Blackfolds

Roberto Emparan
ICREA & U. Barcelona
Based on RE, T. Harmark, V. Niarchos, N. Obers to appear
Taiwan String Theory Workshop 2009
Why study D>4 gravity?
• Main motivation (at least for me!):
– better understanding of gravity (i.e. of what
spacetime can do)
• In General Relativity in vacuum
R
µν
=0
∃ only one parameter for tuning: D
– BHs exhibit novel behavior for D>4
– They may admit a simpler description as D →∞
Why study D>4 gravity?
• Also, for applications to:
– String/M-theory
– AdS/CFT
(+its derivatives: AdS/QGP, AdS/cond-mat etc)
– Math
– TeV Gravity (bh’s @ colliders…)
– etc
• Many findings will be "answers waiting for a
question"
• After the advances of the 1960's-70's, classical black
holes have been widely regarded as essentially
understood...
• ...so the only interesting open problems should concern
quantum properties of black holes
• But recently (~8 yrs ago) we’ve fully realized how little
we know about black holes – even classical ones! – and
their dynamics in D>4
E.g. we’ve learned that some properties of black holes are
• 'intrinsic': Laws of bh mechanics ...
• D-dependent: Uniqueness, Topology, Shape, Stability ...
• Black hole dynamics becomes richer in D=5, and much
richer in D r 6
• Activity launched initially by two main
results:
– Gregory-Laflamme (GL)-instability, its
endpoint, and inhomogeneous phases
– Black rings, non-uniqueness, non-spherical
topologies
• I’ll focus mostly on simplest set up:
vacuum, R
µν
= 0, asymptotically flat
solutions
• Why is D>4 richer?
– More degrees of freedom
– Rotation:
• more rotation planes
• gravitational attraction ⇔centrifugal repulsion
– ∃ ∃∃ ∃ extended black objects: black p-branes
– Both aspects combine to give rise to horizons w/
more than one length scale:
new dynamics appears when two scales differ
FAQ's
x
2
x
4
x
6
x
1
x
3
x
5

O(D−1) V U(1)
⌊(D−1)/2⌋

GM
r
D−3
+
J
2
M
2
r
2
• Why is D>4 harder?
– More degrees of freedom
– Axial symmetries:
• U(1)'s at asymptotic infinity appear only every 2 more
dimensions (asymptoticaly flat – or AdS – cases)
• in D=4,5: stationarity+2 rotation symmetries is enough to
reduce to integrable 2D σ-model (but only in vacuum, not in AdS)
• if D>5 not enough symmetry to reduce to 2D σ-model
– 4D approaches not tailored to deal with horizons with
more than one length scale: need new tools
FAQ's
What's the goal?
• In D=4:
– 1960-70's: Golden Age of Black Holes (4D)
• remarkably simple objects
• amenable to detailed theoretical study
– Much progress prompted by Kerr's exact solution (1963)
– Complete classification: uniqueness theorems
• In D>4: initial aim: try to repeat 4D success
– Construct exact solutions
– Classify all possible black holes (even if no uniqueness)
What's the goal?
• In D=5: exact solns+ classification
– quite successful in integrable sector w/ three commuting
isometries
• In Dr6:
– Dynamics too rich to be captured by looking for exact solutions
– only simplest bhs
– Classification becomes increasingly hopeless – and less
interesting
– Develop qualitatively new tools
– Focus on capturing new dynamics
(analogy to DBI branes)
4D black holes
• Static: Schwarzschild
• Stationary: Kerr
Horizon: ∆(r
hor
)=0, r
hor
∈ R
+
⇒ ⇒ ⇒ ⇒ GM r a
⇒ ⇒⇒ ⇒upper bound on J for given M: J b GM
2
a =
J
M
µ = 2GM
ds
2
= −
_
1 −
µ
r
_
dt
2
+
dr
2
1 −
µ
r
+ r
2
_

