Arxiv:2103.15834V2 [Hep-Ph] 31 Mar 2021

Direct detection of dark energy: the XENON1T excess and future prospects

Sunny Vagnozzi,1, 2, ∗ Luca Visinelli,3, 4, 5, † Philippe Brax,6, ‡ Anne-Christine Davis,7, 1, § and Jeremy Sakstein8, ¶ 1Kavli Institute for Cosmology (KICC), University of Cambridge, Madingley Road, Cambridge CB3 0HA, United Kingdom∗∗ 2Institute of Astronomy (IoA), University of Cambridge, Madingley Road, Cambridge CB3 0HA, United Kingdom 3Istituto Nazionale di Fisica Nucleare (INFN), Laboratori Nazionali di Frascati, C.P. 13, I-100044 Frascati, Italy∗∗ 4Tsung-Dao Lee Institute (TDLI), Shanghai Jiao Tong University, 200240 Shanghai, China 5Gravitation Astroparticle Physics Amsterdam (GRAPPA), University of Amsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands 6Institute de Physique The´orique(IPhT), Universit´eParis-Saclay, CNRS, CEA, F-91191, Gif-sur-Yvette Cedex, France 7Department of Applied Mathematics and Theoretical Physics (DAMTP), Center for Mathematical Sciences, University of Cambridge, CB3 0WA, United Kingdom 8Department of Physics & Astronomy, University of Hawai‘i, Watanabe Hall, 2505 Correa Road, Honolulu, HI, 96822, USA (Dated: August 24, 2021) We explore the prospects for direct detection of dark energy by current and upcoming terrestrial dark matter direct detection experiments. If dark energy is driven by a new light degree of freedom coupled to matter and photons then dark energy quanta are predicted to be produced in the Sun. These quanta free-stream towards Earth where they can interact with Standard Model particles in the detection chambers of direct detection experiments, presenting the possibility that these experiments could be used to test dark energy. Screening mechanisms, which suppress fifth forces associated with new light particles, and are a necessary feature of many dark energy models, prevent production processes from occurring in the core of the Sun, and similarly, in the cores of red giant, horizontal branch, and white dwarf stars. Instead, the coupling of dark energy to photons leads to production in the strong magnetic field of the solar tachocline via a mechanism analogous to the Primakoff process. This then allows for detectable signals on Earth while evading the strong constraints that would typically result from stellar probes of new light particles. As an example, we examine whether the electron recoil excess recently reported by the XENON1T collaboration can be explained by chameleon-screened dark energy, and find that such a model is preferred over the background-only hypothesis at the 2.0σ level, in a large range of parameter space not excluded by stellar (or other) probes. This raises the tantalizing possibility that XENON1T may have achieved the first direct detection of dark energy. Finally, we study the prospects for confirming this scenario using planned future detectors such as XENONnT, PandaX-4T, and LUX-ZEPLIN.

I. INTRODUCTION proposed as alternatives to the CC, within most of which the cosmic acceleration phenomenon is ascribed to addi- More than two decades after the discovery that the ex- tional (typically scalar) degrees of freedom, either in the pansion of the Universe is accelerating [1,2], the nature form of new fundamental particles and fields, or modifica- of the dark energy (DE) component driving this phe- tions to General Relativity (GR). Competing models pre- nomenon and making up 70% of the energy budget dict a variety of novel cosmological phenomena which will of the Universe remains a∼ mystery [3–7]. What is per- be actively sought for by the next generations of cosmo- haps the simplest theoretical DE candidate, a cosmologi- logical surveys, including but not limited to the Simons arXiv:2103.15834v3 [hep-ph] 23 Aug 2021 cal constant (CC) resulting from the collective zero-point Observatory [11, 12], CMB-S4 [13, 14], Euclid [15, 16], energy of quantum fields, suffers from a severe disagree- DESI [17], the Vera C. Rubin Observatory Legacy Sur- ment between its theoretical value suggested from quan- vey of Space and Time [18], and the Nancy Grace Roman tum field theory considerations, and the tiny value in- Space Telescope [19]. ferred from cosmological observations [8–10]. This stag- gering discrepancy goes under the name of cosmological In all but the simplest models, the scalar field leads to constant problem [9, 10]. Several DE models have been a modification of gravity where the associated new light scalar degree of freedom couples to matter [20–22]. Such theories predict the existence of fifth forces that are os-

∗ tensibly excluded by solar system tests of GR [23, 24]. To [email protected] circumvent this problem, realistic models inevitably need † [email protected] ‡ [email protected] to include some form of screening mechanism [6, 25–28]. § [email protected] Screening mechanisms dynamically suppress fifth forces ¶ [email protected] in the solar system without the need to tune the model ∗∗ S.V. and L.V. contributed equally to this work parameters, resorting to environmental effects in the case 2 of several (but not all) of these mechanisms. This then will lay out the formalism for computing the expected allows for strong deviations from GR on cosmological signal in the associated detectors following from the scat- scales, which could modify the growth of structure. The tering/absorption of DE scalars produced in the solar ability to account simultaneously for cosmic acceleration tachocline by nuclei or electrons (depending on the spe- and satisfy solar system tests of GR have made mod- cific detector details). We will assume that the DE scalar ified gravity (MG) theories with screening mechanisms possesses a coupling to photons, allowing for its produc- leading science targets for current and upcoming cosmo- tion from photons in the tachocline. We will also use logical surveys [29]. In addition to their indirect effects the environmental dependence of the couplings in the ef- on the cosmological background expansion and structure fective theory for the scalar fluctuations. We note that formation [30, 31],1 DE and MG theories with screening screening mechanisms of the chameleon, symmetron and mechanisms have the attractive feature that they are also Damour-Polyakov types show a clear dependence of these amenable to direct detection of their associated effects. coupling on the local density, i.e. on the environment. On The distinctive fifth forces they predict can be searched the other hand, screening mechanisms of the K-mouflage for in laboratory experiments [29, 37, 38] and astrophys- and Vainshtein types depend on more global features of ical environments [27, 28, 39–51], and their observation the matter distribution, such as the total mass of the would unequivocally confirm the hypothesis that DE is stars, with little or no dependence on the local mat- linked to a modification of GR. ter density [76]. In order to provide a well-studied and The purpose of this paper is to broaden the scope of well-constrained example, we will later specialise to the new physics accessible to terrestrial dark matter (DM) chameleon mechanism. Theories that screen using the direct detection experiments by exploring their potential chameleon mechanism [29, 71, 77, 78] do so by increasing to detect DE quanta produced in the Sun, and to open the mass of the DE scalar in dense environments, mak- up a new frontier for the direct detection of dark en- ing their fifth force too short-ranged to be relevant in the ergy. Screening mechanisms suppress the production of solar system or in terrestrial fifth force searches. DE quanta in the core of the Sun due to the large scalar As a case study, we will apply our methodology to the mass or weak coupling to matter, but they can be pro- XENON1T DM direct detection experiment [79], which duced in regions of strongly magnetised plasmas through recently reported a 3.3σ excess in the electron re- couplings of the scalar to photons. In particular, they can ∼ 2 coil data above their expected background, in the energy be produced in the solar tachocline, in a manner anal- range 1 7 keV [80]. The XENON1T collaboration finds ogous to the Primakoff process for axions [53, 54]. This that a fit− to the signal which includes a solar axion com- possibility opens up yet another exciting avenue for di- ponent is preferred over the background-only hypothesis rectly detecting DE fluctuations: DE particles produced at a significance of 3.4σ. Alternative explanations pro- in the solar tachocline could be observed in dark mat- posed by the collaboration include an additional tritium ter (DM) direct detection experiments. To date, searches background and a neutrino magnetic moment, both of for DE-like particles in direct detection experiments have which are preferred with a significance of 3.2σ. Unfor- mostly focused on axion-specific experiments such as the tunately, for the solar axion (and, to a minor but still CERN Axion Solar Telescope (CAST) [55–58] and the important extent, for the neutrino magnetic moment), Axion Dark Matter Experiment (ADMX) [59]. Other the parameters that best fit the XENON1T signal are detection techniques currently being considered are at strongly excluded by astrophysical observations of stel- different stages of realisation, or have started to gather lar evolution [81], particularly by constraints from the data [60–66]. For additional details, see Refs. [67–70]. cooling of horizontal branch stars in globular clusters, All well-known screening mechanisms can be clas- the cooling of white dwarfs, and the tip of the red gi- sified into chameleon [71], symmetron [72], Damour- ant branch I-band magnitude of globular clusters and Polyakov [73], K-mouflage [74], and Vainshtein [75] galaxies. The difficulties faced by these three scenarios screened theories. In order to encapsulate the diversity have spurred the proposal of various new physics models exhibited by screening mechanisms, we will use an effec- that might also be able to explain the XENON1T excess, tive theory for the fluctuations of the scalar about the with varying degree of motivation or plausibility: for a background profile due to the environment that includes list of works in this direction, we refer the reader to e.g. operators relevant for each mechanism, including generic Refs. [81–135]. couplings to the standard model (SM). Using this, we Applying our formalism to XENON1T, we find that chameleon-screened DE provides a fit to the signal of only slightly lower quality than the aforementioned sce- 1 See e.g. Refs. [32–36] for recent works studying the effect of di- narios, and is preferred over the background-only hypoth- rect couplings of DE to baryons on both local and cosmological esis at a significance of 2.0σ. The stellar bounds that observables. are debilitating for the solar axion interpretation of the 2 The tachocline is the turbulent shear layer located at the base of the solar convection zone, marking the transition between XENON1T signal are avoided by chameleon-screened DE the radiative interior and the differentially rotating convective theories due to the associated fifth force screening, as well zone [52]. The radial position of the tachocline is approximately as by the dependence of the mass of the scalar field on the 0.7R , with R the solar radius. local density of matter [53, 54]. In particular, the astro- 3 physical objects from which the strongest axion bounds signal in the XENON1T detector. All the codes associ- derive — red giant, horizontal branch, and white dwarf ated with this work are made publicly available online at stars — have cores whose density is larger than that of github.com/lucavisinelli/XENONCHAM. the solar tachocline by a factor of > 106. As a conse- quence, chameleon-screened DE scalars are too heavy to be thermally produced within these objects, and cannot II. DARK ENERGY EFFECTIVE THEORY IN act as a new source of energy loss. This raises the tan- THE SOLAR SYSTEM talizing possibility that XENON1T, originally devised to detect DM, may instead have detected dark energy, or Here, we work under the assumption that dark energy in any case a propagating scalar degree of freedom of a arises as a result of the cosmological dynamics of a sin- theory of modified gravity. gle scalar field, which we denote by ϕ. This scalar does The possibility that chameleon-screened scalars might not have to be a fundamental field. Instead, it could indeed be at the origin of the XENON1T signal is in- arise as a low energy degree of freedom emerging from dependently testable by a number of upcoming low- more involved dynamics at higher energy. For example, threshold DM direct detection experiments. These ex- the scalar could be the St¨uckelberg field of the broken periments are all sensitive to recoil energies of (100 eV) time-diffeomorphism symmetry of the cosmological back- or lower, and are expected to detect the chameleon-O ground [147]. Our further assumption is that this scalar induced signal at much higher statistical significance is involved in generating the acceleration of the Universe than XENON1T. Examples of experiments in this and at the same time couples to gravity and/or the SM in class, some of which make use of cryogenic semicon- a manner that is ghost- and pathology-free, as explicitly ductor detectors, include SuperCDMS [136], CDEX- realised e.g. in models of the Horndeski [148], beyond 10 [137], DAMIC [138], CRESST-III [139], PICO [140], Horndeski (GLPV) [149, 150], or Degenerate Higher- LUX-ZEPLIN [141], EDELWEISS [142], SENSEI [143], Order Scalar-Tensor (DHOST) classes of scalar-tensor PandaX-II [144], PandaX-4T [145], and of course theories [151, 152]. These theories are among the leading XENONnT [146] among others. We will demonstrate candidate DE theories accompanying a modification of that these experimental setups, often discussed in the gravity (see Refs. [3,6, 20–22, 31, 153] for more general context of searches for light (sub-GeV or even MeV) DM reviews concerning MG theories and cosmological tests scattering off nuclei or electrons, are equally well-placed thereof). to detect the signal of solar chameleons.3 The rest of this paper is organized as follows. Our Multi-messenger astronomy involving a gravitational effective theory for screened dark energy in the solar sys- wave (GW) and an associated optical counterpart can tem is presented and discussed in Sec.II. The specific be used to constrain the properties of theories of modi- chameleon model we specialise to in this work is pre- fied gravity. To date, there has been one such confirmed event. In 2017, the LIGO/Virgo collaboration observed sented in Sec.III, where we calculate its flux from produc- the gravitational wave event GW170817 [154], resulting tion in the solar tachocline, and detection cross-section in from a binary neutron star merger. Simultaneously, the DM direct detection experiments. In Sec.IV we discuss Fermi Gamma-ray Burst Monitor and the INTEGRAL the data analysis method we use to confront our model Anti-Coincidence Shield spectrometer detected the short against the XENON1T signal. In Sec.V we present the gamma-ray burst GRB170817A [155, 156], which was results following from this analysis. We critically dis- identified as being the electromagnetic counterpart to cuss these results and prospects for probing chameleons GW170817. The joint GW170817/GRB170817A detec- in current and future DM direct detection experiments tion restricts the speed of GWs to differ from the speed in Sec.VI. Finally, we draw our concluding remarks in 15 Sec.VII. In addition, our paper contains three technical of light by no more than one part in 10 , setting strong appendices. AppendixA is devoted to a more detailed constraints on theories of the Horndeski [157–163], be- discussion of the chameleon mechanism. AppendixB re- yond Horndeski [164, 165], and DHOST types [166– vises the production of chameleon-screened DE scalars 170], although these constraints are subject to important in the Sun, as well as the resulting flux and energy spec- caveats [171, 172]. Other important constraints on such trum of solar chameleons on Earth. AppendixC presents theories arise from potential instabilities in GW back- details concerning the computation of the cross-section grounds [173, 174], astrophysical bounds [27, 28, 175– for what we refer to as the “chameleo-electric effect”, a 177], and cosmological constraints [178–181]. Neverthe- process which is analogous to the photoelectric and axio- less, large regions of parameter space remain observa- electric effects for photons and axions respectively, and tionally and theoretically viable [182–186]. Since we will which is relevant for the computation of the expected work at the level of the effective theory in the solar sys- tem, our formalism is insensitive to the details of the fundamental theory responsible for driving cosmic accel- eration. We will thus assume that the aforementioned 3 In the following we will use the word chameleon for all chameleon- bounds are satisfied. The chameleon-screened dark en- screened scalars. When the original chameleon scalar with an ergy theories that we study in this paper predict that the inverse power potential is meant, this will be explicitly specified. speed of gravitational waves is luminal so the bounds are 4 automatically satisfied without the need to tune param- included this coupling with strength βγ . This term is of eters. Attempts to embed our more general solar system critical importance since it allows for the production of effective theory into a covariant model which can be ex- scalars in the solar tachocline. The couplings to mat- tended to cosmological scales should ensure that these ter arise from the Jordan frame metric, i.e. the metric bounds are satisfied. We defer this study to future work. coupling SM particles to the scalar when the graviton is We begin by expanding the field ϕ(~x,t) around some canonically normalised in the underlying covariant the- background value ϕ0(~x,t) as ϕ(~x,t) = ϕ0(~x,t) + φ(~x,t), ory: where φ(~x,t) is a spacetime-dependent perturbation. De-   J φ X pending on the nature of the fundamental theory, ϕ (~x,t) g = 1 + 2βi(ϕ0) + 2ci(ϕ0) gµν 0 µν M M 4 could either be the field in the cosmological background, Pl the field sourced by the Milky Way, or even different in di(ϕ0) +2 ∂µφ∂ν φ . (4) the Earth and the Sun. The latter scenario is realised by M 4 the chameleon screening mechanism. The relevant part From a quantum field theory perspective, it is unlikely of the action for studying solar phenomenology comprises that the couplings to each matter species are universal, two terms. The first is the quadratic part of a scalar hence our choice to treat the couplings βi, ci, and di field action with background-dependent kinetic and mass as being species-specific. The term multiplying gµν is terms referred to as the conformal factor and the term multi- Z   plying ∂µφ∂ν φ is referred to as the disformal factor. We 4 1 µν 1 2 2 S = d x√ g Z (ϕ0)∂µφ∂ν φ m (ϕ0)φ , therefore refer to the βis and cis as conformal couplings, − −2 − 2 and dis as disformal couplings (ci can also be referred to (1) as kinetic-conformal coupling). µν where Z is the kinetic matrix. We have not included The effective theory in Eq. (3) includes environmen- self-interaction terms in the form of a polynomial in tal variations of the various coupling constants via their φn, n 3 as they are not relevant to the scalar pro- ≥ dependence on the background field ϕ0, and therefore duction mechanism we will consider. In principle, each on the local matter density when the energy-momentum of the monomials in this expansion have ϕ0-dependent tensor of matter is dominated by non-relativistic species, couplings. The second relevant part of the action is the such as in the late-time Universe after matter-radiation coupling to matter and photons equality, or in astrophysical situations through the viri- Z h alised matter density. In particular, we have allowed the 4 φ µν Sφi = d x√ g βi Ti + kinetic matrix Z (ϕ ) to be non-diagonal and to depend − M 0 Pl on the background field. This structure typically emerges i µν ∂µϕ∂ν ϕ 1 µν ρσ from ghost-free higher-derivative couplings of the field to ci (ϕ0) Ti + Ti di µν (ϕ0)∂ρφ∂σφ , (2) M 4 M 4 curvature tensors, and gives rise to the Vainshtein [75] where MPl is the reduced Planck mass, M is the UV- and K-mouflage [74] screening mechanisms. We have also µν allowed the mass to be background field-dependent. This cutoff of the theory, Ti is the energy-momentum ten- µν gives rise to the chameleon mechanism [71, 77]. Note that sor for SM particle species i, with Ti = gµν Ti its µν m(ϕ0) is the mass of fluctuations about ϕ0. The mass trace, and with background-dependent tensors ci (ϕ0) of ϕ in the cosmological background is instead (H0), and dµνρσ(ϕ0). Here we assume for simplicity’s sake O µν µν µνρσ µρ νσ so that ϕ can act as a DE scalar. Finally, we have al- that ci (ϕ0) η and di (ϕ0) η η , where the proportionality∝ coefficients are assumed∝ to be density- lowed the coupling constants βi, ci, and di to be field- dependent and species-dependent constants. As a result dependent too. This field-dependence is utilised by the we shall concentrate on the following action symmetron [72] and dilaton [188] mechanisms. In this work, we shall focus on theories that utilise Z  µν 4 1 µν the chameleon mechanism, and therefore set Z (ϕ0) = S = d x√ g Z (ϕ0)∂µφ∂ν φ µν − −2 η , while taking βi, ci, and di to be background field- independent. We also introduce the energy scales Mi 1 2 2 φ µν 1/4 ≡ m (ϕ0)φ βγ Fµν F M/d , which depend on the species being considered −2 − M i Pl (i). Applying these simplifications to the action in  # X φ X di(ϕ0) µν Eq. (3), the effective theory considered is: + βi(ϕ0) Ti +ci(ϕ0) 4 Ti + 4 Ti ∂µφ∂ν φ , MPl M M Z  i 4 1 µ S = d x√ g ∂µφ∂ φ (3) − −2

