Where everything connects.

QFT document 23: the duality revolution. How the five string theories become one theory. T-duality, S-duality, U-duality. The emergence of 11-dimensional M-theory. Branes as fundamental objects. The web of connections that makes string theory a single framework.

Document 22 presented five consistent superstring theories in 10 dimensions. This was embarrassing; we wanted one theory of quantum gravity, not five. The 1990s “second superstring revolution” showed that the five theories are different limits of a single underlying theory, connected by a web of dualities. The unified theory, whose low-energy limit is 11-dimensional supergravity, is called M-theory.

This reframing was profound. Before dualities:

• String theory seemed like 5 competing candidates
• Non-perturbative physics was mostly inaccessible
• Strong coupling was a frontier of ignorance

After dualities:

• All 5 are unified into M-theory
• Strong-weak dualities make non-perturbative physics tractable
• What seems “elementary” in one description is “composite” in another

The dualities also revealed that strings aren’t the whole story; branes are equally fundamental objects. In M-theory, membranes (M2-branes) and 5-branes (M5-branes) are the basic building blocks; strings emerge as branes wrapped on circles.

Prerequisites

• Document 22 (string theory foundations)
• Document 19 (solitons and Montonen-Olive duality)
• Document 21 (extended SUSY)

Conventions

• Continuing from document 22
• $\ell_s = \sqrt{\alpha'}$ (string length)
• $g_s$ (string coupling)
• $\ell_P = g_s^{1/4}\ell_s$ (10D Planck length)
• For 11D: $\ell_{11}$ (11D Planck length)

Table of Contents

1. What Is a Duality?
2. T-Duality: The First Duality
3. T-Duality Connects IIA and IIB
4. S-Duality: Strong-Weak Coupling
5. Type IIB S-Duality and $SL(2,\mathbb{Z})$
6. The Heterotic/Type I Duality
7. The Emergence of 11 Dimensions
8. M-Theory and 11D Supergravity
9. The Duality Web
10. Branes in M-Theory
11. BPS States and Exact Results
12. Beyond Perturbative Strings
13. Appendix: Duality Reference

1. What Is a Duality?

The Concept

A duality is an equivalence between two seemingly different physical theories. They describe the same physics; same observables, same physical predictions; but use different mathematical descriptions, different “elementary” objects, and often different couplings.

Key examples we’ve already seen:

• Electric-magnetic duality (document 19): $F \leftrightarrow \tilde F$ in pure electromagnetism
• Montonen-Olive duality (document 21): $\mathcal{N}=4$ SYM at coupling $g$ ↔ $\mathcal{N}=4$ SYM at $4\pi/g$
• Seiberg duality (document 21): $SU(N_c)$ ↔ $SU(N_f - N_c)$ in $\mathcal{N}=1$

Why Dualities Matter

They make strong coupling tractable. A strongly-coupled theory in one description is weakly-coupled in its dual. Calculations that are impossible in one become straightforward in the other.

They unify seemingly different theories. What look like distinct fundamental structures are often different faces of the same underlying thing.

They constrain physics. Any proposed theory must be consistent with all its dualities. This is a powerful filter.

They redefine “elementary.” In one description, electron is fundamental, monopole is composite. In the dual, monopole is fundamental, electron is composite. Neither is more real.

Types of Duality in String Theory

T-duality: Relates string theories on compactifications of different sizes. Small compactification radius ↔ large compactification radius.

S-duality: Relates theories at strong coupling to theories at weak coupling. $g_s \leftrightarrow 1/g_s$ (or similar).

U-duality: Combined T + S dualities. Relates more dramatic-seeming transformations.

Gauge/gravity duality (AdS/CFT): (Document 24); connects gauge theories to gravitational theories in higher dimensions.

String theory’s duality web involves all of these, with beautiful structural relations between them.

2. T-Duality: The First Duality

The Setup

Consider a string theory with one spatial dimension compactified on a circle of radius $R$. Points $x$ and $x + 2\pi R$ are identified.

A closed string can:

1. Carry momentum in the compact direction: momentum quantized as $p = n/R$ for integer $n$
2. Wind around the circle: winding number $w$, contributing an additional term to the energy

The total string energy is:

$E^2 = M^2 + \left(\frac{n}{R}\right)^2 + \left(\frac{wR}{\alpha'}\right)^2$

Plus oscillator contributions.

