on the topology of black hole event horizons in higher...

59
Introduction and Motivations Hawking’s Theorem Revisited Five-Dimensional Black Holes Six-Dimensional Black Holes Comments on Higher Dimensions Open Questions and Outlook On The Topology of Black Hole Event Horizons in Higher Dimensions Yaron Oz (Tel-Aviv University) General Relativity In Higher Dimensions, Hebrew University, February 2007 Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensi

Upload: others

Post on 13-Aug-2020

3 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

On The Topology of Black Hole Event Horizonsin Higher Dimensions

Yaron Oz (Tel-Aviv University)

General Relativity In Higher Dimensions,Hebrew University, February 2007

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 2: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Outline

1 Introduction and Motivations

2 Hawking’s Theorem Revisited

3 Five-Dimensional Black Holes

4 Six-Dimensional Black Holes

5 Comments on Higher Dimensions

6 Open Questions and Outlook

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 3: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

In four dimensions the topology of the event horizon of anasymptotically flat stationary black hole is uniquelydetermined to be the two-sphere S2 (Hawking).Hawking’s theorem: The integrated Ricci scalar curvatureR with respect to the induced metric h on the event horizonMH , is positive ∫

MH

RdS > 0 (1)

This condition applied to two-dimensional manifoldsdetermines uniquely the topology.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 4: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

Another way to determine the topology of the event horizonis via the so called topological censorship (Friedman,Schleich, Witt:1993): the region of space-time outside theblack hole should have a simple topology.Mathematically it requires the horizon to be cobordant to asphere via a simply connected oriented cobordism.For a two-dimensional horizon it means that there is asimply connected three-dimensional oriented manifoldwhose boundary is the oriented disjoint union of thehorizon and the two-sphere.Topological censorship implies that the topology of theevent horizon is that of the two-sphere S2 (Chrusciel,Wald:1994).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 5: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

The classification of the topology of the event horizons inhigher dimensions is more complicated.For instance, for five-dimensional asymptotically flatstationary black holes, in addition to the known S3 topologyof event horizons, stationary black hole solutions withevent horizons of S2 × S1 topology (Black Rings) havebeen constructed (Emparan, Reall:2001).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 6: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

We will consider the topology of event horizons indimensions higher than four.First, we reconsider Hawking’s theorem and show that itcontinues to hold in higher dimensions.Using this and Thurston’s geometric types classification ofthree-manifolds, we find that the only possible geometrictypes of event horizons in five dimensions are S3 andS2 × S1.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 7: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

In six dimensions we use the requirement that the horizonis cobordant to a four-sphere, Friedman’s classification oftopological four-manifolds and Donaldson’s results onsmooth four-manifolds, and show that simply connectedevent horizons are homeomorphic to S4 or S2 × S2.We find allowed non-simply connected event horizonsS3 × S1 and S2 × Σg (Σg is a genus g Riemann surface),and event horizons with finite non-abelian first homotopygroup, whose universal cover is S4.We will make some comments on the classification indimensions higher than six.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 8: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

This talk is based on the paper: ”On the Topology of BlackHole Event Horizons in Higher Dimensions”, with C.Helfgott and Y. Yanay, JHEP 0602 (2006) 025.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 9: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

Galloway and Schoen (2006) have shown that genericallythe event horizon admits a metric of positive scalarcurvature (positive Yamabe type).This, together with the fact that closed orientable3-manifolds of positive Yamabe type are connected sumsof S3 and S2 × S1, specifies the topologies of eventhorizons in five dimensions.The theorem does not hold for Ricci flat horizons, and forflat horizons in four and five dimensions.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 10: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Introduction and Motivations

Fields Medalists that will cast in this talk:

Thom (1958): CobordismSmale (1966): h-cobordismThurston (1982): Topology of 3-manifoldsDonaldson (1986): Smooth 4-manifoldsFreedman (1986): Topological 4-manifoldsPerelman (2006, declined): Geometrization conjecture

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 11: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

Let us reconsider Hawking’s theorem. We will find that alsofor asymptotically flat stationary black holes in dimensionshigher than four the integrated Ricci scalar curvature Rwith respect to the induced metric h on the event horizonMh, is positive.Idea of the proof: Consider a closed 2-surface Σ, which isan intersection of the event horizon and a space-likehypersurface. Hawking shows that if Σ is not S2, then itcan be deformed to an outer trapped surface, theexpansion θ < 0 outside the black hole.However, any closed trapped surface must lie inside theblack hole.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 12: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

