An Outline of String Theory

Miao Li

Institute of Theoretical Physics Beijing, China

Contents
I. Background II. Elements of string theory III. Branes in string theory IV. Black holes in string theoryholography-Maldacenas conjecture

I. Background
1. The world viewed by a reductionist
Lets start from where Feynmans lecture starts A drop of water times enlarged 10^9

H
O

Feynman was able to deduce a lot of things from a single sentence: All forms of matter consist of atoms. 1. Qualitative properties of gas, liquid 2. Evaporation, heat transport (to cool your Soup, blow it) 3. Understanding of sounds, waves

Atomic structure

Electron, point-like

H:
10^{-8}cm
Nucleus 10^{-13} cm

Theory: QED (including Lamb shift) Interaction strength:

Dirac:

QED explains all of chemistry and most of physics.

Periodic table of elements, chemical reactions, superconductors, some of biology.

Sub-atomic structure

Nucleus of H=proton
u d u

u=2/3 U(1), d=-1/3 U(1), in addition, colors of SU(3)

Neutron:
d u d

Interaction strengths
QED Size of H=Compton length of electron/=

Strong interaction
Size of proton=Compton length of quark/

So the strong interactions are truly strong, perturbative methods fail. QCD is Still unsolved

Another subatomic force: weak interaction -decay How strong (or how weak) is weak interaction?

Depends on the situation. For quarks:

-mass of u-quark -mass of W-boson

Finally, gravity, the weakest of all four interactions

-mass of proton -Planck mass

(so

Summary:

Strong interaction-SU(3) Yang-Mills

Electromagnetic Weak interaction Gravity SU(2)XU(1)

To asses the possibility of unification, lets Take a look at

2. A brief history of amalgamation of physical theories. Movement of earthly bodies. Movement of celestial bodies.
Newtonian mechanics + universal gravitation. 17th century.

Mechanics Heat, thermodynamics Atomic theory, statistical mechanics of Maxwell, Boltzmann, Gibbs, 19th century.

Electrodynamics Magnetism Light, X-rays, -rays

Faraday, Maxwell, 19th century.

Quantum electrodynamics Weak interaction Semi-unification, Weinberg-Salam model. The disparity between 10^{-2} and 10^{-6} is solved by symmetry breaking in gauge theory. 1960s-1970s (`t Hooft, Veltman, Nobel prize in 1999, total Five Nobel medals for this unification.)

Although eletro-weak, strong interaction appear as different forces, they are governed by the same universal principle: Quantum mechanics or better Qantum field theory valid up to

Further, there is evidence for unification of 3 forces:

(a) In 4 dimensions,
goes up with E goes down with E (b) runs as powers of E if there are large compact dimensions ( )

3. Difficulty with gravity

Gravity, the first ever discovered interaction, has resisted being put into the framework of quantum field theory. So, we have a great opportunity here! Why gravity is different?

There are many aspects, here is a few.

(a) The mediation particle has spin 2.

Thus

amplitude=

The next order to the Born approximation amplitude=

(b) According to Einstein theory, gravity is geometry. If geometry fluctuates violently, causal structure is lost. (c) The existence of black holes. (c1) The failure of classical geometry.

singularity

(c2) A black hole has a finite entropy, or a state of a black hole can not be specified by what is observed outside. Hawking radiation, is quantum coherence lost?

Curiously, the interaction strength at the horizon is not .

The larger the BH, the weak the interaction.

GR predicts the surface gravity be

Curiously,
Size of black hole=Compton length/

or

To summarize, the present days accepted picture of our fundamental theory is

4. The emergence of string theory A little history Strong interaction is described by QCD, however, the dual resonance model was invented to describe strong interaction first, and eventually became a candidate of theory of quantum gravity.

Initially, there appeared infinitely many resonant states ( ,,)

None of the resonant states appears more fundamental than others. In calculating an amplitude, we need to sum up all intermediate states: =
n

Denote this amplitude by A(s,t) : (a)

(b) Analytically extend A(s,t) to the complex plane of s, t, we must have

Namely

This is the famous s-t channel duality.