The Symmetry

T-duality exchanges momentum and winding:

$n \leftrightarrow w, \quad R \leftrightarrow \frac{\alpha'}{R} = \frac{\ell_s^2}{R}$

Under this exchange, the energy formula is invariant. The physics at radius $R$ is the same as the physics at radius $\alpha'/R$!

The self-dual radius is $R_{\rm self} = \sqrt{\alpha'} = \ell_s$, where $R = \alpha'/R$ is automatic.

The Physical Meaning

T-duality is genuinely surprising:

A very small circle is physically the same as a very large circle. When $R \to 0$: strings can wind many times at low energy (small $wR/\alpha'$). From this dual perspective, the dimension looks “large”; it’s traded for winding modes.

There’s no smallest length. Below $R = \ell_s$, you’re just probing the “other side” of the T-dual; physics at $R < \ell_s$ equals physics at $R > \ell_s$. The string length is a natural minimum distance.

This is utterly unlike point-particle physics. Point particles have momentum but not winding. They see arbitrarily small distances. Strings, because they can wind, see a minimum length.

Deriving T-Duality

For the simplest case (flat compactification, no D-branes), T-duality can be derived from a change of worldsheet coordinates:

$X^{25}_L \to X^{25}_L, \quad X^{25}_R \to -X^{25}_R$

(Keep left-movers, flip right-movers in the compactified direction.) This transformation preserves the energy-momentum tensor (hence is a symmetry of the worldsheet theory) but maps $R \to \alpha'/R$ and $n \leftrightarrow w$.

The derivation extends to more complex cases (tori, fibered compactifications, with D-branes, etc.), though with additional subtleties.

T-Duality for Open Strings

For open strings, T-duality is more subtle:

In the direction $X$ where we T-dualize: open string endpoints must satisfy Dirichlet boundary conditions instead of Neumann. The endpoints become fixed.

This is the physical meaning of D-branes! T-dualizing Type IIA gives Type IIB (and vice versa), and $Dp$-branes get mapped to $D(p\pm 1)$-branes depending on whether you T-dualize along or transverse to the brane.

T-duality and D-branes are intimately related; one can’t be separated from the other.

3. T-Duality Connects IIA and IIB

The Key Relation

Type IIA string on a circle of radius $R$ ↔ Type IIB string on a circle of radius $\alpha'/R$

These are literally the same theory; just described with different radii and different T-duality conventions.

What Happens to Branes

T-duality acts on D-branes in a specific way. If we T-dualize along a direction the brane wraps:

• $Dp$-brane wrapping the circle → $D(p-1)$-brane (no longer wrapping)
• $Dp$-brane transverse to the circle → $D(p+1)$-brane (now wrapping)

So T-duality maps:

• IIA D0 → IIB D1 (if you T-dualize a direction)
• IIA D2 → IIB D3
• IIA D4 → IIB D5
• IIA D6 → IIB D7

This is why IIA has even-dimensional D-branes and IIB has odd-dimensional ones; they’re T-dual descriptions.

What Happens to SUSY

IIA has $\mathcal{N}=2$ non-chiral SUSY. IIB has $\mathcal{N}=2$ chiral SUSY. T-duality maps one to the other because the action on fermion zero modes on the worldsheet flips the chirality of one of the supercharges.

Physical Consequence

There’s no distinction between “IIA at radius $R$” and “IIB at radius $\alpha'/R$.” They’re the same theory. What looked like 5 separate string theories is actually fewer because IIA and IIB are connected by T-duality.

T-Duality in More General Settings

T-duality extends to:

Tori: String on $T^n$ at various radii → string on $T^n$ at T-dual radii. The T-duality group becomes $O(n, n; \mathbb{Z})$.

Orbifolds and orientifolds: Related string theories with identifications.

Non-compact fibered geometries: Mirror symmetry (for Calabi-Yau compactifications) is essentially T-duality applied to a fibration. IIA on a Calabi-Yau is T-dual to IIB on the “mirror” Calabi-Yau (SYZ conjecture).

The “mirror symmetry” conjecture connects IIA and IIB compactifications in deep, mathematically rich ways.

4. S-Duality: Strong-Weak Coupling

The Concept

S-duality relates two theories where the coupling constant is inverted:

$g_s \leftrightarrow \frac{1}{g_s}$

(or a variant thereof). Strong coupling in one description is weak coupling in the dual.