Consider a stationary n-dimensional space-time M with ametric g. M is required to be regular predictable, i.e. itsfuture is predictable from a Cauchy surface.Denote by (Y1, Y2) two future-directed null vectorsorthogonal to MH , normalized as

Y a1 Y2a = −1 (2)

We take Y1 to be the future-directed null vector pointing outof the horizon, and Y2 to be the vector pointing into thehorizon.We now deform the event horizon by moving each point onit a parameter distance ω along an orthogonal nullgeodesic with tangent vector Y a

2 .

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 13: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

Following the same steps as in the original proof, onederives the equation

d θ

dω= pb;d hbd −RacY a

1 Y c2 +RadcbY d

1 Y c2 Y a

2 Y b1 +

papa − Y a1 ;c hc

dY d2 ;bhb

a (3)

where θ ≡ Y a1 ;bhb

a and pa = −habY2c;bY c1 .

The (positive definite) induced metric on the horizon reads

hab = gab + Y1aY2b + Y2aY1b (4)

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 14: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

Using a rescaling:

Y1 → Y ′1 = eyY1, Y2 → Y ′

2 = e−yY2 (5)

pa → p′a = pa + haby;b, we get

d θ′

dω′

∣∣∣∣∣ω=0

= pb;d hbd + y;bd hbd −RacY a1 Y c

2 +

RadcbY d1 Y c

2 Y a2 Y b

1 + p′ap′a (6)

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 15: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

Using the Gauss-Godazzi equations, evaluated on thehorizon yields

R = Rijkl hik hjl = R− 2RijklY i1Y j

2Y k1 Y l

2 + 4RijY i1Y j

2 (7)

where R is the Ricci scalar associated to the inducedmetric, and the unhatted quantities are the curvaturetensors of the full metric.Also, we use Einstein’s equations and the dominant energycondition:

12R+RabY a

1 Y b2 = 8πTabY a

1 Y b2 ≥ 0 (8)

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 16: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

We get:

d θ′

dω′

∣∣∣∣∣ω=0

=

∫MH

(−12

R + 8πTabY a1 Y b

2 )dS + p′ap′a (9)

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 17: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Hawking’s Theorem Revisited

Suppose d θ′

dω′

∣∣∣ω=0

is positive. We then take ω′ to be a smallnegative value, thereby looking at a surface slightly outsidethe horizon on which θ′ is now negative.Such a surface is an outer trapped surface, which isforbidden in a stationary regular predictable space-timesatisfying the dominant energy condition.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 18: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The horizons are three-manifolds. Thurston introducedeight geometric types in the classification ofthree-manifolds.There are eight basic homogeneous geometries, up to anequivalence relation, called geometric types. Out of thesetypes one constructs geometric structures, which arespaces that admit a complete locally homogeneous metric.Geometrization conjecture, proven by Perelman: Anycompact and oriented three-manifold has a decompositionas a connected sum of these basic geometric types.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 19: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

Consider an orientable, connected, complete and simplyconnected Riemannian three-manifold X which ishomogeneous with respect to an orientation preservinggroup of isometries G. The eight geometric types classify(X , G). The equivalence relation (X , G) ∼ (X ′, G′) holdswhen there is a diffeomorphism of X onto X ′, which takesthe action of G onto the action of G′.Out of these types one constructs spaces (geometricstructures) M ' X/Γ where Γ is a subgroup of G. Here theaction of Γ is discontinuous, discrete and free. M is locallyhomogeneous with respect to the metric on (X , G). It isisometric to the quotient of X by Γ.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 20: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The first three types in the classification are based on thethree constant curvature spaces, the 3-sphere S3

(Spherical geometry), which has a positive scalar curvatureand isometry group G = SO(4), the Euclidean space R3