A simple formula satisfying (a) and (b) is the famous Veneziano amplitude

polynomial in t: t^J, J-spin of the intermediate state

linear trajectory

This remarkable formula leads us to

String theory For simplicity, consider open strings (to which Veneziano amplitude corresponds)

Ground state

v=c

v=c

An excited state

v=c

v=c

To calculate the spectrum of the excited states, We look at a simple situation (Neuman->Dirichlet)

Let the tension of the string be T, according to Heisenberg uncertainty relation

Now or

If

, then Casimir effect

The above derivation ignores factors such as 2s, s. More generally, there can be

We discovered the linear trajectory.

Morals:
(a) There are infinitely many massive states resulting from a single string (Q.M. is essential) (b) If we have only bosonic strings, no internal colors, we can have only integral spins. spin 1: gauge bosons spin 2: graviton (c) To have a massless gauge boson, a=-1. To have a massless graviton, a=-2 (need to use closed strings).

II. Elements of string theory

1. First quantized strings, Feynman rules Particle analogue Action

A classical particle travels along the shortest path, while a quantum particle can travel along different paths simultaneously, so we would like to compute

Generalization to a string

T dS

tension of the string Minkowski area element

dS

Curiously, string can propagate consistently only when the dimension of spacetime is D=26
Why is it so? We have the string spectrum

Each physical boson on the world sheet contributes to the Casimir energy an amount a=-1/24. When n=1, we obtain a spin vector field with # of degrees D-2 For

A tachyon! This breaks Lorentz invariance, so only for D=26, Lorentz invariance is maintained.

But there is a tachyon at n=0, bosonic string theory is unstable.

Unstable mode if E is complex

For a closed string

(There are two sets of D-2 modes, left moving and right moving: )

For n=2, we have a spin 2 particle, there are however only D(D-3) such states, it ought to be massless to respect Lorentz invariance, again D=26. Interactions In case of particles, use Feynman diagram to describe physical process perturbatively:

+
+

Associated to each type of vertex

more legs

there is a coupling constant The only constraint on these couplings is renormalizability. Associated with each propagator =

Or

By analogy, for string interaction + +

The remarkable fact is that for each topology there is only one diagram.

While for particles, this is not the case, for example = +

Surely, this is the origin of s-t channel duality. One can trace this back to the fact that there is unique string interaction vertex:

Rejoining or splitting

The contribution of a given diagram is

n=# of vertices = genus of the world sheet. In case of the closed strings +

Again, there is a unique diagram for each topology, the vertex is also unique = The open string theory must contain closed Strings =

The intermediate state is a closed string, unitarity requires closed strings be in the spectrum. There is a simple relation between the open string and the closed string couplings.

Emission vertex=

Now

Emission vertex=

Thus,

2. Gauge interaction and gravitation = massless open strings = massless closed strings Define the string scale

Yang-Mills coupling

by dimensional analysis. Gravitational coupling

So

If there is a compact space

D=4+d =volume of the compact space

We have

Since in 4 dimensions

, we have

Phenomenologically, scale, so .

at the unification

We see that in order to raise the string scale, say , we demand . With the advent of D-branes, in the T-dual picture this Implies
Large extra dimensions

3. Introducing fermions, supersymmetry In order to incorporate spin etc into the string spectrum, one is led to introducing fermions living on the world sheet. Again, the particle analogue is

The same as what Dirac did.

( )

Similarly, one introduces world sheet.

on the

This led to the discovery of supersymmetry for the first time in the western world (2D) (independent of Golfand and Lihktman)
Two sectors (a) Ramond sector

(b) Neveu-Schwarz sector

The Ramond sector contains spacetime fermions Zero mode

The Neveu-Schwarz sector contains bosons

Now the on-shell condition

is modified to (open string)

n-integer in R sector n-half integer in NS sector

D=10: NS: n=1/2, massless gauge bosons R: n=0, massless fermions

8 bosons + 8 fermions =supermultiplet in 10D.

Spacetime supersymmetry is a consequence.

In a way, we can say the following (a) Bosonic strings are strings moving in the ordinary spacetime , but quantum mechanics disfavors pure bosons, they are unstable.

(b) Superstrings move in superspace or , no way to avoid SUSY!

