betaFeedback

WorksPhysics and Astronomy

Mass-Varying Dark Matter, Naturally

Abstract

Ultralight scalars that vary dark-matter masses require protection from quantum corrections. Cyclic ZN\mathbb{Z}_N symmetries provide this protection but also suppress variation of the mean mass of equally populated partners. Unsuppressed variation typically requires asymmetric reheating through symmetry-breaking couplings that threaten radiative control. We show that scalar-induced mass splittings built into these constructions can instead bias relic abundances through symmetric interactions. The mechanism applies to thermal and nonthermal relics that reach internal equilibrium after production, and permits mass variation over a broad range of timescales.

Elementary particle masses need not be constant. The Brout–Englert–Higgs mechanism [1, 2, 3] taught us that masses can arise from a background field. If that field evolves, masses can drift or oscillate, with diverse phenomenological consequences. The relevant timescales range from slow evolution driven by scalars with masses near today’s Hubble parameter H0∼10−33 eVH_0 \sim10^{-33}\,\mathrm{eV} to rapid oscillations of sub-eV scalars.

On cosmological timescales, scalar-dependent dark-matter (DM) masses exchange energy with the scalar, as in coupled-quintessence models [4]. This exchange can slow the dilution of the DM energy density, yielding an older Universe [5] or an effective phantom equation of state for dark energy (DE) [6]. Mass-varying neutrinos coupled to a light scalar may jointly drive cosmic acceleration [7], while electron-mass variation near recombination may help resolve the Hubble tension [8, 9, 10]. Mass variation can also affect DM production and freeze-out [11, 12]. Oscillating scalar backgrounds can also modulate Standard Model (SM) parameters, shifting atomic transition frequencies [13, 14, 15], distorting neutrino oscillations [16, 17, 18], or periodically varying the quark mixing matrix [19].

The microscopic theories underlying these phenomena typically face a common naturalness problem. The interactions that generate mass variation also entail radiative corrections that destabilize the scalar potential, altering its dynamics [20, 21, 22, 23]. A shift symmetry can protect the scalar potential [24], but an appreciable variation of physical particle masses requires significant symmetry breaking that threatens the small curvature of the potential [25].

Mirror [26] and cyclic ZN\mathbb{Z}_N symmetries [23, 27, 28], as well as monodromy [29], can separate these effects. In ZN\mathbb{Z}_N constructions, a pseudo–Nambu–Goldstone boson (pNGB) couples to NN symmetry-related fermions. Individual physical masses vary linearly in the dimensionless shift-symmetry-breaking spurion ϵ\epsilon, whereas the field-dependent radiative potential first appears at order ϵN\epsilon^N. Sufficiently large NN thus permits appreciable individual mass variations alongside a radiatively stable light scalar.

The response of a finite-density background, however, depends on the particle populations. For nonrelativistic particles, the energy density is governed by their population-weighted mean mass, whose scalar dependence can be much weaker than that of each physical mass. Equal populations of the ZN\mathbb{Z}_N partners suppress this collective response to O(ϵN)\mathcal{O}(\epsilon^N), as in the radiative potential. Unequal populations can evade this cancellation, but generating them through sector-dependent reheating couplings explicitly breaks ZN\mathbb{Z}_N and risks destabilizing the scalar potential.

Our key observation is that relic populations need not respect the cyclic symmetry of the microscopic interactions. A homogeneous scalar background, for instance selected during inflation, spontaneously breaks ZN\mathbb{Z}_N and splits the partner masses. Nonrelativistic relic formation can imprint this splitting onto the relative sector abundances, allowing an O(ϵ)\mathcal{O}(\epsilon) collective matter response without loss of radiative protection.

In this Letter, we develop a general framework for radiatively stable mass-varying DM, encompassing slowly rolling and oscillating scalars. We quantify the collective mass modulation and its dependence on conversion decoupling, showing how the same description applies to thermal and nonthermal relics that establish internal equilibrium after production. A companion paper [30] applies this framework to interacting DM and scalar DE, where energy exchange can produce an apparent phantom crossing [6] motivated by DESI observations [31, 32, 33]. We also discuss extensions to mass-varying neutrinos and electrons.

