Inert Two-Higgs Doublet Model (IDM)

• The IDM is a two-Higgs doublet extension characterized by an exact Z2 symmetry that segregates the active and inert scalar doublets.
• It features an inert scalar sector where only one doublet acquires a vacuum expectation value, leading to stable dark matter candidates and unique decay patterns.
• Experimental constraints from collider searches, direct detection, and electroweak precision tests tightly correlate IDM parameters, defining specific mass hierarchies and portal interactions.

The inert two-Higgs doublet model (IDM), also called the Inert Doublet Model or Dark 2HDM, is a two-Higgs-doublet extension of the Standard Model with an exact discrete $Z_2$ symmetry under which the Standard Model fields and one scalar doublet are even, while the second scalar doublet is odd. In the inert vacuum only the $Z_2$-even doublet acquires a vacuum expectation value, so the $Z_2$-odd doublet does not mix with the Standard-Model-like Higgs sector and has no Yukawa couplings to fermions. Its lightest neutral state is therefore stable and can serve as a dark-matter candidate (Kalinowski et al., 2019).

1. Symmetry structure, field content, and notation

In the standard formulation the scalar sector contains two $SU(2)_L$ doublets with hypercharge $Y=1/2$, commonly denoted either by $(\Phi_1,\Phi_2)$ or by $(\Phi_S,\Phi_D)$. The exact $Z_2$ action is

$\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$

or equivalently

$\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$

Only the $Z_2$0-even doublet acquires the electroweak vacuum expectation value $Z_2$1, while the inert doublet has no vev and no renormalisable couplings to fermions. After electroweak symmetry breaking the physical spectrum consists of the Standard-Model-like Higgs boson $Z_2$2, a neutral CP-even inert scalar $Z_2$3, a neutral CP-odd inert scalar $Z_2$4, and a charged pair $Z_2$5 (Zarnecki et al., 2020).

Literature conventions are not fully uniform. Some analyses denote the lightest neutral CP-even inert scalar by $Z_2$6 instead of $Z_2$7, and emphasize that the physics is symmetric under $Z_2$8 (He et al., 2024). In the standard IDM convention used in many collider and dark-matter studies, $Z_2$9 is chosen as the lightest $Z_2$0-odd state.

State	$Z_2$1 parity / role	Dominant transitions
$Z_2$2	even, SM-like Higgs	SM-like couplings
$Z_2$3	odd, neutral, often DM	stable if lightest odd state
$Z_2$4	odd, neutral	$Z_2$5
$Z_2$6	odd, charged	$Z_2$7

The exact $Z_2$8 symmetry has two immediate consequences. First, inert states are pair-produced in gauge interactions. Second, tree-level decays of inert scalars into Standard Model fermions through Yukawa couplings are absent. This sharply distinguishes the IDM from conventional $Z_2$9-symmetric 2HDMs in which both doublets acquire vevs and the physical CP-even states are mixtures of the two doublets (0911.2457).

2. Scalar potential, inert vacuum, and mass relations

In standard IDM notation the scalar potential is

$SU(2)_L$0

With $SU(2)_L$1 and $SU(2)_L$2, the tree-level masses are

$SU(2)_L$3

$SU(2)_L$4

with

$SU(2)_L$5

A particularly useful identity is

$SU(2)_L$6

so the sign and magnitude of $SU(2)_L$7 determine the ordering and splitting of the neutral inert scalars (Zarnecki et al., 2020).

The inert vacuum is not only a choice of field basis but a dynamical condition. Tree-level bounded-from-below requirements are commonly written as

$SU(2)_L$8

A widely used condition ensuring that the inert vacuum is the global minimum is

$SU(2)_L$9

For $Y=1/2$0, one analysis obtained the additional bound

$Y=1/2$1

from the requirement that the inert minimum exist and be global (Swiezewska, 2012).

Precision electroweak data constrain the mass splittings among $Y=1/2$2, $Y=1/2$3, and $Y=1/2$4, chiefly through the oblique parameters $Y=1/2$5 and $Y=1/2$6. Global scans therefore tend to favor partial custodial patterns such as $Y=1/2$7 or $Y=1/2$8, especially once collider and dark-matter constraints are imposed. In a scan allowing the IDM to furnish only a subdominant fraction of the dark matter, the combined constraints led to a relatively strong mass degeneracy in the dark scalar sector for masses below $Y=1/2$9, a minimal dark-scalar mass scale of about $(\Phi_1,\Phi_2)$0, and a hierarchy $(\Phi_1,\Phi_2)$1 across the surviving parameter space (Ilnicka et al., 2015).

