1. CAST: Recent Results & Future Outlook
XXIX Workshop on Recent Advances in Particle Physics and Cosmology
Patras, 14th-16th April 2011
CAST: Recent Results & Future Outlook
Thomas Papaevangelou
IRFU – CEA Saclay
for the
CAST Collaboration

2. Outline Axions CAST Results Detector Upgrades Near future outlook
Towards a NGAH
Conclusions

3. Axions & the strong CP problem
CP violation is allowed in strong interactions:
Such violation is not observed experimentally!
 the neutron EDM: dn < 3 × e cm  θ ≤ 10-10
why is θ so small?
 the strong-CP problem the only outstanding flaw in QCD
The most natural solution:
Peccei-Quinn introduced a global U(1)PQ symmetry broken at a scale fPQ,
Appearance of a new field : Axion field
 A new pseudoscalar particle the axion, with mass:
All the axion couplings are inversely proportional to fPQ2 ( fα2)

4. Axion summary L = g (E•B) a Axions couple to photons
Axion properties
neutral, practically stable,
very low mass,
very low interaction cross-sections
 Nearly invisible to ordinary matter
 excellent candidates for dark matter
Axions couple to photons
L = g (E•B) a
(PRIMAKOFF EFFECT)
QCD models:

5. Axion origin Cosmological axions (cold dark matter) Solar axions
Hot Stellar plasmas are a powerful source of axions…
The closest stellar plasma available is: the Sun
Solar surface axion luminosity
Axion flux on earth
Ea = 4.2 keV

6. Cosmological & Astrophysical observations  bounds on axion properties
Άξιον εστί
Discovery of an axion:
New elementary particle
Solution of strong CP problem
Dark matter particle
New solar physics
 answer to solar mysteries, e.g.:
- Solar corona heating problem,
- Flares
- Unexpected solar X-rays, …
Cosmological & Astrophysical observations  bounds on axion properties
Sun lifetime.
Stellar evolution. Globular clusters. 𝒈 𝒂𝜸 ≤ 𝟏𝟎 −𝟏𝟎 𝑮𝒆 𝑽 −𝟏
White dwarf cooling. Axion would affect rotation frequency evolution.
Supernova physics. Effect on neutrino burst duration.
Overclosure. Too much DM, 𝒎 𝒂 ≥𝟏𝝁𝒆𝑽/ 𝒄 𝟐

7. Experimental axion searches
 search for axion – photon interaction (coupling constant gaγ & axion mass ma)
Laser experiments (laboratory):
Photon regeneration (“invisible light shining through wall”)
Polarization experiments (PVLAS)
Search for dark matter axions:
Microwave cavity experiments (ADMX)
Search for solar axions:
Bragg + crystal (SOLAX, COSME, DAMA)
Helioscopes (SUMIKO, CAST)

8. CAST: axion helioscope “a la Sikivie”
Axions would be produced in the Sun’s core and re-converted to x-rays inside an intense magnetic field. P. Sikivie, Phys. Rev. Lett. 51, 1415–1417 (1983)
CAST is using a prototype superconducting LHC dipole magnet able to track the Sun for about 1.5 hours during Sunrise and Sunset.
Operation at T=1.8 K, I=13,000A, B=9T, L=9.26m
Expected signal
X-Ray excess during tracking
at 1-10 keV region
CAST sensitivity per detector
0.3 counts/hour for
gaγγ= GeV-1 and A = 14.5 cm2

9. CAST: a helioscope “a la Sikivie”
Axions would be produced in the Sun’s core and re-converted to x-rays inside an intense magnetic field. P. Sikivie, Phys. Rev. Lett. 51, 1415–1417 (1983)
CAST = a difficult experiment:
- superconducting ( quenches!)
- the only(!?) telescope at 1.8K
- moving / alignment
- Cryo Fluid Dynamics of buffer gas  tracking
- low background X-ray detectors
CAST is using a prototype superconducting LHC dipole magnet able to track the Sun for about 1.5 hours during Sunrise and Sunset.
Operation at T=1.8 K, I=13,000A, B=9T, L=9.26m
Expected signal
X-Ray excess during tracking
at 1-10 keV region
Low X-ray background detectors required to observe a signal.
CAST sensitivity per detector
0.3 counts/hour for
gaγγ= GeV-1 and A = 14.5 cm2

