betaFeedback

WorksPhysics and Astronomy

Exothermic Higgs portal: from LUX-ZEPLIN recoil to Gamma Rays

Abstract

We show that a minimal extension of the Standard Model (SM) by a single complex scalar ϕϕ provides a viable inelastic exothermic, thermal relic, interpretation of the recently reported nuclear-recoil candidate from LUX-ZEPLIN (LZ). The mass terms and Higgs-portal interactions split ϕϕ into two nearly degenerate real scalars, ϕ1,2ϕ_{1,2}. For mass-splitting of a few hundred keV, ϕ2ϕ_2 is cosmologically long-lived and the two states have comparable relic abundances. Taking ϕ1,2ϕ_{1,2} masses to be ∼100 \sim100\,GeV, with off-diagonal Higgs coupling fixed by relic abundance, the exothermic down-scattering yields an order-one event expectation for the LZ exposure, without relying on the high-velocity tail of the dark matter velocity distribution required in endothermic scenarios. This scenario can be tested using different nuclear targets and with indirect dark matter searches: ϕ2→ϕ1γγϕ_2\toϕ_1γγ decays give a broad sub-MeV photon continuum that can be probed by INTEGRAL/SPI and COSI; ϕ1ϕ2ϕ_1ϕ_2 annihilation produces a GeV-scale continuum, extending up to ∼\sim100 GeV, that is probed by Fermi-LAT observations of dwarf galaxies; current CMB constraints are satisfied and annihilation among dark matter particles captured in the Sun is strongly suppressed. Next-generation experiments could probe correlated scattering, decay, and annihilation signatures of this thermal Higgs-portal scenario.

Introduction

The LUX-ZEPLIN (LZ) Collaboration recently reported a nuclear-recoil candidate with energy ER=248±23stat±23sysE_R = 248 \pm23_{\mathrm{stat}} \pm23_{\mathrm{sys}} keV in a 2.84 tonne-year exposure [1]. Relative to the background-only hypothesis, the event reaches a maximum local significance of 3.4σ3.4\sigma among the interactions tested by LZ, and a global significance of 2.6σ2.6\sigma once the look-elsewhere effect is included [1]. Such a single high recoil energy is difficult to explain with elastic scattering of halo-bound non-relativistic dark matter (DM), since this predicts a larger event rate at lower recoil energies. Endothermic inelastic scattering, χ1A→χ2A\chi_1 A \to\chi_2 A, on the other hand, imposes an additional kinematic threshold which suppresses the low-energy recoils and thus shifts the recoil spectrum toward higher energies [2], in agreement with the observed LZ event. That is, for χ1A→χ2A\chi_1 A \to\chi_2 A scattering, the lighter DM state χ1\chi_1 needs to have enough kinetic energy to produce the heavier state χ2\chi_2. For mass splitting of a few hundred keV this cuts out the bulk of the χ1\chi_1 velocity distribution, resulting in a visibly harder recoil spectrum.

A particularly well-motivated supersymmetric realization of the endothermic inelastic scenario is the Higgsino, whose two neutral states can form a nearly degenerate pseudo-Dirac pair and naturally give rise to inelastic scattering. This possibility has recently been studied in the context of the observed LZ high-recoil event and in related electroweak constructions [3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21]. Similar inelastic kinematics has also been explored in singlet-doublet and inert-doublet sectors, neutrino models, dark-photon and other U(1)U(1) extensions, axion-like portals, transition-dipole models, neutrino-mass constructions, freeze-in scenarios, and effective descriptions [22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52].

A broad range of alternative interpretations has been proposed in response to the LZ candidate. These include elastic-scattering scenarios with nonstandard momentum or spin dependence, as well as scalar, axion, magnetic, and extended-Higgs interactions [53, 54, 55, 56, 57]. Other proposals invoked obscure dark matter or nonstandard incident populations [58, 59, 60, 61, 62, 63, 64, 65], dark-matter absorption [66], neutrino-disappearance processes [67, 68, 69], and neutrino-induced nuclear recoils [70, 71, 72]. Further possibilities involving composite dark sectors, dark QCD, extra dimensions, or modified gravity have been explored in Refs. [73, 74, 75, 76].

Here we instead consider exothermic inelastic scattering [77, 78, 79], in which a population of the heavier state that survives to the present day down-scatters as χ2A→χ1A\chi_2 A \to\chi_1 A. Exothermic interpretations of the LZ candidate have been proposed in Refs. [37, 45, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90]. A particularly simple realization of inelastic DM is obtained by extending the Standard Model (SM) with a complex scalar singlet that interacts with the visible sector through the doublet Higgs (HH) portal. Small symmetry-breaking terms split the complex field into two nearly degenerate real mass eigenstates, ϕ1\phi_1 and ϕ2\phi_2 (playing the role of χ1\chi_1 and χ2\chi_2 above), for which the leading couplings to the visible sector can be due to off-diagonal Higgs interactions [91, 92, 93, 94, 95, 96, 97, 98, 99, 100],

Lint=−gϕ1ϕ2(H†H−vh22).(1)\mathcal{L}_{\mathrm{int}} = -g\phi_1\phi_2\left(H^\dagger H - \frac{v_h^2}{2}\right). \tag*{(1)}

After electroweak symmetry breaking, this generates the hϕ1ϕ2h\phi_1\phi_2 coupling responsible for inelastic scattering. The same off-diagonal interaction also sets the relic abundances of ϕ1\phi_1 and ϕ2\phi_2 via freeze-out of coannihilations, ϕ1ϕ2→SM\phi_1\phi_2 \to\mathrm{SM}, and inelastic scatterings, ϕ1+SM↔ϕ2+SM\phi_1 + \mathrm{SM} \leftrightarrow\phi_2 + \mathrm{SM}.

For small enough mass splitting Δm\Delta m, ϕ2\phi_2 can be stable on cosmological timescales. For instance, for Δm<2me\Delta m < 2m_e, the decay ϕ2→ϕ1e+e−\phi_2 \to\phi_1 e^+e^- is kinematically forbidden, and the dominant decay channel is ϕ2→ϕ1γγ\phi_2 \to\phi_1\gamma\gamma, induced via emission of an off-shell Higgs boson. Since the available @body ends_mid=1 energy is shared between the two photons, this decay produces a broad sub-MeV spectrum rather than a monochromatic line. Such connections between the LZ high-recoil event and sub-MeV photon signals have been explored in other inelastic dark-matter scenarios [5, 23, 83, 87, 101]. In the minimal model, however, the strength of the diphoton signal is fully predicted, since the decay is induced by the same off-diagonal Higgs interaction as the nuclear recoil. Existing INTEGRAL/SPI data on photon emissions and future observations with the Compton Spectrometer and Imager (COSI) [102, 103] therefore provide a complementary probe of this scenario.

The inelastic interpretations of the LZ event typically face stringent constraints from DM solar capture and subsequent annihilation. For the endothermic scenario, for instance, the larger velocities reached by DM particles inside the Sun make the inelastic endothermic transitions more easily accessible. For Higgsino DM, these larger velocities can allow endothermic scattering on heavy solar nuclei, followed by capture and annihilation predominantly into W+W−W^{+}W^{-} and ZZZZ. The resulting high-energy neutrino flux leads to strong constraints from IceCube and Super-Kamiokande [104, 105, 106, 107, 108] (for extension to broader classes of electroweak multiplets, see Ref. [109]).

The situation is qualitatively different for an off-diagonal scalar Higgs portal. In regions of parameter space in which the exothermic capture ϕ2A→ϕ1A\phi_{2}A \to\phi_{1}A is efficient, while the reverse transition ϕ1A→ϕ2A\phi_{1}A \to\phi_{2}A is suppressed, most of the captured DM solar population is in the form of the lighter state. This then reduces the importance of tree-level mixed annihilation channel ϕ1ϕ2→h∗→SM\phi_{1}\phi_{2} \to h^{*} \to\mathrm{SM}.

Even if at tree level only the off-diagonal portal, Eq. (1), is present, the diagonal interactions are generated at one-loop. These then give rise to elastic scattering and same-state annihilation of the individual dark states. The ϕ1ϕ1\phi_{1}\phi_{1} annihilation in the Sun can be more important than the ϕ1ϕ2\phi_{1}\phi_{2} one, if ϕ1\phi_{1} population dominates. However, as we will show later, even in this case the annihilation rate into neutrinos is still many orders of magnitude below the current experimental solar-neutrino sensitivity, in the region of parameter space relevant for LZ interpretation.

The paper is organized as follows. Section II introduces the inelastic scalar Higgs-portal model and describes the thermal evolution that determines the relic abundance and the surviving excited-state population. In Sec. III, we calculate the exothermic xenon recoil rate, map the thermal-relic parameter space predicting order-one selected events, and then study a representative benchmark and its finite-momentum nuclear response. Section IV examines the correlated signals from radiative decay and present-day annihilation, confronting them with INTEGRAL/SPI, Fermi-LAT dwarf-galaxy, and CMB constraints, and discusses complementary collider and direct-detection probes. In Sec. V, we study solar capture, the nonthermal evolution of the captured populations, and the resulting annihilation rates and their implications for solar-neutrino searches. We conclude in Sec. VI.

The Inelastic Higgs-Portal Dark Matter Model

The inelastic scalar Higgs-portal model [99, 100] extends the SM by a complex scalar singlet ϕ\phi coupled to the Higgs doublet HH. The tree-level scalar potential involving HH and ϕ\phi is given by

V(H,ϕ)=−μh2H†H+λh(H†H)2+m02∣ϕ∣2+12(ρ02ϕ2+ρ0∗2ϕ∗2)+(κ∣ϕ∣2+12(ηϕ2+η∗ϕ∗2))H†H.(2)V(H,\phi)=-\mu_h^2 H^\dagger H+\lambda_h(H^\dagger H)^2+m_0^2|\phi|^2+\frac{1}{2}\left(\rho_0^2\phi^2+\rho_0^{*2}\phi^{*2}\right)+\left(\kappa|\phi|^2+\frac{1}{2}\left(\eta\phi^2+\eta^*\phi^{*2}\right)\right)H^\dagger H. \tag*{(2)}

The parameters ρ02\rho_0^2 and η\eta may in general be complex. If they are zero, the singlet sector possesses a global U(1)U(1) symmetry, which is otherwise broken to a remaining Z2Z_2 symmetry ϕ→−ϕ\phi\to-\phi (and similarly for the real components, ϕ1,2→−ϕ1,2\phi_{1,2} \to-\phi_{1,2}). The lightest Z2Z_2-odd state is therefore stable.

After electroweak symmetry breaking (EWSB),

H=(G+,vh+h+iG02)T,vh=246 GeV.(3)H=\left(G^+,\frac{v_h+h+iG^0}{\sqrt{2}}\right)^T,\qquad v_h=246\ \mathrm{GeV}. \tag*{(3)}

Defining

m2≡m02+κvh22,ρ2≡ρ02+ηvh22,(4)m^2\equiv m_0^2+\frac{\kappa v_h^2}{2},\qquad\rho^2\equiv\rho_0^2+\frac{\eta v_h^2}{2}, \tag*{(4)}

the complex singlet splits into two real mass eigenstates, ϕ1\phi_1 and ϕ2\phi_2, with

mϕ1,22=m2∓∣ρ2∣,(5)m_{\phi_{1,2}}^2=m^2\mp|\rho^2|, \tag*{(5)}

where we use the convention mϕ2>mϕ1m_{\phi_2}>m_{\phi_1} and define Δm≡mϕ2−mϕ1>0\Delta m\equiv m_{\phi_2}-m_{\phi_1}>0. For Δm≪mϕ1,2\Delta m\ll m_{\phi_{1,2}} the mass splitting is given by

Δm≃∣ρ2∣m.(6)\Delta m\simeq\frac{|\rho^2|}{m}. \tag*{(6)}

Here, ρ2\rho^2 is a complex parameter since both ρ02\rho_0^2 and η\eta are complex. These parameters can be decomposed into their real and imaginary parts as follows:

ρ02=ρ0,R2+iρ0,I2,η=ηR+iηI,ρ2=ρR2+iρI2.(7)\rho_0^2=\rho_{0,R}^2+i\rho_{0,I}^2,\qquad\eta=\eta_R+i\eta_I,\qquad\rho^2=\rho_R^2+i\rho_I^2. \tag*{(7)}

The complex field ϕ\phi similarly decomposes in terms of the mass eigenstates as

ϕ=αϕ1+βϕ2,ϕ∗=α∗ϕ1+β∗ϕ2,(8)\phi=\alpha\phi_1+\beta\phi_2,\qquad\phi^*=\alpha^*\phi_1+\beta^*\phi_2, \tag*{(8)}

where α=e−iθ/2\alpha=e^{-i\theta}/\sqrt{2} and β=ie−iθ/2\beta=ie^{-i\theta}/\sqrt{2}. The mixing angle θ\theta diagonalizes the singlet mass matrix via the rotation

U=(cos⁡θ−sin⁡θsin⁡θcos⁡θ),(9)\mathcal{U}=\begin{pmatrix}\cos\theta&-\sin\theta\\ \sin\theta&\cos\theta\end{pmatrix}, \tag*{(9)}

with

cos⁡2θ=−ρR2∣ρ2∣,sin⁡2θ=−ρI2∣ρ2∣.(10)\cos2\theta=-\frac{\rho_R^2}{|\rho^2|},\qquad\sin2\theta=-\frac{\rho_I^2}{|\rho^2|}. \tag*{(10)}

In the mass basis the Higgs-portal interactions are

LϕH=−(f1ϕ12+f2ϕ22+gϕ1ϕ2)(H†H−vh22),(11)\mathcal{L}_{\phi H}=-\left(f_1\phi_1^2+f_2\phi_2^2+g\phi_1\phi_2\right)\left(H^\dagger H-\frac{v_h^2}{2}\right), \tag*{(11)}

where

f1=κ2+12ηRcos⁡2θ+12ηIsin⁡2θ,(12)f_1=\frac{\kappa}{2}+\frac{1}{2}\eta_R\cos2\theta+\frac{1}{2}\eta_I\sin2\theta, \tag*{(12)}
f2=κ2−12ηRcos⁡2θ−12ηIsin⁡2θ,(13)f_2=\frac{\kappa}{2}-\frac{1}{2}\eta_R\cos2\theta-\frac{1}{2}\eta_I\sin2\theta, \tag*{(13)}
g=ηRsin⁡2θ−ηIcos⁡2θ.(14)g=\eta_R\sin2\theta-\eta_I\cos2\theta. \tag*{(14)}

The f1f_1 and f2f_2 control diagonal Higgs interactions, and gg the off-diagonal ones. We focus on the predominantly off-diagonal regime,

∣f1,2∣≪∣g∣,(15)\lvert f_{1,2}\rvert\ll\lvert g\rvert, \tag*{(15)}

which differs qualitatively from the conventional real-singlet Higgs portal [91, 92, 93, 110, 111, 112, 113], which corresponds to the f1≠0f_1\ne0, f2=g=0f_2=g=0 limit.

