PROTOPHOBIC 8Be TRANSITION EVIDENCE FOR A NEW 17 MEV BOSON Flip Tanedo arXiv:1604.07411 & work in progress SLAC Dark Sectors 2016 (28 — 30 April) with Jonathan Feng, Bart Fornal, Susan Gardner, Iftah Galon, Jordan Smolinsky, & Tim Tait
BART SUSAN IFTAH 1 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 week ending A 6.8σ nuclear transitionPRL 116, 042501 (2016) anomalyPHYSICAL REVIEW LETTERS 29 JANUARY 2016
shape of the resonance [40], but it is definitelyweek different ending pairs (rescaled for better separation) compared with the PRL 116, 042501 (2016) PHYSICAL REVIEW LETTERS 29 JANUARY 2016 from the shape of the forward or backward asymmetry [40]. simulations (full curve) including only M1 and E1 con- Therefore, the above experimental data make the interpre- tributions. The experimental data do not deviate from the Observation of Anomalous Internal Pairtation Creation of the in observed8Be: A Possible anomaly Indication less probable of a as Light, being the normal IPC. This fact supports also the assumption of the Neutralconsequence Boson of some kind of interference effects. boson decay. The deviation cannot be explained by any γ-ray related The χ2 analysis mentioned above to judge the signifi- A. J. Krasznahorkay,* M. Csatlós, L. Csige, Z. Gácsi,background J. Gulyás, M. either, Hunyadi, since I. Kuti, we B. cannot M. Nyakó, see any L. Stuhl, effect J. Timár, at off cance of the observed anomaly was extended to extract the T. G. Tornyi,resonance, and Zs. where Vajta the γ-ray background is almost the same. mass of the hypothetical boson. The simulated angular Institute for Nuclear Research, Hungarian Academy ofTo Sciences the best (MTA of Atomki), our knowledge, P.O. Box 51, the H-4001 observed Debrecen, anomaly Hungary can correlations included contributions from bosons with 2 notT. J. have Ketel a nuclear physics related origin. masses between m0c 15 and 17.5 MeV. As a result Nikhef National Institute for Subatomic Physics,The Science deviation Park 105, observed 1098 XG Amsterdam, at the bombarding Netherlands energy of of the χ2 analysis, we¼ determined the boson mass to be 2 Ep 1.10 MeV and at Θ ≈ 140° has a significance of 6.8 m0c 16.70 0.35 stat MeV. The minimum value for A. Krasznahorkay¼ 2¼ Æ ð Þ CERN, CH-1211 Geneva 23, Switzerland and Institute forstandard Nuclear Research,deviations, Hungarian corresponding Academy of to Sciences a background (MTA Atomki), fluc- the χ =f was 1.07, while the values at 15 and 17.5 MeV 12 P.O. Box 51, H-4001tuation Debrecen, probability Hungary of 5.6 × 10− . On resonance, the M1 were 7.5 and 6.0, respectively. A systematic error caused by (Received 7 April 2015;contribution published 26 should January be 2016) even larger, so the background the instability of the beam position on the target, as well as Electron-positron angular correlations wereshould measured decrease for the isovector faster than magnetic in other dipole cases, 17.6 MeV which would the uncertainties in the calibration and positioning of the π π (J 1þ, T 1) state → ground state (J make0þ, T the0) deviation and the isoscalar even magnetic larger dipoleand more 18.15 significant. MeV detectors is estimated to be 6°, which corresponds to ¼ ¼ ¼ ¼ ΔΘ (Jπ 1 , T 0) state → ground state transitions in 8Be. Significant− enhancement relative to the internal ¼ ¼ þ ¼ The eþe decay of a hypothetical boson emitted iso- 0.5 MeV uncertainty in the boson mass. pair creation was observed at large angles intropically the angular from correlation the target for the has isoscalar been transition simulated with together a with Since, in contrast to the case of 17.6 MeV isovector confidence level of > 5σ. This observation could possibly be due to nuclear reaction interference− effects or might indicate that, in an intermediatethe step, normal a neutral IPC isoscalaremission particle of eþe withpairs. a mass The of sensitivity of transition, the observed anomalous enhancement of the 2 theπ angular correlation measurements to the mass of the 18.15 MeV isoscalar transition could only be explained by 16.70 0.35 stat 0.5 syst MeV=c and J 1þ was created. Æ ð ÞÆ ð Þ assumed¼ boson is illustrated in Fig. 4. also assessing a particle, then it must be of isoscalar nature. DOI: 10.1103/PhysRevLett.116.042501 Taking into account an IPC coefficient of 3.9 × 10−3 for The invariant mass distribution calculated from the the 18.15 MeV M1 transition [32], a boson to γ branching measured energies and angles was also derived. It is shown DOI: 10.1103/PhysRevLett.116.042501 −6 ratio of 5.8 × 10 was found for− the best fit and was then2 in Fig. 5. Recently, several experimental anomalies were discussed particle decaying into an eþe pair, the angular correlation flipas . possible tanedo signatures@ uci for . a edu new light particle [1–3]. SomeusedEVIDENCE forbecomes the other FOR sharply boson A 17 peaked massesMEV at NEW larger in Fig. BOSON angles,4. the correlation14 The dashed line shows the result of the simulation predictions suggest light neutral bosons in the 10 MeVAccording– angle of to a two-particle the simulations, decay is the 180° contribution in the center-of-mass of the performed for M1 23%E1 mixed IPC transition (the 10 GeV mass range as dark matter candidates, whichassumedsystem. boson should be negligible for asymmetric pairs mixing ratio was determinedþ from fitting the experimental 8 couple to electrons and positrons [4–7], to explainwith the 0.5 ≤Toy populate≤ 1.0. the The 17.6, open and circles 18.15 MeV with1þ errorstates bars in Be in angular correlations), the dotted line shows the simulation anomalies. A number of attempts were made to find such selectively,j j we used the 7Li p; γ 8Be reaction at the Fig. 4 show the experimental data obtainedð Þ for asymmetric for the decay of a particle with mass of 16.6 MeV=c2 while particles [1,8–17]. Since no evidence was found, limits Ep 0.441, and 1.03 MeV resonances [29]. Proton beams were set on their mass and their coupling strength to from¼ a 5 MV Van de Graaff accelerator with typical current the dash-dotted line is their sum, which describes the 2 ordinary matter. In the near future, ongoing experiments are of -11.0 μA impinged on 15 μg=cm thick LiF2 and experimental data reasonably well. 10 2 − expected to extend those limits to regions in mass and 700 μg=cm thick LiO2 targets evaporated on 10 μmAl In conclusion, we have measured the eþe angular coupling strength which are so far unexplored. All of them backings. correlation in internal pair creation for the M1 transition − are designed to exploit the radiative production of the so- The eþe pairs were detected by five plastic ΔE E depopulating the 18.15 MeV state in 8Be, and observed a detector telescopes similar to those built by Stiebing− and called dark photons (γ0) by a very intense electron or peaklike deviation from the predicted IPC. To the best of positron beam on a high-Z target [18–23]. co-workers [34], but we used larger telescope detectors in In the present work, we reinvestigated the anomalies combination with position sensitive detectors to signifi- =17.6 MeV =15.6 MeV 2 observed previously in the internal pair creation of iso- cantly increase the coincidence efficiency by2 about 3 orders c c =16.6 MeV 0 0 2 3