Why This Helps

String theory perturbation theory is an expansion in $g_s$:

$\text{Amplitude} = \sum_n g_s^{2n-2}\mathcal{A}_n$

At weak coupling ($g_s \ll 1$): perturbation theory works. Tree-level ($n = 0$) dominates, loops ($n \geq 1$) are suppressed.

At strong coupling ($g_s \gg 1$): perturbation theory breaks down. The series doesn’t converge in any useful sense.

With S-duality: strong coupling in one description = weak coupling in the dual. You can still do perturbation theory, just in the dual variables.

Conditions for S-Duality

Not every theory has an S-duality. You need specific structure; usually extended SUSY helps. Some examples:

$\mathcal{N}=4$ SYM (document 21): Exact S-duality from Montonen-Olive conjecture. $\tau \leftrightarrow -1/\tau$.

Type IIB string theory: S-duality is exact (see next section).

Type I / Heterotic $SO(32)$: S-dual to each other.

Type IIA at strong coupling: Grows an extra dimension, becomes M-theory (section 7).

Heterotic $E_8 \times E_8$ at strong coupling: Also grows an extra dimension (interval), becomes Hořava-Witten theory.

Each of these S-dualities is a specific exact relation that’s been extensively checked.

How S-Duality Is Established

For most string S-dualities, the evidence includes:

BPS spectrum matching. Electric and magnetic BPS states must pair up consistently under $g_s \to 1/g_s$.

Low-energy supergravity matching. The 10D supergravity actions are related by specific field redefinitions.

Compactification consistency. Compactifying both sides gives matching dual 4D or lower-dimensional theories.

String amplitudes. In specific cases (like constant-$\tau$ slices of Type IIB), explicit computations confirm duality.

Anomaly cancellation. Dualities must preserve anomaly structures.

Multiple independent checks typically converge on the same duality relation.

5. Type IIB S-Duality and $SL(2,\mathbb{Z})$

The Axion-Dilaton

In Type IIB, there are two massless scalar fields:

• Dilaton $\Phi$, controlling the string coupling via $g_s = e^{\Phi}$
• Axion (RR 0-form) $C_0$

Combine them into the axion-dilaton:

$\tau = C_0 + \frac{i}{g_s}$

This is a complex number that parameterizes vacua of Type IIB.

The Duality Symmetry

Type IIB has an exact $SL(2, \mathbb{Z})$ duality group acting on $\tau$:

$\tau \to \frac{a\tau + b}{c\tau + d}, \quad \begin{pmatrix}a & b \\ c & d\end{pmatrix} \in SL(2, \mathbb{Z})$

This means Type IIB at coupling $\tau$ is literally the same theory as Type IIB at coupling $\tau' = (a\tau + b)/(c\tau + d)$.

The Generators

The $SL(2, \mathbb{Z})$ is generated by two transformations:

$T$ transformation: $\tau \to \tau + 1$. This shifts the axion $C_0 \to C_0 + 1$. Physically trivial (just a shift of a periodic field).

$S$ transformation: $\tau \to -1/\tau$. When $C_0 = 0$: this gives $g_s \to 1/g_s$. This is the “strong-weak duality” piece.

What Gets Mapped

Under $S$ transformation ($\tau \to -1/\tau$):

• F-string (fundamental string) ↔ D1-string (D-brane string)
• D3-brane is self-dual (invariant)
• NS5-brane ↔ D5-brane
• D7-brane ↔ appropriate other 7-brane

So “the fundamental string” isn’t uniquely defined; it’s just one among a family of 1-dimensional objects, with others being D1-strings and more. Under duality, any of them could play the role of “fundamental.”

(p, q)-Strings

More generally, there’s a whole family of strings labeled by coprime integers $(p, q)$:

• $(1, 0)$: fundamental string
• $(0, 1)$: D1-string
• $(p, q)$: a bound state of $p$ F-strings and $q$ D1-strings

These $(p, q)$-strings transform into each other under $SL(2, \mathbb{Z})$. The $(1, 0)$ → $(p, q)$ transformation uses the element $\begin{pmatrix}p & * \\ q & *\end{pmatrix}$ (with appropriate second column making determinant 1).

