Axion searches with the EDELWEISS Ge bolometers

C. Nones – CEA/DSM/IRFU/SPP on behalf of the EDELWEISS collaboration

1 Outline

 The EDELWEISS-II experiment in a nutshell –> Valentin Kozlov’s talk for more details

 The EDELWEISS Ge hybrid detectors

 Different axion sources

 Axion detection with EDW Ge detectors

 New results on different channels

 Conclusions E. Armengaud et al. – in preparation

2 EDELWEISS-II in a nutshell

EDELWEISS-II ~ 50 people / 4 countries

It is located at the Underground Laboratory of Modane.

Main goal: direct detection of WIMPs with Ge FRANCE ITALIE double-readout (heat + ionization) detectors 4800 mwe LSM (Fréjus)

 EDELWEISS experiment shows that this Altitudes 1228 m 1263 m 1298 m technology is mature for detection of Distances 0 m 6210 m 12 868 m WIMPs and rare events in general.

 EDELWEISS has obtained competitive results in the search for WIMPs, also in the controversial low-mass region.

 The low raw background achieved (including electron recoil signals) close to threshold makes EDELWEISS an interesting set-up to search for axions and ALPs.

3 The EDELWEISS hybrid detectors

1st generation – GeNTD 320 gr - All planar electrodes

 Simultaneous measurement • Heat @ 20 mK with Ge/NTD thermometer • Ionization @ few V/cm with Al electrodes  Evt by evt identification of the recoil

 Q=Eionization/Erecoil • Q=1 for electron recoils • Q≈0.3 for nuclear recoils

nd 2 generation - ID400 – used in this analysis - Keep the EDW-I NTD phonon detector - Modify the E field near the surfaces with interleaved electrodes: - Biases to have an electric field ~ horizontal near the surface and ~ vertical in the bulk MFid~40-50% - The rings are alternately connected by ultra-sonic bonded wires.

Φ 70mm, H 20mm, 410g → Easy cuts on « veto » + guard 14 concentric electrodes (width 100µm, electrodes define the fiducial zone4 spacing 2mm) without beveled edge. Active recoil discrimination

Realization of the concept in the Ionization signal amplitude R = EDELWEISS Phonon signal amplitude technology

Axions region

R set to 1 for ionization signals

WIMPs region (populated by a fast neutron source) Axions? Why not…

 One can use Ge detectors originally built for WIMP dark matter (or 0νββ)

 Beauty of the method: it is possible to piggyback on the existing Ge experiments

EDW is also sensitive to electronic recoils down to 2.5 keV 0.25 c/keV/kg/d @ 12 keV Excellent electron 0.8 keV FWHM recoil background at 2.5 keV threshold low energies (use

fiducial volume Rate[c/keV/kg/d] discrimination)

Axions can generate an electronic recoil: we can detect or constrain axions

Electron recoil energy [keV] 6 Axion searches – Different studied sources

−9 −1 −11 eff −7 1) The Sun gAγ = 10 GeV , gAe = 10 and gAN = 10 Bremsstrahlung Compton Axio RD 1013 • Primakoff production (gAγ) C+B+RD

) Primakoff

-1 14.4 keV 57 • Nuclear de-excitation ( Fe) (gAN) keV -1 s -2 1012

• Bremsstrahlung and cm Compton scattering (gAe) Flux ( • Axio-recombination and 1011 Axio-deexcitation (gAe) 0 2 4 6 8 10 12 14 16 18 20 New flux evaluation by J. Redondo Energy (keV)

We are sensitive to different axion-couplings to particles

2) The Galactic halo

Hypothesis: Axions make up all of galactic dark matter and they have a mass in the keV region. 7 Axion detection with Ge hybrid detectors

1) Axio-electric effect 2) Bragg’s diffraction

Incoming solar axion

Outgoing photon

Axio-electric effect: it is the equivalent of a photo-electric effect with the absorption of gAe an axion instead of a photon.

2 2 ⎛ 2 / 3 ⎞ gAe 3E β σAe (E) = σpe (E)⋅ ⋅ 2 ⋅ ⎜1 − ⎟ β 16παme ⎝ 3 ⎠ gAγ 8

€ Summary table: production & detection

Production Detection Signature

Primakoff channel Bragg diffraction Energy depencence

gAγ gAγ and time correlator

Compton, Bremsstrahlung, Axio-electric Energy dependence axio- effect (spectral bump) recombination&deexcitation gAe gAe Solar 57Fe de-excitation Axio-electric Energy dependence eff gAexgAN effect (line @ 14.4 keV) gAe Axions as dark matter Axio-electric Energy dependence

effect gAe (line @ mA)

9 Data selection: EDELWEISS-II

Axions can trigger an electronic recoil: event selection in the electron recoil Electron recoil band band

Use of the fiducial volume to discriminate surface events

Selection of periods with good baseline on heat and fiducial ionization: homogeneous data set of 450 kg d

Define “best” energy fiducial estimator by combining heat and fiducial ionization

Threshold definition: impose online trigger efficiency > 50% and other cut efficiency > 95%

3 detectors have a threshold @ 2.5 keV 2 @ 3 keV annd 5 @ 3.5 keV. FWHM at low energy is 0.8 keV. 10 An example: the 14.4 keV case (1) Production:14.4 keV monochromatic Detection: axio-electric effect axions emitted in the M1 transition of 57Fe nuclei R14.4 = Φ14.4 × σa × W × ε

We assume a Poisson background and construct a likelihood function €

Flux 3 eff 2 Extraction of a limit 2 -1 [cm *s ] Φ14.4 = β ⋅ Φ0 ⋅ (gAN ) 0.4 0.35

