Neutral Triple Gauge Couplings (nTGCs)

• Neutral triple gauge couplings (nTGCs) are trilinear interactions among neutral electroweak bosons (ZZZ, ZZγ, Zγγ) that appear first at dimension-8 in SMEFT, offering null tests of the Standard Model.
• They serve as sensitive probes of physics beyond the Standard Model, with effects observable in the high-energy tails of collider kinematics and interference patterns.
• Collider experiments use both rate-based and angular observables to constrain nTGC parameters, linking EFT coefficients to potential UV completions and new dynamics.

Neutral triple gauge couplings (nTGCs) are trilinear interactions among neutral electroweak gauge bosons, conventionally $ZZZ$, $ZZ\gamma$, and $Z\gamma\gamma$. In the renormalizable Standard Model they are absent at tree level, and in the Standard Model Effective Field Theory (SMEFT) they first arise at dimension 8 rather than dimension 6. They therefore occupy a distinctive position in collider phenomenology: unlike charged triple gauge couplings, they are simultaneously null tests of the Standard Model at tree level and direct probes of genuinely higher-dimensional electroweak dynamics (Senol et al., 2019, Liu et al., 2024).

1. Standard-Model status and conceptual role

Charged triple gauge couplings such as $WW\gamma$ and $WWZ$ are present already at tree level because they follow from the non-Abelian $SU(2)_L\times U(1)_Y$ structure. Neutral vertices do not share this status. In the formulation used in HL-LHC nTGC studies, “The tree-level vertices of three neutral gauge bosons are not allowed since it violates the underlying $SU(2)_L\times U(1)_Y$ symmetry” (Senol et al., 2019). This statement underlies the standard classification of nTGCs as a clean beyond-the-Standard-Model probe.

A second structural point is their loop suppression inside the Standard Model. In a hadron-collider analysis of CP-violating neutral couplings, the one-loop Standard-Model contributions to the CP-even form factors were quoted as $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$, while the CP-odd structures $f_4^V,\ h_1^V,\ h_2^V$ vanish at one loop (Biekötter et al., 2021). Measurable CP-odd nTGCs would therefore indicate CP-violating new physics in the gauge sector rather than a Standard-Model background.

A common misconception is that neutral and charged TGCs are on the same EFT footing. They are not. Multiple collider studies treat the absence of nTGCs at dimension 6 as central: in SMEFT they first appear at dimension 8, so their observation would point to a sector whose leading imprint on neutral gauge self-interactions is already beyond the conventional dimension-6 analysis (Liu et al., 2024, Ellis et al., 26 Jun 2025).

2. EFT bases, anomalous-coupling parametrizations, and operator content

A frequently used gauge-invariant dimension-8 basis for nTGC phenomenology contains four Higgs–gauge operators,

$$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$

$ZZ\gamma$0

$ZZ\gamma$1

$ZZ\gamma$2

with effective interactions written schematically as $ZZ\gamma$3 (Senol et al., 2019). In one phenomenological convention, “The coefficients $ZZ\gamma$4 (CP conserving) and $ZZ\gamma$5, $ZZ\gamma$6, $ZZ\gamma$7 (CP violating) of dimension-eight operators describe anomalous Neutral Triple Gauge Couplings (aNTGC)” (Senol et al., 2019). Later work on CP-violating form factors instead organizes the CPV sector through operators $ZZ\gamma$8, so the precise CP labeling depends on basis and convention (Ellis et al., 17 Apr 2025).

In parallel, collider analyses often use a general anomalous-coupling language. In that framework, $ZZ\gamma$9 parameterize $Z\gamma\gamma$0 and $Z\gamma\gamma$1, while $Z\gamma\gamma$2 parameterize $Z\gamma\gamma$3 and $Z\gamma\gamma$4. The CP classification commonly used in LHC fits is

$Z\gamma\gamma$5

with $Z\gamma\gamma$6 (Biekötter et al., 2021).

Three parametrization layers are now standard in the literature.

Framework	Representative parameters	Source
Gauge-invariant dimension-8 Higgs–gauge basis	$Z\gamma\gamma$7	(Senol et al., 2019)
General anomalous-coupling framework	$Z\gamma\gamma$8	(Biekötter et al., 2021)
Gauge-symmetric parameterization model	$Z\gamma\gamma$9 and correlated $WW\gamma$0	(Collaboration, 9 Dec 2025)

Beyond the minimal four-operator set, a complete off-shell CP-even description with two Higgs-doublet fields was formulated as a set of 7 dimension-8 operators generating off-shell nTGCs (Ellis et al., 2024). For collider-specific scans, broader bases have also been used: 6 bosonic operators in composite-signal studies of $WW\gamma$1 (Semushin et al., 2024), and 14 dimension-8 operators, including Higgs-related and pure-gauge structures, in multi-TeV muon-collider analyses (Xie et al., 11 Jul 2025).