To simplify the phenomenology we choose sin⁡2θ=0\sin2\theta=0, implying

ρ0,I2=−vh22ηI.(16)\rho_{0,I}^2=-\frac{v_h^2}{2}\eta_I. \tag*{(16)}

The scalar potential in terms of (h,ϕ1,ϕ2)(h,\phi_1,\phi_2) then takes the form

V0=−12μh2h2+14λhh4+12(m02+κ2h2)(ϕ12+ϕ22)+12(ρ0,R2+ηR2h2)(ϕ12−ϕ22)−ηI2(h2−vh2)ϕ1ϕ2+14(λ1ϕ14+λ2ϕ24+λ12ϕ12ϕ22).(17)V_0=-\frac{1}{2}\mu_h^2h^2+\frac{1}{4}\lambda_hh^4+\frac{1}{2}\left(m_0^2+\frac{\kappa}{2}h^2\right)(\phi_1^2+\phi_2^2)+\frac{1}{2}\left(\rho_{0,R}^2+\frac{\eta_R}{2}h^2\right)(\phi_1^2-\phi_2^2)-\frac{\eta_I}{2}(h^2-v_h^2)\phi_1\phi_2+\frac{1}{4}\left(\lambda_1\phi_1^4+\lambda_2\phi_2^4+\lambda_{12}\phi_1^2\phi_2^2\right). \tag*{(17)}

In the remainder of the manuscript we will impose for simplicity f1=f2=λ12=0f_1=f_2=\lambda_{12}=0. For sin⁡2θ=0\sin2\theta=0, the tree-level portal relations reduce to

ηI=−g,κ=ηR=0,(18)\eta_I=-g,\qquad\kappa=\eta_R=0, \tag*{(18)}

and the underlying mass parameters are

m02=12(mϕ12+mϕ22),m_0^2=\frac{1}{2}\left(m_{\phi_1}^2+m_{\phi_2}^2\right),
ρ0,R2=12(mϕ12−mϕ22).(19)\rho_{0,R}^2=\frac{1}{2}\left(m_{\phi_1}^2-m_{\phi_2}^2\right). \tag*{(19)}

In this limit, the tree-level interaction Lagrangian is

Lint=−gϕ1ϕ2(H†H−vh22),(20)\mathcal{L}_{\mathrm{int}}=-g\phi_1\phi_2\left(H^\dagger H-\frac{v_h^2}{2}\right), \tag*{(20)}

which after EWSB generates the hϕ1ϕ2h\phi_1\phi_2 and hhϕ1ϕ2hh\phi_1\phi_2 vertices. The unbroken Z2Z_2 acts identically on the two mass eigenstates, ϕ1,2→−ϕ1,2\phi_{1,2}\to-\phi_{1,2}. As a result, the lighter state ϕ1\phi_1 is stable, whereas ϕ2\phi_2 is a slightly heavier state.

Thermal relic abundances

Note that the dark-sector self-couplings λ1,2\lambda_{1,2} do not affect the relic abundance or any of the signals considered here, beyond ensuring that the potential is bounded from below. With f1=f2=λ12=0f_1=f_2=\lambda_{12}=0, a sufficient tree-level boundedness condition is

λh>0,λ1,2≥0,g2≤2λhλ1λ2.(21)\lambda_h>0,\qquad\lambda_{1,2}\ge0,\qquad g^2\le2\lambda_h\sqrt{\lambda_1\lambda_2}. \tag*{(21)}

The phenomenology we are interested in is fully specified by

{mϕ1,Δm,g}.(22)\{m_{\phi_1},\Delta m,g\}. \tag*{(22)}

A central feature of the off-diagonal Higgs portal is that the same interaction responsible for inelastic nuclear scattering also governs the thermal evolution of the dark sector. That is, in the f1=f2=0f_1=f_2=0 limit, the dominant tree-level annihilation process is

ϕ1ϕ2→h(∗)→SM,(23)\phi_1\phi_2\to h^{(*)}\to\mathrm{SM}, \tag*{(23)}

where “SM” denotes any SM final state. Including also the subdominant diagonal annihilation channels, the two-state system in chemical equilibrium is described by the coannihilation combination of cross sections [114],

⟨σv⟩eff=r12⟨σv⟩11+2r1r2⟨σv⟩12+r22⟨σv⟩22,(24)\langle\sigma v\rangle_{\mathrm{eff}}=r_1^2\langle\sigma v\rangle_{11}+2r_1r_2\langle\sigma v\rangle_{12}+r_2^2\langle\sigma v\rangle_{22}, \tag*{(24)}

where ⟨σv⟩ij\langle\sigma v\rangle_{ij} are the thermal averages of ϕiϕj\phi_i\phi_j annihilation cross sections, while

ri=nin1+n2,(25)r_i=\frac{n_i}{n_1+n_2}, \tag*{(25)}

are the number density fractions, with nin_i (i=1,2i=1,2), the number densities. The factor of two in (24) counts the distinguishable mixed-state initial pairs.

A sub-MeV splitting is smaller than the freeze-out temperature, Δm≪Tf\Delta m\ll T_f, and thus for T∼TfT\sim T_f the ratio of the two equilibrium densities is close to one

n2eqn1eq≃(mϕ2mϕ1)3/2e−Δm/T≃1,(26)\frac{n_2^{\mathrm{eq}}}{n_1^{\mathrm{eq}}}\simeq\left(\frac{m_{\phi_2}}{m_{\phi_1}}\right)^{3/2}e^{-\Delta m/T}\simeq1, \tag*{(26)}

and the two states participate with nearly equal abundances during freeze-out.

For fixed mϕ1m_{\phi_1} and Δm\Delta m, requiring the observed thermal relic abundance determines the off-diagonal portal coupling gg. We determine the total abundance and the relative populations of the two states by solving the coupled Boltzmann equations, including mixed annihilation, elastic scattering with the thermal bath, bath-induced ϕ1↔ϕ2\phi_1\leftrightarrow\phi_2 conversions, number-conserving pair conversions, and the corresponding inverse processes.

Figure 1 shows the coupling required to reproduce the observed relic abundance as a function of the dark-matter mass, for a representative splitting of Δm=600 keV\Delta m=600\ \mathrm{keV}. We consider the mass range 65 GeV≤mϕ1≤150 GeV65\ \mathrm{GeV} \leq m_{\phi_1} \leq150\ \mathrm{GeV}, extending from just above the Higgs resonance (mϕ1≳mh/2m_{\phi_1} \gtrsim m_h/2) to masses where annihilation into electroweak gauge bosons and Higgs bosons becomes important. Moving away from the Higgs resonance, the coupling first increases and reaches a maximum near mϕ1≃73 GeVm_{\phi_1} \simeq73\ \mathrm{GeV}. It then decreases as annihilation into WW-boson pairs, initially through off-shell bosons and subsequently through on-shell production, becomes more efficient. The opening of the ZZZZ and hhhh channels produces further changes at higher masses. Including the precise value of the small mass splitting, Δm≲1 MeV\Delta m \lesssim1\ \mathrm{MeV}, requires a relative adjustment in gg of less than 10−310^{-3}.

Off-diagonal Higgs-portal coupling as a function of dark-matter mass

Figure 1. Off-diagonal Higgs-portal coupling gg required by the observed thermal relic abundance, ΩDMh2=0.120\Omega_{\mathrm{DM}}h^2 = 0.120, as a function of mϕ1m_{\phi_1} for Δm=600 keV\Delta m = 600\ \mathrm{keV}. The vertical dotted lines indicate the W+W−W^+W^-, ZZZZ, and hhhh thresholds.

Exothermic Scattering and the LZ High-Recoil Candidate

An exothermic interpretation of the LZ event requires three ingredients: a mass splitting capable of generating recoils of a few hundred keV, a substantial heavier-state population surviving until the present epoch, and an interaction strength sufficient to yield an observable scattering rate. The relic-density calculation fixes the portal coupling and predicts the surviving heavier-state fraction. We now combine these results with the recoil kinematics and nuclear response to map the expected LZ signal, before analyzing one representative point in detail.

Exothermic recoil kinematics

For a target nucleus of mass mAm_A, the minimum incident velocity required to produce a recoil energy ERE_R is

vmin⁡(ER)=∣mAER/μ2A−Δm∣2mAER,μ2A=mϕ2mAmϕ2+mA.(27)v_{\min}(E_R) = \frac{\left|m_A E_R/\mu_{2A} - \Delta m\right|}{\sqrt{2m_A E_R}}, \qquad\mu_{2A} = \frac{m_{\phi_2}m_A}{m_{\phi_2}+m_A}. \tag*{(27)}

A characteristic feature of exothermic scattering is that even for zero minimal velocity, vmin⁡=0v_{\min}=0, there is a nonzero recoil energy,

ER,A∗=μ2AmAΔm.(28)E_{R,A}^{*} = \frac{\mu_{2A}}{m_A}\Delta m. \tag*{(28)}

For dark-matter masses of order 100 GeV100\ \mathrm{GeV} and splittings of several hundred keV, Eq. (28), the above characteristic recoil scale is in the few-hundred-keV range. Note that ER,A∗E_{R,A}^{*} is comparable but does not in general coincide with the maximum of the observed recoil spectrum: the DM halo velocity distribution, nuclear response, and detector efficiency also enter in this.

Higgs-mediated nuclear scattering

The off-diagonal Higgs interaction mediates coherent scalar scattering from nuclei. At leading nonrelativistic order, the differential cross section for DM scattering on xenon isotope with mass number AA is given by

dσAdER=g2mA8πmϕ22v2(AfNmN)2mh4WMA(q)WMA(0).(29)\frac{\mathrm{d}\sigma_A}{\mathrm{d}E_R} = \frac{g^2m_A}{8\pi m_{\phi_2}^2v^2}\frac{(Af_Nm_N)^2}{m_h^4}\frac{W_M^A(q)}{W_M^A(0)}. \tag*{(29)}

where the recoil four-momentum squared is q2=2mAERq^2=2m_AE_R. The factor (AfNmN)2(Af_Nm_N)^2 is the squared coherent nuclear charge at q=0q=0 and WMA(q)/WMA(0)=F2(q)W_M^A(q)/W_M^A(0)=F^2(q) is the normalized momentum dependence of the nuclear response. Here, mNm_N is the nucleon mass, fNf_N parametrizes the isoscalar Higgs–nucleon scalar coupling, and WMA(q)W_M^A(q) is the isosinglet one-body nuclear response associated with the coherent scalar interaction (we ignore the sub-leading isotriplet contribution). We adopt fN=0.308f_N=0.308 and evaluate the recoil spectrum using the natural xenon isotope mixture and the normalized nuclear responses implemented in WIMpy_NREFT [115, 116, 117, 118, 119]. For the rare isotopes without tabulated responses, we employ the Helm form factor; their contribution to the total rate is below the percent level.

The differential recoil rate per unit detector mass is

dRdER=∑AξAmAn2∫d3v flab(v,t)dσAdERΘ[v−vmin⁡(ER)].(30)\frac{\mathrm{d}R}{\mathrm{d}E_R} = \sum_A \frac{\xi_A}{m_A} n_2 \int\mathrm{d}^3v\, f_{\mathrm{lab}}(\boldsymbol{v},t)\frac{\mathrm{d}\sigma_A}{\mathrm{d}E_R}\Theta[v-v_{\min}(E_R)]. \tag*{(30)}

with ξA\xi_A the isotope mass fraction, v=∣v∣v=|\boldsymbol{v}| the incoming DM velocity, and flab(v,t)f_{\mathrm{lab}}(\boldsymbol{v},t) the normalized laboratory-frame velocity distribution. The step function enforces the minimum-velocity condition for a given recoil energy. We adopt the Standard Halo Model with a truncated Maxwellian velocity distribution and v0=238 km s−1v_0=238\ \mathrm{km}\,\mathrm{s}^{-1}, vesc=544 km s−1v_{\mathrm{esc}}=544\ \mathrm{km}\,\mathrm{s}^{-1}, transformed to the Earth’s frame [120]. The local number density of ϕ2\phi_2 is

n2=r2ρ⊙mˉ,mˉ=r1mϕ1+r2mϕ2.(31)n_2 = \frac{r_2\rho_\odot}{\bar{m}}, \qquad\bar{m}=r_1m_{\phi_1}+r_2m_{\phi_2}. \tag*{(31)}

where ρ⊙\rho_\odot is the local dark-matter mass density and rir_i are the present-day number fractions (r1=1−r2r_1=1-r_2). The local DM density is taken to be ρ⊙=0.3 GeV cm−3\rho_\odot= 0.3\ \mathrm{GeV}\,\mathrm{cm}^{-3} (nuclear-scattering rate scales linearly with it). For the parameter scan in Figure 2, we also include the reverse endothermic process ϕ1A→ϕ2A\phi_1 A \to\phi_2 A, with the corresponding incident density and scattering kinematics. Its contribution is negligible in the region predicting an order-one LZ event yield.