vector (17.6 MeV) and isoscalar (18.15 MeV) M1 tran- c m of magnitude. E detectors of 38 × 45 × 1 mmm and the E Δ 0 800 8 3 sitions in Be [24–29]. The expected signature of the detectors of 78 × 60 × 70 mm werem placed perpendicu- anticipated particle is a very characteristic angular corre- larly to the beam direction at azimuthal angles of 0°, 60°, 700 − lation of the eþe pairs from its decay [30,31]. The angular 120°, 180°, and 270°. These angles ensured homogeneous correlation between the e and e− emitted in the internal IPCC (relative unit) 600 þ acceptance of the e e− pairs as a function of the correlation pair creation (IPC) drops rapidly with the separation angle þ angle. The positions of the hits were determined by 500 θ [32,33]. In striking contrast, when the transition takes =16.6 MeV −13 multiwire-2 proportional counters (MWPC) [35] placed in 2 c place by emission of a short-lived (τ < 10 s) neutral 10 400 0
front of the ΔE and E detectors. m The target strip foil was perpendicular to the beam 300 IPC, M1+E1
direction. The telescope detectors were placed around the (Weighted Counts/0.5 MeV) Published by the American Physical Society under the terms of 80 90 100 110 120 130 140 150 160 170 vacuum chamber made of a carbon-fiber tube. A detailed e+e- 200 the Creative Commons Attribution 3.0 License. Further distri- Θ (deg.) N bution of this work must maintain attribution to the author(s) and description of the experimental setup is published else- the published article’s title, journal citation, and DOI. where [36]. 100 − FIG. 4. Experimental angular eþe pair correlations measured 7 0 in the Li p; e e− reaction at E 1.10 MeV with 9 10 11 12 13 14 15 16 17 18 0031-9007=16=116(4)=042501(5) 042501-1ð þ PublishedÞ by the Americanp ¼ Physical Society −0.5 ≤ y ≤ 0.5 (closed circles) and y ≥ 0.5 (open circles). me+e- (MeV) The results of simulations of boson decayj j pairs added to those of IPC pairs are shown for different boson masses as described in FIG. 5. Invariant mass distribution derived for the 18.15 MeV the text. transition in 8Be.
042501-4 1138 M. J. SAVAGE, B.W. FILIPPONE, AND L. W. MITCHELL 37
sistent in yield, energy, and angular correlation with decays with a 44% branch to the ( J"=0+, T =0) ' —100 ppm of F contamination in the target and the ground state and a 56% branch to the (J =2+, T =0) 95-keV-wide resonance in ' F(p, a)' 0' at 1.720 MeV first excited state at 3.04 MeV, both of which promptly (Ref. 30). The broad peak appearing on the high-energy decay to two a particles. There are also small branches side of the ' N peak is present with the beam off and is to the isospin mixed levels at 16.6 and 16.9 MeV which due to cosmic-ray interactions. To produce the angular- also a decay. The a particles are stopped in the lexan correlation spectrum shown in Fig. 5(a), an energy re- window at the end of the target chamber, and are not quirement was imposed on the total energy spectrum seen by the detector. The transition to the ground state is such that only events in the upper-half of the full energy pure M1 but has a AT=1 isospin admixture of the order peak were accepted. With this cut, the contribution from of 10%; ' however, the ratio of the E2 to M1 com- ' F pairs and cosmic rays was negligible. The efficiency ponents of the transition to the first excited state has not of the detector to M1 pairs, ' ' passing the energy cuts, been experimentally determined. The isospin-triplet state from the 9.17-MeV state was calculated to be 3.7)& 10 at 17.4 MeV, for which the decay to the first excited state and hence, from the y-ray yield observed during the data has an E2 component two orders of magnitude lower acquisition, the expected number of IPC pairs was than the M1 component, has similar nuclear structure (6.6+0.6) X 10, in good agreement with the (6.0 to the 18.1-MeV state; however, hT =0, M1 transitions +0.6)X10 actually seen. The uncertainty quoted in- are inhibited in self-conjugate nuclei by 1 to 2 orders of cludes the uncertainty in the NaI efficiency and in the magnitude. The inhibition is due to the near cancella- efficiency of the software cuts. tion of the isoscalar and isovector magnetic moments in ' As for the scalar-particle case, we define the upper lim- the sum LM' '=p~+IM", where p' is the isoscalar magnetic it on particle decay to be the branch which increases I, moment of the nucleon and p~ and p" are the proton and of the fit between the data and the Monte Carlo simula- neutron magnetic moments, respectively. It is therefore tion by 2. The X, for a Monte Carlo simulation without not easy to estimate the amount of E2 in this transition particle emission [shown in Fig. 5(a)] is 1.5. Figure 5(b) to the first excited state; however, the uncertainty does shows the correlation spectrum that would be produced if not significantly effect the expected number of pairs be- this state had a 4)&10 branch to a 1.8-MeV particle cause the branching ratio of internal pairs for E2 and M1 that decayed to e+e, which is our quoted limit. transitions is comparable. Transitions involving pseu- doscalar particles are not inhibited a priori, and so are in- 2. Isoscalar transition in Be cluded in the Monte Carlo simulations of the experiment. It is important to note that the uncertainty in the M 1 The third transition studied was the decay of the component of the decay to the broad (I =1.5 MeV) first 18.15-MeV (J =1+, T =0) level in Be. The level excited state has little effect ( & 15%) on our sensitivity to scheme for the relevent states in Be is shown in Fig. 6. particle emission because particle decay to this state The 18.1-MeV state, which is apparently the isospin- would produce a broad feature, rather than a peak, in the singlet partner of the triplet state at 17.4 MeV (Ref. 33}, correlation spectrum. The bulk of our sensitivity comes from the ground-state transition which is pure M1, and so for the purpose of deriving limits on particle emission we assume that the first-excited-state transition is pure 8-Beryllium Levels E (Mev) E2 (i.e., no particle emission). The excitation function of the Li(p, y)'Be reaction in- many other states not shown dicated a large direct capture yield to the first excited 1+;0 18.15 state which reduced the sensitivity of the measurement 1;1 17 64 2+;0+ 1 16.92 since the direct capture proceeds via an E 1 (Ref. 36) tran- Li+p 2+;0+1 16.63 sition for which 0 particles cannot be emitted. For an 17.25 MeV E1 transition the IPC to y-ray ratio is somewhat greater —— than that of an M1: I ~z&/I z 4.3X10 compared M1 M1 with I I&/I ~=3.5X10 for the 18.15-MeV transi- 4+.0 11.4 tion. The ratio of the yield of high-energy y rays on resonance to that off resonance was used to estimate the 44% 56% 67K 33% direct capture contribution to the pairs and was used as input for the Monte Carlo simulation. Even with the un- certainty in the nature of the multipolarity of the radia- tion to the first excited state, the number of pairs expect- ed was 464+50, which is significantly less than the 1001 2+;0 3.04 events recorded. The summed-energy spectrum [Fig. 7(a}]is broad and shows little structure (except for a pos- sible small contribution from ' F contamination) in con- o+-o 0.00 trast with the expected spectrum [Fig. 7(b)] that is Savage et al. Phys. Rev. D37 (1987) 11348Be skewed3 toward the high-energy end. The maximum of flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 FIG. 6. Level scheme of Be showing the branches of interest the Monte Carlo spectrum occurs at 14 MeV and not 18 from the two J =1+ states at —18-MeV excitation. MeV as naively expected for an 18-MeV transition, since Experiment & interpretation
ATOMKI PAIR SPECTROMETER
FT, based on Gulyás et al. (1504.00489) 4 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 A 6.8σ anomaly: two measurements
-1 10
m=15.6 MeV 800 symmetric e m=16.6 MeV 700 m=17.6 MeV dN/d θ 600 + e - 500 [per 0.5 MeV]
ee 400 backgroundm=16.6 MeV
300
asymmetric e 200 -2 ÷1.9 Counts, N 10 + 100 e - signal 0 80 90 100 110 120 130 140 150 160 170 9 10 11 12 13 14 15 16 17 18
Opening Angle [Deg] Invariant Mass, mee [MeV]