Modeling protected mass variation— To separate individual mass variation from radiative corrections, we consider a complex scalar Φ\Phi coupled to NN Dirac fermions, with canonical kinetic terms and mass interactions

Lψ=−∑k=0N−1(M+ye2πik/NΦ)ψˉkLψkR+h.c.(1)\mathcal{L}_{\psi}=-\sum_{k=0}^{N-1}\left(M+y e^{2\pi i k/N}\Phi\right)\bar{\psi}_{kL}\psi_{kR}+\mathrm{h.c.} \tag*{(1)}

These preserve a cyclic ZN\mathbb{Z}_N symmetry under ψk→ψk+1\psi_k \to\psi_{k+1} and Φ→e2πi/NΦ\Phi\to e^{2\pi i/N}\Phi, with indices understood modulo NN. For M=0M=0, they also preserve a continuous U(1)U(1) symmetry under Φ→eiαΦ\Phi\to e^{i\alpha}\Phi and ψkR→e−iαψkR\psi_{kR} \to e^{-i\alpha}\psi_{kR}. The mass MM is the sole explicit U(1)U(1)-breaking spurion in this sector, formally transforming as M→eiαMM \to e^{i\alpha}M. Field redefinitions allow MM and yy to be taken real and nonnegative.

An invariant scalar potential spontaneously breaks U(1)U(1) at a scale ff. Below the radial-mode mass, Φ=(f/2)eiφ/f\Phi=(f/\sqrt{2})e^{i\varphi/f} with φ\varphi a pNGB. Defining θk≡θ+2πk/N\theta_k \equiv\theta+2\pi k/N, θ≡φ/f\theta\equiv\varphi/f, and m≡yf/2m\equiv yf/\sqrt{2}, the physical fermion masses are

mψ(θk)=M1+ϵ2+2ϵcos⁡θk.(2)m_{\psi}(\theta_k)=M\sqrt{1+\epsilon^2+2\epsilon\cos\theta_k}. \tag*{(2)}

The dimensionless spurion ϵ≡m/M≥0\epsilon\equiv m/M\ge0 controls their scalar dependence. For ϵ≪1\epsilon\ll1, each mass therefore admits an O(ϵ)\mathcal{O}(\epsilon) fractional modulation over field excursions of order unity [34].

The fermion-induced vacuum potential instead sums over the full symmetry-related spectrum. At one loop, the Coleman–Weinberg (CW) potential is [35]

VCW(θ)=−∑kmψ4(θk)16π2[log⁡mψ2(θk)μ2−32].(3)V_{\mathrm{CW}}(\theta)=-\sum_k\frac{m_{\psi}^4(\theta_k)}{16\pi^2}\left[\log\frac{m_{\psi}^2(\theta_k)}{\mu^2}-\frac{3}{2}\right]. \tag*{(3)}

For N>2N>2, its scalar-dependent part is finite and independent of the renormalization scale μ\mu [27, 36]. The cyclic symmetry cancels all harmonics whose orders are not multiples of NN [28], leaving, at leading order in ϵ\epsilon and up to an additive constant,

VCW(θ)≃(−ϵ)NM44π2(N−1)(N−2)cos⁡(Nθ),(4)V_{\mathrm{CW}}(\theta)\simeq\frac{(-\epsilon)^N M^4}{4\pi^2(N-1)(N-2)}\cos(N\theta), \tag*{(4)}

whose curvature correction δmφ2(θ)≡f−2∂θ2VCW\delta m_\varphi^2(\theta)\equiv f^{-2}\partial_\theta^2V_{\mathrm{CW}} begins at O(ϵN)\mathcal{O}(\epsilon^N) and is bounded by

∣δmφ2(θ)∣≲N2(N−1)(N−2)ϵNM44π2f2.(5)\left|\delta m_\varphi^2(\theta)\right|\lesssim\frac{N^2}{(N-1)(N-2)}\frac{\epsilon^N M^4}{4\pi^2f^2}. \tag*{(5)}

Radiative stability alone does not ensure that the available mass modulation is dynamically realized. Oscillations confined to one CW minimum explore only δθ≲O(1/N)\delta\theta\lesssim\mathcal{O}(1/N). An additional potential with a longer field-space period can support larger excursions. Such a contribution can arise from an anomalous breaking of the global ZNZ_N symmetry by a confining gauge sector, generating, for example,

Vnp(θ)=χtop[1−cos⁡(θ−δ)].(6)V_{\mathrm{np}}(\theta)=\chi_{\mathrm{top}}\left[1-\cos(\theta-\delta)\right]. \tag*{(6)}

Here χtop\chi_{\mathrm{top}} is the topological susceptibility of the confining sector in this cosine approximation, and δ\delta is a phase offset relative to the CW contribution. The topological interactions underlying Equation (6) preserve ZNZ_N in perturbation theory. Provided the new sector introduces no other explicit breaking, the fermion-loop cancellation remains intact, as corrections that violate ZNZ_N must vanish with this nonperturbative breaking.

