Skip to main content
Log in

Singular points of Hölder asymptotically optimally doubling measures

  • Published:
Calculus of Variations and Partial Differential Equations Aims and scope Submit manuscript

Abstract

We consider the question of how the doubling characteristic of a measure determines the regularity of its support. The question was considered by David et al. (Commun Pure Appl Math 54:385–449, 2001) for codimension 1 under a crucial assumption of flatness, and later by Preiss et al. (Calc Var PDE’s 35:339–363, 2009) in higher codimension. However, these studies leave open the question of what can be said about the geometry of the support of such measures in a neighborhood about a nonflat point. Here we answer the question for codimension-1 Hölder doubling measures in \(\mathbb {R}^4\) by constructing parametrizations in a neighborhood of a nonflat point by the Kowalski–Preiss cone. These parametrizations extend to be \(C^{1,\beta }\) diffeomorphisms on an open set. Some of our parametrization techniques build on ideas from: David et al. (Mem AMS 215, 2012; Commun Pure Appl Math 54:385–449, 2001), Preiss et al. (Calc Var PDE’s 35:339–363, 2009) and Taylor (Annals Math 103:489–539, 1976).

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7

We’re sorry, something doesn't seem to be working properly.

Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.

References

  1. Badger, M.: Flat points in zero sets of harmonic polynomials and harmonic measure from two sides. J. Lond. Math. Soc. 87(1), 111–137 (2013)

    Article  MATH  MathSciNet  Google Scholar 

  2. Badger, M.: Harmonic polynomials and tangent measures of harmonic measure. Rev. Mat. Iberoam. 27(3), 841–870 (2011)

    Article  MATH  MathSciNet  Google Scholar 

  3. Chatelin, François: Eigenvalues of Matrices. Wiley, West Sussex (1993)

    MATH  Google Scholar 

  4. Chousionis, V., Garnett, J., Le, T., Tolsa, X.: Square functions and uniform rectifiability. Preprint. arXiv:1401.3382

  5. David, G., Kenig, C.E., Toro, T.: Asymptotically optimally doubling measures and Reifenberg flat sets with vanishing constant. Commun. Pure Appl. Math. 54, 385–449 (2001)

    Article  MATH  MathSciNet  Google Scholar 

  6. David, G., DePauw, T., Toro, T.: A generalization of Reifenberg’s theorem in \({\mathbb{R}}^{3}\). GAFA 18, 1168–1235 (2008)

    MATH  MathSciNet  Google Scholar 

  7. Kenig, C.E., Toro, T.: Free boundary regularity for harmonic measures and Poisson kernels. Ann. Math. (2) 150(2), 369–454 (1999)

  8. Kowalski, O., Preiss, D.: Besicovitch-type properties of measures and submanifolds. J. Reine Angew. Math. 379, 115–151 (1987)

    MATH  MathSciNet  Google Scholar 

  9. Mattila, P.: Geometry of sets and measures in Euclidean space, fractals and rectifiability. In: Cambridge Studies in Advanced Mathematics, vol. 44. Cambrdige University Press, Cambridge (1995)

  10. Preiss, D.: Geometry of measures in \({\mathbb{R}}^{n}\): distribution, recifiability, and densities. Ann. Math. 125, 537–643 (1987)

    Article  MATH  MathSciNet  Google Scholar 

  11. Preiss, D., Tolsa, X., Toro, T.: On the smoothness of Hölder doubling measures. Calc. Var. PDE’s 35, 339–363 (2009)

    Article  MATH  MathSciNet  Google Scholar 

  12. Taylor, J.: The structure of singularities in soap-bubble-like and soap-film-like minimal surfaces. Ann. Math. 103, 489–539 (1976)

    Article  MATH  Google Scholar 

  13. Tolsa, X.: Uniform measures and uniform rectifiability. arXiv:1310.0658

Download references

Acknowledgments

The author was partially supported by NSF DMS 0838212 during this research.

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Stephen Lewis.

Additional information

Communicated by L. Simon.