Krasznahorkay et al. Phys. Rev. Lett 116 (2016) 042501 5 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 Sanity Checks
1. Bump, not monotonically decreasing background
2. Opening angle and invariant mass agree (17 MeV)
3. Bump disappears off resonance not interference with other decays
4. Bump disappears for asymmetric energies consistent with kinematics for on-shell particle
5. Large energy splitting, wouldn’t see in other nuclei
6 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 1138 M. J. SAVAGE, B.W. FILIPPONE, AND L. W. MITCHELL 37
sistent in yield, energy, and angular correlation with decays with a 44% branch to the ( J"=0+, T =0) ' —100 ppm of F contamination in the target and the ground state and a 56% branch to the (J =2+, T =0) 95-keV-wide resonance in ' F(p, a)' 0' at 1.720 MeV first excited state at 3.04 MeV, both of which promptly (Ref. 30). The broad peak appearing on the high-energy decay to two a particles. There are also small branches side of the ' N peak is present with the beam off and is to the isospin mixed levels at 16.6 and 16.9 MeV which due to cosmic-ray interactions. To produce the angular- also a decay. The a particles are stopped in the lexan correlation spectrum shown in Fig. 5(a), an energy re- window at the end of the target chamber, and are not quirement was imposed on the total energy spectrum seen by the detector. The transition to the ground state is such that only events in the upper-half of the full energy pure M1 but has a AT=1 isospin admixture of the order peak were accepted. With this cut, the contribution from of 10%; ' however, the ratio of the E2 to M1 com- ' F pairs and cosmic rays was negligible. The efficiency ponents of the transition to the first excited state has not of the detector to M1 pairs, ' ' passing the energy cuts, been experimentally determined. The isospin-triplet state from the 9.17-MeV state was calculated to be 3.7)& 10 at 17.4 MeV, for which the decay to the first excited state and hence, from the y-ray yield observed during the data has an E2 component two orders of magnitude lower acquisition, the expected number of IPC pairs was than the M1 component, has similar nuclear structure (6.6+0.6) X 10, in good agreement with the (6.0 to the 18.1-MeV state; however, hT =0, M1 transitions +0.6)X10 actually seen. The uncertainty quoted in- are inhibited in self-conjugate nuclei by 1 to 2 orders of cludes the uncertainty in the NaI efficiency and in the magnitude. The inhibition is due to the near cancella- efficiency of the software cuts. tion of the isoscalar and isovector magnetic moments in ' As for the scalar-particle case, we define the upper lim- the sum LM' '=p~+IM", where p' is the isoscalar magnetic it on particle decay to be the branch which increases I, moment of the nucleon and p~ and p" are the proton and of the fit between the data and the Monte Carlo simula- neutron magnetic moments, respectively. It is therefore tion by 2. The X, for a Monte Carlo simulation without not easy to estimate the amount of E2 in this transition particle emission [shown in Fig. 5(a)] is 1.5. Figure 5(b) to the first excited state; however, the uncertainty does shows the correlation spectrum that would be produced if not significantly effect the expected number of pairs be- this state had a 4)&10 branch to a 1.8-MeV particle cause the branching ratio of internal pairs for E2 and M1 that decayed to e+e, which is our quoted limit. transitions is comparable. Transitions involving pseu- doscalar particles are not inhibited a priori, and so are in- 2. Isoscalar transition in Be cluded in the Monte Carlo simulations of the experiment. It is important to note that the uncertainty in the M 1 The third transition studied was the decay of the component of the decay to the broad (I =1.5 MeV) first 18.15-MeV (J =1+, T =0) level in Be. The level excited state has little effect ( & 15%) on our sensitivity to scheme for the relevent states in Be is shown in Fig. 6. particle emission because particle decay to this state The 18.1-MeV state, which is apparently the isospin- would produce a broad feature, rather than a peak, in the singlet partner of the triplet state at 17.4 MeV (Ref. 33}, correlation spectrum. The bulk of our sensitivity comes from the ground-state transition which is pure M1, and so for the purpose of deriving limits on particle emission Not a dark Higgs we assume that the first-excited-state transition is pure E (Mev) E2 (i.e., no particle emission). The excitation function of the Li(p, y)'Be reaction in- dicated a large direct capture yield to the first excited ANGULAR MOMENTUM 1+;0 18.15 state which reduced the sensitivity of the measurement 1;1 17 64 2+;0+ 1 16.92 since the direct capture proceeds via an E 1 (Ref. 36) tran- ` =1 Li+p 2+;0+1 16.63 sition for which 0 particles cannot be emitted. For an 17.25 MeV E1 transition the IPC to y-ray ratio is somewhat greater than that of an M1: I ~z&/I ——4.3X10 compared PARITY z with I I&/I ~=3.5X10 for the 18.15-MeV transi- ` 4+.0 11.4 P =( ) PBe PX tion. The ratio of the yield of high-energy y rays on + + resonance to that off resonance was used to estimate the - 44% 56% 67K 33% direct capture contribution to the pairs and was used as input for the Monte Carlo simulation. Even with the un- certainty in the nature of the multipolarity of the radia- Decay is forbidden tion to the first excited state, the number of pairs expect- up to parity violation ed was 464+50, which is significantly less than the 1001 2+;0 3.04 events recorded. The summed-energy spectrum [Fig. 7(a}]is broad and shows little structure (except for a pos- sible small contribution from ' F contamination) in con- o+-o 0.00 trast with the expected spectrum [Fig. 7(b)] that is 8Be skewed toward the high-energy end. The maximum of FIG. 6. Level scheme of Be showing the branches of interest the Monte Carlo spectrum occurs at 14 MeV and not 18 from the two J =1+ states at —18-MeV excitation. 7 MeV as naively expected for an 18-MeV transition, since flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 6.2 Physics Motivation 135
fa GeV 1014 1013 1012 1011 1010 109 108 107 106 105 104 103 102 101 1
String decay $ % 2.0 Hot!DM CMB BBN
Misalignment Telescope EBL 1.5 Cold DM Burst Duration ! SK ! SN1987A 1.0 ! gae Globular Clusters gaΓ
WD cooling hint White Dwarfs gae 0.5 " # gae Solar Neutrino flux g"aΓ # 0.0 ADMX ADMX!II IAXO Helioscopes " # Beam Dump " # " # !0.5 10!7 10!6 10!5 10!4 10!3 10!2 10!1 1 10 102 103 104 105 106 107 Not an axion-likema eV particle -2 + - Vacuum $ % e +e -> g+inv. -4 Birefringence Beam Dump largely ruled out -6 LSW SN1987a for many decades in ALP— couplings
D -8 � 1
- Solar n Helioscopes Telescopes HB GeV -
@ 10
˝ SN g-burst REAPR, ALPS-II IAXO a g xion See also: 10 -12 WD cooling hint BBN Log Yale 1512.03069 CMB -14 ADMX-HF
Cavities ADMX Microwave EBL
-16 non-Thermal DM DM X axion - Rays KSVZ -18 Thermal -12 -10 -8 -6 -4 -2 0 2 4 6 8 10
Log10 mALP eV Figure 6-1.HewettParameter et al. space “Fundamental for axions (top) Physics and axion-like at the particles Intensity (ALPs) Frontier” (bottom). 1205.2671 In the bottom plot, the QCD axion models lie within an order of magnitude from the explicitly shown “KSVZ” axion 8 line. Coloredflip regions . tanedo are: experimentally@ uci . edu excluded regions@ (darkDEVIDENCE green), constraints FOR A 17 fromMEV astronomicalNEW BOSON 14 observations (gray) or from astrophysical or cosmological arguments (blue), and sensitivity of planned experiments (light green). Shown in red are boundaries where ALPs can account for all the dark matter produced either thermally in the big bang or non-thermally by the misalignment mechanism.
