An Ultra-low-background detector for axion search

1. An Ultra-low-background detector for axion search T. Papaevangelou1, S. Aune1, T. Dafni2, G. Fanourakis3,E. Ferrer Ribas1,J. GalánLacarra2, T. Geralis3, I. Giomataris1, F. J. Iguaz2, I.G. Irastorza2, K. Kousouris3, J. Morales2, J.P. Mols1, J. Ruz2, A. Tomás2 1IRFU, Centre d’Etudes de Saclay, Gif sur Yvette CEDEX, France 2Instituto de Física Nuclear y Altas Energías, Zaragoza, Spain 3Institute of Nuclear Physics, NCSR Demokritos, Athens, Greece 4th Symposium on Large TPCs for Low Energy Rare Event Detection, Paris 2008

2. Outline • Axions • The CAST Experiment • The Micromegas for solar axion search in CAST • An Ultra-low-background Micromegas • Conclusions - Outlook

3. Cleaning up a long standing problem in theoretical physics (Wilczek) CP violation in QCD Solution of the U(1)A problem from t’Hooft (θ-vacuum) resulted in a CP-violating term in QCD lagrangian: Experimental consequence: prediction of electric dipole moment for the neutron in contradiction with experimental results Peccei-Quinn (1977) proposed an elegant solution: Why so small?  Strong CP problem new PQ symmetry spontaneously broken at an energy scale fα (not fixed) The AXION appears as the Nambu-Goldstone boson of the spontaneous breaking of the PQ symmetry (Weinberg and Wilczek, 1978) PQ symmetry is not exact  axions acquire mass • 1st scenario: • fa∼ fEW → ma ∼ O(1MeV) • Visible axions, but… ruled out by experiments • 2nd scenario: • fa>> fEW → ma << O(1MeV) • Invisible axions, what CAST is looking for

4. Axion summary • Axion properties • Neutral pseudoscalar, • practically stable, • very low mass, • spin-parity 0- • very low interaction cross-sections •  Nearly invisible to ordinary matter •  excellent candidates for dark matter (together with WIMPS) Axions couple to photons L = g(E•B)a (PRIMAKOFF EFFECT) Astrophysical & cosmologicalcontraints on ma, fasmall window in ga − ma plane left Axions are viable candidates for Cold and Hot DarkMatter ( 10−6eV ≤ ma ≤ 1.05 eV)

5. Axion searches • Cosmological implications (stellar energy loss, overclosure…) • Experimental • Laboratory searches (laser experiments, PVLAS, OSQAR, etc) • Dark matter axion searches (microwave cavity experiments, ADMX ) • Solar axion searches (helioscopes Tokyo, CAST )  search for axion – photon interaction (coupling constant gaγγ & axion mass ma) The principle of axion detection with helioscopes (Sikivie 1983) Signal: excess of x-rays during alignment over background A thermal photon converts into an axion in the Coulomb fields of nuclei and electrons in the solar plasma The axion converts into a photon under a strong transverse magnetic field in the laboratory (inverse process)

6. Coherent axion – photon conversion Axion to photon conversion probability: Coherence condition: qL< π L magnet’s length, q the momentum transfer For the CAST magnet coherence conversion is lost for ma > 0.02 eV Coherence can be restored using a buffer gas inside the magnet bores (3He, 4He) But: only a very narrow axion mass range is probed for a given gas density • A broad mass range can be tested by changing the gas density (pressure). • For every pressure setting there is a new discovery potential !!! For P~50 mbar Δma ~ 10-4eV !!!

7. CAST CAST CAST CAST CAST, who we are 24 Institutes, ~80 scientists The CAST Collaboration CEA Saclay, France – CERN, Switzerland – Dogus University, Turkey – INFN University of Trieste, Italy – Lawrence Livermore National Laboratory, USA – Max-Planck-Institut für extraterrestrische Physik, Germany – Max-Planck-Institut für Physik, Germany – Max-Planck-Institut für Physik Werner-Heisenberg-Institut, Germany – Max-Planck-Institut für Solare Physik, Germany – National Center for Scientific Research Demokritos, Greece – National Technical University of Athens, Greece – Rudjer Boskovic Institute, Croatia – Institute for Nuclear Research (Moscow), Russia – TU Darmstadt, Germany – University of British Columbia, Canada – University of Chicago, USA – Universität Frankfurt, Germany – Universität Freiburg, Germany – Universität Zürich, Switzerland – University of Florida, USA – University of Patras, Greece – University of Thessaloniki, Greece – Universidad de Zaragoza, Spain

