QFT: Modern Frontiers

Quantum Field Theory » Modern Frontiers

Standard quantum field theory — Lagrangians, Feynman diagrams, renormalization — is enormously successful but conceals deep structure. Over the last few decades a set of reformulations and discoveries has revealed that scattering amplitudes are far simpler than the diagrams suggest, that strongly coupled gauge theories can be dual to gravity in one higher dimension, that some classical symmetries are unavoidably destroyed by quantization, and that field theory itself carries the seeds of quantum gravity. This page is a companion to the main QFT page; it assumes that material (gauge theory, the path integral, renormalization, spontaneous symmetry breaking) and develops the modern frontier in full.

The unifying theme is that physics is simpler than its standard formulation. Feynman diagrams are a redundant, gauge-dependent bookkeeping of an object — the on-shell amplitude — that is far more compact. Spacetime locality and unitarity may be emergent rather than fundamental. And the boundary of a region of spacetime can encode everything inside it. These are not separate curiosities: the amplitudes program, holography, and anomaly inflow keep turning out to be different views of the same mathematics.

Amplitudes over diagrams. On-shell recursion and unitarity reconstruct amplitudes without ever writing a Lagrangian or a loop integral by brute force.
Holography. A gravitational theory in d+1 dimensions is equivalent to a field theory living on its d-dimensional boundary.
Anomalies. Symmetries that survive classically but are broken by the quantum measure — fatal when gauged, predictive when global.
Toward quantum gravity. Effective field theory, asymptotic safety, and holography frame how (and whether) gravity becomes a consistent QFT.

What You’ll Find on This Page

Section	What it covers
The Modern Amplitudes Program	Spinor-helicity, BCFW recursion, generalized unitarity, the amplituhedron
AdS/CFT and Holography	The duality dictionary, the holographic principle, applications
Anomalies	Chiral and gauge anomalies, anomaly cancellation, ‘t Hooft matching
Connections to Quantum Gravity	Gravity as an EFT, asymptotic safety, holographic entropy, the double copy

The Modern Amplitudes Program

The textbook recipe for a scattering amplitude — draw all Feynman diagrams, assign propagators and vertices, integrate — is correct but spectacularly inefficient. The number of diagrams grows factorially, individual diagrams are gauge-dependent and littered with unphysical longitudinal polarizations, and yet the final gauge-invariant answer is frequently a single short expression. The amplitudes program asks: can we compute the physical answer directly, exploiting only the properties an amplitude must have — Lorentz invariance, locality (poles only where intermediate particles go on shell), and unitarity?

The Parke-Taylor surprise

The cleanest illustration is gluon scattering. The tree-level amplitude for two gluons of negative helicity and any number of positive-helicity gluons (a "maximally helicity-violating", or MHV, amplitude) requires summing hundreds of Feynman diagrams for even a handful of gluons. Parke and Taylor (1986) conjectured — and it was later proven — that the entire answer collapses to one term. For n gluons with negative-helicity legs i and j, the color-ordered amplitude is just a ratio of spinor brackets. Hundreds of diagrams, one line. That gap between the complexity of the method and the simplicity of the answer is what the whole program exists to explain.

Spinor-Helicity Formalism

For massless particles the natural variables are not four-momenta but the two-component spinors into which a null momentum factorizes. A null momentum $p_\mu$ (with $p^2 = 0$) can be written as an outer product of a left-handed and a right-handed Weyl spinor:

\[p_{\alpha\dot\alpha} = p_\mu \sigma^\mu_{\alpha\dot\alpha} = \lambda_\alpha \tilde{\lambda}_{\dot\alpha}\]