Predicted efficiency-weighted number of xenon recoils in the $(m_{\phi_1},\Delta m)$ plane

Figure 2. Predicted efficiency-weighted number of xenon recoils, NϵN_\epsilon, in the (mϕ1,Δm)(m_{\phi_1},\Delta m) plane for the 2.84 tonne-year LZ exposure. The portal coupling gg is fixed by the thermal relic abundance, ΩDMh2=0.120\Omega_{\mathrm{DM}}h^2 = 0.120, while the present-day state fractions evolve from the thermal evolution. Both exothermic and endothermic scattering contribute to the recoil yield. The black dotted, red dotted, and black dashed contours correspond to Nϵ=0.5N_\epsilon= 0.5, 11, and 22, respectively. The white hatched region is disfavored by INTEGRAL/SPI observations of Galactic photon emission, while the purple hatched region is disfavored by Fermi-LAT observations of 42 dwarf spheroidal galaxies, see Sec. IV. The recoil rates use the Standard Halo Model, the finite-momentum one-body scalar nuclear response, and the LZ nuclear-recoil efficiency.

Thermal parameter space and expected LZ yield

For fixed DM mass and splitting, the relic-density calculation determines the Higgs-portal coupling gg and the present-day population of ϕ2\phi_2. Eq. (30) then predicts the xenon recoil spectrum and rate without independently adjusting the scattering strength.

To estimate the number of selected events, we use the nuclear-recoil efficiency reported by LZ in Supplemental Fig. S2 of Ref. [1]. The digitized efficiency ϵ(ER)\epsilon(E_R) includes trigger, reconstruction, and analysis-selection effects. The corresponding event expectation is

Nϵ=Efsci∫SϵdER ϵ(ER)dRdER,(32)N_\epsilon= \mathcal{E} f_{\mathrm{sci}} \int_{\mathcal{S}_\epsilon} dE_R\,\epsilon(E_R)\frac{dR}{dE_R}, \tag*{(32)}

where E=2.84\mathcal{E} = 2.84 tonne-year is the LZ exposure and Sϵ=[5.4,299.9]\mathcal{S}_\epsilon= [5.4, 299.9] keV is the recoil-energy interval covered by the digitized efficiency. The LZ analysis reports accidental prompt- and delayed-veto fractions of 0.01%0.01\% and 2.86%2.86\%, respectively [1]. We account for these losses through the survival factor fsci=1−0.0001−0.0286=0.9713f_{\mathrm{sci}} = 1 - 0.0001 - 0.0286 = 0.9713, applied separately from the digitized efficiency.

Figure 2 shows an extended region in which the predicted LZ event yield is of order unity, between the Nϵ=0.5N_\epsilon= 0.5 and 22 contours. The shape of this region reflects the mass dependence of the relic-density coupling gg, shown in Figure 1, together with the recoil kinematics and nuclear response. Near mϕ1∼70m_{\phi_1} \sim70–7575 GeV, where the required coupling is largest, larger mass splittings are needed to obtain the same event yield. At larger dark-matter masses, the order-one-event region shifts toward smaller splittings.

The indirect-detection constraints restrict different parts of this parameter space. The radiative decay ϕ2→ϕ1γγ\phi_2 \to\phi_1\gamma\gamma produces a sub-MeV photon continuum, which we constrain using INTEGRAL/SPI measurements of diffuse hard X-ray and soft gamma-ray emission from the inner Galaxy [121, 122]. Complementary constraints arise from Fermi-LAT observations of 42 dwarf spheroidal galaxies, which probe gamma rays from present-day dark-matter annihilation [123]. We use the predicted mass-dependent annihilation spectra in evaluating the dwarf-galaxy constraint.

To define the indirect-detection boundaries in Figure 2, we introduce

RSPI=τSPIrecτγγ,Rdwarf=⟨σv⟩equiv⟨σv⟩limit,(33)\mathcal{R}_{\mathrm{SPI}} = \frac{\tau_{\mathrm{SPI}}^{\mathrm{rec}}}{\tau_{\gamma\gamma}}, \qquad\mathcal{R}_{\mathrm{dwarf}} = \frac{\langle\sigma v\rangle_{\mathrm{equiv}}}{\langle\sigma v\rangle_{\mathrm{limit}}}, \tag*{(33)}

where τSPIrec\tau_{\mathrm{SPI}}^{\mathrm{rec}} is the reconstructed lower lifetime limit from INTEGRAL/SPI, and ⟨σv⟩limit\langle\sigma v\rangle_{\mathrm{limit}} is the observed 95%95\% confidence-level upper limit on the equivalent annihilation cross section obtained from the 42-dwarf likelihood. The SPI constraint is evaluated for the predicted radiative spectrum, while the dwarf-galaxy constraint uses the mass-dependent annihilation spectrum, with only weak dependence on Δm\Delta m. The boundaries correspond to RSPI=1\mathcal{R}_{\mathrm{SPI}} = 1 and Rdwarf=1\mathcal{R}_{\mathrm{dwarf}} = 1, respectively, with R>1\mathcal{R} > 1 defining the corresponding disfavored regions. The SPI constraint disfavors large mass splittings, while the dwarf-galaxy bound disfavors mϕ1≲72m_{\phi_1} \lesssim72 GeV over the scanned range. The calculations underlying these constraints are discussed in Sec. IV.

Representative benchmark and recoil spectrum

For a detailed discussion of the recoil spectrum and the correlated signals, we consider a benchmark point from the region shown in Figure 2,

mϕ1=100 GeV,Δm=600 keV,g≃0.0515.(34)m_{\phi_1} = 100\ \mathrm{GeV}, \qquad\Delta m = 600\ \mathrm{keV}, \qquad g \simeq0.0515. \tag*{(34)}

The value of gg is from the relic-density condition, ΩDMh2=0.120\Omega_{\mathrm{DM}}h^2 = 0.120, which for this benchmark yields nearly equal present-day ϕ1\phi_1 and ϕ2\phi_2 populations,

(r1,r2)t0≃(0.501, 0.499).(35)(r_1,r_2)_{t_0} \simeq(0.501,\,0.499). \tag*{(35)}

The appreciable population of the heavier state provides incident particles for exothermic scattering. A similar near-equal population is obtained throughout the mass and splitting range in 0.5<Nϵ<20.5 < N_{\epsilon} < 2 in Figure 2.

One-loop-induced same-state annihilation does not appreciably modify the thermal picture, so that these contributions can be safely neglected. For instance, the ϕ1ϕ1→W+W−\phi_1\phi_1 \to W^+W^- contribution is about 10−710^{-7} times smaller than due to the mixed annihilation rate.

For this benchmark, the characteristic xenon recoil scale in Eq. (28) is

ER⋆≃270 keV,(36)E_R^\star\simeq270\ \mathrm{keV}, \tag*{(36)}

which lies near the energy of the LZ candidate. Figure 3 shows the full exposure-normalized truth-level recoil energy distribution (i.e., without taking efficiencies into account), obtained from

Predicted xenon recoil spectrum

Figure 3. Predicted xenon recoil spectrum dNtruth/dERdN_{\mathrm{truth}}/dE_R for the 2.84 tonne-year LZ exposure, with mϕ1=100 GeVm_{\phi_1}=100\ \mathrm{GeV}, Δm=600 keV\Delta m=600\ \mathrm{keV}, and g≃0.0515g\simeq0.0515, using the Standard Halo Model and finite-momentum nuclear response. The spectrum is evaluated at truth level, without detector efficiency or energy smearing.

dNtruthdER=EdRdER,E=2.84 tonne yr.(37)\frac{dN_{\mathrm{truth}}}{dE_R} = \mathcal{E}\frac{dR}{dE_R}, \qquad\mathcal{E}=2.84\ \mathrm{tonne\,yr}. \tag*{(37)}

While the vmin⁡=0v_{\min}=0 recoil energy is ER⋆≃270 keVE_R^\star\simeq270\ \mathrm{keV}, the predicted recoil spectrum reaches its maximum near 180 keV. This behavior arises from the interplay between the halo velocity distribution and the finite-momentum nuclear response. In particular, the pronounced diffraction minimum of the adopted xenon response near 270 keV suppresses the rate around ER⋆E_R^\star, shifting the spectral maximum toward lower recoil energies.

At the reported LZ candidate energy of 248 keV, the predicted differential event density is

dNtruthdER∣248 keV≃1.8×10−3 events/keV,(38)\left.\frac{dN_{\mathrm{truth}}}{dE_R}\right|_{248\ \mathrm{keV}} \simeq1.8 \times10^{-3}\ \mathrm{events/keV}, \tag*{(38)}

approximately 7% of the maximum spectral density. Thus, while the mass splitting generates recoils in the relevant energy range, the nuclear response favors a substantial population at lower recoil energies. Integrating the truth-level spectrum over the recoil-energy range 5.4–590 keV gives Ntruth≃2.13N_{\mathrm{truth}}\simeq2.13, see Table I.

QuantityValue
mϕ1m_{\phi_1}100 GeV
Δm\Delta m600 keV
gg0.0515 ≃\simeq 0.05
ΩDMh2\Omega_{\mathrm{DM}}h^20.120
r2(t0)r^2(t_0)≃\simeq 0.5
Ntruth(5.4N_{\mathrm{truth}}(5.4–590 keV)590\ \mathrm{keV})2.13
NϵN_{\epsilon}1.63
Nϵ(100N_{\epsilon}(100–200 keV)200\ \mathrm{keV})1.18
Nϵ(215N_{\epsilon}(215–269.9 keV)269.9\ \mathrm{keV})0.216
τγγ\tau_{\gamma\gamma}4.49×1023 s4.49 \times10^{23}\ \mathrm{s}
τγγ/τSPIrec\tau_{\gamma\gamma}/\tau_{\mathrm{SPI}}^{\mathrm{rec}}41.4
⟨σv⟩equiv (cm3 s−1)\langle\sigma v\rangle_{\mathrm{equiv}}\ (\mathrm{cm^3\,s^{-1}})2.26×10−262.26 \times10^{-26}

Table I. Benchmark parameters and predictions for xenon scattering, radiative decay, and present-day annihilation. Event counts refer to the 2.84 tonne-year LZ exposure. NtruthN_{\mathrm{truth}} is obtained without detector efficiency, while NϵN_{\epsilon} is an efficiency-weighted true-recoil expectation, not an event-level likelihood result. NϵN_{\epsilon} includes the digitized nuclear-recoil efficiency over its available range, 5.4≤ER/keV≤299.95.4 \le E_R/\mathrm{keV} \le299.9. The lifetime τγγ\tau_{\gamma\gamma} uses leading-order hadronic matching, and τSPIrec\tau_{\mathrm{SPI}}^{\mathrm{rec}} is the reconstructed INTEGRAL/SPI lower limit.

Integrating over the available digitized efficiency range, 5.4≤ER/keV≤299.95.4 \le E_R/\mathrm{keV} \le299.9, we obtain

Nϵ≃1.63.(39)N_{\epsilon} \simeq1.63. \tag*{(39)}

The integration extends beyond the high-energy 50% efficiency point at 269.9 keV [1], using the available digitization up to 299.9 keV. Contributions above this true-recoil energy are not included.

The recoil-energy distribution provides further information beyond the total event rate. Approximately 1.18 efficiency-weighted events are expected between 100 and 200 keV, compared with about 0.22 events between 215 and 269.9 keV. The latter interval includes the reported candidate energy and extends to the high-energy 50% efficiency point. While a dedicated likelihood analysis for the present scenario is beyond the scope of this paper, the predicted recoil spectrum and event normalization appear to give a reasonable description of the LZ event.

Finite-momentum nuclear response

The predicted recoil spectrum is sensitive to the finite-momentum scalar nuclear response, particularly at the high recoil energies relevant to the LZ candidate. At ER=248 keVE_R=248\ \mathrm{keV}, the momentum transfer in xenon is q≃0.25 GeVq\simeq0.25\ \mathrm{GeV}, and the spatial distribution of nucleons is important. To examine the dependence on the nuclear response, we compare our central prediction with the phenomenological Helm form factor [124, 125], which increases the efficiency-weighted event expectation from Nϵ=1.63N_{\epsilon} = 1.63 to 4.70 events. The chiral-EFT nuclear structure factors of Refs. [126, 127], which also include two-nucleon contributions, instead yield 1.61 efficiency-weighted events. These differences reflect the high sensitivity of the predicted spectrum to the momentum dependence of the nuclear response, particularly near its diffraction minima.

Note that the characteristic exothermic recoil scale depends on the target nucleus through Eq. (28). High-recoil searches using xenon, argon, germanium, and other targets could therefore provide complementary tests of this scenario [22, 82]. Other proposed probes of inelastic interpretations of the LZ event include annual modulation [64], heavy-element paleodetectors [83, 128], and searches in large neutrino detectors [128]. The sensitivity of endothermic interpretations to the high-velocity tail of the halo, including perturbations induced by the Large Magellanic Cloud, has also been studied [65, 128, 129].

Correlated Signatures

The interaction in Eq. (1) governs the thermal relic abundance, the scattering on xenon, but also the strength of correlated signatures such as the radiative decay of the heavier scalar ϕ2\phi_2 and the annihilation of ϕ1ϕ2\phi_1\phi_2 into SM particles. Below, we first examine such indirect photon signals and then discuss complementary collider searches.

Radiative decay of 2

An exothermic scattering signal requires an appreciable population of the heavier scalar ϕ2\phi_2 to survive until the present epoch. Although ϕ2\phi_2 is not absolutely stable, its decay is strongly suppressed by the small mass splitting. For Δm=600 keV<2me\Delta m = 600\ \mathrm{keV} < 2m_e, the channel ϕ2→ϕ1e+e−\phi_2 \to\phi_1 e^+e^- is kinematically inaccessible. The leading decay of ϕ2\phi_2 is then ϕ2→ϕ1γγ\phi_2 \to\phi_1\gamma\gamma, which produces a continuous photon spectrum. Related diphoton transitions in multicomponent scalar and inelastic fermionic dark-matter models have been investigated in Refs. [99, 130].