2
+ sin
2
θdφ
2
_
ds
2
= −dt
2
+
µr
Σ
_
dt + a sin
2
θdφ
_
2
+
Σ

dr
2
+ Σdθ
2
+ (r
2
+ a
2
) sin
2
θdφ
2
Σ = r
2
+ a
2
cos
2
θ , ∆ = r
2
− µr + a
2
,
Uniqueness theorem: End of the story
Multi-bhs not rigorously ruled out, but physically unlikely to be stationary
(eg multi-Kerr can't be balanced)
J
extremal Kerr J = GM
2
non-singular, finite area
A
H
Kerr black holes
GM
2
4D black holes
(fixed mass)
• Kerr bound J b GM
2
⇒ ⇒⇒ ⇒ 4D bh horizons
characterized by a single scale:
r
0
∼ GM ∼ a
• Horizons are always round ("plump"),
i.e., w/out large distortions
4D black holes
Black holes in D>4
• Schwarzschild is easy:
• Dynamics very similar to 4D:
– unique among static AF bhs Gibbons+Ida+Shiromizu
– classically linearly stable Ishibashi+Kodama
Tangherlini 1963
ds
2
= −
_
1 −
µ
r
D−3
_
dt
2
+
dr
2
1 −µ/r
D−3
+r
2
dΩ
(D−2)
µ ∝ M
• Solves R
µν
= 0
• Compactify: x
i
∼ x
i
+ L
i
different length scales on horizon: r
0
, L
i
Black strings & branes
black p-brane
Take a d-diml black hole
(Ricci-flat)
add p extra dim
horizon topology
R
p
x S
d-2
• Not asymp flat bhs, but necessary to understand them
ds
2
d+p
= ds
2
d
(black hole) +
p

i=1
dx
i
dx
i
r
0
L
– Gregory-Laflamme instability
– Non-uniform (=inhomogeneous=lumpy) black strings
GL zero-frequency mode = static perturbation
new static inhomogeneous solutions
λ
GL
≃ r
0
Gubser
Wiseman
L<r
0
r
0
L
GL instability
L>r
0
stable unstable
Separate scales New phenomena
Rotation in Dr5
• Myers+Perry (1986): rotating black hole
solutions with angular momentum in an arbitrary
number of planes
e.g. D=5,6: J
1
, J
2
D=7,8: J
1
, J
2
, J
3
etc
• They all have spherical topology
x
2
x
4
x
6
S
D−2
x
1
x
3
x
5

J
1
J
2
J
3
… J
⌊(D−1)/2⌋
Ultra-spinning black holes in Dr5
(single spin case)
gravitational
centrifugal

r
2
−1 = −
µ
r
D−3
+
a
2
r
2
ds
2
= −dt
2
+
µ
r
D−5
Σ
_
dt +a sin
2
θ dφ
_
2
+
Σ

dr
2
+ Σdθ
2
+ (r
2
+a
2
) sin
2
θ dφ
2
+r
2
cos
2
θ dΩ
2
(D−4)
Σ = r
2
+a
2
cos
2
θ , ∆ = r
2
+a
2

µ
r
D−5
,
µ ∝ M
a ∝
J
M
Myers+Perry 1986
Quadratic equation: fix µ, then a
can't be too large for real root
4D: 2a b µ
5D: a
2
b µ
⇒upper bound on J for given M
Horizon: ∆=0 ∆ = r
2
+ a
2

µ
r
D−5
D=4, 5:

r
a
r
0
r
∼r
2
Dr6:
For fixed µ there is an outer event
horizon for any value of a
⇒ ⇒⇒ ⇒No upper bound on J for given M
⇒ ⇒⇒ ⇒∃ ultra-spinning black holes
must have a zero ∀a
∼ −r

(D−5)
Horizon: ∆=0 ∆ = r
2
+ a
2

µ
r
D−5

J
D = 5
upper (extremal) limit
on J is singular
J
D r 6
no upper limit on J
for fixed M
J
D = 5
extremal, singular:
µ=a
2
⇒ r
+
=0
• Single spin MP black holes:
A
H
A
H
J
D r 6
a
H
→0, j →∞
r
2
+
+a
2

µ
r
D−5
+
= 0
• Consider a p r
+
∆(r =r
+
)=0
Ultra-spinning regime
ds
2
= −dt
2
+
µ
r
D−5
Σ
_
dt +a sin
2
θ dφ
_
2
+
Σ

