Effective versions of the positive mass theorem

Effective versions of the positive mass theorem

Alessandro Carlotto ETH Institute for Theoretical Studies, Clausiusstrasse 47, 8092 Zurich, Switzerland Otis Chodosh DPMMS, University of Cambridge, Wilberforce Road, Cambridge, CB3 0WB, United Kingdom  and  Michael Eichmair Faculty of Mathematics, University of Vienna, Oskar-Morgenstern-Platz 1, 1090 Vienna, Austria
Abstract.

The study of stable minimal surfaces in Riemannian -manifolds with non-negative scalar curvature has a rich history. In this paper, we prove rigidity of such surfaces when is asymptotically flat and has horizon boundary. As a consequence, we obtain an effective version of the positive mass theorem in terms of isoperimetric or, more generally, closed volume-preserving stable CMC surfaces that is appealing from both a physical and a purely geometric point of view. We also include a proof of the following conjecture of R. Schoen: An asymptotically flat Riemannian -manifold with non-negative scalar curvature that contains an unbounded area-minimizing surface is isometric to flat .

1. Introduction

The geometry of stable minimal and volume-preserving stable constant mean curvature surfaces in asymptotically flat -manifolds with non-negative scalar curvature is witness to the physical properties of the space-times containing such as maximal Cauchy hypersurfaces; see e.g. [Penrose:1965, Schoen-Yau:PMT1, Christodoulou-Yau:1988, Huisken-Yau:1996, Bray:1997, Bray:2001, Huisken-Ilmanen:2001]. The transition from positive to non-negative scalar curvature of , which is so crucial for physical applications, is a particularly delicate aspect in the analysis of such surfaces. Here we establish optimal rigidity results in this context that apply very generally. We apply them to obtain a precise understanding of the behavior of large isoperimetric or, more generally, closed volume-preserving stable constant mean curvature surfaces in that extends the results of S. Brendle, J. Metzger, and the third-named author [stableCMC, isostructure, hdiso, offcenter]. In combination with existing literature, this yields a rather complete analogy between the picture in and classical results in Euclidean space.

We review the standard terminology and conventions that we use here in Appendix A. In particular, we follow the convention that stable minimal surfaces are by definition two-sided.

To provide context, we recall a celebrated application of the second variation of area formula due to R. Schoen and S.-T. Yau [Schoen-Yau:1978, Theorem 6.1]. Assume (for contradiction) that we are given a metric of positive scalar curvature on the -torus . Using results from geometric measure theory, one can find a closed surface of non-zero genus that minimizes area in its homology class with respect to this metric. In particular, is a stable minimal surface. Using the function in the stability inequality (17), we obtain that

 0≥∫Σ|h|2+Ric(ν,ν).

We may rewrite the integrand as

 |h|2+Ric(ν,ν)=12(|h|2+R)−K,

using the Gauss equation (18). It follows that

 ∫ΣK>0

which is incompatible with the Gauss-Bonnet formula. Thus does not admit a metric of positive scalar curvature.

This crucial mechanism — positive ambient scalar curvature is incompatible with the existence of stable minimal surfaces of most topological types — is at the heart of another fundamental result proven by R. Schoen and S.-T. Yau, the positive mass theorem [Schoen-Yau:PMT1]: If is asymptotically flat with horizon boundary and non-negative integrable scalar curvature, then its ADM-mass is non-negative. Moreover, the ADM-mass vanishes if and only if is isometric to Euclidean space. Using an initial perturbation, they reduce the proof of non-negativity of the ADM-mass to the special case where is asymptotic to Schwarzschild with horizon boundary and positive scalar curvature. If the mass is negative, then the coordinate planes with respect to the chart at infinity act as barriers for area minimization in the slab-like region they enclose in , provided is sufficiently large. Using geometric measure theory, one finds an unbounded complete area-minimizing boundary in this slab. Such a surface has quadratic area growth. Using the logarithmic cut-off trick in the second variation of area (observing the decay of the ambient Ricci curvature to handle integrability issues), it follows as before that

 0<12∫Σ|h|2+R=∫ΣK.

A result of S. Cohn-Vossen shows that . Using that is area-minimizing in a slab, they argue that is asymptotic to a horizontal plane and conclude that the geodesic curvature of the circles in converges to as .111An alternative argument for this part of the proof that also generalizes to stable minimal surfaces with quadratic area growth was given in [stableCMC, Proposition 3.6]. The strategy of [stableCMC] is exploited in the proof of Theorem 1.3 below. The Gauss-Bonnet formula gives that

 ∫ΣK=0,

Observe that this line of reasoning cannot establish the rigidity part (only Euclidean space has vanishing mass) of the positive mass theorem. Conversely, a beautiful idea of J. Lohkamp [Lohkamp:1999, Section 6] shows that the rigidity assertion of the positive mass theorem implies the non-negativity of ADM-mass in general. Indeed, he shows that it suffices to prove that an asymptotically flat Riemannian -manifold with horizon boundary and non-negative scalar curvature is flat if it is flat outside of a compact set.

The ideas of R. Schoen and S.-T. Yau described above are instrumental to our results here. We record the following technical variation on their work as a precursor of Theorems 1.2 and 1.3 below.

Proposition 1.1 (Section 6 in [isostructure]).

Let be an asymptotically flat Riemannian -manifold. Assume that is the unbounded component of an area-minimizing boundary in , and that the scalar curvature of is non-negative along . Then is totally geodesic and the scalar curvature of vanishes along this surface. Moreover, for all sufficiently large, intersects transversely in a nearly equatorial circle. The Gauss curvature is integrable and .