The full vacuum potential is V0=VCW+VnpV_0=V_{\mathrm{CW}}+V_{\mathrm{np}}, with the nonperturbative term setting the bare mass squared mφ2≡χtop/f2m_\varphi^2\equiv\chi_{\mathrm{top}}/f^2 at its minimum. Radiative control requires ∣δmφ2(θ)∣≪mφ2|\delta m_\varphi^2(\theta)|\ll m_\varphi^2 throughout field space. For N≫1N\gg1, the CW correction becomes parametrically comparable to the bare mass squared near

N≃2log⁡[M2/(2πfmφ)]log⁡(1/ϵ).(7)N\simeq\frac{2\log\left[M^2/(2\pi f m_\varphi)\right]}{\log(1/\epsilon)}. \tag*{(7)}

with each additional sector reducing the correction by approximately a factor ϵ\epsilon. As an example, M=100 GeVM=100\ \mathrm{GeV}, f=1010 GeVf=10^{10}\ \mathrm{GeV}, ϵ=0.1\epsilon=0.1, and mφ=10−20 eVm_\varphi=10^{-20}\ \mathrm{eV} give a maximum correction of about 3%3\% for N=46N=46. With this hierarchy, the construction allows O(ϵ)\mathcal{O}(\epsilon) individual mass variation from initial displacement of order unity. The collective matter response further depends on the sector populations, as we discuss next.

Collective matter response

In a dilute background of nonrelativistic ψ\psi fermions, the effective scalar potential is,

Veff(θ)=V0(θ)+mψeff(θ)nψ,(8)V_{\mathrm{eff}}(\theta)=V_0(\theta)+m_{\psi}^{\mathrm{eff}}(\theta)n_\psi, \tag*{(8)}

where nψ=∑knkn_\psi=\sum_k n_k is the total number density across the NN sectors, and

mψeff(θ)=∑k=0N−1ηkmψ(θk),ηk=nknψ,(9)m_{\psi}^{\mathrm{eff}}(\theta)=\sum_{k=0}^{N-1}\eta_km_\psi(\theta_k),\qquad\eta_k=\frac{n_k}{n_\psi}, \tag*{(9)}

is the population-averaged effective fermion mass. The effective potential governs scalar evolution through its derivative at fixed sector number densities. In a homogeneous Universe, the matter contribution can hold the scalar near a density-selected minimum, while its dilution can allow the vacuum potential to drive further evolution and change the effective DM mass.

We characterize the collective matter response entering this dynamics by the rms fractional modulation

Δψ≡⟨mψeff 2⟩θ−⟨mψeff⟩θ2⟨mψeff⟩θ,(10)\Delta_\psi\equiv\frac{\sqrt{\langle m_{\psi}^{\mathrm{eff}\,2}\rangle_\theta-\langle m_{\psi}^{\mathrm{eff}}\rangle_\theta^2}}{\langle m_{\psi}^{\mathrm{eff}}\rangle_\theta}, \tag*{(10)}

where ⟨X⟩θ=(2π)−1∫02πdθX(θ)\langle X\rangle_\theta=(2\pi)^{-1}\int_0^{2\pi}d\theta X(\theta) denotes an angular average at fixed relative populations. Thus, Δψ\Delta_\psi measures the mass modulation available across field space, while the variation realized in time depends on the scalar trajectory. Although each physical fermion mass mψ(θk)m_\psi(\theta_k) depends on the value of the pNGB field at order ϵ\epsilon (see Equation (2)), their average need not exhibit a comparable variation. Indeed, for a ZNZ_N-symmetric background with universal ηk=1/N\eta_k=1/N, a shift θ→θ+2π/N\theta\to\theta+2\pi/N merely permutes the equally weighted sectors and only harmonics that are multiple of NθN\theta survive, yielding, mψeff(θ)=N−1∑kmψ(θk)≡mψsym(θ)m_\psi^{\mathrm{eff}}(\theta)=N^{-1}\sum_km_\psi(\theta_k)\equiv m_\psi^{\mathrm{sym}}(\theta) with