dr
2
+ Σdθ
2
+ (r
2
+a
2
) sin
2
θ dφ
2
+r
2
cos
2
θ dΩ
2
(D−4)
r
+
a
A

=
_
dθdφ

g
θθ
g
φφ
|
r
+
= Ω
2
(r
2
+
+a
2
) ≃ Ω
2
a
2
−→ ℓ

∼ a
A

=
_
dΩ
(D−4)
(r
+
cos θ)
D−4
= Ω
D−4
(r
+
cos θ)
D−4
−→ ℓ

∼ r
+
(a ∼ J/M ∼ j)
A
7
A

Ultra-spinning regime
T
a
T ∼ 1/r
+
black membrane-like
Kerr-like
a ∼r
+
• Black membrane limit: a →∞
finite (transverse) horizon: keep r
+
finite
finite mass density: keep µ/a
2
finite
r
+
a
Multiple spins
• If all spins similar j
1
∼j
2
∼…∼j
⌊(D-1)/2⌋
⇒similar to Kerr
• ∃ ultra-spinning regimes if one j
i
(even D)
or two j
i
(odd D) are much smaller than the
rest
• Qualitatively understood as in single-spin:
p ultra-spins limit to black 2p-branes
bend into a circle
spin it up
black string:
Schw
4D
x R
Horizon topology S
1
x S
2
Exact solution available – and fairly simple
RE+Reall 2001
S
2
The forging of the ring (in D=5)
Metric
ψ-rotation
S
2
G(ξ) = (1 −ξ
2
)(1 +νξ)
S
1
φ
x
ν ∼R
2
/R
1
→ shape, λ/ν → velocity
equilibrium → λ(ν)
Parameters λ, ν, R
F(ξ) = 1 +λξ
ds
2
= −
F(y)
F(x)
_
dt −C(λ, ν) R
1 +y
F(y)

_
2
+
R
2
(x −y)
2
F(x)
_

G(y)
F(y)

2

dy
2
G(y)
+
dx
2
G(x)
+
G(x)
F(x)

2
_
x = const
y = const
"Ring coordinates" x, y
y = y
H
ψ
φ
5D: one-black hole phases
thin black ring
1
fat black ring
Myers-Perry black hole
(naked singularity)
(slowest ring)
3 different black holes with the same value of M,J
(fix scale M)
27
32
A
H
J
2
Other properties
J
J
T

thin
thin
fat
fat
MP bh
MP bh
fat rings ~ ("drilled-through") MP black holes
thin rings ~ (circular) black strings
Black ring to boosted black string
• Ultra-spinning limit:
R→∞
finite thickness: νR ∼ r
0
finite
finite velocity: λ/ν finite
⇒ boosted black string, with limiting boost velocity
v = tanh α =
1

2
R
r
0
Hints of a pattern?
• Ultra-spinning MP bhs and black rings:
– rotation unbounded: greatly distorts the horizon and introduces a
second length scale J/M, much larger than the thickness r
0
– limit to a (boosted) black string/brane
Ultra-spinning, black brane-like, two length scales
Moderately-spinning, Kerr-like, one length scale
D=5

J
T
J
Dr6

T
A
H
A
H
Questions?
4D vs hi-D Black Holes: Size matters
• 4D BHs: not only unique, spherical, and stable,
but also dynamics depends on a single scale:
r
0
∼GM
– true even if rotating: Kerr bound J b GM
2
• Main novel feature of D>4 BHs: in some regimes
they're characterized by two separate scales
– No upper bound on J for given M in D>4
Length scales J/Mand (GM)
1/(D−3)
can differ
arbitrarily
• All this has led to realize that hi-D bhs have
qualitatively new dynamics that was
unsuspected from experience with 4D bhs
• 4D bhs only possess short-scale (∼r
0
) dynamics
• Hi-D bhs: need new tools to deal with long-
distance (∼R pr
0
) dynamics
• Natural approach: integrate out short-distance
physics, find long-distance effective theory
• Two different (but essentially equivalent)
techniques for this:
– Matched Asymptotic Expansion (Mae)
• GR-minded approach, well-developed, produces explicit
perturbative metric, but cumbersome
– Classical Effective Field Theory (Cleft)
• QFT-inspired (Feynman diagrams), efficient for calculating
physical magnitudes, but not yet developed for extended
objects
• To leading order, no difference
• Leading order = blackfold approach
Goldberger+Rothstein
Kol
locally (r ` R)
equivalent to
Lorentz-transformed
black p-brane
r
0
Blackfolds: long-distance effective
dynamics of hi-d black holes
• Black p-branes w/ worldvolume = curved submanifold
of spacetime
M
p
s
n
Horizon: M
p
x s
n
R
• Long-distance dynamics:
find brane spacetime embedding x
µ

α
)
• General Classical Brane Dynamics: Carter
Given a source of energy-momentum, in probe approx,
– Newton's force law: F=ma
– Nambu-Goto-Dirac eqns: T
µν
= Tg
µν
K
ρ
=0: minimal surface
Long-distance theory