This web of strings is all unified by the $SL(2, \mathbb{Z})$ symmetry.

Exact Non-Perturbative Formulation

Type IIB has S-duality as an exact symmetry, not just an approximate statement. This means Type IIB is defined non-perturbatively, at least in principle. The full Type IIB theory is invariant under $\tau \to -1/\tau$; not just the low-energy limit.

This is one reason AdS/CFT works: Type IIB on $AdS_5 \times S^5$ has the full $SL(2, \mathbb{Z})$ duality, which corresponds to the $SL(2, \mathbb{Z})$ S-duality of $\mathcal{N}=4$ SYM.

6. The Heterotic/Type I Duality

The Relation

Another S-duality:

Heterotic $SO(32)$ at coupling $g_H$ ↔ Type I at coupling $g_I = 1/g_H$

Both have gauge group $SO(32)$, but they were thought to be completely different theories. Heterotic has only closed strings; Type I has both open and closed. Their massless spectra look different.

Yet at strong coupling in one, weak coupling in the other, they describe the same physics.

The Physical Picture

When you take Heterotic $SO(32)$ to strong coupling:

• The heterotic string itself becomes heavy (its mass scales up)
• What was a loop expansion becomes dominated by non-perturbative objects
• These objects turn out to be… Type I D-branes and closed strings!

The “solitonic spectrum” of one theory is the “perturbative spectrum” of the other.

Evidence

The matching of BPS spectra:

• Heterotic: has specific BPS states (related to T-duality spectrum)
• Type I: has D-strings, D5-branes, etc.
• Under S-duality: these match up consistently

The low-energy $\mathcal{N}=1$ 10D supergravity + SYM actions are related by specific field redefinitions that implement the duality.

What This Tells Us

The Heterotic/Type I duality suggests:

• Two apparently different string constructions (open vs. closed; different gauge structures) can give the same physics
• Non-perturbative string dynamics unifies what perturbation theory separates
• Our sense of what’s “the string” is itself perturbative; non-perturbatively, strings aren’t even uniquely defined

7. The Emergence of 11 Dimensions

The Setup

Type IIA has a parameter $g_s$ (the string coupling). What happens as $g_s \to \infty$?

At first glance: this is a strongly-coupled regime where everything is hard. But the answer is surprising.

D0-Branes and the Extra Dimension

In Type IIA, there are $D0$-branes; point-like objects with mass $m_{D0} = 1/g_s\ell_s$.

At weak coupling ($g_s \ll 1$): $D0$-branes are very heavy, essentially invisible at low energies.

At strong coupling ($g_s \to \infty$): $D0$-brane mass goes to zero!

A massless state appears in the spectrum at strong coupling. This is unusual; normally you expect a UV completion that just handles the strong coupling, not new massless states.

The Kaluza-Klein Interpretation

The light $D0$-branes behave like Kaluza-Klein momentum states of a compactified theory. Specifically:

$m_{D0} = \frac{n}{R_{11}}$

where $R_{11}$ is the radius of an emerging 11th dimension!

Matching $1/(g_s\ell_s)$ to $1/R_{11}$ gives:

$R_{11} = g_s\ell_s$

So the 11th dimension is small at weak string coupling and grows at strong coupling. At $g_s \to \infty$, this extra dimension becomes macroscopic.

The Conclusion

Type IIA string theory at strong coupling is actually an 11-dimensional theory. The “strongly coupled” regime isn’t a strange limit; it’s just a different corner of the parameter space where an extra dimension opens up.

The 11D theory whose 10D limit (with small extra dimension) gives Type IIA is called M-theory.

Hořava-Witten

Similarly, Heterotic $E_8 \times E_8$ at strong coupling grows an 11th dimension. But this dimension is an interval (not a circle), with $E_8$ gauge fields living on each of the two boundaries.

This is Hořava-Witten theory (1996). It connects heterotic $E_8 \times E_8$ to M-theory on $\mathbb{R}^{10}\times S^1/\mathbb{Z}_2$ (an orbifold circle = interval).

8. M-Theory and 11D Supergravity

What Is M-Theory?

M-theory is the 11-dimensional theory that has Type IIA and Heterotic $E_8\times E_8$ string theories as specific 10D limits. It’s the unified framework containing all 5 superstring theories as different 10D projections or limits.