The spinors carry the helicity information directly. One defines the antisymmetric Lorentz-invariant brackets

\[\langle i\, j \rangle = \epsilon_{\alpha\beta}\, \lambda_i^\alpha \lambda_j^\beta, \qquad [i\, j] = \epsilon_{\dot\alpha\dot\beta}\, \tilde{\lambda}_i^{\dot\alpha} \tilde{\lambda}_j^{\dot\beta}\]

so that the Mandelstam invariant is $s_{ij} = (p_i + p_j)^2 = \langle i\, j\rangle [j\, i]$. In these variables the Parke-Taylor MHV amplitude for $n$ gluons with negative-helicity legs $i$ and $j$ is

\[A_n^{\text{MHV}} = \frac{\langle i\, j\rangle^4}{\langle 1\, 2\rangle \langle 2\, 3\rangle \cdots \langle n\, 1\rangle}\]

Every Lorentz dot product, every polarization sum, every gauge artifact has vanished. The denominator is simply the cyclic product of the adjacent-leg brackets, and the numerator is fixed by little-group weights (helicity). This compactness is the empirical hint that a Lagrangian is the wrong starting point.

BCFW On-Shell Recursion

The Britto-Cachazo-Feng-Witten (BCFW) recursion relations build any tree amplitude from lower-point on-shell amplitudes — never an off-shell Feynman vertex. The trick is to deform two of the external momenta by a complex parameter $z$ while keeping them on shell and conserving total momentum:

\[\hat{\lambda}_i = \lambda_i, \quad \hat{\tilde\lambda}_i = \tilde\lambda_i - z\,\tilde\lambda_j, \qquad \hat{\lambda}_j = \lambda_j + z\,\lambda_i, \quad \hat{\tilde\lambda}_j = \tilde\lambda_j\]

The deformed amplitude $A_n(z)$ is a rational function of $z$ whose only singularities are simple poles where an internal propagator goes on shell. If $A_n(z) \to 0$ as $z \to \infty$ (true for gluons and gravitons with suitable shifts), Cauchy’s theorem gives the physical amplitude $A_n = A_n(0)$ as a sum over the residues:

\[A_n = \sum_{\text{factorizations } I} A_L(z_I)\, \frac{1}{P_I^2}\, A_R(z_I)\]

Each term is a product of two strictly on-shell, lower-point amplitudes glued by a single propagator $1/P_I^2$, evaluated at the $z_I$ that puts $P_I$ on shell. Locality (the pole structure) and unitarity (the factorization onto physical sub-amplitudes) are the only inputs. The Lagrangian, gauge fixing, and ghost fields never appear.

Why this works — and what it teaches

BCFW says a tree amplitude is fixed by its singularities: where it blows up (on-shell intermediate particles) and how it factorizes there. Lagrangian locality is downgraded from an input to an output. This is the technical seed of a radical idea pursued in the amplitudes community — that spacetime locality and unitarity might be emergent, derived consequences of a more primitive structure rather than axioms imposed from the start.

Generalized Unitarity at Loop Level

Tree-level recursion is only half the story; loops carry the quantum corrections. The optical theorem already tells us that the imaginary (absorptive) part of a loop amplitude is a phase-space integral of products of lower amplitudes — unitarity in its original form. Generalized unitarity sharpens this by cutting several internal lines at once. Putting $k$ propagators on shell ($1/(\ell^2 - m^2) \to -2\pi i\,\delta^+(\ell^2 - m^2)$) isolates a particular coefficient in the decomposition of the loop amplitude onto a basis of scalar master integrals:

\[A_n^{\text{1-loop}} = \sum_i c_i\, I_i^{\text{box}} + \sum_j d_j\, I_j^{\text{triangle}} + \sum_k e_k\, I_k^{\text{bubble}} + (\text{rational})\]

In four dimensions a maximal (quadruple) cut freezes the loop momentum completely, reducing the box coefficient $c_i$ to a pure product of four on-shell tree amplitudes — an algebraic operation, no integration required. The integrals $I_i$ themselves are known once and for all. So a one-loop amplitude is reconstructed from trees plus a fixed integral basis, again bypassing the diagrammatic expansion. This is the engine (in tools like the unitarity method and its automation) behind modern next-to-leading-order predictions for LHC processes.

