1
|
- Tadashi Takayanagi (IPMU)
- Mainly based on the papers
- T. Nishioka (Kyoto U.)=
and
TT JHEP 0701:090, 2007
- M. Fujita, T. Nishioka
(Kyoto U.) and TT JHEP 0809:016, 2008
- Also I thank very much=
the collaborations with
- S.Ryu (Berkeley), M.
Rangamani, V. Hubeny(Durham),
- M.Headrick (Stanford),=
A. Karch , E. Thompson
(Washington)
- T. Azeyanagi, T. Hirata
(Kyoto U.)
|
2
|
- What is
entanglement entropy (EE) ?
-  =
;
A measure how much a given quantum state is
-  =
;
quantum mechanically entangled.
- (A)=
It is universal in that it =
is
well defined in any quantum
-  =
;
mechanical system.
- (B)=
In
QFT, it has the properties of
-  =
;
both entropy and correlation functions.
|
3
|
|
4
|
- Various Applications
- ・Quantum Information and Quantum Computing
-  =
;
EE =3D the amount of quantum information
-  =
;
(i.e. a quantum version of Shannon’s entropy)
- ・Condensed matter physics
- EE =3D how much
difficult to perform a computer simulation
-  =
;
This gets divergent at phase transition point !
-  =
;
A new quantum order parameter !
-  =
;
(`Wilson-loops for cond-mat people’)
|
5
|
|
6
|
- In this talk we are interested in applications to quantum
- gravity, especially string theory.
- Consider a holography in general spacetimes:
|
7
|
- Definition of entanglement entropy
- Divide a given quantum system into two parts A and B.
- Then the total Hilbert space becomes factorized
- We define the reduced density matrix for=
A by
- taking trace over the Hilbert space of B .
|
8
|
- Now the entanglement entropy is
defined by the
- von-Neumann entropy
- For example, this quantity is remarkably useful in a
- topological system (e.g. the quantum Hall liquid) as the
- correlation functions become trivial, while it does not.
-  =
; &n=
bsp;  =
; &n=
bsp;  =
;
[05’ Kitaev-Preskill, Levin-Wen]
|
9
|
- In this talk we consider the entanglement entropy in
- quantum field theories on (d+1) dim. spacetime
- Then, we divide &n=
bsp;
into A and B by specifying the
- boundary  =
; &n=
bsp;
.
-  =
; &n=
bsp;
A  =
;
B
|
10
|
- The entanglement entropy (EE) measures
- how A and B are entangled quantum mechanically.
- (1) EE is the en=
tropy
for an observer who is
- only accessible to the subsystem A and not to B.
- EE is a sort of a `non-local version of correlation
-
functions’, which captures some topological information=
.
-  =
; &n=
bsp;  =
; &n=
bsp;  =
;
(cf. Wilson loops)
- (3) EE is proportional=
to the
degrees of freedom.
-
It is non-vanishing even at zero temperature.
|
11
|
- Ex. Entanglement entropy in 3D QFT
- The general structure:
|
12
|
- Comment: Relation to boundary entropy
- The entanglement entropy in 2D CFT with a boundary
- looks like [Calabrese-Cardy 04’]
- where c is the central charge and g is the boundary
- entropy introduced by Affleck-Ludwig.
-  =
; &n=
bsp;
[Holographic Calculation:&nb=
sp;
Azeyanagi-Karch-Thompson-TT 07’
- =
288; based
on the AdS3 Janus by
Bak-Gutperle-Hirano 07’]
|
13
|
- In this way we learned that the entanglement entropy
- can capture the `quantum degrees of freedom’ of
- the ground state.
- This suggests that the entanglement entropy can be an
- order parameter for quantum many body systems.
- Since it has a universal definition, the holography implies
- that it is a nice and general order parameter
- for quantum gravity.
|
14
|
- ①Introduction=
li>
- ② Hologra=
phic
Entanglement Entropy
- ③ A New O=
rder
Parameter of
-  =
;
Confinement/deconfinement Phase Transition
- ④ Geometr=
ic
Entropy and Phase Transition
- ⑤ Entangl=
ement
Entropy in Time-Dependent Systems
- ⑥ Conclus=
ions
|
15
|
- Setup: AdS/CFT correspondence in Poincare Coordinate
|
16
|
- Divide the space N is into A and B.
- (2) Extend their
boundary to
the entire AdS space.
-
This defines a d dimensional surface.
