The Axion Dark Matter experiment (ADMX) Phase 0

Transcription

1 3rd Joint ILIAS CERN DESY Axion WIMPs The Axion Dark Matter experiment (ADMX) Phase 0 Steve Asztalos, LLNL June, 2007

2 Collaboration ADMX is a five institution collaboration Lawrence Livermore National Laboratory Co-spokesperson (K. van Bibber) Staff scientist (D. Kinion) Post doc (G. Carosi), Principal Investigator (S. Asztalos) University of Washington Co-spokesperson (L. Rosenberg) Graduate student (M. Hotz), post doc (G. Rybka - summer 07) University of Florida D. Tanner and P. Sikivie Former graduate student (L. Duffy now at LANL) University of California Berkeley J. Clarke National Radio Astronomy Observatory R. Bradley

3 Collaboration expertise Lawrence Livermore National Laboratory Experimental site since 1995 Operations Phase I design SQUID development University of Washington Operations Analysis University of Florida Theoretical guidance Cryogenic engineering High resolution analysis University of California Berkeley Ground-breaking SQUID design National Radio Astronomy Observatory Low-noise amplifiers

4 The Axion

5 Axion-photon coupling Primakoff interaction g aγγ The axion, like the π 0, has a two-photon coupling The free-space (γγ) lifetime is irrelevantly long (τ ~10 50 sec) J π = 0 L~ E B But it can also be converted into a single real photon in EM field This photon then carries the total energy of the axion This Primakoff interaction is the basis for the most sensitive experiments to search for the axion

6 Properties of the Axion

7 TSP s* fine-tuning problem

8 TSP s hypothesis, and first ADMX unsuccessful experiment invisible pooltable straightening mechanism

9 Summary of past laboratory searches: ADMX A heavy axion is excluded For example: SLAC E137 (Bjorken et al.) lifetime of a γγ (sec) 20 GeV electrons earth shield axions produced here via Primakoff effect a γγ detector f PQ must be considerably greater than the weak scale

10 A key insight

11 A high-q ADMX search for relic oscillations

12 Completing the analogy f l PQ-symmetry breaking scale Pendulum length Quanta m a (ω) ~ f 1 ~ l 1/2 Couplings g ~ f 1 ~ l 1 Total energy Ω a (E) ~ f 7/6 ~ l

13 How to detect dark-matter axions (Sikivie, 1983) Superconducting magnet Ultra-low noise microwave receiver High-Q microwave cavity

14 The first-generation experiments RBF, UF 1980 s From W. Wuensch et al., Phys. Rev. D40 (1989) 3153 The first-generation experiments already came within a factor of of the desired sensitivity a stunning achievement Figure 2

15 The axion as dark matter candidate Cosmological abundance Local halo density Max. likelihood density to multicomponent Milky Way galaxy with all constraints: Rotation curve Virial velocity Projected areal disk density Microlensing optical depth Gates, E.J., G. Gyuk, M.S. Turner, Ap.J. Lett. 449 L123 (1995) 6μeV Ω a = ( ) m a 7/6 ρ halo = GeV /cm 3 The cavity search assumes that axions constitute some or all of the dark matter, but that is a soft assumption for a sufficiently light axion

16 The signal is the total energy of the axion The axion mass range is scanned by tuning the cavity Resonance condition: hν = m a c 2 [1 + O(β 2 ~10-6 )] There may be fine structure in the axion signal

17 Axion ADMX halo dark matter a unique quantum system Axionic dark matter is very dense Milky Way density: Thus if m a ~10μeV: ρ halo 450 MeV cm 3 ρ # cm 3 Axionic dark matter is highly coherent β virial 10 3 λ De Broglie 100 m Δβ flow 10 7 λ Coherence 1000 km The microwave cavity experiment measures the total energy of the axion, thus revealing both Doppler motion and coherence of the axion fluid

18 The ADMX parameter space is bounded KSVZ DFSZ

19 The ADMX parameter space

20 The ADMX radiometer eqn.* dictates the strategy But integration time limited to ~ 100 sec * Dicke, 1946 System noise temp. now T S = T + T N ~ K P sig ~ ( B 2 V Q cav )( g 2 m a ρ a ) ~ watts

21 Basic formulae Signal power: Scanning rate: Q L = Q 0 /(1+β) Loaded Q-value; β coupling δf = f - f 0 C lmn Offset from central Cavity form-factor Δf f step n Cavity bandwidth Frequency tuning steps Overlapping tuning steps Note both the power and scanning rate depend linearly on Q L

22 Rules-of-thumb for optimizing the experiment For scanning at a fixed coupling g aγγ 1 f df dt (B2 V ) 2 1 T S 2 For scanning at a fixed sweep rate g 1 B 2 V T S Ideally one wants sufficiently low temperature such that one can: (i) Be sensitive to the most pessimistic model axion (e.g. DFSZ ) (ii) Which only occupies a fraction of the halo density (e.g. 10% ) (iii) Finish the whole works in a tractable time (e.g.10 yrs )

23 ADMX Axion hardware

24 Axion hardware (cont d)

25 Microwave cavity basics (I) Required/desired features: Cover ~100 MHz to ~100 GHz Practical tuning, ± 50% High quality factor, Q ~ 10 5 High cavity form-factor, C = O(1) Minimal mode-crossings Minimal mode-localization Simplest right circular cavity, TM 010 : E z = J 0 (kr) (empty) f 0 = GHz / R[m] C 010 = 0.69 hν = mc 2 m a = μev. f[ghz]

26 Microwave cavity basics (II) Cavity form-factor C lmn (overlap of E, B ext ): Cavity quality, Q lmn : Q = O (Volume) (Surface Area) (Skin Depth) For uniform B = B 0 : C(TM 010 ) ~ 0.69 Much smaller for TM 0n0 TE, TEM identically 0 In high B-field, low-t: Must be copper (not SC!) Anomalous skin depth limit Try to use the TM 010 -like mode for all configurations Q limited to few 10 5, but we reach the theoretical max

27 Microwave cavity basics (III) Tuning Tuning rods, radial offset Mode-crossings E z for TM 010 mode; two metal rods half-way from center Metal - up; dielectric - down Keep longitudinal symmetry Keep cavity aspect ratio L/R low But can walk-around crossings

28 Multiple Cavity Ops 20 cm Piezo motors have low power dissipation, work at low temperatures and high magnetic fields

29 Microwave ADMX amplifiers

30 Heterojunction ADMX FET ( HEMTs ) & balanced design Donor layer (Al x Ga 1-x As) separate from gate layer (GaAs), thus eliminating impurity scatterers Electrons propagate ballistically across the 2D channel (0.25μ length, 300μ wide) Thus noise is very low Resonant cavities represent a complex frequency dependent input impedance Z 0 (w-w 0 ) Hybrid design minimizes input reflection, providing broad-band match to the complex cavity load There is a small penalty in noise

31 The ADMX world s quietest receiver by 10 4! P02552-ljr-u-018 We are systematics-limited for signals of W 10-3 of DFSZ axion power!

32 The ADMX axion receiver, and high-resolution search

33 The real receiver

34 Sample data and candidates

35 Limits on axion models and local axion halo ADMX density KSVZ ApJ Lett 571 (2002) 27 ρ Halo PRL 80 (1998) 2043 PRD 64 (2001) PRD 69 (2004) (R) Plausible models have been excluded at the halo density over an octave in mass range

36 Results of a high-resolution analysis PRL 95 (9) (2005) 2000 s 52 s Measured power in environmental (radio) peak same in both channels

37 What ADMX if the axion is found?