Fundamental Physics at the Intensity Frontier Not your mother’s dark photon
2 KLOE ε WASA
) σ (3 2) e − 10-5 (g
HADES
(g−2) APEX " 0.011 µ A1 10-6 ⇡ BaBar
E774 NA48/2 Proposal: separate 10-7 E141
2 u, d and e couplings 10 0 10 2 ⇡ X mA’ (MeV/c ) Figure 4: Obtained upper limits at 90% CL! on the mixing parameter ε2 versus the DP mass + − mA′ ,comparedtootherpublishedexclusionlimitsfrommesondecay, beam dump and e e NA46/2 1504.00607collider experiments [16–22]. Also shown is the band where the inconsistency9 of theoretical and experimental values of muon (g 2) reduces to less than 2 standard deviations, as well as the flip . tanedo @ uci . edu EVIDENCE− FOR A 17 MEV NEW BOSON 14 region excluded by the electron (g 2) measurement [2, 23, 24]. −
± the mass range 2me
Conclusions
Asearchforthedarkphoton(DP)productionintheπ0 γA′ decay followed by the prompt → A′ e+e− decay has been performed using the data sample collected by the NA48/2 experiment → in 2003–2004. No DP signal is observed, providing new and morestringentupperlimitsonthe mixing parameter ε2 in the mass range 9–70 MeV/c2.Incombinationwithotherexperimental searches, this result rules out the DP as an explanation for the muon (g 2) measurement − under the assumption that the DP couples to quarks and decays predominantly to SM fermions. The NA48/2 sensitivity to the dark photon production in the K± π±A′ decay has also been → evaluated.
12 π0-phobia = p+-phobia To avoid NA46/2, prohibit π0 decay to X� X X 1 1 ⇡0 = uu¯ dd¯ ⇡0 = upu¯ dd¯ p2 Np=2 1 Goldstone uu¯ dd¯ n p of SU(2)L SU(2)R 2 ⇥ ✓ ◆
FROM QUARK CONTENT FROM CHIRAL PERT. THEORY
QuQu0 QdQd0 =0 p N = Qd0 = 2Qu0 n ✓ ◆
- For spin-1 10 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 Production and Decay 0.03
PRODUCTION 8 8 3 Br( Be⇤ Be X) 2 p~X 6 ! =("p + "n) | | 5.6 10 Br(8Be⇤ 8Be ) p~ 3 ⇡ ⇥ ! | |
DECAY
5 "e & 1.4 10 “Dark photon” with ⇥ separate couplings g " e i ⌘ i 11 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 4
its bounds are derived from X-bremsstrahlung from the initial p beam and ⇡0 decays to X bosons [24]. Both of these are suppressed in protophobic models. The CHARM experiment at CERN also bounds the param- eter space through searches for ⌘, ⌘0 X , followed by + ! X e e [25]. At the upper boundary of the region ex- cluded! by CHARM, the constraint is determined almost completely by the parameters that enter the X decay length, and so the dark photon bound on " applies to 5 "e and requires "e > 2 10 . A similar, but weaker constraint can be| derived| ⇥ from LSND data [26–28]. There are also bounds on the neutrino charge "⌫ .Inthe present case, where "e is non-zero, a recent study of B L gauge bosons [29] finds that these couplings are most stringently bounded by precision studies of⌫ ¯ e scat- tering from TEXONO for the mX of interest here [30]. Reinterpreted for the present case, these studies require 1/2 5 "⌫ "e . 7 10 . There are also bounds from co- herent| | neutrino-nucleus⇥ scattering. Dark matter experi- ments with Xe target nuclei require a B L gauge boson FIG. 2. The 8Be signal region, along with current constraints 5 discussed in the text (gray) and projected sensitivities of fu- to have coupling gB L . 4 10 [31]. Rescaling this to 8 the current case, given Z =⇥ 54 and A = 131 for Xe, we ture experiments in the (mX , "e)plane.Forthe Be signal, 1/2 4 the other couplings are assumed to be in the ranges given in find "⌫ "n < 2 10 . Eq. (10); for all other contours, the other couplings are those | | ⇥8 To explain the Be signal, "n must be significantly of a dark photon. larger than "e. Nevertheless, the⌫ ¯ e scattering con- 8 8 8 straint provides a bound on "⌫ that is comparable to or explains the Be anomaly by Be⇤ Be X, followed by + ! stronger than the ⌫ N constraint throughout parameter X e e , consistent with existing constraints. For "e ! | | space, and so we use the⌫ ¯ e constraint below. Note near the upper end of the allowed range in Eq. (10) and also that, given the range of acceptable " , the bounds "µ "e ,theX boson also solves the (g 2)µ puzzle, e | | ⇡ | | on "⌫ are more stringent than the bounds on "e, and so reducing the current 3.6 discrepancy to below 2 [9]. + B(X e e ) 100%, justifying our assumption above. Conclusions. We find evidence in the recent obser- ! ⇡ + Although not our main concern, there are also bounds vation of a 6.8 anomaly in the e e distribution of on second-generation couplings. For example, NA48/2 nuclear 8Be decays for a new vector gauge boson. The also derives bounds on K+ ⇡+X, followed by X new particle mediates a fifth force with a characteristic + ! ! e e [10]. However, this branching ratio vanishes for length scale of 12 fm. The requirements of the signal, massless X and is highly suppressed for low mX . For along with the many constraints from other experiments mX = 17 MeV, the bound on "n is not competitive with that probe these low energy scales, constrain the mass those discussed above [9, 11]. KLOE-2 also searches for and couplings of the boson to small ranges: its mass is Production (quark) couplings ⌘X followed by X e+e and excludes the dark mX 17 MeV, and it has milli-charged couplings to ! ! 3 ⇡ photon parameter " 7 10 [32]. This is similar up and down quarks and electrons, but with relatively . ⇥ numericallyShaded Regions to bounds discussed above, and the strange suppressed (and possibly vanishing) couplings to protons (and neutrinos) relative to neutrons. If its lepton cou- quark chargeAllowed"s can be chosen to satisfy this constraint. In summary, in the extreme protophobic case with plings are approximately generation-independent, the 17 MeV vector boson may simultaneously explain the exist- mX 17 MeV, the charges are required to satisfy ⇡ 2 4 3 ing 3.6 deviation from SM predictions in the anomalous "n < 2.5 10 and 2Barbieri10 g " e i ⌘ i UCI IPC 1604.07411 12 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 4 its bounds are derived from X-bremsstrahlung from the initial p beam and ⇡0 decays to X bosons [24]. Both of these are suppressed in protophobic models. The CHARM experiment at CERN also bounds the param- eter space through searches for ⌘, ⌘0 X , followed by + ! X e e [25]. At the upper boundary of the region ex- cluded! by CHARM, the constraint is determined almost completely by the parameters that enter the X decay length, and so the dark photon bound on " applies to 5 "e and requires "e > 2 10 . A similar, but weaker constraint can be| derived| ⇥ from LSND data [26–28]. There are also bounds on the neutrino charge "⌫ .Inthe present case, where "e is non-zero, a recent study of B L gauge bosons [29] finds that these couplings are most stringently bounded by precision studies of⌫ ¯ e scat- tering from TEXONO for the mX of interest here [30]. Reinterpreted for the present case, these studies require 1/2 5 "⌫ "e . 