mψsym(θ)≃M[F(ϵ)−(−ϵ)NπN3/2cos⁡(Nθ)].(11)m_\psi^{\mathrm{sym}}(\theta)\simeq M\left[F(\epsilon)-\frac{(-\epsilon)^N}{\sqrt{\pi N^{3/2}}}\cos(N\theta)\right]. \tag*{(11)}

where F(ϵ)≃1+O(ϵ2)F(\epsilon) \simeq1 + O(\epsilon^2) is independent of θ\theta and assuming the large NN limit. Thus, similarly to the CW potential, the matter response is strongly suppressed, with Δψsym≃ϵN/2πN3\Delta_{\psi}^{\mathrm{sym}} \simeq\epsilon^{N}/\sqrt{2\pi N^3}, despite the variation of the individual fermion masses.

Evading this symmetry suppression requires a background that populates the sectors unequally. For general population asymmetries δηk≡ηk−1/N\delta\eta_k \equiv\eta_k - 1/N,

mψeff(θ)=mψsym(θ)+M∑ℓ∉NZFℓ(ϵ)∣Aℓ∣cos⁡(ℓθ+ζℓ),(12)m_{\psi}^{\mathrm{eff}}(\theta) = m_{\psi}^{\mathrm{sym}}(\theta) + M \sum_{\ell\notin N\mathbb{Z}} F_\ell(\epsilon)|A_\ell| \cos(\ell\theta+ \zeta_\ell), \tag*{(12)}

where the sum runs over integers ℓ\ell that are not multiples of NN,

Aℓ≡∑k=0N−1δηke2πiℓk/N,ζℓ≡arg⁡Aℓ(13)A_\ell\equiv\sum_{k=0}^{N-1} \delta\eta_k e^{2\pi i\ell k/N}, \qquad\zeta_\ell\equiv\arg A_\ell \tag*{(13)}

and Fℓ(ϵ)≡2∑j≥0C(1/2,j+ℓ)C(1/2,j)ϵ2j+ℓF_\ell(\epsilon) \equiv2\sum_{j\geq0} C(1/2,j+\ell)C(1/2,j)\epsilon^{2j+\ell} with C(n,k)C(n,k) denoting binomial coefficients. The AℓA_\ell are Fourier moments of the population imbalance around the sector circle, measuring the incomplete cancellation of each mass harmonic. For illustration, consider a mixture of a symmetric population and a fraction pp concentrated in sector k⋆k_\star, δηk=p(δkk⋆−1/N)\delta\eta_k = p(\delta_{kk_\star} - 1/N) where 0≤p≤10 \leq p \leq1. Then Aℓ=pe2πiℓk⋆/NA_\ell= p e^{2\pi i\ell k_\star/N} for ℓ∉NZ\ell\notin N\mathbb{Z}, giving

mψeff(θ)≃mψsym(θ)+ϵpMcos⁡(θ+2πk⋆N),(14)m_{\psi}^{\mathrm{eff}}(\theta) \simeq m_{\psi}^{\mathrm{sym}}(\theta) + \epsilon pM \cos\left(\theta+ \frac{2\pi k_\star}{N}\right), \tag*{(14)}

and Δψ≃[(Δψsym)2+ϵ2p2/2]1/2\Delta_\psi\simeq[(\Delta_{\psi}^{\mathrm{sym}})^2 + \epsilon^2p^2/2]^{1/2}. Thus even a small population imbalance p≫(2/ϵ)Δψsymp \gg(\sqrt{2}/\epsilon)\Delta_{\psi}^{\mathrm{sym}} can overcome the symmetry suppression, and restores an appreciable O(ϵp)O(\epsilon p) mass modulation for moderate pp.

Asymmetric reheating can generate unequal fermion populations. If the DM states possess efficient interactions with a common SM bath, those can repopulate depleted ZN\mathbb{Z}_N partners and wash out the excess. Preserving it generally requires replicating the SM, coupling each partner to its own bath, and suppressing intersector energy transfer. Constraints on extra radiation then favor strongly asymmetric reheating, approaching the single-sector limit p→1p \to1. Non-universal reheating couplings, however, explicitly break ZN\mathbb{Z}_N. The resulting sector-dependent radiative corrections can spoil the cancellation of lower CW harmonics and reintroduce ultraviolet sensitivity in the scalar potential. The transmission of this breaking to the fermion-scalar sector must therefore be tightly controlled to preserve the shallow potential.

