AdS/CFT Correspondence

Two theories that could not look less alike turn out, the conjecture says, to be one theory written in two languages. On one side is quantum gravity — strings, gravitons, and black holes — living inside a negatively curved spacetime that wraps back on itself, anti-de Sitter space. On the other side is an ordinary quantum field theory with no gravity at all, painted onto the lower-dimensional boundary of that spacetime, yet holding every last bit of the bulk's gravitational information. When Juan Maldacena wrote this down in 1997 he made the holographic principle concrete and, crucially, calculable: a theory of gravity in $(d+1)$ dimensions is exactly a conformal field theory in $d$ . The prototype is sharp. Take $N$ coincident D3-branes in type IIB string theory; their near-horizon geometry is $\mathrm{AdS}_5\times S^5$ , and the theory on their worldvolume is four-dimensional $\mathcal{N}=4$ super Yang-Mills with gauge group $SU(N)$ . The two descriptions are the same physics, so they must agree, and the parameters lock as $L^4/\alpha'^2=g_{\mathrm{YM}}^2N\equiv\lambda$ . The operational heart is the dictionary of Gubser, Klebanov, Polyakov and Witten: each bulk field's boundary value sources a boundary operator, the two partition functions are equal, and the on-shell gravitational action is the field theory's generating functional. Because it is a strong/weak duality, classical Einstein gravity computes strongly coupled quantum matter nothing else can touch — the shear viscosity of a quark-gluon plasma, the entropy of a black hole, the entanglement of a boundary region. No proof exists; the mountain of passed tests is why almost everyone believes it anyway.

Key takeaways

The AdS/CFT correspondence conjectures that a theory of quantum gravity in $(d+1)$ -dimensional anti-de Sitter space is exactly equivalent to a conformal field theory living on its $d$ -dimensional boundary — the sharpest realization of holography.
The prototype is type IIB string theory on $\mathrm{AdS}_5\times S^5$ dual to $\mathcal{N}=4$ super Yang-Mills with gauge group $SU(N)$ , with parameters locked by $L^4/\alpha'^2=g_{\mathrm{YM}}^2N=4\pi g_sN$ .
The GKP-Witten dictionary equates the bulk partition function with the CFT generating functional; a bulk field of mass $m$ is dual to an operator of dimension $\Delta$ obeying $m^2L^2=\Delta(\Delta-d)$ , and the radial coordinate is the renormalization-group scale.
It is a strong/weak duality: classical supergravity (valid at $N\to\infty$ , $\lambda\to\infty$ ) computes the strongly coupled, large- $N$ field theory that perturbation theory cannot reach.
Concrete payoffs include the universal ratio $\eta/s=\hbar/(4\pi k_B)$ , the Bekenstein-Hawking entropy of AdS black holes from boundary states, and the Ryu-Takayanagi identification of entanglement with minimal-surface area.
Though unproven, it passes countless quantitative tests and now links quantum gravity, field theory, condensed matter, and quantum information.
Key idea: a gravitating bulk in $(d+1)$ -dimensional anti-de Sitter spacetime is exactly equivalent to a non-gravitating conformal field theory on its $d$ -dimensional boundary — the gravitational partition function equals the CFT generating functional, $Z_{\mathrm{grav}}[\phi_0]=\langle e^{\int_{\partial}\phi_0\,\mathcal{O}}\rangle_{\mathrm{CFT}}$ , with bulk boundary data sourcing boundary operators.

Where this lives today

AdS/CFT has stopped being a conjecture people merely test and become a tool people use. In heavy-ion physics it supplies the benchmark $\eta/s=1/4\pi$ against which the "perfect liquid" quark-gluon plasma at RHIC and the LHC is measured, with holographic hydrodynamics predicting higher transport coefficients that feed collision simulations. In quantum information the Ryu-Takayanagi formula turned "spacetime is built from entanglement" into a research program, realizing the bulk as a quantum error-correcting code and resolving the black-hole information paradox — via the island formula and the Page curve — inside a manifestly unitary boundary theory. In condensed matter, "AdS/CMT" models strange metals and quantum-critical transport where quasiparticles have dissolved. And the deepest problem it sharpened is still open: how to take the flat-space limit $L\to\infty$ and read an ordinary scattering amplitude — or our own nearly-flat universe — out of a holographic description.