After electroweak symmetry breaking, the interaction in Eq. (20) generates the coupling −gvhϕ1ϕ2h-g v_h\phi_1\phi_2 h. The transition to two photons proceeds through electrically charged particles coupled to the Higgs sector. The diphoton invariant mass, s=(k1+k2)2=2k1⋅k2s=(k_1+k_2)^2=2k_1\cdot k_2, with k1k_1 and k2k_2 the two photon momenta, is limited to the range 0≤s≤(Δm)20\le s\le(\Delta m)^2, well below the physical Higgs mass, so that the Higgs can be integrated out. The ϕ2→ϕ1γγ\phi_2\to\phi_1\gamma\gamma transition amplitude is [131, 132, 133]

M=ϵ1∗μϵ2∗ν4αgcγ(s)mh2[(k1⋅k2)gμν−k2μk1ν],(40)\mathcal{M}=\epsilon_1^{*\mu}\epsilon_2^{*\nu}\frac{4\alpha g c_\gamma(s)}{m_h^2}\left[(k_1\cdot k_2)g_{\mu\nu}-k_{2\mu}k_{1\nu}\right], \tag*{(40)}
cγ(s)=α8π[P(s)+H(s)],(41)c_\gamma(s)=\frac{\alpha}{8\pi}\left[\mathcal{P}(s)+\mathcal{H}(s)\right], \tag*{(41)}

where mhm_h is the Higgs mass, ϵi\epsilon_i are the photon polarization vectors, and α\alpha the fine-structure constant. Here, P(s)\mathcal{P}(s) denotes the perturbative charged-particle contribution, while H(s)\mathcal{H}(s) accounts for the low-energy hadronic contribution to the scalar current.

The perturbative contribution is

P(s)=A1(s4mW2)+∑f=e,μ,τ,c,b,tNcfQf2A1/2(s4mf2),(42)\mathcal{P}(s)=A_1\left(\frac{s}{4m_W^2}\right)+\sum_{f=e,\mu,\tau,c,b,t}N_c^f Q_f^2 A_{1/2}\left(\frac{s}{4m_f^2}\right), \tag*{(42)}

where mWm_W and mfm_f are the W-boson and fermion masses, QfQ_f is the electric charge in units of the positron charge, and Ncf=1N_c^f=1 for leptons and 3 for quarks. For all but the electron, the loop functions can be evaluated in the zero-momentum limit, A1(0)=−7A_1(0)=-7 and A1/2(0)=4/3A_{1/2}(0)=4/3. The momentum dependence of the electron loop must be retained,

A1/2(x)=2[x+(x−1)f(x)]x2,A_{1/2}(x)=\frac{2[x+(x-1)f(x)]}{x^2},
f(x)=arcsin⁡21x,(43)f(x)=\arcsin^2\frac{1}{\sqrt{x}}, \tag*{(43)}

since for the benchmark s/(4me2)≃0.35s/(4m_e^2)\simeq0.35 at the kinematic endpoint.

For the nonperturbative contributions from Higgs coupling to the light uu, dd, ss quarks we reuse the LO ChPT results of Ref. [133]. Translating to our notation, we get

H(0)=2327.(44)\mathcal{H}(0)=\frac{23}{27}. \tag*{(44)}

Including the momentum dependence of the pion- and kaon-loop contributions changes the integrated decay width by less than 10−610^{-6} for Δm=600 keV\Delta m=600\ \mathrm{keV}, much less than the size of the neglected NLO ChPT corrections, which are of order (mπ,K/4πfπ,K)2∼O(20%)(m_{\pi,K}/4\pi f_{\pi,K})^2\sim\mathcal{O}(20\%).

The differential ϕ2→ϕ1γγ\phi_2\to\phi_1\gamma\gamma decay width is

dΓγγds=g2∣cγ(s)∣2s2λ(mϕ22,mϕ12,s)64π3mϕ23(mh2−s)2,(45)\frac{\mathrm{d}\Gamma_{\gamma\gamma}}{\mathrm{d}s}=\frac{g^2|c_\gamma(s)|^2s^2\sqrt{\lambda(m_{\phi_2}^2,m_{\phi_1}^2,s)}}{64\pi^3m_{\phi_2}^3(m_h^2-s)^2}, \tag*{(45)}

where λ(a,b,c)=a2+b2+c2−2ab−2ac−2bc\lambda(a,b,c)=a^2+b^2+c^2-2ab-2ac-2bc is the Källén function. In terms of fractional photon energy, x=Eγ/Δmx=E_\gamma/\Delta m, the normalized distribution is (dNγ/dx)=140x3(1−x)3(\mathrm{d}N_\gamma/\mathrm{d}x)=140x^3(1-x)^3 [134], which peaks near Eγ≃Δm/2E_\gamma\simeq\Delta m/2, see Figure 4.

Normalized differential photon-energy distribution

Figure 4. Normalized differential photon-energy distribution, (1/Nγ) dNγ/dEγ(1/N_\gamma)\,dN_\gamma/dE_\gamma, from ϕ2→ϕ1γγ\phi_2 \to\phi_1\gamma\gamma in the ϕ2\phi_2 rest frame, for mϕ1=100m_{\phi_1}=100 GeV and Δm=600\Delta m=600 keV. The distribution integrates to unity, with Nγ=2N_\gamma=2 photons per decay. No instrumental response is applied.

In the small-splitting limit, i.e., for approximately constant cγ(s)≃cγ(0)c_\gamma(s)\simeq c_\gamma(0), the total decay width is

Γγγ≃g2∣cγ(0)∣2(Δm)7210π3mh4mϕ22.(46)\Gamma_{\gamma\gamma}\simeq\frac{g^2|c_\gamma(0)|^2(\Delta m)^7}{210\pi^3m_h^4m_{\phi_2}^2}. \tag*{(46)}

The strong dependence on the mass splitting, Γγγ∝(Δm)7\Gamma_{\gamma\gamma}\propto(\Delta m)^7, leads to a cosmologically stable ϕ2\phi_2, despite its large coupling to the Higgs, and the corresponding induced couplings to the electrically charged SM sector. For the benchmark in Table 1,

τγγbench≃4.5×1023 s.(47)\tau_{\gamma\gamma}^{\mathrm{bench}}\simeq4.5\times10^{23}\ \mathrm{s}. \tag*{(47)}

For comparison, omitting the hadronic contribution, setting H=0\mathcal{H}=0, while retaining the perturbative loops gives τγγ≃1.4×1024\tau_{\gamma\gamma}\simeq1.4\times10^{24} s. In either case, the benchmark lifetime is much longer than the age of the Universe, so radiative depletion of the excited population is negligible.

The radiative ϕ2\phi_2 decay signal can be searched for observationally. Assuming that ϕ1\phi_1 and ϕ2\phi_2 follow the same Galactic DM density profile, the differential photon intensity from ϕ2\phi_2 decay in the direction Ω\Omega is given by

dΦγdEγ dΩ=r24πmˉτγγdNγdEγD(Ω),(48)\frac{d\Phi_\gamma}{dE_\gamma\,d\Omega}=\frac{r_2}{4\pi\bar{m}\tau_{\gamma\gamma}}\frac{dN_\gamma}{dE_\gamma}D(\Omega), \tag*{(48)}

where

D(Ω)=∫l.o.s.dℓ ρDM(ℓ,Ω)(49)D(\Omega)=\int_{\mathrm{l.o.s.}}d\ell\,\rho_{\mathrm{DM}}(\ell,\Omega) \tag*{(49)}

is the line-of-sight integral of the Galactic DM density. We model the Galactic DM distribution using a Navarro–Frenk–White (NFW) density profile with scale radius rs=9.98r_s=9.98 kpc. We take the Sun–Galactic-center distance to be RGC=8.34R_{\mathrm{GC}}=8.34 kpc and normalize the profile to ρDM(RGC)=ρ⊙=0.3\rho_{\mathrm{DM}}(R_{\mathrm{GC}})=\rho_\odot=0.3 GeV cm−3^{-3}, consistent with the local density adopted in § 1.2 [135, 136].

The predicted photon continuum lies within the hard X-ray and soft gamma-ray energy range accessible to the Spectrometer aboard the INTEGRAL satellite (SPI), which employs high-resolution germanium detectors [137]. Analyses of 16 years of SPI observations have examined diffuse emission from the inner Galaxy over the energy range 30 keV–8 MeV [121]. The observational analyses employ spatial templates and models of astrophysical emission, including diffuse inverse-Compton radiation, unresolved sources, and positron-annihilation emission, to constrain additional photon contributions associated with Galactic dark-matter decay. In particular, Ref. [134] performed a likelihood-based analysis of related two- and three-photon continua from inelastic DM, including the astrophysical background components in the spectral fit (see also [121]).

To constrain the radiative decay of the heavier scalar, we use the NFW-template spectrum and energy-redistribution matrix released with the 16-year INTEGRAL/SPI analysis [121, 122]. The release contains 24 measured-energy bins spanning 30 keV–8 MeV and a response matrix defined on 377 true-energy bins. The comparison uses the Galactic region ∣l∣,∣b∣≤47.5∘|l|,|b|\leq47.5^\circ, for which the integrated decay flux is

dΦγROIdEγ=r22DROI4πmˉτγγdNγdEγ,(50)\frac{d\Phi_\gamma^{\mathrm{ROI}}}{dE_\gamma}=\frac{r_2^2D_{\mathrm{ROI}}}{4\pi\bar{m}\tau_{\gamma\gamma}}\frac{dN_\gamma}{dE_\gamma}, \tag*{(50)}

where DROI=∫ROIdΩ D(Ω)D_{\mathrm{ROI}}=\int_{\mathrm{ROI}}d\Omega\,D(\Omega). For the halo parameters specified above, integration over ∣l∣,∣b∣≤47.5∘|l|,|b|\leq47.5^\circ, with the line-of-sight integral extended to infinity, gives DROI≃8.85×1022D_{\mathrm{ROI}}\simeq8.85\times10^{22} GeV cm−2^{-2}. For the numerical evaluation, we use r2=0.49924≃0.50r_2=0.49924\simeq0.50. We integrate the Galactic photon flux over the true-energy bins and apply the released energy-redistribution matrix to obtain the predicted channel rates RiDMR_i^{\mathrm{DM}}. These are compared with the NFW-template rates RiR_i and their statistical uncertainties σi\sigma_i. Using a χ2\chi^2 constructed from all 24 measured energy bins gives one-sided 95% C.L. upper bound

τSPIrec≳1.1×1022 s.(51)\tau_{\mathrm{SPI}}^{\mathrm{rec}}\gtrsim1.1\times10^{22}\ \mathrm{s}. \tag*{(51)}

The benchmark lifetime in Eq. (47) is approximately 40×40\times longer. Although the reference benchmark lies below the reconstructed INTEGRAL/SPI limit, larger mass splittings can give a substantially enhanced photon signal, since Γγγ∝g2(Δm)7/mϕ22\Gamma_{\gamma\gamma}\propto g^2(\Delta m)^7/m_{\phi_2}^2. This is particularly relevant at lower DM masses, where the relic abundance requires a larger portal coupling and an order-one LZ event yield occurs for splittings approaching 1 MeV. The SPI lifetime constraint therefore excludes part of the LZ favored band in the (mϕ,Δm)(m_\phi,\Delta m) plane, for higher mass splitting values, shown as white hatched excluded region in Figure 2.

Note that in obtaining these results, we neglect inter-bin correlations and do not propagate systematic uncertainties from the template and background modeling. A dedicated analysis of related inelastic dark-matter photon continua, including uncertainties in astrophysical background components, was performed in Ref. [134], which also considers a diphoton continuum with the same leading-order spectral shape as ours, although a direct comparison of lifetime limits requires consistent signal normalization, astrophysical assumptions, and statistical treatment.

Prospects for COSI

The photon continuum from ϕ2→ϕ1γγ\phi_2\to\phi_1\gamma\gamma falls partly within the 0.2–5 MeV energy range of the planned COSI mission [138]. For instance, for the benchmark in Table 1 about 82.7% of emitted photons are above the COSI energy threshold.

Recent forecasts based on all-sky simulations indicate that two years of COSI observations could strengthen the lower lifetime limits for Galactic dark-matter decay lines by more than an order of magnitude over much of the COSI energy band [138]. Assuming that comparable sensitivity improvements can be achieved for continuum decays, COSI is expected to probe a substantial part of the LZ-favored region shown in Figure 2, providing an important test of this scenario.

Annihilation and indirect limits

The mixed annihilation process ϕ1ϕ2→h∗→SM\phi_1\phi_2 \to h^* \to\mathrm{SM} that controls the thermal relic abundance, also contributes to late-time indirect-detection signals. This process contains a nonvanishing ss-wave component, with the low-velocity cross section

⟨σv⟩12≃4.52×10−26 cm3 s−1.(52)\langle\sigma v\rangle_{12} \simeq4.52 \times10^{-26}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}. \tag*{(52)}

One can set a constraint on the ϕ1ϕ2\phi_1\phi_2 annihilation rate per unit volume for distinguishable initial states, n1n2⟨σv⟩12n_1n_2\langle\sigma v\rangle_{12}, by recasting the bounds on annihilation cross section ⟨σv⟩equiv\langle\sigma v\rangle_{\mathrm{equiv}} of a self-conjugate DM, by identifying

n1n2⟨σv⟩12=12(ρDMmˉ)2⟨σv⟩equiv,(53)n_1n_2\langle\sigma v\rangle_{12} = \frac{1}{2}\left(\frac{\rho_{\mathrm{DM}}}{\bar{m}}\right)^2\langle\sigma v\rangle_{\mathrm{equiv}}, \tag*{(53)}

which then implies

⟨σv⟩equiv=2r1r2⟨σv⟩12.(54)\langle\sigma v\rangle_{\mathrm{equiv}} = 2r_1r_2\langle\sigma v\rangle_{12}. \tag*{(54)}

For r1≃r2≃1/2r_1 \simeq r_2 \simeq1/2 we thus have

⟨σv⟩equiv≃2.26×10−26 cm3 s−1.(55)\langle\sigma v\rangle_{\mathrm{equiv}} \simeq2.26 \times10^{-26}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}. \tag*{(55)}

The factor 2r1r22r_1r_2 in (54) accounts for both the relative populations and the distinguishable initial states. At the Table I benchmark mass, the ϕ1ϕ2→h∗→SM\phi_1\phi_2 \to h^* \to\mathrm{SM} annihilation produces predominantly W+W−W^+W^- and ZZZZ in roughly 3:13:1 proportion, with other SM final states contributing below percent level.

We derive the dwarf-galaxy constraint shown in Figure 2 using the released energy-bin likelihoods from the 14-year Fermi-LAT analysis [123]. For each dark-matter mass, we construct the predicted gamma-ray spectrum using the mass-dependent annihilation rates and branching fractions, including off-shell electroweak channels below threshold and Higgs-pair final states where kinematically accessible. We combine the likelihoods of the 42 dwarf spheroidal galaxies in the Benchmark sample, which contains 30 galaxies with measured JJ factors and 12 with estimated values. The JJ-factor uncertainties are included by profiling the corresponding nuisance parameters.