Dynamical population imbalance— The required population asymmetry can instead arise dynamically from nonrelativistic relic production, with all DM partners

coupled symmetrically to a single SM bath. We assume that θ\theta remains approximately frozen during population formation. The ZN\mathbb{Z}_N cancellation suppresses leading thermal forces when the partners are relativistic and equally populated, while Hubble friction can suppress evolution driven by the nonperturbative potential provided mφ≪Hm_\varphi\ll H during this epoch. At a generic scalar background value θi\theta_i, for instance randomly selected by inflation, their different masses allow production and decoupling processes to distinguish them as they become nonrelativistic. Unequal relic populations can therefore emerge without introducing ZN\mathbb{Z}_N-breaking interactions or additional SM copies, avoiding the reheating-induced corrections discussed above.

The population imbalance generally depends on the specific relic-production mechanism. This dependence can nevertheless be erased if number-conserving intersector conversions establish internal thermal equilibrium and remain efficient after the total abundance is fixed and into the nonrelativistic regime. All partners then share a common temperature and chemical potential. The latter cancels from their relative abundances, giving [37]

ηk=ηkeq≡mk3/2e−mk/T′∑jmj3/2e−mj/T′,(15)\eta_k = \eta_k^{\mathrm{eq}} \equiv\frac{m_k^{3/2}e^{-m_k/T'}}{\sum_j m_j^{3/2}e^{-m_j/T'}}, \tag*{(15)}

where mk≡mψ(θi+2πk/N)m_k \equiv m_\psi(\theta_i + 2\pi k/N), even when the total density differs substantially from its zero-chemical-potential equilibrium value. Here T′T' is the fermion temperature, equal to the SM temperature while kinetic equilibrium with the bath holds. The decoupling of conversions at a temperature Td′T'_d in the nonrelativistic regime freezes a mass-dependent population bias, provided subsequent redistribution and species-dependent depletion are negligible. The equilibrium form of the frozen fractions is therefore independent of the initial production mechanism, instead it is set by the fermion temperature and scalar background at conversion decoupling. If these conditions fail, the abundances must instead be obtained from the coupled Boltzmann equations, as illustrated in studies of coannihilation without chemical equilibrium [38].

For thermal freeze-out [39, 40], the same interactions that allow the partners to annihilate into common SM bath particles can also mediate intersector conversions. These conversions can remain efficient after annihilation freeze-out, maintaining internal equilibrium while the total comoving abundance is conserved. For nonthermal production, such as freeze-in [41, 42], sufficiently strong intersector interactions, for example mediated by a common dark radiation, can thermalize the fermions and equilibrate their relative populations after production, even if their interactions with the SM never establish thermal equilibrium [43].

The mass modulation amplitude retains a memory of the scalar background at conversion decoupling. For small ϵ\epsilon and large NN, expanding the equilibrium weights in Equation (15) while retaining their exponential bias gives

mψeff(θ)≃mψsym(θ)−ϵMI1(q)I0(q)cos⁡(θ−θi),(16)m_{\psi}^{\mathrm{eff}}(\theta) \simeq m_{\psi}^{\mathrm{sym}}(\theta)-\epsilon M\frac{I_1(q)}{I_0(q)}\cos(\theta-\theta_i), \tag*{(16)}

where I0I_0 and I1I_1 are modified Bessel functions of the first kind, q≡ϵ(xd−3/2)q \equiv\epsilon(x_d-3/2) and xd≡min⁡[mk]/Td′x_d \equiv\min[m_k]/T'_d. The full derivation is given in the End Matter. For q≪1q \ll1, the modulation scales as Δψ∝ϵ2(xd−3/2)\Delta_\psi\propto\epsilon^2(x_d-3/2), while for q≳1q \gtrsim1, it becomes of order ϵ\epsilon.

Figure 1 shows the amplitude of the effective-mass dependence on the scalar across a broad range of ultralight scalar masses compatible with radiative control, ∣δmϕ2(θ)∣<mϕ2|\delta m_\phi^2(\theta)| < m_\phi^2, for several values of NN. The modulation and minimal scalar mass are linked through their common dependence on ϵ\epsilon, with the latter set by the CW potential. The physical scalar mass, set predominantly by the nonperturbative potential, can lie above it. Increasing NN allows substantial modulation to coexist with an extremely light, radiatively protected scalar. As conversion decoupling occurs deeper in the nonrelativistic regime, the thermal curves approach the single-sector population scenario, as for ϵd≫1\epsilon_d \gg1, Boltzmann suppression strongly favors the lighter partners. This shows that the modulation associated with maximally asymmetric reheating can be dynamically reproduced without replicating the SM or introducing additional ZN\mathbb{Z}_N-breaking interactions.