3. Amplitude structure, interference patterns, and sensitive observables

The generic EFT decomposition of a neutral-diboson amplitude is

$WW\gamma$2

In the HL-LHC $WW\gamma$3 study this was emphasized as an interference-driven strategy: the quadratic term scales as $WW\gamma$4, while the interference scales as $WW\gamma$5, so “for large $WW\gamma$6, the interference term dominates” (Senol et al., 2019). By contrast, an LHC-wide fit phrased directly in anomalous couplings found that for neutral TGCs the Standard-Model–BSM interference is polarization-suppressed and numerically negligible, so the bounds are often dominated by the quadratic contribution and by the highest transverse-momentum bins (Biekötter et al., 2021). The two statements are not contradictory; they refer to different operator bases, observables, and kinematic regimes.

At hadron colliders, the most common rate-sensitive observables are the hard tails of $WW\gamma$7 and $WW\gamma$8 production. In $WW\gamma$9, the photon $WWZ$0 spectrum develops visible deviations from the Standard Model above $WWZ$1 GeV, becoming pronounced at $WWZ$2 GeV, while the invariant mass $WWZ$3 deviates significantly for $WWZ$4 GeV (Senol et al., 2019). In ATLAS-inspired neutral-coupling fits, the constraints stem almost entirely from the last bin of $WWZ$5 in $WWZ$6 and $WWZ$7, and from the last bin of $WWZ$8 in $WWZ$9 and $SU(2)_L\times U(1)_Y$0 (Biekötter et al., 2021).

Angular information supplies additional discrimination. In the HL-LHC $SU(2)_L\times U(1)_Y$1 analysis, the variable

$SU(2)_L\times U(1)_Y$2

was defined as the polar angle in the $SU(2)_L\times U(1)_Y$3 rest frame with respect to the $SU(2)_L\times U(1)_Y$4 direction in the $SU(2)_L\times U(1)_Y$5 rest frame, and its shape deformation under $SU(2)_L\times U(1)_Y$6 and $SU(2)_L\times U(1)_Y$7 was used in a binned $SU(2)_L\times U(1)_Y$8 fit (Senol et al., 2019). For explicitly CP-violating nTGC studies in $SU(2)_L\times U(1)_Y$9 production, special angular variables,

$SU(2)_L\times U(1)_Y$0

and matrix-element-based optimal observables were constructed from the interference term

$SU(2)_L\times U(1)_Y$1

with the optimal-observable method giving the strongest expected limits in that analysis (Semushin et al., 26 Mar 2025).

This split between high-tail rate analyses and interference-based CP-sensitive observables is one of the defining methodological features of the nTGC literature.

4. Hadron-collider measurements and projections

The HL-LHC $SU(2)_L\times U(1)_Y$2 study at $SU(2)_L\times U(1)_Y$3 TeV used FeynRules, MadGraph5_aMC@NLO, Pythia 6, Delphes 3.3.3 with an ATLAS card, and ROOT/ExRootAnalysis. With cuts including $SU(2)_L\times U(1)_Y$4 GeV and $SU(2)_L\times U(1)_Y$5 GeV, the projected 95% C.L. limits at $SU(2)_L\times U(1)_Y$6 fb$SU(2)_L\times U(1)_Y$7 were

$SU(2)_L\times U(1)_Y$8

for $SU(2)_L\times U(1)_Y$9, degrading modestly for $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$0 and $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$1 systematics (Senol et al., 2019). The same study quoted improvement factors of about 5 for $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$2 and 3 for $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$3 relative to then-current ATLAS Run-2 $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$4 constraints (Senol et al., 2019).

A broader HL-LHC reinterpretation in anomalous-coupling language combined $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$5, $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$6, $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$7, and $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$8. At $f_5^V,\ h_3^V,\ h_4^V \sim 10^{-4}$9 fb$f_4^V,\ h_1^V,\ h_2^V$0, it obtained

$f_4^V,\ h_1^V,\ h_2^V$1

and

$f_4^V,\ h_1^V,\ h_2^V$2

at 95% C.L. (Biekötter et al., 2021). In that framework the constraints on nTGCs stem almost entirely from the high-$f_4^V,\ h_1^V,\ h_2^V$3 tails, and the bounds are nearly symmetric because the quadratic anomalous contribution dominates (Biekötter et al., 2021).