We obtain the observed one-sided 95% confidence-level upper limit on the equivalent annihilation cross section by requiring an increase of 2.71 in the profiled −2ln⁡L-2\ln\mathcal{L} relative to its minimum, with a nonnegative signal normalization. Since the predicted spectra depend only weakly on Δm\Delta m, we evaluate the likelihood as a function of dark-matter mass at a representative splitting of 600 keV and apply the resulting limits across the scanned splitting range. For the benchmark in Table I, we obtain

⟨σv⟩equiv<3.34×10−26 cm3 s−1(95% C.L.).(56)\langle\sigma v\rangle_{\mathrm{equiv}} < 3.34 \times10^{-26}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}\qquad(95\%\ \mathrm{C.L.}). \tag*{(56)}

The predicted value, ⟨σv⟩equiv≃2.26×10−26 cm3 s−1\langle\sigma v\rangle_{\mathrm{equiv}} \simeq2.26 \times10^{-26}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}, lies below this limit. Across the thermal parameter space, the predicted annihilation signal exceeds the reconstructed limit for mϕ1≲72 GeVm_{\phi_1} \lesssim72\ \mathrm{GeV}, as shown in Figure 2.

A combined analysis of dwarf-spheroidal observations by Fermi-LAT, HAWC, H.E.S.S., MAGIC, and VERITAS provides complementary constraints on annihilation into electroweak gauge bosons [139]. The reported limits depend appreciably on the adopted dwarf-galaxy JJ factors. Because the benchmark predicts a specific mixture of W+W−W^+W^- and ZZZZ final states, an exact application of this result requires a likelihood-level evaluation using the corresponding mixed gamma-ray spectrum.

Annihilation around recombination also deposits energy into the photon–baryon plasma. The relevant energy-injection parameter is

pann=feff⟨σv⟩equivmˉ,(57)p_{\mathrm{ann}} = f_{\mathrm{eff}}\frac{\langle\sigma v\rangle_{\mathrm{equiv}}}{\bar{m}}, \tag*{(57)}

where fefff_{\mathrm{eff}} is the effective energy-deposition efficiency for the predicted annihilation products. We compute the injected photon and electron–positron spectra per ϕ1ϕ2\phi_1\phi_2 annihilation, including the branching fractions quoted above, and convolve them with the deposition-efficiency functions of Ref. [140]. This gives feff≃0.179f_{\mathrm{eff}}\simeq0.179, pann≃4.05×10−29 cm3 s−1 GeV−1p_{\mathrm{ann}}\simeq4.05\times10^{-29}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}\,\mathrm{GeV}^{-1}. This prediction lies about an order of magnitude below the 95% upper limits pann<3.2×10−28 cm3 s−1 GeV−1p_{\mathrm{ann}} < 3.2\times10^{-28}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}\,\mathrm{GeV}^{-1} from Planck (TT,TE,EE+lowE+lensing) [141] and pann<3.6×10−28 cm3 s−1 GeV−1p_{\mathrm{ann}} < 3.6\times10^{-28}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}\,\mathrm{GeV}^{-1} from the combination of Planck and ACT DR6 primary CMB, CMB lensing, and DESI BAO data (P-ACT-LB) [142]. For this comparison, we use the state fractions obtained from the coupled cosmological evolution at recombination. They differ negligibly from their present-day values because state-changing reactions and radiative decay are ineffective over the intervening epoch.

Although the diagonal Higgs couplings vanish at tree level, processes with identical incoming scalars are not strictly absent. In particular, tt- and uu-channel ϕ2\phi_2 exchange mediates

ϕ1ϕ1→hh∗→hffˉ,(58)\phi_1\phi_1 \to h h^* \to h f\bar{f}, \tag*{(58)}

where ff denotes a Standard Model fermion. For the benchmark, its low-velocity cross section is ⟨σv⟩11→hffˉ≃4.55×10−33 cm3 s−1\langle\sigma v\rangle_{11\to h f\bar{f}} \simeq4.55\times10^{-33}\ \mathrm{cm}^{3}\,\mathrm{s}^{-1}. Including the identical-pair factor 1/21/2 and the relative abundances, this channel contributes approximately 5×10−85\times10^{-8} of the Galactic ϕ1ϕ2\phi_1\phi_2 annihilation source. The loop-induced ϕ1ϕ1→W+W−\phi_1\phi_1\to W^+W^- contribution is suppressed by a similar factor. Both are negligible for the dwarf-galaxy and CMB comparisons above. Their relative importance differs in the Sun, where capture produces a much larger population of ϕ1\phi_1 than of ϕ2\phi_2; see § 1.4.

Collider and other probes

Collider searches provide an additional probe of the off-diagonal Higgs portal. If mϕ1+mϕ2<mhm_{\phi_1}+m_{\phi_2}<m_h, the mixed pair can be produced through the on-shell Higgs decay h→ϕ1ϕ2h\to\phi_1\phi_2. For particles that escape the detector, the corresponding partial width is constrained by Higgs-rate measurements and searches for invisible Higgs decays [95, 143, 144, 145, 146]. This is not an issue for our benchmark and for masses shown in Figure 2, since in all these cases mϕ1+mϕ2>mhm_{\phi_1}+m_{\phi_2}>m_h. The diagonal channels h→ϕiϕih\to\phi_i\phi_i are also absent at tree level because the diagonal portal couplings are set to f1=f2=0f_1=f_2=0.

The leading collider production channel of ϕ1ϕ2\phi_1\phi_2 is therefore off-shell Higgs production accompanied by initial-state radiation,

pp→h∗+j→ϕ1ϕ2+j,(59)pp\to h^*+j\to\phi_1\phi_2+j, \tag*{(59)}

resulting in a monojet signature (ϕ2\phi_2 is stable on collider scales) [147, 148, 149, 150]. Due to additional 3 body final state phase space suppression, existing collider limits on this topology are not expected to compete with the direct and indirect probes for the moderate portal coupling g≃0.05g\simeq0.05 considered here.

The singlet states can also contribute radiatively to off-shell Higgs processes such as gg→h∗→ZZgg\to h^*\to ZZ, modifying the high-invariant-mass diboson spectrum [151, 152], though for the typical values of gg considered here these are again expected to be below experimental sensitivity.

Finally, the characteristic exothermic recoil scale depends on the target nucleus, see Eq. (28). High-recoil searches with xenon, argon, germanium, and other targets could therefore provide complementary tests of this scenario.

Solar capture and annihilation

DM entering the Sun can scatter on solar nuclei, ϕ1A→ϕ2A\phi_1A\to\phi_2A and vice versa, and become gravitationally bound. Such scattering changes the relative populations of ϕ1\phi_1 and ϕ2\phi_2, so that these in general no longer equal the Galactic abundances. Moreover, because diagonal elastic scattering arises only radiatively, repeated collisions need not lead to common thermal distributions. We follow the captured particles along their nonthermal orbits, allowing for subsequent inelastic transitions and possible ejection, rather than assuming a thermal spatial distribution. The numerical solar analysis below focuses on the benchmark in Table I.

Inelastic capture and solar population evolution

The first of the two possible off-diagonal portal scatterings

ϕ2A→ϕ1A,ϕ1A→ϕ2A,(60)\phi_2A\to\phi_1A,\qquad\phi_1A\to\phi_2A, \tag*{(60)}

is exothermic and provides the dominant capture channel. For a stationary target, the reverse endothermic transition requires

w≥wthr,A=2Δmμ1A,μ1A=mϕ1mAmϕ1+mA,(61)w\geq w_{\mathrm{thr},A}=\sqrt{\frac{2\Delta m}{\mu_{1A}}},\qquad\mu_{1A}=\frac{m_{\phi_1}m_A}{m_{\phi_1}+m_A}, \tag*{(61)}

where ww is the local dark-matter–nucleus relative speed. For the benchmark in Table I and a stationary 56Fe{}^{56}\mathrm{Fe} nucleus,

wthr,Fe≃1800 km s−1,(62)w_{\mathrm{thr},\mathrm{Fe}}\simeq1800\ {\rm km\ s}^{-1}, \tag*{(62)}

which exceeds the maximum incident speed at the solar center, wmax⁡≃1600 km s−1w_{\max}\simeq1600\ {\rm km\ s}^{-1}, for the adopted halo distribution and solar potential.

For stationary iron nuclei, the endothermic transition is therefore kinematically inaccessible at the benchmark. In the numerical calculation, we include thermal nuclear motion in both halo capture and subsequent transitions along captured orbits. Thermal motion removes the strict stationary-target threshold, although the contribution from iron to endothermic halo capture remains negligible. For the adopted solar composition, endothermic transitions are dominated by heavier trace nuclei. These processes affect the small captured excited-state population, but the resulting mixed annihilation rate remains strongly suppressed, as discussed below.

Exothermic scattering does not necessarily lead to capture, since the outgoing ϕ1\phi_1 can remain gravitationally unbound. We define C2→1C_{2\to1} as the number of incident ϕ2\phi_2 particles per unit time that scatter into gravitationally bound ϕ1\phi_1 orbits. In terms of the halo speed distribution, this rate can be written as

C2→1=n2,∞∑A∫0R⊙4πr2 dr∫0∞du f⊙(u)u wχ Ω2→1,A(wχ,r),(63)C_{2\to1}=n_{2,\infty}\sum_A\int_0^{R_\odot}4\pi r^2\,dr \int_0^\infty du\,\frac{f_\odot(u)}{u}\,w_\chi\,\Omega_{2\to1,A}(w_\chi,r), \tag*{(63)}

where

n2,∞=ρ2DMmˉ,wχ2=u2+vesc2(r).(64)n_{2,\infty}=\frac{\rho_{2\mathrm{DM}}}{\bar m},\qquad w_\chi^2=u^2+v_{\mathrm{esc}}^2(r). \tag*{(64)}

Here f⊙(u)f_\odot(u) is the unit-normalized halo speed distribution in the solar frame. The quantity Ω2→1,A\Omega_{2\to1,A} is the scattering rate per incident particle on nuclear species AA, averaged over the thermal target distribution and integrated over final states in which the outgoing ϕ1\phi_1 is gravitationally bound. We evaluate Eq. (63) using the adopted solar density and composition profiles, the nuclear form factors, and the exact inelastic two-body kinematics [153]. For the benchmark,

C2→1≃9.3×1019 s−1.(65)C_{2\to1}\simeq9.3\times10^{19}\ {\rm s}^{-1}. \tag*{(65)}

The phase-space point following the capture collision fixes the initial bound orbit. We evolve these orbits through subsequent passages through the solar interior, allowing for further scattering, conversion, and ejection. For a constant capture rate, the present-day density is

nicap(r,t⊙)=C2→1∫0t⊙da pi(r,a),(66)n_i^{\mathrm{cap}}(\mathbf{r},t_\odot)=C_{2\to1}\int_0^{t_\odot} da\,p_i(\mathbf{r},a), \tag*{(66)}

where t⊙t_\odot is the age of the Sun and pi(r,a) d3rp_i(\mathbf{r},a)\,d^3r is the probability that a particle captured a time aa earlier occupies the volume d3rd^3r in state ϕi\phi_i. This probability includes subsequent conversion and ejection.

Approximately 72% of the particles captured as ϕ1\phi_1 undergo no further state-changing collision in the resulting orbital histories. The spatial densities are therefore constructed by weighting each orbit by the fraction of its period spent in each radial interval, rather than by imposing a common isothermal profile. Radiative conversion is likewise negligible on solar timescales. Even if a bound ϕ2\phi_2 remained in that state for the full age of the Sun, t⊙t_\odot, the decay time is orders of magnitude larger, t⊙/τγγ≃3.2×10−7t_\odot/\tau_{\gamma\gamma}\simeq3.2\times10^{-7}.

Solar annihilation and neutrino constraints

Let nicap(r)n_i^{\mathrm{cap}}(\mathbf{r}) denote the present-day number density of gravitationally bound ϕi\phi_i particles. For annihilations occurring within the solar interior, we define

I11cap=∫r<R⊙d3r [n1cap(r)]2,(67)I_{11}^{\mathrm{cap}}=\int_{r<R_\odot}d^3r\,\left[n_1^{\mathrm{cap}}(\mathbf{r})\right]^2, \tag*{(67)}
I12cap=∫r<R⊙d3r n1cap(r)n2cap(r).(68)I_{12}^{\mathrm{cap}}=\int_{r<R_\odot}d^3r\,n_1^{\mathrm{cap}}(\mathbf{r})n_2^{\mathrm{cap}}(\mathbf{r}). \tag*{(68)}

The orbital calculation gives

I11cap≃5.3×1041 cm−3,I12cap≃5.8×1030 cm−3.(69)I_{11}^{\mathrm{cap}}\simeq5.3\times10^{41}\ \mathrm{cm}^{-3},\qquad I_{12}^{\mathrm{cap}}\simeq5.8\times10^{30}\ \mathrm{cm}^{-3}. \tag*{(69)}

Solar capture is dominated by ϕ2A→ϕ1A\phi_2 A\to\phi_1 A, leaving most captured dark matter in the lighter state. Subsequent endothermic scattering can regenerate ϕ2\phi_2, but these particles spend relatively little time in the excited state before converting back or being ejected. The resulting small ϕ2\phi_2 population accounts for the hierarchy I12cap≪I11capI_{12}^{\mathrm{cap}}\ll I_{11}^{\mathrm{cap}}. The rate for annihilation between the two captured populations is

Γ12cap=I12cap⟨σv⟩12≃2.6×105 s−1,(70)\Gamma_{12}^{\mathrm{cap}}=I_{12}^{\mathrm{cap}}\langle\sigma v\rangle_{12}\simeq2.6\times10^{5}\ \mathrm{s}^{-1}, \tag*{(70)}

where we used ⟨σv⟩12\langle\sigma v\rangle_{12} in (52). The small value of Γ12cap\Gamma_{12}^{\mathrm{cap}} is caused by the limited spatial overlap of the captured ϕ1\phi_1 and ϕ2\phi_2 populations.