Figure 1. Fractional fermion mass modulation Δψ≃ϵ∣A1∣/2\Delta_\psi\simeq\epsilon|A_1|/\sqrt{2} versus the minimal scalar mass mφm_\varphi compatible with radiative control for the indicated values of NN. Solid curves show the one-sector population (p=1p=1) from asymmetric reheating; other curves show the asymmetric population resulting from thermalization for several values of xd≃min⁡[mk]/Td′x_d \simeq\min[m_k]/T'_d. The mφm_\varphi axis scales as (M2/f)(M^2/f).

Conclusions and outlook— Therefore an appreciable scalar dependence of the population-averaged fermion mass can coexist with a radiatively protected ultralight scalar. The nonperturbative potential sets the scalar mass and permits large field excursions, while the cyclic symmetry suppresses the fermion-induced radiative correction. Unequal relic populations restore the collective matter response without introducing perturbative ZN\mathbb{Z}_N-breaking interactions or SM replicas. If number-conserving intersector conversions maintain internal equilibrium after the total abundance is fixed, the relative populations are set by the fermion temperature and scalar background at conversion decoupling.

The framework accommodates both thermal and nonthermal relic production, provided intersector conversions remain efficient after production and the scalar is approximately stationary while the relative populations are established. Once these populations freeze, the effective DM mass responds to subsequent scalar rolling or oscillations. Two regimes remain to be explored for either production mechanism. If the scalar oscillates during production, the evolving mass spectrum can affect the population bias. If conversions are inefficient during production and do not subsequently establish internal equilibrium, the final relative abundances retain a dependence on the production processes themselves.

Extensions to electrons and active neutrinos are challenging. First, additional chiral fermions must be introduced to cancel anomalies induced by the replicated chiral SM leptons. Moreover, light partners carrying SM electroweak charges face stringent bounds, for instance from precision measurement of the ZZ width. Placing these partners in separate SM copies avoids these bounds but requires an asymmetric thermal history to suppress extra radiation. Additional explicit ZN\mathbb{Z}_N breaking introduced through sector-dependent reheating must then be sufficiently isolated from the fermion–scalar sector to preserve radiative protection.

Acknowledgment— We thank ChatGPT for useful comments on the manuscript. S.J.L. is partly supported by the Samsung Science and Technology Foundation under Project Number SSTF-BA2106-06. BY was partly supported by the NSF grant PHY-2309456.

References

Here we derive the effective mass for thermally distributed partner populations and show how the large-NN limit simplifies its dependence on the scalar field. Working at leading order in ϵ\epsilon, we retain the full ϵxd\epsilon x_d dependence of the Boltzmann factor while neglecting corrections of order ϵ2xd≪1\epsilon^2 x_d \ll1. The abundance fractions in Equation (15) then become

ηkeq≃1NZN(q,θi)e−qcos⁡(θi+2πk/N),(17)\eta_k^{\rm eq} \simeq\frac{1}{N Z_N(q,\theta_i)} e^{-q\cos(\theta_i+2\pi k/N)}, \tag*{(17)}

where

ZN(q,θi)≡1N∑k=0N−1e−qcos⁡(θi+2πk/N).(18)Z_N(q,\theta_i) \equiv\frac{1}{N}\sum_{k=0}^{N-1} e^{-q\cos(\theta_i+2\pi k/N)}. \tag*{(18)}

After conversion decoupling, these frozen populations give

mψeff(θ)≃mψsym(θ)−ϵM[∂log⁡ZN∂qcos⁡(θ−θi)+1q∂log⁡ZN∂θisin⁡(θ−θi)].(19)m_\psi^{\rm eff}(\theta) \simeq m_\psi^{\rm sym}(\theta)-\epsilon M\left[\frac{\partial\log Z_N}{\partial q}\cos(\theta-\theta_i)+\frac{1}{q}\frac{\partial\log Z_N}{\partial\theta_i}\sin(\theta-\theta_i)\right]. \tag*{(19)}

Expanding the Boltzmann factor in modified Bessel functions and performing the root-of-unity sum yields

