sign in sign up
<<
Home , Lamb shift, Standard Model

LAPTH-028/17, MIT-CTP/4929

Probing new -independent interactions through precision in with few

C´edricDelaunay,1 Claudia Frugiuele,2 Elina Fuchs,2 and Yotam Soreq3 1Laboratoire d’Annecy-le-Vieux de Physique Th´eoriqueLAPTh, CNRS – USMB, BP 110 Annecy-le-Vieux, F-74941 Annecy, France 2Department of and , Weizmann Institute of , Rehovot 7610001, Israel 3Center for , Massachusetts Institute of Technology, Cambridge, MA 02139, U.S.A. The very high precision of current measurements and theory of spectral lines in few- atoms allows to efficiently probe the existence of exotic between electrons, and . We investigate the sensitivity to new spin-independent interactions in transition fre- quencies (and their isotopic shifts) of hydrogen, helium and some helium-like ions. We find that present data probe new regions of the -carrier couplings to electrons and neutrons around the MeV range. We also find that, below few keV, the sensitivity to the electron coupling in precision spectroscopy of helium and is comparable to that of the anomalous magnetic moment of the electron. Finally, we interpret our results in the dark- model where a new gauge is kinetically mixed with the photon. There, we show that helium transitions, combined with the anomalous magnetic moment of the electron, provide the strongest indirect bound from laboratory experiments above 100 keV.

I. INTRODUCTION ing the experimental precision to a bound on NP is the theory uncertainty, which is by far larger than the exper- imental precision. However, the high accuracy achieved Fundamental interactions of known elementary parti- in frequency measurements of narrow atomic transitions cles are well described by the the (SM) in heavy elements can be exploited to probe NP provided of . Nevertheless, the SM cannot be the new observables largely insensitive to theory uncertain- complete description of Nature since it does not account ties are identified. For example, Refs. [16–19] proposed to for oscillations, does not provide a viable dark bound new interactions between neutrons and electrons candidate and cannot explain the asym- by testing linearity of King plots [20] of isotope shift (IS) metry of the . Moreover, the SM suffers from measurements. several hierarchy problems, such as the stability of the The situation is different for atoms and ions with few Higgs mass to quantum corrections and the strong CP electrons and a small number of . There, QED problem. In addition, it contains intriguing puzzles re- calculations are carried out to high accuracy. For in- lated to the observed large hierarchies in the charged stance, low excited states in helium can be calculated and mixing angles. However, non up to O(α6) corrections, which yields theoretical predic- of these pending issues which call for new physics (NP) tions often more accurate than the measurements, see beyond the SM indicate a specific scale at which e.g. Refs. [21, 22]. In addition, nuclear finite size ef- the associated NP phenomena will manifest themselves. fects are also well described thanks to the extraction of Hence, a broad experimental program must be pursued. the nuclear radius from electron scattering experiments While accelator-based experiments search directly for and muonic spectroscopy. This allows to directly new over many orders of the mass and couplings, compare theory and experiment in order to probe NP precision low-energy measurements may also reveal in- interactions [23–26]. arXiv:1709.02817v1 [hep-ph] 8 Sep 2017 directly in a complementary approach the existence of In this work we derive new limits on spin-independent new phenomena. Precise table-top exper- NP interactions between the / and the iments are promising in this regard. For example, ob- electron from optical frequency measurements, as well servables which violate discrete symmetries of QED, as as from frequency shifts between different isotopes, in in -violating transitions [1–8], are a powerful tool hydrogen and helium atoms, helium-like ions such as to probe NP [9]. lithium and nitrogen. In addition, we constrain new Parity-conserving transitions are also interesting electron-electron interactions via precision spectroscopy probes of NP which, however, require higher theoretical of helium and positronium atoms. We compare the re- control. Frequency measurements of narrow atomic tran- sulting bounds on the product of the electron and neu- sitions in heavy elements are possible within a few×10−16 tron couplings and on the electron coupling separately, relative accuracy [10, 11] thanks to the optical frequency with existing indirect constraints and future prospects comb technique [12, 13]. Moreover, the experimen- from IS measurement in heavier systems with many elec- tal uncertainty in some systems now reaches the 10−18 trons. Furthermore, we perform global fits using all avail- level [14, 15], thus indicating the possibility of significant able transitions to constrain simultaneously the electron, improvement in the near future. A limitation in translat- neutron and proton couplings in a model-independent 2

2 way. Finally, as a phenomenological application, we in- the leading finite nuclear size effects, hr iA being the nu- terpret the bounds within the model. clear radius squared and Fi is the field-shift (FS) constant. The IS between two isotopes A and A0 for this transition is then described as II. CONSTRAINING NEW 0 0 SPIN-INDEPENDENT INTERACTIONS IN A,A A,A 2 0 νi = νi,0 + Fiδhr iA,A0 + yeynXi(A − A ) , (5) ATOMS 0 A,A A A0 2 2 2 where νi ≡ νi − νi and δhr iA,A0 ≡ hr iA − hr iA0 . Consider a new force mediated by a boson φ of mass Higher-order effects of nuclear and finite mφ and spin s = 0, 1, 2 with spin-independent couplings magnetic radius are not resolvable by the current exper- to the electron, proton and neutron ye, yp and yn, respec- imental accuracy [21, 30] and are therefore omitted in tively. At atomic , the exchange of φ is described Eqs. (4) and (5). by an effective potential between a nucleus N of charge Absolute frequencies and IS in hydrogen and helium Z and mass number A and its bound electrons as are calculated to high accuracy in the limit of point-like (−1)s+1 e−mφr nuclei. However, the full theoretical is often V (r) = y y , (1) N 4π e N r limited by the uncertainty related to the finite nuclear size effects. For a comparison between the experimental where yN ≡ ypZ +(A−Z)yn and r is the electron-nucleus value and the QED prediction in the point-nucleus limit, distance. we define For atoms with more than one electron, like helium, an A A A effective potential between electron pairs is also induced ∆i ≡ νi,exp − νi,0 . (6)

2 −mφr12 s+1 ye e The NP contribution generically depends on the three Ve(r12) = (−1) , (2) 4π r12 coupling constants, ye, yp, and yn, and the mediator mass, m . At fixed m the product y y can be probed where r ≡ |~r − ~r | is the distance between electron 1 φ φ e n 12 1 2 independently of y from a single IS measurement (for and electron 2, with ~r , ~r describing their positions rel- p 1 2 any transition i) using Eq. (5) ative to the nucleus. 0 The total frequency shift induced by the above poten- A,A 2 tials for a transition i between atomic states a and b (with ∆i − Fiδhr iA,A0 yeyn = 0 , (7) Eb > Ea) is described by first-order perturbation theory Xi(A − A ) as 0 0 where ∆A,A ≡ ∆A − ∆A using Eq. (6). Hence the NP A s+1 2  i i i δNPνi = (−1) yeyN Xi + ye Yi , (3) bound depends on the change in the mean-square nu- 2 clear charge radius, δhr i 0 , which is measured either where X and Y are overlap integrals with the electronic A,A i i in electron scattering or in muonic atom spectroscopy wavefunctions depending only on the φ boson mass. In experiments. Whenever applicable, the latter typically this paper, we focus on electronic transitions in helium- yields much more precise values of the charge radii. In like and hydrogen atoms, for which the X and Y func- i i principle, the charge radius determination via electron tions are calculated within first-order perturbation the- scattering may be affected by NP. However, we find that ory using non-relativistic wavefunctions as detailed in NP is only noticeable there for large coupling values that AppendicesB andC. While QED calculations rely on are already excluded by more sensitive probes. Hence much more sophisticated wavefunctions, the use of the the charge radius extraction from electron scattering can- non-relativistic ones is a sufficient approximation to the not be contaminated by NP. Muonic atom spectroscopy dominant NP effects. Helium wavefunctions are very well measurements are more sensitive to NP contributions, approximated by antisymmetrized combinations of hy- especially in the keV–MeV mass range, and existing con- drogenic wavefunctions with effective nuclear charges ac- straints on the y y coupling product do not rule out counting for the electronic screening [27], except for the µ n the possibility of NP contaminations in this region [31– 2S spin-singlet state where an accurate description of 33]. Therefore, for simplicity we will henceforth assume the inter-electron repulsion requires the use of Hylleraas y = 0. functions [28, 29]. µ Alternatively, the charge radius dependence can be Taking into account the NP contribution in Eq. (3), eliminated using an IS measurement in a second tran- the theory prediction for the frequency of an electronic sition, yielding transition i in an isotope A is given by 0 0 A A 2 A A,A A,A νi =νi,0 + Fihr iA + δNPνi , (4) F2∆1 − F1∆2 yeyn = 0 , (8) (F2X1 − F1X2)(A − A ) A where νi,0 is the dominant contribution calculated in the point-like nucleus limit (including spin effects and nu- which, besides X1,2, depends only on quantities known clear ), whereas the second term describes theoretically with high accuracy, namely F2/F1 and 3