µν µν 1 2 2 φ µν where X = g ∂µφ∂ν φ and F is the photon field- m (ϕ0)φ βγ Fµν F strength tensor.− The sum in Eq. (3) includes photons, −2 − MPl  # but since their energy-momentum tensor is traceless they X φ X 1 µν do not contribute to the first two terms. A direct coupling + βi Ti + ci 4 Ti + 4 Ti ∂µφ∂ν φ . MPl M M to photons through their field-strength tensor can arise i i through quantum anomalies [187], and we have therefore (5) 5

The mechanism by which DE is produced in the Sun does not lead to a substantial improvement in the fit to and either scatters or is absorbed in DM direct detec- the XENON1T signal (which would otherwise warrant its tion experiments is the following. The Sun is necessarily inclusion as a free parameter), as the disformal detection screened, implying that the mass of the scalar in the Sun channel dominates over the conformal one(s), for reasons 3 is m = m(ϕ ) > 10 H0 [189, 190]. In practice, the high we will discuss in Sec.IV below. The choice of setting density of the Sun with respect to the cosmological back- ci = 0 for the purposes of this work should be viewed 29 ground (ρ /ρc 10 ) implies that m is in fact much as a simplifying working assumption, which sets a min- heavier than this∼ deep inside the Sun. The exact value imal phenomenologically working model: (re)-including is model-dependent, as it depends on the exact density ci would not change our results, nor the goodness of the dependence of the scalar’s mass. The high mass prevents fit. the efficient production of chameleons via Compton or bremsstrahlung processes in the Sun’s core via a Boltz- mann suppression. However, the direct coupling to pho- tons allows for production in the magnetic field of the III. MODEL tachocline via a mechanism analogous to the Primakoff process for axions [191]: we review the chameleon pro- duction process in AppendixB. The relevant operators A. Theoretical considerations for this production process are: As it stands, our effective theory is still too general Z  1 φ S 4x√ g F F µν β F F µν to calculate the production and detection processes be- φγ = d µν γ µν 2 − − 4 − MPl cause we need to determine the free functions m (φ0) = 2 2 1  m (~x) = m (ρ), βi(ρ), and di(ρ). There are two possi- + T µν ∂ φ∂ φ , (6) M 4 γ µ ν bilities for fixing the spatial-dependence of the scalar’s γ mass. The first is to parameterise our ignorance by as- 1/4 suming a functional form for its density and constraining where M = M/d is the energy scale related to the γ γ the parameters associated with this choice. The second disformal coupling to photons. As chameleons are pro- is to provide an explicit model and thereby to calculate duced in the tachocline and not in the core, we need to the spatial-dependence explicitly. We opt for the sec- impose that the mass of the chameleon in the core be ond choice for three reasons. First, it is not guaran- larger than the local temperature. This can be achieved teed that an arbitrary fitting function will reproduce the using the density-dependence of the mass, as the ratio dynamics of any fundamental theory, so it is not clear between the densities in the core and in the tachocline is what information is lost by making such a choice. Sec- around two orders of magnitude. ond, chameleon models are well-constrained by labora- Once produced, solar chameleons free-stream out of tory and astrophysical probes, so choosing a well-studied the Sun, with a fraction passing through the Earth, and model allows us to explore complementarity with these an even smaller fraction through the chambers of DM bounds, and to determine the feasibility of our scenario direct detection experiments. In these chambers the mass by identifying the existence of regions of parameter space is mostly determined by the detector’s geometry [71, 77] where our model can simultaneously satisfy these bounds and is typically small (m m , where m is the mass DC e e and successfully explain the XENON1T signal. Finally, of the electron). Chameleons can therefore be treated as chameleons have β and d constant, so we can exemplify massless particles in the chambers of DM direct detection i i our scenario using a simple minimal model. experiments for all intents and purposes. Chameleons passing through the detector chamber can scatter off or Chameleon theories are subject to a no-go theo- be absorbed by the particles utilised for the detection, rem [190] that excludes the possibility of self-acceleration via couplings of the form defined strictly as acceleration in the Jordan frame but not the Einstein frame in the complete absence of any Z  φ X 1  cosmological constant. Dark energy scenarios where S 4x√ g β T c T T µν ∂ φ∂ φ . φi = d i i + i 4 i + 4 i µ ν the acceleration is driven by the scalar potential, i.e. a − MPl M Mi (7) quintessence-like explanation, are not excluded by this These couplings give rise to what we refer to as the theorem, but require a tuning of an overall additive con- “chameleo-electric effect”, and correspondingly to elec- stant. Of course, such a tuning is also present in self- tron recoils in the (keV) range. Similar interactions acceleration scenarios as this additive constant is arbi- with neutrons and protonsO will instead give rise to atomic trarily set to zero. The generalized couplings (disformal recoils, which we will not explicitly consider in this paper. and kinetic-conformal, XT ) that we consider here are In the subsequent analysis, we will neglect the coupling additional potential caveats to the no-go theorem since ci, which controls the strength of the kinetic-conformal they were not considered when deriving it [192]. Other coupling XTi. We find that this coupling has an effect caveats are discussed in Ref. [190]. similar to that of the conformal coupling bi at the level As discussed in more detail in AppendixA, the density- of detection. Moreover, from a statistical perspective, it dependence of the chameleon’s mass arises because its 6 dynamics are governed by an effective potential where d = 1 A.U. is the Earth-Sun distance, and Rt 0.7R is the tachocline radial coordinate. The flux of∼   βmφ chameleons produced in the Sun, dΦ/dω, is calculated in Veff (φ) = V (φ) + ρ exp , (8) MPl AppendixB, see Eq. (B13). In principle, one could also consider production from where ρ is the density of the matter species coupled to the toroidal magnetic modes deeper within the Sun. How- chameleon, βm is the strength of such coupling, and V (φ) ever, we note that there are significant uncertainties as- is the bare potential, which would govern the dynamics sociated to the strength and profile of these modes [206– if the field were not coupled to matter. A common model 208]. The equipartition value for the large-scale toroidal for the bare potential is the power-law chameleon [71, 77, magnetic field due to shearing is 1 T [206–208]. As 78] leading to a density-dependent mass at the minimum ∼ we later assume a strength Bt = 30 T for the tachocline φmin of the effective potential magnetic field, we expect the associated contribution to 2 the chameleon flux relative to the contribution we con- 2 d Veff (φ) m = sidered to be suppressed by a factor (1/30)2 (10−3), φ dφ2 ≈ O φ=φmin and to appear at higher energies than those of interest for 2+n   1+n the XENON1T excess. While the strength of these modes βm ρ = n(1 + n)Λ4+n , (9) may be amplified locally by up to (102) their equiparti- nM Λ4+n Pl tion values in the convection zoneO to form sunspots [206], where Λ is an energy scale and n is a power-law index the significant uncertainties at play prevent us from fully (V (φ) φ−n). Note that both n > 0 (inverse power- quantifying the impact of these processes on our re- law) and∝ n < 0 (power-law) can lead to the chameleon sults. We thus conservatively choose to neglect the ef- behaviour provided n = 1, 2, or an odd negative inte- fect of toroidal magnetic modes on chameleon produc- 6 − − ger. We also assume that βmφ/MPl 1, in order for the tion deeper within the Sun, noting that these could lead  excursion of the chameleon field not to exceed MPl/βm. to subdominant features in the event rate at higher en- Note that the swampland conjectures (see Ref. [193]) lead ergies, but deferring a full study to follow-up work. to a lower bound on βm & cV/ρ, where c is a constant of We have not considered other more “standard” produc- order unity [194]. tion mechanisms, which in the parameter space spanned are subdominant. For example, chameleons could be pro- duced through so-called ABC reactions (atomic recom- B. Production in the Sun bination and deexcitations, bremsstrahlung, and Comp- ton), in a similar fashion to axions [209–211]. How- Chameleons can be resonantly produced in a dense ever, by virtue of the chameleon mass being density- magnetised plasma when the chameleon mass matches dependent, we can always find large regions of param- the plasma frequency of the environment. This process, eter space where these processes are kinematically dis- governed by the chameleon-photon coupling in Eq. (6), favoured (in particular by adjusting the energy scale Λ). occurs in the Sun within a narrow shell whose location de- As discussed at the end of AppendixA, this can be 2 2 pends on the chameleon rest mass [54, 195]. Chameleon achieved by requiring that meff (ρcore) & Tcore, where 2 2 2 2 production can also occur through non-resonant pro- meff mφ ωPl, with ωPl the plasma frequency squared ≡ − 3 cesses, which occur in all magnetised regions inside the given by Eq. (B6), ρcore 150 g/cm the Sun’s core den- '7 Sun. Here we adopt the non-resonant production mech- sity, and Tcore 1.5 10 K the Sun’s core temperature. ' × anism and consider a magnetic field profile B = B(r), Typically we expect that the mass of the chameleons where r is the radial coordinate. 4 For the solar model, scale like m(ρ) ρα where α = 3/4 for chameleons with ' we adopt the profiles described in Ref. [198], which has n = 1. This implies that the mass deep inside the core also been used to derive the formula for the Primakoff increases typically by two orders of magnitude compared flux used by the XENON1T collaboration, see Ref. [199]. to the mass in the tachocline. If this condition is satis- We also note that there is some disagreement in the field fied, production of solar chameleons in the deeper regions between different solar models, see e.g. Refs. [200–205]. of the Sun is kinematically forbidden, with the overall The resulting differential flux per unit energy of solar flux being dominated by Primakoff-like production in the chameleons on Earth, resulting from isotropic production tachocline. in the Sun, is given by From Eq. (9), we see that the previous condition trans- lates into constraints on βm, Λ, and n. Fixing βm and 2 dΦEarth πRt dΦ n > 0, we find that this condition sets an upper limit = 2 , (10) dω 4πd dω on the allowed value of Λ. Later in our analysis we will 2 consider βm 10 and n = 1, since we find that the ' event rate in XENON1T is mostly sensitive to βγ and Me, and only weakly sensitive to βm, n, Λ, and Mγ . For 4 Note that we do not consider couplings of chameleons to longi- this choice of βm and n, we find that production of solar tudinal plasmons, as recently considered in Refs. [196, 197]. chameleons in the Sun’s core is kinematically forbidden 7 as long as Λ . 1 µeV. We remark again that fixing Λ to where the expression for the flux at Earth is given in other values has no appreciable effect on the XENON1T Eq. (10), and where the number of atoms per ton of 27 −1 event rate, and hence on our results. xenon is NXe = 4.6 10 ton , so that the differential event rate is expressed× in units of ton−1 yr−1 keV−1. We have appended the subscript th to the differential event C. Chameleon detection rate per production energy in Eq. (11) to stress that this is a theoretical event rate, which depends on the under- To reach the XENON1T detector and leave detectable lying chameleon parameters through the dependence of imprints, solar chameleons with energies ω & (keV) dΦEarth/dω and σφe in Eq. (12) on these parameters. need to propagate through various dense mediaO un- Comparing the theoretical event rate against the event scathed. In general, chameleons with incoming energy rate measured by XENON1T will allow us to set con- ω will be able to traverse a dense barrier of energy ρ pro- straints on the underlying chameleon parameters. vided ω & m(ρ). Let us focus on the benchmark point While we focus on the XENON1T apparatus, we stress we discussed above and which we consider throughout that the results in AppendixB and AppendixC, from 2 the paper, where βm 10 ,Λ . 1 µeV, and n = 1, so which we derive Eqs. (11)-(12), are more broadly appli- that m ρ3/4. We thus' need to ensure that chameleons cable. In particular, they can be applied to future DM make it∝ through the densest material along their path. direct detection experiments, which we discuss in more The highest density involved in the problem is that of detail in Sec.VIB. Note, moreover, that because we used lead, which the XENON1T detector is partially made of, an effective action, the expression for the cross-section we 3 and whose density is ρPb 10 g/cm . For the above compute, as well as its derivation, are general results that ∼ parameters, we find m(ρPb) 0.6 keV, meaning that can be applied to any mass and coupling and for different ∼ chameleons with energies ω & 0.6 keV are able to reach detector setups. Our goals are now to explore whether the XENON1T detector. This is sufficient to ensure that solar chameleons are able to account for at least part chameleons are able to fit the XENON1T excess, which of the observed low-recoil excess observed in XENON1T occurs at energies higher than this cut-off. Moreover, as and, if so, to determine benchmark examples of solar rocks and the tachocline have densities respectively one chameleon parameters which provide an adequate fit to and two orders of magnitude lower than that of lead, the XENON1T signal. this also ensures that chameleons are able to escape the As discussed in Sec.II, we do not include the kinetic- tachocline and propagate through the rock which makes conformal ci coupling (i.e. the term proportional to XTi), up Gran Sasso (being mostly made of limestone, with since we find a posteriori that including this operator density ρ 3 g/cm3). does not lead to a substantial improvement in fit to the In the XENON1T∼ detector, solar chameleons can be XENON1T signal, which would instead be required to absorbed by electrons via the chameleo-electric effect. warrant its inclusion as a free parameter. If we were to This is the chameleon analogue of the photoelectric and include this term, a fit to XENON1T data would strongly axio-electric effects for photons and axions respectively. prefer setting the associated coupling to zero, leading to The cross-section for the chameleo-electric effect in DM the one-parameter extension not being preferred from direct detection experiments, σφe, is computed in Ap- a statistical point of view. This may not be the case pendixC, see in particular Eq. (C30). The resulting for other DE models but it is for the specific case of differential event rate per unit production energy ω for chameleon DE. chameleon absorption by electrons in the XENON1T de- The physical reason why this coupling worsens the fit tector is given by: to the XENON1T signal is that the associated cross- section does not scale fast enough with energy ω. This   Z dR dR0(ωR) leads to the corresponding peak in the resulting event = (ω) Θ(ω ωR) dωR , (11) dω th dωR − rate being below the 2 keV required to explain the XENON1T excess. On≈ the other hand, the disformal where Θ(ω) is the energy resolution of the detector and coupling leads to additional powers of ω in the associated (ω) is the XENON1T detection efficiency, given in Fig. 2 cross-section, thereby moving the peak to the correct po- 5 of Ref. [80]. The “raw” differential event rate per unit sition to explain the XENON1T excess. production energy of chameleons in the XENON1T de- tector, i.e. not taking into account energy resolution and detection efficiency effects, is given by: IV. ANALYSIS

dR0(ω) dΦEarth = NXe σφe , (12) dω dω In principle, the parameter space describing pro- duction in the Sun and subsequent detection in the XENON1T detector of solar chameleons is six- dimensional, and spanned by the following parameters: 5 As described below Eq. (1) in Ref. [80], the efficiency does not the coupling to matter (in this case electrons) βe βm, enter the integral in the convolution function. the scale governing the disformal coupling to electrons≡ 8