10. The 1st Solar Cycle of CAST!!!
Proposal approved by CERN (13th April 2000)
Commissioning (2002)
CAST Phase I: vacuum operation ( ) completed
CAST Phase II: (2005–2011)
4He run, (2005–2006) completed
0.02 eV < ma < 0.39 eV
3He run ( ) ongoing data taking
0.39 eV <ma<~1.20 eV
Low energy axions (2007 – 2011) in parallel with the main program
~ few eV range
5th April 2011: Proposal to SPSC for Start of 2nd solar cycle (!?)
13th April 2011: CAST completed one solar cycle

11. CAST published results
For ma < 0.02 eV:
gαγ < 0.88 × GeV-1
JCAP04(2007)010, CAST Collaboration
PRL (2005) 94, , CAST Collaboration
For ma < 0.39 eV typical upper limit:
gαγ < 2.2 × GeV-1
JCAP 0902:008,2009, CAST Collaboration
CAST byproducts:
High Energy Axions: Data taking with a HE calorimeter (JCAP 1003:032,2010)
14.4 keV Axions: TPC data (JCAP 0912:002,2009)
Low Energy (visible) Axions: Data taking with a PMT/APD. (arXiv: )
CAST experimental limit dominates in most of the favored (cosmology/astrophysics) parameter space

12. Annual Workshops

13. Extending sensitivity to higher axion masses…
Axion to photon conversion probability:
Vacuum:
Γ=0, mγ=0
Coherence condition: qL < π ,
For CAST phase I conditions (vacuum), coherence is lost for ma > 0.02 eV.
With the presence of a buffer gas it can be restored for a narrow mass range:
with
New discovery potential for each density (pressure) setting
For P~50 mbar Δma ~ 10-3 eV

14. Working with a buffer gas…
Precise knowledge and reproducibility of each pressure setting is essential
Gas density homogeneity along the magnet bore during tracking is critical
Superfluid 1.8 K guaranties temperature stability along the cold bore
Hydrostatic pressure effects are not critical
However:
There are parts of the magnet bores, outside the magnetic field region, that are in higher temperatures. These temperatures can vary during tracking and may depend on the buffer gas density
There is also a volume of pipe work in high temperature directly connected with the cold bores
To face that situation we:
Measure precisely the amount of gas injected into the magnet with a metering volume kept at stable temperature (typically 36 oC)
During 4He phase we kept the cold windows that confine the gas at T=120K, making the effective “dead volume” negligible
However at higher densities (3He phase) heat conduction towards the magnet does not allow high window temperatures. Dead volumes become significant and more effects appear (gas convection) . Comprehending the gas behavior is critical for the data analysis!

15. Understanding 3He dynamics
A number of additional temperature and pressure sensors have been placed in several points of the magnet and the gas system
 Input to Computational Fluid Dynamics simulations
in CAST temperature and density conditions 3He is not an ideal gas (Van der Waals forces)
convergence between simulation results & experimental data
Knowledge of gas density / setting reproducibility possible
Gas density stable along magnet bore
Coherence length slowly decreases with increasing density
No discrete pressure settings but continuous coverage

16. Preliminary - 3He result
Density variation during a tracking  4He phase analysis not easy to be used.
new formulation of the unbinned likelihood:
1st term: expected number of axions. Depends on exposure time
2nd term: Depends on the gas density at the moment a count occurred
Preliminary
Preliminary

17. Preliminary - 3He result
Density variation during a tracking  4He phase analysis not easy to be used.
new formulation of the unbinned likelihood:
1st term: expected number of axions. Depends on exposure time
2nd term: Depends on the gas density at the moment a count occurred
Exceed expectations!
Detector improvements
Preliminary
Preliminary

18. CAST detectors (phase I & phase II-4He)
Sunrise side
CCD + X-Ray Telescope (prototype for the ABRIXAS Space mission)
The X-Ray Telescope is focusing a 43 mm x-ray beam to 3mm
S/B improvement by ~150
Sunset side
Unshielded Micromegas
Shielded TPC,
covering both magnet bores
Typical rates
TPC
85 counts/h (2-12 keV) MM
25 counts/h (2-10 keV)
CCD
0.18 counts/h (1-7 keV)

19. Micromegas evolution
New manufacturing technique:
Microbulk Micromegas.
High radio-purity materials (kapton, Cu & Plexiglas)
Potential for ultra-low background rates.
: Replacement of TPC and sunrise Micromegas by new technology Micromegas Implementation of shielding
Unshielded Micromegas
(classic technology)
Sunset side
Sunrise side
Typical new MM rate: ~2 c/h
Shielded Micromegas
(bulk and microbulk technology)
Nominal values during data taking periods
Special shielding conditions
in Canfranc Underground Lab

20. < 2·10-7 counts keV-1 cm-2 s-1 [2-7 keV] (~1 count/day)
Canfranc
1 week with 109Cd source
0.2 Hz trigger rate
0.005 Hz trigger rate
First approach to final background limited only by intrinsic radioactivity (from microbulk, chamber materials, inner shielding):
< 2·10-7 counts keV-1 cm-2 s-1 [2-7 keV] (~1 count/day)
This result proves that background levels >20 times lower that current CAST MM nominal background are possible via shielding improvement.