0.3

0.25

0.2

0.15

€ Rate (c/kg.d.keV) 0.1 R<0.038 c/kg/d 0.05

0 12 13 14 15 16 17 18 Fiducial recoil energy (keV) eff −17 gAe × gAN < 4.7⋅ 10 11

€ An example: the 14.4 keV case (2)

10-13 Model independent limit

10-14 eff

eff gAe × gAN vs. Axion mass AN 10-15 x g Ae g

10-16 € 10-17 0 2 4 6 8 10 12 14 -6 10 KSVZ (S=0.5) - EDW Axion mass [keV] 2 DFSZ (S=0.5; cos `DFSZ=1) - EDW |g | - KSVZ (E/N=0) -7 Ae 2 10 gAe DFSZ (cos `DFSZ=1)

Model dependent limit 10-8

-9 Limits on gAe assuming values predicted by 10 Ae

the DFSZ and KSVZ models for the g 0 3 10-10 couplings g AN and g AN. Assumptions: S = 0.5 for the flavor- -11 10 singlet axial vector matrix element in both 2 -12 models and cos βDFSZ = 1 for the DFSZ 10 model. 10-13 0.001 0.01 0.1 1 10 12 100 Axion mass [keV] Dark Matter axions

Hypothesis: Axion make up all of galactic dark matter and they have a mass in the keV region.

15 vA 9⋅ 10 keV The flux does not depend on any axion ΦDM = ρDM = −2 −1 β coupling. mA cm ⋅ s mA

GeV ≈ 0.3 -7 cm3 10

-8 Solar neutrinos € Axions are not relativistic: β ~10-3 10

€ 10-9 Signal: a line at the axion mass in

-10 the recoil spectrum 10Ae DAMA g

10-11 EDELWEISS −12 CoGeNT gAe <1.05 ×10 @ ma =12.5 keV 10-12 CDMS

10-13 1 10 102 € Axion mass (keV) 13 gAe limits with the EDELWEISS-II data

14 gAe limits: EDELWEISS-II & the others

---- model dependent limit RED curves – EDW results 10-6 Borexino

10-7 Derbin KSVZ Cuore R&D DFSZ 10-8

10-9 Derbin ae g 10-10 XMASS

Solar neutrinos 10-11 CoGent

CDMS 10-12 Red giant

10-13 10-5 10-4 10-3 10-2 10-1 1 10 102 103 104 105 Axion mass (keV) Competitive with astrophysical limit 15 gAγ: Primakoff effect

Production: Primakoff process in the Sun Detection: Bragg diffraction in Ge crystals

The typical transferred momentum has a wavelength close to the interatomic spacing A Bragg pattern arises -> strong enhancement of the signal Average energy of about 4.2 keV and a Incoming negligible intensity solar axion above 10 keV Outgoing photon

4 gAγ 16 The expected rate in 1 day

2 g 2 ˜ 3 V dΦ Aγ 2 1 0 ˜ R(E ,t,α) = 2(2π) 2 ∑ 2 sin(2θ) 2 S(G)FA (G) W (E A,E ) va G dE A 16π G

4 ⎛ g 108 ⎞ Aγ ⋅ ˜ ˜ = ⎜ −1 ⎟ ⋅ R (E ,t,α) ≡ λ⋅ R (E ,t,α) ⎝ GeV ⎠

€ Bragg condition Time and energy variation of the signal € =

€

−8 -1 gAγ =1⋅ 10 GeV 17

€ The statistical analysis

The vertical axis of the bolometer tower is aligned with the [001] axis of each detector (one degree precision) BUT the individual azimuthal orientation α of each detector was not measured. Axion rate averaged over time Build a time χ ∝ R(E,t) − R correlation function ∑ events ⎛ 8 ⎞ 4 gAγ ⋅ 10 Dependency on the azimuthal orientation λ = ⎜ −1 ⎟ ⎝ GeV ⎠ Combine all 10 detectors with unknown € orientations: € → MC simulation of the experiment` → Scan over orientations and possible couplings → Rule out coupling from the comparison between simulation and real data.

Energy analysis window: 3-8 keV

18 gAγ limits with the EDELWEISS-II data −9 -1 gAγ < 2.13⋅ 10 GeV

10-8

KSVZ E/N=0 € SOLAX and COSME CDMS )

-1 DAMA 90% CL EDELWEISS 10-9 GeV (

a

A Axion models g CAST

HB stars 10-10

10-3 10-2 10-1 1 10 102 Axion mass (eV) ⎛ b ⎞1 / 8 Room of improvement ? gAγ ∝ ⎜ ⎟ ⎝ Mt⎠ 19

€ Conclusions (1)

Bkg data from the EDELWEISS-II detectors, originally used for WIMP search, have been used to constrain the couplings of axion-like particles within different scenarios. Production Detection Coupling limits

-9 -1 Primakoff channel Bragg diffraction gAγ<2.13x10 GeV gAγ gAγ

-11 Compton, Bremsstrahlung, Axio-electric gAe<2.56x10 axio- effect recombination&deexcitation gAe gAe

57 eff -17 Solar Fe de-excitation Axio-electric gAexgAN <4.7x10 eff gAexgAN effect gAe

-12 Axions as dark matter Axio-electric gAe<1.05x10 effect gAe @ 12.5 keV

20 Conclusions (2)

EDELWEISS-II data: axion mass limits

Channel 57Fe C-B-RD Primakoff

KSVZ 154 eV

DFSZ 7.93 eV

EDW solar axion searches

21