7 10 . There are also bounds from co- herent| | neutrino-nucleus⇥ scattering. Dark matter experi- ments with Xe target nuclei require a B L gauge boson FIG. 2. The 8Be signal region, along with current constraints 5 discussed in the text (gray) and projected sensitivities of fu- to have coupling gB L . 4 10 [31]. Rescaling this to 8 the current case, given Z =⇥ 54 and A = 131 for Xe, we ture experiments in the (mX , "e)plane.Forthe Be signal, 1/2 4 the other couplings are assumed to be in the ranges given in find "⌫ "n < 2 10 . Eq. (10); for all other contours, the other couplings are those | | ⇥8 To explain the Be signal, "n must be significantly of a dark photon. larger than "e. Nevertheless, the⌫ ¯ e scattering con- 8 8 8 straint provides a bound on "⌫ that is comparable to or explains the Be anomaly by Be⇤ Be X, followed by + ! stronger than the ⌫ N constraint throughout parameter X e e , consistent with existing constraints. For "e ! | | space, and so we use the⌫ ¯ e constraint below. Note near the upper end of the allowed range in Eq. (10) and also that, given the range of acceptable " , the bounds "µ "e ,theX boson also solves the (g 2)µ puzzle, e | | ⇡ | | on "⌫ are more stringent than the bounds on "e, and so reducing the current 3.6 discrepancy to below 2 [9]. + B(X e e ) 100%, justifying our assumption above. Conclusions. We find evidence in the recent obser- ! ⇡ + Although not our main concern, there are also bounds vation of a 6.8 anomaly in the e e distribution of on second-generation couplings. For example, NA48/2 nuclear 8Be decays for a new vector gauge boson. The also derives bounds on K+ ⇡+X, followed by X new particle mediates a fifth force with a characteristic + ! ! e e [10]. However, this branching ratio vanishes for length scale of 12 fm. The requirements of the signal, massless X and is highly suppressed for low mX . For along with the many constraints from other experiments mX = 17 MeV, the bound on "n is not competitive with that probe these low energy scales, constrain the mass those discussed above [9, 11]. KLOE-2 also searches for and couplings of the boson to small ranges: its mass is Decay (lepton) couplings ⌘X followed by X e+e and excludes the dark mX 17 MeV, and it has milli-charged couplings to ! ! 3 ⇡ photon parameter " 7 10 [32]. This is similar up and down quarks and electrons, but with relatively . ⇥ Shaded Regions numerically to bounds discussedKLOE-2 above, and the strange suppressed (and possibly vanishing) couplings to protons quark charge " can be chosen to satisfy this constraint. (and neutrinos) relative to neutrons. If its lepton cou- E141 Ruled Out s + In summary, in thee extremee protophobic X case with plings are approximately generation-independent, the 17 Lower bound MeV vector boson may simultaneously explain the exist- mX 17 MeV, the charges are! required to satisfy on εe: decay ⇡ 2 4 3 ing 3.6 deviation from SM predictions in the anomalous "n < 2.5 10 and 2 10 < "e < 1.4 10 , | | 1/⇥2 5 ⇥ | | ⇥ magnetic moment of the muon. It is also interesting to inside dump and "⌫ "e . 7 10 . Combining these with Eqs. (5) and| (7),| we find⇥ that a protophobic gauge boson with note that couplings of this magnitude, albeit in an ax- 0 + ial vector case, may resolve a 3.2 excess in ⇡ e e first-generation charges PROTOPHOBIA ! 8 decays [33, 34]. Be 8 1 3 To confirm the Be signal, the most direct approach " = " 3.7 10 Lower bound u 3 n ⇡ ± ⇥ would be to look for other nuclear states that decay to on εe: decay 2 3 discrete gamma rays with energies above 17 MeV through " = " 7.4 10 in detector d 3 n ⇡⌥ ⇥ M1 or E1 electromagnetic transitions. Unfortunately, 4 3 8 2 10 "e 1.4 10 the Be system is quite special and, to our knowledge, . . 8 8 ⇥ | | ⇥ the Be⇤ and Be⇤ states yield the most energetic such 1/2 5 0 " " 7 10 (10) | ⌫ e| . ⇥ gamma rays of all the nuclear states. UCI IPC 1604.07411 13 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 Future experiments See also: reach from SeaQuest SUSAN this sesson MMAPS @ Cornell? UCI IPC 1604.07411 14 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 BART SUSAN IPC IFTAH FLIP ATOMKI PAIR SPECTROMETER Extra Slides Internal22 PairJ. Gulyás Conversion et al. / Nuclear Instruments and Methods in Physics Research A 808 (2016) 21–28 -1 4. Overview of pair spectrometers 10 Electric & magnetic NuclearMagnetic de-excitationβ ray spectrometers were used first for internal pair fi 4 transitions (ℓ-wave) formation studies [30–34]. Maximal detection ef ciency of 10À byfor off-shell electron–positron photon pair detection was achieved for a few cases [32,34]. Improvement of the pair resolution by improvement of -2 the momentum resolution (to 1.3%) with smaller particle trans- 6 10 mission reduced the efficiency to 5 10 À . An important advance  E0 [33] in the use of intermediate-image pair spectrometer was provided by the installation* of a specially designed spiral baffle system which selected� electron–positron internal pairs emitted at E1 large relative angles (50 θ 90 ). IPCC 1r r 1 The next generation of internal-pair spectrometers used two dN/d θ -3 E2 10 dE=dx E scintillator-detector telescopes for the detection of the þ M1 electron–positron pairs in quadruple coincidence [35,36]. A multi- detector (six scintillation electron telescopes plus an annular Si(Li) particle detector) high-efficiency pair spectrometer was built by M2 Birk and co-workers [37]. An experimental pair-line efficiency of -4 Smooth, 28% and a sum-peak energy resolution of 12% for the 6.05 MeV E0 10 monotonically decreasing pair line in 16O were achieved. Schumann and Waldschmidt have detected internal pair spec- 0 20 40 60 80 100 120 140 160 180 tra in the energy range of 2.8–6.5 MeV from an (n,γ) reaction with Θ (degree) a combination of super-conducting solenoid transporter plus Si Opening Angle θ [deg] (Li)-detector spectrometer [38]. The pair-line efficiency of the Fig. 1. Calculated angular correlations of eþ eÀ pairs obtained from IPC for different spectrometer [39] was large, but it had a very limited dis- multipolarities and a transition energy of Eγ 17 MeV based on the expressions of ¼ crimination power for different multipolarities in this energy GulyásRose [2]. et al. NIM 1504.00489 region. 