We also mention that other proofs of the positive mass theorem (including that of E. Witten [Witten:1981] based on the Bochner formula for harmonic spinors and that of G. Huisken and T. Ilmanen [Huisken-Ilmanen:2001] based on inverse mean curvature flow) have been given.

The discoveries of R. Schoen and S.-T. Yau have incited a remarkable surge of activity investigating the relationship between scalar curvature, locally area-minimizing (or stable minimal) surfaces, and the physical properties of spacetimes evolving from asymptotically flat Riemannian -manifolds according to the Einstein equations. This has lead to spectacular developments in geometry and physics. We refer the reader to [Fischer-Colbrie-Schoen, Schoen:1984, Fischer-Colbrie:1985, Anderson-Rodriguez:1989, Corvino:2000, Bray:2001, Huisken-Ilmanen:2001] to gain an impression of the wealth and breadth of the repercussions.

The following rigidity result for scalar curvature was first proven by the first-named author under the additional assumption of quadratic area growth for the surface . Subsequently, the quadratic area growth assumption was removed independently (in the form of Theorem 1.2 below) by the first-named author [Carlotto] and (in the form of Theorem 1.3 below) in a joint project of the second- and third-named authors. The proof of Theorem 1.3 is included in this paper.

Theorem 1.2 ([Carlotto]).

Let be an asymptotically flat Riemannian -manifold with non-negative scalar curvature. Let be a non-compact properly embedded stable minimal surface. Then is a totally geodesic flat plane and the ambient scalar curvature vanishes along . Such a surface cannot exist under the additional assumption that is asymptotic to Schwarzschild with mass .

Theorem 1.3.

Let be a Riemannian -manifold with non-negative scalar curvature that is asymptotic to Schwarzschild with mass and which has horizon boundary. Every complete stable minimal immersion that is proper is an embedding of a component of the horizon.

To obtain these results, it is necessary to understand how non-negative scalar curvature keeps in check the a priori wild behavior at infinity of the minimal surface. This difficulty does not arise in the original argument by R. Schoen and S.-T. Yau. The proofs of Theorems 1.2 and 1.3 use properness in a crucial way. Moreover, the embeddedness assumption is essential in the derivation of Theorem 1.2 in [Carlotto].

In spite of their geometric appeal, we cannot apply Theorems 1.2 and 1.3 to prove effective versions of the positive mass theorem such as Theorem 1.10 below. This is intimately related to the fact that properness is not preserved by convergence of immersions. Our first main contribution here is the following technical result that rectifies this:

Theorem 1.4.

Let be an asymptotically flat Riemannian -manifold with non-negative scalar curvature. Assume that there is an unbounded complete stable minimal immersion that does not cross itself. Then admits a complete non-compact properly embedded stable minimal surface.

Using this, we obtain the following substantial improvement of Theorems 1.2 and 1.3:

Theorem 1.5.

Let be a Riemannian -manifold with non-negative scalar curvature that is asymptotic to Schwarzschild with mass and which has horizon boundary. The only non-trivial complete stable minimal immersions that do not cross themselves are embeddings of components of the horizon.

For the proof of Theorem 1.4, we develop in Section 4 a general procedure of extracting properly embedded top sheets from unbounded complete stable minimal immersions that do not cross themselves. The method depends on a purely analytic stability result — Corollary C.2 — that restricts the topological type of complete stable minimal immersions into .

The proof of the positive mass theorem suggests the following conjecture [Schoen:talk, p. 48] of R. Schoen: An asymptotically flat Riemannian manifold with non-negative scalar curvature that contains an unbounded area-minimizing surface is isometric to Euclidean space. We include here a proof of this conjecture and that of a related rigidity result for slabs, both due to the second- and third-named authors:

Theorem 1.6.

The only asymptotically flat Riemannian -manifold with non-negative scalar curvature that admits a non-compact area-minimizing boundary is flat .

We recall the precise meaning of area-minimizing boundaries in Appendix LABEL:sec:am.

Theorem 1.7.

Let be an asymptotically flat Riemannian -manifold with non-negative scalar curvature and with horizon boundary. Any two disjoint connected unbounded properly embedded complete minimal surfaces in bound a region that is isometric to a standard Euclidean slab .

The proofs of Theorem 1.6 and 1.7 are inspired by the recent refinement due to G. Liu [Liu:2013] of a strategy of M. Anderson and L. Rodríguez [Anderson-Rodriguez:1989] to prove rigidity results for complete manifolds with non-negative Ricci curvature.

We point that that we may excise the slab in the conclusion of Theorem 1.7 from to produce a new smooth asymptotically flat Riemannian -manifold with non-negative scalar curvature that contains a properly embedded totally geodesic flat plane along which the ambient scalar curvature vanishes.

For comparison, we recall the following consequence of a gluing result due to the first-named author and R. Schoen:

Theorem 1.8 ([Carlotto-Schoen]).

There exists an asymptotically flat Riemannian metric with non-negative scalar curvature and positive mass on such that on .

The coordinate planes with in Theorem 1.8 are stable minimal surfaces. In particular, the area-minimizing condition in Theorem 1.6 cannot be relaxed. We also see that the condition that be asymptotic to Schwarzschild in Theorem 1.5 is necessary.

There is a rich theory of rigidity results for (compact) minimal surfaces in Riemannian -manifolds with a lower scalar-curvature bound. We refer the reader to the papers [Cai-Galloway:2000, Bray-Brendle-Eichmair-Neves:2010, Bray-Brendle-Neves:2010, MarquesNeves:min-max-rigidity-3mflds, MaximoNunes:hawking-rigidity, Nunes:hyperbolic-rigidity, Ambrozio:free-bdry-rigidity, MicallefMoraru] for several recent results in this direction, and to the introductions of these papers for a complete overview.