The physical picture

Picture the bulk as a stack of nested copies of the boundary field theory, one at each radial depth, each coarse-grained to a coarser energy scale than the last. The boundary at infinity is the ultraviolet, where the resolution is sharpest; moving inward integrates out short-distance modes, so the radial direction literally is renormalization-group flow rendered as geometry. Deep in the interior a black-brane horizon marks the far infrared — a thermal state of the boundary theory, its Hawking temperature equal to the field-theory temperature and its horizon area equal to the thermal entropy. The first figure lays out this dictionary in one frame: where the CFT lives, what the radial direction means, and how a bulk field reaches down to source a boundary operator.

Figure

The holographic dictionary — a gravitating AdS bulk in d+1 dimensions is dual to a conformal field theory on its d-dimensional boundary; the radial coordinate z is the renormalization-group scale, with z→0 the ultraviolet boundary and a black-brane horizon deep inside encoding a thermal CFT state at temperature T

Holography sounds paradoxical: a volume seems to hold far more information than its surface. AdS defuses this with geometry — the warp factor $L^2/z^2$ blows up so fast near the boundary that a slab of bulk, in Planck units, scales like the boundary area, not its own volume. The maximum entropy in any region is bounded by that area, exactly the law black hole entropy obeys, so a surface theory has just enough capacity. What makes the duality both powerful and treacherous is that it is strong/weak: when the field theory is an intractable strongly coupled plasma, the gravity side is a tame classical geometry. You can only ever solve the easy side — which is why the correspondence hands you genuinely new physics, not a redundant cross-check. The second figure shows the bridge a practitioner reaches for first: the mass-dimension relation turning a bulk field's mass into a boundary operator's scaling dimension.

Figure

The mass-dimension relation Δ_± = d/2 ± √(d²/4 + m²L²) for AdS₅ with boundary d = 4: a massless bulk field m² = 0 gives a marginal operator Δ = 4, the value m²L² = −3 gives Δ = 3 or 1, and the Breitenlohner-Freedman bound m²L² = −4 gives the degenerate Δ = 2, below which AdS becomes unstable

The core equations

Anti-de Sitter space in Poincaré coordinates is the arena. It is the maximally symmetric solution of the vacuum Einstein equations with a negative cosmological constant, and the single conformal factor $L^2/z^2$ is what encodes both the RG scale and the holographic bound.

\mathrm{d}s^2 = \frac{L^2}{z^2}\left(\eta_{\mu\nu}\,\mathrm{d}x^\mu \mathrm{d}x^\nu + \mathrm{d}z^2\right) \qquad (1)

where: $L$ — AdS curvature radius; $z$ — radial holographic coordinate, with $z\to0$ the boundary and $z\to\infty$ the Poincaré horizon; $x^\mu$ — the $d$ boundary directions $\mu=0,\dots,d-1$ ; $\eta_{\mu\nu}$ — flat Minkowski metric of the boundary. Reading it: each fixed- $z$ slice is Minkowski space rescaled by $L/z$ , so approaching the boundary magnifies short distances — the ultraviolet. The negative curvature is fixed entirely by $L$ : AdS has constant sectional curvature $-1/L^2$ , and threading the metric through Einstein's equation pins the cosmological constant to a single value.

R_{\mu\nu\rho\sigma} = -\frac{1}{L^2}\left(g_{\mu\rho}g_{\nu\sigma}-g_{\mu\sigma}g_{\nu\rho}\right),\qquad \Lambda = -\frac{d(d-1)}{2L^2} \qquad (2)

where: $R_{\mu\nu\rho\sigma}$ — Riemann tensor of $\mathrm{AdS}_{d+1}$ ; $\Lambda$ — cosmological constant; $d$ — boundary dimension, so the bulk has $d+1$ . Reading it: the isometry group is $SO(d,2)$ — exactly the conformal group of $d$ -dimensional Minkowski space, the first structural hint that the boundary theory is a CFT. The prototype fixes how the two sides' parameters relate: matching the near-horizon geometry of $N$ D3-branes to their worldvolume theory ties the AdS radius, string scale, and gauge coupling together.

