Flavour Deconstruction in Gauge Theories

• Flavour Deconstruction is a framework that deconstructs the Standard Model gauge symmetry into family-specific copies, creating flavor hierarchies through sequential symmetry breaking.
• It employs link fields and higher-dimensional operators to generate lighter-family masses and mixings while naturally suppressing flavor-changing neutral currents.
• Realizations range from supersymmetric and minimal Abelian setups to composite-Higgs models, predicting distinctive experimental and cosmological signatures at TeV scales.

Flavour Deconstruction (FD) is a class of non-universal gauge extensions in which the Standard Model (SM) gauge symmetry is deconstructed into separate copies, often one for each family or one for the light families and one for the third family, and the SM emerges only after a hierarchical chain of spontaneous symmetry breaking. In these constructions, the Higgs is frequently charged only under the third-family gauge factor, so only third-generation Yukawa couplings are renormalizable, while light-family masses and mixings arise from higher-dimensional operators involving link fields or heavy mediators. The same gauge structure can generate accidental flavour symmetries, suppress dangerous flavour-changing neutral currents (FCNCs), and relate flavour hierarchies to TeV- or multi-TeV-scale new dynamics (Greljo et al., 2024, Barbieri et al., 2023, Isidori, 18 Mar 2025).

1. Genealogy of the framework

An early implementation of the core idea appeared in a supersymmetric setting, where deconstruction was used to construct simple, weakly-coupled supersymmetric models that explain the Standard Model flavor hierarchy and produce a flavorful soft spectrum compatible with precision limits. In that two-site construction, electroweak symmetry breaking is fully natural, the $\mu$-term is dynamically generated with no $B\mu$-problem, and the Higgs mass is easily raised above LEP limits without reliance on large radiative corrections. The resulting superpartner spectrum has the characteristic form of “effective supersymmetry”: the third generation superpartners tend to be light, while the rest of the scalars are heavy (Craig et al., 2011).

Subsequent work reformulated the same organizing principle as a broader flavour framework. In lecture notes and recent model-building papers, FD is described as promoting the SM gauge symmetry above the weak scale to a product such as $G^{[12]}\times G^{[3]}$, or more generally to $G^3$, with separate gauge factors assigned to different families and then broken back to the universal SM group. This recasts flavour hierarchies as a consequence of gauge structure and symmetry breaking rather than as arbitrary hierarchies in elementary Yukawa parameters (Isidori, 18 Mar 2025).

A recurring theme across the literature is a multiscale origin of flavour. Third-generation masses and direct Higgs couplings typically appear at the lowest stage, while first- and second-generation masses emerge only after additional symmetry-breaking steps or after integrating out vector-like fermions and heavy Higgs doublets. This multiscale structure is central both to the explanatory power of FD and to its phenomenological viability (Navarro, 2024).

2. Core field-theoretic mechanism

In a two-site deconstruction, the UV gauge symmetry is enlarged to

$$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$

Bifundamental chiral superfields $\chi,\tilde\chi$ connect the sites and break the product gauge group to the diagonal SM at a scale $f\sim \langle\chi\rangle$, via

$W_\chi = A(\chi\tilde\chi - f^2) \, .$

After this breaking, the low-energy gauge couplings satisfy

$$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$

The crucial step is that extended gauge invariance forbids many Yukawa couplings at the renormalizable level; lighter-generation masses are then generated by operators with link-field insertions, for example

$W \supset \frac{1}{M_*^n}\chi^n H Q \bar u \, ,$

with suppression $B\mu$0, yielding a dynamically realized Froggatt–Nielsen-like hierarchy (Craig et al., 2011).

In the family-by-family formulation, the pattern is often presented as a hierarchical chain,

$B\mu$1

with suppression factors

$B\mu$2

When the Higgs is charged only under the third-family factor, only the third family has renormalizable Yukawa couplings, while light-family masses and CKM entries arise from higher-dimensional operators involving the symmetry-breaking scalars (Greljo et al., 2024).

A closely related two-site $B\mu$3 construction places the first and second generation quark doublets on “site 1”, the third generation on “site 2”, and the Higgs also on site 2. Renormalizable Yukawa couplings are then only possible for third generation quarks, whereas the light generations get masses and CKM elements through higher-dimensional operators with insertions of the link field $B\mu$4,

$B\mu$5

This reproduces the generic FD logic in a particularly economical form (Capdevila et al., 2024).

3. Principal realizations

The literature contains several distinct but structurally related realizations of FD.