The “M” is deliberately ambiguous; originally proposed by Witten as standing for “Magic,” “Mystery,” or “Membrane,” reflecting that the theory wasn’t (and still isn’t) fully understood.

11D Supergravity

At low energies (below the 11D Planck scale), M-theory reduces to 11-dimensional supergravity; a well-defined field theory:

$\mathcal{L}_{11D} = \sqrt{-g}R - \tfrac{1}{2}|F_4|^2 - \text{(CS term)} + \text{fermions}$

where $F_4 = dA_3$ is the field strength of a 3-form gauge field $A_3$, and there’s a specific Chern-Simons term.

Particle content:

• Graviton (metric in 11D)
• Three-form gauge field $A_3$ (bosonic)
• Gravitino $\Psi^\mu_\alpha$ (fermionic)

11D has maximum SUSY: 32 real supercharges (since max in any dim is 32). This is $\mathcal{N} = 1$ in 11D.

Uniqueness

11D supergravity is the maximum-dimensional SUSY theory allowed. In higher dimensions, SUSY algebras force spins $>2$ to appear, which can’t be consistent (weakly-coupled) field theories.

So 11D is the “top” of the SUSY tower. M-theory lives there.

Relating 11D to 10D Strings

To Type IIA: Compactify 11D M-theory on a circle $S^1$. The radius $R_{11}$ is related to Type IIA coupling via $R_{11} = g_s\ell_s$.

To Heterotic $E_8\times E_8$: Compactify M-theory on $S^1/\mathbb{Z}_2$ (an interval). Each boundary carries $E_8$ gauge fields.

These aren’t two different theories; they’re two different compactifications of the same 11D M-theory.

M-Theory on Calabi-Yau

For 4D physics:

• Compactify M-theory on a $G_2$-holonomy manifold (7-dimensional, preserves $\mathcal{N}=1$ in 4D)
• Or compactify on Calabi-Yau 3-fold × circle, giving $\mathcal{N}=2$ in 4D

The $G_2$ compactification is favored for phenomenology. It gives chiral fermions and the right symmetry structure.

M-Theory Is Not Yet Fully Formulated

Here’s the honest situation: M-theory is a conjectural 11D theory, defined by:

• Its low-energy limit (11D supergravity)
• Its dualities (relations to 10D string theories)
• Various limits and compactifications

But a complete non-perturbative definition doesn’t exist. We don’t have “the M-theory Lagrangian” or a first-principles construction.

Various proposals:

• Matrix theory (BFSS): 11D M-theory as a matrix model in specific limits
• $(2,0)$ theory: M5-brane worldvolume theory as a definition of 6D $(2,0)$ theory, which has M-theoretic significance
• AdS/CFT approaches: Define M-theory via its boundary CFT in $AdS_7\times S^4$ or $AdS_4\times S^7$

Each has merits, but a complete non-perturbative definition of M-theory remains an open problem.

9. The Duality Web

The Picture

Here’s how the 5 superstring theories + M-theory are connected:

                    M-THEORY (11D)
                    /          \
                   /            \
              Compactify     Compactify
              on S^1          on S^1/Z_2
              /                  \
      Type IIA (10D)       Heterotic E8xE8 (10D)
         |                        |
      T-duality              T-duality
         |                        |
      Type IIB (10D)       Heterotic SO(32) (10D)
                                  |
                                S-duality
                                  |
                                Type I (10D)
                                  |
                              also S-dual to
                                  |
                           Heterotic SO(32)

Plus:

• T-duality connects IIA and IIB
• S-duality connects Type I and Heterotic $SO(32)$
• Various compactifications provide more connections
• Mirror symmetry connects IIA/IIB on Calabi-Yau pairs

Every theory is connected to every other by some chain of dualities.

U-Duality

For compactified theories, T and S dualities combine into U-duality:

$U = T\cdot S\cdot T\cdot S\cdot \ldots$

The combined symmetry group depends on the compactification. For Type IIB on $T^5$ (5-torus), U-duality is $E_6(\mathbb{Z})$. For Type IIB on $T^6$: $E_7(\mathbb{Z})$. For Type IIB on $T^7$: $E_8(\mathbb{Z})$.

These large exceptional symmetry groups are striking. They constrain the physics powerfully.