\frac{L^4}{\alpha'^2} = 4\pi g_s N = g_{\mathrm{YM}}^2 N \equiv \lambda \qquad (3)

where: $L$ — common radius of $\mathrm{AdS}_5$ and $S^5$ ; $\alpha'=\ell_s^2$ — string length squared; $g_s$ — closed-string coupling; $N$ — number of branes and rank of $SU(N)$ ; $g_{\mathrm{YM}}$ — Yang-Mills coupling, $g_{\mathrm{YM}}^2=4\pi g_s$ ; $\lambda$ — 't Hooft coupling. Reading it: classical supergravity needs both a large radius in string units, $L/\ell_s=\lambda^{1/4}\gg1$ , and weak string coupling, $g_s\sim\lambda/N\ll1$ — that is $\lambda\to\infty$ and $N\to\infty$ , precisely the strongly coupled planar gauge theory. The operational core is a single equation between generating functionals. The bulk partition function, computed with a prescribed boundary value for each field, equals the field-theory partition function with that boundary value inserted as a source.

Z_{\mathrm{string}}\big[\phi(z,x)\big|_{z\to0}=\phi_0(x)\big] = \Big\langle \exp\!\int_{\partial}\!\mathrm{d}^d x\,\phi_0(x)\,\mathcal{O}(x)\Big\rangle_{\mathrm{CFT}} \qquad (4)

where: $Z_{\mathrm{string}}$ — bulk string/gravity partition function; $\phi(z,x)$ — a bulk field; $\phi_0(x)$ — its boundary value, acting as a source; $\mathcal{O}(x)$ — the dual gauge-invariant boundary operator. Reading it: in the supergravity limit a saddle dominates the bulk path integral, so the renormalized on-shell action becomes the generating functional of connected CFT correlators — differentiate in $\phi_0$ to extract them. Which operator a bulk field is dual to is fixed by its mass: solving the free wave equation near the boundary produces two power-law falloffs whose exponents are the roots of a quadratic.

\Delta_\pm = \frac{d}{2} \pm \sqrt{\frac{d^2}{4} + m^2 L^2},\qquad m^2 L^2 \ge -\frac{d^2}{4} \qquad (5)

where: $\Delta$ — conformal scaling dimension of $\mathcal{O}$ ; $m$ — mass of the bulk scalar; $d$ — boundary dimension; the inequality is the Breitenlohner-Freedman bound. Reading it: negative $m^2$ is allowed in AdS as long as it stays above the BF bound, because the negative curvature stabilizes modes that would be tachyonic in flat space; below the bound the geometry is unstable. Extracting a physical observable — the boundary stress tensor — requires the full renormalized gravitational action, since the naive on-shell action diverges near the boundary.

S_{\mathrm{ren}} = \frac{1}{16\pi G_{d+1}}\int \mathrm{d}^{d+1}x\,\sqrt{-g}\,(R - 2\Lambda) + S_{\mathrm{GH}} + S_{\mathrm{ct}},\qquad \langle T_{\mu\nu}\rangle = \frac{2}{\sqrt{-\gamma}}\frac{\delta S_{\mathrm{ren}}}{\delta \gamma^{\mu\nu}} \qquad (6)

where: $G_{d+1}$ — bulk Newton constant; $S_{\mathrm{GH}}$ — Gibbons-Hawking term for a well-posed variational principle; $S_{\mathrm{ct}}$ — local counterterms removing volume divergences (holographic renormalization); $\gamma_{\mu\nu}$ — induced boundary metric; $T_{\mu\nu}$ — CFT stress tensor. Reading it: the geometry near the boundary literally encodes the field theory's energy and momentum, read off as the response of the renormalized action to deforming the boundary metric. Two further entries turn geometry into information and into thermodynamics. The Ryu-Takayanagi formula equates a boundary region's entanglement entropy with a bulk minimal-surface area, and a planar AdS black hole reproduces the temperature and entropy of a thermal state.

S_A = \frac{\mathrm{Area}(\gamma_A)}{4 G_{d+1}},\qquad T = \frac{d\,r_h}{4\pi L^2},\qquad S_{\mathrm{BH}} = \frac{A_h}{4 G_{d+1}} \qquad (7)

where: $S_A$ — entanglement entropy of boundary region $A$ ; $\gamma_A$ — minimal bulk surface anchored on $\partial A$ and homologous to $A$ ; $r_h$ — horizon radius of the planar AdS-Schwarzschild black brane; $A_h$ — horizon area; $T,S_{\mathrm{BH}}$ — Hawking temperature and Bekenstein-Hawking entropy, equal to the CFT temperature and thermal entropy. Reading it: both entanglement and thermal entropy are areas in the bulk, the same $1/4G$ law — geometry made of information. The single most-quoted holographic prediction is the shear viscosity of a strongly coupled plasma, universally small in units of its entropy density.