Twistors and the Amplituhedron

Pushing further, planar amplitudes in maximally supersymmetric $\mathcal{N}=4$ Yang-Mills theory exhibit a hidden geometry. Witten’s twistor-string reformulation mapped amplitudes to curves in twistor space; that line led to the amplituhedron, a positive geometric region in a Grassmannian whose canonical volume form is the amplitude. In this picture there is no Lagrangian, no unitarity sum, and crucially no manifest locality: the amplitude is the volume of a geometric object, and the physical properties emerge as features of that geometry’s boundaries.

Approach	Core idea	What it eliminates
Spinor-helicity	Null momenta as spinor bilinears	Polarization vectors, dot-product clutter
BCFW recursion	Complex momentum shift + Cauchy	Off-shell vertices, gauge artifacts
Generalized unitarity	Multi-line cuts onto an integral basis	Brute-force loop integration
Amplituhedron	Amplitude as a positive geometry’s volume	Locality and unitarity as inputs

The practical payoff is precision collider physics; the conceptual payoff is the growing evidence that the smooth, local spacetime of the Lagrangian is not the most economical description of the physics.

AdS/CFT and Holography

The most influential idea connecting field theory to gravity is the AdS/CFT correspondence (Maldacena, 1997): a conjectured exact equivalence between a quantum gravity theory in an anti-de Sitter (AdS) spacetime and a conformal field theory (CFT) living on its lower-dimensional boundary. It is the sharpest known realization of the holographic principle — the proposal that the information content of a region of space is bounded by, and encoded on, its boundary area rather than its volume.

The holographic principle, from black holes

Ordinary thermodynamic entropy scales with volume — twice the box, twice the entropy. Black holes break this rule. The Bekenstein-Hawking entropy of a black hole scales with the area of its horizon, not the volume it encloses, and a black hole is the most entropic object that can fit in a given region. The conclusion ('t Hooft, Susskind) is that the maximum information in any region is set by its bounding area in Planck units — the world is, in this informational sense, a hologram. AdS/CFT is the concrete model where this is provably true: the "bulk" gravitational degrees of freedom are fully captured by a "boundary" field theory with one fewer dimension.

The Statement of the Duality

The canonical example relates Type IIB string theory on $AdS_5 \times S^5$ to $\mathcal{N}=4$ supersymmetric Yang-Mills theory with gauge group $SU(N)$ living on the four-dimensional boundary. The dictionary connects parameters on the two sides:

\[\frac{L^4}{\ell_s^4} = g_{YM}^2 N = \lambda, \qquad g_s = \frac{g_{YM}^2}{4\pi}\]

Here $L$ is the AdS radius, $\ell_s$ the string length, $g_s$ the string coupling, and $\lambda = g_{YM}^2 N$ the ‘t Hooft coupling. The structure is a strong/weak duality: the gravity description is reliable (weakly curved, classical) precisely when $\lambda$ is large, i.e. when the gauge theory is strongly coupled and perturbation theory fails. This is what makes the correspondence so useful — it turns hard strongly-coupled field-theory questions into tractable classical-gravity calculations, and vice versa.

The GKP-Witten Dictionary

The operational heart of the correspondence is the equality of partition functions. Every bulk field $\phi$ approaches a boundary value $\phi_0$ that acts as a source for a dual boundary operator $\mathcal{O}$, and the generating functionals match:

\[Z_{\text{gravity}}\big[\phi \to \phi_0\big] = \Big\langle \exp\!\Big(\int d^d x\; \phi_0(x)\,\mathcal{O}(x)\Big)\Big\rangle_{\text{CFT}}\]

In the supergravity limit the left side becomes the classical on-shell action, so boundary correlation functions are computed by solving classical bulk equations of motion:

\[Z_{\text{gravity}}\big[\phi \to \phi_0\big] \approx e^{-S_{\text{on-shell}}[\phi_0]}\]

The conformal dimension $\Delta$ of the boundary operator is fixed by the mass $m$ of the dual bulk field through the relation (for a scalar in $AdS_{d+1}$):

\[\Delta(\Delta - d) = m^2 L^2\]

This entry of the dictionary turns the spectrum of bulk excitations into the spectrum of operator dimensions in the CFT.