A,A0 νi,0 . The main drawback of Eq. (8) relative to Eq. (7) of the electron ae ≡ (g−2)e/2 [37, 38], neutron scattering is the possible loss of sensitivity when (F2X1 − F1X2) → on atomic electrons [39] or nuclei [40–43] and fifth force 0 [16, 19]. The latter is rather severe for all mφ when experiments [44, 45], as well as astrophysical constraints close-by transitions with F1 ≈ F2 and X1 ≈ X2 are used, from SN 1987a [46] and globular clusters [47–51]. Note for example in transitions involving different states of the that also spectroscopy in molecular ions and antiprotonic same fine-structure multiplet. Another disadvantage of helium constrains spin-independent interactions of nucle- Eq. (8) is represented by how rapidly the sensitivity to ons [52–55], but it results in weaker bounds than from −2 neutron scattering and is therefore not shown in Fig.1. yeyn weakens at large mass. While Xi ∝ mφ above Some of these constraints can be evaded in specific mod- mφ ∼ O(10 keV), this leading term cancels out in the difference (F X − F X ) ∝ m−3 [19]. Equation (8) els [56–60]. We notice that bounds from few-electron 2 1 1 2 φ atoms provide the strongest (indirect) constraints in the bears resemblance to the method proposed in Ref. [19] region above 300 keV where astrophysical bounds lose for heavy elements. However, there, the less accurate A,A0 sensitivity. Note, however, that direct constraints (not theory calculations for νi,0 are traded for IS measure- shown here) also exist, which are particularly sensitive ments between two additional isotope pairs. for mφ > 1 MeV. Yet, they strongly depend on the as- NP contributions to the electron-electron and electron- sumed branching ratios for relevant φ decay modes, see proton interactions cancel out to first approximation in e.g. Ref. [61] for a review. In the near future a higher the IS. Thus, they can be probed more efficiently in ab- sensitivity to NP is still expected in the King linearity solute frequency measurements, despite the lower abso- test of Yb+ [19] compared to few-electron atom spec- lute accuracy. Helium transitions are sensitive to both troscopy, despite the projected improvements in helium kinds of interaction. In fact a combination of two fre- transitions. In the following subsections, we discuss in 2 quency measurements can resolve the yeyN and ye cou- detail the bounds obtained from IS measurements in he- pling products, thanks to transition-dependent Xi,Yi lium, helium-like ions and hydrogen/deuterium atoms. constants in Eq. (3). Hydrogen frequencies are sensi- tive to electron-proton interactions and have been used previously to probe yeyp [23–25]. However, presently un- A. Helium and helium-like isotope shifts resolved issues related to the well-known proton radius puzzle [34, 35] limits the application of hydrogen spec- We derive here IS bounds using precision spectroscopy troscopy for constraining new atomic forces. Therefore, in two-electron atoms. This includes constraints from we will not use absolute spectroscopy measurements of measurements in helium and helium-like lithium and ni- hydrogen or deuterium as a probe of NP. Instead, we trogen ions, all of which are presented in Fig.2, while the will consider hydrogen-deuterium IS spectroscopy since 2 strongest one is also reported in Fig.1 for comparison in this case there is no tension between δhr i values ex- with constraints from other atoms and different sources. tracted from electronic and muonic measurements [36]. The most accurate IS in helium are measured within a few kHz uncertainty between A = 3, 4 for the 21S − 23S [63] and 23P − 23S [64, 65] transitions around III. BOUNDS FROM ISOTOPE SHIFT 1557 nm and 1083 nm, respectively. While QED calcu- MEASUREMENTS lations in the point-nucleus limit reached sub-kHz ac- curacy, the theory prediction for the IS is limited by 2 Let us first discuss probes of new electron-neutron in- the charge radius difference δhr i3,4 [21]. The latter can teractions. We focus here on spectroscopic probes based be extracted within a few percent from e-He scattering on IS measurements in helium, helium-like and hydro- data [66], gen/deuterium atoms. The comparison of theory to ex- 2 e−scat 2 δhr i3,4 = (1.067 ± 0.065) fm . (9) periment directly probes yeyn independently of the pres- ence of a NP coupling to protons. As shown in TableI (a Using Eq. (9) as an input for the theory predictions of full list of our input values is given in AppendixA), the He IS yields a good agreement between theory and ex- theory uncertainty (in the point-like nucleus limit) is cur- periment for both transitions, thus allowing to constrain rently smaller than the experimental error for transitions NP electron-neutron interactions. involving low excited states, and the sensitivity to NP A higher sensitivity could be reached by combining the is limited by the experimental determination of charge 2 two transitions in order to eliminate δhr i3,4. In that radius differences. Our results are summarized in Fig.1 case, Eq. (8) results in which show the best constraints on yeyn, as a function of the mediator mass, arising from transitions in hydrogen (−51 ± 14) kHz and helium(-like) atoms. yeyn ≈ , (10) (5.7X1557 − X1083) For comparison, we also show constraints derived from King linearity in heavy atoms [19] and from other (non- which is ∼ 4σ away from zero. Thus, it is not justi- atomic) observables. Those include laboratory con- fied, given such a disagreement, to use the above to set straints resulting from the anomalous magnetic moment limits on NP. Note, however, that this large deviation 4

the 21S −23S transition down to O(100 Hz) precision are 1 2S-2S 3,4He(e-scat) expected [68], with a comparable theory improvement. 2S-2S 3,4He(projection) Hence, this would potentially improve sensitivity to NP 0.01 1S-2S HD(e-scat) 1S-2S HD( LS) effects for that transition by two orders of magnitude. As King plot Ca+ shown in Fig.1, this is still weaker than the sensitivity King plot Yb+ (projection) -4 10 globular cluster expected from King linearity violation in ytterbium ions, except for m 10 MeV due to the different scal- - φ & n-scattering on atomice 2 3 -6 ing of the bound with the mediator mass (m versus m ). 10 n-scattering on nuclei+a e φ φ n

y -8 Precision measurements are also achievable in heavier

e 10 y (unstable) helium isotopes. For instance, IS between 3 3 -10 A = 4 and A = 6, 8 isotopes for the 2 S − 3 P transition 10 (389 nm) are measured with ∼ 100 kHz accuracy [69]. However, the situation is different here since there is 10-12 no independent measurement of the 6,8He charge radii and the FS cannot be reliably predicted for the 389 nm 10-14 bounds on new transition. Nevertheless one can still derive an upper electron-neutron interaction bound on NP by saturating the difference between 10-16 theory (assuming a point-like nucleus) and experiment, 10-6 10-3 1 103 2 which corresponds to setting δhr iAA0 = 0 in Eq. (7). mϕ [MeV] 8,4 Since ∆389 = −0.918 MHz [69], the NP contribution is not strongly constrained. Yet, the resulting bound on 0 FIG. 1: Summary of the indirect constraints on a new yeyn is strengthened by a factor of A − A = 4 which electron-neutron interaction: isotope shift spectroscopy of he- makes it comparable to the IS bound from the 1083 nm lium and hydrogen/deuterium (this work) with the charge ra- transition. An order of magnitude improvement could dius determined from electron scattering (“e-scat”) or Lamb be obtained with an independent determination of the shift in muonic atoms (“muon LS”); comparison with King charge radii of A = 6, 8 isotope of helium. plot analyses of IS in heavy atoms (Ca+, Yb+)[19, 62], fifth force experiments [44, 45], electron-neutron scattering [39], Finally, IS in helium-like ions are also well mea- neutron-nucleus scattering [40–43] combined [19] with ae, and globular cluster [50]. The existing bounds are in solid lines, sured. The highest accuracy is obtained in singly-ionized while the projections are in dashed lines. The He projection lithium [70] and five- ionized nitrogen [71]. The assumes a combined theory and experimental uncertainty of measured frequency shifts are between A = 6, 7 in the 100 Hz; the Yb+ projection assumes King linearity at 1 Hz. 23S − 23P transition for Li+, and between A = 14, 15 in the 21S − 23P transition for N5+. We rely on the theory predictions used in the quoted references. This assumes nuclear charge radii determined from electron-scattering is the mere consequence of a known tension between the data [72] with a relative accuracy of ∼ 2 % for lithium two transitions which may originate from underestimated and from electron-scattering data [73] and muonic X-ray uncertainties [21]. Despite this circumstance, it remains line measurements [74] with ∼ 0.5 % for nitrogen. The interesting to observe that in the case that the tension resulting bounds are weaker than the ones from helium. will be resolved by refined QED calculations and/or mea- Further precision measurements with helium-like boron surements, the expected sensitivity to yeyn is stronger by and carbon ions are also underway [21]. 2 e−scat a factor ∼ 6 relative to the use of δhr i3,4 . In the (yet implausible) event that the above deviation is an evidence for a new electron-neutron interaction, the latter should be visible in other atomic systems. For instance, Eq. (10) B. Hydrogen-deuterium shifts would imply a violation of King linearity in ytterbium ion clock transitions at the O(100 Hz) level [19]. Hydrogen-deuterium shifts are complementary probes 2 Alternatively, δhr i3,4 can be extracted with high ac- of new electron-neutron interactions. The most accurate curacy from muonic helium spectroscopy. The CREMA IS measurement is for the 1S − 2S transition (121.6 nm), collaboration is currently conducting mea- with ∼ 10−11 relative uncertainty [75, 76]. The QED surements in muonic He+ aiming at a determination calculation is less precise by a factor of ∼ 60, being of 3,4He charge radii with a relative uncertainty of equally limited by the experimental value of the proton- −4 2 3 × 10 [67]. Assuming this will result in a δhr i3,4 to-electron and deuteron-to-electron mass ratios as well value consistent with e-He scattering and (electronic) as higher-order corrections to the Lamb shift and nu- helium spectroscopic data, the sensitivity to NP will clear polarizability [75]. Additional IS measurements ex- hence be limited by the experimental accuracy in helium ist with lower precision, including the 2S − nS/D transi- IS measurements. Moreover, future IS measurements in tion series for n = 8, 12 states [77–79], and the frequency 5