8. CAST Experiment Solar axions may convert to photons inside the field of a prototype LHC magnet via inversed Primakoff effect Signal: X-Ray excess during tracking (1-10 keV) Low count rate  Low background Detector stability CAST phase II  different pressure setting(s) in every tracking. Each setting is a new experiment! A background rate ~0 cts/h is essential for the discovery potential Axion flux on earth expect 0.3 events/hour for gaγγ= 10-10 GeV-1 and A = 14.5 cm2 • Decommissioned prototype LHC dipole magnet. • Superconducting, operation at T=1.8 K. • Electric current 13,000 A. • Magnetic field: B=9T • Length: L=9.26m. • Low background • Low radioactivity materials (plexiglas, copper, kapton) • Particle recognition • Shielding • X-Ray focusing device

9. CAST detectors (phase I & phase II-4He) Sunset side Sunrise side CCD + X-Ray Telescope (prototype for the ABRIXAS Space mission) Shielded TPC, covering both magnet bores Rotating platform ( Vertical: ±8, Horizontal: ±40) ~90 min solar tracking during sunrise/sunset Unshielded Micromegas

10. CAST Phase I result TPC subtracted spectrum MM subtracted spectrum Data taking during 2003 and 2004 (total 12 months) Result from CAST phase I: CCD spectrum (ma < 0.02 eV) JCAP04(2007)010, CAST Collaboration PRL (2005) 94, 121301, CAST Collaboration

11. CAST Phase II-4He result (to be published in JCAP) TPC typical daily data (1 pressure step) • Spent at least one full tracking per pressure setting • Measure / calculate corresponding backgrounds • Compare with tracking rates • During 2005/2006 we have covered 160 4He pressure settings, reachingP = 13.43 mbar  ma~ 0.39 eV/c2 Every day is a “new” experiment!!! CCD tracking count distribution Micromegas count rate vs pressure mean background 0.27 cts/tracking!

12. CAST Phase I & II-4He result to be published in JCAP CAST has entered the theoretically favored axion model region! CAST experimental limit dominates in the most of the favored (cosmology/astrophysics) parameter space

13. Micromegas in CAST Phase I (2003-2004) & He-4 phase (2005-2006) one conventional not shielded Micromegas was used (sunrise) Readout :192 x 192 strips, 350 μm pitch Gas mixture: 95% Ar, 5% Isobutane @ 1 bar (flamable) X-Ray detection Threshold: ~0.6keV Background ~510-5 events keV-1s-1cm-2 Position sensitivity ~100 μm

14. Micromegas Technologies • Conventional technology • The pillars are attached to the mesh. A supporting ring or frame is adjusting the mesh on top of the readout plane • Typical dimensions: mesh thickness 5 μm, gap 50 μm • Bulk technology • The pillars are attached to a woven mesh and to the readout plane • Typical dimensions: mesh thickness 30 μm, gap 100 μm • Microbulk technology • The pillars are constructed by chemical process of a kapton foil, that is attached to the mesh and to the readout plane • Typical dimensions: mesh thickness 5 μm, gap 50 μm

15. Bulk technology Woven Inox mesh 30 µm Readout plane + mesh all in one vacrel 128 µm Well established technique Readout pads Advantages: Uniformity, Reachable resolution (~18% @ FWHM, limited by the thickness of the mesh), very robust, low noise due to lower capacity, easy to construct Disadvantages: higher risetime, sensitivity to pressure variations, stability pad pillar

16. Microbulk technology Micromesh 5µm copper Readout plane + mesh all in one Kapton 50 µm Readout pads Innovative technique, well developed during the last year Advantages: Uniformity, reachable resolution (better than 13% @ FWHM), stability at long term runs, less sensitivity to pressure variations, background rejection Disadvantages: Higher electronic noise due to higher capacity, complexity of the manufacturing process, fragility Mesh 5 µm of copper Holes 30 µm diameter Pitch 100 µm Pads of 400 μm

17. Micromegas in CAST, 3He Phase Sunrise line: it has been redesigned for Shielding X-Ray focusing optics A Bulk detector replaced the conventional one Readout :106 x 106 strips, 550 μm pitch Gas mixture: 97.7% Ar, 2.3% Isobutane @ 1.44 bar (non flammable – increased efficiency) • Sunset side: • the TPC was replaced by one Bulkand one Microbulk detector • Readout :106 x 106 strips, 550 μm pitch • Gas mixture: 97.7% Ar, 2.3% Isobutane @ 1.44 bar Under construction at LLNL

18. Detector Performance Long term stability of gain Energy resolution 5 % Isobutane 2.3 % Isobutane 18% FWHM Spatial resolution Calibration @ Panter with X-Ray telescope Image from a lead foil with pinholes Efficiency