\frac{\eta}{s} = \frac{1}{4\pi}\,\frac{\hbar}{k_B} \qquad (8)

where: $\eta$ — shear viscosity of the boundary plasma; $s$ — entropy density; $\hbar,k_B$ — reduced Planck and Boltzmann constants. Reading it: this is the Kovtun-Son-Starinets value, conjectured as a lower bound for all fluids; in natural units it is just $1/4\pi\approx0.0796$ , and the quark-gluon plasma at RHIC and the LHC sits astonishingly close to it.

Derivation AThe decoupling argument that turns D3-branes into a duality

The goal is to see why a stack of D-branes forces gauge theory and gravity to be the same thing. Maldacena's trick is to describe one system — $N$ coincident D3-branes in flat ten-dimensional type IIB string theory — in two low-energy limits that must agree: in one the object is a gauge theory, in the other a curved geometry. Both contain the same decoupled sector, so that sector's two names must be equal, and the equality is the correspondence. We take the low-energy limit twice, once from open strings and once from closed.

Step 1. In the open-string picture, the light excitations are open strings whose endpoints live on the branes. At energies far below the string scale only the massless modes survive, and their interactions are governed by four-dimensional $\mathcal{N}=4$ super Yang-Mills with gauge group $SU(N)$ ; the closed strings in the surrounding bulk decouple as gravity becomes free at low energy. Why: a stack of $N$ branes carries a $U(N)$ gauge field from strings stretched between them, and the low-energy truncation is exactly the maximally supersymmetric gauge theory.

Step 2. In the closed-string picture the same branes are heavy objects that curve spacetime, sourcing the extremal black 3-brane metric $\mathrm{d}s^2=H^{-1/2}\eta_{\mu\nu}\mathrm{d}x^\mu\mathrm{d}x^\nu+H^{1/2}(\mathrm{d}r^2+r^2\mathrm{d}\Omega_5^2)$ with harmonic function $H(r)=1+L^4/r^4$ . Why: the branes carry Ramond-Ramond flux and tension, and general relativity answers that stress-energy with a warped geometry whose warp factor is the same $L$ that will become the AdS radius.

Step 3. Take the near-horizon limit $r\ll L$ , where $H\to L^4/r^4$ , and substitute $z=L^2/r$ . The metric collapses to $\mathrm{d}s^2=(L^2/z^2)(\eta_{\mu\nu}\mathrm{d}x^\mu\mathrm{d}x^\nu+\mathrm{d}z^2)+L^2\mathrm{d}\Omega_5^2$ — exactly $\mathrm{AdS}_5\times S^5$ with both radii equal to $L$ . Why: close to the branes the constant in $H$ is negligible, and the leftover geometry is the maximally symmetric AdS throat times a five-sphere, the arena of equation (1).

Step 4. In each description the low-energy physics splits into a free bulk-graviton sector plus a strongly interacting "throat" sector, and the free pieces match trivially. Equating what remains gives $\mathcal{N}=4$ $SU(N)$ SYM $\Longleftrightarrow$ type IIB string theory on $\mathrm{AdS}_5\times S^5$ , with the radius set by $L^4/\alpha'^2=g_{\mathrm{YM}}^2N$ . Why: two decompositions of one system must agree term by term, so the two names of the interacting sector are two descriptions of the same theory — the duality.

Go deeper · Derivation B · 1 minOperator dimension from the falloff of a bulk field

Derivation BOperator dimension from the falloff of a bulk field

The goal is to derive the mass-dimension relation of equation (5), the quantitative backbone of the dictionary: it tells you which boundary operator a given bulk field is dual to. We solve the free scalar wave equation in the Poincaré metric near the boundary, look for power-law behavior in the radial coordinate, and read the two exponents — one the source, one the response, their difference fixed by the mass. The result also exposes why AdS tolerates negative masses that flat space forbids.

Step 1. Write the free scalar equation $(\Box-m^2)\phi=0$ with $\Box=\frac{1}{\sqrt{-g}}\partial_M(\sqrt{-g}\,g^{MN}\partial_N)$ in the metric (1), and try a boundary power law $\phi\sim z^{\Delta}$ for a mode with no boundary momentum. Why: near $z\to0$ the radial derivatives dominate, so the leading behavior is a pure power of $z$ , and the exponent is what we want to determine.