cancels out, thus making it directly sensitive to Lamb isotopes transition δ ν σ σ σ 2 NP νexp ν0 δhr i L shift (LS) corrections. As a result, νLS becomes less sen- 21S − 23S +9 ± 14 2.4 0.19 14 3He/4He sitive to NP with an interaction range longer than the 23P − 2S −2 ± 78 3.3 0.9 78 −1 −1 atomic size ∼ a0 = (αme) ≈ (4 keV) . Since all 1S − 2S +76 ± 61 0.02 0.9 61 transitions in the 2S −nS/D series have comparable sen- H/D 2S − 12D +1.2 ± 10 9.3 4.2 sitivity to NP, we consider only the 2S − 12D transition for illustration. Here again, the FS contributions are least known the-

TABLE I: Allowed NP contributions δNPν for the most accu- oretically as they are limited by the charge radius differ- 2 rate IS measurements in helium and hydrogen isotopes, along ence δhr i2,1 between the deuteron and the proton. The with the standard uncertainties from experiment, σνexp , QED latter can be extracted either from electron scattering 1 calculation (point-nucleus limit, σν0 ) and charge radius differ- data , which yields [82] ence extracted from electron-scattering data, σδhr2i. Only for 2S −12D in H/D, σν refers to the complete theory prediction 2 e−scat 2 0 δhr i2,1 =(3.764 ± 0.045) fm , (12) including the FS. All numbers are in kHz. For references and input values, see TablesIII andIV. or muonic hydrogen/deuterium spectroscopy [36]

2 µ 2 1 δhr i2,1 = (3.8112 ± 0.0034) fm . (13)

2S-2S 3,4He(e-scat) 3,4 Note that the charge radius differences in Eqs. (12) 0.01 2S-2S He(projection) 2S-2P 3,4He(e-scat) and (13) are consistent within uncertainties, despite the 2S-2S/2S-2P 3,4He (still puzzling) significant discrepancies between muonic 4,8 10-4 2S-3P He and electronic determinations of the proton [34, 35] and 6,7 + 2 µ 2S-2P Li (e-scat) deuteron [36] radii. Using δhr i2,1 to predict the FS con- 2S-2P 14,15N 5+ (e-scat) tribution yields a sensitivity to NP larger by a factor ∼ 13 10-6 2 e−scat relative to δhr i2,1 , assuming the radii extraction from n y

e muonic spectroscopy is not affected by a possible NP cou- y pling to . 10-8 The IS bounds on a new electron-neutron interaction from hydrogen/deuterium are summarized in Fig.3. 10-10

10-12 IV. BOUNDS FROM ABSOLUTE FREQUENCY isotope shift bounds MEASUREMENTS helium(-like) atoms 10-14 10-6 10-3 1 103 While IS are only sensitive to electron-neutron inter-

mϕ [MeV] actions, absolute frequencies can also probe the electron- proton and, in atoms with more than one electron or in FIG. 2: Isotope shift bounds on new electron-neutron inter- positronium, electron-electron interactions. As we dis- action from helium spectroscopy. The solid red and blue lines cussed above, by measuring two transitions one can ex- are the limits from the 1557 nm and 1083 nm transitions, re- tract ye and yN separately, and the combination with the spectively, using the charge radii from e-He scattering. The IS data will also allow for a separation of yp from yn. dotted-dashed purple line is an illustration of the potential In case φ couples both to protons and neutrons with a limit obtained by combining the two transitions (due to a similar strength, as in a Higgs portal or gauged B−L, the 4σ tension, see text for details). The solid orange line is sensitivity to probe NP with IS is expected to be stronger the upper bound derived from IS in unstable isotopes. The than from the absolute frequency measurements. This dashed red line represents the projected sensitivity with the can be understood as follows. In light atoms, the NP con- 1557 nm transition using charge radii from muonic helium and tributions to the IS and to the absolute frequency are of assuming a combined experimental and theory uncertainty of 100 Hz. the same order. However, typically the absolute accuracy of IS data (theory and experiment) is better by at least an order of magnitude than the absolute frequency data, see AppendixA. Thus, for yn ≈ yp IS measurements are differences [80] a more sensitive to NP than the absolute frequencies. 1 νL ≡ ν − ν , (11) LS 2S−4L 4 1S−2S with L = S, D. The latter is constructed such that 1 We use here the proton radius value extracted from the so-called the leading contribution from Coulomb-like potentials Mainz data [81]. 6

1 1

4 1S-2S(e-scat) 2S-2S He 2P-3D 4He 1S-2S(muon LS) 0.01 0.1 2S-3D 4He 2S-12D(e-scat) 2S-2P 4He 1S-2S ortho-Ps νS -4 LS 10 ae () νD LS 0.01 ae (vector)

10-6 n

e -3 y

y 10 e y 10-8

10-4 10-10

10-5 10-12 isotope shift bounds electron-electron hydrogen/deuterium interaction bounds 10-14 10-6 10-6 10-3 1 103 10-6 10-3 1 103

mϕ [MeV] mϕ [MeV]

FIG. 3: Isotope shift bounds on new electron-neutron interac- FIG. 4: The upper bound on ye as function of mφ from 3 3 tion from hydrogen/deuterium spectroscopy. The blue (pur- 1 S1 − 2 S1 in positronium [83, 84] (cyan), helium data [21] ple) line is the limit from the 1S − 2S transition using charge and a comparison to electron magnetic moment, ae, for a radii from electron scattering (muonic spectroscopy). The or- scalar (dark gray) and vector (light gray). ange and green lines are limits derived from the 2S−12D tran- S,D sition and the νLS observables, respectively, using electron- scattering data. See text for details. which makes it more prone to cancelation against addi- tional contributions from other states present in a com- plete NP model. Note that the helium bounds in Fig.4 It is important to distinguish between the case of a are evaluated by assuming no electron-nucleus interac- generic new force coupled to the electron and to the nu- tions. We have verified that marginalizing over the lat- cleus (including the proton) and the case of a dark photon ter does not significantly change the bounds. The bounds (kinetic mixing) where the charges are proportional to from positronium and helium are comparable and below the electric charges and a more careful treatment of the few keV are weaker than the bound from ae only a factor definition of the electromagnetic coupling, α, is required, of few. see Ref. [25].

B. Model-independent bounds on ye, yp and ye A. Bounds on ye from helium and positronium Here we combine observables from different atoms to The electron-electron interaction can be probed in probe the NP couplings yp, yn and ye independently. In atoms with more than one electron, the simplest is he- order to do so we perform a global fit based on a χ2 func- lium, or in purely electronic systems such as positronium. tion constructed from IS in hydrogen and helium as well 3 3 2 Starting with the positronium, the 1 S1 −2 S1 interval is as absolute transition frequencies in helium. Our χ is measured at the 10−9 level [83] in a agreement with the composed of the 21S − 23S and 23P − 2S IS between theory prediction of Ref. [84]. For helium we combine 3He and 4He, the helium/deuterium IS in the 1S − 2S S all the transitions that are given in Table II of Ref. [21], transition and the νLS observable, and the absolute fre- where the agreement between theory and experiment is quencies considered in SectionIVA. We present in Fig.5, better than 2σ; the full list is given in AppendixA. Thus, the 95 % CL contours in the ye − yp and ye − yn planes we use the above to put upper bounds on ye as function of for several values of mφ. For each pair of couplings, we the force-carrier mass mφ . The results are presented in marginalize over the third coupling yn and yp, respec- Fig.4, where we also added the constraint from the elec- tively. The generic shape of the bounds is understood as tron magnetic moment, ae, for comparison. This shows follows. Since the overlap integrals Xi,Yi for electron- that ae is still the strongest probe among the three. Yet nucleus and electron-electron interactions are of compa- the φ contribution to ae enters only at the loop level rable order when yp,n & ye, absolute frequencies and IS 7