19. Micromegas readout Mesh: 1 GHz FADC, 2.5 μsec (MATACQ) Strips:Integrated charge at each strip in groups of 96 (Gasiplex) 55Fe • 1 – 10 keV X-Rays: Localized energy deposition (<1 mm)  • Short risetime and pulse width for the analogue signal • Charge in one cluster of few strips per axis • Pattern recognition algorithm can be applied to reduce background 55Fe cosmic cosmic 55Fe

20. Pattern recognition Define signal characteristics for 55Fe X-Rays (daily calibration runs) Strips: multiplicity, width, topology of clusters Mesh: risetime, width, amplitude, integral of signals Examine distributions Define cuts from bins with content above a relative value Apply cuts in data runs Sequential Multivariate analysis Neural networks Optimize between efficiency and rejection Data reduction >103 55Fe data cut 55Fe data cut 55Fe data

21. Background of new micromegas in CAST The new background levels imply 2-3 expected bkg counts per pressure setting (~2800 sec solar tracking per setting) compared to 12 (Micromegas) and 40 (TPC) from 4He phase The combined performance of the three detectors is comparable with the Telescope-CCD system concerning discovery potential 2007: Implementation of shielding (Pb+ Cd + Cu + Polyethylene) Background during Summer 2008 from the sunset B4 detector • The background level (after cuts) is reduced by a factor 3 to 5 • Implementation of a telescope in the sunrise side can decrease the background by a factor ~100

22. The Ultra-low-background detector The bulk detectors were build on 400 μm of PCB for technical reasons Possibly increased intrinsic background The bulk detectors, due to the bigger amplification gap, have potentially reduced rejection efficiency The first microbulks up to M8 had shown several defects, since the technique was still under development Possibly reduced rejection efficiency A new series of defect free microbulk detectors was constructed The new detectors were installed in CAST around October 24th

23. New Detector Performance Software efficiency @ 6 keV Calibration @ CAST with 55Fe from ~1 m distance Long term stability of gain Software efficiency @ 3 keV Background count distribution Resolution stability

24. The background level of the new detector was initially ~1×10-5 s-1 keV-1 cm-2 The count rate appeared to drop with time On the 4th of November the count rate increased suddenly and then started dropping again It was found that the Nitrogen flow was interrupted due to a refilling intervention by the cryo group Several days later the background reached an unexpected low level of ~2×10-7 s-1 keV-1 cm-2 implying ~0.05 counts/hour for the energy range 1-7 keV An intervention to face a noise problem (proven to be a faulty NIM power supply, because of power cuts) caused background increase  RADON Background of the new detector Attempt to seal shielding Power cuts Nitrogen interruption Nitrogen interruption installation noise

25. The intrinsic background level of the detector is minimal due to clean materials: only plexiglas, capton, copper for the detector andmylar, aluminum and stainless steel for the window Muons are efficiently rejected by the offline analysis External radiation is stopped by the shielding (4 cm archeological lead, Cd, 5mm Copper, external polyethylene + Nitrogen flow + detector plexiglas) Under these conditions the background is determined by the fluorescence lines (Copper, Iron and escape peaks) – the spectra during the different phases seem to be proportional These lines are mainly excited by radon radioactivity  extremely important to eliminate it from inside the shielding (for the moment tightness is achieved using aluminum tape!!!) The radon effect at a surface experiment Trigger rate during a 3-day run Count rate during a 3-day run. 5 counts!!!

26. CONCLUSIONS • The CAST experiment is taking data since 2003 resulting to a dominant experimental limit over almost all the theory favored range • During phase II CAST has entered the axion model region. In this phase detectors with low background level are critical • The micromegas detector has several advantages in rare event detection such as axion searches: • Low intrinsic background due to • Low radioactivity materials • Particle discrimination • Spatial resolution • Stability in long term runs • Good energy resolution • During ~6 year operation in CAST, micromegas has shown a low and stable background level, contributing significantly in the achieved results. Recent upgrades are leading to better performance and a dominant contribution to the future results is expected. • The latest detectors have shown extremely promising performance for phase II and the (possible) future CAST campagnes

27. OUTLOOK • The ultra-low-background level of the new detectors brings up new possibilities: • The micromegas has at least twice higher efficiency than the CCD + telescope system (which dominated phase I limit) and lower background level • Returning to phase I conditions for 6 months only (equivalent with 7 CCD-telescope systems) would easily lead to a coupling constant limit gaγγ < 4×10-11 GeV-1 • In parallel we are looking for of an advanced x-ray optics focusing device with enhanced efficiency • The implementation of such a device will lead to background reduction of more than a factor of 100. Depending on the optics transmission the coupling constant could be pushed down to gaγγ 1×10-11 GeV-1and any single count would be interpreted as a potential signal!!! (0.15 expected counts in 6 months) • The detector can be easily adapted (using proper windows) for low energy axion searches

28. END

29. ( ) The Strong CP Problem Solution of the U(1)A problem from t’Hooft (θ-vacuum) resulted in a CP-violating term in QCD lagrangian: Experimental consequence: prediction of electric dipole moment for the neutron: (A = 0.04 – 2.0) (90% CL) BUT experimental result... So, (QCD vacuum + EW quark mixing) The strong CP violating term could be suppressed easily! All needed is a massless quark! …no quark is massless…

30. Cleaning up a long standing problem in theoretical physics (Wilczek) The Peccei – Quinn solution Peccei-Quinn (1977) proposed an elegant solution to this problem: new global chiral U(1) symmetry (PQ) spontaneously broken at scale fα θ is not anymore a constant, but a field  the axion a(x) The AXION appears as the Nambu-Goldstone boson of the spontaneous breaking of the PQ symmetry (Weinberg and Wilczek (1978) PQ symmetry is not exact  axions acquire mass • 1st scenario: • fa∼ fEW → ma ∼ O(1MeV) • Visible axions, but… ruled out by experiments • 2nd scenario: • fa>> fEW → ma << O(1MeV) • Invisible axions, what CAST is looking for

31. Axion origin • Cosmological axions (cold dark matter) • Solar axions Stellar plasmas may be a powerful source of axions… The closest stellar plasma available is: the Sun Solar surface axion luminosity Axion flux on earth G.G. Raffelt, PRD33 (1986) 897 La = g102 1.85  10-3 L

32. Axion to photon conversion Axion to photon conversion probability: Vacuum: Γ=0, mγ=0 Coherence condition: qL< π L magnet’s length, q the momentum transfer CAST has ~100 times higher conversion probability than other helioscopes! For CAST phase I conditions (vacuum), coherence is lost for ma > 0.02 eV In the presence of a buffer gas inside the magnet bores the photon acquires an effective mass mγ > 0. The momentum transfer is now Coherence is restored for a narrow mass range:

33. Extending sensitivity to higher axion masses… • mγcan be adjusted by changing the gas pressure: or (T=1.8 K) • Every specific pressure of the gas allows the test of a specific axion mass. • The higher the pressure, the higher the photon effective mass, the higher the axion mass tested. • For every pressure setting there is a new discovery potential !!! For P~50 mbar Δma ~ 10-4eV !!!

34. CAST, Physics program • CAST Phase I:vacuum operation, completed(2003 - 2004) • CAST Phase II:4He run, completed(2005 – 2006) • 4He vapor pressure < 16.4 mbar • P<13.4 mbar, 160 steps, • 0.02 eV < ma < 0.39 eV • 3He run, commissioning inNov. 2007data taking startedinMar. 2008 • 3He vapor pressure < 135.6 mbar • P~120 mbar, ~1000 steps, • 0.39 eV <ma<~1.20 eV • Low energy axions(2007 – 2010) in parallel with the main program i.e. the X-ray range ~few eV range and 5 eV – 1 keV range • CAST NEXT ??

35. Micromegas Detector Two-region gaseous detector: Conversion region Primary ionization Charge drift Amplification region Charge multiplication Readout layout Strips (1/2 D) Pixels Separated by a Micromesh Very strong and uniform electric field Drift field typical 102-3 V/cm Amplification field typical 104-5 V/cm Amplification gap: 50-100 mm Mesh signal Pixels / strips signals γ Giomataris, Charpak (1996)

36. 4He gas system – filling the magnet bores (in operation in 2005 and 2006) Controlled injection of 4He in the bores Precise measurement of injected gas quantity (measuring volume in controlled P,T) Precise monitoring of gas P,T High reproducibility precision (< 0.01 mbar) No thermoacoustic oscillations Cold Windows installed to contain the gas in the bores Transparent to X-rays and visible light Resistance to magnet quenches Same detector setup During 2005/2006 we covered 160 4He pressure steps, reaching P = 13.43 mbar  ma~ 0.39 eV/c2 CAST Phase II (4He)

37. 3He Phase gas system • Storage region • Metering region • Axion conversion region • Expansion region (Recovery) Technical Design Report [CERN-SPSC-2006-029] • Requirements: • Avoid loss of 3He • Absence of TAO’s • Precision/Reproducibility • Multiple stepping / Ramping of density • Constraints: • Limited space • Weight restrictions • Custom components • Magnet movement • Instrumentation