The much larger captured ϕ1\phi_1 population also permits annihilation of two ϕ1\phi_1 particles. For a final state XX, the corresponding event rate is

Γ11cap,X=12I11cap⟨σv⟩11→X,(71)\Gamma_{11}^{\mathrm{cap},X}=\frac{1}{2}I_{11}^{\mathrm{cap}}\langle\sigma v\rangle_{11\to X}, \tag*{(71)}

where the factor 1/21/2 avoids double counting pairs of identical incoming particles. At tree level, two off-diagonal portal vertices generate

ϕ1ϕ1→hh∗→hffˉ(72)\phi_1\phi_1\to hh^*\to hf\bar{f} \tag*{(72)}

through intermediate ϕ2\phi_2 exchange. Although suppressed by the three-body phase space and the off-shell Higgs propagator, this process has the low-velocity rate coefficient

⟨σv⟩11→hffˉ≃4.6×10−33 cm3s−1,(73)\langle\sigma v\rangle_{11\to h f\bar{f}}\simeq4.6\times10^{-33}\ \mathrm{cm}^3\mathrm{s}^{-1}, \tag*{(73)}

and therefore gives

Γ11cap,hffˉ=12I11cap⟨σv⟩11→hffˉ≃1.2×109 s−1.(74)\Gamma_{11}^{\mathrm{cap},h f\bar{f}}=\frac{1}{2}I_{11}^{\mathrm{cap}}\langle\sigma v\rangle_{11\to h f\bar{f}}\simeq1.2\times10^{9}\ \mathrm{s}^{-1}. \tag*{(74)}

The captured ϕ1\phi_1 population also annihilates through the loop-induced channel

ϕ1ϕ1→W+W−.(75)\phi_1\phi_1\to W^+W^-. \tag*{(75)}

The renormalized one-loop amplitude includes the induced diagonal Higgs form factor together with the contact, gauge, and Goldstone contributions required by electroweak gauge invariance, retaining their interference. The resulting threshold rate coefficient is

⟨σv⟩11→WW≃2.2×10−33 cm3s−1.(76)\langle\sigma v\rangle_{11\to WW}\simeq2.2\times10^{-33}\ \mathrm{cm}^3\mathrm{s}^{-1}. \tag*{(76)}

Its velocity dependence is negligible over the range populated by the captured orbits. Eq. (71) then gives

Γ11cap,WW=12I11cap⟨σv⟩11→WW≃5.9×108 s−1.(77)\Gamma_{11}^{\mathrm{cap},WW}=\frac{1}{2}I_{11}^{\mathrm{cap}}\langle\sigma v\rangle_{11\to WW}\simeq5.9\times10^{8}\ \mathrm{s}^{-1}. \tag*{(77)}

The hffˉh f\bar{f} rate is approximately twice the loop-induced W+W−W^+W^- rate. Although both ϕ1ϕ1\phi_1\phi_1 rate coefficients are roughly seven orders of magnitude below ⟨σv⟩12\langle\sigma v\rangle_{12}, their event rates are larger than Γ12cap\Gamma_{12}^{\mathrm{cap}} because I11capI_{11}^{\mathrm{cap}} greatly exceeds I12capI_{12}^{\mathrm{cap}}.

The calculated annihilation channels remove particles from the captured population at a rate

2(Γ12cap+Γ11cap,hfˉf+Γ11cap,WW)C2→1≃3.9×10−11,(78)\frac{2\left(\Gamma_{12}^{\mathrm{cap}}+\Gamma_{11}^{\mathrm{cap},h\bar{f}f}+\Gamma_{11}^{\mathrm{cap},WW}\right)}{C_{2\to1}}\simeq3.9\times10^{-11}, \tag*{(78)}

where the factor of two counts the two captured particles removed in each event. Annihilation between captured particles therefore has a negligible effect on the solar population, and the calculated captured-pair channels do not establish capture–annihilation equilibrium.

The preceding rates involve two particles drawn from the captured solar population. A separate contribution arises when a captured ϕ1\phi_1 annihilates with an unbound Galactic ϕ2\phi_2 passing through the Sun,

ϕ1capϕ2halo→h∗→W+W−, ZZ,…(79)\phi_1^{\mathrm{cap}}\phi_2^{\mathrm{halo}}\to h^*\to W^+W^-,\ ZZ,\ldots \tag*{(79)}

Denoting the gravitationally focused density of the incident population by n2halo(r)n_2^{\mathrm{halo}}(\mathbf{r}), its solar-interior rate is

Γ12cap−halo=∫r<R⊙d3r n1cap(r)n2halo(r)⟨σvrel⟩12,r.(80)\Gamma_{12}^{\mathrm{cap-halo}}=\int_{r<R_\odot}d^3r\,n_1^{\mathrm{cap}}(\mathbf{r})n_2^{\mathrm{halo}}(\mathbf{r})\langle\sigma v_{\mathrm{rel}}\rangle_{12,r}. \tag*{(80)}

The velocity average combines the nonthermal captured distribution with the incident halo phase-space distribution transported through the solar potential [54]. This contribution is not contained in I12capI_{12}^{\mathrm{cap}} and does not enter Eq. (78); each such event removes only one captured particle. Its evaluation requires the position-dependent overlap of the captured orbits with the gravitationally focused halo distribution and is therefore kept separate from the numerical rates quoted above.

For comparison with neutrino-telescope searches, we use the calculated ϕ1ϕ1→W+W−\phi_1\phi_1 \to W^+W^- component. Expressing the released mχ=100m_\chi= 100 GeV, pure-W+W−W^+W^- result from the ten-year IceCube analysis as an annihilation-rate limit, using the capture prescription and equilibrium normalization adopted in that analysis, gives [155]

ΓICWW,90≃5.4×1021 s−1(81)\Gamma_{\mathrm{IC}}^{WW,90} \simeq5.4 \times10^{21}\ \mathrm{s}^{-1} \tag*{(81)}

(90% C.L.). Here the equilibrium relation is used only to translate the normalization of the IceCube signal template and is not imposed on the captured population of the present model. The calculated W+W−W^+W^- component satisfies

Γ11cap,WWΓICWW,90≃1.1×10−13.(82)\frac{\Gamma_{11}^{\mathrm{cap,WW}}}{\Gamma_{\mathrm{IC}}^{WW,90}} \simeq1.1 \times10^{-13}. \tag*{(82)}

The hffˉhf\bar{f} event rate lies more than twelve orders of magnitude below this rate (though the pure-W+W−W^+W^- IceCube template cannot be applied directly because its neutrino spectrum is different). The captured-particle–incident-halo contribution produces a mixture of electroweak final states and has not been evaluated numerically. A complete solar-neutrino prediction would require its normalization and spatial distribution, together with the channel-dependent neutrino yields, propagation through the Sun, and detector response. The explicitly calculated captured-pair contributions are nevertheless far below present neutrino-telescope sensitivity.

Conclusions

We have studied a minimal thermal realization of exothermic dark matter in which two nearly degenerate real scalars interact with the SM predominantly through an off-diagonal Higgs portal. For a representative benchmark with mϕ1≃100m_{\phi_1} \simeq100 GeV and Δm≃600\Delta m \simeq600 keV, the observed relic abundance requires a portal coupling g≃0.05g \simeq0.05, while the subsequent conversion history leaves approximately half of the dark matter in the excited state today.

In the ϕ2A→ϕ1A\phi_2 A \to\phi_1 A scattering the ϕ2−ϕ1\phi_2-\phi_1 mass splitting can produce nuclear recoils in the few-hundred-keV range without requiring particles from the extreme high-velocity tail of the halo. With the interaction strength fixed by the relic-density requirement, the benchmark gives an order-one efficiency-weighted event expectation over the available true-recoil efficiency range for the reported LZ exposure.

The same interaction gives rise to the radiative transition ϕ2→ϕ1γγ\phi_2 \to\phi_1\gamma\gamma, linking the recoil signal to a broad sub-MeV photon continuum. The predicted lifetime is sufficiently long to preserve the excited population and exceeds the lower bound recast from the released INTEGRAL/SPI spectrum. Present-day mixed annihilation also lies below the Fermi-LAT dwarf-galaxy and CMB limits; the margin in the dwarf-galaxy comparison is modest and already excludes part of the parameter space where ϕ1\phi_1 mass approaches half of the Higgs mass region.

Solar capture produces qualitatively different two-state dynamics. Exothermic scattering preferentially populates the ground state, while inefficient regeneration of the excited state suppresses mixed annihilation between captured particles. Consequently, the tree-level three-body channel ϕ1ϕ1→hffˉ\phi_1\phi_1 \to hf\bar{f} and the loop-induced channel ϕ1ϕ1→W+W−\phi_1\phi_1 \to W^+W^- exceed the captured mixed-state annihilation rate despite their much smaller rate coefficients. The resulting annihilation rates are well below the IceCube bounds.

The central prediction of the above explanation of the LZ event is the connection between a thermally determined two-state population, target-dependent high-energy nuclear recoils, and a photon continuum signal in indirect DM searches, such as in COSI. Higher-statistics high-recoil measurements in xenon and complementary nuclear targets, together with sub-MeV gamma-ray observations, could thus test the proposed scenario.

Acknowledgments. SR acknowledges many useful discussions at Fermilab. The work of SR and JZ at the University of Cincinnati has been supported by DOE Grants Nos. DE-SC0011784 and DE-SC0026301 and NSF Grant No. OAC-2411215. JZ also acknowledges support in part from NSF Grants Nos. OAC-2103889 and OAC-2417682.