21. Background level in CAST similar to Canfranc:
Shielding CERN
Partial front
shielding
Full front
ULB MMs: ongoing measurements
Background level in CAST similar to Canfranc:
(2.75  0.5)x10-7 s-1 cm-2 keV-1 or 1-2 counts/day @ 2-7 keV

22. Design of new Micromegas
Main features
Improved shielding
Radiopurity of materials
Upgraded electronics: include time information for each strip detector becomes TPC
Present design
Radon is avoided with a leak-tight vessel of lead, which also works as shielding.
Stainless steel and plexiglass have been replaced by copper and peek.
A new drift cage guarantees an uniform drift field.
Design being finalised with input from simulation, Canfranc tests and measurements at CAST.

23. Porosity 12-50%, hole  35-200 nm, thickness ~50 μm
Operating in the sub-kev range
Low threshold
Transparent Windows
Argon: good absorption efficiency in the energy range 1-7 keV
Neon: higher gain (> factor 10)  Single electron detection!!!
Good absorption up to 3 keV
Mixture of Ar-Ne can be used to cover all range keV
Nanotube Porous aluminum membrane  fraction of incident photons can be transmitted either directly or “channeled” through the pores
Small gas diffusion
Porosity 12-50%, hole  nm, thickness ~50 μm
Spectrum taken with MM & UV lamp 5eV
Ne absorption efficiency
CCD: windowless!
E [eV]
Prototype
N. Meidinger, private communication (2011)

24. 101st Meeting of the CERN / SPSC
Motivated by new physics cases
Supported by detector improvements
101st Meeting of the CERN / SPSC
CAST Physics Proposal to SPSC
K. Zioutas on behalf of CAST
and
in collaboration with
D. Anastassopoulos, O. Baker, M. Betz, P. Brax, F. Caspers, J. Jaeckel,
Lindner, Y. Semertzidis, N. Spiliopoulos, S. Troitsky, A. Vradis. 
CERN
5th April 2011 

25. Revisiting vacuum phase
12 months of data taking in Phase I conditions (vacuum) with the new detectors  CAST sensitivity well below astrophysical limits (~5×10-11 GeV-1)
CAST phase I limit determined by X-Ray telescope
Micromegas developments  3 additional high performance detectors
12 calendar months data taking with existing MM
calendars months data
taking ULB MM
(~10-7 cts /keV/cm2/s)
KSVZ
in parallel with paraphoton & chameleon runs
preparation of new detectors for a future experiment

26. Repeat 4He runs with new detectors
existing MM
KSVZ
ULB MM (10-7 cts/keV/cm2/s )
KSVZ
existing MM
4 months
8 months
12 months
16 months
mbar: 6, 12 & 18 calendar months
(1.5, 3 and 4.5 trackings/step)
ULB MM (10-7 cts/keV/cm2/s)
KSVZ
6 months
12 months
18 months
significant improvement in background wrt. 2006
crossing axion KSVZ model
could start in autumn (with present detectors)
no competition in sight

27. Solar Paraphotons Hidden Sector particles  Theoretically motivated
kinetic mixing: γ ↔ γ’ oscillations
 NO magnetic field!  NO cold bores needed
 Vacuum path length relevant for oscillations
 upstream in front of the detector
 a good sensitivity requires: 3 ULB MMs & FS pnCCD
& solar tracking!!!
GRID measurements
Sun filming Sep. 2010

28. Off-pointing solar tracking
Solar Paraphotons
Converted from thermal photons inside the Sun due to kinetic mixing
Region of interest for CAST < m < 10 eV
Resonant production in a shell inside sun for m > eV
Clear signature …….. ring images r/Ro > 0.7
CAST XRT/CCD - our solar imaging system
resonance:
thin slice brighter
no resonance:
central part brighter
Off-pointing solar tracking
Normal solar tracking