1 1 m m m m m ϕ = ϕ = ϕ = ϕ = 10 100 ϕ = 100 1 10 MeV MeV MeV m MeV MeV ϕ = 1 m MeV ϕ = 100 0.01 0.01 keV

m ϕ = m 100 ϕ = 1 keV keV n -4 p -4 y 10 y 10

m ϕ = 1 keV 10-6 10-6

10-8 10-8 10-8 10-6 10-4 0.01 1 10-8 10-6 10-4 0.01 1

ye ye

FIG. 5: The allowed region 95 % CL contours for fixed mediator mass (as indicated in the plot) from the global χ2 analysis of helium and hydrogen absolute frequencies as well as their isotope shifts. The left (right) panel shows contours of yn (yp) vs ye, marginalizing over yp (yn).

constrain the products yeyp and yeyn, respectively, lead- atomic transitions together or one transition with ae, the ing to contours at 45 degrees. The latter are then trun- anomalous magnetic moment of the electron. Figure6 cated once ye reaches a large value (typically ye & yp,n) shows the 95 % CL bounds that we derived from helium 2 so that the ye term dominates the NP contribution to and positronium, each combined with ae, as well as exist- absolute frequencies in helium and the bounds become ing bounds from hydrogen spectroscopy [23–25]. We find independent of yp or yn. Note that helium absolute fre- that helium and positronium bounds surpass the known quencies are in principle sensitive to a possible relative hydrogen bounds above ∼ 100 keV. We chose to present sign between the nuclear and the electron couplings. We only indirect constraints from atomic spectroscopy on the checked that either sign yields very similar bounds and kinetic-mixing parameter since those do not depend on thus, for simplicity, we present global-fit results for posi- the A0 decay mode. In the sub-MeV region, these atomic tive couplings only. probes are the most sensitive ones, after the LSND neu- trino detector which directly searches for A0 in the 3 pho- tons decay [86], and the study of star cooling in globular C. Atomic bounds on kinetic mixing clusters which excludes, for mA0 . 300 keV, mixing pa- rameter values far below the displayed range of  in Fig.6. For m 0 1 MeV, the sensitivity of atomic spectroscopy For the sake of illustration, we apply now our result A & is also much weaker compared to probes based on A0 de- to a specific NP model, that of a kinetically mixed mas- cay (either visibly or invisibly) as in electron beam-dump sive , the dark photon, denoted as A0 [85]. experiments or colliders, like BaBar (see Ref. [61] for a As a result of the mixing between the photon and the review). In conclusion, for m 0 0.3 MeV, the most sen- dark photon, A0 couples to the electromagnetic current, A & sitive indirect probe of dark photon is from combining a and its couplings to the protons, electrons and neutrons e with atomic transitions in helium. are yp,n,e = e, 0, −e, respectively, where e is the QED gauge and  is a mixing parameter. Since all A0 couplings are determined by a single param- eter, a single atomic transition would suffice to probe it. V. DISCUSSION However, when mA0 . 1/a0, the dark photon induces a 1/r atomic potential which is not distinguishable from In this work we study the sensitivity to new spin- the Coulomb one and the A0 effect is a mere redefinition independent forces of hydrogen and helium-like atoms of the fine-structure constant, α → (1 + 2)α. Hence, in considering both absolute frequency and isotope shift this regime, we need at least two observables to probe the measurements. We exploit the accuracy of both the dark photon, one of them being used to fix α. We follow measurements and the theoretical predictions achieved in here the procedure of Ref. [25] and combine either two these systems [21, 22]. We demonstrate for the first 8

Acknowledgments 1

He 2S-2S/ae He 2P-3D/ae We thank D. Budker, R. Ozeri, K. Pachucki and ortho-Ps 1S-2S/ae G. Perez for useful discussions and careful reading of 0.1 H 1S-2S/Rydberg H 1S-2S/ae the manuscript. We thank G. B´elanger,J. Berengut, H 2S-2P A. Falkowski, O. Hen, J. Jaeckel and M. Safranova for H 1S-2S/2S-8S 0.01 globular cluster useful discussions and correspondence. CD is supported by the program Initiative d’Excellence of Grenoble-Alpes University under grant Contract Number ANR-15-IDEX- -3 ϵ 10 02. The work of YS is supported by the Office of High Energy Physics of U.S. Department of Energy (DOE) under grant Contract Number DE-SC0012567. 10-4

Appendix A: Experimental Data and Theoretical Prediction 10-5 atomic bounds on dark photon In this appendix we provide all experimental data and 10-6 theoretical predictions that have been used in this paper. 10-6 10-3 1 103 We start with the nuclear charge radii in Tab.II used m A' [MeV] throughout the paper.

FIG. 6: Comparison of atomic (indirect) constraints on the element A rA [fm] method ref. mixing  between the photon and the dark photon A0 de- 1 0.8791 ± 0.0079 [88] pending on its mass mA0 . The red/blue (cyan) lines refer e-scat. to He (positronium) transitions obtained in this work. For 2 2.130 ± 0.010 [88] H/D comparison, previous bounds from H spectroscopy are shown 1 0.84087 ± 0.00039 [34, 35] µ-spec. for 1S − 2S combined with Rydberg states (purple) or ae 2 2.12562 ± 0.00078 [36] (pink) [23, 24, 87] and from one or the combination of two H 3 1.973 ± 0.016 [66] transitions (orange, green) [25]. The range of mA0 ≤ 300 keV He e-scat. 4 1.681 ± 0.004 [66] is excluded by globular clusters [50]; for mA0 & 1 MeV several direct, decay-dependent bounds also apply. 6 2.589 ± 0.039 Li e-scat. [89] 7 2.444 ± 0.042 14 2.560 ± 0.011 µ-spec. [74] N 15 2.612 ± 0.009 e-scat. [73] the power of isotope shift measurements in few-electrons atoms to constrain models where the new degree of free- TABLE II: Charge radii rA used for the IS bounds for differ- dom, φ, couples not only to the proton, but also to the ent elements with isotopes A, obtained via electron-scattering neutron as, for instance, in the B − L and Higgs por- experiments (e-scat.) or spectroscopy in muonic atoms (µ- spec.). tal models. The derived bounds represent, to date, the strongest laboratory bound on yeyn for mφ & 100 eV. For masses heavier than 300 keV, where astrophysical probes are ineffective, isotope shift spectroscopy in few-electron In Tab.III we continue with the input values used for atoms constrains new regions of the parameter space the IS bounds in Tab.I and Fig.2 of Sect.III, see also in a model-independent way. Previous works on spin- Tab. 5 of Ref. [21]. Finally, we provide the values of independent new interactions [23–25] focused on hydro- measurements and calculations of absolute frequencies in gen which is only sensitive to new interactions between Tab.IV. the electron and the proton. The highly precise spec- troscopy of helium has the advantage to probe also the Appendix B: Helium wavefunction electron coupling alone, reaching a sensitivity compara- ble to ae below few keV. (See Ref. [26] for similar results regarding spin-dependent electron-electron interactions.) In this appendix we specify the approximate wavefunc- Furthermore, we show that current precision in positro- tions we use in the helium calculations. It is conventional nium spectroscopy has comparable constraining power. to label the states of helium as 2S+1 The present work emphasizes how the effort in improv- n LJ (B1) ing the knowledge of the nuclear size has the indirect ef- fect of improving the sensitivity to new spin-independent corresponding to the following electronic configuration forces between the constituents of the atoms. (1s)(nl). L is the total orbital momentum, S is the spin 9

0 AA0,exp AA0,th 2 A, A transition i νi [kHz] ref. νi,0 [kHz] Fi [kHz/fm ] ref. 21S1/2 − 23S3/2 -8 034 286.259±2.4 -8 034 065.91±0.19 -214.66±0.02 3,4He [90–92] [30, 93–95] 23P − 2S -33 668 444.7±3.2 -33 667 149.3±0.9 -1 212.2 ± 0.1 4,8 He 2S − 3P2 64 701 466±52 [96] 64 702 409 1 008 [96] H/D 1S − 2S 670 994 334.605 ± 0.015 [76] 670 999 566.90±0.89 -1 369.88 [76] 14,15 1 3 N 2 S0 − 2 P1 649 418 424.16±29 979.2 [71] 649 469 388.8 ± 269 812.8 [71]

6,7 3 3 Li 2 P0 − 2 S1 3 474 773±55 [70] 34 747 876 [70]

AA0,exp TABLE III: Theoretical and experimental input values for the IS bounds: measured IS νi , theory prediction for a point- AA0,th like nucleus νi,0 , and the field shift constant Fi. See also Ref. [21]. For the IS of 2S − 12D in H/D, see the absolute frequencies in Tab.IV. F1S−2S for H/D is an approximate value obtained from the quoted FS in Ref. [76].

exp th element transition i νi [kHz] ref. νi [kHz] ref H 799 191 727 402.8 ± 6.7 [77] 799 191 727 409.1 ± 3.0 2S1/2 − 12D3/2,5/2 [97] D 799 409 184 967.6 ± 6.5 [77] 799 409 184 973.4 ± 3.0

