research36 min studyLesson 840 in Mathematics

Minimal surfaces and regularity thresholds

Separate mass minimization from stationarity and track hypersurface dimension at the first singular cone.

On this page

What you’ll understand

  • Distinguish stationary, stable, and locally mass-minimizing geometric objects.
  • State codimension-one interior regularity in both hypersurface and ambient dimensions.
  • Compute a singular cone’s mass ratio without treating that computation as a proof of minimality.

Weak minimizers solve an existence problem before they solve a smoothness problem. Regularity theory then asks where their supports become smooth hypersurfaces and how large the exceptional singular set can be. The sharp codimension-one threshold uses the dimension of the hypersurface, not the ambient dimension. The Simons cone lies exactly at the first dimension where a singular minimizing example is possible, and its link geometry permits an explicit density computation.

Stationarity is weaker than minimization

An interior stationary surface has zero first variation under compactly supported deformations, corresponding to zero mean curvature on a smooth piece. Stability adds nonnegative second variation, while local mass minimization compares with every admissible compactly supported integral-current change having the same boundary. Minimization implies stability and stationarity where those variations make sense, but the reverse implications are not general. A critical-point equation alone does not certify an absolute minimum.

Calibration gives one direct certificate for special minimizers. A closed k-form with comass at most one bounds the mass of every oriented competitor. If its integral equals the mass of a candidate and is unchanged under the allowed boundary-preserving comparison, the candidate minimizes. Planar patches with their standard orientation provide a concrete example. The argument uses orientation and a declared competitor class, rather than only the visual flatness of the surface.

dω=0,∥ω∥comass≤1,T(ω)=M(T)⟹M(S)≥M(T) for the stated homologous competitors.d\omega=0,\quad\|\omega\|_{\mathrm{comass}}\le1,\quad T(\omega)=\mathbf M(T)\quad\Longrightarrow\quad \mathbf M(S)\ge\mathbf M(T)\text{ for the stated homologous competitors}.
Calibration needs equality of the form integrals for competitors, justified by boundary or homology conditions and Stokes.↔ Scroll wide equations horizontally.

Interior hypersurface regularity has a seven-dimensional threshold

Let T be a locally mass-minimizing integral m-current in Euclidean (m+1)(m+1)-space. Away from its boundary support, its singular set is empty for m at most six, locally discrete for m equal to seven, and has Hausdorff dimension at most m minus seven for larger m. The regular part is a smooth embedded hypersurface, with locally constant integer multiplicity on its sheets. These are interior results; boundary regularity requires additional assumptions and is not included here.

Equivalently, the ambient threshold is seven dimensions for guaranteed interior smoothness, with the first isolated singular minimizing hypersurface possible in ambient dimension eight. Higher codimension does not inherit this threshold: integral area minimizers there have a different regularity theory and may have lower-dimensional branching phenomena. Nor should the statement be transferred to arbitrary stationary varifolds, nonintegral currents, or anisotropic energies without checking the relevant theorem.

m≤6: sing⁡T=∅;m=7: sing⁡T locally discrete;m≥8: dim⁡Hsing⁡T≤m−7.m\le6:\ \operatorname{sing}T=\varnothing;\qquad m=7:\ \operatorname{sing}T\text{ locally discrete};\qquad m\ge8:\ \dim_H\operatorname{sing}T\le m-7.
Here m is the hypersurface dimension and the Euclidean ambient dimension is m+1; the statement is away from spt(∂T).↔ Scroll wide equations horizontally.

Tangent cones and densities expose a singular model

For minimizing currents, rescaling around an interior point produces tangent cones along suitable subsequences. Monotonicity controls normalized mass ratios in boundary-free balls. At a regular point the ratio tends to the local integer multiplicity; a multiplicity-one plane has density one. A density larger than one does not by itself prove singularity, because a smooth sheet with larger multiplicity also has that density. The geometry of the cone must still be considered.

The classical cone {∣x∣=∣y∣}⊂R4×R4\{|x|=|y|\}\subset\mathbb R^4\times\mathbb R^4 is the boundary of {∣x∣<∣y∣}\{|x|<|y|\}. The Bombieri–De Giorgi–Giusti theorem makes it a locally minimizing integral hypersurface, not merely a stationary cone. It is smooth away from the vertex and nonplanar, so it cannot be smooth at its conical vertex: a cone smooth there would equal its tangent plane. Its minimizing property is a substantial theorem; the calculation below only checks its dimension and normalized mass.

Θm(∥T∥,0)=lim⁡r↓0M(T⌞Br)ωmrm.\Theta^m(\|T\|,0)=\lim_{r\downarrow0}\frac{\mathbf M(T\llcorner B_r)}{\omega_mr^m}.
Density measures normalized mass and includes multiplicity; it is not an unconditional test for smoothness.↔ Scroll wide equations horizontally.