4. Five different string theories in 10 dimensions.

Consistency conditions allow for only 5 different string theories (it appears that we have a complete list, thanks to duality)

4.1 open superstring or type I string theory

Characteristics:

(a) There are open strings, whose massless modes are super Yang-Mills in 10D.

(b) As we said, there must be closed strings (unitarity). The massless modes are N=1 SUGRA in 10D.
(c) One can associate a charge to an end of an open string.

fundamental representation of G,

anti-fundamental rep of G

Combined, they form the adjoint rep of G. G can be U(N), Sp(N), SO(N). For U(N), the two ends are different, therefore one may label the orientation of the string. For Sp(N) and SO(N), the two ends are identical, thus the string is un-oriented.

(d) Further, anomaly cancellation

G= SO(32)

Type I theory is also chiral. 4.2 Closed superstring, type IIA For a closed string: and

The left movers are independent of the right movers.

or superposition of them.

two sets of

matrices.

Therefore, two basic choices One choice: chiral anti-chiral

We have type IIA superstring theory, no chirality. Thus, it appears that it has nothing to do with the real world. The massless modes = type IIA SUGRA.

4.3 Type IIB superstring theory If

chiral chiral

We have type IIB string theory, it is chiral. Although type IIB theory is chiral, it has no gauge group, it appears to be ruled out by Nature too.

4.4 Two heterotic string theories L: 10D superstring R: 26D bosonic string 26=10+16 Naively, it leads to gauge group Gauge symmetry is enhanced: or , but the

In the heterotic theory, there is only one the theory is chiral.

Remarkably, the low energy sector of the SO(32) heterotic theory is identical to that of type I theory, is this merely coincidence?
Some lessons we learned before the summer of 1994:
1. String theory is remarkably rigid, it must have SUSY, it must live in 10D. There are only 5 different theories. Even the string coupling constant is dynamical.

2. It has too many consistent vacuum solutions, to pick up one which describes our world, we have to develop nonperturbative methods. 3. It tells us that some concepts of spacetime are illusion, for instance T-duality tells us that a circle of radius R is equivalent to a circle of radius 1/R (in string unit). Sometimes, even spaces of different topologies are equivalent.
4. The theory is finite. The high energy behavior is extremely soft.

The more the energy, the larger the area S.

is small.

5. There are a lot of things unknown to us, we must be modest (such as, what about the cosmological constant?)

What we could not do before 1994:

1. Any nonpertubative calculation.

2. What happens to black holes, what happens to singularities.

3. No derivation of the standard model.

III. Branes in String/M theory 1. Why branes? In the past, it was often asked that if one can replace particles by strings, why not other branes such as membranes? The answer to this question were always:

(a) We know how to quantize particles and strings, while we inevitably end up with inconsistency in quantizing other objects.

(b) Perturbative string theory is unitary, no need to add to the spectrum other things.

Thus (a) and in particular (b) sounds like a no-go theorem.

To avoid this no-go theorem, we need to look up no other than quantum field theory. (a) In some QFT, there are solitons, these objects can be quantized indirectly by quantizing fluctuations of original fields in the soliton background.

(b) A theory may be unitary perturbatively, but nonperturbatively the S-matrix may not be unitary (showing up in resummation of a divergent series). Such inconsistency arises in particular when new stable particles exist, their masses are heavy when g is small.

Some stable particles can be associated with conservation of charge.

For example, when there is an Abelian gauge field

Happily, for a oriented closed string there is also a gauge field

Of course, when the space has a simple topology, there is no conserved charge

string

If there is a circle and the string is wrapped on it, there is a charge.

This is just conservation of winding number.

In a string theory, there is a variety of other high rank gauge fields, for instance, the so called Ramond-Ramond tensor field: But the perturbative states, strings, are not coupled to them directly. Are these fields wasted? There is a plausible argument for the existence of p-brane coupled to C .

One can always find a black-brane solution with a long-ranged

p+1
horizon r

When , there is no apparent function source for . In other words, the source is the smeared fields carried by the BH solution. This avoids the apparent paradox that perturbative fields carry no charge. If , will stop at , black brane decays, but it

A soliton charged under

, stable.