Representative hadron-collider results span both EFT-coefficient and anomalous-coupling languages.

Channel and setup	Observable language	Representative 95% C.L. result
HL-LHC $f_4^V,\ h_1^V,\ h_2^V$4, $f_4^V,\ h_1^V,\ h_2^V$5 fb$f_4^V,\ h_1^V,\ h_2^V$6	$f_4^V,\ h_1^V,\ h_2^V$7	$f_4^V,\ h_1^V,\ h_2^V$8, $f_4^V,\ h_1^V,\ h_2^V$9 TeV$$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$0 (Senol et al., 2019)
HL-LHC $$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$1 combination, $$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$2 fb$$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$3	$$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$4	$$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$5, $$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$6, $$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$7 (Biekötter et al., 2021)
FCC-hh $$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$8, $$\mathcal{O}_{BW} = iH^{\dagger}B_{\mu\nu}W^{\mu\rho}\{D_{\rho},D^{\nu}\}H,$$9 TeV, $ZZ\gamma$00 ab$ZZ\gamma$01	$ZZ\gamma$02	$ZZ\gamma$03, $ZZ\gamma$04 TeV$ZZ\gamma$05 (Yilmaz, 2021)

At a 100 TeV FCC-hh, the $ZZ\gamma$06 channel yielded projected one-parameter limits

$ZZ\gamma$07

for $ZZ\gamma$08 ab$ZZ\gamma$09 and no systematics (Yilmaz et al., 2019). The $ZZ\gamma$10 channel at the same machine improved these to

$ZZ\gamma$11

respectively (Yilmaz, 2021).

Run-3 data have already begun to enter this territory experimentally. CMS measured the $ZZ\gamma$12 fiducial cross section at $ZZ\gamma$13 TeV with 34.8 fb$ZZ\gamma$14 as

$ZZ\gamma$15

in agreement with the Standard-Model prediction $ZZ\gamma$16 pb, and set nTGC limits both in the traditional vertex-parameterization model and, for the first time in this channel, in the gauge-symmetric parameterization model (Collaboration, 9 Dec 2025). The observed combined limits included

$ZZ\gamma$17

at 95% C.L. (Collaboration, 9 Dec 2025).

Methodologically, one hadron-collider study argued that neutral-coupling sensitivity can be increased even at fixed luminosity by including EFT effects in backgrounds. In $ZZ\gamma$18, the “composite anomalous signal” prescription found that the dominant beyond-the-Standard-Model background contribution comes from $ZZ\gamma$19, improving one-dimensional limits by up to $ZZ\gamma$20 in the linear+quadratic EFT model and up to $ZZ\gamma$21 in the linear EFT model (Semushin et al., 2024).

5. Electron–positron and muon colliders

Lepton colliders probe nTGCs in a more differential and often more interference-sensitive regime. For $ZZ\gamma$22 at CEPC, a detector-level analysis at $ZZ\gamma$23 GeV and $ZZ\gamma$24 ab$ZZ\gamma$25 found expected 95% C.L. sensitivities to form factors at the level $ZZ\gamma$26, corresponding to new-physics scales of about $ZZ\gamma$27–$ZZ\gamma$28 TeV for order-one Wilson coefficients (Liu et al., 2024). The same paper emphasized that nTGCs are “pure” dimension-8 effects in SMEFT and exploited the full spin correlations in $ZZ\gamma$29 (Liu et al., 2024).

Later $ZZ\gamma$30 studies extended both the operator basis and the treatment of form factors. For CP-violating nTGCs, a new form-factor formulation compatible with spontaneous electroweak symmetry breaking led to the relation

$ZZ\gamma$31

and projected sensitivities for future $ZZ\gamma$32 colliders ranging from $ZZ\gamma$33 at $ZZ\gamma$34 GeV to $ZZ\gamma$35 at $ZZ\gamma$36 TeV, with associated new-physics scales ranging from $ZZ\gamma$37 to $ZZ\gamma$38 (Ellis et al., 17 Apr 2025). Beam polarization was found to improve these probes.