17 flip . tanedo @ uci . edu 4 EVIDENCE FOR A 17 TheMEV Debrecen NEW BOSON superconducting solenoid14 transporter plus two- neutral boson. A limit of r4:1 10 À was obtained for the boson  Si(Li)-detector electron spectrometer was also adapted for to γ-ray branching ratio [25–29]. internal-pair studies [40]. The observed pair-line efficiency for two detectors operated in sum-coincidence mode was 35%, while the energy resolution was 0.6% at 2 MeV. A similar spectrometer built 2. Internal Pair Creation (IPC) by Kibédi and co-workers [41] and has been used recently for internal pair studies [42]. It was predicted [2–4] that the angular correlation between the A highly segmented phoswich array of plastic scintillators was eþ eÀ pairs (emitted in IPC) peaks at 01 and drops rapidly with the constructed for measurements of eþ eÀ pairs emitted in high- correlation angle (Θ) as shown in Fig. 1. energy electromagnetic transitions in nuclei by Montoya and co- The above calculations show that the angular correlations at workers [43]. Electron (positron) energies of 2–30 MeV can be small separation angles are almost independent of the multi- measured by each individual element, with a total transition energy resolution of δE/E 13% for a 20 MeV transition. The array polarity of the radiation, whereas at large separation angles, they ¼ covers 29% of the full solid angle and its efficiency is 1.6% for a depend critically upon the multipole order. Thus, it is important to 6 MeV E0 internal pair decay, and 1.1% for an 18 MeV E1 transition. measure angular correlations efficiently at large angles. A positron–electron pair spectroscopy instrument (PEPSI) was designed to measure transitions in the energy region of 10– 40 MeV by Buda and co-workers [44]. It consists of Nd2Fe14B fi 3. The two-body decay of a boson permanent magnets forming a compact 4π magnetic lter con- sisting of 12 positron and 20 electron mini-orange-like When a nuclear transition occurs by emission of a short-lived spectrometers. 13 A ΔE E multi-detector array was constructed by Stiebing and τo10 À s neutral particle, the annihilation into an eþeÀ pair is À ð Þ co-workers [45] from plastic scintillators for the simultaneous anti-parallel (i.e. Θcm 1801) in the center of mass system. In the ¼ measurement of energy and angular correlation of e þeÀ pairs laboratory system, their angular distribution is sharply peaked produced in internal pair conversion (IPC) of nuclear transitions up (FWHM 10 ) at intermediate angles due to the Lorentz boost and o 1 to 18 MeV. The array was designed to search for deviations from provides a unique signature for the existence and a measure for IPC stemming from the creation and subsequent decay into e þeÀ the mass of an intermediate boson. In order to search for such an pairs of a hypothetical short-lived neutral boson. The spectrometer anomaly in the angular correlation, we need a spectrometer with consisted of six ΔE E scintillator detector telescopes. The size of À sufficient angular resolution. the ΔE detectors, which determines the solid angle, were The invariant mass can be determined approximately from the 22 22 1 mm3 and placed at 110 mm from the target. The   correlation angle Θ between eþ and eÀ and from their energies in angular resolution of the spectrometer was ΔΘ 151, and the ¼ 5 the following way[26]: efficiency for one pair of telescopes was 3 10 À . The fixed  mounting angles of the telescopes made possible investigating 15 m2 1 y2 E2 sin 2 Θ=2 ; 1 correlation angles simultaneously. The investigated angular range ð À Þ ð Þ ð Þ extended from 201 to 1311. where E E þ E À 1:022 MeV is the transition energy and In this paper, we present a novel eþ eÀ pair spectrometer ¼ þ þ y E þ E À = E þ E À , with E þðÀÞ indicating the kinetic energy equipped with multi-wire proportional chambers and large ¼ð À Þ ð þ Þ of the positron (electron) in the laboratory system. volume plastic scintillator telescopes placed as close to the target A 6.8σ anomaly: opening angle 1 7Li(p,e+e-)8Be dN/d θ -1 10 ×10 PROTON ENERGY Ep=1.20MeV ×4 IPCC (relative unit) Ep=1.10MeV ×2 -2 E =1.04MeV 10 p ×1 Ep=0.80MeV 40 60 80 100 120 140 160 OpeningΘ (deg.) Angle θ [deg] Krasznahorkay et al. Phys. Rev. Lett 116 (2016) 042501 18 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 week ending PRL 116, 042501 (2016) PHYSICAL REVIEW LETTERS 29 JANUARY 2016 − 16 × 102 eþe pairs of the 6.05 MeV transition in O, and of the 4.44 and 15.11 MeV transitions in 12C excited in the 1200 O 11 12 16 (a) B p; γ C reaction (Ep 1.6 MeV) were used to cali- ð Þ ¼ 7 + - 8 brate the telescopes. A ϵrel 20% HPGe detector was also 1000 Li(p,e e ) Be ¼ E =441 keV used at 50 cm from the target to detect the 477.61 keV γ ray p 7 in the Li p; p0γ reaction [37], to monitor the Li content of the targetð as a functionÞ of time. 800 In order to check the effective thickness of the targets during the long runs, the shape (width) of the high-energy γ 600 rays was measured by a 100% HPGe detector. In the case of 5 the broad 18.15 MeV ( 168 keV) resonance, the energy × Γ Counts/channel of the detected γ rays is¼ determined by the energy of the 400 proton at the time of its capture (taking into account the energy loss in the target), so the energy distribution of the γ 200 rays reflects the energy distribution of the protons. The Be 14.6 MeV Be 17.6 MeV 8 8 intrinsic resolution of the detector was less than 10 keV at week ending PRL 116, 042501 (2016)0 PHYSICAL REVIEW LETTERS 29 JANUARY 2016 17.6 MeV and the line broadening caused by the target Decays0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 thickness was about 100 keV, allowing us a reliable− E 16 (MeV) × 102 eþe pairs of the 6.05 MeV transition insum O, and of the monitoring. 4.44 and 15.11 MeV transitions in 12C excited in the 1200 The raw spectra were continuously monitored during the O 11 12 (b) 16 (a) B p; γ C reaction (Ep 1.6 MeV) were used to cali- whole experiment. The counting rates were reasonablyð lowÞ ¼ 7 + - 8 brate the telescopes. A ϵrel 20% HPGe detector was also 1000 Li(p,e e ) Be ¼ E =441 keV and not challenging the electronics. We observedused only at a 50 cm from the target to detect the 477.61 keV γ ray p few percent gain shifts of the energy detectors, but7 1 in the Li p; p0γ reaction [37], to monitor the Li contentE1 of otherwise the whole spectrometer was stable duringthe target the ð as a functionÞ of time. 800 typically 1 week long experiments performed atIn each order to check the effective thickness of the targets bombarding energy. The targets were changed every 8 h. during the long runs, the shape (width) of the high-energy γ 600 The acceptance as a function of the correlation anglerays was in measured by a 100% HPGe detector. In the caseM1+0.02E1 of 5 comparison to isotropic emission was determined fromthe broad the 18.15 MeV ( 168 keV) resonance, the energy × Γ Counts/channel − ¼ 400 same data set by using uncorrelated eþe pairs of differentof the detected γ raysIPCC (relative unit) is determinedGate: by14.6 the MeV energy of the single electron events [36], and used to determineproton the at the time of its-1 capture (taking into account the 10 angular correlations of different IPC transitions Gate: 17.6 MeV energy loss in the target), so the energy distribution of the γ 200 simultaneously. rays reflects the energy distribution of the protons. The Be 14.6 MeV Be 17.6 MeV − M1 8 8 Figure 1 shows the total energy spectrum of eþeintrinsicpairs resolution of the detector was less than 10 keV at 0 measured at the proton absorption resonance of 44117.6 keV (a) MeV and the line broadening caused by the target 0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 − 40 60 80 100 120 140 160 and the angular correlations of the eþe pairs originatedthickness was about 100 keV, allowing us a reliable Θ (deg.) Esum (MeV) from the 17.6 MeV 1þ → 01þ isovector M1 transitionmonitoring. and the 14.6 MeV 1þ → 21þ transition (b). The raw spectraFIG. 