Boundary (CFT)	Bulk (gravity in AdS)
Conserved current $J^\mu$	Gauge field $A_\mu$
Stress tensor $T^{\mu\nu}$	Metric $g_{\mu\nu}$ (graviton)
Scalar operator of dimension $\Delta$	Scalar field of mass $m^2 L^2 = \Delta(\Delta-d)$
Global symmetry	Gauge symmetry in the bulk
Finite temperature	Black hole / black brane in AdS
Entanglement entropy	Minimal-surface area (Ryu-Takayanagi)

Applications

Because the duality maps strong coupling to weak curvature, it has become a calculational tool well beyond its string-theory origins:

Quark-gluon plasma. The shear-viscosity-to-entropy-density ratio of a strongly coupled plasma is computed from a black brane in AdS, yielding the famous bound $\eta/s = 1/(4\pi)$ in natural units, remarkably close to what the QGP produced at RHIC and the LHC appears to satisfy.
Condensed matter. “AdS/CMT” models strange metals, holographic superconductors, and non-Fermi-liquid behavior using charged black holes, capturing physics that has no quasiparticle description and is therefore inaccessible to standard perturbation theory.
Quantum information. The Ryu-Takayanagi formula identifies boundary entanglement entropy with bulk minimal-surface area (see below), and holographic error-correcting codes have reframed how bulk locality emerges from boundary entanglement.

The correspondence remains a conjecture — there is no proof — but it has passed an enormous number of nontrivial checks and now functions both as a window into quantum gravity and as a practical engine for strongly coupled field theory.

Anomalies

A quantum anomaly is a symmetry of the classical action that fails to survive quantization: the classical Noether current is conserved on the equations of motion, but the conservation law acquires a nonzero right-hand side once the path-integral measure is included. Whether an anomaly is welcome or fatal depends entirely on whether the broken symmetry is global or gauge.

The Chiral (ABJ) Anomaly

The prototype is the Adler-Bell-Jackiw anomaly. Classically, a massless Dirac fermion coupled to electromagnetism has a conserved axial current $j^\mu_5 = \bar\psi \gamma^\mu \gamma^5 \psi$. But the quantum triangle diagram with one axial and two vector currents does not vanish, and the conservation law is corrected to

\[\partial_\mu j^\mu_5 = \frac{e^2}{16\pi^2}\, \epsilon^{\mu\nu\rho\sigma} F_{\mu\nu} F_{\rho\sigma}\]

This is not an approximation that higher loops modify — the Adler-Bardeen theorem states the one-loop coefficient is exact. Fujikawa gave the deepest interpretation: the anomaly is the non-invariance of the fermionic path-integral measure $\mathcal{D}\psi\,\mathcal{D}\bar\psi$ under a chiral rotation. The Jacobian is not unity, and its logarithm is precisely the anomaly density above.

An anomaly that is measured

The chiral anomaly is not a formal subtlety — it makes a sharp prediction. The decay rate of the neutral pion into two photons, $\pi^0 \to \gamma\gamma$, is controlled almost entirely by the anomaly. A naive symmetry argument forbids or strongly suppresses the decay; the anomaly restores it, and the predicted rate (including a factor of the number of quark colors, $N_c = 3$) matches experiment. The agreement is one of the cleanest confirmations that the color quantum number takes three values and that anomalies are physically real.