The stability is due to

(a) (b) is conserved. implies naked singularity.

The p-brane will be called D-brane, or multiple D-branes. Their tension is large when g small.

They can be viewed as a collective excitation of strings, but there is another beautiful interpretation!

2. Emergence of D-branes D is shorthand for Dirichlet. In a closed string theory, the ends of a open string are stuck on a D-brane. Namely, these ends are confined in the bulk. (The brane is like a defect in a superconductor.)

We argued that there must be fundamental branes saturating the BPS bound .

If is continuous, as the classical solution suggests, we have the trouble for accounting a continuous spectrum.
Fortunately, some time ago, it was proven that must be quantized, according to a generalized Dirac quantization condition.

Denote

dual to

rank=8-p

rank=p+2

Thus
Some unit

Both

and

are quantized.

We said that the microscopic description of a fundamental p-brane is D-brane. We now follow the route that Polchinski originally followed to see how this description emerges in string theory. 2.1 T-duality To understand the logic behind D-branes, we need to review T-duality.

There are waves on a circle:

There are also winding states on a circle:

Define a new radius such that

Then

That is, wave modes winding modes. We cannot distinguish a string theory on a circle of radius R from another string theory on a circle of radius . T-duality.

2.2 T-duality for open strings Starting with an open string theory which contains closed strings automatically. How do we map open string wave modes? An open string can couple to a gauge field tangent to a circle:

if The natural interpretation is

Thus, an open string wave mode is mapped to a winding mode with ends attached to something: D-branes. Boundary conditions on the ends of the string are Dirichlet. In the original theory momentum is conserved, thus in the dual theory winding number is conserved, the ends stick to branes.

In the original theory, winding is not conserved, no such quantum number.

2.3 Brane tension

emission absorption

Open string channel

Closed string channel

The old idea of s-t channel duality:

=
one-loop tree-level

From the open string perspective, the interaction between 2 D-branes :

Amplitude= vacuum fluctuations, independent of g

From the closed string perspective

amplitude =

But

Exact formula is

2.4 Effective theory on D-branes

Open string fluctuations longitudinal to Dbranes: gauge fields; Open string fluctuations traverse to D-branes: scalar fields; Fermions = Goldstone modes.

The position of a D-brane = vev of scalars A geometric interpretation of the Higgs mechanism:

massless

massive

3. Branes as solitonic solutions Back to the field.

(generalization of

We use the action

Postulate a solution breaking

Breaking

Further,

The solution is

When r large so When r small There is no pt-like source for . That is, the all non-linear structure of fields serve as a smeared source-just like the monopole solution in a broken gauge theory.

The mass, or rather the tension

While

It is interesting to note that there is a formal horizon:

But there is no entropy

So this black brane is more or less a pure state. We know that it is the ground state of N coincident D-branes.

4. Implications for string dualities In type IIA string theory, there is soliton with mass so How to understand the theory when There is an additional circle of radius so is a K-K mode of graviton. ? pt-like

 Type IIB theory, there is

D-string Bound states of D-strings + F-strings: (p,q)-dyonic strings. This is implied by the SL(2,Z) duality. In type I theory, there is also

Another kind of D-string, this is the heterotic string.

The list continues

Type I

SO(32) or

Heterotic

SO(32)

heterotic string

32 free fermions
16 bosons

IV. Black holes in string theory 1. Basics In real world, only a very massive collapsing body can form a black hole due to the fact that the basic matter constituents are fermions. Small black holes could (and perhaps did) form in early universe.

In an ideal situation, such as a free scalar field, any mass of black hole can form.

The typical black hole (in 4D)

No signal can escape from the horizon.

 Black hole no hair theorem Outside a black hole, one can measure only a few conserved quantities, associated to long range fields: Mass, angular momentum, charge Gravitational field, EM field

 Classical information loss

Black hole

 Bekenstein-Hawking entropy Due to the no-hair theorem and the second law of thermodynamics, a black hole must have entropy.
State 1, state 2, state 3, state 1 billion

The same black hole

An interesting theorem proven in 60s and 70s:

A = area of black hole never decreases.