Realization	Gauge structure	Salient features
Supersymmetric two-site (Craig et al., 2011)	$B\mu$6	Link-field Yukawa suppression; natural EWSB; “effective supersymmetry” spectrum
Deconstructed hypercharge (Davighi et al., 2023)	$B\mu$7	Accidental $B\mu$8; single $B\mu$9; finite naturalness
Minimal flavour deconstruction (Barbieri et al., 2023)	Universal $G^{[12]}\times G^{[3]}$0 plus flavour non-universal Abelian sector	Two-step breaking at $G^{[12]}\times G^{[3]}$1 and $G^{[12]}\times G^{[3]}$2; no small Yukawa coupling, $G^{[12]}\times G^{[3]}$3
$G^{[12]}\times G^{[3]}$4 deconstruction (Capdevila et al., 2024)	$G^{[12]}\times G^{[3]}$5	Massive vector triplet; accidental $G^{[12]}\times G^{[3]}$6
Tri-hypercharge / tri-unified viewpoint (Navarro, 2024)	$G^{[12]}\times G^{[3]}$7, or $G^{[12]}\times G^{[3]}$8	Separate hypercharge per family; $G^{[12]}\times G^{[3]}$9 yields $G^3$0 in the UV
Composite-Higgs FD (Covone et al., 2024)	Non-universal $G^3$1 and $G^3$2	Higgs as pNGB; flavoured gauge bosons and top partners at the TeV scale

These realizations differ in how much of the gauge sector is deconstructed. Some confine flavour non-universality to the Abelian sector, guided by a criterion of minimality; others deconstruct non-Abelian electroweak or Pati–Salam-like structures; still others embed the idea into a composite-Higgs sector or into a tri-$G^3$3 ultraviolet completion with cyclic permutation symmetry $G^3$4 (Navarro, 2024).

Within the supersymmetric two-site model, two benchmark Higgs assignments were emphasized. In the “vector-like Higgs” model, both $G^3$5 and $G^3$6 live on the same node as the third generation, while the first two generations live on the other node. In the “chiral Higgs” model, $G^3$7 and $G^3$8 are split across the two nodes, so that the $G^3$9-term is generated only after link breaking and $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$0 remains naturally small (Craig et al., 2011).

In minimal Abelian FD, the extended gauge group

$$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$1

is reduced to the SM hypercharge in two steps, at scales $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$2 and $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$3, with $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$4. The explicit purpose is to reproduce the charged-fermion flavour pattern without introducing tiny Yukawa couplings by hand (Barbieri et al., 2023).

4. Accidental symmetries, FCNC suppression, and characteristic signals

A central structural result of FD is the emergence of accidental flavour symmetries. In two-site $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$5 deconstruction, the gauge and field content give an accidental global flavour symmetry $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$6. In deconstructed hypercharge and related constructions, the light generations instead enjoy an accidental $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$7 symmetry,

$$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$8

These symmetries suppress dangerous FCNCs because flavour violation is controlled by the same small mixings that generate the observed masses and CKM structure, so new effects arise predominantly through mixing with the third generation and are CKM suppressed, resembling Minimal Flavor Violation in their parametric form (Capdevila et al., 2024, Davighi et al., 2023).

The low-energy spectrum is correspondingly distinctive. The $$G_{SM}^{(1)} \times G_{SM}^{(2)} = [SU(3)_1 \times SU(2)_1 \times U(1)_1] \times [SU(3)_2 \times SU(2)_2 \times U(1)_2] \, .$$9 model yields a massive vector triplet $\chi,\tilde\chi$0, with couplings to fermions and the Higgs determined by site assignment. One of the benchmark lepton arrangements can give a sizeable lepton flavour universal effect in the Wilson coefficient $\chi,\tilde\chi$1 while naturally suppressing contributions to $\chi,\tilde\chi$2, and simultaneously predicts a mild positive shift in the $\chi,\tilde\chi$3 boson mass (Capdevila et al., 2024). In the detailed phenomenology of minimal FD, the lightest new neutral vector boson is $\chi,\tilde\chi$4, and current experimental limits (ATLAS, CMS) exclude $\chi,\tilde\chi$5 below about $\chi,\tilde\chi$6–$\chi,\tilde\chi$7 TeV for most of the parameter space. The same framework predicts $\chi,\tilde\chi$8 at leading order in $\chi,\tilde\chi$9 and organizes the pattern of tree-level FCNCs through the accidental symmetries and flavour rotations (Barbieri, 2024).

Deconstructed hypercharge has a similarly sharp phenomenology. Its low-energy spectrum is dominated by a single $f\sim \langle\chi\rangle$0 gauge boson with chiral and flavour non-universal couplings, with mass as light as a few TeV thanks to the $f\sim \langle\chi\rangle$1 symmetry. The model unavoidably leads to large positive shifts in the $f\sim \langle\chi\rangle$2-boson mass, as well as an enhancement in $f\sim \langle\chi\rangle$3. A future electroweak precision machine such as FCC-ee is described as having the reach to fully exclude the natural parameter space (Davighi et al., 2023).

5. Neutrinos, leptons, and CP structure

FD was initially most successful in the quark and charged-lepton sectors, but large PMNS mixing posed a challenge. If only SM groups such as $f\sim \langle\chi\rangle$4 and $f\sim \langle\chi\rangle$5 are deconstructed, right-handed neutrinos remain gauge singlets; the Dirac Yukawa matrix is then hierarchical while the Majorana mass matrix is anarchic, so the seesaw formula

$f\sim \langle\chi\rangle$6

inherits the hierarchy and tends to produce small mixing, contrary to observation. A key result is that neutrino anarchy can arise when FD is applied instead to carefully chosen extended subgroups such as $f\sim \langle\chi\rangle$7 or $f\sim \langle\chi\rangle$8, so that right-handed neutrinos are charged and the hierarchy in $f\sim \langle\chi\rangle$9 cancels that in $W_\chi = A(\chi\tilde\chi - f^2) \, .$0 (Greljo et al., 2024).