H 1 4797338 ± 10 4 797 329 ± 5 (4S1/2 − 2S1/2) − (2S − 1S) D 4 4 801 693 ± 20 4 801 692 ± 5 [80] [80] H 1 6 490 144 ±24 6 490 128 ± 5 (4S1/2 − 2D5/2) − (2S − 1S) D 4 6 494 841 ± 41 6 494 816 ± 5 1 3 2 S0 − 2 S1 192 510 702 145.6 ± 1.8 192 510 703 400 ± 800 3 3 2 P0 − 3 D1 510 059 755 352 ± 28 510 059 754 000 ± 700 4He [21] [21] 3 3 2 S1 − 3 D1 786 823 850 002 ± 56 786 823 848 400 ± 1 300 1 1 2 S0 − 2 P1 145 622 892 886 ± 183 145 622 891 500 ± 2 300 3 3 Ps 1 S1 − 2 S1 1 233 607 216 400 ± 3200 [83] 1 233 607 222 180 ± 580 [84]

TABLE IV: Measurements and predictions of absolute transition frequencies in H, D, He and positronium (Ps). The H and D values are used for the H/D IS in Sect.IIIB. The middle part of the table summarizes the experimental and theoretical 1 frequencies of the Lamb shift (4L − 2S1/2) − 4 (2S − 1S) for L = S, D. The lower part of the table is used for constraining ye in Fig.4: the transitions in 4He with an agreement of better than 2σ between theory and experiment, as well as a transition in Ps. 10 and J ≤ L + S is the total angular momentum. Since state Zi Za one electron is always in the (1s) orbital L = l, and 1 S = 0, 1 corresponding to the singlet and triplet states, (2) S 2.08 1.21 respectively. (2)3S 2.01 1.53 For two-electron systems in the non-relativistic limit, (2)3P 2.00 0.97 the spin and spatial parts of the wavefunction are factor- 3 ized. The spin singlet (S = 0) state is (2) P 1.99 1.09 n ≥ 3 2 1 | ↑↓i − | ↓↑i |S = 0, mS = 0i = √ , (B2) 2 TABLE V: Effective nuclear charges for the excited states of helium under consideration. Zi is the charge of the core (1s) while the spin triplet (S = 1) is with components electron and Za is the charge of the excited electron. These are obtained by variational methods using the non-relativistic |S = 1, mS = 1i = | ↑↑i , hydrogen wavefunctions as trial functions. When the electron | ↑↓i + | ↓↑i is excited to a n = 3 or higher orbital, the screening of the |S = 1, mS = 0i = √ , core electron is found nearly perfect. 2

|S = 1, mS = −1i = | ↓↓i . (B3) combination of angular momentum using the Clebsch- Using an antisymmetrized combination of hydrogenic or- Gordan coefficients CJ,mJ as L,m,S,mS bitals, the spatial part of the wavefunction takes the form L S X X |n2S+1L i = CJ,mJ |ψ i|S, m i. S 1 1 J,mJ L,m,S,mS nLm S h~r1, ~r2|ψnlmi = √ × 2 π p2(1 + N ) m=−L mS =−S nl (B8)  S  Fnl(r1, r2)Ylm(Ω2) + (−1) Fnl(r2, r1)Ylm(Ω1) , (B4) We use the above wavefunctions for all helium states √ with the exceptions of the 21S, 23S and 23P states where where the 1/(2 π) prefactor is the Y00 spherical har- we use non-relativistic wavefunctions based on Hyller- monic from the 1s electron, and Nnl ensures that the radial part of the wavefunction is canonically normal- aas functions taken from Refs. [98, 99] in order to better describe the repulsion between the two electrons. This ized. We write the Fnl’s as products of non-relativistic 1 hydrogen radial wavefunctions as turns out to be of particular importance for the 2 S state. Indeed, for spin-singlet states, the spatial part of Fnl(r1, r2) = R10(r1,Zi)Rnl(r2,Za) , (B5) the wavefunction is symmetric under the exchange of the two-electrons so that the wavefunction in Eq. (B4) may where (in units of a0 = 1) over-estimate the electronic density is in the region where the electrons are close to each other, r ∼ r . Hylleraas s 1 2 2Z 3/2 (n − l − 1)! functions then provide a more accurate description of the R (r, Z)= e−ρ/2ρlL2l+1 (ρ) , nl n 2n(n + l)! n−l−1 electron repulsion effect by introducing an explicit depen- dence on the inter-electronic distance r in the wavefunc- (B6) 12 tion. The spatial part of wavefunction is then taken to α be of the form where ρ = 2rZ/n, Lk (x) are the generalized Laguerre 1 polynomials of degree k, and Zi and Za are the ef- S h~r1, ~r2|ψ i = √ [Fnl(r1, r2, r12)Ylm(Ω2)+ fective nuclear charges for the core (1s) and valence nlm N (nl) electrons, respectively. An important point is that S  +(−1) Fnl(r2, r1, r12)Ylm(Ω1) (B9) Zi,a 6= Z = 2 because of screening effects; they depend on the electronic configuration considered, see TableV. where N is a normalization constant and Fnl now de- The Rnl’s form an orthonormal basis for a fixed Z. How- pends on r12 and is expanded on Hylleraas functions as ever, because the two electrons effectively feel a different l − κ (s−σt) Fnl(r1, r2, r12) = [κ(s + t)] e 2 nuclear charge (Zi 6= Za), there is an overall normaliza- k tion constant for S waves because of the cross term in X × c φ (κs, −κt, κu) , (B10) the square of Eq. (B4) i i i=1 2 Z  with s ≡ r1 + r2, t ≡ r2 − r1, u ≡ r12 and φi(s, t, u) = S 2 Nnl = (−1) δl,0 drr R10(r, Zi)Rnl(r, Za) , (B7) spi tqi uri . It is convenient to reorganize the Hylleraas terms according to their powers of r12. We then write which vanishes for Zi = Za and n ≥ 2 by orthogonality the radial function in Eq. (B10) as of Rnl(r, Z). k The total wavefunction for a fixed J and (its projec- X i Fnl(r1, r2, r12) = fi(r1, r2)r12 . (B11) tion) mJ = −J...J are then constructed from L × S i=0 11

Appendix C: Overlap integrals for helium ˆ ˆ In this appendix we give the analytical expressions for Xi ≡Xa − Xb " n #" n # the overlap integrals for the case of helium, i.e. the elec- Z e e −mφri 1 Y 3 X e tronic NP coefficients Xi and Yi. = d ri 4π ri k=1 i=1  2 2 × |Ψa(~r1, . . . , ~rne )| − |Ψb(~r1, . . . , ~rne )| , (C1) 1. Electron-nucleus interactions 2 where ne is the number of bound electrons and |Ψ| is the Let us consider the potential of Eq. (1) between the electron wavefunction density. Using hydrogenic wave- nucleus and its bound electron with the above helium functions in Eq. (B4) the contributions from each state wavefunctions. In first-order perturbation theory we find is

1 Z −mφr  2 2 ˆ 2S+1 Xn LJ,m = − drre R10(r, Zi) + RnL(r, Za) J 4π(1 + NnL) Z Z  S −mφr 2 + 2(−1) δL,0 drre R10(r, Zi)RnL(r, Za) × drr R10(r, Zi)RnL(r, Za) . (C2)

For the case of Hylleraas wavefunctions in Eq. (B9), we with F (α, β; γ; x) denoting the Gauss hypergeometric need the following expansion of r12 raised to the power function and (ξ)s ≡ Γ(ξ + s)/Γ(ξ). We then find k on spherical harmonics

∞ l k X (k) X ∗ r12 = 4π Hl (r1, r2) Ylm(Ω1)Ylm(Ω2), (C3) Z  −mφr1 −mφr2  l=0 m=−l 1 2 2 e e ˆ 2S+1 Xn LJ,m = − r1r2dr1dr2 + J 4πN r1 r2 where the coeffcients can be written in closed form in Z S 2 terms of hypergeometric functions [100] × dΩ1dΩ2|ψnlm| , (C5)

n  l (n) (−n/2)l r> r< Hl (r1, r2) = (1/2)l 2l + 1 r>  2  r< where the square of the spatial wavefunction integrated × F l − n/2, −(n + 1)/2; l + 3/2; 2 , (C4) r> over the angular variables is

Z S 2 2 X (i+j) dΩ1dΩ2|ψnlm| =f0(r1, r2) + H0 (r1, r2)fi(r1, r2)fj(r1, r2) + [r1 ↔ r2] i+j≥1   S X (i+j) + 2(−1) f0(r1, r2)f0(r2, r1)δl0 + Hl (r1, r2)fi(r1, r2)fj(r2, r1) . (C6) i+j≥1