Gauge Anomalies and Their Cancellation

When the anomalous current is a global symmetry, the anomaly is a feature — it encodes real physics like the pion decay. When it is a gauge symmetry, the anomaly is a catastrophe: gauge invariance is what removes the unphysical polarizations of the gauge bosons, and an anomaly destroys it, wrecking unitarity and renormalizability. A consistent chiral gauge theory therefore requires its gauge anomalies to cancel.

The anomaly of a gauge group is proportional to a group-theoretic factor $A^{abc} = \text{Tr}\big[T^a {T^b, T^c}\big]$ summed over all chiral fermions in the theory. Consistency demands

\[\sum_{\text{chiral fermions}} \text{Tr}\big[T^a \{T^b, T^c\}\big] = 0\]

In the Standard Model this is a remarkable, nontrivial constraint. The hypercharge and mixed gauge-gravitational anomalies cancel only because the quark and lepton charges within a single generation conspire — summed over color, the contributions of the $(u,d)$ quarks and the $(\nu,e)$ leptons add to zero. This is widely read as evidence that quarks and leptons belong together in complete generations (and motivates grand unification, where they sit in a single representation that is automatically anomaly-free).

‘t Hooft Anomaly Matching

Global anomalies obey a powerful constraint discovered by ‘t Hooft. Imagine weakly gauging a global symmetry that has an anomaly. The anomaly coefficient is computed from the fundamental (UV) degrees of freedom — the quarks and gluons, say. But the same coefficient must be reproduced by whatever degrees of freedom describe the theory in the deep infrared, after strong-coupling effects like confinement have set in:

\[\mathcal{A}_{\text{UV}}\big[\text{fundamental fermions}\big] = \mathcal{A}_{\text{IR}}\big[\text{bound states / composites}\big]\]

Because the anomaly is renormalization-group invariant, it acts as a bookkeeping device that survives across all scales. In QCD, anomaly matching constrains the spectrum of massless composite states (e.g. it is consistent with, and partly requires, the existence of the pions as Goldstone bosons of chiral symmetry breaking). More generally, ‘t Hooft matching is one of the very few exact handles on strongly coupled, non-perturbative dynamics — it lets us rule out proposed low-energy descriptions that fail to reproduce the UV anomaly.

Type	Symmetry broken	Consequence
Chiral (ABJ) anomaly	Global axial $U(1)$	Physical — predicts $\pi^0 \to \gamma\gamma$
Gauge anomaly	Local gauge symmetry	Fatal — must cancel for consistency
Mixed gauge-gravitational	Gauge $\times$ diffeomorphisms	Constrains charge assignments
‘t Hooft anomaly	Weakly gauged global symmetry	Matched UV $\leftrightarrow$ IR, constrains spectra

Connections to Quantum Gravity

Quantum field theory and gravity famously resist marriage: naive quantization of the metric produces a non-renormalizable theory. Yet field theory supplies several of the most productive frameworks for thinking about quantum gravity — some that tame the problem within QFT, some that reach beyond it.

Gravity as an Effective Field Theory

The statement “gravity is non-renormalizable” is true but easy to overstate. Treated as an effective field theory, general relativity is perfectly predictive at energies far below the Planck scale. One organizes the action as an expansion in curvature with a tower of higher-derivative terms,

\[S = \int d^4x\, \sqrt{-g}\,\Big(\frac{1}{16\pi G} R + c_1 R^2 + c_2 R_{\mu\nu}R^{\mu\nu} + \cdots\Big)\]

and the non-renormalizable couplings are suppressed by powers of $E/M_{\text{Pl}}$. This is enough to compute genuine quantum-gravitational corrections unambiguously — for example the leading quantum correction to the Newtonian potential between two masses,

\[V(r) = -\frac{G m_1 m_2}{r}\left(1 + \alpha\,\frac{G(m_1 + m_2)}{r c^2} + \beta\,\frac{G\hbar}{r^2 c^3} + \cdots\right)\]

where the last term is the leading quantum correction with a universal, calculable coefficient. The EFT viewpoint clarifies that the problem is not that quantum gravity is meaningless at low energy, but that it requires a UV completion — new physics or a new framework — near $M_{\text{Pl}}$.