Theorem 1.6 plays a role in the classification of initial data sets that admit a global static potential. Let be a connected Riemannian manifold that admits a non-constant function with , where

 L∗f=−(Δf)g+∇df−fRic

is the formal adjoint of the linearisation of the scalar curvature operator at . We recall from e.g. [Corvino:2000] that when is asymptotically flat, then its scalar curvature vanishes and the condition that is equivalent to

 ∇df=fRic and Δf=0,

implying that the spacetime

 (Mo×R,g−f2dt⊗dt) where Mo={x∈M:f(x)>0}

is a static solution of the vacuum Einstein equations. More generally, G. Galloway and P. Miao show in [Galloway-Miao:2014] that when is an asymptotically flat Riemannian -manifold — possibly with several ends — such that vanishes on the boundary of , then every unbounded component of the (necessarily regular) level set is an absolutely area-minimizing plane. As observed in Section 4 of [Galloway-Miao:2014], Theorem 1.6 shows that such unbounded components can only exist when is flat and is a linear function. Together with Corollary 1.1 of [Miao-Tam:2015] and the refinement of the results of G. Bunting and A. K. M. Masood-ul-Alam [Bunting-Masood-ul-Alam:1987] in Proposition 4.1 of [Miao-Tam:2015], both due to P. Miao and L.-T. Tam, one obtains the following classification result:

Corollary 1.9.

Let be an asymptotically flat Riemannian -manifold, possibly with several ends, that admits a non-constant function with that vanishes on the boundary of . Then is isometric to either flat , or, for some , either Schwarzschild

 ⎛⎝{x∈R3:|x|≥m/2},(1+m2|x|)43∑i=1dxi⊗dxi⎞⎠

or the doubled spatial Schwarzschild geometry

 ⎛⎝R3∖{0},(1+m2|x|)43∑i=1dxi⊗dxi⎞⎠.

We are grateful to L. Ambrozio and P. Miao for valuable discussions concerning this point.

We now turn our attention to the role played by closed volume-preserving CMC surfaces in asymptotically flat manifolds.

In their groundbreaking paper [Huisken-Yau:1996], G. Huisken and S.-T. Yau have shown that the complement of a certain (large) compact subset of a Riemannian -manifold that is asymptotic to Schwarzschild with mass admits a foliation by closed volume-preserving CMC spheres where has (outward) mean curvature . Importantly, they observed that each leaf is characterized uniquely by its mean curvature among a large class of surfaces, justifying reference to as the canonical foliation of the end of . In [Qing-Tian:2007], J. Qing and G. Tian have given a delicate improvement of this characterization showing that is in fact the only closed volume-preserving stable CMC sphere of mean curvature in that encloses . These results of [Huisken-Yau:1996, Qing-Tian:2007] are perturbative in nature in that only surfaces far out in the chart at infinity are considered. They lie very deep even in the special case of the exact Schwarzschild (spatial) geometry

 (1) ⎛⎝R3∖¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯Bm2(0),(1+m2|x|)43∑i=1dxi⊗dxi⎞⎠.

We mention the spectacular recent characterization [Brendle:2013] by S. Brendle of closed embedded constant mean curvature surfaces in Schwarzschild as the centered coordinate spheres in this context.

In the next two main results, we investigate the question of global uniqueness results for large volume-preserving stable CMC surfaces in asymptotically flat manifolds.

Theorem 1.10.

Let be an asymptotically flat Riemannian -manifold with non-negative scalar curvature and horizon boundary. Assume that contains no properly embedded totally geodesic flat planes along which the ambient scalar curvature vanishes. Let be compact. There is so that every connected closed volume-preserving stable CMC surface with

 area(Σ)≥α

is disjoint from .

In conjunction with the uniqueness results from [Huisken-Yau:1996, Qing-Tian:2007], we obtain the following consequence:

Corollary 1.11.

Let be a Riemannian -manifold with non-negative scalar curvature that is asymptotic to Schwarzschild with mass and which has horizon boundary. Let . Every connected closed volume-preserving stable CMC surface that contains and which has sufficiently large area is part of the canonical foliation.

Theorem 1.10 was proven by the third-named author and J. Metzger in [stableCMC] under the (much) stronger assumption that has positive scalar curvature. As we have already mentioned, our improvement here is closely tied to the generality of Theorem 1.4.

In [offcenter], S. Brendle and the third-named author have constructed examples of Riemannian -manifolds asymptotic to Schwarzschild with positive mass that contain a sequence of larger and larger volume-preserving stable CMC surfaces that diverge to infinity together with the regions they bound. Thus, in the uniqueness results of [Huisken-Yau:1996, Qing-Tian:2007], a proviso that the surfaces enclose some given set is certainly necessary. When the assumption of Schwarzschild asymptotics is dropped, the examples in Theorem 1.8 show even more dramatically that some such a condition is necessary to obtain uniqueness results. Theorem 1.10 extends the results of [Huisken-Yau:1996, Qing-Tian:2007] optimally in this sense.

We remark that much progress has been made recently in developing analogues of the results of [Huisken-Yau:1996, Qing-Tian:2007] in general asymptotically flat Riemannian -manifolds, see e.g. [Huang:2010, Ma:2011, Nerz:2014].