2. Electron-electron interactions ample [101],

∞ l Consider the NP potential between the bounded elec- −mr12 e X X ∗ trons, Ve(r12) see Eq. (1). It is useful to expand the =4π Gl(r1, r2, m) Ylm(Ω1)Ylm(Ω2) , (C7) r12 Yukawa potential over spherical harmonics as, see for ex- l=0 m=−l 12 where the coefficients are Gl(r1, r2, m). The first-order perturbation theory result is Il+1/2(mr<) Kl+1/2(mr>) Gl(r1, r2, m) = √ √ , (C8) r< r> Yi ≡Yˆa − Yˆb " n #  n  Z e e −mφrij with I and K the modified Bessel functions of the first 1 Y 3 X e and second kind respectively and r (r ) is the greater = d ri   > < 4π rij k=1 i>j (lesser) of r1 and r2. For Hylleraas wavefunctions which involve additional powers of r it will be convenient  2 2 12 × |Ψa(~r1, . . . , ~rne )| − |Ψb(~r1, . . . , ~rne )| . (C9) to use Eq. (C7) as a “generating functional” in order k−1 −mr12 to derive the expansion of any r12 e functions Using the expansion of Eq. (C7) and the hydrogenic (for k ≥ 1) by differentiating k−times the coefficients wavefunctions from Eq. (B4) we find

Z 1 2 2 2 2 ˆ 2S+1 Yn LJ,m = − dr1dr2r1r2 G0(r1, r2)R10(r1,Zi) RnL(r2,Za) J 4π(1 + NnL) S  + (−1) GL(r1, r2)R10(r1,Zi)R10(r2,Zi)RnL(r1,Za)RnL(r2,Za) . (C10)

where only G0,L coefficients in the Yukawa expan- For the case of Hylleraas wavefunctions in Eq. (B9), sion of Eq. (C7) are needed. Note that the inte- we find grand above no longer depends on m0 and m hence S Z the sum over Clebsch-Gordan coeffecients squared gives ˆ 1 2 2 Yn2S+1L = − r1r2dr1dr2 P J,mJ 2 J,mJ 0 [C 0 ] = 1 by orthonormality. Finally, 4πN m ,mS L,m ,S,mS Z −mr12 note that the shift in Eq. (C10) is indepedent of J and e S 2 × dΩ1dΩ2 |ψnlm| , (C11) mJ , which is expected since the potential in Eq. (1) is r12 invariant under rotations. We also used the fact the Gl(r1, r2) = Gl(r2, r1) to simplify the expression. where the angular integral simplifies to

Z −mr12 e X (i+j) dΩ dΩ |ψS |2 = G (r , r )[f (r , r )f (r , r ) + f (r , r )f (r , r )] 1 2 r nlm 0 1 2 i 1 2 j 1 2 i 2 1 j 2 1 12 ij S X (i+j) + 2(−1) Gl (r1, r2)fi(r1, r2)fj(r2, r1) , (C12) ij

where (k) indicates the kth differentiation with respect to wavefunctions on the electronic NP constants by cal- m, culating the ratios to the respective quantity based on hydrogen-like wavefunctions, XHylleraas/XH−like, k i i (k) k ∂ Gl(r1, r2, m) Hylleraas H−like G (r , r , m) ≡ (−1) . (C13) Yi /Yi . l 1 2 ∂mk In Fig.7 we evaluate the impact of the Hylleraas

[1] S. G. Porsev, K. Beloy, and A. Derevianko, “Precision [2] K. Tsigutkin, D. Dounas-Frazer, A. Family, J. E. determination of electroweak coupling from atomic Stalnaker, V. V. Yashchuk, and D. Budker, parity violation and implications for particle physics,” “Observation of a Large Atomic Parity Violation Effect Phys. Rev. Lett. 102 (2009) 181601. in Ytterbium,” Phys. Rev. Lett. 103 (2009) 071601. 13

B. Roberts, “Revisiting parity non-conservation in cesium,” Phys. Rev. Lett. 109 (2012) 203003, arXiv:1207.5864 [hep-ph]. 100 [7] B. M. Roberts, V. A. Dzuba, and V. V. Flambaum, 2S-2S 2S-2P “Parity and Time-Reversal Violation in Atomic 50 Systems,” Ann. Rev. Nucl. Part. Sci. 65 (2015) 63–86, arXiv:1412.6644 [physics.atom-ph]. [8] K. P. Jungmann, “Symmetries and fundamental H - like i interactions-selected topics,” Hyperfine Interact. 227 / X (2014) 5–16. 10 [9] Particle Data Collaboration, K. A. Olive et Hylleraas i

X al., “Review of Particle Physics,” Chin. Phys. C38 5 (2014) 090001. [10] R. M. Godun, P. B. R. Nisbet-Jones, J. M. Jones, S. A. King, L. A. M. Johnson, H. S. Margolis, K. Szymaniec, Hylleraas-to-hydrogenoid S. N. Lea, K. Bongs, and P. Gill, “Frequency Ratio of wavefunction ratio Two Optical Clock Transitions in 171Yb+ and 1 Constraints on the Time Variation of Fundamental Constants,” Phys. Rev. Lett. 113 (Nov, 2014) 210801. 10-6 10-3 1 103 [11] N. Huntemann, M. Okhapkin, B. Lipphardt, m [MeV] ϕ S. Weyers, C. Tamm, and E. Peik, “High-Accuracy Optical Clock Based on the Octupole Transition in 171 + 50 Yb ,” Phys. Rev. Lett. 108 (Feb, 2012) 090801. [12] J. Reichert, M. Niering, R. Holzwarth, M. Weitz, T. Udem, and T. W. H¨ansch, “Phase Coherent 2S-2S -Ultraviolet to Radio Frequency Comparison 2S-2P 10 with a Mode-Locked Laser,” Phys. Rev. Lett. 84 (Apr, 2000) 3232–3235. 5 [13] J. Reichert, R. Holzwarth, T. Udem, and T. H¨ansch, H - like i “Measuring the frequency of light with mode-locked

/ Y lasers,” Communications 172 no. 1, (1999) 59 – 1 68. Hylleraas i [14] N. Huntemann, C. Sanner, B. Lipphardt, C. Tamm, Y 0.50 and E. Peik, “Single-Ion Atomic Clock with 3 × 10−18 Systematic Uncertainty,” Phys. Rev. Lett. 116 (Feb, 2016) 063001. [15] B. J. Bloom, T. L. Nicholson, J. R. Williams, S. L. 0.10 Hylleraas-to-hydrogenoid wavefunction ratio Campbell, M. Bishof, X. Zhang, W. Zhang, S. L. 0.05 Bromley, and J. Ye, “An Optical Lattice Clock with Accuracy and Stability at the 10−18 Level,” Nature -6 -3 3 10 10 1 10 506 (2014) 71–75, arXiv:1309.1137 m ϕ [MeV] [physics.atom-ph]. [16] C. Delaunay, R. Ozeri, G. Perez, and Y. Soreq, “Probing The Atomic Higgs Force,” FIG. 7: Comparison of the electronic NP coefficients Xi,Yi based on hydrogen-like and Hylleraas wavefunctions depend- arXiv:1601.05087 [hep-ph]. [17] C. Frugiuele, E. Fuchs, G. Perez, and M. Schlaffer, ing on the mediator mass mφ for the transitions 2S − 2S and 2P − 2S in helium. “Constraining New Physics Models with Isotope Shift Spectroscopy,” Phys. Rev. D96 no. 1, (2017) 015011, arXiv:1602.04822 [hep-ph]. [18] C. Delaunay and Y. Soreq, “Probing New Physics with Isotope Shift Spectroscopy,” arXiv:1602.04838 [3] K. Tsigutkin, D. Dounas-Frazer, A. Family, J. E. [hep-ph]. Stalnaker, V. V. Yashchuk, and D. Budker, “Parity [19] J. C. Berengut et al., “Probing new light violation in atomic ytterbium: Experimental force-mediators by isotope shift spectroscopy,” sensitivity and systematics,” Phys. Rev. A81 (Mar, arXiv:1704.05068 [hep-ph]. 2010) 032114. [20] W. H. King, “Comments on the Article “Peculiarities [4] N. Leefer, L. Bougas, D. Antypas, and D. Budker, of the Isotope Shift in the Samarium Spectrum”,” J. “Towards a new measurement of parity violation in Opt. Soc. Am. 53 no. 5, (May, 1963) 638–639. dysprosium,” 2014. arXiv:1412.1245 [21] K. Pachucki, V. Patk´oˇs,and V. Yerokhin, “Testing [physics.atom-ph]. fundamental interactions on the helium atom,” Phys. [5] O. O. Versolato et al., “Atomic parity violation in a Rev. A95 no. 6, (2017) 062510, arXiv:1704.06902 single trapped radium ion,” Hyperfine Interact. 199 [physics.atom-ph]. no. 1-3, (2011) 9–19. [22] S. G. Karshenboim, “Precise physics of simple atoms,” [6] V. A. Dzuba, J. C. Berengut, V. V. Flambaum, and AIP Conf. Proc. 551 (2001) 238–253, 14