1. were Measured continuouslyMixing total with monitored energy E1 transition spectrum during (a) the and angular (b) − The Monte Carlo simulations of the experimentwhole were experiment.correlation The (b) counting of the rateseþe werepairs reasonably originated from low the decay of performed using the GEANT code. Target chamber,and target not challengingthe 17.6 MeV the electronics. resonance compared We observed with the only simulated a angular Krasznahorkay et al. Phys. Rev. Lett 116 (2016) 042501 backing, windows, and detector geometries were includedfew percentcorrelations gain shifts[36] ofassuming the energyM1 (full detectors, curve) and butM1 1.4%E1 1 19 − þ E1 14 in order to model the detector response to eþe pairsotherwise and γ themixed whole transitionsflip spectrometer . tanedo (dashed@ was line).uci stable. edu during the EVIDENCE FOR A 17 MEV NEW BOSON − rays. The scattering of the eþe pairs and the effectstypically of the 1 week long experiments performed at each and it adds to the M1 decay of the resonance. Previously, external pair creation (EPC) in the surrounding materialsbombarding energy. The targets were changed every 8 h. pure M1 transitions from the decay of the 17.6 MeV were also investigated. Besides the IPC process,The the acceptance as a function of the correlation angle in resonance were assumed [24 26], which is reasonable for M1+0.02E1 background of γ radiation, EPC, and multiplecomparison lepton to isotropic emission was determined– from the scattering were considered in the simulations to facilitate the resonance itself, but not− for the underlying background. same data set by using uncorrelated e e pairs of different IPCC (relative unit) a thorough understanding of the spectrometer and the The contribution of the directþ capture depends on the target Gate: 14.6 MeV single electron events [36], and used to determine the -1 thickness if the energy loss of the beam in the target is 10 detector response [36]. angular correlations of different IPC transitions Gate: 17.6 MeV larger than the width of the resonance. The dashed For the 17.6 MeV transition we observed asimultaneously. slight simulated curve in Fig. 1(b) is obtained by fitting a small deviation from the simulated internal pair conversionFigure 1 shows the total energy spectrum of e e− pairs M1 (2.0%) E1 contribution to the dominantþ M1 one, which correlation (IPCC) curve at angles above 110°, butmeasured without at the proton absorption resonance of 441 keV (a) any structure, and the deviation could be fully explainedand the by angulardescribes correlations the experimental of the e e data− pairs reasonably originated well. 40 60 80 100 120 140 160 The 18.15 MeV resonanceþ is isoscalar and much broader Θ (deg.) admixing some E1 component typical for the background.from the 17.6 MeV 1 → 0 isovector M1 transition and ( 168þkeV) 1þ[29], than the one at 17.6 MeV The background originates from the direct (nonresonant)the 14.6 MeVΓ 1 → 2 transition (b). ( ¼þ12.2 1þkeV) [29] and its strength is more distributed.FIG. 1. Measured total energy spectrum (a) and angular proton capture and its multipolarity is dominantly E1 [38], Γ − The Monte Carlo¼ simulations of the experiment were correlation (b) of the eþe pairs originated from the decay of performed using the GEANT code. Target chamber, target the 17.6 MeV resonance compared with the simulated angular backing,042501-2 windows, and detector geometries were included correlations [36] assuming M1 (full curve) and M1 1.4%E1 − þ in order to model the detector response to eþe pairs and γ mixed transitions (dashed line). − rays. The scattering of the eþe pairs and the effects of the external pair creation (EPC) in the surrounding materials and it adds to the M1 decay of the resonance. Previously, were also investigated. Besides the IPC process, the pure M1 transitions from the decay of the 17.6 MeV background of γ radiation, EPC, and multiple lepton resonance were assumed [24–26], which is reasonable for scattering were considered in the simulations to facilitate the resonance itself, but not for the underlying background. a thorough understanding of the spectrometer and the The contribution of the direct capture depends on the target detector response [36]. thickness if the energy loss of the beam in the target is For the 17.6 MeV transition we observed a slight larger than the width of the resonance. The dashed deviation from the simulated internal pair conversion simulated curve in Fig. 1(b) is obtained by fitting a small correlation (IPCC) curve at angles above 110°, but without (2.0%) E1 contribution to the dominant M1 one, which any structure, and the deviation could be fully explained by describes the experimental data reasonably well. admixing some E1 component typical for the background. The 18.15 MeV resonance is isoscalar and much broader The background originates from the direct (nonresonant) (Γ 168 keV) [29], than the one at 17.6 MeV ¼ proton capture and its multipolarity is dominantly E1 [38], (Γ 12.2 keV) [29] and its strength is more distributed. ¼ 042501-2 ATOMKIThe completed spectrometer Pair Spectrometer from A.J. Krasznahorkay; slideplayer.com/slide/6112261/ 20 flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 week ending PRL 116, 042501 (2016) PHYSICAL REVIEW LETTERS 29 JANUARY 2016 shape of the resonance [40], but it is definitely different pairs (rescaled for better separation) compared with the from the shape of the forward or backward asymmetry [40]. simulations (full curve) including only M1 and E1 con- Therefore, the above experimental data make the interpre- tributions. The experimental data do not deviate from the tation of the observed anomaly less probable as being the normal IPC. This fact supports also the assumption of the consequence of some kind of interference effects. boson decay. The deviation cannot be explained by any γ-ray related The χ2 analysis mentioned above to judge the signifi- Significancebackground either, since we cannot see any effect at off cance of the observed anomaly was extended to extract the resonance, where the γ-ray background is almost the same. mass of the hypothetical boson. The simulated angular To the best of our knowledge, the observed anomaly can correlations included contributions from bosons with 2 not have a nuclear physics related origin. masses between m0c 15 and 17.5 MeV. As a result The deviation observed at the bombarding energy of of the χ2 analysis, we¼ determined the boson mass to be 2 Ep 1.10 MeV and at