µ
T
µν
= 0 ⇒ T
µν
K
µν
ρ
= 0
extrinsic curvature
or, with external force: F
ρ
= T
µν
K
µν
ρ
• But, what is T
µν
for a blackfold?
• Short-distance physics determines
effective stress-energy tensor:
– blackfold locally Lorentz-equivalent to
black p-brane of thickness (s
n
-size) r
0
– In region r
0
` r field linearizes
⇒approximate brane by equivalent
distributional source T
µν

α
)
• Black p-brane
Lorentz-transform:
(e.g., black string boosted black string)
ds
2
= −
_
1 −
r
n
0
r
n
_
dt
2
+
p

i=1
dz
2
i
+
dr
2
1 −
r
n
0
r
n
+r
2
dΩ
2
n+1
r
0
z
i
T
tt
= r
n
0
(n + 1)
T
ii
= −r
n
0
T
µν
→T
µν
= r
n
0
_
(n + 1)a
t
µ
a
t
ν

p
i=1
a
i
µ
a
i
ν
¸
(t, z
i
) = σ
µ
, σ
µ
→a
µ
ν
σ
ν
, a
µ
ν
∈ O(1, p)
• Position-dependent thickness r
0

α
) and
"boosts" a
µ
ν

α
) ⇒T
µν

α
)
• Impose global blackness condition:
– Uniform surf gravity κ & angular velocities Ω
i
(eliminate thickness and boost parameters)
• Solve K
µν
ρ

α
)T
µν

α
;κ, Ω
i
)=0
⇒ determine x
µ

α
;κ, Ω
i
)
A Blackfold Bestiary
"Bestiary: a descriptive or
anecdotal treatise on
various real or mythical
kinds of animals, esp. a
medieval work with a
moralizing tone.''
Tune boost to equilibrium
Horizon S
1
x s
D−3
"small" transverse sphere ∼ r
0
• Simplest example: black rings in Dr5
K
µν
ρ
T
µν
= 0
T
11
R
= 0
T
11
= r
D−3
0
_
(D − 3) sinh
2
σ −1
¸
⇒sinh
2
σ =
1
D −3
R
r
0
(in D=5 reproduces value from exact soln)
Tune boost to equilibrium
Horizon T
p
x s
n+1
"small" transverse sphere ∼ r
0
• p-brane curved into T
p
:
K
µν
ρ
T
µν
= 0
T
11
R
1
=
T
22
R
2
= · · · = 0
T
ii
= r
n
0
(n(a
t
i
)
2
− 1)
⇒a
t
i
= −
1

n
, a
t
t
=
_
n +p
n
n=D−p−3
• Axisymmetric blackfolds
• Simple analytic solutions:
– even p: ultraspinning MP bh, with p/2 ultraspins
– odd p : round S
p
, with all (p+1)/2 rotations equal
• I'll illustrate two simple cases of each
(possibly rotations along all axes)
• Ultra-spinning 6D MP bh as blackfold
• Embed black two-brane in 3-space
ds
2
=dz
2
+dρ
2

2

2
, as z =const
to obtain planar blackfold P
2
xs
2
• Find size r
0
(ρ) of s
2
& boost α(ρ) of locally-equiv 2-brane:
– Soln: α(ρ)→∞, r
0
(ρ)→0 at ρ
max
=1/Ω
• Disk D
2
fibered by s
2
: topology S
4
: like 6D MP bh!
• All physical magnitudes match those of the ultraspinning
6D MP bh
ρ
φ
locally equiv to boosted black 2-brane
• S
3
xs
n+1
black hole as blackfold
(n r2)
• Embed three-brane in a space containing
ds
2
=dr
2
+r
2
dΩ
2
(3)
as r =R
• Solution exists if |Ω
1
|=|Ω
2
|=(3/(3+n))
1/2
R
−1
size of s
n
r
0
= const
• If |Ω
1
|>|Ω
2
| then numerical solution for
r =R(θ) : non-round S
3
φ
1
φ
2
S
3
In Progress:
• Sphere products:
• Static minimal blackfolds:
If no boost (no local Lorentz transf)
T
ij
= −Pg
ij
(spatial) ⇒ K
ρ
=0: minimal submanifold
D =

i
p
i
+n + 3
e.g.: hyperboloid
non-compact static blackfold

p
i
∈odd
S
p
i
×s
n+1
"small" transverse sphere ∼ r
0
Blackfold dynamics as 1st Law
• For stationary blackfolds, compute M, J
i
, A
H
, by
integrating stress-energy tensor T
tt
, T
ti
, and horizon
area element
• Consider M[x
µ
], J
i
[x
µ
], A
H
[x
µ
] as functionals of
embedding x
µ