arXiv:hep-ph/0007278 [hep-ph]. [40] R. Barbieri and T. E. O. Ericson, “Evidence Against [23] S. G. Karshenboim, “Constraints on a long-range the Existence of a Low Mass from spin-independent interaction from precision atomic Neutron-Nucleus Scattering,” Phys. Lett. 57B (1975) physics,” Phys. Rev. D82 (2010) 073003, 270–272. arXiv:1005.4872 [hep-ph]. [41] H. Leeb and J. Schmiedmayer, “Constraint on [24] S. G. Karshenboim, “Precision physics of simple atoms hypothetical light interacting from low-energy and constraints on a light boson with ultraweak neutron experiments,” Phys. Rev. Lett. 68 (1992) coupling,” Phys. Rev. Lett. 104 (2010) 220406, 1472–1475. arXiv:1005.4859 [hep-ph]. [42] V. V. Nesvizhevsky, G. Pignol, and K. V. Protasov, [25] J. Jaeckel and S. Roy, “Spectroscopy as a test of “Neutron scattering and extra short range Coulomb’s law: A Probe of the ,” Phys. interactions,” Phys. Rev. D77 (2008) 034020, Rev. D82 (2010) 125020, arXiv:1008.3536 [hep-ph]. arXiv:0711.2298 [hep-ph]. [26] F. Ficek, D. F. J. Kimball, M. Kozlov, N. Leefer, [43] Yu. N. Pokotilovski, “Constraints on new interactions S. Pustelny, and D. Budker, “Constraints on exotic from neutron scattering experiments,” Phys. Atom. spin-dependent interactions between electrons from Nucl. 69 (2006) 924–931, arXiv:hep-ph/0601157 helium fine-structure spectroscopy,” Phys. Rev. A95 [hep-ph]. no. 3, (2017) 032505, arXiv:1608.05779 [44] M. Bordag, U. Mohideen, and V. M. Mostepanenko, [physics.atom-ph]. “New developments in the Casimir effect,” Phys. Rept. [27] C. Eckart, “The Theory and Calculation of Screening 353 (2001) 1–205, arXiv:quant-ph/0106045 Constants,” Phys. Rev. 36 (Sep, 1930) 878–892. [quant-ph]. [28] E. A. Hylleraas, “Uber¨ den Grundzustand des [45] M. Bordag, G. L. Klimchitskaya, U. Mohideen, and Heliumatoms,” Zeitschrift f¨urPhysik 48 no. 7, (Jul, V. M. Mostepanenko, “Advances in the Casimir 1928) 469–494. effect,” Int. Ser. Monogr. Phys. 145 (2009) 1–768. [29] E. A. Hylleraas, “Neue Berechnung der Energie des [46] G. Raffelt, “Limits on a CP-violating scalar Heliums im Grundzustande, sowie des tiefsten Terms - interaction,” Phys. Rev. D86 (2012) von Ortho-Helium,” Zeitschrift f¨urPhysik 54 no. 5, 015001, arXiv:1205.1776 [hep-ph]. (May, 1929) 347–366. [47] Collaboration, W. M. Yao et [30] K. Pachucki and V. A. Yerokhin, “Theory of the al., “Review of Particle Physics,” J. Phys. G33 (2006) Helium Isotope Shift,” Journal of Physical and 1–1232. Chemical Reference Data 44 no. 3, (2015) 031206, [48] J. A. Grifols and E. Masso, “Constraints on Finite arXiv:1503.07727 [physics.atom-ph]. Range Baryonic and Leptonic Forces From Stellar [31] Y.-S. Liu, D. McKeen, and G. A. Miller, Evolution,” Phys. Lett. B173 (1986) 237–240. “Electrophobic Scalar Boson and Muonic Puzzles,” [49] J. A. Grifols, E. Masso, and S. Peris, “Energy Loss Phys. Rev. Lett. 117 no. 10, (2016) 101801, From the Sun and RED Giants: Bounds on Short arXiv:1605.04612 [hep-ph]. Range Baryonic and Leptonic Forces,” Mod. Phys. [32] D. Tucker-Smith and I. Yavin, “Muonic hydrogen and Lett. A4 (1989) 311. MeV forces,” Phys. Rev. D83 (2011) 101702, [50] E. Hardy and R. Lasenby, “Stellar cooling bounds on arXiv:1011.4922 [hep-ph]. new light particles: plasma mixing effects,” JHEP 02 [33] I. Beltrami et al., “New Precision Measurements of the (2017) 033, arXiv:1611.05852 [hep-ph]. Muonic 3-D (5/2) - 2p (3/2) X-ray Transition in 24Mg [51] J. Redondo and G. Raffelt, “Solar constraints on and 28Si: Test and Search for hidden re-visited,” JCAP 1308 (2013) 034, Muon - Interactions Beyond QED,” Nucl. arXiv:1305.2920 [hep-ph]. Phys. A451 (1986) 679–700. [52] E. J. Salumbides, W. Ubachs, and V. I. Korobov, [34] R. Pohl et al., “The size of the proton,” Nature 466 “Bounds on fifth forces at the sub-Angstrom length (2010) 213–216. scale,” J. Molec. Spectrosc. 300 (2014) 65, [35] A. Antognini et al., “Proton Structure from the arXiv:1308.1711 [hep-ph]. Measurement of 2S − 2P Transition Frequencies of [53] W. Ubachs, J. C. J. Koelemeij, K. S. E. Eikema, and Muonic Hydrogen,” Science 339 (2013) 417–420. E. J. Salumbides, “Physics beyond the Standard [36] CREMA Collaboration, R. Pohl et al., “Laser Model from hydrogen spectroscopy,” spectroscopy of muonic deuterium,” Science 353 arXiv:1511.00985 [physics.atom-ph]. no. 6300, (2016) 669–673. [54] J. Biesheuvel, J.-P. Karr, L. Hilico, K. S. E. Eikema, [37] Particle Data Group Collaboration, C. Patrignani W. Ubachs, and J. C. J. Koelemeij, “High-precision et al., “Review of Particle Physics,” Chin. Phys. C40 spectroscopy of the HD+ at the 1-p.p.b. no. 10, (2016) 100001. level,” B 123 no. 1, (Dec, 2016) 23. [38] D. Hanneke, S. F. Hoogerheide, and G. Gabrielse, [55] “Physics beyond the Standard Model from hydrogen “Cavity Control of a Single-Electron Quantum spectroscopy,” Journal of Molecular Spectroscopy 320 Cyclotron: Measuring the Electron Magnetic (2016) 1 – 12. Moment,” Phys. Rev. A83 (2011) 052122, [56] B. Feldman and A. E. Nelson, “New regions for a arXiv:1009.4831 [physics.atom-ph]. chameleon to hide,” JHEP 08 (2006) 002, [39] S. L. Adler, R. F. Dashen, and S. B. Treiman, arXiv:hep-ph/0603057 [hep-ph]. “Comments on Proposed Explanations for the mu - [57] A. E. Nelson and J. Walsh, “Chameleon vector Mesic Atom x-Ray Discrepancy,” Phys. Rev. D10 bosons,” Phys. Rev. D77 (2008) 095006, (1974) 3728. arXiv:0802.0762 [hep-ph]. [58] C. Burrage, “Supernova Brightening from 15