- Pick up a minimal area surface and call this .=
li>
- The E.E. is given by naively applying the
-
Bekenstein-Hawking formula
-  =
; &n=
bsp;
as if  =
;
were an event horizon.
|
17
|
- Holographic Proof of Strong Subadditivity
-  =
; &n=
bsp;  =
; &n=
bsp;
[06’ Hirata-TT, 07’ Headrick-TT]
- The strong subadditivity is known as the most important
- inequality satisfied by EE.  =
;
[Lieb-Ruskai 73’]
|
18
|
- The holographic proof of this inequality is very quick !
|
19
|
- EE from AdS3/CFT2
- Consider AdS3 in the global coordinate
- In this case, the minimal surface is a geodesic line which
- starts at  =
; &n=
bsp;
and ends at
- (  =
; &n=
bsp;  =
;
) .
|
20
|
- The length of  =
;
, which is denoted by  =
;
, is found as
- Thus we obtain the prediction of the entanglement entropy
- where we have employed the celebrated relation
-  =
; &n=
bsp;  =
; &n=
bsp;  =
;
[Brown-Henneaux 86’]
|
21
|
- Furthermore, the UV cutoff a is related to
via
- In this way we reproduced the known formula
-  =
; &n=
bsp;
[94’ Holzhey-Larsen-Wilczek, 04’ Calabrese-Cardy =
]
- (In 2D CFT, we can analytically compute EE in various
- setups owing to the conformal map technique.
- But this is not so in higher dimensions.)
|
22
|
- Finite temperature case
- We assume the length of the total system is infinite.
- In this case, the dual gravity background is the BTZ black
- hole and the geodesic distance is given by
- This again reproduces the known formula at finite T.
|
23
|
- Geometric Interpretation
-  =
;
(i) Small A  =
; &n=
bsp;
(ii) Large A
|
24
|
|
25
|
- Cf. Wilson loop at finite temperature
|
26
|
- (3-1) Entanglement Entropy in 4D Gauge Theories
- We define the entanglement entropy in 4D Yang-Mills
- theory by dividing the space manifold into A and B as follows
|
27
|
|
28
|
- General Form of the Holographic Entanglement Entropy
|
29
|
|
30
|
- From the previous analysis, we learned the behavior
- where c is a constant. This
behavior is always true in
- any 4D CFT as is clear from the dimensional analysis.
- Moreover, the coefficient c is proportional to the central
- charge of the CFT, as far as it has its AdS dual. For large
- N gauge theories, it behaves like  =
;
.
|
31
|
- (3-2) Entanglement Entropy and Confinement
- Now we would like to move on to the confining gauge
- theories.
- When the width l is ve=
ry
small, the entanglement entropy
- probes the ultraviolet physics i.e. asymptotically free.
- On the other hand, when l is large, it probes the IR
- confinement phase (i.e. mass gap).
- Therefore, we expect that the entanglement entropy should
- detect the confinement/deconfinement transition.
|
32
|
- In summary, we expect the following behavior
|
33
|
- (3-3) Holographic Calculation
- To get a simple holographic dual of a confining gauge
- theory, we consider the Scherk-Schwarz compactification
- of AdS background.
|
34
|
- This background is unstable and should decay into
- a stable background with the same asymptotic
- AdS geometry. This stable solution is known as
- the AdS Soliton. &nbs=
p;
[Witten 98’, Horowitz-Silverstein 06’]
- &nbs=
p;
Closed string
- &nbs=
p; &=
nbsp;
tachyon
|
35
|
- The metric of AdS soliton is given by the double Wick
- rotation of the AdS black hole solution.
|
36
|
- We consider the entanglement entropy defined by
- Since the radius in the x1 direction vanishes smoothly at
- r=3Dr0, the minimal surface can end at r=3Dr0.
|
37
|
-  =
; &n=
bsp;  =
; &n=
bsp;
[Nishioka-TT 06’, Klebanov-Kutasov-Murugan 07’]=
li>
|
38
|
- In this way we learned that the entanglement entropy plays
- a role of order parameter of confinement/deconfinement
- transition.
- Our holographic analysis is convenient because we can
- detect the phase transition only from the analysis of zero
- temperature supergravity solution.
- Klebanov-Kutasov-Murgan 07’ showed that this
- phenomenon is rather general for other holographic
- solutions dual to confining gauge theories such as
- the Klebanov-Strassler solution.