Appendix

Appendix

Lemma 29

Suppose that \(U,V\subseteq \mathbb {R}^n\) are open sets, \(\Gamma \subseteq \mathbb {R}^n\), \(0<m<n\), and \(\psi \in C^2(U,V)\) is bijective and satisfies

$$\begin{aligned} 0<\lambda \leqslant \frac{|\psi (x)-\psi (y)|}{|x-y|}\leqslant \Lambda \quad \text{ for } \text{ all } x,y\in U \quad \text{ and } \quad ||D^2\psi ^a||_\infty = \sup _{x\in U}||D_x^2\psi ||<\infty . \end{aligned}$$
(298)

Let \(z\in \Gamma \cap U\), \(B(z,r)\subseteq U\), P be a plane of dimension m through z, and set \(\widetilde{P} = D_z\psi (P-z) + \psi (z)\), \(\widetilde{\Gamma }= \psi (\Gamma )\). Then

$$\begin{aligned} d^{\psi (z),\lambda r}\left( \widetilde{\Gamma },\widetilde{P}\right) \leqslant \frac{||D^2 \psi ||_\infty }{2\lambda }r + \frac{\Lambda }{\lambda }d^{z,r}\left( \Gamma ,P\right) . \end{aligned}$$
(299)

Proof

Without loss of generality, take \(z = \psi (z) = 0\). Let \(P\subseteq \mathbb {R}^n\) be an m-plane through 0 and set \(d = d^{0,r}\left( \Gamma ,P\right) .\) Note that \(\lambda \leqslant \frac{|D_0\psi (v) |}{|v|} \leqslant \Lambda \) for all \(v\in \mathbb {R}^n {\setminus } \{0\}\) by (298). Further, note that we have

$$\begin{aligned} B(0,\lambda r)\subseteq \psi (B(0,r)). \end{aligned}$$
(300)

Let \(y\in \widetilde{\Gamma }\cap B(0,\lambda r)\). Then by (300) and bijectivity, we have that there exists \(x\in B(0,r)\cap \Gamma \) such that \(y=\psi (x)\). By \(d = d^{0,r}\left( \Gamma ,P\right) \), we get that there exists \(p\in P\) such that \(|p-x|\leqslant rd.\) Let \(\widetilde{p} = D_0\psi (p)\in \widetilde{P}\). We compute

$$\begin{aligned} |\widetilde{p}-y|&= |D_0\psi (p) - \psi (x)|\leqslant |D_0\psi (p) - \psi (p)| + |\psi (p) - \psi (x)|. \end{aligned}$$
(301)

By (298) and Taylor expansion, we get that

$$\begin{aligned} |\widetilde{p}-y|\leqslant \frac{||D^2\psi ||_\infty }{2}|p-0|^2 + \Lambda |p-x|\leqslant \frac{||D^2\psi ||_\infty }{2}r^2 + \Lambda rd. \end{aligned}$$
(302)

Because for every \(y\in B(0,\lambda r)\cap \widetilde{\Gamma }\), there exists \(\widetilde{p}\in \widetilde{P}\) satisfying (302), we get that

$$\begin{aligned} \widetilde{d}^{0,\lambda r}\left( \widetilde{\Gamma },\widetilde{P}\right) \leqslant \frac{||D^2\psi ||_\infty }{2\lambda }r + \frac{\Lambda }{\lambda } d . \end{aligned}$$
(303)

By Lemma 2, we get that (303) tells us (299). \(\square \)

Lemma 30

For \(a\in \mathcal {C}{\setminus } \{0\}\), and \(A>0\) large enough, there exists a neighborhood \(U\supseteq B(a, 2|a|/{A})\), \(V\subseteq \mathbb {R}^3\) open, \(I\ni 0\) an open interval, and a smooth coordinate map \(\psi ^a: U\rightarrow V\times I\) such that \(V\times \{0\} = \psi ^a(\mathcal {C}\cap U)\) and \(\widetilde{\pi }= \psi ^a\circ \pi \circ (\psi ^a)^{-1}\) is orthogonal projection onto \(\mathbb {R}^3\times \{0\}\) (where \(\pi \) is the same map defined in Sect. 5; see (189)). Further, \(\psi ^a\) satisfies the estimates

$$\begin{aligned} \frac{1}{2}\leqslant \frac{|\psi ^a(x)-\psi ^a(y)|}{|x-y|}\leqslant 2\quad \text{ for } \text{ all } \,\,x,y\in U \end{aligned}$$
(304)

and

$$\begin{aligned} ||D^2\psi ^a||_\infty = \sup _{x\in U} ||D^2_x\psi ^a|| \leqslant \frac{C}{|a|} \end{aligned}$$
(305)

for some C independent of a.