Thus, S of the black hole must be ~ A So, Bekenstein reasoned S=A But, what is ?

Bekenstein argued, using an infalling massive spin particle, that . This differs from the correct value . Hawking discovered Hawking radiation and computed Use

 Thermodynamics Zero-th law: there is a temperature. First law: Second law

Third law: T=0 is impossible.

 Quantum information loss

Radiation, mixed state

2. Black holes in string theory Pre D-brane era Almost no string theorits believed in the claim of Hawking, that QM breaks down, and Einstein wins anyway.

Perturbative string theory is important in dealing with such a situation, to quote Susskind:

String theory perhaps has to solve itself before solving the information loss paradox-Scientific American.

There were a few proposals. An incomplete list:

(a) It appears that some nonlocality must be involved in order for the radiation carries away information. String theoy has some nonlocality built in. (b)

Strikingly similar to D-branes.

(c) Susskind-Horowitz-Polchinski correspondence principle For a massive string oscillation level

So But for a bh

Horowitz-Polchinski suggested (post-D-brane) that in order to form a bh, G must be tuned on. But in 4D: or for The correspondence point: for we have string and for we have a bh.

Schematically

lng
BH phase String phase

lnN Phase transition line?

3. Black holes in string theory-D-brane age 3.1 Near extremal black D-branes The pure D-brane solution

There is no entropy on the pure branes. Exciting the branes

hot gas

Near extremal black brane Thus At the horizon

Horizon area =

Specified to p=3 is independent of

Counting the entropy of a free Yang-Mills gas, one finds

The discrepancy is due to the large effective coupling on the black brane:

p=3 is called non-dilaton black brane, since

In general
For 6>p>3, theories are sufficient complex. For p=2, not much research exists For p=1, Hashimoto-Izthaki For p=0, ML

3.2 Extremal black holes (branes) Strominger-Vafa A black hole in 5D

T5: D5-branes waves

D1-branes

T4

Physical picture:
D5-D1 open strings species

The classical solutions

and other gauge fields, where The horizon volume

fixed at r=0 expands at r=0

To compute entropy, we also need

So Exact result:

Thus the # of states is

Microscopic origin: A 1D gas of open strings

In the weak coupling limit:

For a boson or a fermion:

The exact formula (Cardy) is

For a boson c=1, for a fermion c=1/2. For the system of the D1-D5 strings

This result is valid even in case of the large

by extrapolating BPS states.

Further develoments: (a) 4 charged BH in 4D. (b) Near extremal BH by adding left moving modes. (c) Hawking radiation.

The idea of Hawking radiation viewed in Dbrane picture is simple:

D-brane calculation reproduces Hawkings formula (Das-Mathur)

(d) Grey-body factor

.
Potential due to the background

Maldacena-Strominger, complete agreement. Are there magic nonrenormalization theorem?

Maldacea conjecture: The supergravity (or string theory) is dual to the CFT on the branes. The fact that the near horizon geometry is AdS is the initial strong motivation for this conjecture. In the D1-D5 case

Need large Geometry:

to have semi-classical

Need small

Another much-studied case is D3-branes, AdS5XS5:

4. Beyond D-branes 4.1 Horowitz-Polchinskis correspondence Curvature ~ String states or brane states BHs

Entropy matches ~ O(1) coefficient. No need of D-brane charges.

4.2 Matrix BH .
boost Gas of D0-branes

Qualitatively understood: Banks et al., Horowitz-Martinec, ML, ML&Martinec

But in order to compute exact coefficient, need to solve many body problem accurately.

4.3 AdS Can study near extremal BH only ( c>0 ). But provides an opportunity to study formation and evaporation of BH accurately. One may also study singularity. Technically unlikely to be solved in the near future.

Both 4.2 and 4.3 are under the influence of Dbranes.

5. BH problem is unsolved (a) Counting entropy for Schwarzschild BH honestly, accurately. (b) Dynamic process of formation of BH in Dbrane picture or AdS/CFT , information puzzle (c) Counting entropy for near-extremal BH accurately for p<3. (d) For p=3, understand .

(e) Prove the existence of gas transition.

BH phase

(d) Matrix BH need to be studied further