References

  1. [1]D. S. Akerib et al. (LZ), Search for dark matter particle interactions in an extended nuclear recoil energy window with the LUX-ZEPLIN (LZ) experiment (2026), arXiv:2609.02823 [hep-ex].arxiv.org/abs/2609.02823
  2. [2]D. Tucker-Smith and N. Weiner, Phys. Rev. D 64, 043502 (2001), arXiv:hep-ph/0101138.
  3. [3]J. Fan and M. Reece, Higgsino Above the Sea of Fog (2026), arXiv:2609.01504 [hep-ph].arxiv.org/abs/2609.01504
  4. [4]K. Freese and D. P. Theodosiopoulos, Higgsino Dark Matter Interpretation of the LUX-ZEPLIN 248 keV Nuclear-Recoil Event (2026), arXiv:2609.01583 [hep-ph].arxiv.org/abs/2609.01583
  5. [5]L. Wu, Y. Zhang, and B. Zhu, TeV Higgsino Dark Matter from LZ Nuclear Recoil to Fermi-LAT Gamma Rays (2026), arXiv:2609.01590 [hep-ph].arxiv.org/abs/2609.01590
  6. [6]W. Yin, A PQ-Symmetric High-Scale SUSY Interpretation of the LZ High-Energy Recoil (2026), arXiv:2609.01892 [hep-ph].arxiv.org/abs/2609.01892
  7. [7]Y. Nomura, Dark Matter as the Z_2 Partner of the Standard Model Higgs Boson (2026), arXiv:2609.02505 [hep-ph].arxiv.org/abs/2609.02505
  8. [8]M. Di Mauro, Dark Matter at the Kinematic Edge: Interpreting the 248 keV LZ Nuclear-Recoil Candidate (2026), arXiv:2609.02608 [hep-ph].arxiv.org/abs/2609.02608
  9. [9]L. Visinelli, A Peccei-Quinn Origin for Inelastic Electroweak Dark Matter after LUX-ZEPLIN (2026), arXiv:2609.02807 [hep-ph].arxiv.org/abs/2609.02807
  10. [10]J. Smirnov, S. Griffith, and J. F. Beacom, Inelastic Signatures of Electroweak Dark Matter (2026), arXiv:2609.04144 [hep-ph].
  11. [11]X. Du and F. Wang, TeV Higgsino Interpretation of the LZ High-Recoil Event with Intermediate-Scale Electroweak Gauginos (2026), arXiv:2609.04163 [hep-ph].arxiv.org/abs/2609.04163
  12. [12]S. Bisal, J. Cao, and F. Li, Higgsino Dark Matter Interpretation of the LZ High-Recoil Event in the GNMSSM with TeV-Scale Gauginos (2026), arXiv:2609.07811 [hep-ph].arxiv.org/abs/2609.07811
  13. [13]K. Cheung, S. K. Kang, and R. Kumar, From LUX-ZEPLIN to Colliders: Probing Higgsino Dark Matter (2026), arXiv:2609.08712 [hep-ph].arxiv.org/abs/2609.08712
  14. [14]K. Langhoff, Heavy Higgsino Interpretation of the LZ Event (2026), arXiv:2609.09835 [hep-ph].
  15. [15]D. Frolovsky and S. V. Ketov, Higgsino dark matter in the Starobinsky supergravity with the MSSM in light of the LUX-ZEPLIN event (2026), arXiv:2609.11241 [hep-ph].arxiv.org/abs/2609.11241
  16. [16]P. K. Paul, S. K. Sahoo, N. Sahu, and S. Sharma, Resurrecting Electroweak Dark Matter via Type-II Seesaw in light of recent LZ Event (2026), arXiv:2609.22063 [hep-ph].arxiv.org/abs/2609.22063
  17. [17]I. I. Khan, A. Muhammad, G. Mustafa, F. Atamurotov, A. Abdujabbarov, and M. Khan, LZ-Motivated Pseudo-Dirac Higgsinos in the Supersymmetric 331 Model with a Supersymmetric SU(6) GUT Model (2026), arXiv:2609.23691 [hep-ph].arxiv.org/abs/2609.23691
  18. [18]J. Unwin, Heavy Higgsino Dark Matter at the Scale of the Vanishing Higgs Quartic (2026), arXiv:2609.28740 [hep-ph].arxiv.org/abs/2609.28740
  19. [19]H. Murayama and B. Noether, Heavy Inert Doublet Recoils LZ and IceCube (2026), arXiv:2609.28819 [hep-ph].arxiv.org/abs/2609.28819
  20. [20]S. Roy, P. Schwaller, and C. E. M. Wagner, Interpreting the High-Recoil LUX-ZEPLIN Event with Bino-/Singlino-like and Higgsino Dark Matter (2026), arXiv:2609.40234 [hep-ph].arxiv.org/abs/2609.40234
  21. [21]S. J. Lee and T. Youn, The LZ inelastic splitting as a probe of anomaly mediation (2026), arXiv:2610.13930 [hep-ph].arxiv.org/abs/2610.13930
  22. [22]L. Su, J. M. Yang, and W.-N. Yang, Inelastic Dark Matter Signature at High Recoil Energy in LUX-ZEPLIN and CRESST (2026), arXiv:2609.01475 [hep-ph].arxiv.org/abs/2609.01475
  23. [23]K. Yamashita, Inelastic Dark Photon Dark Matter for the LUX-ZEPLIN High-Recoil Event and the Galactic Halo Gamma-Ray Excess (2026), arXiv:2609.02868 [hep-ph].arxiv.org/abs/2609.02868
  24. [24]H. M. Lee, Inelastic dark matter and baryon flavor symmetry in light of LUX-ZEPLIN (LZ) experiment (2026), arXiv:2609.06171 [hep-ph].arxiv.org/abs/2609.06171
  25. [25]L. Wang and Y. Xiao, The Inert Doublet Model of Dark Matter and the LUX-ZEPLIN High-Recoil Event (2026), arXiv:2609.06571 [hep-ph].arxiv.org/abs/2609.06571
  26. [26]P. Das, B. Karmakar, S. Mahapatra, and P. K. Paul, Inelastic Self-interacting Dark Matter and LUX-ZEPLIN 248 keV Event in a Dirac Modular Inverse Seesaw (2026), arXiv:2609.06825 [hep-ph].arxiv.org/abs/2609.06825
  27. [27]N. Okada and O. Seto, Inelastic B = L scalar dark matter and the LUX-ZEPLIN event (2026), arXiv:2609.06909 [hep-ph].arxiv.org/abs/2609.06909
  28. [28]W. Ahmed and G. K. Leontaris, A Dark-Dimension Origin of Geometric Inelastic Dark Matter: The LUX-ZEPLIN High-Recoil Event and Multi-Target Tests (2026), arXiv:2609.07138 [hep-ph].arxiv.org/abs/2609.07138
  29. [29]X.-Y. Du, W. Huang, and K. Xie, Pseudo-Dirac Inelastic Dark Matter in the Leptophobic U(1)B Model: Confronting the LUX-ZEPLIN High-Recoil Event with Collider Searches (2026), arXiv:2609.07225 [hep-ph].arxiv.org/abs/2609.07225
  30. [30]D. Bandyopadhyay, D. Borah, and P. Borah, LZ nuclear recoil event from inelastic singlet-doublet scalar dark matter (2026), arXiv:2609.07451 [hep-ph].arxiv.org/abs/2609.07451
  31. [31]D. Borah, S. K. Sahoo, N. Sahu, and S. Sharma, Inelastic Singlet-Doublet Fermion Dark Matter in light of the 248 keV LZ event (2026), arXiv:2609.07800 [hep-ph].arxiv.org/abs/2609.07800
  32. [32]G.-W. Yuan, B. Zhang, W.-Y. Cao, L. Feng, and R. Yang, ALP-mediated inelastic dark matter and the LUX-ZEPLIN high-recoil candidate event LZ230616 (2026), arXiv:2609.08893 [hep-ph].arxiv.org/abs/2609.08893
  33. [33]P. Zhu, G. Dalla Valle Garcia, X.-G. Wang, A. W. Thomas, and M. J. White, Endothermic dark matter with a light dark photon and the LUX-ZEPLIN high-energy nuclear-recoil candidate (2026), arXiv:2609.09015 [hep-ph].arxiv.org/abs/2609.09015
  34. [34]P. Asadi, A. Batz, P. J. Fox, S. D. Homiller, and G. D. Kribs, For Whom the Xenon Recoils: Magnetic Inelastic Dark Baryons (2026), arXiv:2609.09107 [hep-ph].arxiv.org/abs/2609.09107
  35. [35]S. J. Lee and T. Youn, Mixing-suppressed inelastic dark matter: a minimal model for the LZ 248 keV event (2026), arXiv:2609.09138 [hep-ph].arxiv.org/abs/2609.09138
  36. [36]Y. He, Transition magnetic-dipole dark matter and the LZ230616 high-recoil candidate (2026), arXiv:2609.10453 [hep-ph].arxiv.org/abs/2609.10453
  37. [37]Z.-T. Fan, H.-J. He, Y.-C. Wang, and Y. Zhao, Inelastic Dark Matter and High-Energy Recoil Signatures in LZ (2026), arXiv:2609.10491 [hep-ph].arxiv.org/abs/2609.10491
  38. [38]R. Kumar and H. K. Prajapati, Generalized Chiral U(1)B−L with Inelastic Scalar Dark Matter for the LZ 248 keV Event (2026), arXiv:2609.10827 [hep-ph].arxiv.org/abs/2609.10827
  39. [39]H. Okada, Y. Shigekami, and J.-J. Wu, Can a minimal radiative seesaw explain the LZ 248 keV event? (2026), arXiv:2609.13038 [hep-ph].arxiv.org/abs/2609.13038
  40. [40]D. Cabo-Almeida, F. Costa, D. Feitiera, and V. Oliveira, A Freeze-In Interpretation of the LZ High-Energy Nuclear Recoil Event (2026), arXiv:2609.13130 [hep-ph].
  41. [41]P. Borah, S. Mahapatra, and N. Nath, Inelastic Dark Matter at LZ from Radiative Dirac Neutrino Mass Paradigm (2026), arXiv:2609.15027 [hep-ph].arxiv.org/abs/2609.15027
  42. [42]A. Das and T. Nomura, Effect of inelastic scalar dark matter in hidden U(1) scenario after the LZ nuclear recoil (2026), arXiv:2609.15600 [hep-ph].arxiv.org/abs/2609.15600
  43. [43]X.-G. He, X. Hong, and S. Jeesun, Hydrophilic inelastic freeze-in dark matter in q1 − q2 gauge extension and the high energy LZ event (2026), arXiv:2609.15714 [hep-ph].arxiv.org/abs/2609.15714
  44. [44]J. Lian and J. M. Yang, Explain the LZ High-Energy Recoil Event with Inelastic Sneutrino Dark Matter in Supersymmetry (2026), arXiv:2609.15742 [hep-ph].arxiv.org/abs/2609.15742
  45. [45]H. An, F. Gao, J. Liu, M. Liu, and C. Xu, Cosmological Constrained Axion-Portal Inelastic Dark Matter for the LZ Event (2026), arXiv:2609.17412 [hep-ph].arxiv.org/abs/2609.17412
  46. [46]N. Nagata and T. T. Yanagida, Asymmetric Inelastic Dark Matter and the LUX-ZEPLIN event (2026), arXiv:2609.18564 [hep-ph].arxiv.org/abs/2609.18564
  47. [47]N. Okada and D. Raut, Endothermic Z’-Portal Dark Matter: LZ-LHC Complementarity (2026), arXiv:2609.21011 [hep-ph].arxiv.org/abs/2609.21011
  48. [48]W. Ahmed and G. K. Leontaris, Kaluza–Klein Dark-Photon Mediation of Inelastic Dark Matter at LUX-ZEPLIN (2026), arXiv:2609.22739 [hep-ph].
  49. [49]J. Tang and P.-H. Gu, Inelastic dark matter mediated by natural dark photon (2026), arXiv:2609.28415 [hep-ph].
  50. [50]K. Langhoff and H. Xiao, Sneaky Neutrino Scattering at LZ (2026), arXiv:2609.28616 [hep-ph].
  51. [51]D. Choudhury, J. Das, and D. Sachdeva, Inelastic Dark Matter from a Higher-Dimensional Operator: The LUX-ZEPLIN High-Recoil Event and Thermal Relic (2026), arXiv:2610.00645 [hep-ph].
  52. [52]A. Capolupo, G. Pisacane, and R. Serao, Solar capture and annihilation of fermionic and scalar dark matter motivated by the LUX-ZEPLIN high-recoil candidate (2026), arXiv:2609.28218 [hep-ph].
  53. [53]J. Unwin, Axion Portal Dark Matter and the LUX-ZEPLIN High-Recoil Event (2026), arXiv:2609.04186 [hep-ph].
  54. [54]G. Arcadi, M. di Mauro, A. Djouadi, and F. Queiroz, A possible interpretation of the LUX-ZEPLIN recoil event in the 2HD+a scenario (2026), arXiv:2609.17196 [hep-ph].
  55. [55]B. De, The 248 keV LZ Recoil: A Possible Hint of Non-SM-Like Quark Yukawa Couplings with a Scalar-Portal Dark Matter (2026), arXiv:2609.23096 [hep-ph].
  56. [56]I. Khan, S. Capozziello, G. Mustafa, F. Botorov, A. Abdujabbarov, F. Atamurotov, and P. Channuiе, Elastic toroidal vector dark matter through a dark photon in the LUX-ZEPLIN high recoil window (2026), arXiv:2609.25114 [hep-ph].
  57. [57]F. Elahi and P. Schwaller, A Vector-Like Lepton Interpretation of the High-Energy Nuclear Recoil Candidate in LUX-ZEPLIN (2026), arXiv:2609.08998 [hep-ph].
  58. [58]J.-H. Liang, Z. Liu, V. Q. Tran, and Y. Xu, LZ Nuclear-Recoil Excess from Boosted Light Magnetic Dipole-dark Matter (2026), arXiv:2609.06756 [hep-ph].
  59. [59]K. Kannike, M. Raidal, and A. Strumia, Boosted dark particles and the LZ nuclear recoil event (2026), arXiv:2609.07742 [hep-ph].
  60. [60]H. Alhazmi, D. Kim, K. Kong, J.-C. Park, and S. Shin, High-Energy Nuclear Recoils from Boosted Dark Matter for the LZ 248-keV Event: Beyond the Halo-Dependent High-Velocity Tail (2026), arXiv:2609.06890 [hep-ph].
  61. [61]M. Heikinheimo and N. Zimmermann, Cosmic ray boosted dark matter with momentum dependent interactions can explain the LZ 248 keV event (2026), arXiv:2609.11600 [hep-ph].
  62. [62]S. Mahapatra and P. K. Paul, Boosted or Inelastic? Discriminating Interpretations of the LZ 248 keV Event (2026), arXiv:2609.14799 [hep-ph].
  63. [63]B. Chauhan, S. Sahasrabuddhe, and M. Sen, A closer look at the LZ 248 keV event through the lens of cosmic-ray boosted dark matter (2026), arXiv:2609.24982 [hep-ph].
  64. [64]C. McCabe, Seasonal dark matter from the LUX-ZEPLIN high-energy event (2026), arXiv:2609.04181 [hep-ph].
  65. [65]C. A. J. O’Hare, The high-velocity dark matter halo of the Milky Way in light of the LZ 248 keV event (2026), arXiv:2609.21444 [hep-ph].
  66. [66]Y. Lou and C.-T. Lu, Fermionic Dark Matter Absorption and the High-Energy Event in LUX-ZEPLIN (2026), arXiv:2609.01592 [hep-ph].
  67. [67]M. Aghaie and A. Strumia, Neutron disappearance and the LZ nuclear recoil event (2026), arXiv:2609.09037 [hep-ph].
  68. [68]J. Lee, F. Takahashi, and Y.-D. Tsai, Nuclear Recoils from Invisible Neutron-Pair Annihilation and the LZ event (2026), arXiv:2609.12045 [hep-ph].
  69. [69]P. Uttayarat, J. Julio, and R. Primulando, DM induced neutron disappearance as the origin of the LZ nuclear recoil event (2026), arXiv:2609.15933 [hep-ph].