ZN(q,θi)=I0(q)+2∑ℓ≥1(−1)NℓINℓ(q)cos⁡(Nℓθi).(20)Z_N(q,\theta_i)=I_0(q)+2\sum_{\ell\geq1}(-1)^{N\ell}I_{N\ell}(q)\cos(N\ell\theta_i). \tag*{(20)}

These higher-order terms encode the dependence of the frozen populations on the formation angle θi\theta_i. All harmonics of the evolving field θ\theta whose orders are multiples of NN remain confined to mψsym(θ)m_\psi^{\rm sym}(\theta) and cancel exactly in the population-dependent difference. At large NN limit and fixed qq, INℓ(q)∼(q/2)Nℓ/(Nℓ)!→0I_{N\ell}(q)\sim(q/2)^{N\ell}/(N\ell)! \to0, so that ZN→I0(q)Z_N\to I_0(q) and

mψeff(θ)≃mψsym(θ)−ϵMI1(q)I0(q)cos⁡(θ−θi).(21)m_\psi^{\rm eff}(\theta)\simeq m_\psi^{\rm sym}(\theta)-\epsilon M\frac{I_1(q)}{I_0(q)}\cos(\theta-\theta_i). \tag*{(21)}

References

  1. [1]F. Englert and R. Brout, Phys. Rev. Lett. 13, 321 (1964).
  2. [2]P. W. Higgs, Phys. Rev. Lett. 13, 508 (1964).
  3. [3]G. S. Guralnik, C. R. Hagen, and T. W. B. Kibble, Phys. Rev. Lett. 13, 585 (1964).
  4. [4]L. Amendola, Phys. Rev. D 62, 043511 (2000), arXiv:astro-ph/9908023DOI
  5. [5]G. W. Anderson and S. M. Carroll, in 1st International Conference on Particle Physics and the Early Universe (1997) pp. 227–229, arXiv:astro-ph/9711288.
  6. [6]S. Das, P. S. Corasaniti, and J. K. Khoury, Phys. Rev. D 73, 083509 (2006), arXiv:astro-ph/0510628.DOI
  7. [7]R. Fardon, A. E. Nelson, and N. Weiner, JCAP 2004 (10), 005, arXiv:astro-ph/0309800.DOI
  8. [8]M. Baryakhtar, O. Simon, and Z. J. Weiner, Phys. Rev. D 110, 083505 (2024), arXiv:2405.10358 [astro-ph.CO].DOI
  9. [9]N. Schöneberg, V. Poulin, A. G. Ferrari, F. Finelli, J. Lesgourgues, L. Morelli, M. R. Mosbech, R. K. Sharma, and T. Simon, The H₀ World Cup. I. Summary of the baseline group stage results (2026), arXiv:2607.13282 [astroph-CO].arxiv.org/abs/2607.13282
  10. [10]N. Schöneberg, V. Poulin, A. G. Ferrari, F. Finelli, J. Lesgourgues, L. Morelli, M. R. Mosbech, R. K. Sharma, and T. Simon, The H₀ World cup. II. A comprehensive competition between proposed Hubble tension solutions (2026), arXiv:2607.13283 [astro-ph.CO].arxiv.org/abs/2607.13283
  11. [11]L. Heurtier and H. Partouche, Phys. Rev. D 101, 043527 (2020), arXiv:1912.02828 [hep-ph].
  12. [12]S. Ferrante, M. Perelstein, and B. Yu, Phys. Rev. Lett. 136, 201804 (2026), arXiv:2511.16731 [hep-ph].DOI
  13. [13]A. Arvanitaki, J. Huang, and K. Van Tilburg, Phys. Rev. D 91, 015015 (2015), arXiv:1405.2925 [hep-ph].DOI
  14. [14]Y. V. Stadnik and V. V. Flambaum, Phys. Rev. Lett. 115, 201301 (2015), arXiv:1503.08540 [astro-ph.CO].DOI
  15. [15]A. Banerjee, G. Perez, M. Safronova, I. Savaroy, and A. Shalit, JHEP 2023 (10), 042, arXiv:2211.05174 [hep-ph].DOI
  16. [16]G. Krnjaic, P. A. N. Machado, and L. Necib, Phys. Rev. D 97, 075017 (2018), arXiv:1705.06740 [hep-ph].DOI
  17. [17]J. Liao, D. Marfatia, and K. Whisnant, JHEP 2018 (04), 136, arXiv:1803.01773 [hep-ph].DOI
  18. [18]M. Losada, Y. Nir, G. Perez, I. Savaroy, and Y. Shpilman, JHEP 2023 (03), 032, arXiv:2205.09769 [hep-ph].DOI