The $ZZ\gamma$39 channel adds a feature unavailable to $ZZ\gamma$40: direct access to the pure $ZZ\gamma$41 sector. In a gauge-consistent $ZZ\gamma$42 study, the operator

$ZZ\gamma$43

was identified as a pure $ZZ\gamma$44 combination, in contrast to $ZZ\gamma$45, which probes the mixed $ZZ\gamma$46 sector (Ellis et al., 26 Jun 2025). With visible and invisible $ZZ\gamma$47 decays, angular observables, machine learning, and 5 ab$ZZ\gamma$48, the unpolarized 2$ZZ\gamma$49 reach at $ZZ\gamma$50 TeV was

$ZZ\gamma$51

improving to

$ZZ\gamma$52

with polarized beams (Ellis et al., 26 Jun 2025). The same analysis reported machine-learning improvements of $ZZ\gamma$53–$ZZ\gamma$54 for $ZZ\gamma$55, $ZZ\gamma$56–$ZZ\gamma$57 for $ZZ\gamma$58, $ZZ\gamma$59–$ZZ\gamma$60 for $ZZ\gamma$61, and $ZZ\gamma$62–$ZZ\gamma$63 for $ZZ\gamma$64 (Ellis et al., 26 Jun 2025).

Muon colliders have produced two complementary nTGC programs. In photon-induced $ZZ\gamma$65, a 14 TeV analysis with 20 ab$ZZ\gamma$66 reported best 95% C.L. limits

$ZZ\gamma$67

without systematic uncertainty (Spor, 2022). A separate high-energy study of $ZZ\gamma$68 over $ZZ\gamma$69–$ZZ\gamma$70 TeV considered 14 dimension-8 operators, found that annihilation dominates over vector-boson fusion in this process at TeV scales, and reported that the $ZZ\gamma$71 polarization enhances sensitivity to several operators, with two pure-gauge operators giving the most stringent expected constraints (Xie et al., 11 Jul 2025).

6. Gauge-consistent interpretation, UV completion, and open issues

One of the main theoretical controversies in the subject is not whether nTGCs are useful, but how they should be parameterized. The older vertex-parameterization model respects only $ZZ\gamma$72, whereas more recent formulations impose full electroweak gauge symmetry in the broken phase. CMS explicitly compared these two languages: a direct VPM bound on $ZZ\gamma$73 at the level of $ZZ\gamma$74 translated into a gauge-symmetric bound of order $ZZ\gamma$75, so the VPM limit appears about 40–50 times tighter numerically, but the gauge-symmetric result is the theoretically robust one (Collaboration, 9 Dec 2025). In that framework the electroweak symmetry also imposes correlations such as

$ZZ\gamma$76

for the operator $ZZ\gamma$77 (Collaboration, 9 Dec 2025).

The same issue reappears in the modern form-factor program. For CP-violating nTGCs, a new extended basis was introduced precisely because the conventional $ZZ\gamma$78 structure leads to high-energy behavior incompatible with the equivalence theorem unless accompanied by an additional form factor $ZZ\gamma$79 satisfying $ZZ\gamma$80 (Ellis et al., 17 Apr 2025). For $ZZ\gamma$81, a parallel effort formulated $ZZ\gamma$82 and $ZZ\gamma$83 in a way that is explicitly compatible with spontaneous symmetry breaking and with dimension-8 matching (Ellis et al., 26 Jun 2025).

EFT validity is another recurrent theme. Since the sensitivity is driven by the far tails, overflow bins can matter numerically. In the anomalous-coupling fits of $ZZ\gamma$84 and $ZZ\gamma$85, adding the overflow bin would tighten $ZZ\gamma$86 and $ZZ\gamma$87 limits by about $ZZ\gamma$88, but for the higher-derivative $ZZ\gamma$89 it can halve the limit; the same study emphasized that this makes EFT interpretation more delicate in the highest-energy region (Biekötter et al., 2021). Hadron-collider EFT studies based on $ZZ\gamma$90 GeV or $ZZ\gamma$91 GeV improved sensitivity, but generally did not present explicit unitarity bounds (Senol et al., 2019).

Finally, UV completion studies have clarified what kinds of heavy dynamics can generate nTGCs. In a renormalizable model with vector-like heavy fermions, one-loop matching produces a complete set of 7 dimension-8 CP-even operators generating off-shell nTGCs, and heavy–light mixing yields extra logarithmic corrections that cannot be accommodated by conventional form-factor parametrizations (Ellis et al., 2024). This suggests that sufficiently precise nTGC measurements do more than constrain contact terms: they can begin to discriminate between purely local EFT deformations and UV structures with nontrivial momentum dependence.

Taken together, these developments place nTGCs at the intersection of collider precision physics, higher-dimensional EFT, and UV-model diagnosis. Their experimental hallmark is a combination of hard diboson tails and distinctive angular structure; their theoretical hallmark is that they are genuine dimension-8 electroweak effects whose robust interpretation requires gauge-consistent operator and form-factor bases (Biekötter et al., 2021, Ellis et al., 2024).