  70. [70]S. Jeesun and A. Majumdar, Atmospheric neutrino up-scattering explanation of LZ 2026 excess (2026), arXiv:2609.04185 [hep-ph].
  71. [71]A. Chattaraj, A. Majumdar, D. K. Papoulias, and R. Srivastava, Can Elastic Neutrino Scattering Account for the LZ230616 Event? (2026), arXiv:2609.10504 [hep-ph].
  72. [72]V. Brdar and D. S. Chattopadhyay, A Narrow Neutrino Window for the LZ Event (2026), arXiv:2609.30255 [hep-ph].
  73. [73]J. Sheng and K. Zhang, The LUX-ZEPLIN Event as Hyperfine Spectroscopy of Composite Dark Matter (2026), arXiv:2609.23477 [hep-ph].
  74. [74]H. Jung and S. C. Park, Dark Diffraction at LZ from a Screened Neutral Composite Baryon (2026), arXiv:2609.25723 [hep-ph].
  75. [75]F. Sannino and J. Turner, Interpreting the LZ 248 keV Event using Dark QCD (2026), arXiv:2609.24988 [hep-ph].
  76. [76]V. S. H. Lee and L. Randall, A Warped Extra Dimensional Candidate for the LZ 248 keV Event (2026), arXiv:2609.09136 [hep-ph].
  77. [77]B. Batell, M. Pospelov, and A. Ritz, Phys. Rev. D 79, 115019 (2009), arXiv:0903.3396 [hep-ph].
  78. [78]D. P. Finkbeiner, T. R. Slatyer, N. Weiner, and I. Yavin, JCAP 09, 037 (2018), arXiv:0903.1037 [hep-ph].
  79. [79]P. W. Graham, R. Harnik, S. Rajendran, and P. Saraswat, Phys. Rev. D 82, 063512 (2010), arXiv:1004.0937 [hep-ph].
  80. [80]J. B. Dent and J. L. Newstead, Exothermic and Endothermic Inelastic Dark Matter Interpretations at LZ: Sideband Constraints and Future Prospects (2026), arXiv:2609.04673 [hep-ph].
  81. [81]C. H. de Lima, Exothermic Dark Matter at LZ (2026), arXiv:2609.05204 [hep-ph].arxiv.org/abs/2609.05204
  82. [82]H. Baer and V. Barger, Exothermic dark matter and the 248 keV nuclear recoil in LUX-ZEPLIN (2026), arXiv:2609.06153 [hep-ph].
  83. [83]M. Yang, Q.-f. Wu, K.-K. Duan, Z.-Q. Xia, Y.-L. S. Tsai, and Y.-Z. Fan, A Common ~55 GeV Inelastic Dark Matter Origin of the LZ, Fermi-LAT, and AMS-02 Excesses and Paleo-Detector Prospects (2026), arXiv:2609.06640 [hep-ph].
  84. [84]C.-Y. Xing, Galactic Endothermic Production and Exothermic Detection of Excited Dark Matter: Implications for LUX-ZEPLIN (2026), arXiv:2609.17935 [hep-ph].
  85. [85]C. Gemmell, D. Hooper, and G. Krnjaic, A Simple Dark Matter Model to Explain the LZ Event and Galactic Center Excess (2026), arXiv:2609.26570 [hep-ph].
  86. [86]V. Oliveira, The LZ event with a two-state dark matter halo (2026), arXiv:2609.35650 [hep-ph].
  87. [87]Y. Jho, S. Kim, and S. C. Park, Decaying Dark Matter for the Exothermic Recoil at LZ (2026), arXiv:2609.37561 [hep-ph].
  88. [88]R. Sheikh, T. Toma, and K. Tsumura, Exothermic nuclear recoils from thermal quasi-pNGB dark matter (2026), arXiv:2610.09516 [hep-ph].arxiv.org/abs/2610.09516
  89. [89]K. Deka, Exothermic dark-photon dark matter and the LZ high-recoil event: finite-reheating constraints (2026), arXiv:2610.18974 [hep-ph].arxiv.org/abs/2610.18974
  90. [90]J. Kim, J. Kim, and P. Ko, Exothermic Dark Matter with a Dark Higgs: The LZ 248 keV Event and the Galactic Center Excess (2026), arXiv:2610.25423 [hep-ph].arxiv.org/abs/2610.25423
  91. [91]V. Silveira and A. Zee, Phys. Lett. B 161, 136 (1985).
  92. [92]J. McDonald, Phys. Rev. D 50, 3637 (1994), arXiv:hep-ph/0702143.
  93. [93]C. P. Burgess, M. Pospelov, and T. ter Veldhuis, Nucl. Phys. B 619, 709 (2001), arXiv:hep-ph/0011335.DOI
  94. [94]K. Ghorbani and H. Ghorbani, Phys. Rev. D 93, 055012 (2016), arXiv:1501.00206 [hep-ph].arxiv.org/abs/1501.00206
  95. [95]J. A. Casas, D. G. Cerdeño, J. M. Moreno, and J. Quilis, JHEP 05, 036, arXiv:1701.08134 [hep-ph].
  96. [96]B. Díaz Sáez, J. Lahiri, and K. Möhlring, JCAP 10, 001, arXiv:2404.19057 [hep-ph].
  97. [97]J. A. Casas, G. A. Gómez Vargas, J. M. Moreno, J. Quilis, and R. Ruiz de Austri, JCAP 06, 031, arXiv:1711.10957 [hep-ph].
  98. [98]J. Guo, Y. He, J. Liu, and X.-P. Wang, JHEP 04, 024, arXiv:2211.01164 [hep-ph].
  99. [99]D. Hooper, G. Krnjaic, D. Rocha, and S. Roy, Phys. Rev. D 113, 115009 (2026), arXiv:2507.22975 [hep-ph].
  100. [100]S. Roy, Phys. Rev. D 113, 055040 (2026), arXiv:2509.19982 [hep-ph].DOI
  101. [101]M. Dutra, J. P. Neto, and C. Siqueira, Sub-MeV gamma-ray lines from an endothermic interpretation of the LZ recoil candidate (2026), arXiv:2609.33952 [hep-ph].arxiv.org/abs/2609.33952
  102. [102]J. P. Roques et al., Astron. Astrophys. 411, L91 (2003), arXiv:astro-ph/0310793.DOI
  103. [103]J. A. Tomsick et al., PoS ICRC2023, 745 (2023), arXiv:2308.12362 [astro-ph.HE].arxiv.org/abs/2308.12362
  104. [104]M. Pospelov and H. Ramani, Strong Constraints on Higgsino Dark Matter from Solar Capture (2026), arXiv:2609.02775 [hep-ph].arxiv.org/abs/2609.02775
  105. [105]D. Bose et al., Not so good vs for Higgsino dark matter as LZ excess: stringent limits from Super-Kamiokande and IceCube (2026), arXiv:2609.07807 [hep-ph].arxiv.org/abs/2609.07807
  106. [106]M. Di Mauro and H. Shaikh, Solar Capture Tests of Inelastic Dark Matter after the LZ High-Recoil Event (2026), arXiv:2609.06760 [hep-ph].
  107. [107]T. T. Q. Nguyen, T. Linden, and D. Hooper, Solar Neutrino Constraints on Inelastic Dark Matter Scattering in Light of Recent LUX-ZEPLIN Observations (2026), arXiv:2609.11833 [hep-ph].arxiv.org/abs/2609.11833
  108. [108]A. Ghosh, I. Chavez, and C. Kelso, Confronting the Higgsino Interpretation of the LZ Event with Astrophysical Uncertainties and the Solar Capture Constraints (2026), arXiv:2609.15321 [hep-ph].arxiv.org/abs/2609.15321
  109. [109]J. P. Luengas, M. Pospelov, and H. Ramani, Constraints on Complex Electroweak WIMPs from Solar Capture (2026), arXiv:2609.38316 [hep-ph].arxiv.org/abs/2609.38316
  110. [110]J. M. Cline, K. Kainulainen, P. Scott, and C. Weniger, Phys. Rev. D 88, 055025 (2013), [Erratum: Phys.Rev.D 92, 039906 (2015)], arXiv:1306.4710 [hep-ph].DOI
  111. [111]G. Arcadi, M. Dutra, P. Ghosh, M. Lindner, Y. Mambrini, M. Pierre, S. Profumo, and F. S. Queiroz, Eur. Phys. J. C 78, 203 (2018), arXiv:1703.07364 [hep-ph].DOI
  112. [112]P. Ghosh, T. Ghosh, and S. Roy, JHEP 10, 057, arXiv:2211.15640 [hep-ph].arxiv.org/abs/2211.15640
  113. [113]S. Roy, Phys. Rev. D 111, 015037 (2025), arXiv:2212.11230 [hep-ph].arxiv.org/abs/2212.11230
  114. [114]K. Griest and D. Seckel, Phys. Rev. D 43, 3191 (1991).
  115. [115]A. L. Fitzpatrick, W. Haxton, E. Katz, N. Lubbers, and Y. Xu, JCAP 02, 004, arXiv:1203.3542 [hep-ph].DOI
  116. [116]N. Anand, A. L. Fitzpatrick, and W. C. Haxton, Phys. Rev. C 89, 065501 (2014), arXiv:1308.6288 [hep-ph].
  117. [117]J. B. Dent, L. M. Krauss, J. L. Newstead, and S. Sabharwal, Phys. Rev. D 92, 063515 (2015), arXiv:1505.03117 [hep-ph].DOI
  118. [118]R. Catena, K. Fridell, and M. B. Krauss, JHEP 08, 030, arXiv:1907.02910 [hep-ph].DOI
  119. [119]B. J. Kavanagh and T. D. P. Edwards, WIMpy_NREFT v1.2 [computer software] (2024).
  120. [120]C. McCabe, JCAP 02, 027, arXiv:1312.1355 [astro-ph.CO].
  121. [121]F. Calore, A. Dekker, P. D. Serpico, and T. Siegert, Constraints on light decaying dark matter candidates from 16 yr of INTEGRAL/SPI observations (2023), [Erratum: Mon.Not.Roy.Astron.Soc. 538, (2025), Erratum: Mon.Not.Roy.Astron.Soc. 547, stago45 (2026)], arXiv:2209.06299 [hep-ph].DOI
  122. [122]T. Siegert, J. Berteaud, F. Calore, P. D. Serpico, and C. Weinberger, Astron. Astrophys. 660, A130 (2022), arXiv:2202.04574 [astro-ph.HE].DOI
  123. [123]A. McDaniel, M. Ajello, C. M. Karwin, M. Di Mauro, A. Drlica-Wagner, and M. A. Sánchez-Conde, Phys. Rev. D 109, 063024 (2024), arXiv:2311.04982 [astro-ph.HE].arxiv.org/abs/2311.04982
  124. [124]R. H. Helm, Phys. Rev. 104, 1466 (1956).
  125. [125]J. D. Lewin and P. F. Smith, Astropart. Phys. 6, 87 (1996).
  126. [126]M. Hoferichter, P. Klos, J. Menéndez, and A. Schwenk, Phys. Rev. D 94, 063505 (2016), arXiv:1605.08043 [hep-ph].DOI
  127. [127]M. Hoferichter, P. Klos, J. Menéndez, and A. Schwenk, Phys. Rev. D 99, 055031 (2019), arXiv:1812.05617 [hep-ph].DOI
  128. [128]P. W. Graham, J. P. Luengas, H. Ramani, and S. S. Y. Wong, Testing the inelastic dark matter interpretation of the LZ event with heavy-element experiments (2026), arXiv:2610.21041 [hep-ph].arxiv.org/abs/2610.21041
  129. [129]N. Bozorgnia, S. Contreras, G. B. Gelmini, A. C. Kamaha, J. Reynoso-Cordova, and Y. Xu, LMC-Perturbed LZ Dark Matter Landscape (2026), arXiv:2609.36184 [hep-ph].arxiv.org/abs/2609.36184
  130. [130]A. Ghosh, A. Ibarra, T. Mondal, and B. Mukhopadhyaya, JCAP 01, 011, arXiv:1909.13292 [hep-ph].
  131. [131]W. J. Marciano, C. Zhang, and S. Willenbrock, Phys. Rev. D 85, 013002 (2012), arXiv:1109.5304 [hep-ph].DOI
  132. [132]M. Herrero and R. A. Morales, Phys. Rev. D 102, 075040 (2020), arXiv:2005.03537 [hep-ph].DOI
  133. [133]F. Feruglio, G. Levati, and R. Ziegler, JHEP 05, 307, arXiv:2512.20695 [hep-ph].arxiv.org/abs/2512.20695
  134. [134]G. Krnjaic, D. McKeen, R. Mizuta, G. Mohlabeng, D. E. Morrissey, and D. Tuckler, Phys. Rev. D 112, 115039 (2025), arXiv:2509.19428 [hep-ph].arxiv.org/abs/2509.19428
  135. [135]J. F. Navarro, C. S. Frenk, and S. D. M. White, Astrophys. J. 462, 563 (1996), arXiv:astro-ph/9508025.
  136. [136]J. F. Navarro, C. S. Frenk, and S. D. M. White, Astrophys. J. 490, 493 (1997), arXiv:astro-ph/9611107.DOI
  137. [137]G. Vedrenne, J.-P. Roques, V. Schönfelder, P. Mandrou, G. G. Lichti, A. von Kienlin, B. Cordier, S. Schanne, J. Knödlseder, G. Skinner, P. Jean, F. Sanchez, P. Caraveo, B. Teegarden, P. von Ballmoos, L. Bouchet, P. Paul, J. Matteson, S. Boggs, C. Wunderer, P. Leleux, G. Weidenspointner, P. Durochoux, R. Diehl, A. Strong, M. Cassé, M. A. Clair, and Y. André, A&A 411, L63 (2003). (2023), arXiv:2303.01214 [hep-ex].
  138. [138]R. Anthony-Petersen et al., MeV dark matter decay forecasts for the COSI mission (2026), arXiv:2610.17217 [astro-ph.CO].
  139. [139]S. Abdollahi et al. (Fermi-LAT, HAWC, H.E.S.S., MAGIC, VERITAS), JCAP 03, 035, arXiv:2508.20229 [astro-ph.HE].arxiv.org/abs/2508.20229
  140. [140]T. R. Slatyer, Phys. Rev. D 93, 023527 (2016), arXiv:1506.03811 [hep-ph].DOI
  141. [141]N. Aghanim et al. (Planck), Astron. Astrophys. 641, A6 (2020), [Erratum: Astron.Astrophys. 652, C4 (2021)], arXiv:1807.06209 [astro-ph.CO].DOI
  142. [142]E. Calabrese et al. (Atacama Cosmology Telescope), JCAP 11, 063, arXiv:2503.14454 [astro-ph.CO].arxiv.org/abs/2503.14454
  143. [143]D. Curtin et al., Phys. Rev. D 90, 075004 (2014), arXiv:1312.4992 [hep-ph].DOI
  144. [144]G. Aad et al. (ATLAS), JHEP 08, 104, arXiv:2202.07953 [hep-ex].
  145. [145]A. Tumasyan et al. (CMS), Eur. Phys. J. C 83, 933 (2023), arXiv:2303.01214 [hep-ex].
  146. [146]G. Aad et al. (ATLAS), Phys. Lett. B 842, 137963 (2023), arXiv:2301.10731 [hep-ex].
  147. [147]J. J. Fox, R. Harnik, J. Kopp, and Y. Tsai, Phys. Rev. D 85, 056011 (2012), arXiv:1109.4398 [hep-ph].
  148. [148]G. Aad et al. (ATLAS), Phys. Rev. D 103, 112006 (2021), arXiv:2102.10874 [hep-ex].
  149. [149]A. Tumasyan et al. (CMS), JHEP 11, 153, arXiv:2107.13021 [hep-ex].
  150. [150]A. Tumasyan et al. (CMS), Phys. Rev. D 105, 092007 (2022), arXiv:2201.11585 [hep-ex].
  151. [151]D. Gonçalves, T. Han, and S. Mukhopadhyay, Phys. Rev. Lett. 120, 111801 (2018), [Erratum: Phys.Rev.Lett. 121, 079902 (2018)], arXiv:1710.02149 [hep-ph].DOI
  152. [152]D. Gonçalves, T. Han, and S. Mukhopadhyay, Phys. Rev. D 98, 015023 (2018), arXiv:1803.09751 [hep-ph].DOI
  153. [153]M. Blennow, S. Clementz, and J. Herrero-Garcia, Eur. Phys. J. C 78, 386 (2018), [Erratum: Eur.Phys.J.C 79, 407 (2019)], arXiv:1802.06880 [hep-ph].DOI
  154. [155]R. Abbasi et al. (IceCube), Search for High-Energy Neutrinos From the Sun Using Ten Years of IceCube Data (2025), arXiv:2507.08457 [hep-ex].arxiv.org/abs/2507.08457

Paper details

vlnc:W_s2cutczw
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC0-1.0
Usage

views · 0

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