3. Higgs portal, dark matter, and viable parameter regimes

The phenomenology of the lightest inert scalar is governed by the Higgs portal and by electroweak gauge interactions. The $(\Phi_1,\Phi_2)$2 coupling is

$(\Phi_1,\Phi_2)$3

and, when kinematically open, the invisible Higgs width is

$(\Phi_1,\Phi_2)$4

in conventions where $(\Phi_1,\Phi_2)$5 is a real scalar (Krawczyk et al., 2013). The same coupling controls the leading tree-level spin-independent direct-detection rate, schematically

$(\Phi_1,\Phi_2)$6

so direct-detection experiments strongly constrain $(\Phi_1,\Phi_2)$7 (Cembranos et al., 2010).

Several dark-matter regimes recur throughout the IDM literature. For low masses, annihilation through the Higgs funnel near $(\Phi_1,\Phi_2)$8 can reproduce the relic density with very small $(\Phi_1,\Phi_2)$9. In the intermediate regime around and above the $(\Phi_S,\Phi_D)$0-threshold, gauge annihilation into $(\Phi_S,\Phi_D)$1 and $(\Phi_S,\Phi_D)$2 becomes efficient, often assisted by coannihilation with $(\Phi_S,\Phi_D)$3 and $(\Phi_S,\Phi_D)$4 when the spectrum is compressed. In the heavy regime, $(\Phi_S,\Phi_D)$5, quasi-degenerate inert states and gauge-driven coannihilations again become central (Krawczyk et al., 2013).

A recent precision study of Higgs-strahlung in the IDM distinguished two surviving dark-matter regimes once relic density and direct-detection limits are simultaneously imposed: a low-mass region with $(\Phi_S,\Phi_D)$6 and a high-mass region with $(\Phi_S,\Phi_D)$7. In the same analysis, direct detection implied $(\Phi_S,\Phi_D)$8 few $(\Phi_S,\Phi_D)$9 at $Z_2$0, while heavy nearly degenerate spectra were dominated by a one-loop gauge contribution and still passed current limits (He et al., 2024).

This structure often produces a misconception that the IDM is generically unconstrained because the dark sector is inert. The opposite is closer to the present status: Higgs invisible decays, relic density, direct detection, electroweak precision observables, and LEP/LHC searches together carve out only specific mass hierarchies and portal strengths. Historically, LEP II reinterpretations excluded approximately

$Z_2$1

at $Z_2$2 confidence, up to the LEP kinematic limit (0810.3924).

4. Collider phenomenology: gauge production, compressed spectra, and precision probes

Because inert scalars do not couple to fermions, collider production is gauge-driven. At hadron colliders the leading channels are Drell–Yan processes such as

$Z_2$3

followed by

$Z_2$4

At lepton colliders the canonical channels are

$Z_2$5

through $Z_2$6-channel $Z_2$7 exchange and

$Z_2$8

through $Z_2$9-channel $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$0 exchange. Near threshold, scalar-pair production is $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$1-wave suppressed, with the characteristic $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$2 behavior (Zarnecki et al., 2020).

The collider kinematics are controlled by the splittings $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$3 and $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$4. If these are below $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$5 or $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$6, the gauge bosons in $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$7 and $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$8 are off shell, leading to soft leptons or jets. This is precisely the regime in which LEP and many LHC searches lose efficiency, but it is also the regime favored by coannihilation and electroweak precision constraints.

Run-1 LHC reinterpretations of opposite-sign dilepton plus missing-energy searches excluded $\Phi_1 \to \Phi_1,\qquad \Phi_2 \to -\Phi_2,$9 up to about $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$0 for $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$1, and up to about $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$2 in the most favorable configurations, with a clear complementarity between off-shell-$\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$3 SUSY-like dilepton searches and $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$4-type searches (Sengupta, 2015). Full Run-2 reinterpretations sharpened the picture. A mono-$\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$5 search optimized for the 2HDM+a topology was shown to have limited sensitivity to the IDM because the dominant $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$6 topology usually yields much softer $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$7 bosons and missing transverse energy than the targeted benchmark, so even larger IDM rates can evade the search. In contrast, the VBF invisible-Higgs channel provides the leading collider constraint on the Higgs portal, and soft-lepton analyses are particularly effective for compressed spectra: with full Run 2 data, points with $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$8 and $\phi_1 \to \phi_1,\qquad \phi_2 \to -\phi_2.$9 were excluded in the compressed regime (Lahiri et al., 28 Nov 2025).