1/4 (hereafter “electron disformal scale”) Me = M/de , the We can view βeff as being the chameleon equivalent of the coupling to photons βγ , the scale governing the disformal product gaγ gae for the case of solar axions produced via coupling to photons (hereafter “photon disformal scale”) the Primakoff effect in the Sun and detected via the axio- 1/4 Mγ = M/dγ , and finally the energy scale Λ and power- electric effect in the XENON1T detector. In particular, law index n describing the chameleon self-interaction po- the expected event rate in the XENON1T detector is 2 tential as given in Eq. (A12). proportional to βeff , as we derive in Eq. (C31) in the To simplify our analysis, we set Λ = 1 µeV and n = 1. appropriate limit discussed in the Appendix. We now proceed to analyze the XENON1T measure- As explained earlier, requiring Λ . (µeV) ensures that the chameleon’s effective mass is sufficientlyO high in the ments to verify whether solar chameleons can fit these core of the Sun so that production of chameleons through measurements, and whether our previous expectations the usual Compton and bremsstrahlung mechanisms is are met. We perform a Bayesian statistical analysis to kinematically suppressed. On the other hand, n = 1 constrain the four chameleon parameters, which we col- corresponds to the best-studied chameleon model, with lectively denote by θ βγ ,Mγ , βe ,Me . The likeli- hood (θ d) to observe≡ the { data d given a} certain set of bounds typically only being reported for this specific L | choice [37]. These extensive studies have excluded a large model parameters θ is region of parameter space [29], and it is thus of interest to  χ2(θ, d) (θ d) = exp , (14) explore whether XENON1T is able to probe part of the L | − 2 remaining parameter space of this model. In any case, we 2 have explicitly verified that fixing Λ and n to other val- where the χ function entering the likelihood is given by 2 ues affects chameleon production and the resulting event " dR dR # X ( )th(θ)+B0 ( )meas(d) rate well below the %-level, and thus has no appreciable χ2(θ, d)= dω − dω ,(15) σ2 effects on our results. i i β M β These choices leave us with 4 parameters: e, e, γ , with the sum being performed over the energy bins ωi at and Mγ . However, we anticipate that XENON1T will be which XENON1T measure their event rates. In Eq. (15), β M mostly sensitive to γ and e, and very weakly sensitive (dR/dω)th denotes the theoretical event rate given by β M to e and γ , for the following reasons. Firstly, we ex- Eq. (11), (dR/dω)meas denotes the rate as measured pect the disformal detection channel to dominate over the by XENON1T (black re-binned datapoints of Fig. 4 in conformal one, as the former scales with a higher power of Ref. [80]), and B0 denotes the background model (red energy than the latter, see Eq. (C30). This feature moves curve of Fig. 4 in Ref. [80]), which is itself a function of the peak in the event rate towards higher energies, bet- energy. The XENON1T background model is described ter fitting the XENON1T excess. Therefore, detection in in more detail in Sec. IIIB of Ref. [80] (see in particu- XENON1T is mostly controlled by the electron disformal lar their Table 1 and Fig. 3), and includes contributions scale Me rather than the coupling βe, with the associated from ten different components, ranging from 214Pb to so- 8 cross-section scaling as 1/Me , see Eq. (C30). Second, if lar neutrinos. We refer the reader to Ref. [80] for more a we require that Mγ (keV) so that Primakoff produc-  O detailed discussion of B0. tion in horizontal branch stars does not dominate over For simplicity, we only consider the first 16 bins, in the neutrino losses [212], production in the Sun will predom- recoil energy range 1.5 keV ω 16.5 keV. We do not 6 . R . inantly proceed through the conformal channel. As a consider bins beyond the 16th for two reasons: result, production will mostly be controlled by the pho- 1. the theoretical solar chameleon event rate drops ton coupling βγ rather than the photon disformal scale quickly beyond the third bin, partly due to the lim- Mγ . In particular, the associated production flux scales 2 ited width of the differential chameleon flux (see as βγ , see Eq. (B13). Finally, the expected event rate in XENON1T, which is the only quantity we can directly Fig.2 below), and to the effects of energy resolu- compare to observations, depends only on the product tion and detector efficiency, entering in Eq. (11) of the production flux and the detection cross-section, as through Θ(ω) and (ω); can be clearly seen in Eq. (12). This product scales as 2. the measured rate in the bins beyond the third is 2 8 βγ /Me . We therefore expect that the XENON1T mea- highly consistent with the background model B0, surements will predominantly constrain the following pa- and therefore does not call for new physics expla- rameter combination, which we denote by βeff , and refer nations: the only exceptions are a few anomalous to as the effective coupling: bins (the 17th, the 20th, the 24th, the 26th, and the th 4 29 bins respectively), which none of the proposed keV  βeff βγ . (13) theoretical models (including the solar axions, neu- ≡ Me trino magnetic moment, and tritium explanations invoked by the XENON1T collaboration) have been able to explain.7 6 Note that the bounds on Mγ derived in Ref. [212] do not directly apply to our scenario, as they were derived assuming that the scalar is massless. 7 Note that also the 11th and 14th bins lie ≈ 1σ above the back- 9

Of these 16 bins, the second and third deviate the most To sample the posterior distribution of the chameleon from the XENON1T background model B0, and therefore parameters we use Markov Chain Monte Carlo (MCMC) contribute the most to the excess. methods. We make use of the cosmological sampler We impose flat priors on log βγ [0; 11], Montepython3.3 [217], configured to act as a generic 10 ∈ log (Mγ /keV) [0; 25], log βe [1; 2], and sampler. The convergence of the generated MCMC 10 ∈ 10 ∈ log10(Me/keV) [0; 25]. For βγ , the upper prior edge is chains is monitored through the Gelman-Rubin parame- motivated by the∈ latest results from the Kinetic Weakly ter R 1 [218], and we require R 1 < 0.01 for the chains Interacting Slim Particles (KWISP) detector on the to be− considered converged. − CAST axion search experiment at CERN, which finds Finally, we also quantify the significance of the pref- 11 βγ < 10 [58], whereas the lower prior edge is arbi- erence (if any) for the solar chameleon model over the trary (we have checked that extending it to lower values background-only model B0. We do so by adopting the does not qualitatively affect our final results). Due to the same test statistic q(s) used by the XENON1T collabora- weak sensitivity of the XENON1T measurements to βe, tion, with s symbolically denoting the signal parameters, we have chosen a narrow prior for this parameter. We see Eq. (17) in Ref. [80]. This is essentially a profile have explicitly verified that extending the prior further log-likelihood test statistic. The statistical significance has no effects on our results. Similarly, fixing βe (for of the preference (if any) for the solar chameleon model 2 instance to βe = 10 ) would also have no effect on our is then determined by q(0), i.e. comparing the difference results, see later discussion below Eq. (16). in goodness-of-fit of the best-fit solar chameleon model We allow the photon and electron disformal scales to relative to a B0-only fit. span the range between the keV scale and the Planck scale. It is worth noting that limits on the disformal scale M obtained from collider searches, indicating M & V. RESULTS (100 GeV) as for instance in Refs. [213–216] (including O works from one of us), only apply to the chameleon-quark We now analyze the XENON1T data using the disformal coupling scale, and not to the scale governing methodology, priors, and likelihood described in Sec- the coupling to photons and electrons. The strongest tionIV. We perform an MCMC run on the four- bound on the scalar-photon disformal coupling comes dimensional parameter space spanned by βγ , βe, Mγ , and from demanding that Primakoff production of scalars in Me. This MCMC run confirms our earlier expectation horizontal branch stars does not dominate over neutrino that we are only sensitive to the parameter combination losses, and the strongest bound on the scalar-electron dis- of βγ and Me given by βeff in Eq. (13), while not being formal coupling similarly derives from demanding that sensitive to βe and Mγ . We shall discuss the obtained losses from Compton and bremsstrahlung production do constraints on βeff later on. not significantly alter the properties of these objects. In A very important result of our analysis is that we are both cases, the bounds impose Me,Mγ (0.1 GeV) as & O able to identify regions/benchmark points in parameter derived by one of us in Ref. [212]. Note, however, that space which provide a good fit to the XENON1T sig- these bounds only apply in the limit where the scalar’s nal (to be quantified shortly). One such benchmark ex- mass can be neglected. This is not the case in our sce- ample is presented in Fig.1, where the blue curve is nario since we impose meff (ϕ ) > T , where T is the obtained by fixing the chameleon parameters to βe = core temperature of the Sun. The bounds derived in 2 3.6 10 10 , Me = 10 keV, βγ = 10 , Mγ = 1000 TeV, Ref. [212] therefore do not directly apply to our case. Λ = 1 µeV, and n = 1. The black data points de- Moreover, as we will discuss in Sec.VIA, we expect pro- note the XENON1T measurements, and the grey curve is duction within these stellar objects to be kinematically the XENON1T background model B0 (the measurements suppressed for the benchmark point in parameter space and background are taken from Ref. [80]). Overall, the we considered. We consequently conservatively choose resulting fit to the XENON1T signal is good, with a best- to allow Me and Mγ to be as low as (keV), but not 2 O fit χmin 13.2 for 16 datapoints. Moreover, we find an any lower. As discussed at the end of AppendixB, for ' χ2 χ2 . improvement in fit of min B0 4 0 with respect to Me ,Mγ . (keV), the back-reaction effect of the dis- − ' − 2 O a B0-only fit, with the latter delivering χ 17.2. Un- formal coupling on the scalar field profile in the Sun can B0 der this signal model, and using the previously' discussed become non-negligible, resulting in the break-down of the q(s) profile log-likelihood test statistic, a B -only fit to resulting production flux computation. The (keV) scale 0 the signal is rejected at a significance of 2.0σ. which determines whether or not this effectO is negligible is set by the maximal temperature reached within the The detection rate in Eq. (12) depends on the spec- Sun. trum of chameleons on Earth resulting from Primakoff- like production in the tachocline, as given by Eq. (10). In Fig.2 we show this solar chameleon flux on Earth using the same benchmark choice of chameleon parameters as

ground model B0. The solar axion explanation of the low-energy in Fig.1. Note that the production flux remains high excess improves the fit to the 14th bin through the contribution even at energies below 1 keV, an aspect which will have 57 of Fe axions, see Fig. 7b in Ref. [80]. important consequences for our subsequent discussion. 10

108 ]

100 1 7 − 10

80 keV

keV) 6 1 10 − × s

2 5 yr 60 − 10 ×

[cm 4 (t 10

/ 40

B0 /dω 103 20 B0+Solar chameleons Earth Events B0+Solar axions Φ 102 XENON1T measurements d 0 0 2 4 6 8 10 12 14 16 1 2 3 4 5 ωR [keV] ω [keV]