(Euclidean geometry) with R = 0 and isometry groupG = R3 × SO(3) and the hyperbolic space H3 (Hyperbolicgeometry) with R < 0 and isometry group G = PSL(2, C).Of these three geometric types, only the S3 type satisfiesthe curvature condition and is allowed as an horizon.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 21: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The next two types are based on S2 × R and H2 × R. Ofthese two geometric types, only the S2 × R type satisfiesthe curvature condition and is allowed as an horizon. Inthis allowed class we have S2 × S1.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 22: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The last three geometric types are Nil geometry, Solgeometry and the universal cover of the Lie groupSL(2, R).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 23: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The Nil geometry: this is the geometry of thethree-dimensional Lie group of 3× 3 real upper triangularmatrices of the form 1 x z

0 1 y0 0 1

(10)

under matrix multiplication (Heisenberg group).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 24: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

We can think about Nil as (x , y , z) ∈ R3 with themultiplication

(x , y , z) · (x ′, y ′, z ′) = (x + x ′, y + y ′, z + z ′ + xy ′) (11)

The Nil metric (the left-invariant metric on R3) is given by

ds2 = dx2 + dy2 + (dz − xdy)2 (12)

and has R = −12 . The Nil geometry type does not satisfy

the curvature condition and is not allowed as an horizon.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 25: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The Sol geometry: this is the geometry of the Lie groupobtained by the semidirect product of R with R2. We canthink about Sol as (x , y , z) ∈ R3 with the multiplication

(x , y , z) · (x ′, y ′, z ′) = (x + e−zx ′, y + ezy ′, z + z ′) (13)

The left-invariant Sol metric is given by

ds2 = e2zdx2 + e−2zdy2 + dz2 (14)

and has R = −2. The Sol geometry type does not satisfythe curvature condition and is not allowed as an horizon.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 26: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

The ˜Sl(2, R) geometry: this is the geometry of theuniversal covering of the three-dimensional Lie group of all2× 2 real matrices with determinant one Sl(2, R). The˜Sl(2, R) geometry type has negative R and is not allowed

as an horizon.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 27: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Five-Dimensional Black Holes

Summary: We find that only two geometric types areallowed horizons in five dimensions: the S3 geometric typeand the S2 × R geometric type.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 28: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Consider now six-dimensional stationary black holes inasymptotically flat space-times.We will use topological censorship together withFriedman’s classification of topological four-manifolds andDonaldson’s results on smooth four-manifolds, in order toclassify possible event horizons.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 29: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

First we note that oriented cobordism from the eventhorizon to a four-sphere S4 exists if and only if the horizonmanifold Mh has vanishing Pontrjagin and Steifel-Whitneynumbers.In general, two smooth closed n-dimensional manifolds arecobordant iff all their corresponding Steifel-Whitneynumbers are equal (Thom).If in addition we require the cobordism to be oriented then(when n = 4k ) their corresponding Pontrjagin numbers areequal.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 30: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

We start by considering simply connected event horizons,that is

Π1(Mh) = 0 (15)

This, in particular, implies that the cohomology groupsH1(MH) (which is the abelianization of the first homotopygroup), and H3(MH) (by Hodge duality) vanish.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 31: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

One can use the second cohomology group H2(MH) todefine an intersection form

Q(α, β) =

∫MH

α ∧ β (16)

where α, β ∈ H2(MH). Q is the basic topological invariantof a compact four-manifold.Q is symmetric, non-degenerate withrank(Q) = b2 ≡ dim H2(MH), and can be diagonalizedover R.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 32: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Since the four-sphere S4 has zero second cohomologygroup, all its intersection numbers vanish and

Q(S4) = (0) (17)

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 33: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

The signature σ of a four-manifold is defined by thedifference of positive and negative eigenvalues of Q.It can be expressed using the Hirzebruch signaturetheorem as

σ(MH) =13

∫MH

p1 (18)

where p1 is the first Pontrjagin class

p1 = − 18π2 TrR ∧ R (19)

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 34: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Since p1(S4) vanishes, topological censorship implies thenthat the signature of Q(MH) vanishes

σ(Q(MH)) = 0 (20)

Mathematically, the signature is cobordant invariant.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 35: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Consider next the Stiefel-Whitney classes

ωi ∈ H i(MH , Z2) (21)

For a compact, simply-connected orientable manifoldω1 = ω3 = 0.ω2 is the obstruction to a spin-structure.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 36: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Although ω2(S4) = 0, oriented cobordism does not implythat the second Stiefel-Whitney class of MH is zero. Inother words, MH is not necessarily a spin manifold.The reason is that we require the Stiefel-Whitney numbersand not the classes to vanish.Here it means that