D. Christodoulou and S.-T. Yau [Christodoulou-Yau:1988] have noted that the Hawking mass of volume-preserving stable CMC spheres in asymptotically flat Riemannian -manifolds with non-negative scalar curvature is non-negative. This observation makes these surfaces particularly appealing from a physical standpoint. Geometrically, they arise in the variational analysis of the fundamental question of isoperimetry. The results described above beg the question whether each leaf of the canonical foliation has least area for the volume it encloses, and whether it is uniquely characterized by this property. This global uniqueness result was established by J. Metzger and the third-named author in [isostructure]. (In exact Schwarzschild, this was proven by H. Bray in his dissertation [Bray:1997].) Unlike the results based on stability that we have described above, the existence and global uniqueness of isoperimetric regions of large volume has been verified in higher dimensions as well [hdiso].

The definition of the ADM-mass through flux integrals as in (15) and that of related physical invariants that canonically associated with an asymptotically flat Riemannian -manifold is suggested by the Hamiltonian formalism of general relativity. The fact that the positive mass theorem was a longstanding open problem is witness to the elusive nature of these concepts. Over the past two decades, in a quest for quasi-local versions of these notions, considerable effort has been spent on recasting these concepts in terms of geometric properties of . A spectacular advance in this direction is the development of an isoperimetric notion of mass by G. Huisken. Recall the classical fact that a small geodesic ball in a Riemannian manifold that is centered at a point of positive scalar curvature bounds more volume than a Euclidean ball of the same surface area. An explicit computation gives that large centered coordinate balls in Schwarzschild (which is scalar-flat) have the same property, and that the “isoperimetric deficit” encodes the mass. G. Huisken has introduced the concept of isoperimetric mass

 mISO=limr→∞2area(Sr)(vol(Br)−area(Sr)3/26√π)

which does not involve derivatives of the metric at all. In [Fan-Shi-Tam:2009], X.-Q. Fan, P. Miao, Y. Shi, and L.-F. Tam have shown that

if the scalar curvature of is integrable. Their derivation is based on a striking integration by parts. Thus, if , then large coordinate balls in contain more volume than balls of the same surface area in Euclidean space. Together with the positive mass theorem, this leads to a remarkable large scale manifestation of non-negative scalar curvature. We note that this implies that, in the examples constructed by R. Schoen and the first-named author that we described above, sufficiently large spheres in the Euclidean half-space, though evidently volume-preserving stable CMC surfaces, are not isoperimetric. We include the following consequence of this discussion, which sharpens [hdiso, Theorem 1.2] of J. Metzger and the third-named author:

Theorem 1.12.

Let be an asymptotically flat Riemannian -manifold with horizon boundary, integrable scalar curvature, and positive ADM-mass. For all sufficiently large there is an isoperimetric region of volume , i.e., there is a bounded region of volume that contains the horizon such that

 (2) area(∂ΩV)=inf{area(∂Ω):Ω⊂M smooth open region containing the % horizon, volume V}.

The region is smooth away from a thin singular set of Hausdorff dimension .

Assume now that and that the scalar curvature of is non-negative. Remarkably, isoperimetric regions exist in for all volumes in this case. This follows from a beautiful observation due to Y. Shi [Shi:isoIMCF], as we explain in Appendix LABEL:sec:appendixisoallvolume. It is natural to wonder about the behavior of for large volumes . For simplicity of exposition, we assume for a moment that has empty boundary. Let where . It has been shown in [isostructure] that these surfaces either diverge to infinity as , or that alternatively a subsequence of these surfaces converges geometrically to a non-compact area-minimizing boundary . In view of Theorem 1.6, the latter is impossible unless is flat . We arrive at the dichotomy that large isoperimetric regions in are either drawn far into the asymptotically flat end, or they contain the center of the manifold.

Corollary 1.13.

Let be an asymptotically flat Riemannian -manifold with non-negative scalar curvature and positive mass. Let be a bounded open subset that contains the boundary of . There is so that for every isoperimetric region of volume , either or is a thin smooth region that is bounded by the components of and nearby stable constant mean curvature surfaces.

Note that the conclusion of the corollary is wrong for flat . When the scalar curvature of is everywhere positive, this result was observed as Corollary 6.2 of [isostructure]. The role of Theorem 1.6 here is that of Theorem 1.5 in the proof of Corollary 1.11.

Acknowledgments. The first-named author wishes to express his gratitude to Richard Schoen for introducing him, with great professionality and unparalleled enthusiasm, to the mathematical challenges of general relativity. He also thankfully acknowledges the support of André Neves through his ERC Starting Grant. The second-named author would like to convey his deepest thanks to his advisor Simon Brendle for his invaluable support and continued encouragement. His research was supported in part by an NSF fellowship DGE-1147470 as well as the EPSRC grant EP/K00865X/1. The third-named author expresses his gratitude to Hubert Bray, Simon Brendle, Greg Galloway, Gerhard Huisken, Jan Metzger, and Richard Schoen. A part of this paper was written up during his invigorating stay of two months at Stanford University, which was supported by their Mathematical Sciences Research Center. The second- and third-named authors would also like to thank the Erwin-Schrödinger-Institute of the University of Vienna for its hospitality during the special program “Dynamics of General Relativity: Numerical and Analytic Approaches” in the summer of 2011. It is a pleasure to sincerely congratulate R. Schoen on the occasion of his 65th birthday.

2. Sheeting of volume-preserving stable CMC surfaces

Proposition 2.1.

Let be a homogeneously regular Riemannian -manifold with non-negative scalar curvature . Assume that there is a bounded open set and a sequence of connected closed volume-preserving stable CMC surfaces in with

 (3) limk→∞area(O∩Σk)=∞.

There exists a totally geodesic stable minimal immersion that does not cross itself. Moreover, with the induced metric is conformal to the plane and the ambient scalar curvature vanishes along this immersion.

Proof.