Future $Z_2$00 colliders substantially improve coverage because the gauge production mechanism is fixed and the environment is clean. In a study of future linear colliders, leptonic channels with $Z_2$01 yielded discovery reaches of about $Z_2$02 at $Z_2$03, $Z_2$04 at $Z_2$05, and $Z_2$06 at $Z_2$07 for $Z_2$08 production, while the $Z_2$09 channel reached $Z_2$10, $Z_2$11, and $Z_2$12, respectively. At high-energy CLIC, pure leptonic reach saturates, but semi-leptonic $Z_2$13 final states extend the discovery potential to roughly $Z_2$14 (Zarnecki et al., 2020).

Precision Higgs measurements probe the same model from a different angle. In the IDM, leading-order and NLO QCD corrections to $Z_2$15 are Standard-Model-like, while inert effects enter only through electroweak one-loop corrections. In dark-matter-consistent freeze-out scenarios the resulting deviations in $Z_2$16 and $Z_2$17 rates are only at the per-mil to few-$Z_2$18 level; in dark-matter-relaxed scenarios they can reach a few percent and may become observable at the HL-LHC, especially in $Z_2$19 where photon-PDF systematics are less severe (He et al., 2024).

5. Higgs observables, electroweak phase transition, and naturalness

The charged inert scalar modifies $Z_2$20 at one loop, while $Z_2$21 or $Z_2$22 enlarge the total Higgs width if kinematically open. In a dedicated analysis with $Z_2$23, an enhancement $Z_2$24 was found to be impossible if invisible channels are open. For closed invisible channels, $Z_2$25 required

$Z_2$26

together with $Z_2$27 and typically $Z_2$28 (Krawczyk et al., 2013). More recent global Higgs analyses instead use the measured diphoton signal strength to restrict $Z_2$29 and $Z_2$30, making very large loop effects increasingly difficult to maintain (He et al., 2024).

The IDM is also one of the simplest scalar extensions in which a strong first-order electroweak phase transition can occur. Finite-temperature analyses show that large splittings between the dark-matter state and the heavier inert scalars enhance the bosonic cubic terms in the thermal effective potential and strengthen the transition. Benchmark studies found $Z_2$31 for configurations such as

$Z_2$32

with $Z_2$33 in the region simultaneously compatible with relic density and XENON-100 (Krawczyk et al., 2013). A more detailed effective-potential treatment concluded that a sufficiently strong first-order transition is generically possible across several dark-matter mass regimes, but that achieving both a strong transition and the observed thermal relic abundance is possible only in the Higgs-funnel regime once collider and direct-detection constraints are imposed (Blinov et al., 2015).

This does not by itself establish electroweak baryogenesis. The minimal IDM with exact $Z_2$34 lacks new CP-violating sources, so a successful baryogenesis scenario requires additional CP violation beyond the minimal inert setup (Blinov et al., 2015).

The question of naturalness is more contentious. Although the IDM is often grouped with broader 2HDM-motivated BSM constructions, a one-loop Veltman-style cancellation of quadratic divergences was shown to be incompatible with the bounded-from-below conditions in the exact inert model. In that sense, the IDM cannot satisfy the one-loop quadratic-divergence cancellation requirement within its minimal exact-$Z_2$35 realization (0910.4068).

6. Extensions, cosmological deformations, and present outlook

Two directions recur in current IDM research: symmetry-based extensions that explain the exact $Z_2$36, and nonstandard cosmologies that alter the relic-density calculation without changing collider-scale interactions.

A local $Z_2$37 completion can replace the imposed $Z_2$38 by a spontaneously broken gauge symmetry whose remnant stabilizes the dark matter. In that construction the usual IDM $Z_2$39 term is forbidden at the renormalisable level and generated effectively after $Z_2$40 breaking, while new annihilation channels such as $Z_2$41 and $Z_2$42 open up. This allows $Z_2$43, unlike the usual IDM (Ko et al., 2014).

A different modification keeps the particle content of the IDM but changes the pre-BBN expansion history. In a kination-like era with a stiff equation of state $Z_2$44, freeze-out occurs earlier and the relic density increases. Under standard cosmology the IDM is typically underabundant in roughly the $Z_2$45 window because gauge annihilation is too efficient. With kination, the range

$Z_2$46

can be reopened for

$Z_2$47

while still satisfying current experimental constraints (Bélanger et al., 17 Dec 2025).

The contemporary picture is therefore highly structured rather than minimal in practice. In the strict IDM, direct detection constrains $Z_2$48, Higgs observables constrain invisible and loop-induced decays, electroweak precision data constrain splittings, and collider searches are most sensitive either to clean gauge production at future $Z_2$49 machines or to compressed-spectrum strategies such as soft-lepton and VBF channels at the LHC (Kalinowski et al., 2019). A plausible implication is that the IDM is best viewed not as a single benchmark point but as a constrained framework in which dark matter, electroweak symmetry breaking, and collider signatures remain tightly correlated.