FIG. 1. Benchmark example of a solar chameleon fit to the FIG. 2. Solar chameleon differential flux per unit energy on XENON1T signal (event rate in units of ton−1 yr−1 keV−1 Earth, in units of cm−2 s−1 keV−1, resulting from isotropic versus recoil energy in units of keV), representative of the production in the solar tachocline. The chameleon parameters 2 3.6 10 best achievable fit within this scenario. The chameleon pa- are fixed to βe = 10 , Me = 10 keV, βγ = 10 , Mγ = 2 3.6 10 rameters are fixed to βe = 10 , Me = 10 keV, βγ = 10 , 1000 TeV, Λ = 1 µeV, and n = 1. This is an example of set Mγ = 1000 TeV, Λ = 1 µeV, and n = 1. The black of parameters which maximises the resulting quality of fit to data points denote the XENON1T measurements, the grey the XENON1T signal, as shown in Fig.1 (blue curve). curve is the XENON1T background model B0, and the blue curve gives the event rate predicted by the solar chameleon model with parameters fixed to the aforementioned values. through different mechanisms: these include ABC reac- The fit improves with respect to a background-only fit by tions [209–211], the Primakoff effect [219], and the 57Fe ∆χ2 ' −4.0, with a total χ2 = 13.2 for 16 datapoints. Vari- ous combinations of the chameleon parameters can lead to a transition line [220]. These differences appear clearly in fit of identical quality to the benchmark fit shown here, which Fig.1, where the solar chameleon model (blue line) shows one peak at ω 2 keV, while the solar axion model (red is almost entirely governed by the combination of βγ and Me ≈ given by βeff in Eq. (13), see also Eq. (16). Also shown for line) shows three distinct peaks at three different energies comparison (red curve) is the best-fit result for the signal de- corresponding to the different production mechanisms. rived from the solar axion model discussed by the XENON1T These features result in a better fit of the solar axion collaboration in Ref. [80], see Fig. 7b therein. model to the third, fourth, fifth, sixth, and 14th bins, improving the overall fit and significance at which the model is preferred over the background. In particular, Although the fit in Fig.1 is visually adequate, some ABC reactions are responsible for the considerably im- features require further investigation. In fact, focusing proved fit to the bins from the third to the sixth. These on the second and third bins, i.e. the two bins where the differences between the solar chameleon and solar axion measured event rate deviates the most from B0, the qual- scenarios can be distinguished in future DM direct detec- ity of the fit is only slightly worse that of the solar axion, tion experiments by their different spectra. We further neutrino magnetic moment, and tritium explanations in- discuss the prospects for detection of solar chameleons in voked by XENON1T (see Figs. 7a-7d in Ref. [80]). For future DM direct detection experiments in SectionVIB. comparison, the predicted signal resulting from the best- Although the measured recoil rate in the first bin fit solar axion model is given by the red curve in Fig.1. is perfectly in line with the background model, the However, as discussed in Ref. [80], the background model chameleon model overshoots this first point by just over is rejected at a significance of more than 3σ within all one standard deviation. The reason is that the solar these signal models, much stronger than our 2.0σ. chameleon differential flux on Earth remains apprecia- The paramount difference between the solar axion and ble at low energies (keV), see Fig.2. This results in  O solar chameleon models is the available production chan- an increase in the event rate over B0 at lower energies, nels. The effects of this are evident by comparing the blue leading to a poorer fit to the first bin, which is perfectly and red curves in Fig.1 (for solar chameleons and solar in line with B0 and would not call for any additional axions respectively). As discussed in the Introduction, contributions to the fit. screening prevents the production of chameleons in the The reason why the signal beyond the third bin is core through bremsstrahlung, Primakoff, and Compton completely dominated by the background model B0 effects and proceeds via the Primakoff effect [54] in the (which at that point is in very good agreement with the tachocline. Axions, on the other hand, are not affected XENON1T measurements) is that the production flux, by screening, and are produced in the core of the Sun after peaking at an energy slightly below ω = 1 keV, 11 quickly drops for higher energies (see Fig.2). In other plies that (dR/dω)th +B0 B0 for all i in the numerator words, in the sum in the numerator of Eq. (15), one has of Eq. (15), regardless of' the choice of model parame- (dR/dω)th + B0 B0 (dR/dω)meas for i 4, and ters. This behaviour leads to an extended “plateau” in ' ' ≥ therefore (dR/dω)th + B0 (dR/dω)meas 1, resulting parameter space where the likelihood is completely flat in small contributions to the− χ2 for i 4 in the numera- with χ2 17.2. Along the plateau, the quality of the tor of Eq. (15). For this reason, computing≥ the χ2 over all resulting' fit to the XENON1T signal is not only iden- 16 bins might be a misleading goodness-of-fit metric. For tical for any choice of parameters, but also identical to the same choice of benchmark parameters as mentioned the quality of a B0-only fit. The goodness-of-fit along above, and in Fig.1, the contribution of the first three the plateau is worse by only ∆χ2 = +4.0 with respect to bins to the total χ2 is 6.5, which better quantifies the the solar chameleon best-fit (for which χ2 13.2, see ' min ' imperfect fit to the first three bins visible in Fig.1. Fig.1). This explains the shallow plateau in the log 10 βeff Finally, let us discuss how XENON1T constrains the posterior for large negative values of log10 βeff . effective coupling βeff , given by Eq. (13). We treat βeff as We find that the best fit to the XENON1T signal oc- −4.5 a derived parameter whose posterior distribution we infer curs for a value of βeff 10 . This implies ' from our MCMC samples of the four fundamental param-   Me eters. The normalized posterior distribution for log βeff 10 log10 βγ 4 log10 4.5 . (16) is given in Fig.3, which shows that XENON1T is indeed − keV ' − able to set meaningful constraints on this parameter. Because of the shape of the log10 βeff posterior shown in Fig.3 (the tail can extend indefinitely to large negative values of log βeff ), we do not quote a confidence interval 1.0 10 on log10 βeff . Rather, we use Eq. (16) as indicative of what combinations of βγ and Me lead to a good fit to the 0.8 XENON1T signal, of quality identical to that shown in Fig.1. We have also verified our earlier expectation that

0.6 βe and Mγ play a negligible role in our analysis. We have max 2 fixed these parameters to βe = 10 and Mγ = 1000 TeV

P/P respectively, and repeated the analysis. Doing so, we find 0.4 essentially the same posterior for log10 βeff as shown in Fig.3, which was instead previously obtained by varying

0.2 all four parameters. Demanding that solar chameleons fit the XENON1T signal, and combining the relation in Eq. (16) and the 0.0 11 10 8 6 4 2 0 upper limit of βγ < 10 from CAST [58] sets an upper − log− β = log− (β ) 4− log (M /keV)− 10 eff 10 γ − 10 e limit on Me . 10 MeV. In other words, if Me & 10 MeV, for any allowed value of βγ the event rate will be com- FIG. 3. Normalized posterior distribution for log β , the pletely dominated by B0, and we will find ourselves along 10 eff β combination of the photon coupling βγ and the electron dis- the plateau for large negative values of log10 eff in Fig.3. formal scale Me to which the XENON1T measurements are This upper limit on Me is ostensibly in contradiction with most sensitive, see Eqs. (13,16). A value of log10 βeff ' −4.5 the lower limit one obtains by demanding that horizontal is required to provide a good fit to the XENON1T signal, of branch stars are not affected by the disformal coupling, quality identical to that shown in Fig.1. which sets Me & 0.1 GeV, as found by one of us [212]. However, we note that the limit obtained in Ref. [212] is The shape of the log10 βeff posterior, with a tail as not applicable to our case, as it was obtained assuming log βeff 0 and a plateau for large negative val- that the scalar is massless. This is clearly not the case 10 → ues of log10 βeff , is worth explaining further. Moving in our scenario, given the constraints we have imposed log βeff 0 means that either βγ is being raised or on the chameleon mass within the Sun, which in turn 10 → Me is being lowered. In other words, either or both suppresses production, making it so that the bounds ob- solar chameleon production and detection are being en- tained in Ref. [212] do not apply. This is generically true hanced. Enhancing the signal sufficiently will make the for all of the relevant limits in Ref. [212], which were all total event rate too large compared to the XENON1T derived assuming a massless scalar. measurements, and hence increasingly unlikely, leading The upper limit Me . 10 MeV is nominally in tension to the tail in the log βeff posterior as log βeff 0. with LEP constraints, which require Me 3 GeV as de- 10 10 → & On the other hand, moving log10 βeff towards large neg- rived by one of us in Ref. [212]. However, care must be ative values means that either or both solar chameleon taken with this bound, as it again was derived assum- production and detection are being suppressed. Within ing a massless chameleon. A proper re-evaluation of the this regime, the resulting event rate would be too low bounds of Ref. [212] would require a dedicated analysis, compared to the XENON1T detector background, and for instance determining the field profile in the LEP pipe therefore the total signal is dominated by B0. This im- simultaneously including conformal and disformal cou- 12 plings, a calculation which is well beyond the scope of this from an incoherent sum of solar-produced axions and paper. Therefore, in continuing our exciting program for chameleons. This could be beneficial for the solar ax- the direct detection of dark energy quanta, re-evaluating ion model, as it might be able to alleviate the tension the LEP bounds is of paramount importance, and will be in the axion-electron coupling between the XENON1T the subject of follow-up work. 8 results and astrophysical constraints [81]. Within this Let us summarize the main findings of this Section: scenario, the lower-energy excess would be fitted by so- lar chameleons, plus a smaller contribution from solar 1. The expected event rate in the XENON1T detector axions, allowing for a lower value of gae. On the other is sensitive to the combination of the photon cou- hand, the higher-energy end would be fit by the solar ax- β M pling γ and the electron disformal scale e given ion via its couplings to photons and nucleons. We defer β by the effective coupling eff in Eq. (13). the study of this interesting hybrid possibility to follow- 2. On the other hand, XENON1T has only very weak up work. sensitivity to βe, Mγ , Λ, and n.

3. A value of log10 βeff 4.5 is required to fit the VI. DISCUSSION XENON1T signal well.' − This leads to an improve- χ2 . B B ment in of 4 0 compared to a 0-only fit ( 0 In this section, we discuss other experimental bounds .'σ excluded at 2 0 ), and a quality of fit as shown in on our scenario, stellar bounds in particular, and the Fig.1. In no region of parameter space can the prospects for detecting dark energy in current and quality of the fit to the XENON1T signal be better planned dark matter direct detection experiments. than that shown in Fig.1. As CAST requires βγ < 1011, demanding that solar chameleons explain the XENON1T signal and therefore log βeff 4.5 10 ' − A. Stellar cooling constraints implies Me . 10 MeV. 4. Given the previous points 1. and 2., we can iden- The stellar bounds on the axion-electron and axion- tify various combinations of the chameleon param- photon coupling are debilitating for the solar axion in- eters which lead to a fit of identical quality to that terpretation of the XENON1T excess [81, 222]. The shown in Fig.1. There is therefore a large window situation with chameleons is different. The paramount of parameter space which can account for part of difference between the two models is the environment- the XENON1T excess, while remaining consistent dependence of the chameleon’s mass. This ensures that with laboratory and astrophysical tests. chameleons are not produced in the Sun’s core since Compton and bremsstrahlung processes are kinemati- 5. With respect to the solar axion model invoked by cally suppressed, leading to a severe Boltzmann suppres- XENON1T, the statistical significance of the pref- sion. Instead, chameleons are produced in the strong erence for the solar chameleon model is lower be- magnetic field of the solar tachocline. Similarly, the cores cause of the poorer fits to the first bin, as well of red giant, horizontal branch (HB), and white dwarf as to a few bins at higher energies (due to the stars are denser than the Sun’s by a significant factor, larger available number of production channels for implying an even stronger suppression in these objects. solar axions). However, we stress that the solar It is possible that some of these objects may have strong chameleon model is not excluded by other bounds, magnetic fields [223], but without dedicated individual unlike the solar axion model. observations and detailed stellar modelling, it is not pos- Our overall conclusion is that solar chameleons are able sible to derive quantitative constraints on chameleons. to provide an adequate fit to the XENON1T signal. Additionally, in Tab.I we report typical core densi- This raises the tantalizing possibility that XENON1T, ties and temperatures for these objects, alongside the originally constructed to detect dark matter, may have chameleon mass for the benchmark parameter space achieved the first direct detection of dark energy quanta. point we have considered throughout the paper. As In principle, one may also consider a hybrid chameleon- we see, production of chameleons within these objects axion scenario where both particles are present and con- is strongly kinematically suppressed, even more so than tribute to the XENON1T signal, which thus results within the Sun, implying that stellar cooling constraints are evaded. Therefore, the bounds obtained by one of us in Ref. [212], derived assuming a massless chameleon, may be safely evaded. 8 Should the model be in tension with LEP constraints, extensions Finally, we note that the literature is rich with astro- which can alleviate this tension are possible. For instance, along physical (stellar and galactic) bounds on chameleons, see the lines of Ref. [221], one could entertain the possibility of an environmentally-dependent disformal coupling. This effectively e.g. Refs. [27, 28, 224–229]. These refer to searches for amounts to an extension of the chameleon mechanism retaining the effects of fifth forces rather than chameleon particle the background field-dependence in the disformal term in Eq. (4), production. Chameleon models predicting fifth forces in which we dropped for simplicity. astrophysical objects occupy a different region of param- 13

Stellar object ρcore Tcore mcore Experiment Exposure Electron recoil Events / yr (typical) (typical) background (expected) [g/cm3] [keV] [keV] (ton×yr) (ton×yr×keV)−1 Sun 150 1.3 6 XENON1T [80] 0.65 76.0 20 White dwarfs 106 O(1) ∼ 6000 XENONnT [146] 20.0 12.3 180 Red giants 5 × 105 O(10) ∼ 4000 PandaX-4T [145] 5.6 18.0 130 Horizontal branch stars 5 × 104 O(10) ∼ 100 LUX-ZEPLIN [141] 15.0 14.0 250

TABLE I. Typical core densities and temperatures for stellar TABLE II. Expected exposure in units of ton×yr (tonne-year) objects of interest: the Sun, white dwarfs, red giants, and hor- and electron recoil background in units of (ton × yr × keV)−1 izontal branch stars. The final column reports the chameleon for recoil energies . 10 keV, expected for the upcoming mass within the core of these objects for the benchmark pa- XENONnT [146], PandaX-4T [145], and LUX-ZEPLIN [141] rameter space point we have considered throughout the paper, experiments, which will be able to confirm or disprove the 2 where βe = 10 , Λ = 1 µeV, and n = 1. It is clear that pro- XENON1T excess. The last column reports the number of duction of stellar chameleons is strongly kinematically sup- excess events that are expected per year in each detector, in pressed within these objects, as mcore  Tcore. the energy range (1 − 30) keV. eter space than those that give rise to chameleon parti- either confirm or disprove it. These experiments include cle production in the solar tachocline considered in this XENONnT (the planned upgrade to XENON1T) [146], work. The underlying reason for this is that fifth forces as well as PandaX-4T [145] and LUX-ZEPLIN [141], all are only relevant in astrophysical bodies of radius R if of which use a dual-phase xenon time projection cham- m(ϕ0)R 1, where ϕ0 is the background field in that ∼ ber. In TableII we report the expected exposure in units body. The chameleon theories accessible to the direct de- of ton yr and electron recoil background in units of tection experiments that we have discussed in this work (ton ×yr keV)−1 for each of these experiments. have m(ϕ )R 1, implying a fifth force range too × ×  We focus on the benchmark point in parameter space small to affect stellar structure. For this reason, it is considered in Sec.V, as well as Figs.1 and2. Adopt- generally the case that astrophysical fifth force searches ing these values, the expected excess number of events do not constrain our proposed scenario. These theories per year due to a hypothetical signal is about 20 for may be subject to the bounds arising from laboratory XENON1T, 180 for XENONnT, 130 for PandaX-4T, and tests [29, 37], although such bounds are highly model- 250 for LUX-ZEPLIN. Note that we have not considered dependent. The specific model studied in this work is the effects of background noise or energy resolution in ob- able to simultaneously satisfy all experimental bounds taining these estimates of excess number of events since and account for part of the XENON1T signal. the energy resolution specifications for these future ex- periments are currently unavailable. As is clear from TableII, all of these next-generation B. Other dark matter direct detection experiments xenon-based experiments will have a background level B0 a factor of 5-6 lower than current levels, while the Recent DM direct detection searches prior to exposure will increase≈ by over an order of magnitude. XENON1T did not report any excess over the expected This combination of lower background and increased ex- background. For example, the PandaX-II experiment posure results leads to the extremely high number of ex- with an exposure of 27 ton day placed an upper limit pected events. By virtue of this, future experiments will ≈ × −12 on the axion-electron coupling gae . 4.35 10 for an be able to confirm or disprove our hypothesis that solar 9 × axion mass ma . 1 keV [230]. A competitive limit has chameleons are the origin of the XENON1T signal with also been placed by the LUX-ZEPLIN collaboration with extremely high statistical significance. More generally, it an exposure of 11.2 ton day, which lead to the result will be possible to test at high significance whether the −12 × gae 3.5 10 at 90% confidence level [232]. These . × XENON1T excess is due to a statistical fluke, a back- exposures are all significantly lower than the XENON1T ground contaminant, or new physics such as the scenario exposure of 0.65 ton yr, which could explain why these considered here. experiments did not× observe any low-energy excess. Various upcoming experiments plan to search for the signal reported by the XENON1T collaboration, and will VII. CONCLUSIONS