It follows from (20) and (3) that the mean curvatures of the surfaces tend to as . By Lemma D.2, the second fundamental forms of the surfaces are bounded independently of . Passing to a subsequence if necessary, we can find such that

 (4) limk→∞area(Br(p)∩Σk)=∞

for all . Choose base points for the submanifolds with Passing to a convergent subsequence, we obtain a complete minimal immersion with base point such that . As it is the limit of embedded surfaces, this immersion does not cross itself. Its second fundamental form is bounded. In particular, the area of small geodesic balls in is bounded below uniformly in terms of the radius. We see from (3) that is non-compact.

Let be the universal cover of . Let be a point such that . Consider the immersion .

In the argument below, we denote the second fundamental forms of the submanifolds and the immersion by and by respectively. Let be open, bounded, connected, and simply connected with . Fix sufficiently small.

Using the curvature bounds and (4), upon passing to a further subsequence, we see that there are components of that are geometrically close to one another, where is strictly increasing in . In fact, we can choose points contained in these components such that as for every . Using the maximum principle, we see that for every , the submanifolds with respective base points converge to to an immersion which agrees with after passing to the universal cover. It follows that we can find such that in as for every , and such that

 Σjk={expujk(x)ν(φ(x)):x∈U}

are disjoint subsets of for every .

Assume that there is a point in where for some . Let be a subset as above that contains this point. Fix sufficiently large. Then, for each , this implies that the surface contains a subset where whose area is bounded below independently of . Because can be taken arbitrarily large, this contradicts (20). It follows that is totally geodesic and .

To see that is stable, it is enough to show that every bounded open subset admits a positive Jacobi function. The argument below is similar to [Simon:1987, p. 333], [Meeks-Rosenberg:2005, p. 732], or [Meeks-Rosenberg:2006, p. 493]. We may assume that is simply connected and that . By the argument above, contains two disjoint pieces that appear as small graphs above whose unit normals approximately point in the same direction. The defining functions of these graphs are ordered. They tend to zero in as . These functions satisfy the same graphical prescribed constant mean curvature equation on . Hence, their difference is a positive solution of a linear uniformly elliptic partial differential equation. By the Harnack principle, the supremum and the infimum of this solution are comparable on small balls. As in [Simon:1987, p. 333], we may rescale and pass to a subsequence that converges to a positive solution of the Jacobi equation on .

It follows from [Fischer-Colbrie-Schoen, Theorem 3 (ii)] that with the induced metric is conformal to the plane. ∎

3. Bounded complete stable minimal immersions

Proposition 3.1.

Let be an asymptotically flat Riemannian -manifold with horizon boundary. Every complete minimal immersion with uniformly bounded second fundamental form is either unbounded or an embedding of a component of the horizon.

Proof.

Assume that the trace of the immersion is contained in a compact set. Let be the union of the horizon and the closure of . There is a closed minimal surface in that contains . To see this, let large be such that and such that the mean curvature of the coordinate sphere with respect to the outward pointing unit normal is bounded below by .

Let . Consider the functional

 Ω↦FH(Ω)=area(∂Ω)−Hvol(Ω)

on

 A={Ω:Ω⊂M is open with smooth boundary and S⊂Ω⊂Br}.

The curvature bounds from Lemma D.2 together with the completeness of the immersion ensure that acts as an effective geometric barrier for the minimization of this functional in the following sense: There is small depending on such that given with

 dist(∂Ω,∂(Br∖S))<δ

there is with

 dist(∂~Ω,∂(Br∖S))≥δ

such that

 FH(~Ω)

This follows from a classical calibration argument, see for example [Duzaar-Fuchs:1990, Lemma 7.2], based on vector fields as described in Lemma LABEL:lem:effectivebarrier. Standard arguments of geometric measure theory, see for example [Duzaar-Fuchs:1990, Fuchs:1991], imply that there is a minimizer of . Its boundary is a closed hypersurface in with constant (outward) mean curvature that is strongly stable, i.e., its Jacobi operator is non-negative definite. We obtain that

 area(ΣH)≤area(Sr)

from direct comparison. In conjunction with strong stability, we obtain uniform curvature estimates for from e.g.  [Schoen-Simon-Yau:1975] or [Schoen-Simon:1981]. It follows that the Hausdorff distance between and the horizon tends to zero as , since otherwise we could find (by extraction of a convergent subsequence) a closed minimal surface in that is not a component of the horizon. In particular, the trace of the original immersion is contained in a component of the horizon. Since the components are spheres, it follows that the immersion is an embedding. ∎

Remark 3.2.

The proof of the preceding lemma should be compared to those of [Huisken-Ilmanen:2001, Lemma 4.1] and [GAH, Theorem 4.1]. The key point is to recognize that the trace of the immersion acts as a barrier for area minimization.

4. Top sheets

Lemma 4.1.

Let be an asymptotically flat Riemannian -manifold. Let be an unbounded complete stable minimal immersion that does not cross itself. For every there is so that for all there is a plane through the origin in the chart at infinity with

 sup{r−1dist(φ(x),Π)+|projΠ(ν(x))|:x∈φ−1(Sr)}<ε.
Proof.

All rescalings take place in the chart at infinity.

Suppose, for a contradiction, that for some there is a sequence such that

 sup{r−1kdist(φ(x),Π)+|projΠ(ν(x))|:p∈φ−1(Srk)}≥ε

for every plane through the origin. Let be points with . It follows from Proposition E.4 that there is a plane through the origin so that, after passing to a subsequence, the rescaled immersions

 φ−1(M∖¯¯¯¯¯¯B1)→R3∖¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯B1/rk(0) given by x↦φ(x)/rk

with respective base points converge to an immersion

 φ1:Σ1→R3∖{0}

with . Let be points such that and

 (5) r−1kdist(φ(y∗k),Π1)+|projΠ1(ν(y∗k))|≥ε2.

