Extended Higgs Sectors: Alignment, Custodial Symmetry, and Decoupling

An extended Higgs sector is physically specified by representations, a complete potential, the selected vacuum, mass eigenstates, and Yukawa tensors—not by basis-dependent Lagrangian labels. In a two-doublet benchmark, alignment fixes the Standard-Model-like gauge direction, custodial symmetry constrains invariant mass/coupling combinations, flavor safety constrains Yukawa alignment, and decoupling requires a gauge-invariant heavy scale distinct from large quartics.

Required background. Consistency Checklist for Standard Model Extensions supplies the structural gates. Higgs Self-Interactions and the Scalar Potential supplies the electroweak vacuum and scalar-coupling conventions.

Helpful background. Vacuum Orbits and Unbroken Subgroups explains why gauge-related vacuum representatives are not distinct physical vacua.

A bounded two-doublet benchmark

Take two hypercharge-$+1/2$ doublets and a CP-conserving, softly broken $\mathbb Z_2$ potential,

$$\begin{aligned} V={}&m_{11}^2\Phi_1^\dagger\Phi_1+m_{22}^2\Phi_2^\dagger\Phi_2 -m_{12}^2(\Phi_1^\dagger\Phi_2+\text{h.c.})\\ &+\frac{\lambda_1}{2}(\Phi_1^\dagger\Phi_1)^2 +\frac{\lambda_2}{2}(\Phi_2^\dagger\Phi_2)^2 +\lambda_3(\Phi_1^\dagger\Phi_1)(\Phi_2^\dagger\Phi_2)\\ &+\lambda_4|\Phi_1^\dagger\Phi_2|^2 +\frac{\lambda_5}{2}\left[(\Phi_1^\dagger\Phi_2)^2+\text{h.c.}\right], \end{aligned}$$

with all displayed coefficients real. This fixture deliberately omits hard $\mathbb Z_2$ breaking; the general two-doublet model contains additional couplings and phases.

For a neutral vacuum,

$$\langle\Phi_1^0\rangle=\frac{v_1}{\sqrt2}, \qquad \langle\Phi_2^0\rangle=\frac{v_2}{\sqrt2}, \qquad v^2=v_1^2+v_2^2,$$

define $s_\beta=v_2/v$, $c_\beta=v_1/v$, $\lambda_{345}=\lambda_3+\lambda_4+\lambda_5$, and $M^2=m_{12}^2/(s_\beta c_\beta)$. After the three gauge Goldstones are removed, the physical fields are $H^\pm$, a CP-odd scalar $A$, and two CP-even scalars $h,H$. Their CP-odd and charged masses are

$$m_A^2=M^2-\lambda_5v^2, \qquad m_{H^\pm}^2=M^2-\frac{\lambda_4+\lambda_5}{2}v^2.$$

In the neutral CP-even basis $(\rho_1,\rho_2)$, after using stationarity,

$$\mathcal M_{\rm even}^2= \begin{pmatrix} M^2s_\beta^2+\lambda_1v^2c_\beta^2 & (-M^2+\lambda_{345}v^2)s_\beta c_\beta\\ (-M^2+\lambda_{345}v^2)s_\beta c_\beta & M^2c_\beta^2+\lambda_2v^2s_\beta^2 \end{pmatrix}.$$

An orthogonal rotation by $\alpha$ gives $m_h^2\le m_H^2$. The gauge couplings satisfy

$$\frac{g_{hVV}}{g_{hVV}^{\rm SM}}=\sin(\beta-\alpha), \qquad \frac{g_{HVV}}{g_{hVV}^{\rm SM}}=\cos(\beta-\alpha),$$

for $V=W,Z$. These ratios are physical; $\tan\beta$ alone is not basis invariant in the general model. It becomes meaningful only when a Yukawa or symmetry structure selects a doublet basis Davidson and Haber 2005, §§II–III.

Boundedness and vacuum selection

For the restricted CP-conserving quartic above, necessary and sufficient bounded-from-below conditions are

$$\lambda_1>0, \quad \lambda_2>0, \quad \lambda_3>-\sqrt{\lambda_1\lambda_2}, \quad \lambda_3+\lambda_4-|\lambda_5|>-\sqrt{\lambda_1\lambda_2}.$$

They do not prove that the chosen neutral extremum is the global minimum. Solve the stationarity equations, test the charged and neutral physical Hessians, and compare every normal, charge-breaking, and CP-breaking stationary point allowed by the fixture. For a more general potential, reuse neither these inequalities nor this CP classification without rederiving them.

Perturbative unitarity is a matrix condition. Construct the high-energy scalar–scalar $2\to2$ coupled-channel amplitudes, diagonalize the $J=0$ partial-wave matrix, and require

$$|\operatorname{Re}a_{0,k}|\le\frac12$$

for every eigenchannel. An individual rule such as $|\lambda_i|<4\pi$ is at most a screening heuristic and can miss dangerous linear combinations Kanemura, Kubota, and Takasugi 1993, pp. 155–158.