Most of our knowledge about dark energy (DE) arises 9 A more recent analysis with an exposure of 100.7 ton×day found from cosmological measurements which are mainly sen- −12 a similar result, gae . 4.6 × 10 [231]. sitive to its gravitational effects. Yet searching for non- 14 gravitational signatures of DE by directly detecting DE in terrestrial laboratories. quanta would be an extremely important step towards understanding the physics powering cosmic acceleration. In this paper, our aim has been to broaden the scope of ACKNOWLEDGMENTS new physics accessible to terrestrial dark matter (DM) di- rect detection experiments by investigating the intriguing We are grateful to Djuna Croon, Samuel D. McDermott, possibility that these instruments may be able to detect Rouven Essig, Katherine Freese, Chris Kelso, Jason Ku- DE quanta via their couplings to matter. Specifically, mar, Xudong Sun, and Sebastian Trojanowski for sev- we have envisaged a scenario wherein DE particles pro- eral enlightening discussions. S.V. acknowledges support duced in the strong magnetic fields of the solar tachocline from the Isaac Newton Trust and the Kavli Foundation travel to Earth and are absorbed by electrons or nuclei in through a Newton-Kavli Fellowship, and by a grant from terrestrial DM detectors. In this paper, focusing on DE the Foundation Blanceflor Boncompagni Ludovisi, n´ee scalars including screening mechanisms of the chameleon Bildt. S.V. acknowledges a College Research Associ- type, we have laid out the formalism for computing the ateship at Homerton College, University of Cambridge. expected signal from such a process, demonstrating that L.V. acknowledges support from the European Union’s it can lead to measurable recoils in the (keV) range, O Horizon 2020 research and innovation programme under well within the sensitivity of current and upcoming DM the Marie Sk lodowska-Curie grant agreement “TALeNT” direct detection experiments. No. 754496 (H2020-MSCA-COFUND-2016 FELLINI), We have applied our results to the XENON1T ex- as well as support from the NWO Physics Vrij Pro- periment, which recently reported a 3.3σ excess in gramme “The Hidden Universe of Weakly Interact- ≈ their electron recoil data at recoil energies of 1-2 keV. ing Particles” with project number 680.92.18.03 (NWO ≈ We have shown that solar chameleons can explain the Vrije Programma), which is (partly) financed by the XENON1T excess (see blue curve in Fig.1), and are Dutch Research Council (NWO). P.B. acknowledges sup- preferred over the background-only hypothesis at a sig- port from the Institut Pascal at Universit´eParis-Saclay nificance of 2.0σ. Our results have been obtained with the support of the P2I and SPU research depart- ≈ using the code which we make publicly available at ments and the P2IO Laboratory of Excellence (pro- github.com/lucavisinelli/XENONCHAM. gram “Investissements d’avenir” ANR-11-IDEX-0003-01 Compared to the much discussed solar axion interpre- Paris-Saclay and ANR-10-LABX-0038), as well as by tation of the XENON1T signal, the statistical prefer- the IPhT. A.C.D. acknowledges partial support from ence for solar chameleons is lower, mostly due to the the STFC Consolidated Grants No. ST/P000673/1, reduced number of available production channels in the No. ST/P000681/1, and No. ST/T000694/1. This work Sun (chameleons are only produced through Primakoff- was performed using resources provided by the Cam- like processes in the tachocline). However, the stellar bridge Service for Data Driven Discovery (CSD3) op- cooling constraints which are debilitating for the solar erated by the University of Cambridge Research Com- axion model do not apply to solar chameleons due to the puting Service (www.hpc.cam.ac.uk), provided by Dell environment-dependence of the chameleon mass, which EMC and Intel using Tier-2 funding from the Engineer- within the dense environments of red giants and white ing and Physical Sciences Research Council (capital grant dwarfs results in Compton and bremsstrahlung produc- EP/P020259/1), and DiRAC funding from the Science tion processes being suppressed. and Technology Facilities Council (www.dirac.ac.uk). We have also studied prospects for testing this explana- tion in future DM direct detection experiments. If solar chameleons are indeed at the origin of the XENON1T ex- Appendix A: Chameleon Screening cess then this will be confirmed at very high significance in upcoming experiments such as XENONnT, PandaX- Chameleon-screened theories provide an explicit exam- 4T, and LUX-ZEPLIN. What is perhaps more important ple where the mapping between the effective theory in is that future low-threshold DM direct detection experi- the solar system given in the main text and a more com- ments will be well-suited to detect the signatures of DE plete theory can be calculated [189] (see Refs. [29, 233] particles produced within the Sun. for more details about screening). In this case, the scalar There are several avenues for future research in this φ is canonically normalised with a scalar potential V (φ). direction. While our study has focused on chameleon- The interaction with matter is obtained via the Jordan J 2 screened scalars, we stress that the effective theory ap- frame metric gµν = A (φ)gµν . We treat the disformal proach we have adopted is quite general, and our formal- and derivative interactions as negligible perturbations ism can therefore be applied (with appropriate modifica- compared to the dominant effect due to the conformal tions) to several other physical scenarios of interest. More rescaling given by A(φ). This is valid provided the sup- generally, we hope that this paper will stimulate further pression scales of the disformal and derivative interac- research aimed towards enabling direct detection of dark tions are large enough. In the Einstein frame, where the energy, searching for non-gravitational signatures of dark graviton and scalar are canonically normalized but the energy, and unraveling the physics of cosmic acceleration coupling to matter is non-minimal, the dynamics of φ 15 depend on the effective potential and to integrate (A6) Z ρ Veff (φ) = V (φ) + ρA(φ) . (A1) φ(ρ) A(ρ)β(ρ) = φ0 2 2 dρ . (A9) MPl − ρ0 mφ(ρ)MPl where ρ is the conserved matter density in the Einstein frame, related to the density in the Einstein frame as This provides a one-to-one relationship between the den- ρE = Aρ. This relationship between ρE and ρ fol- sity of matter and the value of the minimum φ(ρ). More- lows from the non-conservation of the Einstein-frame en- over from (A6) we get ergy momentum tensor which, for pressureless matter, gives [234] dA−1(ρ) β2(ρ) = , (A10) dρ m2 (ρ)M 2 β φ Pl ρ˙E + 3hEρE = ρEφ˙ , (A2) mPl allowing one to obtain A(ρ) as a decreasing function of ρ, i.e. A−1 is an increasing function of ρ with a positive whereρ ˙ = uµD ρ is the Einstein-time derivative along E µ E derivative. Finally the minimum equation gives the trajectories of the matter particles with 4-velocities µ u . The local Hubble rate in the Einstein frame is hE = V β2 ρ A2 ρ µ d ( ) ( ) Dµu /3 which reduces to the cosmological Hubble rate = ρ 2 2 , (A11) dρ − mφ(ρ)MPl hE = H on large scales. From Eq. (A2), we deduce that from which we can find V (ρ). Hence eliminating ρ be- ρ˙ + 3h ρ = 0 , (A3) E tween A(ρ), V (ρ), and φ(ρ), one can reconstruct A(φ) which expresses the local conservation of the matter den- and V (φ). −4 The potential V (φ) is defined up to an integration con- sity ρ. The Jordan frame matter density ρJ = A ρE is stant, i.e. the screening properties of the models do not conserved in the Jordan frame where dtJ = AdtE and hE A˙ depend on an additive cosmological constant. This con- the local Hubble rate is given by h = + 2 , i.e. J A A stant has to be tuned to generate the appropriate accel- eration of the Universe. As the screening properties are dρJ + 3h ρ = 0 . (A4) independent of this choice, this has no effects on the re- dt J J J sults obtained in this paper. As a specific example, let The Jordan frame matter density is deemed to be the us consider the inverse power law potential. “physical” density as, in the local Jordan frame where Λ4+n the Jordan metric is nearly Minkowskian, the Lagrangian V (φ) = V0 + n , (A12) of the standard model reduces to the usual one. As long φ as A 1, the difference between ρ and ρJ is negligible. ≈ where we have included the constant V0 which needs to be Chameleon-screened models have a minimum of the adjusted to fit the current dark energy value. As we have φ ρ effective potential ( ) which depends on the conserved seen, the screening properties only depend on the the ρ matter density in the Einstein frame. The minimum inverse power law part. We note that models could arise equation reads from the strong dynamics of a confining supersymmetric dark sector at higher energy [235]. dV dA + ρ = 0 , (A5) The configuration of the field at which the potential φ φ d φ(ρ) d φ(ρ) is minimized φ(ρ), and the chameleon rest mass squared obtained from the curvature of the effective potential, which can be used to obtained a parametric description m2 , are given by of the value φ(ρ). Taking the derivative of the minimum φ equation with respect to ρ leads to 1 n M Λ4+n  1+n φ(ρ) = Pl , (A13) β ρ 2 dφ(ρ) dA m mφ(ρ) = , (A6) 2+n dρ − dφ φ(ρ)  β ρ  1+n m2 (ρ) = n(1 + n)Λ4+n m , (A14) φ nM Λ4+n where we have defined the effective mass Pl m 2 2 This is the mass φ that we use as a template in the 2 d V d A main text [see Eq. (9)]. mφ(ρ) = 2 + ρ 2 . (A7) dφ φ(ρ) dφ φ(ρ) Finally this reconstruction procedure of V (φ) and A(φ) from mφ(ρ) and β(ρ) allows one to design models where It is convenient to introduce the effective coupling the production of scalars in the tachocline is favoured compared to very deep inside the Sun or in other even d ln A β ρ M , denser stars. In the Sun, all that is required is that in ( ) = Pl (A8) −3 dφ the core where the density ρcore 150 g cm is large φ(ρ) ≈ 16

−3 compared to the one in the tachocline ρtach 1 g cm , 238], so that photon propagation in this region proceeds the production of scalars is kinematically forbidden.≈ For through a random walk process, which can be described chameleons produced in matter, e.g. by the Primakoff as a Poisson diffusion process with mean free path λ. For process deep in the electric field of a nucleus, their effec- a typical distance ` between two scatterings, the total tive mass is modified by the presence of the surround- number of scatterings per unit time is c/`. For a given 2 ∼ ing plasma and their mixing with photons as meff = length path `, the differential probability of conversion 2 2 m (ρcore) ω where the plasma frequency is defined in in the solar interior is [54] φ − Pl Eq. (B6). For large enough densities, the mass mφ(ρcore) Z +∞ r −`/λ is generically larger than the plasma frequency which dPφ d` ls e 1/2 10 = Pγ→φ , (B7) scales as ρ . As a result, only scalars of momenta dR 0 ` ` λ k2 T 2 m2 can be produced. When m T , core eff eff & core ¯ 10 production' − is highly suppressed. This also applies to the where ls = ct 3 10 cm in the tachocline, i.e. ap- ' × ¯ Compton and bremsstrahlung processes involving the di- proximately one light-second. Here, t is the typical time rect coupling between scalars and electrons where the such that the photon flux at the tachocline nγ,t =v ¯n¯t same kinematical obstruction is at play. This mechanism wheren ¯t is the photon number density at the tachocline ¯ 1/2 was first proposed in Ref. [236]. andv ¯ = (cλ/t) is the typical radial velocity of pho- tons due to their Brownian motion. The differential flux of chameleons per unit energy emitted by the Sun is Appendix B: Production of chameleons in the Sun dΦ Z R dP = dR p (R) n (R) φ , (B8) ω γ γ R Given the chameleon-photon disformal coupling in d 0 d

Eq. (6), the probability of a photon in the uniform mag- where nγ (R) is the photon flux profile and the photon B netic field converting into a chameleon after a distance spectrum pγ (R) depends on the temperature profile of ` is given by [195] the plasma T = T (R) as

2 2 4∆B 2 ` 1 ω Pγ→φ = 2 2 sin , (B1) pγ (R) = . (B9) 4∆B + (∆pl ∆a) `ω 2ζ(3) T 3 exp(ω/T ) 1 − − where the coefficients are We model the magnetic field profile inside the Sun as a 2β B thin shell around the solar tachocline, where the magnetic γ , ∆B = (B2) field is taken to be constant with a value Bt = 30 T. The MPl √1 + b2 thin shell around the tachocline has radius R = 0.7R , ω2 t pl where R is the radius of the Sun, and the thickness ∆pl = , (B3) 2ω ∆R = 0.01R . The integrand in Eq. (B8) at the 2 2 2 2 2
m + 2b ω 1 Bz /B tachocline is then ∆ = φ − , (B4) a 2 r 2ω (1 + b ) dΦ Z +∞ d` ls e−`/λ 2 = ∆R pγ,t nγ,t Pγ→φ , (B10) `ω = . (B5) dω 0 ` ` λ p 2 2 4∆ + (∆pl ∆a) B 21 −2 −1 − where nγ,t = nγ (Rt) 10 cm s and pγ,t is the In the expressions above, ω is the energy of the produced expression in Eq. (B9)≈ evaluated at the tachocline tem- 2 4 chameleon, the dimensionless parameter b = Bt /Mγ is perature T 0.2 keV, and we have used Eq. (B7) which the ratio of the magnetic field in the solar tachocline Bt expresses the≈ differential probability of conversion. In- to the UV-cutoff scale of the effective theory, and the serting Eq. (B1) into Eq. (B10) we obtain plasma frequency is given in terms of the electron number 2 r   density ne and the electron mass me as dΦ ∆R 4∆B ls `ω = pγ,t nγ,t 2 2 , dω λ 4∆ + (∆pl ∆a) `ω I λ 2 4π ne 8 2  ne  B − ωpl = (2.0 10 GHz) 23 −3 . (B6) (B11) me ≈ × 10 cm where the integral over y = `/`ω has been performed The quantity Bz is the z-component of the magnetic field exactly. For any constant a, we find which we fix by assuming an isotropic magnetic field dis- Z +∞ 2 r q  2 2 sin y −ay π p tribution as Bz = B /3. a y e a a2 √ a . ( ) d 3/2 = + + 4 2 The thickness of the tachocline is much larger than the I ≡ 0 y 2 − main free path of photons in the region, λ 0.3 cm [237, (B12) ≈ For most of the region of the parameter space we ex- plore, the relation `ω λ holds, for which the integral  in Eq. (B12) is (`ω/λ) √π. 10 I ≈ For the inverse power law chameleons with n = 1, their mass For the region of parameters allowed we have ∆B scales as ρ3/4.  ∆pl. We also assume Mγ (keV), which corresponds  O 17 to b 1. In this limit, the expression for the solar We decompose the electron free field as chameleon flux in Eq. (B11) reduces to Z 3 X d p −ip·x ip·x
3/2 ψ u p b e v p c e , 2 1/2 2 ! = p6 s( ) s + s( ) s dΦ ∆R 32β ls B ω `  2Ep = p n γ ω . s γ,t γ,t 2 2 2 (C3) dω λ MPl ωpl mφ I λ − s (B13) where the subscript labels the spinor component and 2 2 us(p) and vs(p) are Dirac spinors following the normali- In the limit where mφ ωpl, Eq. (B13) reduces to P P  sation condition s usu¯s =p+me and s vsv¯s =p me. The operator b (p) and its6 adjoint satisfy the6 − anti- 32β2 ls1/2B2   s dΦ ∆R γ 3/2 `ω commutation relation = pγ,t nγ,t 2 3 ω . (B14) dω λ MPlωpl I λ † 0 (3) 0 bs(p), bs0 (p ) = δ (p p )δss0 , (C4) The computation we have just outlined is valid as long { } − as the effect of the disformal coupling on the scalar field and similarly for the operator cs(p), where curly brackets profile in the Sun is negligible. In other words, that the denote the anti-commutation of the two operators. We back-reaction effect due to the disformal coupling can be have defined d3p = d3p/(2π)3. 6 neglected. The disformal coupling leads to a contribution The non-relativistic electron bound state is to the kinetic term of the scalar field proportional to P/M 4 [239–243], with P T 4 the pressure of the solar X Z i ψ = d3p χ ϕ(p)b e−ip·x , (C5) photon gas, at a temperature∝ T (where, in our units, the s s s 6 proportionality factor is smaller than unity). Therefore, requiring that the disformal coupling does not modify the where the Dirac spinor χs(p) has the antiparticle entries field-profile in the Sun is tantamount to requiring that equal to zero and ϕ(p) is a non-relativistic wave func- Mi Tcore, and therefore Mi (keV). The condition & & O tion in the momentum representation. We consider the in the tachocline is weaker as its temperature is lower ground state of a bound electron, than the core temperature by an order of magnitude. In the analysis of Sec.IV, we have imposed priors  3/2 1 Z −Zr/a which ensure that Mγ ,Me & (keV), to satisfy the pre- ϕ(r) = e 0 , (C6) vious bound. In addition, we noteO that for the benchmark √π a0 point of parameter space we have discussed throughout the paper, and in particular in Figs.1 and2, we have where Z is the atomic number (Z = 131 for xenon) and 3.6 a0 is the Bohr radius. In momentum space, we obtain set Me = 10 keV and Mγ = 1000 TeV, such that the bound is well satisfied. Z 8√π Z 5/2 ϕ(p)= d3re−ip·rϕ(r)= , 2 2 2 (p + (Z/a0) ) a0 Appendix C: Chameleo-electric cross-section (C7) where the wave function is normalised such that In this Appendix we derive an expression for the detec- Z tion cross-section of chameleons from the analogue of the d3p ϕ(p) 2 = 1 . (C8) photoelectric effect. The cross-section for the “chameleo- 6 | | electric” effect receives contributions from each of the three terms in the last line of the Lagrangian in Eq. (5). We first consider the disformal coupling between φ and the electron,