Asymptotic Safety

One candidate UV completion stays entirely within quantum field theory. The asymptotic safety scenario (Weinberg) proposes that gravity’s couplings flow to a nontrivial fixed point of the renormalization group at high energy:

\[\beta(g_i) = \mu\,\frac{d g_i}{d\mu} \;\xrightarrow{\;\mu \to \infty\;}\; 0 \quad \text{at a non-Gaussian fixed point}\]

If such an interacting fixed point exists and only a finite number of directions flow into it, the theory is predictive up to arbitrarily high energies despite being perturbatively non-renormalizable — much as a familiar renormalizable theory is controlled by the free (Gaussian) fixed point. Functional renormalization-group calculations provide suggestive evidence for the fixed point, though its existence in the full theory remains unproven.

Holographic Entanglement Entropy

AdS/CFT supplies the most concrete bridge. The Ryu-Takayanagi formula computes the entanglement entropy of a region $A$ of the boundary CFT as the area of a minimal surface $\gamma_A$ in the bulk that is anchored on the boundary of $A$:

\[S_A = \frac{\text{Area}(\gamma_A)}{4 G_N}\]

The structural resemblance to the Bekenstein-Hawking black-hole entropy is not a coincidence — it is the same formula, and it ties the information in the boundary theory directly to geometry in the bulk. This relation underlies the modern slogan that “entanglement builds spacetime”: the connectivity and smoothness of the bulk geometry are encoded in the pattern of entanglement of the boundary degrees of freedom, and severing that entanglement disconnects the bulk. Combined with quantum error correction, it has reshaped how the emergence of a gravitational, gravitating spacetime from a non-gravitational field theory is understood.

The Double Copy

A striking, almost algebraic connection between gauge theory and gravity is the double copy (Bern-Carrasco-Johansson). It builds on color-kinematics duality: gauge-theory amplitudes can be arranged so that the kinematic numerators $n_i$ satisfy the same Jacobi-like identities as the color factors $c_i$. Once in that form, replacing color by a second copy of kinematics turns a gauge-theory amplitude into a gravity amplitude:

\[A_n^{\text{gauge}} = g^{n-2} \sum_i \frac{c_i\, n_i}{D_i} \quad\longrightarrow\quad M_n^{\text{gravity}} = \Big(\frac{\kappa}{2}\Big)^{n-2} \sum_i \frac{n_i\, \tilde n_i}{D_i}\]

Schematically, “gravity = gauge $\times$ gauge”. This is not just an amplitude curiosity: it has produced state-of-the-art predictions for the gravitational two-body problem relevant to gravitational-wave observatories, deriving classical post-Minkowskian dynamics by squaring gauge-theory amplitudes. It also hints that gravity is, at some deep level, the square of a simpler gauge theory — tying the amplitudes program back to the quantum-gravity question.

Amplitudes are simple. On-shell recursion and unitarity reconstruct amplitudes from physical data alone; the Lagrangian’s complexity is redundant.
Holography is exact (in AdS). A gravity theory in the bulk equals a CFT on the boundary; strong coupling on one side is weak curvature on the other.
Anomalies are double-edged. Global anomalies predict real physics ($\pi^0 \to \gamma\gamma$); gauge anomalies must cancel, constraining the particle content.
Gravity is an EFT. General relativity is predictive below $M_{\text{Pl}}$ and needs only a UV completion — asymptotic safety, strings, or holography.
Entanglement builds geometry. Ryu-Takayanagi ties boundary entanglement entropy to bulk minimal-surface area — spacetime emerges from information.
Gravity = gauge squared. The double copy turns gauge amplitudes into gravity amplitudes, powering gravitational-wave two-body predictions.