〈ω22, µ〉 = 0 (22)

where µ ∈ H4(M, Z2) is the so called fundamentalhomology class.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 37: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

The intersection form Q is actually defined on the latticeH2(MH , Z ) and is a unimodular (det(Q) = ±1) symmetricbilinear form over the integers.One says that Q is of even type if

Q(α, α) ∈ 2Z (23)

for all α ∈ H2(MH , Z ).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 38: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Fact: If ω2 = 0 then Q is even.Summary: event horizons which are spin manifolds arecharacterized as topological four-manifolds by anintersection form Q(MH), which has vanishing signatureand is of even type.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 39: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

When the event horizons are not spin manifolds, ω2 6= 0and Q(MH) is odd.In this case there are two topological four-manifolds MHfor a given intersection form.They are distinguished by the Kirby-Siebenmann invariantα(MH), which is zero if MH × S1 is smooth and one ifMH × S1 is not smooth.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 40: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

In the classification of possible intersection forms of MHwe distinguish two cases:

(i) Q(MH) is positive definite.

(ii) Q(MH) is indefinite.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 41: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Consider first the case when Q(MH) is positive definite.Theorem (Donaldson): If Q(MH) is even then MH ishomeomorphic to the four-sphere S4.If Q(MH) is odd then MH is homeomorphic to aconnected sum of positively oriented CP2’s.H2(CP2, Z ) has one generator and Q(CP2) = (1). Thus,the signature of the connected sum in nonzero, and this isnot an allowed horizon.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 42: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

If Q(MH) is indefinite, then if it is even it can be written as

Q(MH) = aE8 + bH, a, b ∈ Z b 6= 0 (24)

where E8 is the Cartan matrix of the Lie algebra E8 and

H =

(0 11 0

)(25)

is the intersection form of S2 × S2.H2(S2 × S2, Z ) is generated by a = S2 × pt andb = pt × S2. The intersection matrix Q(S2 × S2) in thebasis (a, b) is H.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 43: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Since the signature σ(E8) = 8 and we require thatσ(Q(MH)) = 0 this implies that a = 0.The basic case is b = 1 and MH is S2 × S2. If we takeb > 1 we will get a connected sum of S2 × S2.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 44: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

We should note, however, that by using connected sumsthere is a way to construct other event horizons whoseintersection form has vanishing signature and is of eventype.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 45: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Consider for instance a K 3. Its intersection form isQ(K 3) = −2E8 + 3H. Since its signature is nonzero, K 3 isnot an allowed horizon.However, if we take a connected sum of the K 3 and −K 3,where −K 3 has an opposite orientation we get an evenintersection form with vanishing signature, sinceQ(−K 3) = −Q(K 3).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 46: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Theorem (Freedman):If Q(MH) is indefinite and odd then MH is a connectedsum of ±CP2’s, where −CP2 has the opposite orientationof CP2 and Q(−CP2) = −Q(CP2) = (−1).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 47: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Summary: We found that if the horizon is simply connectedthen it is homeomorphic, up to a connected sum operation,to S4 or to S2 × S2. Note that both S4 and S2 × S2 arespin manifolds.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 48: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Consider next the case when MH is not simply connected.When Q(MH) is positive definite, one can relax thecondition that Π1(MH) = 0 by requiring only that there areno non-trivial homomorphisms of Π1(MH) into SU(2). Thisimplies that every flat SU(2) bundle over MH is trivial andthat H1(MH) being the abelianization of Π1(MH) vanishes.This allows Π1(MH) to be any finite simple nonabeliangroup. With this relaxed condition we get four-manifoldevent horizons, whose universal cover is S4.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 49: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

There are three other non-simply connected cases that wewould like to explore.Consider T 4. Its intersection form is Q(T 4) = 3H and it isnot excluded by the previous discussion from being anevent horizon.However, it does not satisfy our curvature condition (atleast with the flat metric).