  19. [19]M. Dine, G. Perez, W. Ratzinger, and I. Savaroy, Phys. Rev. D 111, 015049 (2025), arXiv:2405.06744 [hep-ph].DOI
  20. [20]M. Doran and J. A. Peacock, Phys. Rev. D 66, 043519 (2002), arXiv:astro-ph/0203018.DOI
  21. [21]G. D’Amico, T. Hamill, and N. Kaloper, Phys. Rev. D 94, 103526 (2016), arXiv:1605.00996 [hep-th].DOI
  22. [22]M. C. D. Marsh, Phys. Rev. Lett. 118, 011302 (2017), arXiv:1606.01538 [astro-ph.CO].DOI
  23. [23]D. Brzeminski, Z. Chacko, A. Dev, and A. Hook, Phys. Rev. D 104, 075019 (2021), arXiv:2012.02787 [hep-ph].DOI
  24. [24]J. Goldstone, A. Salam, and S. Weinberg, Phys. Rev. 127, 965 (1962).
  25. [25]Super-renormalizable Higgs couplings provide a technically natural alternative [? ? ].
  26. [26]C. Delaunay, M. Geller, Z. Heller-Algazi, G. Perez, and K. Springmann, Phys. Rev. D 113, 035011 (2026), arXiv:2507.12514 [hep-ph].arxiv.org/abs/2507.12514
  27. [27]J. A. Frieman, C. T. Hill, A. Stebbins, and I. Waga, Phys. Rev. Lett. 75, 2077 (1995), arXiv:astro-ph/9505060.DOI
  28. [28]A. Hook, Phys. Rev. Lett. 120, 261802 (2018), arXiv:1802.10093 [hep-ph].DOI
  29. [29]H. Bagherian, M. Reece, L.-T. Wang, and H. Xiao, A Natural Dark Matter-Dark Energy Interaction from Axion-Monopole Dynamics (2026), arXiv:2610.27059 [hep-ph].arxiv.org/abs/2610.27059
  30. [30]C. Delaunay, S. J. Lee, Y. Yin, and B. Yu, Natural Phantom Crossing from Axion-WIMP Interactions (2026), arXiv:2607.2872 [hep-ph].arxiv.org/abs/2607.28721
  31. [31]A. G. Adame et al. (DESI), JCAP 2025 (02), 021, arXiv:2404.03002 [astro-ph.CO].DOI
  32. [32]K. Lodha et al. (DESI), Phys. Rev. D 112, 083511 (2025), arXiv:2503.14743 [astro-ph.CO].arxiv.org/abs/2503.14743
  33. [33]M. Abdul Karim et al. (DESI), Phys. Rev. D 112, 083515 (2025), arXiv:2503.14738 [astro-ph.CO].
  34. [34]Interestingly, the cyclic fermion sector also admits a geometric realization with φ as a Wilson-line scalar in a deconstructed five-dimensional U(1) gauge theory [? ? ? ], which will be developed elsewhere [? ].
  35. [35]S. R. Coleman and E. J. Weinberg, Phys. Rev. D 7, 1888 (1973).
  36. [36]J. A. Frieman, C. T. Hill, and R. Watkins, Phys. Rev. D 46, 1226 (1992).
  37. [37]K. Griest and D. Seckel, Phys. Rev. D 43, 3191 (1991).
  38. [38]M. Garny, J. Heisig, B. Lulf, and S. Vogl, Phys. Rev. D 96, 103521 (2017), arXiv:1705.02992 [hep-ph].
  39. [39]B. W. Lee and S. Weinberg, Phys. Rev. Lett. 39, 165 (1977).
  40. [40]G. Steigman and M. S. Turner, Nucl. Phys. B 253, 375 (1985).
  41. [41]J. McDonald, Phys. Rev. Lett. 88, 091304 (2002), arXiv:hep-ph/0106249.DOI
  42. [42]L. J. Hall, K. Jedamzik, J. March-Russell, and S. M. West, JHEP 2010 (03), 080, arXiv:0911.1120 [hep-ph].DOI
  43. [43]X. Chu, T. Hambye, and M. H. G. Tytgat, JCAP 2012 (05), 034, arXiv:1112.0493 [hep-ph].DOI

Paper details

vlnc:W_5bp5pzac
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 12

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents