search for solar axions: the cast experimentproblem have been postulated [2, 4], being the most...

Search for solar axions: the CAST experiment

S. Andriamonje1, S. Aune1, K. Barth2, A. Belov3, B. Beltrán4,1,H. Bräuninger5, J. M. Carmona4, S. Cebrián4, J. I. Collar6, T. Dafni7,2,

M. Davenport2, L. Di Lella2,3, C. Eleftheriadis8, J. Englhauser5

G. Fanourakis9, E. Ferrer-Ribas1, H. Fischer10, J. Franz10, P. Friedrich5,T. Geralis9, I. Giomataris1, S. Gninenko3, M. D. Hasinoff11,

F. H. Heinsius10, D. H. H. Hoffmann12, I. G. Irastorza1,4, J. Jacoby13,K. Jakovcic14 D. Kang10, K. Königsmann10, R. Kotthaus15, M. Krcmar14,

K. Kousouris9, M. Kuster7,5 B. Lakic14, C. Lasseur2, A. Liolios8,A. Ljubicic14, G. Lutz15, G. Luzón4, D. W. Miller6, A. Morales4,5,

J. Morales4, A. Ortiz4, T. Papaevangelou2, A. Placci2, G. Raffelt15,J. Ruz4, H. Riege7, P. Serpico15, L. Stewart2, J. D. Vieira6, J. Villar4,

J. Vogel10 L. Walckiers2, K. Zioutas16

1DAPNIA, Centre d’Études Nucléaires de Saclay (CEA-Saclay), Gif-sur-Yvette, France2European Organization for Nuclear Research (CERN), Genève, Switzerland

3Institute for Nuclear Research (INR), Russian Academy of Sciences, Moscow, Russia4Instituto de Física Nuclear y Altas Energías, Universidad de Zaragoza, Zaragoza, Spain

5Max-Planck-Institut für extraterrestrische Physik, Garching, Germany6Enrico Fermi Institute and KICP, University of Chicago, Chicago, IL, USA

7Technische Universität Darmstadt, Institut für Kernphysik, Schlossgartenstrasse 9, 64289Darmstadt

8Aristotle University of Thessaloniki, Thessaloniki, Greece9National Center for Scientific Research “Demokritos”, Athens, Greece

10Albert-Ludwigs-Universität Freiburg, Freiburg, Germany11Department of Physics and Astronomy, University of British Columbia, Vancouver, Canada12Gesellschaft für Schwerionenforschung, GSI-Darmstadt, Plasmaphyisk, Planckstr. 1, 64291

Darmstadt13Johann Wolfgang Goethe-Universität, Institut für Angewandte Physik, Frankfurt am Main,

Germany14Rudjer Boškovic Institute, Zagreb, Croatia

15Max-Planck-Institut für Physik (Werner-Heisenberg-Institut), München, Germany16Physics Department, University of Patras, Patras, Greece.

Abstract. Hypothetical axion-like particles with a two-photon interaction would be produced inthe sun by the Primakoff process. In a laboratory magnetic field they would be transformed into X-rays with energies of a few keV. The CAST experiment at CERN is using a decommissioned LHCmagnet as an axion helioscope in order to search for these axion-like particles. The analysis of the2003 data [1] showed no signal above the background, thus implying an upper limit to the axion-photon coupling of gaγ < 1.16×10−10 GeV−1 at 95% CL for ma � 0.02 eV. The stable operationof the experiment during 2004 data taking allowed us to lower down this parameter to a preliminaryvalue of gaγ < 0.9×10−10 GeV−1.

Keywords: Axions, Dark matter, Solar Physics, Low background

395

Downloaded 04 May 2007 to 137.138.4.27. Redistribution subject to AIP license or copyright, see http://proceedings.aip.org/proceedings/cpcr.jsp

PACS: 95.35.+d; 14.80.Mz; 07.85.Nc; 84.71.Ba

INTRODUCTION

QCD is the universally accepted theory for describing the strong interactions, but it hasone serious blemish: the so-called “strong CP problem”. QCD predicts the existenceof a CP violating term in the standard equations, yet Nature has never exhibited thisbehavior in any experiment [2, 3]. Various theoretical attempts to solve this strong CPproblem have been postulated [2, 4], being the most elegant solution the one proposed byPeccei and Quinn in 1977 [5, 6]. For this they introduced an additional global symmetry,known as the Peccei-Quinn symmetry U(1)PQ, to the Standard Model Lagrangian whichis spontaneously broken at a scale fPQ. Immediately and independently, Weinberg [7]and Wilczek [8] realized that, because U(1)PQ is spontaneously broken, there should bea pseudo-Goldstone boson, “the axion” (or as Weinberg originally referred to it, “thehigglet”). Because U(1)PQ suffers from a chiral anomaly, the axion acquires a smallmass of the order of ma ≈ 6 μeV

(1012 GeV/fPQ

).

A priori the value of the mass of the axion (or equivalently the fPQ scale) is arbi-trary, but it can be constrained using data from various experiments, astrophysical con-siderations (cooling rates of stars) and cosmological arguments (overclosure of the Uni-verse) [9, 10]. Nowadays it is believed to fall inside the so-called “axion mass window”:10−6 eV < ma < 10−3eV.

The interaction strength of axions with ordinary matter (photons, electrons andhadrons) scales [3] as 1/fPQ and so the larger this number, the more weakly the ax-ion couples. The present constraints on its mass make the axion a weakly interactingparticle, therefore a nice candidate for the Dark Matter of the Universe [10].

One generic property of the axions is a two-photon interaction of the form:

Laγ = −14

gaγFνμ Fνμa = gaγE ·Ba (1)

where F is the electromagnetic field-strength tensor, F is its dual, and E and B the elec-tric and magnetic fields. As a consequence axions can transform into photons in externalelectric or magnetic fields [11], an effect that may lead to measurable consequences inlaboratory or astrophysical observations. For example, stars could produce these par-ticles by transforming thermal photons in the fluctuating electromagnetic field of thestellar plasma [12, 9], or axions could contribute to the magnetically induced vacuumbirefringence, interfering with the corresponding QED effect [13, 14]. The PVLAS [15]

1 Attendant speaker, Berta.Beltrá[email protected], Present address: Department of Physics, Queen’s University,Kingston, Ontario K7L 3N6 Canada2 Present address: DAPNIA, Centre d’Études Nucléaires de Saclay (CEA-Saclay), Gif-sur-Yvette, France3 Present address: Scuola Normale Superiore, Pisa, Italy4 Present address: Instituto de Física Nuclear y Altas Energías, Universidad de Zaragoza, Zaragoza, Spain5 Deceased

396


experiment apparently observes this effect, although an interpretation in terms of axion-like particles requires a coupling strength far larger than existing limits.

The sun would be a strong axion source and thus offers a unique opportunity toactually detect such particles by taking advantage of their back-conversion into X-raysin laboratory magnetic fields [16]. The expected solar axion flux at the Earth due to thePrimakoff process is: Φa = g2

aγ ×3.67×1031 cm−2 s−1, with an approximately thermalspectral distribution given by:

dΦa

dEa= g2

aγ ×3.821×1030 (Ea/keV)3

(eEa/1.103 keV−1)cm−2 s−1 keV−1 (2)

and an average energy of 4.2 keV 1.

PRINCIPLE OF DETECTION

A particularly intriguing application of magnetically induced axion-photon conversionsis to search for solar axions using an “axion helioscope” as proposed by Sikivie [16].One looks at the sun through a “magnetic telescope” and places an X-ray detector at thefar end. Inside the magnetic field, the axion couples to a virtual photon, producing a realphoton via the Primakoff effect: a+γvirtual � γ . The energy of this photon is then equalto the axion’s total energy. The expected number of these photons that reach the X-raydetector is:

Nγ =∫

dΦa

dEaPa→γ ST dEa (3)

where dΦa/dEa is the axion flux at the Earth as given by eq.(2), S is the magnet borearea (cm2), T is the measurement time (s) and Pa→γ is the conversion probability of anaxion into a photon. If we take some realistic numbers (gaγ = 10−10 GeV−1,T = 100 hand S = 15 cm2) this number of photons would be nearly 30 events.

The conversion probability in vacuum is given by:

Pa→γ =(

Bgaγ

2

)2

2L2 1− cos(qL)(qL)2 (4)

where B and L are the magnetic field and its length (given in natural units), andq = m2

a/2E is the longitudinal momentum difference between the axion and an X-rayof energy E. The conversion process is coherent when the axion and the photon fieldsremain in phase over the length of the magnetic field region. The coherence conditionstates that [20, 21] qL =< π so that a coherence length of 10 m in vacuum requiresma � 0.02 eV for a photon energy 4.2 keV.

Coherence can be restored for a solar axion rest mass up to ∼ 1eV by filling themagnetic conversion region with a buffer gas [17] so that the photons inside the magnet

1 The spectrum in [17] has been changed to that proposed in [18], however with a modified normalizationconstant to match the total axion flux used here, which is predicted by a more recent solar model [19].

397


pipe acquire an effective mass whose wavelength can match that of the axion. For anappropriate gas pressure, coherence will be preserved for a narrow axion mass window.Thus, with the proper pressure settings it is possible to scan for higher axion masses.

The first implementation of the axion helioscope concept was performed at BNL [20].More recently, the Tokyo axion helioscope [22] with L = 2.3 m and B = 3.9 T hasprovided the limit gaγ < 6.0 × 10−10 at 95% CL for ma � 0.03 eV (vacuum) andgaγ < (6.8−10.9)×10−10 for ma � 0.3 eV (using a variable-pressure buffer gas) [23].Limits from crystal detectors [24, 25, 26] are much less restrictive.

CAST EXPERIMENT

In order to detect solar axions or to improve the existing limits on gaγ an axion he-lioscope (Fig 1) has been built at CERN by refurbishing a decommissioned LHC test

FIGURE 1. The CAST experiemnt at CERN

magnet [21] which produces a magnetic field of B = 9.0 T in the interior of two parallelpipes of length L = 9.26 m and a cross–sectional area S = 2× 14.5 cm2. The apertureof each of the bores fully covers the potentially axion-emitting solar core (∼ 1/10th ofthe solar radius). The magnet is mounted on a platform with ±8◦ vertical movement,allowing for observation of the sun for 1.5 h at both sunrise and sunset. The azimuthrange of 80◦ encompasses nearly the full azimuthal movement of the sun throughout theyear. The time the sun is not reachable is devoted to background measurements. A fullcryogenic station is used to cool the superconducting magnet down to 1.8 K needed forits superconducting operation [27]. The hardware and software of the tracking systemhave been precisely calibrated, by means of geometric survey measurements, in order toorient the magnet to the reachable celestial coordinates. The overall CAST pointing pre-cision is better [28] than 0.01◦ including all sources of inaccuracy such as astronomicalcalculations, as well as spatial position measurements. At both ends of the magnet, threedifferent detectors search for excess X-rays from axion conversion in the magnet whenit is pointing to the sun. Covering both bores of one of the magnet’s ends, a conven-tional Time Projection Chamber (TPC) is looking for X-rays from “sunset” axions. At

398


the other end, facing “sunrise” axions, a second smaller gaseous chamber with novel MI-CROMEGAS (micromesh gaseous structure – MM) [29] readout is placed behind oneof the magnet bores, while in the other one, a X-ray mirror telescope [30] is used witha Charge Coupled Device [31] (pn-CCD) as the focal plane detector. Both the pn-CCDand the X-ray telescope are prototypes developed for X-ray astronomy [32]. The X-raymirror telescope can produce an “axion image” of the sun by focusing the photons fromaxion conversion to a ∼ 6mm2 spot on the pn-CCD. The enhanced signal-to-backgroundratio substantially improves the sensitivity of the experiment.

DATA ANALYSIS AND RESULTS

2003 data tacking

CAST operated for about 6 months from May to November in 2003, during most ofwhich time at least one detector was taking data. An important feature of the CAST datatreatment is that the detector backgrounds are measured with ∼10 times longer exposureduring the non-alignment periods. The use of these data to estimate and subtract the trueexperimental background during sun tracking data is the most sensitive step in the CASTanalysis. To assure the absence of systematic effects, the main strategy of CAST is theuse of three independent detectors with complementary approaches.

These data were compatible with the absence of any signal, thus allowing us to extractthe following limit to the axion-to-photon coupling constant [1]:

gaγ < 1.16×10−10GeV−1(95%CL). (5)

2004 data tacking

All detectors and the magnet moving platform were operating with improved condi-tions during 2004, which lead to a very stable operation during that year, and thereforeto a wider set of data. Because of this the calculated upper limit to the axion-to-photoncoupling constant was lowered to the preliminary value of:

gaγ < 0.9×10−10GeV−1(95%CL). (6)

Thus far, our analysis in both years was limited to the mass range ma � 0.02 eV wherethe expected signal is mass-independent because the axion-photon oscillation length farexceeds the length of the magnet. For higher ma the overall signal strength diminishesrapidly and the spectral shape differs as it can bee seen in figure 2.

SUMMARY

The origin of the axion as a particle that solves the strong CP problem has been reviewed.Some properties of this pseudoescalar particle have been pointed out, among them thefact that it can transform into a photon in external electric or magnetic fields, this being

399


(eV)axionm-510 -410 -310 -210 -110 1 10

)-1

(GeV

γag

-1210

-1110

-1010

-910

-810

-710

Prelim

inar

y-510 -410 -310 -210 -110 1 10

Axion

mod

els

DAMA

SOLAX, COSME

CAST prospectsCAST prospects

Tokyo helioscope

Lazarus et al.

globular clustersCAST 2004

CAST combined limit

FIGURE 2. Exclusion limit at 95% C.L. from the 2004 CAST data. For comparison, the result fromother experiments is also shown. The shaded region represents the theoretical models

the only property of the axion on which CAST relies. The CAST experiment and its firstresults [1]have been presented.