By Proposition E.4, there is a plane through the origin such that a subsequence of the immersions

 φ−1(M∖¯¯¯¯¯¯B1)→R3∖¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯B1/rk(0) given by x↦φ(x)/rk

with respective base points converges to an immersion

 φ2:Σ2→R3∖{0}

with . We must have that because the original immersion does not cross itself. This contradicts (5). ∎

Proposition 4.2.

Let be an asymptotically flat Riemannian -manifold with non-negative scalar curvature. Assume that there is an unbounded complete stable minimal injective immersion Then there is a proper such embedding.

Proof.

All rescalings take place in the chart at infinity.

By Lemma 4.1, after a rotation of the chart at infinity, there is large so that

 sup{dist(φ(x),Π):x∈φ−1(Sr)}≤r/2

where and

 (6) |ν(x)⋅e3|≥12

for all with .

Let be points such that and

 limk→∞φ3(x∗k)=sup{φ3(x):x∈Σ with |φ(x)|=r}.

Here, on . The second fundamental form of the immersion is bounded by Lemma D.2. The pointed immersions with respective base points subconverge to an unbounded complete stable minimal immersion with base point that does not cross itself and such that . It follows from Corollary C.2 that with the induced metric is conformal to the plane. Lemma F.2 shows that is injective. Note that

 (7) ^φ3(^x∗)=sup{^φ3(x):^x∈^Σ with |^φ(^x)|=r}.

Thus is a disjoint union of traces of complete injectively immersed curves. In view of (6), these curves are either infinite spirals or simple and closed. The curve containing is simple and closed by (6) and (7). The preimage of this curve under is simple and closed in . By the maximum principle, the image under of the bounded open region in bounded by is contained in . Finally, a continuity argument using Lemma E.3 gives that is a proper embedding. ∎

5. Proofs of main theorems

Proof of Theorem 1.3.

Any non-compact, proper immersion must have unbounded trace. It follows from Corollary C.2 that with the induced metric is conformal to the plane. The Ricci tensor of the Schwarzschild metric (1) is given by

 m|x|3(1+m2|x|)−2(δij−3xkxℓ|x|2δikδjℓ)dxi⊗dxj.

In conjunction with Lemma E.5, we see that

 (8) Ric(ν,ν)(x)≥m2|φ(x)|−3

holds for all with sufficiently large. Since the immersion is proper, it follows that the negative part of is integrable. Using the conformal invariance of the Dirichlet energy in dimension two, the logarithmic cut-off trick, and Fatou’s lemma, we obtain that

 (9) ∫Σ|h|2+Ric(ν,ν)≤0

from stability. It follows that the function

 x↦|φ(x)|−3

is integrable along the immersion. Using also the Gauss equation (18) and the estimate

 (10) R∘φ(x)=o(|φ(x)|−3) as |φ(x)|→∞,

we see that the Gauss curvature of the immersion is integrable. Rewriting the integrand in (9) using the Gauss equation in the manner of R. Schoen and S.-T. Yau, we conclude that

 12∫Σ|h|2+R∘φ≤∫ΣK.

In particular,

 (11) 0≤∫ΣK.

For sufficiently large, we have that is a smooth bounded region by Lemma E.3. In fact, it follows from the argument in the proof of Lemma E.3 that is connected. The maximum principle gives that is simply connected.

At this point, we argue as in [stableCMC, Proposition 3.6], except that we use limits of pointed immersions instead of limits in the sense of geometric measure theory. By Proposition E.4, the geodesic curvature of the boundary of with respect to the induced metric is given by

 κ=(1+o(1))/r as r→∞.

Moreover,222In fact, either or .

 limsupr→∞length(∂Σr)2πr≥1.

Recall that the Gauss-Bonnet formula reads

 ∫ΣrK+∫∂Σrκ=2π.

By (11), we obtain that

 limsupr→∞length(∂Σr)2πr=1 and ∫ΣK=0.

A modification of the argument in [Fischer-Colbrie-Schoen, p. 209] using the logarithmic cut-off trick in the construction of the test functions shows that ; cf. [Carlotto, p. 11]. This is incompatible with the Gauss equation (18) and the estimates (8) and (10). ∎

Remark 5.1.

The argument from [Fischer-Colbrie-Schoen] applied as in the last step of the preceding proof shows that the surface in Proposition 1.1 is intrinsically flat.

Proof of Theorem 1.4.

The domain with the induced metric is conformal to the plane by Corollary C.2. If the immersion is injective, the result follows from Proposition 4.2. If not, it follows from Remark LABEL:rem:willbeinjective and Lemma LABEL:lem:stability-factors that the immersion factors to an unbounded complete stable minimal immersion through a side-preserving covering . Note that is cylindrical by topological reasons. This is impossible by Corollary C.2. ∎

Proof of Theorem 1.5.

This is immediate from Theorems 1.4 and 1.3, Lemma D.2, and Proposition 3.1. ∎

Proof of Theorem 1.6.

We first deal with the case where the boundary of is empty.

Let be as in Appendix LABEL:sec:cf. Let be such that is convex for all . Every closed minimal surface of is contained in .

Let be a connected unbounded properly embedded and separating surface that is area-minimizing with respect to . Fix a component of the complement of in and choose a point to the following specifications:

• is disjoint from ;

• ;

• intersects in a single component, and the function is decreasing in the direction of the unit normal of this component that is pointing into .

In Appendix LABEL:sec:cf, we construct a family of conformal Riemannian metrics on with the following properties (see also Figure 1):

1. smoothly as ;

2. on ;

3. as quadratic forms on , with strict inequality on ;

4. the scalar curvature of is positive on ;

5. the region is weakly mean-convex with respect to .