Step 2. The $\sqrt{-g}=(L/z)^{d+1}$ and inverse metric $g^{zz}=z^2/L^2$ turn the wave equation into a radial ODE whose indicial equation is $\Delta(\Delta-d)-m^2L^2=0$ . Its two roots are $\Delta_\pm=\tfrac{d}{2}\pm\sqrt{\tfrac{d^2}{4}+m^2L^2}$ . Why: substituting the power law converts the differential equation into an algebraic condition on $\Delta$ , a simple quadratic whose roots are the two allowed falloffs.

Step 3. Near the boundary the general solution is $\phi(z,x)\sim\phi_0(x)\,z^{d-\Delta}+\langle\mathcal{O}(x)\rangle\,z^{\Delta}$ with $\Delta\equiv\Delta_+$ . The slower-decaying, non-normalizable mode $z^{d-\Delta}$ is the source $\phi_0$ that appears in equation (4); the normalizable mode $z^{\Delta}$ carries the operator expectation value $\langle\mathcal{O}\rangle$ . Why: identifying source with the leading coefficient and response with the subleading one is exactly the statement that the boundary value sources $\mathcal{O}$ , and it fixes the operator's dimension to be $\Delta_+$ .

Step 4. Reality of $\Delta$ requires $m^2L^2\ge-d^2/4$ , the Breitenlohner-Freedman bound; in the window $-d^2/4\le m^2L^2\le-d^2/4+1$ both roots are above the unitarity bound $(d-2)/2$ and either quantization is allowed. Why: below the BF bound the exponents go complex, the falloff oscillates without decaying, and the mode signals an instability — negative mass squared is safe in AdS only down to this curvature-set floor. *The numeric anchors are: $m^2L^2=0\Rightarrow\Delta=4$ , $m^2L^2=-3\Rightarrow\Delta=3$ or $1$ , $m^2L^2=-4\Rightarrow\Delta=2$ , consistent with the figure above.*

Go deeper · Derivation C · 2 minThe Hawking-Page transition as deconfinement

Derivation CThe Hawking-Page transition as deconfinement

The goal is to show that a purely gravitational phase transition in the bulk is dual to the confinement-deconfinement transition of the boundary gauge theory — a vivid demonstration that bulk geometry stores boundary thermodynamics. We put the CFT on a spatial sphere at temperature $T=1/\beta$ , so the bulk is global AdS with a thermal circle, and compare the two Euclidean saddles that fill in that boundary; their free energies cross at a critical temperature, and the order- $N$ scaling on either side is the confinement diagnostic.

Step 1. For a CFT on $S^{d-1}\times S^1_\beta$ the Euclidean bulk path integral has two smooth saddles that share the boundary: thermal AdS, which is just global AdS with the time circle identified and no horizon, and the large AdS-Schwarzschild black hole, which caps off the geometry with a horizon. Why: both geometries are regular and asymptote to the same boundary, so both contribute, and the path integral is dominated by whichever has the smaller renormalized action.

Step 2. Computing the two renormalized on-shell actions and forming $\Delta F=F_{\mathrm{BH}}-F_{\mathrm{AdS}}$ , the difference changes sign at a critical horizon radius, giving the Hawking-Page temperature $T_{\mathrm{HP}}=(d-1)/(2\pi L)$ . Why: at low temperature thermal AdS has lower free energy and dominates; above $T_{\mathrm{HP}}$ the black hole wins, a genuine first-order transition with latent heat.

Step 3. The two phases have sharply different free-energy scalings in $N$ . Below $T_{\mathrm{HP}}$ the free energy is order $1$ in $N$ — color-singlet, confined; above it the black hole gives order $N^2$ , the count of liberated gluon degrees of freedom in a deconfined plasma. Why: the black hole horizon area scales as $N^2$ through $1/G\sim N^2$ , so its entropy and free energy carry the full adjoint degree-of-freedom count, exactly what deconfinement releases.

Step 4. Witten identified this bulk transition with the confinement-deconfinement transition of the dual gauge theory. Why: the order parameter — free energy jumping from $O(1)$ to $O(N^2)$ — is precisely the field-theory signature of deconfinement, so a black hole forming in the bulk is the boundary plasma melting. *The critical temperature is $T_{\mathrm{HP}}=(d-1)/2\pi L$ , and the jump in free energy from $O(1)$ to $O(N^2)$ across the transition is the standard diagnostic of deconfinement.*

Limiting cases

The validity and the very form of the correspondence turn on two knobs — the rank $N$ and the 't Hooft coupling $\lambda=g_{\mathrm{YM}}^2N$ — with a fourth entry showing what happens as the curvature is switched off.