The same general framework also admits a non-anarchic alternative. In the minimal tri-hypercharge theory, after decomposing family hypercharges into the corresponding $W_\chi = A(\chi\tilde\chi - f^2) \, .$1 gauge groups, the charged-lepton hierarchy and the right-handed neutrino charge assignments imply the sequential dominance conditions for a natural neutrino mass hierarchy,

$W_\chi = A(\chi\tilde\chi - f^2) \, .$2

In this case the atmospheric angle receives contributions from both the neutrino and charged-lepton sectors, the reactor angle is dominated by charged-lepton $W_\chi = A(\chi\tilde\chi - f^2) \, .$3-$W_\chi = A(\chi\tilde\chi - f^2) \, .$4 mixing, and the solar angle is dominated by the neutrino sector. The model is presented as showing that neutrino anarchy is not the only viable neutrino outcome of gauge flavour deconstruction (Navarro et al., 26 Jun 2025).

The lepton-sector effective theory has also been developed beyond leading order. Starting from the ultraviolet completion of minimal FD, a systematic spurion expansion yields the charged-lepton Yukawa texture and identifies the dominant sources of flavour and CP violation. At leading order, dipole operators are approximately aligned with the Yukawa matrices, but next-to-leading order contributions generically induce physical CP-violating phases and flavour misalignment. Future searches for $W_\chi = A(\chi\tilde\chi - f^2) \, .$5-$W_\chi = A(\chi\tilde\chi - f^2) \, .$6 conversion and the electron EDM can probe scales in the multi-$W_\chi = A(\chi\tilde\chi - f^2) \, .$7 TeV range under natural assumptions on the flavour structure and CP phases, making charged-lepton flavour violation and EDMs complementary probes of the same FD spurion structure (Masiero et al., 1 Jun 2026).

6. Higgs sector, naturalness, and cosmological extensions

In the supersymmetric formulation, the Higgs sector is intertwined with deconstruction. The $W_\chi = A(\chi\tilde\chi - f^2) \, .$8-term can be generated from link-field vevs, for example

$W_\chi = A(\chi\tilde\chi - f^2) \, .$9

while the $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$0-term is not generated at high scale and instead arises radiatively through MSSM running. The Higgs quartic is enhanced by non-supersymmetric $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$1-term corrections after breaking to the diagonal subgroup,

$$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$2

allowing the Higgs mass to be raised above LEP limits without heavy stops or large $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$3-terms (Craig et al., 2011).

Naturalness considerations continue to shape later realizations. In deconstructed hypercharge, only hypercharge is deconstructed, and the smallness of the hypercharge gauge coupling helps control radiative Higgs mass corrections, so the model satisfies finite naturalness criteria. This setup allows flavour to begin being explained at the TeV scale, while dynamics solving the large hierarchy problem can lie at a higher scale up to around $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$4 TeV without worsening the unavoidable little hierarchy problem (Davighi et al., 2023). In the composite-Higgs realization, the Higgs emerges as a pseudo Nambu–Goldstone boson of

$$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$5

and FD suppresses the couplings of the light families to the composite sector by powers of a heavy mass scale. The radiatively generated Higgs potential contains the ingredients needed to justify the unavoidable tuning required to separate electroweak and composite scales, and the model predicts new TeV-scale states within reach of near-future searches (Covone et al., 2024).

Recent work has extended FD into cosmology. Low-scale semi-simple embeddings of $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$6 generically lead to magnetic monopole production when the extended gauge group is broken through intermediate stages containing an unbroken $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$7 factor. For the scales typical of flavour non-universal models, cosmological and astrophysical constraints require low-scale inflation to dilute the monopoles, followed by reheating below the monopole-production scale, typically around $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$8. This establishes a direct connection between flavour physics and the thermal history of the early Universe (Pesut et al., 19 May 2026). At the same time, the link-field scalar sector inherent to FD can support strong first-order phase transitions and produce primordial gravitational waves. The resulting spectra typically peak at higher frequencies than the millihertz range, so a positive observation at LISA is possible but not guaranteed, whereas the signal falls naturally in the range of mid-band proposals (Fabri et al., 15 Sep 2025).

Across these realizations, FD functions less as a single model than as a model-building principle: flavour hierarchies are encoded in gauge non-universality, lighter-family Yukawas emerge only after symmetry breaking, accidental $$\frac{1}{g_i^2}=\frac{1}{g_{(1)i}^2}+\frac{1}{g_{(2)i}^2} \, .$$9-type symmetries control FCNCs, and the same structure propagates into Higgs physics, neutrino phenomenology, collider signatures, and cosmology.