Our 2004 limit improves the best previous laboratory constraints [22] on gaγ by afactor ∼ 6.5 in our coherence region ma � 0.02 eV and surpass the astrophysical limitset by the globular clusters.

In 2005 the so-called second phase of CAST started, where data are taken with avarying-pressure buffer gas in the magnet pipes in order to restore coherence for axionmasses above 0.02 eV. The extended sensitivity to higher axion masses will allow usto enter into the shaded region shown in Fig. 2, which is especially motivated by axionmodels [33], as shown by the “CAST prospects” line.

ACKNOWLEDGMENTS

This work has been performed in the CAST collaboration. We thank our colleagues fortheir support. Also we acknowledge the helpful discussions within the network on directdark matter detection of the ILIAS integrating activity (Contract number: RII3-CT-2003-506222)

400


REFERENCES

1. K. Zioutas et al. [CAST Collaboration], Phys. Rev. Lett. 94 (2005) 1213012. C. Jarlskog(Ed), World Scientific (1989)3. M. S. Turner, Phys. Rep. 197, 67 (1990).4. M. Dine, [arXiv:hep-ph/0011376].5. R. D. Peccei and H. R. Quinn, Phys. Rev. D 16 (1977) 1791.6. R. D. Peccei and H. R. Quinn, Phys. Rev. Lett. 38 (1977) 1440.7. S. Weinberg, Phys. Rev. Lett. 40, 223 (1978).8. F. Wilczek, Phys. Rev. Lett. 40, 279 (1978).9. G. G. Raffelt, Ann. Rev. Nucl. Part. Sci. 49, 163 (1999) [arXiv:hep-ph/9903472].10. S. Eidelman et al. [Particle Data Group], Phys. Lett. B 592 (2004) 1.11. H. Primakoff, Phys. Rev. 81, 899 (1951).12. D. A. Dicus, E. W. Kolb, V. L. Teplitz and R. V. Wagoner, Phys. Rev. D 18 (1978) 1829.13. L. Maiani, R. Petronzio and E. Zavattini, Phys. Lett. B 175 (1986) 359.14. G. Raffelt and L. Stodolsky, Phys. Rev. D 37 (1988) 1237.15. G. Cantatore et al., in Proceedings of the 5th International Workshop on the Identification of Dark

Matter, Edinburgh, UK, 2004 ( to be publised). Also see U.Gastaldi in this volume.16. P. Sikivie, Phys. Rev. Lett. 51, 1415 (1983) [Erratum-ibid. 52, 695 (1984)].17. K. van Bibber et al., Phys. Rev. D 39, 2089 (1989).18. R. J. Creswick et al., Phys. Lett. B 427, 235 (1998).19. J. N. Bahcall, M. H. Pinsonneault and S. Basu, Astrophys. J. 555 (2001) 990 [arXiv:astro-

ph/0010346].20. D. M. Lazarus et al., Phys. Rev. Lett. 69, 2333 (1992).21. K. Zioutas et al., Nucl. Instrum. Meth. A 425 (1999) 480 [arXiv:astro-ph/9801176].22. S. Moriyama et al., Phys. Lett. B 434, 147 (1998).23. Y. Inoue et al., Phys. Lett. B 536, 18 (2002).24. F. T. Avignone et al., Phys. Rev. Lett. 81, 5068 (1998).25. A. Morales et al., Astropart. Phys. 16, 325 (2002).26. R. Bernabei et al., Phys. Lett. B 515 6 (2001).27. K. Barth et al., Proc. 2003 Cryogenic Engineering Conference (CEC) and Cryogenic Materials

Conference (ICMC).28. http://cast.web.cern.ch/CAST/edited_tracking.mov29. Y. Giomataris et al., Nucl. Instrum. Meth. A 376 29 (1996).30. R. Kotthaus et al., arXiv:astro-ph/0511390.31. L. Struder et al., Astron. Astrophys. 365, L18 (2001).32. J. Altmann et al., in Proceedings of SPIE: X-Ray Optics, Instruments, and Mission, 1998, edited by

Richard B. Hoover and Arthur B. Walker, p. 350; J. W. Egle et al., ibid., p. 359; P. Friedrich et al.,ibid., p. 369.

33. D. B. Kaplan, Nucl. Phys. B 260, 215 (1985).

401


search for solar axions: the cast experimentproblem have been postulated [2, 4], being the most...

Documents