Case	Result	Meaning
Classical supergravity	$N\to\infty,\ \lambda\to\infty$	String loops ( $\sim1/N^2$ ) and $\alpha'$ corrections ( $\sim\lambda^{-1/2}$ ) both vanish; the bulk is two-derivative Einstein gravity plus matter, dual to a strongly coupled planar CFT. Nearly every holographic computation lives here.
Classical string theory	$N\to\infty,\ \lambda$ finite	Loops suppressed but curvature reaches the string scale, so all $\alpha'$ corrections matter; dual to planar but finitely coupled $\mathcal{N}=4$ SYM, tractable via integrability of the planar spin chain.
Full quantum gravity	$N$ finite, $\lambda$ finite	Both $1/N$ and $\alpha'$ corrections present — the exact conjectured duality, with no controlled bulk calculation, but the CFT serves as a nonperturbative definition of the bulk quantum gravity.
Flat-space limit	$L\to\infty$	Sending the AdS radius to infinity flattens the bulk; recovering a flat-space S-matrix from AdS correlators is subtle and an active research frontier.

Common trapThe correspondence lets you compute weak-coupling gauge theory from classical gravity

It does the opposite. AdS/CFT is a strong/weak duality: classical supergravity is trustworthy only when the field theory is strongly coupled and large- $N$ ( $\lambda\gg1$ , $N\gg1$ ), because that is exactly where $\alpha'$ and string-loop corrections are suppressed. You cannot feed weakly coupled gauge-theory questions into classical gravity, nor trust bulk string-loop corrections away from their regime — you can only ever solve the easy side of the duality, which is why it delivers new physics rather than a cross-check you could have done by hand. A second trap hides in the name: "AdS/CFT" is not one statement but a whole family of dualities, and the precise dictionary — which operators, which multiplets, which boundary conditions — depends on the specific bulk and boundary theories in play.

Worked examples

1 · Reading operator dimensions off bulk masses (fully worked)

In $\mathrm{AdS}_5$ (so $d=4$ ), find the dimension of the boundary operator dual to a bulk scalar with (a) $m^2=0$ , (b) $m^2L^2=-3$ , and (c) $m^2L^2=-4$ . State which case saturates the Breitenlohner-Freedman bound.

Step 1. The tool is equation (5) with $d=4$ , so $\Delta_\pm=2\pm\sqrt{4+m^2L^2}$ . The BF bound is $m^2L^2\ge-d^2/4=-4$ , and the scalar unitarity bound is $\Delta\ge(d-2)/2=1$ .

Step 2. Case (a), $m^2=0$ : $\Delta_+=2+\sqrt{4}=\mathbf{4}$ . A massless bulk scalar is dual to a marginal operator of dimension $4$ — consistent with the dilaton pairing with the dimension-4 Lagrangian operator $\mathrm{Tr}\,F^2$ in $\mathcal{N}=4$ SYM.

Step 3. Case (b), $m^2L^2=-3$ : $\Delta_\pm=2\pm\sqrt{4-3}=2\pm1$ , giving $\Delta_+=\mathbf{3}$ and $\Delta_-=\mathbf{1}$ . Both roots sit at or above the unitarity bound, so either quantization is legitimate — the same bulk field can be dual to a dimension-3 or a dimension-1 operator depending on boundary conditions.

Step 4. Case (c), $m^2L^2=-4$ : $\Delta_\pm=2\pm\sqrt{4-4}=\mathbf{2}$ , a degenerate double root. This saturates the BF bound; any more negative $m^2$ would make the square root imaginary and destabilize AdS. So the three answers are Δ = 4, Δ = 3 or 1, and Δ = 2 (BF-saturating).

2 · The viscosity bound in the laboratory

Evaluate the KSS ratio $\eta/s$ in SI units and compare it with the quark-gluon plasma produced in heavy-ion collisions.

Step 1. Equation (8) gives $\eta/s=\hbar/(4\pi k_B)$ . Insert $\hbar=1.055\times10^{-34}\ \mathrm{J\,s}$ and $k_B=1.381\times10^{-23}\ \mathrm{J/K}$ : $\eta/s=(1.055\times10^{-34})/(4\pi\times1.381\times10^{-23})=\mathbf{6.08\times10^{-13}\ K\,s}$ .