Alignment, custodial symmetry, and decoupling

The Higgs basis rotates the doublets so that only $H_1$ has a vacuum expectation value. Its scalar potential contains a basis-covariant coefficient $Z_6$ multiplying $(H_1^\dagger H_1)H_1^\dagger H_2$. Exact alignment is

$$Z_6=0 \quad\Longleftrightarrow\quad \cos(\beta-\alpha)=0$$

for the light Standard-Model-like state, apart from a mass degeneracy requiring separate treatment. Alignment can occur without heavy extra scalars through a coupling relation. In the decoupling limit,

$$\cos(\beta-\alpha) \simeq-\frac{Z_6v^2}{m_H^2-m_h^2}, \qquad M^2\gg v^2,$$

so a gauge-invariant $M^2$ suppresses mixing at fixed perturbative quartics Gunion and Haber 2003, §§III–IV. Taking heavy masses large through $\lambda_i v^2$ instead can violate unitarity and leave nondecoupling loop effects.

Custodial symmetry is also an invariant statement about the potential and vacuum. In this restricted CP-conserving basis, $\lambda_4=\lambda_5$ gives the representative degeneracy $m_{H^\pm}=m_A$; it is a sufficient fixture, not the most general custodial realization. Hypercharge and Yukawa interactions break custodial symmetry even when the scalar potential respects it. The basis-independent conditions and alternative mass degeneracies are given by Haber and O’Neil 2011, §§4–5.

Exact aligned fixture

Choose

$$\lambda_1=\lambda_2=1, \quad \lambda_3=\frac12, \quad \lambda_4=\lambda_5=0, \quad \tan\beta=1, \quad M^2=4v^2.$$

The boundedness inequalities hold. The CP-even mass matrix is

$$\frac{\mathcal M_{\rm even}^2}{v^2} =\begin{pmatrix}5/2&-7/4\\-7/4&5/2\end{pmatrix},$$

with eigenvalues $3v^2/4$ and $17v^2/4$. The light eigenvector is proportional to $(1,1)$, exactly the vacuum direction, so it is aligned. Also $m_A^2=m_{H^\pm}^2=4v^2$. These algebraic results do not replace the remaining stationary-point comparison or loop-level RG check.

Yukawa flavor safety

With two identical scalar representations, generic Yukawa matrices $Y_1^f$ and $Y_2^f$ cannot generally be diagonalized simultaneously, producing neutral-scalar flavor change at tree level. Two standard protections are:

• a symmetry that makes each right-handed fermion type couple to only one doublet, realizing natural flavor conservation; or
• proportional/aligned Yukawa matrices in flavor space, whose radiative stability must be checked in the declared theory.

The symmetry conditions for natural flavor conservation are due to Glashow and Weinberg 1977, pp. 1958–1965. A label such as “Type I” or “Type II” is incomplete until the charge assignment, basis, and soft breaking are specified.

Checks and handoffs

• Verify Goldstone counting, positive physical masses, and the trace/determinant of both charged and neutral mass matrices.
• Apply boundedness, vacuum selection, and coupled-channel unitarity as independent tests.
• Express conclusions through masses, invariant couplings, and amplitudes; do not treat a freely rotatable basis angle as observable.
• Recover the Standard Model amplitude in the exact-alignment/decoupling limit after matching inputs, while retaining calculable heavy-threshold effects.
• Keep pole widths and interference when heavy states overlap; current coupling fits and exclusions require versioned likelihoods.

Singlet-only mixing belongs to Higgs-Singlet Scalar Portals. Supersymmetric Higgs potentials belong to the supersymmetry volume, thermal histories to the thermal volume, and live model comparisons to Effective Field Theory and Tests of the Standard Model.

References

• Davidson, Sacha, and Howard E. Haber. “Basis-Independent Methods for the Two-Higgs-Doublet Model.” Physical Review D 72 (2005): 035004. DOI.
• Glashow, Sheldon L., and Steven Weinberg. “Natural Conservation Laws for Neutral Currents.” Physical Review D 15 (1977): 1958–1965. DOI.
• Gunion, John F., and Howard E. Haber. “The CP-Conserving Two-Higgs-Doublet Model: The Approach to the Decoupling Limit.” Physical Review D 67 (2003): 075019. DOI.
• Haber, Howard E., and Deva O’Neil. “Basis-Independent Methods for the Two-Higgs-Doublet Model III: The CP-Conserving Limit, Custodial Symmetry, and the Oblique Parameters S, T, U.” Physical Review D 83 (2011): 055017. DOI.
• Kanemura, Shinya, Takahiro Kubota, and Eiichi Takasugi. “Lee–Quigg–Thacker Bounds for Higgs Boson Masses in a Two Doublet Model.” Physics Letters B 313 (1993): 155–160. DOI.