α
;κ, Ω
i
)
• Then eqs of motion K
µν
ρ
T
µν
=0 are equivalent to
⇒ Stationary blackfold eqs = 1st Law
δM
δx
µ

κ
8πG
δA
H
δx
µ
−Ω
i
δJ
i
δx
µ
= 0
A
H
J
Blackfolds in Dr6 phase diagram
(w/ one spin)
Black rings and Black Saturns as blackfolds
can be described as blackfold
(M = 1)
cannot be
described as blackfold
D=5
Blackfolds in AdS (w/ one spin)
A
H
J
fixed M
J=ML
AdS
Dr6
A
H
J
"compressed"
versions of Λ=0
J=ML
AdS
BPS bound: ML
AdS
r Σ
i
|J
i
|
(could also include Black Saturns)
Thin rings: r
0
`min(R,L
AdS
)
• Stability of blackfolds for long wavelength
(λ p r
0
) perturbations can be analyzed within
blackfold approximation
• But black branes have short-wavelength G-L
instabilities
• Expect blackfolds to be unstable – on quick time
scales, Γ∼1/r
0
Instabilities and non-uniform phases
λ
GL
≃ r
0
/.88
r
0
• Non-uniform static black branes exist:
• Expected to also be unstable below D
*
(∼13)
but stable above D
*
• Use stable non-uniform branes as basis for
blackfolds (~ wiggly cosmic strings)
• These would emit grav waves, but in a much
longer time-scale than GL-instability
⇒ long-lived wiggly blackfolds
~ effective T
µν
~
Other blackfold applications
• Curved strings & branes in gravitational potentials:
– black rings in AdS: ∃ w/ arbitrarily large radius
– black rings in dS: ∃ static ones
– black rings in Taub-NUT (D0-D6 brane interactions)
– black saturns, charged and susy blackfolds
– Probe D-branes at finite temperature
• Dynamical evolution
– Time-dependent evolution
– Coupling to gravitational radiation (quadrupole formula)
Limitations
• Stationarity not guaranteed at higher
orders (i.e. w/ backreaction)
• Misses 'threshold' regime of connections
and mergers of phases, where scales
meet r
0
∼R
resort to extrapolations, numerical analysis,
and other information
• Long-distance dynamics of hi-d black holes efficiently
captured by blackfold approach (extends & includes DBI
approach to D-branes)
– black branes are elastic!
• Change focus:
– search for all Dr6 black hole solutions in closed analytic form is
futile (some may still show up: p=D−4)
– classification becomes increasingly harder at higher D, but also
less interesting
Investigate novel dynamical possibilities
Blackfolds as organizing framework
for hi-d black holes
A conjecture about uniqueness and
stability of hi-d black holes
• Non-uniqueness (=more than one solution for given asymptotic
conserved charges) and instabilities appear only as the two
horizon scales begin to differ, ie when
J
i
/Mt (GM)
1/(D−3)
(for at least one J
i
)
• For smaller spins ∃ only MP bhs (as far as we know), with
properties similar to Kerr, and there is no suggestion of
their instability
• It really looks like it's the appearance of more than one
scale that brings in all that differs from 4D
A conjecture about uniqueness and
stability of hi-d black holes
• Conjecture:
MP black holes are unique and dynamically stable iff
∀i =1,…,⌊(D−1)/2⌋
|J
i
| < α
D
M(GM)
1/(D−3)
,
with α
D
=Ο(1) (to be determined, presumably numerically).
• Applies to stationary black hole solutions w/ connected
non-degenerate horizons, and possibly to multi-black
hole solutions in thermal equilibrium
• Presumably also extends to charged solutions
Conclusions & Outlook
• Hi-D black hole dynamics turns out to be surprisingly rich,
but we have made great progress in recent years
• Our understanding of vacuum 5D black holes is close to
complete (main open problem: 1 or 2 rotation isometries?)
• In Dr6, blackfold approach captures much of the new
physics. It remains to:
– develop blackfold dynamics: we have the tools
– understand effects at threshold of effective description:
connections between phases, GL-like instabilities and
inhomogeneous phases. Need new tools, or numerical analysis
• In AdS, the cosmo length scale makes it difficult to
analyze some regimes of interest – eg large black holes
– again: new approaches, or numerics