1 µν √ g 4 ∂µφ∂ν φ Te , (C1) L ⊃ − Me

1/4 where Me = M/de is the energy scale related to the dis- formal coupling with electrons. The stress-energy tensor associated with an electron four-spinor ψ is i   T µν = ψγ¯ (µDν)ψ D(µψγ¯ ν)ψ , (C2) e 2 − where γµ are the Dirac matrices and a bracket denotes a symmetrisation over the four-indices µ, ν. In the follow- µ FIG. 4. Feynman diagram for the scattering process associ- ing, we adopt Feynman slash notation A = γ Aµ for a † 06 ated with the disformal coupling in Eq. (C1). four-vector Aµ, and we define ψ¯ = ψ γ . 18

The scattering vertex from the disformal coupling in Eq. (C1) is sketched in the Feynman diagram in Fig.4 and amplitude given by

1 0 µ = 4 u¯(p ) γ Aµ χ ϕ(p) , (C9) M −4Me where we have introduced the vector

0 0 ν 0 ν 0 Aµ = kµ (p + p ) kν + kµ (p + p ) kν . (C10) The square of the amplitude summed over the spins of the final states and averaged over the spins of the initial states is 2 2 ϕ(p) 0
= | |8 Tr A(p + me) A (1 + γ ) = |M| 64Me 6 6 6 2 ϕ(p)  2 0 0 µ 0  = | |8 A (me E )+2A (A pµ) , (C11) 16Me − with the cross-section Z 3 3 0 3 0 d p d k d p 2 (4) 0 0 σφe,dis = 6 6 6 δ (k+p k p ) = 2k 2ω0 2E0 |M| 6 − − Z d3p d3k0 1 2 0 0 FIG. 5. Relative orientation of the vectors k, k0, p, in the = 6 6 0 0 δ(ω+E ω E ) = 2k 2ω 2E |M| 6 − − spherical coordinate system chosen. 1 Z (k0)2 dk0 = dydy0dτdpp2 2 . (C12) (2π)5 8kE0ω0 |M| dω0 the cavity R [71, 244, 245], and thus to be of the order of −7 Here, δ(x) = 2πδ(x) and we decomposed the four-vectors mφ 1/R 10 eV, which is much smaller than other 6 ∼ ≈ as follows. We consider the incoming and outgoing energies in the system. For this reason, we neglect mφ µ 0 µ 0 0 chameleon four-vectors k = (ω, k) and (k ) = (ω , k ). in the rest of the computation. Combining Eqs. (C14) µ The electron is described by the systems p = (E, p) and and (C15) we obtain 0 µ 0 0 (p ) = (E , p ), where E = me Eb and Eb is the bind- ing energy of the atomic electron.− In the following, we 2 2 2 0 E + 2kE p me 2kpy k k k0 k0 p p k = − − − k , (C16) make use of the magnitudes = , = , = , | | 2 (k + E px ky0) ≈ and p0 = p0 . We also introduce the| | angle |θ between| | k| − − and p, the| angle| θ0 between k and k0, and the azimuthal 0 where “ ” indicates the limit me k Eb mφ. angle τ between the projections of k and p on the plane We define≈ the product   | |  orthogonal to k. The orientations of these vectors are 0 ν 0 0 sketched in Fig.5. α1 = (p + p ) kν = (E + E )ω (p + p ) k = Since dk0/dω0 = ω0/k0, the expression for the cross- − · = (E0 +E)ω (2py+k k0y0) k , (C17) section in Eq. (C12) is − − Z 0 0 1 0 2 2 where in the last line the spatial part is p = p + k k . σφe,dis = dydy dτ dp p ψ(p) − 8 5 Similarly, we define 128 Me (2π) | | × k0 0 ν 0 0 0 0 0 A2 m E0 A Aµp0
 , α2 = (p + p ) kν = (E + E )ω (p + p ) k = 0 ( e ) + 2 0 µ (C13) − · E k − = (E0 +E)ω0 (2px k0 +ky0) k0 . (C18) where we have used the expression for the amplitude − − 0 squared in Eq. (C11). We then have Aµ = α1 kµ+α2 kµ. In the limit considered, 0 To proceed with the computation, we consider the con- E E me, so we obtain ≈ ≈ servation of the four-vector on shell q q α1 α2 2me ω . (C19) 2 2 p 0 2 2 0 2 2 ≈ ≈ E+ k + mφ = (p ) +me + (k ) +mφ ,(C14) In the definitions of α and α , the temporal part of the (p0)2 (k0)2 p2 =k2 2pk0x+2pky 2kk0y0 , (C15) 1 2 − − − − four-product is the dominant one in the limit considered. 0 0 0 0 where y = cos θ, y = cos θ , x = yy + sin θ sin θ cos τ. The time component of the vector Aµ is In the detector, the mass of the chameleon is expected 0 2 to be set by a resonance condition involving the size of A0 = α1 ω + α2 ω 4me ω , (C20) ≈ 19 where in the last step we used Eq. (C19). The square of can be neglected, the cross-section is the vector Aµ is Z 1 0 2 2 2 4 σφe,dis dydy dτ dp p ψ(p) [32 m ω ] = 8 5 e 2 µ 0µ µ 0
≈ 128Me (2π) | | A = A Aµ = (α1k + α2k ) α1kµ + α2kµ = 2 4 Z 2 4 2 2 2
0 0 0 me ω 2 2 me ω = m α + α + 2α1α2 (ωω kk y ) = dp p ψ(p) = , (C27) φ 1 2 8 4 8π2 M 8 0 0 0 − ≈ Me (2π) | | e 2α1α2 (ωω kk y ) , (C21) ≈ − where the angular integral is trivial as there are no an- where the last approximation assumes a massless gles appearing in Eq. (C26). In the last step, we have µ chameleon at detection, k kµ = 0 or k = ω. Since normalised the wave function according to Eq. (C8). ω0 ω, in the limit considered we have | | A second contribution to the cross section comes from ≈ the conformal term in Eq. (5) A2 8m2 ω4 (1 y0) . (C22) ≈ e − φ √ gβe Te , (C28) L ⊃ − MPl We evaluate the product for which the absorption cross section depends on the µ 0 0 0 A p A E A p photo-electric cross section σphoto in the limit ω mφ µ = 0 =  0 − · 0 0 0 as [210, 246, 247] =(α1 ω +α2 ω)E (α1 k +α2 k) (p + k k ) = 0 0 −0 0 · − 0 0 2 2 =[α1ω (E px+ky k )+α2ω(E py+k k y )] βe ω 2 2 − − − − ≈ σφe,conf = 2 σphoto . (C29) 4me ω , (C23) 2παMPl ≈ We have taken the energy-dependent photoelectric cross 0 0 0 section from Ref. [248]. Note, that we have not consid- where in the last step E E me p , k , k . Using the approximation ≈ ≈  | | | | | | ered the production/detection from the XT coupling and therefore set ce = 0. 2 In terms of the parameters used in the MCMC analy- 0 0 ω E me + (1 y ) , (C24) sis, Eqs. (C27) and (C29) combine to give the following ≈ − me expression for the cross-section together with the expressions in Eqs. (C20), (C22), m2ω4 β2ω2 σ σ σ e e σ . and (C23), we find φe = φe,dis + φe,conf = 2 8 + 2 photo 8π Me 2παMPl (C30) 2 0 µ 0
6 0 2 3 4 A (me E )+2A0 A pµ 8me ω (1 y ) +32me ω , Although in this work we have focused on xenon-based − ≈ − − (C25) detectors such as XENON1T, the results of the compu- µ 0
so the term is dominated by the part 2A0 A pµ . The tation can be applied more broadly to any material. second line in the computation of the cross-section in When the second term can be neglected, the event Eq. (C13) is then rate in the detector given by Eq. (12), with flux given in Eq. (B14), gives k0  2 0 µ 0
 2 4 2 2 1/2 2 2 11/2 A (me E ) + 2A0 A pµ 32 me ω . (C26) dR (ω) ∆R R β ls B m ω E0 k − ≈ 0 N p n eff e , = Xe γ,t γ,t 2 3/2 2 3 8 dω λ d π MPlωpl keV For a chameleon produced in the Sun with an energy ω, (C31) and in the limit in which its effective mass in the detector where βeff has been defined in Eq. (13).

[1] A. G. Riess et al. (Supernova Search Team), Astron. J. arXiv:1205.3421 [gr-qc]. 116, 1009 (1998), arXiv:astro-ph/9805201. [6] A. Joyce, B. Jain, J. Khoury, and M. Trodden, Phys. [2] S. Perlmutter et al. (Supernova Cosmology Project), As- Rept. 568, 1 (2015), arXiv:1407.0059 [astro-ph.CO]. trophys. J. 517, 565 (1999), arXiv:astro-ph/9812133. [7] D. Huterer and D. L. Shafer, Rept. Prog. Phys. 81, [3] S. Nojiri and S. D. Odintsov, eConf C0602061, 06 016901 (2018), arXiv:1709.01091 [astro-ph.CO]. (2006), arXiv:hep-th/0601213. [8] J. Sola, J. Phys. Conf. Ser. 453, 012015 (2013), [4] J. Frieman, M. Turner, and D. Huterer, Ann. Rev. As- arXiv:1306.1527 [gr-qc]. tron. Astrophys. 46, 385 (2008), arXiv:0803.0982 [astro- [9] C. P. Burgess, in 100e Ecole d’Ete de Physique: Post- ph]. Planck Cosmology (2013) arXiv:1309.4133 [hep-th]. [5] K. Bamba, S. Capozziello, S. Nojiri, and S. D. [10] A. Padilla, (2015), arXiv:1502.05296 [hep-th]. Odintsov, Astrophys. Space Sci. 342, 155 (2012), [11] P. Ade et al. (Simons Observatory), JCAP 02, 056 20

(2019), arXiv:1808.07445 [astro-ph.CO]. (2014), arXiv:1312.6016 [astro-ph.CO]. [12] M. H. Abitbol et al. (Simons Observatory), Bull. Am. [43] M. Gronke, C. Llinares, D. F. Mota, and H. A. Astron. Soc. 51, 147 (2019), arXiv:1907.08284 [astro- Winther, Mon. Not. Roy. Astron. Soc. 449, 2837 (2015), ph.IM]. arXiv:1412.0066 [astro-ph.CO]. [13] K. N. Abazajian et al. (CMB-S4), (2016), [44] M. Gronke, D. F. Mota, and H. A. Winther, Astron. arXiv:1610.02743 [astro-ph.CO]. Astrophys. 583, A123 (2015), arXiv:1505.07129 [astro- [14] K. Abazajian et al., Bull. Am. Astron. Soc. 51, 209 ph.CO]. (2019), arXiv:1908.01062 [astro-ph.IM]. [45] M. Vargas dos Santos and D. F. Mota, Phys. Lett. B [15] R. Laureijs et al. (EUCLID), (2011), arXiv:1110.3193 769, 485 (2017), arXiv:1603.03243 [astro-ph.CO]. [astro-ph.CO]. [46] T. Katsuragawa and S. Matsuzaki, Phys. Rev. D 95, [16] L. Amendola et al., Living Rev. Rel. 21, 2 (2018), 044040 (2017), arXiv:1610.01016 [gr-qc]. arXiv:1606.00180 [astro-ph.CO]. [47] T. Katsuragawa and S. Matsuzaki, Phys. Rev. D [17] A. Aghamousa et al. (DESI), (2016), arXiv:1611.00036 97, 064037 (2018), [Erratum: Phys.Rev.D 97, 129902 [astro-ph.IM]. (2018)], arXiv:1708.08702 [gr-qc]. [18] A. Abate et al. (LSST Dark Energy Science), (2012), [48] G. J. Olmo, D. Rubiera-Garcia, and A. Wojnar, Phys. arXiv:1211.0310 [astro-ph.CO]. Rept. 876, 1 (2020), arXiv:1912.05202 [gr-qc]. [19] D. Spergel et al., (2013), arXiv:1305.5422 [astro-ph.IM]. [49] T. Kumar Poddar, S. Mohanty, and S. Jana, Eur. Phys. [20] T. P. Sotiriou and V. Faraoni, Rev. Mod. Phys. 82, 451 J. C 81, 286 (2021), arXiv:2002.02935 [hep-ph]. (2010), arXiv:0805.1726 [gr-qc]. [50] M. C. Straight, J. Sakstein, and E. J. Baxter, Phys. [21] T. Clifton, P. G. Ferreira, A. Padilla, and C. Sko- Rev. D 102, 124018 (2020), arXiv:2009.10716 [gr-qc]. rdis, Phys. Rept. 513, 1 (2012), arXiv:1106.2476 [astro- [51] Y.-D. Tsai, Y. Wu, S. Vagnozzi, and L. Visinelli, ph.CO]. (2021), arXiv:2107.04038 [hep-ph]. [22] S. Nojiri, S. D. Odintsov, and V. K. Oikonomou, Phys. [52] E. A. Spiegel and J. P. Zahn, Astron. Astrophys. 265, Rept. 692, 1 (2017), arXiv:1705.11098 [gr-qc]. 106 (1992). [23] C. M. Will, Living Rev. Rel. 9, 3 (2006), arXiv:gr- [53] P. Brax and K. Zioutas, Phys. Rev. D 82, 043007 (2010), qc/0510072. arXiv:1004.1846 [astro-ph.SR]. [24] J. Sakstein, Phys. Rev. D 97, 064028 (2018), [54] P. Brax, A. Lindner, and K. Zioutas, Phys. Rev. D 85, arXiv:1710.03156 [astro-ph.CO]. 043014 (2012), arXiv:1110.2583 [hep-ph]. [25] J. Khoury, (2010), arXiv:1011.5909 [astro-ph.CO]. [55] S. Baum, G. Cantatore, D. H. H. Hoffmann, M. Karuza, [26] P. Brax, Class. Quant. Grav. 30, 214005 (2013). Y. K. Semertzidis, A. Upadhye, and K. Zioutas, Phys. [27] J. Sakstein, Int. J. Mod. Phys. D 27, 1848008 (2018), Lett. B 739, 167 (2014), arXiv:1409.3852 [astro-ph.IM]. arXiv:2002.04194 [astro-ph.CO]. [56] V. Anastassopoulos et al. (CAST), Phys. Lett. B 749, [28] T. Baker et al., Rev. Mod. Phys. 93, 015003 (2021), 172 (2015), arXiv:1503.04561 [astro-ph.SR]. arXiv:1908.03430 [astro-ph.CO]. [57] V. Anastassopoulos et al. (CAST), JCAP 01, 032 [29] C. Burrage and J. Sakstein, Living Rev. Rel. 21, 1 (2019), arXiv:1808.00066 [hep-ex]. (2018), arXiv:1709.09071 [astro-ph.CO]. [58] S. Arguedas Cuendis et al., Phys. Dark Univ. 26, 100367 [30] K. Koyama, Rept. Prog. Phys. 79, 046902 (2016), (2019), arXiv:1906.01084 [hep-ex]. arXiv:1504.04623 [astro-ph.CO]. [59] G. Rybka et al. (ADMX), Phys. Rev. Lett. 105, 051801 [31] P. G. Ferreira, Ann. Rev. Astron. Astrophys. 57, 335 (2010), arXiv:1004.5160 [astro-ph.CO]. (2019), arXiv:1902.10503 [astro-ph.CO]. [60] J. K. Vogel et al., in 8th Patras Workshop on [32] J. Sakstein, H. Desmond, and B. Jain, Phys. Rev. D Axions, WIMPs and WISPs (2013) arXiv:1302.3273 100, 104035 (2019), arXiv:1907.03775 [astro-ph.CO]. [physics.ins-det]. [33] S. Vagnozzi, L. Visinelli, O. Mena, and D. F. [61] Y. Kahn, B. R. Safdi, and J. Thaler, Phys. Rev. Lett. Mota, Mon. Not. Roy. Astron. Soc. 493, 1139 (2020), 117, 141801 (2016), arXiv:1602.01086 [hep-ph]. arXiv:1911.12374 [gr-qc]. [62] A. Caldwell, G. Dvali, B. Majorovits, A. Millar, G. Raf- [34] J. B. Jim´enez,D. Bettoni, D. Figueruelo, and F. A. felt, J. Redondo, O. Reimann, F. Simon, and F. Stef- Teppa Pannia, JCAP 08, 020 (2020), arXiv:2004.14661 fen (MADMAX Working Group), Phys. Rev. Lett. 118, [astro-ph.CO]. 091801 (2017), arXiv:1611.05865 [physics.ins-det]. [35] K. V. Berghaus, P. W. Graham, D. E. Kaplan, G. D. [63] R. Barbieri, C. Braggio, G. Carugno, C. S. Gallo, Moore, and S. Rajendran, (2020), arXiv:2012.10549 A. Lombardi, A. Ortolan, R. Pengo, G. Ruoso, and [hep-ph]. C. C. Speake, Phys. Dark Univ. 15, 135 (2017), [36] R.-G. Cai, Z.-K. Guo, L. Li, S.-J. Wang, and W.-W. arXiv:1606.02201 [hep-ph]. Yu, Phys. Rev. D 103, 121302 (2021), arXiv:2102.02020 [64] B. M. Brubaker et al., Phys. Rev. Lett. 118, 061302 [astro-ph.CO]. (2017), arXiv:1610.02580 [astro-ph.CO]. [37] C. Burrage and J. Sakstein, JCAP 11, 045 (2016), [65] D. Alesini, D. Babusci, D. Di Gioacchino, C. Gatti, arXiv:1609.01192 [astro-ph.CO]. G. Lamanna, and C. Ligi, (2017), arXiv:1707.06010 [38] K. Homma and Y. Kirita, JHEP 09, 095 (2020), [physics.ins-det]. arXiv:1909.00983 [hep-ex]. [66] D. Alesini et al., (2019), arXiv:1911.02427 [physics.ins- [39] P. Chang and L. Hui, Astrophys. J. 732, 25 (2011), det]. arXiv:1011.4107 [astro-ph.CO]. [67] I. G. Irastorza and J. Redondo, Prog. Part. Nucl. Phys. [40] L. Iorio, JHEP 05, 073 (2012), arXiv:1109.6249 [gr-qc]. 102, 89 (2018), arXiv:1801.08127 [hep-ph]. [41] C. Llinares and D. F. Mota, Phys. Rev. Lett. 110, [68] L. Visinelli and S. Vagnozzi, Phys. Rev. D 99, 063517 151104 (2013), arXiv:1205.5775 [astro-ph.CO]. (2019), arXiv:1809.06382 [hep-ph]. [42] C. Llinares and D. F. Mota, Phys. Rev. D 89, 084023 [69] L. Di Luzio, M. Giannotti, E. Nardi, and L. Visinelli, 21

