as in Theorem 12 were found numerically by Chopp, and they are compact.

    “In the cases of the cube and the octahedron, shrinkers as in Theorem 12 were found numerically by Chopp [Cho94], and they are compact.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [Cho94]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    Under the Local Regularity Theorem hypotheses, for large n, M_n(t) ∩ U is smooth and converges smoothly to M(t) for all t ∈ [a',b].

    “Then, for large n, M n (t) ∩ U is smooth and converges smoothly to M (t) for all t ∈ [a ′ , b].”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites [Whi05]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • May combine two claims.

    Premisecitation

    Ketover proved existence of shrinkers with properties (1) and (2) in [Ket16] for the Platonic case and in [Ket24] for the regular polygon case.

    “Using minimax methods, Ketover proved existence of shrinkers with properties (1) and (2) in [Ket16] (the Platonic case (i)) and in [Ket24] (the regular polygon case (ii)).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites [Ket16]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites [Ket24]

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

20 of 20
  • The paper relies on this without support.

    The area of the evolving surface M(t) is a decreasing function of t.

    “and the fact that area(M (t)) is a decreasing function of t.”
  • The paper relies on this without support.

    The singularities at O and O_∞ at time T have the same genus, namely genus(Σ).

    “Because the singularities at O and O ∞ at time T have the same genus, namely, genus(Σ), 2 genus(Σ) ≤ genus(M ).”
  • The paper relies on this without support.

    A shrinker M is connected.

    “Since M is a shrinker, it is connected.”
  • The paper relies on this without support.

    Smooth graphs never develop singularities.

    “But smooth graphs never develop singularities, and thus T pos (M n ) = ∞ for such n.”
  • The paper relies on this without support.

    Every smooth shrinker without boundary is an unstable minimal surface with respect to the shrinker metric.

    “But then Σ ′ is unstable, a contradiction.(Every smooth shrinker without boundary is an unstable minimal surface with respect to the shrinker metric.)”
  • The paper relies on this without support.

    The family F is connected.

    “Note that F is a connected family.”
  • The paper relies on this without support.

    The set S of all shrinkers at (p, T) is connected with respect to smooth convergence.

    “The set S is connected (with respect to smooth convergence), so it suffices to show that if Σ i ∈ S converges smoothly to Σ ∈ S , then genus(Σ i ) = genus(Σ) for all sufficiently large i.”
7 of 7

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • A shrinker flow is given the opposite scaling in Lemma 58

    Minor

    The paper defines the flow generated by a shrinker using |t|^{1/2}Σ, but Lemma 58 says its flow is |t|^{-1/2}Σ. These are different time-dependent surfaces and cannot both be the stated shrinker flow.

  • Appendix A reverses the tangent-flow rescaling

    Minor

    Lemma 66 defines shrinkers as limits under multiplication by (T−t)^{1/2}, whereas the preliminaries define their blowup limits using (T−t)^{-1/2}. The appendix therefore states a different limit for the collection whose genus it analyzes.

  • The orientation changes sign despite the asserted continuity

    Minor

    The proof of Theorem 59 assigns e3 to the orientation at time zero at every nonorigin point of the plane, then assigns −e3 at those same points for every positive time, explicitly invoking continuity for both statements.

  • The defined Δ0 reaches radius two, not the unit disk

    Minor

    Setting R=0 in the displayed definition of Δ_R leaves wedge regions extending to radius 2. The subsequent description places Δ0 and its boundary in the unit disk and unit circle.

4 of 4

Cites

Flagged first, then by citations

7 of the 33 references couldn’t be matched to a known paper.

Cited by

Still being checked.

Authors
Hoffman, D., Martín, F., & White, B.
Affiliations
  • Stanford University (US)
  • Universidad de Granada (ES)
Corresponding author
Martín, F.
Venue
Mathematische Annalen
Published
2026-10-02
Volume
396
Issue
3
ISSN
  • 0025-5831 (ISSN-L)
  • 1432-1807
OpenAlex
W4416151933
Cited by
0
References
33
FWCI
0.00
Type
journal article
Language
en
Open access
Locations
  • Mathematische Annalen (publishedVersion, cc-by)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
Agencia Estatal de Investigación (PID2024-156031NB-I00)
Topics
  • Physical Sciences › Mathematics › Applied Mathematics › Geometric Analysis and Curvature Flows
  • Physical Sciences › Physics and Astronomy › Astronomy and Astrophysics › Advanced Differential Geometry Research
  • Physical Sciences › Mathematics › Applied Mathematics › Nonlinear Partial Differential Equations
Keywords
mean curvature flow, singularities, blow-up

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-11-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-11-06.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review