Papers
Topics
Authors
Recent
Search
2000 character limit reached

BCJ Relations in Gauge Theory

Updated 17 June 2026
  • BCJ relations are linear identities among color-ordered tree-level scattering amplitudes derived via color–kinematics duality, reducing the independent basis to (n-3)!.
  • They enable the double-copy construction of gravitational amplitudes, linking gauge theory to gravity through systematic amplitude reductions.
  • Generalized using on-shell recursion and algebraic methods, BCJ relations extend to scalar theories, QCD, string theory, and loop-level integrands.

The Bern–Carrasco–Johansson (BCJ) Relations are a system of linear identities among color-ordered tree-level scattering amplitudes, discovered in the context of gauge theory and string theory. They arise from a conjectured duality between color and kinematics (color–kinematics duality), which posits that numerators in cubic diagram expansions of gauge-theory amplitudes can be arranged to obey the same Jacobi identities as color factors. The BCJ relations drastically reduce the number of independent partial amplitudes and underpin the double-copy construction of gravitational amplitudes from gauge-theory inputs.

1. Algebraic Structure and Statement of the BCJ Relations

Let Afull(1,,n)A_{\text{full}}(1,\dots,n) denote the full tree-level gluon amplitude, decomposed as

Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),

where AnA_n are the color-ordered partial amplitudes. The fundamental BCJ relation for nn legs reads

j=2n1(k1k2++j)An(2,3,,j,1,j+1,,n)=0,\sum_{j=2}^{n-1}\left(k_1\cdot k_{2+\cdots+j}\right)A_n(2,3,\dots,j,1,j+1,\dots,n) = 0,

with k2++j=k2+k3++kjk_{2+\cdots+j}=k_2+k_3+\cdots+k_j and Mandelstam invariants s1,2j=2k1k2js_{1,2\cdots j}=2k_1\cdot k_{2\cdots j} (Feng et al., 2011, Brown et al., 2016, Cruz et al., 2015).

This linear system, together with cyclic invariance and Kleiss–Kuijf (KK) relations, reduces the basis of nn-point tree-level amplitudes to (n3)!(n-3)! independent color orderings.

In scalar cubic theories, an analogous relation holds for color-ordered adjoint scalar amplitudes ASA_S: Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),0 where Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),1 (Du et al., 2011).

2. Color–Kinematics Duality and Origin of BCJ Relations

The color–kinematics duality asserts that gauge-theory amplitudes can be arranged such that the structure constants Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),2 of cubic graphs and kinematic numerators Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),3 satisfy identical Jacobi relations: Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),4 At tree level, this duality directly implies the existence of BCJ amplitude relations among partial amplitudes, as linear consequences of the kinematic Jacobi identities (Feng et al., 2011, Brown et al., 2016, Cruz et al., 2015).

The duality is not unique to gluons: scalar cubic theories also satisfy BCJ relations via the antisymmetry and Jacobi identities of their structure constants (Du et al., 2011).

3. On-Shell Proofs and Recursion Techniques

The BCJ relations can be proved using on-shell recursion relations (BCFW shifts). The key steps involve:

  • Performing a complex deformation of a pair of external momenta, for instance, Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),5 with Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),6.
  • Studying the large-Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),7 scaling of deformed amplitudes, where for nonadjacent leg shifts, Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),8, enabling vanishing boundary terms.
  • Constructing a contour integral Afull(1,,n)=σSn/ZnTr(Taσ(1)Taσ(n))An(σ(1,2,,n)),A_{\text{full}}(1,\dots,n)=\sum_{\sigma\in S_n/\mathbb{Z}_n}\mathrm{Tr}\bigl(T^{a_{\sigma(1)}}\cdots T^{a_{\sigma(n)}}\bigr)A_n\bigl(\sigma(1,2,\dots,n)\bigr),9 of the BCJ sum, which vanishes by power counting.
  • Demonstrating that residues at all physical factorization poles cancel among terms in the BCJ sum, yielding a vanishing result and thus proving the relation (Feng et al., 2011, Brown et al., 2016, Jia et al., 2010, Cruz et al., 2015).

In scalar theories, nonadjacent BCFW shifts produce nontrivial boundary terms that are handled by diagrammatic grouping and cancellations via the Jacobi identity (Du et al., 2011).

Supersymmetric theories, notably AnA_n0 SYM, allow all component BCJ relations to be recovered from superfield BCFW recursion due to the absence of boundary contributions and the polynomial nature of component expansion in Grassmann variables (Jia et al., 2010).

4. Structural Reduction: Cyclicity, KK, and KLT Double Copy

A hierarchy links the various amplitude relations:

  • Cyclicity: cyclic invariance of AnA_n1 reduces the basis from AnA_n2 to AnA_n3 amplitudes.
  • Kleiss–Kuijf (KK) relations: further reduction to AnA_n4 via AnA_n5 decoupling identities.
  • BCJ relations: final reduction to AnA_n6 via the fundamental BCJ linear constraints (Feng et al., 2011).

This minimal BCJ basis directly enables the Kawai–Lewellen–Tye (KLT) double-copy construction of gravity tree amplitudes from gauge theory: AnA_n7 where AnA_n8 is the KLT/BCJ kernel composed of Mandelstam invariants (Feng et al., 2011, Brown et al., 2016, Grassi et al., 2011).