Phys. Rept. 870, 1 (2020), arXiv:2003.01100 [hep-ph]. [97] G. Paz, A. A. Petrov, M. Tammaro, and J. Zupan, [70] P. Sikivie, Rev. Mod. Phys. 93, 015004 (2021), Phys. Rev. D 103, L051703 (2021), arXiv:2006.12462 arXiv:2003.02206 [hep-ph]. [hep-ph]. [71] J. Khoury and A. Weltman, Phys. Rev. D 69, 044026 [98] H. M. Lee, JHEP 01, 019 (2021), arXiv:2006.13183 [hep- (2004), arXiv:astro-ph/0309411. ph]. [72] K. Hinterbichler and J. Khoury, Phys. Rev. Lett. 104, [99] Q.-H. Cao, R. Ding, and Q.-F. Xiang, Chin. Phys. C 231301 (2010), arXiv:1001.4525 [hep-th]. 45, 045002 (2021), arXiv:2006.12767 [hep-ph]. [73] T. Damour and A. M. Polyakov, Nucl. Phys. B 423, 532 [100] A. N. Khan, Phys. Lett. B 809, 135782 (2020), (1994), arXiv:hep-th/9401069. arXiv:2006.12887 [hep-ph]. [74] E. Babichev, C. Deffayet, and R. Ziour, Int. J. Mod. [101] K. Nakayama and Y. Tang, Phys. Lett. B 811, 135977 Phys. D 18, 2147 (2009), arXiv:0905.2943 [hep-th]. (2020), arXiv:2006.13159 [hep-ph]. [75] A. I. Vainshtein, Phys. Lett. B 39, 393 (1972). [102] J. Bramante and N. Song, Phys. Rev. Lett. 125, 161805 [76] K. Koyama and J. Sakstein, Phys. Rev. D 91, 124066 (2020), arXiv:2006.14089 [hep-ph]. (2015), arXiv:1502.06872 [astro-ph.CO]. [103] H. An, M. Pospelov, J. Pradler, and A. Ritz, Phys. [77] J. Khoury and A. Weltman, Phys. Rev. Lett. 93, 171104 Rev. D 102, 115022 (2020), arXiv:2006.13929 [hep-ph]. (2004), arXiv:astro-ph/0309300. [104] C. Gao, J. Liu, L.-T. Wang, X.-P. Wang, W. Xue, [78] P. Brax, C. van de Bruck, A.-C. Davis, J. Khoury, and Y.-M. Zhong, Phys. Rev. Lett. 125, 131806 (2020), and A. Weltman, Phys. Rev. D 70, 123518 (2004), arXiv:2006.14598 [hep-ph]. arXiv:astro-ph/0408415. [105] M. Lindner, Y. Mambrini, T. B. de Melo, and [79] E. Aprile et al. (XENON), Eur. Phys. J. C 77, 881 F. S. Queiroz, Phys. Lett. B 811, 135972 (2020), (2017), arXiv:1708.07051 [astro-ph.IM]. arXiv:2006.14590 [hep-ph]. [80] E. Aprile et al. (XENON), Phys. Rev. D 102, 072004 [106] I. M. Bloch, A. Caputo, R. Essig, D. Redigolo, M. Shola- (2020), arXiv:2006.09721 [hep-ex]. purkar, and T. Volansky, JHEP 01, 178 (2021), [81] L. Di Luzio, M. Fedele, M. Giannotti, F. Mescia, arXiv:2006.14521 [hep-ph]. and E. Nardi, Phys. Rev. Lett. 125, 131804 (2020), [107] D. McKeen, M. Pospelov, and N. Raj, Phys. Rev. Lett. arXiv:2006.12487 [hep-ph]. 125, 231803 (2020), arXiv:2006.15140 [hep-ph]. [82] F. Takahashi, M. Yamada, and W. Yin, Phys. Rev. [108] S.-F. Ge, P. Pasquini, and J. Sheng, Phys. Lett. B 810, Lett. 125, 161801 (2020), arXiv:2006.10035 [hep-ph]. 135787 (2020), arXiv:2006.16069 [hep-ph]. [83] K. Kannike, M. Raidal, H. Veerm¨ae, A. Strumia, [109] W. Chao, Y. Gao, and M. j. Jin, (2020), and D. Teresi, Phys. Rev. D 102, 095002 (2020), arXiv:2006.16145 [hep-ph]. arXiv:2006.10735 [hep-ph]. [110] S. Baek, J. Kim, and P. Ko, Phys. Lett. B 810, 135848 [84] G. Alonso-Alvarez,´ F. Ertas, J. Jaeckel, F. Kahlhoe- (2020), arXiv:2006.16876 [hep-ph]. fer, and L. J. Thormaehlen, JCAP 11, 029 (2020), [111] S. Chigusa, M. Endo, and K. Kohri, JCAP 10, 035 arXiv:2006.11243 [hep-ph]. (2020), arXiv:2007.01663 [hep-ph]. [85] B. Fornal, P. Sandick, J. Shu, M. Su, and Y. Zhao, [112] O. G. Miranda, D. K. Papoulias, M. T´ortola, and Phys. Rev. Lett. 125, 161804 (2020), arXiv:2006.11264 J. W. F. Valle, Phys. Lett. B 808, 135685 (2020), [hep-ph]. arXiv:2007.01765 [hep-ph]. [86] C. Boehm, D. G. Cerdeno, M. Fairbairn, P. A. N. [113] N. Okada, S. Okada, D. Raut, and Q. Shafi, Phys. Lett. Machado, and A. C. Vincent, Phys. Rev. D 102, 115013 B 810, 135785 (2020), arXiv:2007.02898 [hep-ph]. (2020), arXiv:2006.11250 [hep-ph]. [114] G. Choi, T. T. Yanagida, and N. Yokozaki, Phys. Lett. [87] K. Harigaya, Y. Nakai, and M. Suzuki, Phys. Lett. B B 810, 135836 (2020), arXiv:2007.04278 [hep-ph]. 809, 135729 (2020), arXiv:2006.11938 [hep-ph]. [115] H.-J. He, Y.-C. Wang, and J. Zheng, JCAP 01, 042 [88] A. Bally, S. Jana, and A. Trautner, Phys. Rev. Lett. (2021), arXiv:2007.04963 [hep-ph]. 125, 161802 (2020), arXiv:2006.11919 [hep-ph]. [116] I. M. Shoemaker, Y.-D. Tsai, and J. Wyenberg, (2020), [89] L. Su, W. Wang, L. Wu, J. M. Yang, and B. Zhu, Phys. arXiv:2007.05513 [hep-ph]. Rev. D 102, 115028 (2020), arXiv:2006.11837 [hep-ph]. [117] H. N. Long, D. V. Soa, V. H. Binh, and A. E. [90] M. Du, J. Liang, Z. Liu, V. Q. Tran, and Y. Xue, Chin. C´arcamoHern´andez, (2020), arXiv:2007.05004 [hep- Phys. C 45, 013114 (2021), arXiv:2006.11949 [hep-ph]. ph]. [91] Y. Chen, M.-Y. Cui, J. Shu, X. Xue, G.-W. Yuan, and [118] C.-W. Chiang and B.-Q. Lu, Phys. Rev. D 102, 123006 Q. Yuan, JHEP 04, 282 (2021), arXiv:2006.12447 [hep- (2020), arXiv:2007.06401 [hep-ph]. ph]. [119] G. Arcadi, A. Bally, F. Goertz, K. Tame-Narvaez, [92] U. K. Dey, T. N. Maity, and T. S. Ray, Phys. Lett. B V. Tenorth, and S. Vogl, Phys. Rev. D 103, 023024 811, 135900 (2020), arXiv:2006.12529 [hep-ph]. (2021), arXiv:2007.08500 [hep-ph]. [93] N. F. Bell, J. B. Dent, B. Dutta, S. Ghosh, J. Ku- [120] Y. Ema, F. Sala, and R. Sato, Eur. Phys. J. C 81, 129 mar, and J. L. Newstead, Phys. Rev. Lett. 125, 161803 (2021), arXiv:2007.09105 [hep-ph]. (2020), arXiv:2006.12461 [hep-ph]. [121] J. Kim, T. Nomura, and H. Okada, Phys. Lett. B 811, [94] J. Buch, M. A. Buen-Abad, J. Fan, and J. S. C. Leung, 135862 (2020), arXiv:2007.09894 [hep-ph]. JCAP 10, 051 (2020), arXiv:2006.12488 [hep-ph]. [122] J. Cao, X. Du, Z. Li, F. Wang, and Y. Zhang, (2020), [95] D. Aristizabal Sierra, V. De Romeri, L. J. Flores, and arXiv:2007.09981 [hep-ph]. D. K. Papoulias, Phys. Lett. B 809, 135681 (2020), [123] D. Borah, S. Mahapatra, D. Nanda, and N. Sahu, Phys. arXiv:2006.12457 [hep-ph]. Lett. B 811, 135933 (2020), arXiv:2007.10754 [hep-ph]. [96] G. Choi, M. Suzuki, and T. T. Yanagida, Phys. Lett. [124] Y. Farzan and M. Rajaee, Phys. Rev. D 102, 103532 B 811, 135976 (2020), arXiv:2006.12348 [hep-ph]. (2020), arXiv:2007.14421 [hep-ph]. [125] L. Zu, R. Foot, Y.-Z. Fan, and L. Feng, JCAP 01, 070 22