Θ ≈ 140° has a significance of 6.8 m0c 16.70 0.35 stat MeV. The minimum value for standard¼ deviations, corresponding to a background fluc- the χ2¼=f was 1.07,Æ whileð Þ the values at 15 and 17.5 MeV tuation probability of 5.6 × 10−12. On resonance, the M1 were 7.5 and 6.0, respectively. A systematic error caused by contribution should be even larger, so the background the instability of the beam position on the target, as well as should decrease faster than in other cases, which would the uncertainties in the calibration and positioning of the make the deviation even larger and more significant. detectors is estimated to be ΔΘ 6°, which corresponds to − ¼ The eþe decay of a hypothetical boson emitted iso- 0.5 MeV uncertainty in the boson mass. tropically from the target has been simulated together with Since, in contrast to the case of 17.6 MeV isovector − the normal IPC emission of eþe pairs. The sensitivity of transition, the observed anomalous enhancement of the the angular correlation measurements to the mass of the 18.15 MeV isoscalar transition could only be explained by assumed boson is illustrated in Fig. 4. also assessing a particle, then it must be of isoscalar nature. Taking into account an IPC coefficient of 3.9 × 10−3 for 21 The invariant mass distribution calculated from the flip . tanedo @ uci . edu EVIDENCE FOR A 17 MEV NEW BOSON 14 the 18.15 MeV M1 transition [32], a boson to γ branching measured energies and angles was also derived. It is shown ratio of 5.8 × 10−6 was found for the best fit and was then in Fig. 5. used for the other boson masses in Fig. 4. The dashed line shows the result of the simulation According to the simulations, the contribution of the performed for M1 23%E1 mixed IPC transition (the assumed boson should be negligible for asymmetric pairs mixing ratio was determinedþ from fitting the experimental with 0.5 ≤ y ≤ 1.0. The open circles with error bars in angular correlations), the dotted line shows the simulation Fig. 4 showj thej experimental data obtained for asymmetric for the decay of a particle with mass of 16.6 MeV=c2 while the dash-dotted line is their sum, which describes the -1 experimental data reasonably well. 10 − In conclusion, we have measured the eþe angular correlation in internal pair creation for the M1 transition depopulating the 18.15 MeV state in 8Be, and observed a peaklike deviation from the predicted IPC. To the best of =17.6 MeV =15.6 MeV 2 2 c c =16.6 MeV 0 0 2 c m m 0 800 m 700 IPCC (relative unit) 600 500 =16.6 MeV -2 2 c 10 400 0 m 300 IPC, M1+E1 (Weighted Counts/0.5 MeV) 80 90 100 110 120 130 140 150 160 170 e+e- 200 Θ (deg.) N 100 − FIG. 4. Experimental angular eþe pair correlations measured 7 0 in the Li p; e e− reaction at E 1.10 MeV with 9 10 11 12 13 14 15 16 17 18 ð þ Þ p ¼ −0.5 ≤ y ≤ 0.5 (closed circles) and y ≥ 0.5 (open circles). me+e- (MeV) The results of simulations of boson decayj j pairs added to those of IPC pairs are shown for different boson masses as described in FIG. 5. Invariant mass distribution derived for the 18.15 MeV the text. transition in 8Be. 042501-4 Pre-History: Extinct 8Be Anomaly 238 F.W,N. de Boer et al./ Physics Letters B 388 (1996) 235-240 b) broad excess .-c 1.5 consistent with BG g likely E1 pollution 61.25 0 .s! =T 1 ll...~...~...~...,...,~.‘,., 20 40 60 80 100 120 140 20 40 60 80 100 120 14(I correlation angle o (degrees) correlation angle w (degrees) Fig. 3. (a) Measured12 angular correlation for the yield of e+e- 8Fig. 4. (a) Measured angular correlation for the yield of e+e- pairs from the reactionC E1’ 'B( p, transitione+e- ) ‘*C, using geometrical de- pairsBe from (17.6 the reaction MeV) 7Li(p, e+e-)‘Be M1 and transition curves similar to tector efficiencies and normalised to the theoretical El-IPC cor- Fig. 3a. (b) Ratio of data and IPC obtained in the same way as relations [ 9 ] (dashed line) in the angular range over this range. inexcluded Fig. 3b. by ATOMKI study The solid line includes effects from EPC [ 131 and multiple scat- tering calculated in a MC simulation (dot-dashed line). (b) Ratio de Boer et al. Phys. Lett. B388 (1996) 235 transitions) and 7Li(p, e+e- ) *Be (Ml -transitions), of the experimental data and the IPC prediction (dashed curve in 22 fliFig.p . 3a).tanedo Open circles@ denoteuci .correlations edu between the six small EVIDENCErespectively. FOR The A 17 data MEV sets NEWare scaled BOSON to unity at the 14 detectors, open triangles between one large detector (7) and small smallest correlation angle measured. The dashed lines detectors, open squares between the other large detector (8) and represent IPC distributions for non-aligned nuclei [ 91 small detectors. The data point at 68’ represents the correlation normalised to the data points at large w (w > 120’)) between the two large detectors (7) and (8) (see also Fig. I). where relative contributions from EPC and multiple scattering are minimal. The latter clearly show up at determined with adequate precision. To account for o < 50” and mainly arise from the carbon-fibre tube differences in the low-energy thresholds of the detec- used as vacuum window for the leptons. tors, in particular for the large detectors, normalisation The shape of these contributions for the particu- factors of 0.78 and 0.95 were applied to all combi- lar geometry of our apparatus (dot-dashed lines in nations of the small telescopes with telescope 7 and Figs. 3a and 4a) has been determined by means of 8 respectively. The detector combination 7-8 conse- GEANT Monte Carlo simulations [ 17 1. At 21 o the quently has been normalised by the product of both number of pairs due to EPC and multiple scattering factors. The level of systematic uncertainties due to was calculated for both reactions to be typically 45% detector, beam and target alignment, as well as due with respect to IPC. However a value of 57% is needed to energy and angular smearing, have been estimated. to achieve good agreement for the 12C data. Consid- The overall effect is expected to be roughly of the ering the uncertainties in the nontrivial determination same size as the statistical errors. This is confirmed by of the EPC contributions this appears acceptable to us. the spread - beyond statistical fluctuations - in data Moreover, the difference does not influence the data points of Figs. 3 and 4 where approximately the same at w > 50” to a significant extent. The solid lines in central w values are obtained from different pairs of Figs. 3a and 4a represent the sum of IPC (normalised detector telescopes. to the data at w > 120”) and EPC, both including mul- In Figs. 3a and 4a the measured angular correla- tiple scattering. As the angular correlations of e+e- tions of e+e- pairs with a sum energy above 5MeV pairs due to IPC decline by almost two orders of mag- are shown for the reactions “B(p, e+e-)‘*C (El- nitude in the w range considered, any possible devia-