29. Cast Paraphoton sensitivity
New Vacuum Run (2x106sec): normal tracking, XRT FOV strongly limits sensitivity.  Off-pointing
 Could be used to look for structure with good ε over reduced region
Ideal XRT
3 MM
More realistic estimate of XRT/CCD
( assume only 60% of sun visible via XRT)
1.0 – 7 keV 2years, 0.3 – 7 keV, 0.1 – 7 keV
ULB & LET MM ( keV)
ε = 15%
14m (black) & 1m (purple)

30. Solar Chameleons
Chameleons are DE candidates to explain the acceleration of the Universe
Chameleon particles can be created by the Primakoff effect in a strong magnetic field. This can happen in the Sun.
The chameleons created inside the sun eventually reach earth where they are energetic enough to penetrate the CAST experiment. Like axions, they can then be back-converted to X-ray photons.
In vacuum, CAST observations lead to stronger constraints on the chameleon coupling to photons than previous experiments.
When gas is present in the CAST pipe, the analogue spectrum of regenerated photons shows characteristic oscillations: ID
 axion helioscope = chameleon helioscope, LE!!
 Low energy threshold: MM + CCD!
+ vacuum
P. Brax, K. Zioutas PRD (2010)

31. CAST sensitivity to Chameleons
The mass of the chameleon in the CAST pipe with vacuum is: mch = 40 μeV
[keV]
vacuum
0.1 mbar
The analogue spectrum [h-1 keV-1] of regenerated photons as predicted to be seen by CAST: matter coupling = 106, B=30T in a shell of width 0.1Rsolar around the tachocline (~0.7Rsolar).
1 mbar
[keV]

32. Towards a new relic axion antenna?
In addition…
Search in the visible:
a, γ’, CH, ...
BaRBE & Transition Edge Sensor (TES)
Towards a new relic axion antenna?
Dielectric waveguide inside the CAST magnet may perform as a new kind of “macroscopic fiber”, being a sensitive detector for relic axions:
 ~ meV rest mass range (experimentally inaccessible)
 Feasibility study of the proposed concept is in progress
>>> theoretical estimates!!

33. Exploiting QCD axion models
> 50 M€ project

34. Axion helioscopes FOM
3 parts drive the sensitivity of an axion helioscope
X-RAY DETECTORS
X-RAY
OPTICS
MAGNET

35. Cross section of the magnet
L. Walckiers’/CERN
suggestion
New magnet “a la ATLAS”
CAST enjoys one of the best existing magnets than one can “recycle” for axion physics
- LHC test magnet
Only way to make a step further is to built a new magnet, specially conceived for this.
Work ongoing, but best option up to know is a toroidal configuration:
Much bigger aperture than CAST: ~0.5-1 m / bore
Lighter than a dipole (no iron yoke)
Bores at room temperature
Cross section of the magnet
ATLAS

36. X-Ray optics Light Path in XMM-Newton Telescope
Large magnet aperture  X-ray focusing device
Background reduction
Detector simplification
XMM mirror module
Collecting area:
~1900cm2 (<150 eV), ~1500cm2 2 keV),
~900cm2 7keV), ~350 cm2 10 keV).

37. Prospects of a NGAH
1990
~2020

38. Goals for the next 1 – 2 years
Magnet
Built a new “magnet”, tailored to our needs
Main goal: B2L2A ~ x1000 better than CAST (desirable), x100 (minimum)
Other construction technical issues  feasibility study, design study.
Work in progress
Optics
Cost-effective large optics (all magnet instrumented)  m2
Detectors
Main goal: background ~ 10-7 keV-1 cm-2 s-1 >>> already reached!?
Platform, general assembly engineering
40-50% Sun coverage?

39. Conclusions
CAST, during its 11 years of existence:
has put the strictest experimental limit on axion searches for a wide ma range
is currently testing axion masses inside the region favored by QCD models
New searches for WISPs can start by 2012
CAST Collaboration has gained much experience on Helioscope Axion Searches
Detector development and research on superconducting magnets can lead to more sensitive helioscopes
In combination with Microwave Cavity experiments (ADMX), a big part of QCD favored model region can be swept up to 2020

40. Thank you!!!

41. Transition Edge Sensor (TES) working principle
TES working principle schematic
Incident energy heats an absorber (the choice of absorber material sets the spectral range of the sensor)
Thermometer: A thin film (e.g., Ir or Ti) at the transition temperature between normally and super- conducting measures the temperature change of the absorber
The thin film (the actual TES) is biased by a voltage: a change in its resistance is sensed as a change in current with amplitude proportional to the energy deposited in the absorber
At the end of an event the absorber slowly thermalizes towards an heat- sink
The background noise is virtually zero since at the operating temperature, mK, there are no “internal” heat sources
View of two sample sensors: the small (50 µm)x(50 µm) square is the actual TES, and the larger two rectangles are Al contact pads.
Output pulse sample from D. Bagliani et al, 
J Low Temp Phys. 151 (2008) 234