Proof

First, we fix an \(a\in \mathcal {C}\), \(|a| = 1\). We define \(\psi ^a\) by defining its inverse. Choose orthonormal coordinates \((z_1, z_2, z_3)\) on \(T_a\mathcal {C}\) centered at a. Let p be orthogonal projection from \(\mathcal {C}\) onto \(T_a\mathcal {C}\), and take \(U'\supseteq B(a, 8/A)\) to be an open set where \(p^{-1}\) is defined. Identifying \(T_a\mathcal {C}\) with \(\mathbb {R}^3\) under the z coordinates, let \(V = U'\cap T_a\mathcal {C}\). Let \(I = (-8/A,8/A)\). Define for \((z,t)\in V\times I\)

$$\begin{aligned} (\psi ^a)^{-1}(z,t) = p^{-1}(z) + t\eta _{p^{-1}(z)}. \end{aligned}$$
(306)

Assume that \(A\geqslant 16\), so that \(8/A\leqslant 1/2\). Note that \(\psi ^a\) is bijective onto \(U = (\psi ^a)^{-1}(V\times I)\), because \(\eta \) is a smooth vector field (away from the \(x_4\) axis), and the point (zt) is the flow after time t of the point \(p^{-1}z\) along the integral curves of \(\eta \). Further, the same comments imply that it is smooth. Then we note that because \(p^{-1}(z)\in \mathcal {C}\), \(V\times \{0\} = \psi ^a(\mathcal {C}\cap U)\). Further, \(\widetilde{\pi }(z,t) = \psi ^a(\pi ( p^{-1}(z) + t\eta _{p^{-1}(z)})) = \psi ^a(p^{-1}(z)) = z\), and so \(\widetilde{\pi }\) is orthogonal projection onto \(V\times \{0\}\).

We now show that (304) holds for \(\psi ^a\) as long as \(U'\) is chosen small enough and 1 / A is chosen small enough. Continuing the identifcation of \(T_a\mathcal {C}\) with \(\mathbb {R}^3\), we set \(e_i\) to be the coordinate vector of \(z_i\), and note that \(D_a\psi ^a\) is the map

$$\begin{aligned} D_a\psi ^a(e_i) = e_i,\quad D_a\psi ^a \eta _a = e_4\quad \text{ for } \, i =1,2,3. \end{aligned}$$
(307)

Because the \(z_i\) are orthonormal and \(\measuredangle (\eta _a, T_a\mathcal {C}) = \pi /4\), we get that

$$\begin{aligned} \langle e_i, e_j\rangle = \delta _{ij}, \quad |\langle e_i, \eta _a\rangle |\leqslant 1/{\sqrt{2}}. \end{aligned}$$
(308)

From (307) and (308), as well as recalling that \(|\nu _a| = 1\), we get that

$$\begin{aligned} \frac{1}{\sqrt{2}} \leqslant \frac{|D_a\psi ^a v|}{|v|}\leqslant \sqrt{2}\quad \text{ for } \, v\in \mathbb {R}^4 {\setminus } \{ 0\}. \end{aligned}$$
(309)

Because \(\psi ^a\) is smooth, it follows from (309) that for \(U'\) small enough,

$$\begin{aligned} \frac{1}{2}\leqslant \frac{|\psi ^a(x) - \psi ^a(y)|}{|x-y|} \leqslant 2 \quad \text{ for } x,y\in U'. \end{aligned}$$
(310)

Thus if A is large enough, \(B(a, 8/A)\subseteq U'\). By definition of U and (310), we have that \(B(a, 2/A)\subseteq U\). Because \(\psi \) is \(C^2\), by potentially restricting to a compactly contained open set, we may assume

$$\begin{aligned} C = ||D^2\psi ^a||_\infty < \infty . \end{aligned}$$
(311)

Let \(b\in \mathcal {C}\), \(|b| = 1\). Then there is a rotation \(O\in O(4)\) taking b to a and fixing \(\mathcal {C}\). Define \(\psi ^b = O^{-1}\circ \psi ^a\circ O\). For \(b\in \mathcal {C}{\setminus } \{0\}\), we define \(\psi ^b = |b|\psi ^{\frac{b}{|b|}}(\cdot /{|b|})\). Note that \(\psi ^{\frac{b}{|b|}}\) satisfies (310) and (311), we have that \(\psi ^b\) satisfies (304) and (305). \(\square \)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Lewis, S. Singular points of Hölder asymptotically optimally doubling measures. Calc. Var. 54, 3667–3713 (2015). https://doi.org/10.1007/s00526-015-0918-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00526-015-0918-y

Mathematics Subject Classification

Navigation