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 50: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Next consider S3 × S1. It is not ruled out by our analysissince it has vanishing Pontrjagin and Steifel-Whitneynumbers. Also, it satisfies the curvature condition.The last examples are Σg × Σh, where Σg and Σh areRiemann surfaces of genus g and h respectively, and wehave assumed that the induced metric decomposes as adirect (unwarped) sum.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 51: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Assuming a product metric, the curvature condition reads

(g − 1)Vol(Σh) + (h − 1)Vol(Σg) < 0 (26)

This can be satisfied by Σh = S2 and

g < 1 +Vol(Σg)

Vol(S2)(27)

We cannot exclude, a priori, the possibility that the ratio ofvolumes can be as large as we want and therefore allgenera g are allowed. We encountered above the caseg = 0, namely the horizon S2 × S2. The intersection formQ(S2 × Σg) = H and it is not ruled out by topologicalcensorship.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 52: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Six-Dimensional Black Holes

Summary: We found that if the event horizon has vanishingfirst homotopy group then it is homeomorphic to S4 orS2 × S2.If the event horizon has finite simple nonabelian firsthomotopy group and positive intersection form, then itsuniversal cover is homeomorphic to S4.We found other allowed non-simply connected casesS3 × S1 with first homotopy group Z and S2 × Σg with firsthomotopy group1 Π1(Σg).

1Π1(Σg) is generated by a1, b1, ..., ag , bg with the relationa1b1a−1

1 b−11 ...agbga−1

g b−1g = 1.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 53: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Comments on Higher Dimensions

The event horizons MH are now closed differentiablen-manifolds with dimension n higher than four, cobordantto the n-sphere Sn.An important concept in differential topology is that ofh-cobordism. Two cobordant n-manifolds M1, M2 areh-cobordant if their inclusion map in the n + 1-dimensionalmanifold W are homotopy equivalent:M1 → W ∼ M2 → W .

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 54: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Comments on Higher Dimensions

The h-cobordism theorem (Smale) implies that if thehorizon manifold MH is h-cobordant to Sn then it isdiffeomorphic to Sn.Note, however, that h-cobordism is a stronger requirementthan what is implied by topological censorship.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 55: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Comments on Higher Dimensions

Let us introduce the concept of spin cobordism.Mathematically, it requires in our context the existence ofan (n + 1)-dimensional compact spin manifold, whoseboundary is the oriented disjoint union of Sn and MH . Inparticular, MH is a spin manifold whose spin structure isinduced from that of the (n + 1)-dimensional manifold.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 56: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Comments on Higher Dimensions

The concept of spin cobordism may be relevant, since weare mainly interested in higher-dimensional black holessolutions to supergravity equations as the low energyeffective description of the superstring equations.Therefore, we would like the geometry to accommodatefermions.Note, that in the classification of the previous section, spincobordism would have implied that the intersection form isof even type.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 57: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Comments on Higher Dimensions

n = 5 : For five-dimensional manifolds with vanishingsecond Steifel-Whitney class , there exists a classificationof all possible closed simply connected manifolds. Themanifolds are in 1-1 correspondence with finitely generatedabelian groups.n = 6 : Six-dimensional closed manifolds with vanishingfirst and second homotopy groups Π1 = 0 and Π2 = 0(2-connected) are homeomorphic to S6 or connected sumof copies of S3 × S3.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 58: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Comments on Higher Dimensions

n = 2k : There are general results which enumerate the(k − 1)-connected 2k -manifolds.n ≥ 5: If MH is n-dimensional compact, simply connectedand spin cobordant to Sn, it is obtained from Sn by doingsurgery on spheres of codimension greater than two.

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions

Page 59: On The Topology of Black Hole Event Horizons in Higher ...old.phys.huji.ac.il/~barak_kol/HDGR/proceedings/Oz.pdf · In four dimensions the topology of the event horizon of an

Introduction and MotivationsHawking’s Theorem RevisitedFive-Dimensional Black HolesSix-Dimensional Black Holes

Comments on Higher DimensionsOpen Questions and Outlook

Open Questions and Outlook

In order to have a complete classification of event horizons inhigher dimensions, additional physical requirements areneeded. This can come, for instance, from:

Refinement of topological censorshipConsiderations of stabilitySupersymmetryAnd, how to find solutions in six dimensions with S2 × S2

or S3 × S1 horizon topology ?

Yaron Oz On The Topology of Black Hole Event Horizons in Higher Dimensions