By taking smaller if necessary, we may assume that all closed minimal surfaces of are contained in .

According to Proposition 1.1, for all sufficiently large, the intersection of with is transverse in a nearly equatorial circle. We denote this circle by . Consider all properly embedded surfaces in that have boundary and which together with bound an open subset of . Using (v) and standard existence results from geometric measure theory, we see that among all these surfaces there is one — call it — that has least area with respect to . This surface is disjoint from by convexity. It has one component with boundary . Its other components are closed minimal surfaces in that are disjoint from . Importantly though, intersects , since otherwise,

 (12) areag(Σρ(t))=areag(t)(Σρ(t))≤areag(t)(Σ∩Bρ)

The strict inequality holds on account of (iii) and because intersects . Observe that (12) violates the area-minimizing property of with respect to .

Using standard convergence results from geometric measure theory, we now find a connected unbounded properly embedded separating surface as a subsequential geometric limit of as . By construction, is contained in where it is area-minimizing with respect to . Moreover, intersects . If intersects , then it also intersects because of (iv) and Proposition 1.1. Passing to a subsequential geometric limit as , we obtain a connected unbounded properly embedded separating surface that is contained in where it minimizes area with respect to . Using now the area-minimizing property of , we see that is in fact area-minimizing in all of . Note that intersects while it is disjoint from . It follows from the maximum principle that and are disjoint.

Repeating this argument with choices of converging to a fixed point on , we obtain a sequence of totally geodesic intrinsically flat planes in (see Proposition 1.1 and Remark 5.1) along which the ambient scalar curvature vanishes and that converge to from one side. Proceeding as in [Simon:1987, p. 333] but using the first variation of the second fundamental form instead of the Jacobi equation, cf. [Anderson-Rodriguez:1989] and [Liu:2013], we obtain a positive function such that

 (13) (∇2Σf)(X,Y)+Rm(ν,X,Y,ν)f=0

for all tangent fields of . Here, is a unit normal field of . Tracing this equation and using that (this follows from the Gauss equation), we obtain that

 ΔΣf=0.

It follows that is a positive constant. Going back to the original equation (13), we see that whenever are tangential to . The Codazzi equation implies that provided that are tangential, and the Gauss equation gives that whenever are tangential. It follows that the ambient curvature tensor vanishes along .

We may repeat this argument, beginning with any surface constructed as above. It follows that an open neighbourhood of in is flat and in fact isometric to standard for some . Moreover, the surfaces in that correspond to where are all area-minimising. Using standard compactness properties of such surfaces and a continuity argument, we conclude that is isometric to flat .

We now turn to the general case where has boundary. Consider with non-compact area-minimizing boundary . The unique non-compact component of is a separating surface. Let and denote the two components of its complement in . Note that the interior of agrees with either (Case 1) or (Case 2) outside of . The proof that is flat in proceeds exactly as above, except for the following change. In Case 1, we let have least area among properly embedded surfaces with boundary that bound together with in and relative to . In Case 2, we let have least area among properly embedded surfaces with boundary that bound together with in and relative to . Theorem 1.6 follows upon switching the roles of and . ∎

Remark 5.2.

The use of the conformal change of metric in this proof is inspired by an idea of G. Liu in his classification of complete non-compact Riemannian -manifolds with non-negative Ricci curvature [Liu:2013]. The observation (12) is crucial in the proof of Theorem 1.6, as we use it to be sure that the surfaces do not run off as . This observation is not needed for Theorem 1.7 below, since the solutions of Plateau problems considered in the proof cannot escape the slab as we pass to the limit. We point out that at a related point in the work of M. Anderson and L. Rodríguez [Anderson-Rodriguez:1989], their assumption of non-negative Ricci curvature is used tacitly in their delicate estimation of comparison surfaces [Anderson-Rodriguez:1989, (1.5)].

Proof of Theorem 1.7.

Since has horizon boundary, is diffeomorphic to the complement of a finite union of open balls with disjoint closures in . Let be the connected region bounded by two disjoint unbounded properly embedded complete minimal surfaces . By solving a sequence of Plateau problems in with boundary on and passing to a subsequential geometric limit as , we obtain an unbounded properly embedded boundary that is contained in where it is area-minimizing with respect to . In particular, every component of is a stable minimal surface. By the maximum principle, if such a component intersects with or , then it coincides with the respective surface. By Theorem 1.2, every unbounded component is a totally geodesic flat plane along which the ambient scalar curvature vanishes. We may now proceed as in the proof of Theorem 1.6. ∎

Proof of Theorem 1.10.

Assume that there exist a compact set and closed volume-preserving stable CMC surfaces with and . Suppose that

 supkarea(Br∩Σk)<∞

for every . Using the methods from [stableCMC] we find an unbounded complete stable minimal surface that is properly embedded. (In fact, the surface has quadratic area growth.) In conjunction with Theorem 1.2, this contradicts our hypothesis.333The proof of Theorem 1.2 simplifies considerably for surfaces with quadratic area growth. Indeed, the arguments in [stableCMC, Sections 3 and 4] show that . It follows from [Fischer-Colbrie-Schoen, p. 209] that is flat with its induced metric. Lemma E.5 is quite elementary for surfaces with quadratic area growth, see the argument in [stableCMC, Lemma 3.5]. Finally, the Gauss equation rearrangement argument applied in the manner of R. Schoen and S.-T. Yau leads to a contradiction.

Assume now that

 supkarea(Br∩Σk)=∞,

for some . Using Proposition 2.1 we obtain a complete stable minimal immersion that does not cross itself and where with the induced metric is conformal to the plane. Such an immersion must be unbounded by Proposition 3.1 and the fact that the components of the horizon are spheres. This contradicts Theorem 1.4. ∎

Remark 5.3.