-  =
; &n=
bsp;
[For global AdS, see Pando-Zayas et.al.08’]
|
39
|
|
40
|
- A Comment
- Under the
closed string tachyon condensation, we can
- show the entang=
lement
entropy decreases in our
- holographic exa=
mples
if we neglect the radiations.
-  =
;
Probably, EE always decreases as the energy
-  =
;
does under closed string tachyon condensations
-  =
;
because the radiations carry them away.
|
41
|
- So far we have obtained the evidence that the
- entanglement entropy can be an order parameter of
- confinement/deconfinement transition.
- To confirm this directly, we would like to work out how
- the holographic entanglement entropy changes under
- the Hagedorn transition.
- In particular, we consider the Hagedorn transition
- in the 4D N=3D4 super Yang-Mills theory on  =
;
.
|
42
|
- However, the direct computation of EE in the compactified
- Yang-Mills turns out to be quite hard.
- To make the calculation simpler, we introduce another
- quantity called geometric entropy, which can be regarded
- as EE after the doubl=
e Wick
rotation.
- It is simply defined as follows:
|
43
|
- We parameterize the three sphere by the metric
- The Zn orbifold is defined by  =
; &n=
bsp;
.
- If we perform a double Wick rotation, the geometric entropy
- becomes the entanglement entropy.
|
44
|
- (4-1) Free Yang-Mills Analysis
- We employ the matrix model description of the free N=3D4
- Yang-Mills. [Aharony-Marsano-Minwalla-Papadodimas-Raamsdonk 0=
3’]
- Then the orbifold partition function is expressed as follows
|
45
|
- We approximate the matrix integral by assuming that
- only the first wave mode of the eigenvalue density
- is non-vanishing.
- The geometric entropy can be found from the free energy
- by the formula
|
46
|
- In this way we plotted the geometric entropy:
- [Vanishing chemical potential]
|
47
|
- (4-2) SUGRA Analysis (strongly coupled YM)
- We can holographicaly compute the geometric entropy
- from the area of a certain minimal surface, just as we
- did for the entanglement entropy.
- Let us calculate the geometric entropy for the AdS black
- hole
- The position of the h=
orizon
is given by
|
48
|
- Since the Zn orbifold introduces the deficit angle on the
- circle defined by  =
;
, the minimal surface we want
- is extended in the  =
;
direction.
- Therefore the holographic geometric entropy becomes
- By subtracting the res=
ult
for the pure AdS5 we get
|
49
|
- In terms of the dimensionless temperature  =
;
,
- we finally obtain the result
- Since the Hagedorn transition occurs at  =
;
,
- the geometric entropy suddenly jumps from zero to
|
50
|
- Below we compare the geometric entropy from gravity
- with that of the free Yang-Mills
- As is clear from the above plots, the geometric entropy
- jumps at the transition temperature.
-  =
;
We can regard it as an order parameter.
|
51
|
|
52
|
- Example: Vaidya-AdS spacetimes (we assume 3 dim.)
- When m(v) is constant, this is the same as BTZ black hole.
- The null energy condition requires
- This background describes a process of black hole
- formation via a collapse of radiating star.
|
53
|
- Since it is asymptotically AdS3, it is dual to a CFT2 in a
- time-dependent background. Thus its entanglement
- entropy should be time-dependent.
- The holographic calculation leads to (we assume m<<1)
- The null energy condition requires that this is a
- monotonically increasing function as in the BH second law.
|
54
|
|
55
|
- The entanglement entropy can be an order parameter of
confinement/deconfinement transitions.
- We also
observed that it decreases at the end point of closed
- tac=
hyon
condensation.
- If =
we are
given a holographic SUGRA solution, we can easily
-
determined if i=
t is
confining or not.
- This viewpoint suggests that EE is a basic quantity also
- in the dual quantum gravity side.
- The area of the apparent horizon may be dual to the
-
time-dependent von-Nuemann entropy.
|
56
|
|
57
|
- Entanglement Entropy and Black Hole Entropy
- There are the following four arguments.
- The similarity of area law [Bombelli 86’, Srednicki 93’]=
- Brane-world BH [Maldacena-Hawking-Strominger 00’, Emparan 06
-  =
; &n=
bsp;  =
; &n=
bsp;  =
; &n=
bsp;  =
;
Rangamani-Hubeny-T 07’]
- (3) AdS Black Holes [Maldacena 01=
8217;,
Azeyanagi-Nishioka-T 07’]
- (4) Minimal surf=
ace
wrapped on horizon [Ryu-T 06’]
|