Step 2. In natural units ( $\hbar=k_B=1$ ) this is simply $1/4\pi=\mathbf{0.0796}$ , a pure number and — the conjecture holds — a lower bound for any fluid. Ordinary fluids sit far above it: water near room temperature has $\eta/s\approx30$ in these units — about $380$ times the bound — and helium near its minimum about $9$ times the bound.

Step 3. The quark-gluon plasma inferred from elliptic-flow data at RHIC and the LHC has $\eta/s\approx0.08$ – $0.2$ , within a factor of two or three of the holographic value — the strongest empirical hint that a real strongly coupled plasma behaves like the dual of a black brane. The strongly coupled QGP is the closest thing in nature to a "perfect fluid," and holography predicts why.

Physics of the portrait

The correspondence has one clear author, but it crowned a decade of holographic thinking — from 't Hooft's large- $N$ counting to the holographic principle of 't Hooft and Susskind — and it was made operational within months by two other groups.

Origins · 1 minJuan Maldacena and how this was discovered

Juan Maldacena

b. 1968 · Argentine theoretical physicist, AdS/CFT

Juan Martín Maldacena was born in Buenos Aires in 1968 and completed his PhD at Princeton in 1996 on black holes in string theory. In late 1997, a young faculty member at Harvard, he wrote "The Large N Limit of Superconformal Field Theories and Supergravity," proposing that type IIB string theory on $\mathrm{AdS}_5\times S^5$ is the same theory as four-dimensional $\mathcal{N}=4$ super Yang-Mills — a claim so productive it became the most-cited paper in high-energy theory. The idea had roots: 't Hooft had shown in 1974 that gauge theories reorganize into a genus expansion in $1/N$ mirroring the string loop expansion, and 't Hooft (1993) and Susskind (1994) argued on black-hole grounds that a gravitating region's information must live on its boundary — the holographic principle. Maldacena's step was to make holography concrete: a specific bulk, a specific boundary, a dictionary. Within months Gubser, Klebanov and Polyakov, and independently Witten, supplied the field-operator dictionary of equation (4) that turned the conjecture into a calculational machine. Now at the Institute for Advanced Study, Maldacena went on to help reframe the black-hole information problem, the emergence of spacetime from entanglement, and the "ER = EPR" link between wormholes and quantum correlations — all extensions of one idea: geometry and quantum matter are two languages for one physics.

History

1993

— 't Hooft proposes the holographic principle from black-hole entropy

1994

— Susskind formulates holography within string theory

1997

— Maldacena conjectures the AdS/CFT correspondence

1998

— Gubser, Klebanov, Polyakov and Witten supply the field-operator dictionary

2006

— Ryu and Takayanagi relate entanglement entropy to bulk minimal surfaces

Going further

Texts and reviews. Maldacena, "The Large N Limit of Superconformal Field Theories and Supergravity" (1998) · Gubser, Klebanov, Polyakov, "Gauge Theory Correlators from Non-Critical String Theory" (1998) · Witten, "Anti De Sitter Space and Holography" (1998) · Aharony, Gubser, Maldacena, Ooguri, Oz, "Large N Field Theories, String Theory and Gravity" (2000) · Ammon & Erdmenger, Gauge/Gravity Duality (2015) · Nastase, Introduction to the AdS/CFT Correspondence (2015) · Wald, General Relativity (1984), Ch. 11 (asymptotic structure) · Carroll, Spacetime and Geometry (2004), Appendix on maximally symmetric spaces. All equations are standard results of the AdS/CFT literature. Key numerical checks: the Poincaré metric (1) solves Einstein's equation with $\Lambda=-d(d-1)/2L^2$ and has constant curvature $-1/L^2$ ; the indicial equation $\Delta(\Delta-d)=m^2L^2$ yields the anchors $m^2L^2=0\Rightarrow\Delta=4$ , $-3\Rightarrow\Delta=3$ or $1$ , $-4\Rightarrow\Delta=2$ ; the parameter relation $L^4/\alpha'^2=g_{\mathrm{YM}}^2N=4\pi g_sN$ follows from the D3-brane near-horizon limit; $\eta/s=\hbar/4\pi k_B=1/4\pi$ in natural units $=6.08\times10^{-13}\ \mathrm{K\,s}$ in SI; and the Hawking-Page temperature is $T_{\mathrm{HP}}=(d-1)/2\pi L$ .