Chameleon-Photon Mixing,” Phys. Rev. D77 (2008) T. Udem, U. D. Jentschura, N. Kolachevsky, and 043009, arXiv:0711.2966 [astro-ph]. T. W. Hansch, “Precision Measurement of the [59] C. Burrage and J. Sakstein, “A Compendium of Hydrogen-Deuterium 1S-2S Isotope Shift,” Phys. Rev. Chameleon Constraints,” JCAP 1611 no. 11, (2016) Lett. 104 (2010) 233001. 045, arXiv:1609.01192 [astro-ph.CO]. [76] U. Jentschura, A. Matveev, C. Parthey, J. Alnis, [60] P. Brax and C. Burrage, “Atomic Precision Tests and R. Pohl, T. Udem, N. Kolachevsky, and T. Hansch, Light Scalar Couplings,” Phys. Rev. D83 (2011) “Hydrogen-deuterium isotope shift: From the 035020, arXiv:1010.5108 [hep-ph]. 1S − 2S-transition frequency to the proton-deuteron [61] J. Alexander et al., “Dark Sectors 2016 Workshop: charge-radius difference,” Phys. Rev. A 83 (Apr, 2011) Community Report,” 2016. arXiv:1608.08632 042505. [hep-ph]. [77] B. de Beauvoir, C. Schwob, O. Acef, L. Jozefowski, [62] F. Gebert, Y. Wan, F. Wolf, C. N. Angstmann, J. C. L. Hilico, F. Nez, L. Julien, A. Clairon, and Berengut, and P. O. Schmidt, “Precision Isotope Shift F. Biraben, “Metrology of the hydrogen and deuterium Measurements in Calcium Ions Using atoms: Determination of the Rydberg constant and Detection Schemes,” Phys. Rev. Lett. 115 (2015) Lamb shifts,” EPJD 12 no. 1, (Sep, 2000) 61–93. 053003. [78] B. de Beauvoir, F. Nez, L. Julien, B. Cagnac, [63] R. van Rooij, J. S. Borbely, J. Simonet, M. D. F. Biraben, D. Touahri, L. Hilico, O. Acef, A. Clairon, Hoogerland, K. S. E. Eikema, R. A. Rozendaal, and and J. J. Zondy, “Absolute Frequency Measurement of W. Vassen, “Frequency Metrology in Quantum the 2S − 8S/D Transitions in Hydrogen and Degenerate Helium: Direct Measurement of the 2 3S1 Deuterium: New Determination of the Rydberg → 2 1S0 Transition,” Science 333 no. 6039, (2011) Constant,” Phys. Rev. Lett. 78 (Jan, 1997) 440–443. 196–198. [79] C. Schwob, L. Jozefowski, B. de Beauvoir, L. Hilico, [64] P. Cancio Pastor, L. Consolino, G. Giusfredi, F. Nez, L. Julien, F. Biraben, O. Acef, J.-J. Zondy, P. De Natale, M. Inguscio, V. A. Yerokhin, and and A. Clairon, “Optical Frequency Measurement of K. Pachucki, “Frequency Metrology of Helium around the 2S − 12D Transitions in Hydrogen and Deuterium: 1083 nm and Determination of the Nuclear Charge Rydberg Constant and Lamb Shift Determinations,” Radius,” Phys. Rev. Lett. 108 (Apr, 2012) 143001. Phys. Rev. Lett. 82 (Jun, 1999) 4960–4963. [65] P. C. Pastor, G. Giusfredi, P. D. Natale, G. Hagel, [80] M. Weitz, A. Huber, F. Schmidt-Kaler, D. Leibfried, C. de Mauro, and M. Inguscio, “Absolute Frequency W. Vassen, C. Zimmermann, K. Pachucki, T. H¨ansch, 3 3 Measurements of the 2 S1 → 2 P0,1,2 Atomic Helium L. Julien, and F. Biraben, “Precision measurement of Transitions around 1083 nm,” Phys. Rev. Lett. 92 the 1S ground-state Lamb shift in atomic hydrogen (Jan, 2004) 023001. and deuterium by frequency comparison,” Physical [66] I. Sick, “Form factors and radii of light nuclei,” Review A 52 no. 4, (1995) 2664. arXiv:1505.06924 [nucl-ex]. [81] A1 Collaboration, J. C. Bernauer, M. O. Distler, [67] A. Antognini et al., “Illuminating the proton radius J. Friedrich, T. Walcher, P. Achenbach, conundrum: The mu He+ Lamb shift,” Can. J. Phys. C. Ayerbe Gayoso, R. B¨ohm,D. Bosnar, L. Debenjak, 89 no. 1, (2011) 47–57. L. Doria, A. Esser, H. Fonvieille, M. G´omezRodr´ıguez [68] W. Vassen, “Ultracold He spectroscopy: QED test, line de la Paz, J. M. Friedrich, M. Makek, H. Merkel, D. G. shapes and nuclear sizes.” Talk at FKK 2017, Warsaw. Middleton, U. M¨uller,L. Nungesser, J. Pochodzalla, [69] P. Mueller et al., “Nuclear charge radius of He-8,” M. Potokar, S. S´anchez Majos, B. S. Schlimme, Phys. Rev. Lett. 99 (2007) 252501, arXiv:0801.0601 S. Sirca,ˇ and M. Weinriefer, “Electric and magnetic [nucl-ex]. form factors of the proton,” Phys. Rev. C 90 (Jul, [70] E. Riis, A. G. Sinclair, O. Poulsen, G. W. F. Drake, 2014) 015206. W. R. C. Rowley, and A. P. Levick, “Lamb shifts and [82] P. J. Mohr, B. N. Taylor, and D. B. Newell, hyperfine structure in 6Li+ and 7Li+: Theory and “CODATA Recommended Values of the Fundamental experiment,” Phys. Rev. A 49 (Jan, 1994) 207–220. Physical Constants: 2010,” Rev. Mod. Phys. 84 (2012) [71] J. K. Thompson, D. J. H. Howie, and E. G. Myers, 1527–1605, arXiv:1203.5425 [physics.atom-ph]. 1 3 “Measurements of the 1s2s S0˘1s2p P1,0 transitions in [83] M. S. Fee, A. P. Mills, S. Chu, E. D. Shaw, heliumlike nitrogen,” Phys. Rev. A 57 (Jan, 1998) K. Danzmann, R. J. Chichester, and D. M. 180–188. Zuckerman, “Measurement of the positronium S-131- [72] C. De Jager, H. De Vries, and C. De Vries, “Nuclear S-231 interval by continuous-wave two-photon charge-and magnetization-density-distribution excitation,” Phys. Rev. Lett. 70 (1993) 1397–1400. parameters from elastic electron scattering,” Atomic [84] A. Czarnecki, K. Melnikov, and A. Yelkhovsky, data and nuclear data tables 14 no. 5-6, (1974) “Positronium S state spectrum: Analytic results at 479–508. O(m alpha**6),” Phys. Rev. A59 (1999) 4316, [73] J. de Vries, D. Doornhof, C. de Jager, R. Singhal, arXiv:hep-ph/9901394 [hep-ph]. S. Salem, G. Peterson, and R. Hicks, “The 15N ground [85] B. Holdom, “Two U(1)’s and Epsilon Charge Shifts,” state studied with elastic electron scattering,” Physics Phys. Lett. 166B (1986) 196–198. Letters B 205 no. 1, (1988) 22 – 25. [86] M. Pospelov and Y.-D. Tsai, “Probing Light Bosons in [74] L. Schaller, L. Schellenberg, A. Ruetschi, and the Borexino-SOX Experiment,” arXiv:1706.00424 H. Schneuwly, “Nuclear charge radii from muonic [hep-ph]. X-ray transitions in beryllium, boron, carbon and [87] M. Pospelov, “Secluded U(1) below the weak scale,” nitrogen,” A 343 (1980) 333 – 346. Phys. Rev. D80 (2009) 095002, arXiv:0811.1030 [75] C. G. Parthey, A. Matveev, J. Alnis, R. Pohl, [hep-ph]. 16

[88] P. J. Mohr, D. B. Newell, and B. N. Taylor, “CODATA Drake, M. Dubois, C. El´eon,G. Gaubert, R. J. Holt, Recommended Values of the Fundamental Physical R. V. F. Janssens, N. Lecesne, Z.-T. Lu, T. P. Constants: 2014,” Rev. Mod. Phys. 88 no. 3, (2016) O’Connor, M.-G. Saint-Laurent, J.-C. Thomas, and 035009, arXiv:1507.07956 [physics.atom-ph]. L.-B. Wang, “Nuclear Charge Radius of 8He,” Phys. [89] I. Angeli and K. Marinova, “Table of experimental Rev. Lett. 99 (Dec, 2007) 252501. nuclear ground state charge radii: An update,” Atomic [97] U. D. Jentschura, S. Kotochigova, E.-O. Le Bigot, P. J. Data and Nuclear Data Tables 99 no. 1, (2013) 69 – 95. Mohr, and B. N. Taylor, “Precise Calculation of [90] G. W. Drake, Atomic, Molecular and Optical Physics. Transition Frequencies of Hydrogen and Deuterium Springer, 2006. Based on a Least-Squares Analysis,” Phys. Rev. Lett. [91] H. A. Schuessler, E. N. Fortson, and H. G. Dehmelt, 95 (Oct, 2005) 163003. “Hyperfine Structure of the Ground State of 3He+ by [98] S.-S. Huang, “The Continuous Absorption Coefficient the Ion-Storage Exchange-Collision Technique,” Phys. of the Helium Atom.,” Astrophys. J. 108 (Nov., 1948) Rev. 187 (Nov, 1969) 5–38. 354. [92] S. D. Rosner and F. M. Pipkin, “Hyperfine Structure [99] J. Traub and H. M. Foley, “Variational Calculations of 3 3 3 of the 2 S1 State of He ,” Phys. Rev. A 1 (Mar, 1970) Energy and for the 2 P State of 571–586. Helium,” Phys. Rev. 116 (Nov, 1959) 914–919. [93] V. c. v. Patk´oˇs,V. A. Yerokhin, and K. Pachucki, [100] R. Sack, “Generalization of Laplace’s expansion to “Higher-order recoil corrections for triplet states of the arbitrary powers and functions of the distance between helium atom,” Phys. Rev. A 94 (Nov, 2016) 052508. two points,” Journal of 5 no. 2, [94] V. Patk´oˇs,V. A. Yerokhin, and K. Pachucki, (1964) 245–251. “Higher-order recoil corrections for singlet states of the [101] P. Serra and S. Kais, “Ground-state stability and helium atom,” Phys. Rev. A95 no. 1, (2017) 012508, criticality of two-electron atoms with screened arXiv:1612.06142 [physics.atom-ph]. Coulomb potentials using the B-splines basis set,” [95] V. c. v. Patk´oˇs,V. A. Yerokhin, and K. Pachucki, Journal of Physics B: Atomic, Molecular and Optical “Higher-order recoil corrections for singlet states of the Physics 45 no. 23, (2012) 235003. http: helium atom,” Phys. Rev. A 95 (Jan, 2017) 012508. //stacks.iop.org/0953-4075/45/i=23/a=235003. [96] P. Mueller, I. A. Sulai, A. C. C. Villari, J. A. Alc´antara-N´u˜nez,R. Alves-Cond´e,K. Bailey, G. W. F.