(2021), arXiv:2007.15191 [hep-ph]. (2017), arXiv:1710.05834 [astro-ph.HE]. [126] A. Karozas, S. F. King, G. K. Leontaris, and [156] A. Goldstein et al., Astrophys. J. Lett. 848, L14 (2017), D. K. Papoulias, Phys. Rev. D 103, 035019 (2021), arXiv:1710.05446 [astro-ph.HE]. arXiv:2008.03295 [hep-ph]. [157] J. Sakstein and B. Jain, Phys. Rev. Lett. 119, 251303 [127] S. Chakraborty, T. H. Jung, V. Loladze, T. Okui, (2017), arXiv:1710.05893 [astro-ph.CO]. and K. Tobioka, Phys. Rev. D 102, 095029 (2020), [158] T. Baker, E. Bellini, P. G. Ferreira, M. Lagos, J. Noller, arXiv:2008.10610 [hep-ph]. and I. Sawicki, Phys. Rev. Lett. 119, 251301 (2017), [128] W.-Y. Keung, D. Marfatia, and P.-Y. Tseng, JHEAp arXiv:1710.06394 [astro-ph.CO]. 30, 9 (2021), arXiv:2009.04444 [hep-ph]. [159] J. M. Ezquiaga and M. Zumalac´arregui, Phys. Rev. [129] R. Foot, (2020), arXiv:2011.02590 [hep-ph]. Lett. 119, 251304 (2017), arXiv:1710.05901 [astro- [130] D. Q. Adams, S. Jacobsen, and C. Kelso, (2020), ph.CO]. arXiv:2011.03079 [hep-ph]. [160] P. Creminelli and F. Vernizzi, Phys. Rev. Lett. 119, [131] A. Aboubrahim, M. Klasen, and P. Nath, JHEP 02, 251302 (2017), arXiv:1710.05877 [astro-ph.CO]. 229 (2021), arXiv:2011.08053 [hep-ph]. [161] S. Arai and A. Nishizawa, Phys. Rev. D 97, 104038 [132] D. Buttazzo, P. Panci, D. Teresi, and R. Ziegler, Phys. (2018), arXiv:1711.03776 [gr-qc]. Lett. B 817, 136310 (2021), arXiv:2011.08919 [hep-ph]. [162] C. D. Kreisch and E. Komatsu, JCAP 12, 030 (2018), [133] H.-J. He, Y.-C. Wang, and J. Zheng, (2020), arXiv:1712.02710 [astro-ph.CO]. arXiv:2012.05891 [hep-ph]. [163] R. Kase and S. Tsujikawa, Int. J. Mod. Phys. D 28, [134] S. Xu and S. Zheng, Eur. Phys. J. C 81, 446 (2021), 1942005 (2019), arXiv:1809.08735 [gr-qc]. arXiv:2012.10827 [hep-ph]. [164] R. Kase and S. Tsujikawa, Phys. Rev. D 97, 103501 [135] M. Dutta, S. Mahapatra, D. Borah, and N. Sahu, Phys. (2018), arXiv:1802.02728 [gr-qc]. Rev. D 103, 095018 (2021), arXiv:2101.06472 [hep-ph]. [165] L. Amendola, M. Kunz, I. D. Saltas, and I. Sawicki, [136] R. Agnese et al. (SuperCDMS), Phys. Rev. D 95, Phys. Rev. Lett. 120, 131101 (2018), arXiv:1711.04825 082002 (2017), arXiv:1610.00006 [physics.ins-det]. [astro-ph.CO]. [137] H. Jiang et al. (CDEX), Phys. Rev. Lett. 120, 241301 [166] N. Bartolo, P. Karmakar, S. Matarrese, and M. Scom- (2018), arXiv:1802.09016 [hep-ex]. parin, JCAP 05, 048 (2018), arXiv:1712.04002 [gr-qc]. [138] A. E. Chavarria et al., Phys. Procedia 61, 21 (2015), [167] A. Casalino, M. Rinaldi, L. Sebastiani, and S. Vagnozzi, arXiv:1407.0347 [physics.ins-det]. Phys. Dark Univ. 22, 108 (2018), arXiv:1803.02620 [gr- [139] A. H. Abdelhameed et al. (CRESST), Phys. Rev. D 100, qc]. 102002 (2019), arXiv:1904.00498 [astro-ph.CO]. [168] A. Ganz, N. Bartolo, P. Karmakar, and S. Matarrese, [140] C. Amole et al. (PICO), Phys. Rev. D 100, 022001 JCAP 01, 056 (2019), arXiv:1809.03496 [gr-qc]. (2019), arXiv:1902.04031 [astro-ph.CO]. [169] A. Casalino, M. Rinaldi, L. Sebastiani, and [141] D. S. Akerib et al. (LUX-ZEPLIN), Phys. Rev. D 101, S. Vagnozzi, Class. Quant. Grav. 36, 017001 (2019), 052002 (2020), arXiv:1802.06039 [astro-ph.IM]. arXiv:1811.06830 [gr-qc]. [142] Q. Arnaud et al. (EDELWEISS), Phys. Rev. Lett. 125, [170] S. Arai, P. Karmakar, and A. Nishizawa, Phys. Rev. D 141301 (2020), arXiv:2003.01046 [astro-ph.GA]. 102, 024003 (2020), arXiv:1912.01768 [gr-qc]. [143] O. Abramoff et al. (SENSEI), Phys. Rev. Lett. 122, [171] C. de Rham and S. Melville, Phys. Rev. Lett. 121, 161801 (2019), arXiv:1901.10478 [hep-ex]. 221101 (2018), arXiv:1806.09417 [hep-th]. [144] Q. Wang et al. (PandaX-II), Chin. Phys. C 44, 125001 [172] E. J. Copeland, M. Kopp, A. Padilla, P. M. Saffin, (2020), arXiv:2007.15469 [astro-ph.CO]. and C. Skordis, Phys. Rev. Lett. 122, 061301 (2019), [145] H. Zhang et al. (PandaX), Sci. China Phys. Mech. As- arXiv:1810.08239 [gr-qc]. tron. 62, 31011 (2019), arXiv:1806.02229 [physics.ins- [173] P. Creminelli, M. Lewandowski, G. Tambalo, and det]. F. Vernizzi, JCAP 12, 025 (2018), arXiv:1809.03484 [146] E. Aprile et al. (XENON), JCAP 11, 031 (2020), [astro-ph.CO]. arXiv:2007.08796 [physics.ins-det]. [174] P. Creminelli, G. Tambalo, F. Vernizzi, and [147] J. Gleyzes, D. Langlois, M. Mancarella, and F. Vernizzi, V. Yingcharoenrat, JCAP 05, 002 (2020), JCAP 08, 054 (2015), arXiv:1504.05481 [astro-ph.CO]. arXiv:1910.14035 [gr-qc]. [148] G. W. Horndeski, Int. J. Theor. Phys. 10, 363 (1974). [175] J. Sakstein, Phys. Rev. Lett. 115, 201101 (2015), [149] M. Zumalac´arregui and J. Garc´ıa-Bellido, Phys. Rev. D arXiv:1510.05964 [astro-ph.CO]. 89, 064046 (2014), arXiv:1308.4685 [gr-qc]. [176] J. Sakstein, Phys. Rev. D 92, 124045 (2015), [150] J. Gleyzes, D. Langlois, F. Piazza, and F. Vernizzi, arXiv:1511.01685 [astro-ph.CO]. Phys. Rev. Lett. 114, 211101 (2015), arXiv:1404.6495 [177] I. D. Saltas and I. Lopes, Phys. Rev. Lett. 123, 091103 [hep-th]. (2019), arXiv:1909.02552 [astro-ph.CO]. [151] D. Langlois and K. Noui, JCAP 02, 034 (2016), [178] J. Renk, M. Zumalac´arregui,F. Montanari, and A. Bar- arXiv:1510.06930 [gr-qc]. reira, JCAP 10, 020 (2017), arXiv:1707.02263 [astro- [152] D. Langlois, Int. J. Mod. Phys. D 28, 1942006 (2019), ph.CO]. arXiv:1811.06271 [gr-qc]. [179] J. Dutta, W. Khyllep, E. N. Saridakis, N. Tamanini, [153] L. Sebastiani, S. Vagnozzi, and R. Myrzakulov, and S. Vagnozzi, JCAP 02, 041 (2018), Adv. High Energy Phys. 2017, 3156915 (2017), arXiv:1711.07290 [gr-qc]. arXiv:1612.08661 [gr-qc]. [180] S. Peirone, G. Benevento, N. Frusciante, and [154] B. P. Abbott et al. (LIGO Scientific, Virgo), Phys. Rev. S. Tsujikawa, Phys. Rev. D 100, 063509 (2019), Lett. 119, 161101 (2017), arXiv:1710.05832 [gr-qc]. arXiv:1905.11364 [astro-ph.CO]. [155] B. P. Abbott et al. (LIGO Scientific, Virgo, Fermi- [181] M. Zumalacarregui, Phys. Rev. D 102, 023523 (2020), GBM, INTEGRAL), Astrophys. J. Lett. 848, L13 arXiv:2003.06396 [astro-ph.CO]. 23

[182] A. Dima and F. Vernizzi, Phys. Rev. D 97, 101302 Starkman, Phys. Lett. B 179, 223 (1986). (2018), arXiv:1712.04731 [gr-qc]. [211] J. Redondo, JCAP 12, 008 (2013), arXiv:1310.0823 [183] D. Langlois, R. Saito, D. Yamauchi, and K. Noui, Phys. [hep-ph]. Rev. D 97, 061501 (2018), arXiv:1711.07403 [gr-qc]. [212] P. Brax and C. Burrage, Phys. Rev. D 90, 104009 [184] M. Crisostomi and K. Koyama, Phys. Rev. D 97, 021301 (2014), arXiv:1407.1861 [astro-ph.CO]. (2018), arXiv:1711.06661 [astro-ph.CO]. [213] P. Brax, C. Burrage, and C. Englert, Phys. Rev. D 92, [185] M. Crisostomi and K. Koyama, Phys. Rev. D 97, 084004 044036 (2015), arXiv:1506.04057 [hep-ph]. (2018), arXiv:1712.06556 [astro-ph.CO]. [214] P. Brax, C. Burrage, C. Englert, and M. Spannowsky, [186] M. Crisostomi, M. Lewandowski, and F. Vernizzi, Phys. Phys. Rev. D 94, 084054 (2016), arXiv:1604.04299 [hep- Rev. D 100, 024025 (2019), arXiv:1903.11591 [gr-qc]. ph]. [187] P. Brax, C. Burrage, A.-C. Davis, D. Seery, [215] M. Aaboud et al. (ATLAS), JHEP 05, 142 (2019), and A. Weltman, Phys. Lett. B 699, 5 (2011), arXiv:1903.01400 [hep-ex]. arXiv:1010.4536 [hep-th]. [216] S. Trojanowski, P. Brax, and C. van de Bruck, Phys. [188] P. Brax, C. van de Bruck, A.-C. Davis, and D. Shaw, Rev. D 102, 023035 (2020), arXiv:2006.01149 [hep-ph]. Phys. Rev. D 82, 063519 (2010), arXiv:1005.3735 [astro- [217] T. Brinckmann and J. Lesgourgues, Phys. Dark Univ. ph.CO]. 24, 100260 (2019), arXiv:1804.07261 [astro-ph.CO]. [189] P. Brax, A.-C. Davis, and B. Li, Phys. Lett. B 715, 38 [218] A. Gelman and D. B. Rubin, Statist. Sci. 7, 457 (1992). (2012), arXiv:1111.6613 [astro-ph.CO]. [219] D. A. Dicus, E. W. Kolb, V. L. Teplitz, and R. V. [190] J. Wang, L. Hui, and J. Khoury, Phys. Rev. Lett. 109, Wagoner, Phys. Rev. D 18, 1829 (1978). 241301 (2012), arXiv:1208.4612 [astro-ph.CO]. [220] S. Moriyama, Phys. Rev. Lett. 75, 3222 (1995), [191] H. Primakoff, Phys. Rev. 81, 899 (1951). arXiv:hep-ph/9504318. [192] J. Noller, JCAP 07, 013 (2012), arXiv:1203.6639 [gr-qc]. [221] P. Brax, Phys. Lett. B 712, 155 (2012), arXiv:1202.0740 [193] G. Obied, H. Ooguri, L. Spodyneiko, and C. Vafa, [hep-ph]. (2018), arXiv:1806.08362 [hep-th]. [222] N. Vinyoles, A. Serenelli, F. L. Villante, S. Basu, [194] P. Brax, C. van de Bruck, and A.-C. Davis, Phys. Rev. J. Redondo, and J. Isern, JCAP 10, 015 (2015), D 101, 083514 (2020), arXiv:1911.09169 [hep-th]. arXiv:1501.01639 [astro-ph.SR]. [195] P. Brax, P. Brun, and D. Wouters, Phys. Rev. D 92, [223] G. P. Briggs, L. Ferrario, C. A. Tout, and D. T. Wickra- 083501 (2015), arXiv:1505.01020 [astro-ph.HE]. masinghe, Mon. Not. Roy. Astron. Soc. 478, 899 (2018), [196] A. Caputo, A. J. Millar, and E. Vitagliano, Phys. Rev. arXiv:1804.09910 [astro-ph.SR]. D 101, 123004 (2020), arXiv:2005.00078 [hep-ph]. [224] A.-C. Davis, E. A. Lim, J. Sakstein, and D. Shaw, [197] C. A. J. O’Hare, A. Caputo, A. J. Millar, and Phys. Rev. D 85, 123006 (2012), arXiv:1102.5278 [astro- E. Vitagliano, Phys. Rev. D 102, 043019 (2020), ph.CO]. arXiv:2006.10415 [astro-ph.CO]. [225] B. Jain, V. Vikram, and J. Sakstein, Astrophys. J. 779, [198] J. N. Bahcall and M. H. Pinsonneault, Phys. Rev. Lett. 39 (2013), arXiv:1204.6044 [astro-ph.CO]. 92, 121301 (2004), arXiv:astro-ph/0402114. [226] A. P. Naik, E. Puchwein, A.-C. Davis, and C. Arnold, [199] M. Kuster, G. Raffelt, and B. Beltran, eds., Axions: Mon. Not. Roy. Astron. Soc. 480, 5211 (2018), Theory, cosmology, and experimental searches. Pro- arXiv:1805.12221 [astro-ph.CO]. ceedings, 1st Joint ILIAS-CERN-CAST axion training, [227] A. P. Naik, E. Puchwein, A.-C. Davis, D. Sijacki, and Geneva, Switzerland, November 30-December 2, 2005, H. Desmond, Mon. Not. Roy. Astron. Soc. 489, 771 Vol. 741 (2008). (2019), arXiv:1905.13330 [astro-ph.CO]. [200] M. Asplund, N. Grevesse, A. J. Sauval, and [228] A. P. Naik, N. W. Evans, E. Puchwein, H. Zhao, P. Scott, Ann. Rev. Astron. Astrophys. 47, 481 (2009), and A. C. Davis, Phys. Rev. D 102, 084066 (2020), arXiv:0909.0948 [astro-ph.SR]. arXiv:2002.05738 [astro-ph.GA]. [201] A. Serenelli, S. Basu, J. W. Ferguson, and M. Asplund, [229] H. Desmond and P. G. Ferreira, Phys. Rev. D 102, Astrophys. J. Lett. 705, L123 (2009), arXiv:0909.2668 104060 (2020), arXiv:2009.08743 [astro-ph.CO]. [astro-ph.SR]. [230] C. Fu et al. (PandaX), Phys. Rev. Lett. 119, 181806 [202] F. L. Villante, A. M. Serenelli, F. Delahaye, and (2017), arXiv:1707.07921 [hep-ex]. M. H. Pinsonneault, Astrophys. J. 787, 13 (2014), [231] X. Zhou et al. (PandaX-II), (2020), 10.1088/0256- arXiv:1312.3885 [astro-ph.SR]. 307X/38/1/011301, arXiv:2008.06485 [hep-ex]. [203] S. Vagnozzi, K. Freese, and T. H. Zurbuchen, Astro- [232] D. S. Akerib et al. (LUX), Phys. Rev. Lett. 118, 261301 phys. J. 839, 55 (2017), arXiv:1603.05960 [astro-ph.SR]. (2017), arXiv:1704.02297 [astro-ph.CO]. [204] A. Serenelli, P. Scott, F. L. Villante, A. C. Vincent, [233] J. Sakstein, Astrophysical Tests of Modified Grav- M. Asplund, S. Basu, N. Grevesse, and C. Pena- ity, Ph.D. thesis, Cambridge U., DAMTP (2014), Garay, Mon. Not. Roy. Astron. Soc. 463, 2 (2016), arXiv:1502.04503 [astro-ph.CO]. arXiv:1604.05318 [astro-ph.SR]. [234] P. Brax, Rept. Prog. Phys. 81, 016902 (2018). [205] S. Vagnozzi, Atoms 7, 41 (2019), arXiv:1703.10834 [235] P. Binetruy, Phys. Rev. D 60, 063502 (1999), arXiv:hep- [astro-ph.SR]. ph/9810553. [206] Y. Fan, Living Reviews in Solar Physics 6, 4 (2009). [236] P. Brax, C. van de Bruck, and A.-C. Davis, Phys. Rev. [207] P. Charbonneau, Living Reviews in Solar Physics 7, 3 Lett. 99, 121103 (2007), arXiv:hep-ph/0703243. (2010). [237] C. Blancard, P. Coss´e, and G. Faussurier, Astrophys. [208] N. J. Wright and J. J. Drake, Nature (London) 535, 526 J. 745, 10 (2011). (2016), arXiv:1607.07870 [astro-ph.SR]. [238] M. Krief, A. Feigel, and D. Gazit, Astrophys. J. 821, [209] G. G. Raffelt, Phys. Rev. D 33, 897 (1986). 45 (2016), arXiv:1601.01930 [astro-ph.SR]. [210] S. Dimopoulos, J. A. Frieman, B. W. Lynn, and G. D. [239] M. Zumalacarregui, T. S. Koivisto, and D. F. Mota, 24

Phys. Rev. D 87, 083010 (2013), arXiv:1210.8016 [astro- arXiv:0707.2801 [hep-ph]. ph.CO]. [245] P. Brax, A.-C. Davis, B. Li, and H. A. Winther, [240] J. Sakstein, JCAP 12, 012 (2014), arXiv:1409.1734 Phys. Rev. D 86, 044015 (2012), arXiv:1203.4812 [astro- [astro-ph.CO]. ph.CO]. [241] J. Sakstein, Phys. Rev. D 91, 024036 (2015), [246] S. Dimopoulos, G. D. Starkman, and B. W. Lynn, Mod. arXiv:1409.7296 [astro-ph.CO]. Phys. Lett. A 1, 491 (1986). [242] H. Y. Ip, J. Sakstein, and F. Schmidt, JCAP 10, 051 [247] M. Pospelov, A. Ritz, and M. B. Voloshin, Phys. Rev. (2015), arXiv:1507.00568 [gr-qc]. D 78, 115012 (2008), arXiv:0807.3279 [hep-ph]. [243] J. Sakstein and S. Verner, Phys. Rev. D 92, 123005 [248] W. J. Veigele, Atom. Data Nucl. Data Tabl. 5, 51 (2015), arXiv:1509.05679 [gr-qc]. (1973). [244] P. Brax, C. van de Bruck, A.-C. Davis, D. F. Mota, and D. J. Shaw, Phys. Rev. D 76, 085010 (2007),