What This Unification Means

There’s one underlying theory. What looked like 5 string theories (and M-theory) is one structure with different perturbative descriptions. Each theory is a weak-coupling expansion around a different point in the full parameter space.

Non-perturbative physics is partially understood. Dualities give us handles on strongly-coupled regimes we’d otherwise be unable to touch.

Predictions must be consistent across dualities. Any observable computed in one description must match any dual description. This is an enormously powerful consistency constraint.

Does This Help With Our Universe?

Unfortunately, dualities don’t directly help select “the” 4D vacuum. All the dualities we’ve discussed are in 10D or 11D. Compactifying to 4D is where the landscape problem enters.

But dualities do help:

• Constrain the structure of 4D theories that can come from strings
• Identify which 4D observables are “non-perturbatively stable”
• Provide handles on strong-coupling dynamics in specific compactifications

Dualities are a tool, not an answer, for the phenomenological problem.

10. Branes in M-Theory

Basic M-Theory Objects

M-theory has a different fundamental structure than strings. Instead of 1-dimensional strings, it has:

M2-branes (also called membranes): 2+1-dimensional objects (2 space + 1 time). These are the “fundamental” objects in M-theory in some sense.

M5-branes: 5+1-dimensional objects. Also fundamental.

How M2 and M5 Give Rise to Strings

When M-theory is compactified on a small circle (giving Type IIA):

M2-brane wrapping the circle → Type IIA F-string (fundamental string). The 2D brane, with one dimension wrapped, looks like a 1D string in 10D.

M2-brane not wrapping → Type IIA D2-brane. A 2D object in 10D.

M5-brane wrapping the circle → Type IIA D4-brane (5D brane becomes 4D).

M5-brane not wrapping → Type IIA NS5-brane (5D brane, different from D-brane).

So strings in Type IIA are really M2-branes in disguise; M-theory membranes that happen to wrap the compactified 11th dimension.

The M5-Brane Worldvolume Theory

$N$ coincident M5-branes support a 6D worldvolume theory with extraordinary properties:

• 6 dimensions (5 spatial + 1 time)
• $\mathcal{N}=(2,0)$ superconformal symmetry (3 BPS bounds, central charges)
• Non-abelian “tensor” gauge fields (not just vector gauge fields; more exotic)
• No Lagrangian description in terms of standard fields
• Essential for understanding 4D $\mathcal{N}=2$ dualities via dimensional reduction

The $(2,0)$ theory is a mysterious but central object in modern theoretical physics. It lacks a standard Lagrangian but is uniquely characterized by its symmetries.

Branes Wrapping Cycles

In compactifications, M-theory branes wrap cycles in the compactification manifold:

• M2-brane wrapping a 2-cycle in $G_2$ manifold → particle in 4D
• M2-brane wrapping a 1-cycle → string in 4D
• M5-brane wrapping a 4-cycle → particle in 4D (a different kind)
• M5-brane wrapping a 2-cycle → domain wall in 4D

Different brane configurations and cycle choices give different 4D states.

This is geometric engineering: physical states in 4D come from geometrical data (cycles, wrapping) in the compactification manifold.

Examples

$G_2$-manifold compactifications of M-theory give phenomenologically interesting 4D theories:

• Chiral fermion generations from M5-branes wrapping 3-cycles
• Gauge fields from M-theory gauge connections
• Yukawa couplings from triple intersections of cycles

Realistic models require specific $G_2$-manifolds with appropriate topology. This is an active area of string phenomenology research.

11. BPS States and Exact Results

Dualities Must Preserve BPS Spectrum

Under any duality, BPS states must be preserved (at least at the level of their mass formulas and charges). The BPS states of Type IIA must correspond to BPS states of Type IIB under T-duality, etc.

This spectrum matching is one of the strongest tests of duality conjectures. If BPS spectra didn’t match, the dualities would be wrong.

BPS State Counting

In many cases, BPS state counting is:

• Protected by SUSY: quantum corrections don’t change the count
• Computable in either dual description: dualities give multiple ways to calculate
• Connected to topological invariants: Donaldson invariants, Gromov-Witten invariants, etc.

This makes BPS counting a powerful tool. Mathematical results from physics:

Strominger-Vafa (1996): BPS black hole entropy from D-brane microstate counting. The number of BPS D-brane configurations matches the Bekenstein-Hawking entropy formula; a major success for string theory as quantum gravity (document 25).