We remark that in the preceding proof, because the immersion at hand is totally geodesic, the argument for “passing to the top sheet” simplifies. Indeed, we obtain the estimate

 |hδ(x)|≤O(|φ(x)|−2)

as for the Euclidean second fundamental form of the immersion. This can be integrated up at infinity to show that the immersion is essentially a union of multi graphs above a fixed plane outside a large compact set.

Proof of Theorem 1.12.

Assume that for a sequence there is no isoperimetric region of volume . The argument in the proof of [hdiso, Theorem 1.2] (see also [Nardulli:existence, Theorem 2] for a much more general version of this line of argument in the case where the horizon is empty) shows that there is a minimizing sequence for

 (14) inf{area(∂Ω):Ω⊂M % smooth open region of volume Vi containing the horizon}

consisting of a divergent sequence of coordinate balls of radii and a residual isoperimetric region , and that the volumes of these residual regions diverges as . Moreover, we have that

 limj→∞n−1rj(Vi)=H(Vi)

where is the (outward) mean curvature scalar of . Let . The blow-down argument in [hdiso] shows that is close to a coordinate ball of radius upon rescaling by when is sufficiently large. We conclude that (14) is almost achieved by the union of two large disjoint coordinate balls of comparable radii provided is sufficiently large. This contradicts the Euclidean isoperimetric inequality. ∎

Appendix A Basic notions and conventions

Consider a complete Riemannian -manifold , possibly with boundary.

We say that is asymptotically flat if there are a compact subset and a chart

 M∖K≅{x∈R3:|x|>12}

so that the components of the metric tensor have the form

 gij=δij+bij

where

 |bij|+|x||∂kbij|+|x|2|∂k∂ℓbij|=o(1) as x→∞.

Such a chart is called a structure at infinity. We always fix such a chart when introducing an asymptotically flat Riemannian manifold and refer to it as the chart at infinity. We also define a smooth positive function

 |⋅|:M→(0,∞)

that coincides with the Euclidean distance from the origin in in the above chart and which on is bounded by . Given , we let

 Br={p∈M:|p|

If the scalar curvature of is integrable, then the limit

 (15) limr→∞116π∫{|x|=r}3∑i,j=1(∂igij−∂jgii)xj|x|

exists. It is independent of the choice of structure at infinity [Bartnik:1986] and called the ADM-mass (after R. Arnowitt, S. Deser, and C. W. Misner [ADM:1961]) of the asymptotically flat manifold .

We say that asymptotically flat has horizon boundary if its only closed minimal surfaces are the components of its boundary. It is known that the boundary components of such are area-minimizing spheres. Moreover, is diffeomorphic to the complement of a finite union of open balls with disjoint closures in . See [Huisken-Ilmanen:2001, Lemma 4.1] and the references therein.

Let . We say that is asymptotic to Schwarzschild with mass if there exists a chart as above such that

 (16) gij=(1+m2|x|)4δij+cij

where

 |x||cij|+|x|2|∂kcij|+|x|3|∂k∂ℓcij|=o(1) as x→∞.

We say that an immersion does not cross itself if given with there are open with and such that and so that the restrictions of to and are embeddings.

The concept of “immersions that do not cross themselves” arises naturally when studying limits of injective immersions of co-dimension one.

Consider a two-sided immersion of a boundaryless surface with unit normal .

Below, we use and to denote the ambient Ricci tensor and scalar curvature, we write and for the (scalar) mean curvature and the second fundamental form of the immersion with respect to the designated unit normal, we denote by the Gauss curvature of the induced metric on , and we compute gradients and lengths and perform integration with respect to the induced metric.

Recall that is a stable minimal immersion if its mean curvature vanishes and

 (17) ∫Σ|∇u|2≥∫Σ(|h|2+Ric(ν,ν))u2 for all u∈C∞c(Σ).

Such immersions arise in area minimization; cf. Appendix LABEL:sec:variationformulae.

Recall that is a volume-preserving stable CMC immersion if its mean curvature is constant and

 ∫Σ|∇u|2≥∫Σ(|h|2+Ric(ν,ν))u2 for all u∈C∞c(Σ) with ∫Σu=0.

Such immersions arise in area minimization with a (relative) volume constraint, i.e. in the isoperimetric problem; cf. Appendix LABEL:sec:variationformulae.

Finally, recall the Gauss equation

 (18) R∘φ=2K+|h|2−H2+2Ric(ν,ν).

We emphasize that in this paper, we adopt the convention that constant mean curvature immersions with non-zero mean curvature and stable minimal immersions are by definition two-sided. The immersions considered here are all of co-dimension one. The domain of a complete immersion is connected by definition.

The notion of convergence for pointed immersions and compactness results in the presence of uniform curvature bounds are used throughout the paper and are reviewed in Appendix B.

Appendix B A compactness result for pointed immersions

For a proof of the following compactness result, see [Cooper].

Lemma B.1 (Limits of immersions).

Let be a complete Riemannian manifold. Let

 {φk:Σk→M}∞k=1

be a sequence of complete constant mean curvature immersions such that

 supksupx∈Σk|hk(x)|<∞.

Assume that there are points such that the limit

 limk→∞φk(x∗k)

of points in exists. There is a complete constant mean curvature immersion

 φ:Σ→M

and a point so that a subsequence of the immersions

 φk:Σk→M with base points x∗k

converges to

 φ:Σ→M with base point x∗

in the sense of pointed immersions. By this we mean that the following holds up to passing to a subsequence. Let be a unit normal field of . There are bounded open subsets and with