42. Relic Axion searches Axion antenna Radiometer PLANCK satellite
Axion  Microwave photon conversion (Primakoff effect) in one bore of CAST magnet
Dielectric waveguide collects and concentrates converted microwave photons
 mono mode operation
Radiometer
The antenna outputs mostly thermal background noise (T=2K) + a faint additional signal from converted Axions.
Radiometer = device to accurately measure small differences in noise power
Do we see a difference in noise power with the magnetic field switched on and off?
PLANCK satellite
- 30 GHz, Bandwidth = 8 GHz, Detector temperature = 20 K, Detector sensitivity = ± 20 μK

43. Filming Twice per year (March/September) direct optical check
A camera on top of the magnet aligned with the bore axis
Corrections for visible light refraction are taken into account
The Magnet Pointing precision is well within our requirements (0.5 mrad)
Filming
CAST Status & Perspectives, Theopisti Dafni (UNIZAR), Moriond2010

44. Tracking system precision
Sun Filming
Several yearly checks cross-check that the magnet is following the Sun with the required precision
Twice a year (March – September)
Direct optical check. Corrected for
optical refraction
Verify that the dynamic Magnet Pointing precision (~ 1 arcmin) is within our aceptance
GRID Measurements
Horizontal and Vertical encoders define the magnet orientation
Correlation between H/V encoders has been established for a number of points (GRID points)
Periodically checked with geometer measurements

45. FOV Problem: XRT r/Ro > 0.6 suppressed due to limited FOV @ LE?
Off-pointing solar tracking
FOV
Normal solar tracking
 Under investigation
 CCD size 
(1cm x 3cm)
Vignetting due to cold bore and large off-axis angles in XRT (1.5 keV)

46. Solar Chameleons - CAST
The mass of the chameleon in eV
in the CAST pipe with vacuum is:
mch = 40 μeV
[keV]
vacuum
0.1 mbar
The analogue spectrum [/hour/keV] of regenerated photons as predicted to be seen by CAST: matter coupling = 106, B=30T in a shell of width 0.1Rsolar around the tachocline (~0.7Rsolar).
1 mbar
[keV]
 Low energy threshold: MM + CCD!
+ vacuum

47. ? Solar chameleons: simulated exclusion plots 0.1mbar 1mbar 10mbar
 log mCH [eV]
Assumption: B=30T in a shell of 0.1Rsolar around ~0.7Rsolar. The analog spectrum has been cut below a mass of ~55eV due to Pb in front of the CAST pipes.
Note: mass = derived quantity!

48. Next generation Helioscope
Coupling constant dependence:
Detector improvements
Dependence  8th root but big margins:
Ultra-Low-Background periods (>1 week each) have been observed with 4 different detectors
Not fully understood yet but:
Ongoing simulations by Zaragoza group
Ongoing background measurements in a controlled environment (Canfranc)
Optimized detector design
New X-ray optics feasible
Cover big aperture
High efficiency (>50%)

49. Next generation Helioscope
Coupling constant dependence:
Magnet improvements
Strong dependence with B,L but small improvement margins for next decade
Increase of tracking time helps but big technical difficulties
Magnet bore’s aperture!
Meetings with CERN magnet experts  alternative configurations
In an axion Helioscope field homogeneity is not as important as in accelerators
An “ATLAS – like” configuration proposed by S. Russenschuck and L.Walckiers is the most promising
Big aperture (~1 m) / multiple bores seem possible
Big magnetic field possible (new superconductive material)
Lighter construction than a dipole
Feasibility studies have started

51. “Politics” NGAH’s 50-100MEUROs fundable? WIMPs vs. Axions
New generation of WIMP dark matter experiments (1+ ton of detector mass): EURECA, XENON1T, DARWIN, in Europe, SuperCDMS, LUX, MAX, etc… in US, are MEUROs projects. NGAH physics case is not inferior to theirs. Rather the opposite. >>> Unique?
WIMPs vs. Axions
WIMP lobby if far larger, but scientifically -> no reason to disregard the “axion option” for DM
NGAH is the next step after CAST.
Helioscope vs. microwave
ADMX is sensitive to QCD axions!
In the search for QCD axions, NGAH, ADMX, Sumico and CAST) can close the gap in gaγγ-m.