Maldacena-Strominger-Witten: Black hole entropy and topological string theory connections.

Gaiotto-Moore-Neitzke: Wall-crossing formulas for BPS spectra in $\mathcal{N}=2$ gauge theories, connecting to cluster algebras and integrable systems.

Exact Non-Perturbative Results

Combining:

• SUSY constraints (especially extended SUSY)
• Holomorphy (from document 21)
• Dualities (from this document)

One can compute many quantities exactly in SUSY string compactifications:

• Effective superpotentials
• Non-perturbative gauge couplings
• Moduli space geometries
• Correlation functions of BPS operators

These exact results make SUSY string compactifications a remarkable laboratory for quantum gravity.

Mathematical Impact

The BPS-duality framework has had enormous impact on mathematics:

• Mirror symmetry: relates Calabi-Yau pairs, discovered through string duality
• Donaldson-Thomas theory: counting sheaves on Calabi-Yau, inspired by string theory
• Gromov-Witten invariants: counting holomorphic curves, physical origin in string amplitudes
• Knot theory: Chern-Simons invariants of knots via WZW models and strings
• Representation theory: insights from BPS spectra and dualities

Many Fields Medals (Witten, Kontsevich, Mirzakhani, Okounkov) have been awarded for work with deep string-theoretic origins.

12. Beyond Perturbative Strings

What Perturbation Theory Tells Us

The perturbative string is well-defined:

• Scattering amplitudes at any loop order (in principle)
• String spectrum exact
• Supergravity corrections from massive string states

But perturbation theory is an expansion in $g_s$. It doesn’t tell us about:

• Strong coupling dynamics
• Vacuum selection
• Black hole microstates (mostly)
• Non-perturbative effects like instantons

Dualities as Non-Perturbative Tools

Dualities bridge this gap. By combining perturbative calculations in different dual descriptions, we can piece together the non-perturbative structure.

Example: Computing the entropy of a black hole in 10D Type IIA.

In 10D Type IIA perturbation theory: black hole is a complicated object, hard to count states.

Via S-duality to 11D M-theory: black hole becomes a simpler M2- or M5-brane configuration wrapping cycles. Counting states on the brane side gives the entropy.

This is how Strominger-Vafa counted black hole microstates.

What’s Still Missing

A first-principles non-perturbative formulation. Something like “the M-theory equation” that defines everything.

Vacuum selection. How to pick a specific 4D vacuum matching our universe.

Understanding dS vacua. Our universe has $\Lambda > 0$. Constructing stable dS in string theory is controversial; some argue it’s impossible.

Experimental tests. Strings at the Planck scale are far from direct detection. Indirect tests (SUSY at TeV, modified gravity, etc.) have found nothing.

Conceptual clarity on what M-theory is. The 11D theory is defined by its low-energy limit and dualities, but a complete characterization is elusive.

The State of the Field

String theory has been enormously productive as a framework:

• Finite perturbative gravity
• AdS/CFT and holography (document 24)
• Black hole microstate counting
• Mathematical physics breakthroughs
• Insights into quantum field theory dualities

But the grand question; “Is string theory the correct description of nature?”; remains open. We don’t know if the full theory makes definite predictions for our universe, or if it’s consistent with observed cosmology (dS vacua), or if it’s uniquely selected by some mathematical principle.

This doesn’t diminish its importance; it’s the most developed candidate framework for quantum gravity. But honest intellectual accounting requires acknowledging the uncertainties.

13. Appendix: Duality Reference

T-Duality

$R \leftrightarrow \frac{\alpha'}{R} = \frac{\ell_s^2}{R}, \quad n \leftrightarrow w$

Connects:

• Type IIA on circle ↔ Type IIB on T-dual circle
• Heterotic $E_8\times E_8$ on torus ↔ Heterotic $SO(32)$ on T-dual torus

S-Duality

$g_s \leftrightarrow \frac{1}{g_s}$

Connects:

• Type IIB ↔ Type IIB (exact $SL(2,\mathbb{Z})$)
• Heterotic $SO(32)$ ↔ Type I
• Type IIA at strong coupling ↔ 11D M-theory