5. Extensions and Generalizations

(a) Scalar, Fermion, and QCD Amplitudes

BCJ relations apply to amplitudes with external matter (quarks, scalars), provided at least one gluon is present. In QCD, the relations take the form

AnA_n9

where 2 is the chosen gluon (Cruz et al., 2015, Brown et al., 2016).

(b) String Theory and Amplitude Monodromy

Open string amplitudes at four points satisfy "stringy" BCJ relations: nn0 reflecting monodromy properties of the integration domain and generalizing the field-theory BCJ relation to all energies. In the nn1 limit, this reduces to the Yang–Mills BCJ identity (Lai et al., 2016).

(c) Gauge Invariance and Symmetry-Based Derivations

Recent proofs demonstrate that BCJ relations are enforced by a symmetry of gauge-theory amplitudes: invariance under momentum-dependent shifts of color factors, termed "color-factor symmetry." This perspective unifies string, field-theory, and diagrammatic proofs and is applicable to amplitudes with arbitrary massless/massive external states in any representation, so long as gluons are present (Brown et al., 2016).

(d) Beyond Flat Space: AdS and Polytopal Generalizations

Color–kinematics duality, and hence BCJ relations, generalize to anti-de Sitter (AdS) correlators by replacing kinematic invariants with differential operators acting on embedding-space contact diagrams. The resulting BCJ-type relations among boundary correlators reduce, in the flat-space limit, to standard field-theory BCJ equations (Diwakar et al., 2021).

Generalizations also arise in twisted de Rham cohomology and positive-geometry approaches, where the number of independent BCJ basis elements matches the number of bounded chambers of accordiohedral polytopes, extending the usual nn2 counting of the CHY/associahedral case (Kalyanapuram, 2020).

6. Loop-Level Generalizations

The extension of BCJ relations to one-loop and higher-loop integrands is an area of active research. For one-loop amplitudes, integrand-level BCJ-type relations relate various partial integrands and are crucial for manifestly gauge-invariant KLT-type double-copy formulas at one loop (Dong et al., 2023, Chester, 2016, He et al., 2016, Cao et al., 29 Sep 2025, He et al., 2015). The key structural property persists, with kinematic numerators and color factors entering one-loop graphs such that algebraic Jacobi identities among color factors induce corresponding relations among kinematic numerators and partial amplitudes.

7. Algebraic and Hopf-Algebraic Interpretations

Recent advances use the formalism of kinematic Hopf algebras to organize BCJ numerators and amplitude relations. The algebra is built from non-commutative generators labeled by flavor and kinematical structure, with the extended quasi-shuffle product encoding all symmetries, Jacobi identities, and amplitude relations. This structure makes the combinatorics and gauge invariance of BCJ numerators manifest and provides closed-form expressions for numerators satisfying all necessary constraints (Chen et al., 2022).


References:

  • (Feng et al., 2011) An Introduction to On-shell Recursion Relations
  • (Du et al., 2011) BCJ Relation of Color Scalar Theory and KLT Relation of Gauge Theory
  • (Brown et al., 2016) Color-factor symmetry and BCJ relations for QCD amplitudes
  • (Cruz et al., 2015) Proof of the fundamental BCJ relations for QCD amplitudes
  • (Jia et al., 2010) U(1)-decoupling, KK and BCJ relations in nn3 SYM
  • (Cachazo, 2012) Fundamental BCJ Relation in N=4 SYM From The Connected Formulation
  • (Chen et al., 2022) Kinematic Hopf algebra for amplitudes and form factors
  • (Lai et al., 2016) The String BCJ Relations Revisited and Extended Recurrence relations of Nonrelativistic String Scattering Amplitudes
  • (Diwakar et al., 2021) BCJ Amplitude Relations for Anti-de Sitter Boundary Correlators in Embedding Space
  • (Kalyanapuram, 2020) On Polytopes and Generalizations of the KLT Relations
  • (Dong et al., 2023) One-loop Bern-Carrasco-Johansson Numerators on Quadratic Propagators from the Worldsheet
  • (He et al., 2016) New Relations for Gauge-Theory and Gravity Amplitudes at Loop Level
  • (Cao et al., 29 Sep 2025) Loop-Level Double Copy Relations from Forward Limits
  • (Mafra et al., 2015) Berends-Giele recursions and the BCJ duality in superspace and components
  • (Grassi et al., 2011) BCJ and KK Relations from BRST Symmetry and Supergravity Amplitudes

Summary Table of Structural Reductions:

Symmetry/Relation Number of Independent Orders Prototype Relation Form
Cyclicity nn4 nn5
KK (Kleiss–Kuijf) nn6 nn7
BCJ nn8 nn9

The BCJ relations are foundational to the modern analytic S-matrix program, directly underlie the color–kinematics duality and the double-copy construction of gravitational amplitudes, and have been generalized to multiple classes of quantum field theories and string models. The algebraic, geometric, and recursion-related structures emerging from these relations unify many aspects of gauge and gravity amplitudes.

Topic to Video (Beta)

No one has generated a video about this topic yet.

Whiteboard

No one has generated a whiteboard explanation for this topic yet.

Follow Topic

Get notified by email when new papers are published related to Bern-Carrasco-Johansson (BCJ) Relations.