The seven-dimensional cone has density 15π/32

For the Simons cone in R8\mathbb R^8, intersect with the unit sphere and use its link S3(1/2)×S3(1/2)S^3(1/\sqrt2)\times S^3(1/\sqrt2). Compute the link area, cone mass in the unit ball, and normalized seven-dimensional density. Use H3(S3(1))=2π2\mathcal H^3(S^3(1))=2\pi^2.

  1. Each radius-1/21/\sqrt2 three-sphere has area 2π2(1/2)32\pi^2(1/\sqrt2)^3. The product link has dimension six and area 4π4(1/2)6=π4/24\pi^4(1/\sqrt2)^6=\pi^4/2. The radial cone therefore has dimension seven, in ambient dimension eight.

  2. The cone volume element contributes radial factor r6drr^6dr, so integrating to radius one divides link area by seven. The unit-ball cone mass is π4/14\pi^4/14. The normalized Euclidean seven-ball volume is ω7=16π3/105\omega_7=16\pi^3/105.

  3. Divide these quantities to get 15π/32≈1.47262215\pi/32\approx1.472622. The ratio is independent of radius because the object is a cone. Its isolated vertex singularity agrees with the m-equal-seven theorem; neither this ratio nor the link-area calculation proves the separate minimizing theorem.

π4/1416π3/105=15π32>1.\frac{\pi^4/14}{16\pi^3/105}=\frac{15\pi}{32}>1.
The denominator is normalized seven-dimensional Euclidean ball volume, not ambient eight-dimensional volume.↔ Scroll wide equations horizontally.
The result

The cone is a seven-dimensional hypersurface in R8, with unit-ball mass π⁴/14 and density 15π/32. Its vertex realizes the first allowed singular dimension.

Explore the model

Interactive lab

Interactive lab · matrix

Calibrate affine planar patches before discussing singular cones

Pause and Reset, then set scale a to 1 and shear b to 1. The planar image of the unit square has area one, equal to its integral of the standard oriented area form. For a fresh comparison Reset and set a to 2 and b to 0; both quantities become two. These flat positive-orientation patches illustrate a calibration equality. The instrument does not display the seven-dimensional cone or prove a regularity theorem. Editing preserves time; set both parameters after Reset for each comparison. For each run, take eight Steps to reach t=4 seconds, when the drawn geometry equals the selected matrix; its readouts describe that selected matrix throughout.

Calibrate affine planar patches before discussing singular conesThe selected matrix is A = [[a,b],[0,1]]. Playback interpolates geometry between the identity and A; readouts always describe A. Its determinant is a and its singular values come from AᵀA. Positive a makes every supported matrix invertible. This two-dimensional family is not a general eigensolver or a learned network. Current readings: Area scale · det A: 1.4; Singular values · σmax, σmin: 1.597, 0.876; Condition number · κ₂(A): 1.823. All parameters are adjustable with the labeled controls below.Matrix A with first row 1.4, 0.6 and second row zero, one; display interpolates between identity and A while final determinant remains 1.4.−4−2024−2−1012x · equal geometric scaleyA = [ 1.4 0.6 ; 0 1 ]Display blend: 0% · final values below
Interpolated geometryOriginal unit squareSelected final transform
Area scale · det A
1.4
Singular values · σmax, σmin
1.597, 0.876
Condition number · κ₂(A)
1.823
t = 0 s

The selected matrix is A = [[a,b],[0,1]]. Playback interpolates geometry between the identity and A; readouts always describe A. Its determinant is a and its singular values come from AᵀA. Positive a makes every supported matrix invertible. This two-dimensional family is not a general eigensolver or a learned network. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forA finite planar calibration check and a high-dimensional singularity theorem have different scopes.

Check your understanding

Which statement correctly interprets the codimension-one interior regularity threshold?

Try it yourself

Use dx∧dydx\wedge dy to certify a radius-two disk in the horizontal plane of three-space against compactly supported oriented integral competitors with the same boundary. Then state the maximum singular-set dimension allowed for minimizing hypersurfaces of dimensions six and nine.

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

The planar disk saturates the comass-one area form, which is exact in Euclidean space. Keep hypersurface dimension separate from ambient dimension.

Compare with the solution

The disk mass and its integral of dx∧dydx\wedge dy are both 4π4\pi. A compactly supported competitor with the same boundary has the same integral by Stokes, using the primitive x dyx\,dy with a suitable cutoff. Its mass is at least that integral, proving minimality. Dimension six in ambient R7 has no interior singularities. Dimension nine in ambient R10 permits a singular set of Hausdorff dimension at most two; this is a bound, not an existence assertion for every minimizer.

Take it with you

  • Mass minimization is stronger than stationarity or stability.
  • The first singular minimizing hypersurface has dimension seven in ambient dimension eight.
  • Density and tangent-cone calculations require dimension and multiplicity bookkeeping.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.