Dimensions

M-Theory ↔ String Connections

• Type IIA = M-theory on $S^1$ with $R_{11} = g_s\ell_s$
• Heterotic $E_8\times E_8$ = M-theory on $S^1/\mathbb{Z}_2$ (interval) with $E_8$ on each boundary

Compactifications

• M-theory on $G_2$: $\mathcal{N}=1$ in 4D
• Type II on Calabi-Yau: $\mathcal{N}=2$ in 4D
• Heterotic on Calabi-Yau: $\mathcal{N}=1$ in 4D (phenomenologically favored)
• F-theory on Calabi-Yau 4-fold: various 4D SUSY

Brane-Particle Correspondence (M-theory)

• M2-brane wrapping 1-cycle in $G_2$: 1D object in 4D
• M2-brane wrapping 2-cycle: 0D (particle) in 4D
• M5-brane wrapping 4-cycle: 0D (particle) in 4D
• M5-brane wrapping 3-cycle: 1D object in 4D

U-Duality Groups

For Type IIB on $T^d$ (d-torus): U-duality group is $E_{d+1}(\mathbb{Z})$ (exceptional Lie group at integer points).

Examples:

• $T^5$: $E_6(\mathbb{Z})$
• $T^6$: $E_7(\mathbb{Z})$
• $T^7$: $E_8(\mathbb{Z})$

Further Reading

• Becker, Becker, Schwarz, String Theory and M-Theory (2007): covers all of this
• Polchinski, String Theory Vol. 2: classic rigorous treatment
• Schwarz’s lectures on M-theory: original perspectives
• Witten, String Theory Dynamics in Various Dimensions (1995): foundational paper
• Townsend, The Eleven-Dimensional Supermembrane Revisited (1995): M-theory dimensions
• Hořava & Witten, Heterotic and Type I String Dynamics From Eleven Dimensions (1996): HW theory

Problems

1. For Type IIA compactified on a circle, derive the relation $R_{11} = g_s\ell_s$ between the 11th dimension and the string coupling.

2. Starting from the 11D supergravity Lagrangian, compute the 10D Type IIA Lagrangian via Kaluza-Klein reduction.

3. Show that T-duality on Type IIA (acting on one circle) gives Type IIB, by tracking the spectrum of massless states.

4. For Type IIB, verify that $SL(2,\mathbb{Z})$ acts correctly on the axion-dilaton and on the set of $(p,q)$-strings.

5. Using heterotic-Type I duality, relate specific massive states: show how a heterotic string state at strong coupling corresponds to a Type I D-string state at weak coupling.

6. In M-theory on $G_2$ manifold, explain how different chiral generations arise from M5-branes wrapping different 3-cycles.

Closing Note

The duality revolution transformed string theory from “5 competing candidates” into “one unified framework.” M-theory in 11 dimensions is the underlying structure, with the 5 superstring theories as specific 10D limits.

What You Now Know

• T-duality, S-duality, U-duality as the key relations
• How Type IIA and IIB are T-dual
• How Type I and Heterotic $SO(32)$ are S-dual
• How strong-coupling Type IIA grows an 11th dimension → M-theory
• 11D supergravity as the low-energy limit of M-theory
• M2-branes and M5-branes as M-theoretic fundamentals
• The duality web connecting everything

The Bigger Picture

String theory is a single theory (M-theory) with many descriptions. Each 10D string theory is a weak-coupling expansion around a specific corner. Dualities tie them all together.

This unification is the central conceptual breakthrough of the second superstring revolution (1995-). It transformed string theory from “5 string theories in need of selection” to “1 framework with many dual descriptions.”

What’s Next

Document 24 covers AdS/CFT and holography; the most concrete realization of string theory’s framework. $\mathcal{N}=4$ SYM at strong coupling is equivalent to classical gravity on $AdS_5\times S^5$. This has:

• Provided quantitative predictions for the quark-gluon plasma
• Inspired “holographic” approaches to condensed matter
• Given quantum gravity a well-defined framework on AdS backgrounds
• Enabled progress on the black hole information paradox

This is where the abstract structures of documents 20-23 become a concrete, calculationally useful tool. It’s arguably the most impactful theoretical development of the past 30 years.

After that: black holes and the information paradox (doc 25), and a survey of quantum gravity approaches (doc 26).

We’re well into the deep territory now. Take a break if you need one. Four more docs in the BSM sequence, and we’ll have covered the core of modern theoretical physics speculation.