Builds on: ADM 3+1 Decomposition of Spacetime · ADM Mass, Momentum and Angular Momentum · Conformal Method for Constraint Solving. Leads to: Holographic Entanglement Entropy · Black Hole Thermodynamics · Holographic Hydrodynamics.

Ask the tutor: Ask me to walk through why $N$ D3-branes force $\mathcal{N}=4$ Yang-Mills to equal string theory on $\mathrm{AdS}_5\times S^5$ , and how far the shear-viscosity bound really constrains the quark-gluon plasma.

Is the AdS/CFT correspondence actually proven?

No — it remains a conjecture, with no mathematical proof that the two theories are identical. What it has is an enormous body of quantitative evidence: matched symmetries and spectra, protected quantities that agree at all couplings, integrability results interpolating between weak and strong coupling, and thermodynamic and correlator checks that line up precisely. Because it is a strong/weak duality, most tests compare a hard strong-coupling field-theory result against an easy classical-gravity one, and they keep agreeing. The consensus is that it is true; the missing piece is a first-principles derivation at finite $N$ and finite coupling.

Does our universe obey AdS/CFT? It is not anti-de Sitter.

Correct — our universe is close to flat and, on the largest scales, slightly positively curved (de Sitter-like), whereas the correspondence is built on negative curvature. AdS is special because its boundary is a well-defined place where a field theory can live and boundary conditions make the problem well-posed; de Sitter and flat space lack such a clean boundary, which is why holography there is far less understood. So AdS/CFT is not directly a theory of our cosmos. Its value is as a controlled laboratory where quantum gravity is defined nonperturbatively, letting us learn general lessons — about black hole entropy, information, and the emergence of spacetime — we hope survive the passage to realistic geometries.

What does "strong/weak duality" mean in practice?

It means the two sides are easy in opposite regimes. Classical supergravity is reliable when the 't Hooft coupling $\lambda$ and the rank $N$ are both large, since then string-loop and $\alpha'$ corrections are negligible — but large $\lambda$ is exactly where the dual gauge theory is strongly coupled and perturbation theory fails. So when one description is a simple classical calculation, the other is intractable, never the reverse. The upshot is a one-way street: you solve the tractable side to learn something new about the intractable one, which is why holography reaches strong-coupling plasma transport no perturbative method can.

How can a bulk field's mass fix a boundary operator's dimension?

Because a bulk field's falloff rate near the boundary is set by its mass, and that rate is what the boundary reads as a scaling dimension. Solving the wave equation near $z=0$ gives two power laws, $z^{d-\Delta}$ and $z^{\Delta}$ , with $\Delta(\Delta-d)=m^2L^2$ : the slower falloff is the source, the faster is the response, and $\Delta$ is exactly the dimension a conformal operator would have. Heavier fields decay faster and map to higher-dimension operators; a massless field maps to a marginal operator of dimension $d$ . Negative $m^2$ is allowed down to the Breitenlohner-Freedman bound because AdS's curvature stabilizes modes flat space would call tachyonic.

Why is a black hole in the bulk the same as heat in the field theory?

A planar AdS black hole (a black brane) has a Hawking temperature set by its horizon and an entropy set by its horizon area, and the dictionary identifies both with the temperature and thermal entropy of the boundary theory. Put the CFT at finite temperature and the dominant bulk geometry grows a horizon whose $T$ and $S=A/4G$ are literally the plasma's temperature and entropy. This is why the Hawking-Page transition — a black hole forming as you heat global AdS — is dual to deconfinement, and why holography turns black-hole thermodynamics into ordinary statistical mechanics of a unitary quantum system, sharpening the information paradox into a problem with a definite answer.

What is the Ryu-Takayanagi formula, and why is it exciting?

It says the entanglement entropy of a boundary region $A$ equals the area of the smallest bulk surface anchored on $A$ 's edge, over $4G$ — the same $1/4G$ that governs black-hole entropy. The thrill is conceptual: entanglement, an abstract quantum-information quantity, becomes a geometric area, so the connectivity of spacetime looks woven from boundary entanglement. Cut the entanglement between two halves and the bulk pinches apart — the origin of "entanglement builds geometry" and of the bulk-as-error-correcting-code picture. It is also the engine behind the island-formula resolution of the information paradox and the recovery of the Page curve.