New micromegas for axion searches in CAST
T. Dafni
a,, S. Aune
b, G. Fanourakis
c, E. Ferrer-Ribas
b, J. Gala´n
a, A. Gardikiotis
d, T. Geralis
c,
I. Giomataris
b, H. Go´mez
a, F.J. Iguaz
a, I.G. Irastorza
a, G. Luzo´n
a, J. Morales
a,1, T. Papaevangelou
b,
A. Rodrı´guez
a, J. Ruz
a,2, A. Toma´s
a, T. Vafeiadis
e, S.C. Yildiz
fa
Laboratorio de Fı´sica Nuclear y Astropartı´culas, University of Zaragoza, Zaragoza, Spain bIRFU, Centre d’ E´tudes de Saclay, CEA, Gif-sur-Yvette, France
cInstitute of Nuclear Physics, NCSR Demokritos, Athens, Greece d
University of Patras, Patras, Greece e
Aristotle University of Thessaloniki, Thessaloniki, Greece f
Do ˘gus- University, Istanbul, Turkey
a r t i c l e
i n f o
Available online 7 July 2010 Keywords:
Gas micropattern detectors Micromegas X-ray detectors Cast Axions Low background
a b s t r a c t
Micromegas detectors have been taking data in the CAST experiment since 2002, occupying one opening (out of the two looking for sunrise axions) of the magnet and showing good performance and stability. Currently, three of the four X-ray detectors used in the experiment are micromegas. The new detectors are of the Microbulk technology, which have attracted a lot of attention because of the advantages they present, among them the low-material construction, high radiopurity and good energy resolution. Here, their performance during the last year will be commented. In particular, the low background levels reached in some detectors have triggered a set of studies in order to understand the effect.
&2010 Elsevier B.V. All rights reserved.
1. Cast
Axions are hypothetical particles which could explain the strong CP problem. They could be produced in the core of stars like the Sun via the Primakoff effect. The CAST experiment (CERN Axion Solar Telescope) has been looking for solar axions since 2002 [1]. It is using a decommissioned LHC-prototype magnet which is 10 m long and can reach a 9 T magnetic field as a converter of axions coming from the Sun into detectable X-rays (Fig. 1). The energy range of the expected signal is between 1 and 10 keV, and its rate is dependent on the very weak axion–photon coupling (Fig. 2). Therefore, low background X-ray detectors are necessary in order to have a high sensitivity. The requirements for such detectors include radiopure components, shielding, good energy and spatial resolution to perform powerful off-line rejection conditions and reduce backgrounds, as well as stability over long periods of operation.
In the first phase of the experiment, the magnet bores were kept in vacuum, allowing CAST to explore the axion mass range up
to 0.02 eV/c2[2,3]. The three different detectors of the experiment at the time (an X-ray telescope coupled to a CCD[4], a Time Projection Chamber [5] and a micromegas [6]) registered no signal of axions, and for this range of masses, CAST has given the most stringent limit on the axion-to-photon coupling constant. In order to extend the search to a wider range of axion rest masses, data have also been acquired by introducing inside the magnet bores a buffer gas (first4He and then 3He). Using 4He allowed CAST in the years 2005–2006 to scan axion masses up to 0.4 eV giving an improved limit for the new scanned region[7,8]. In 2007 the experiment was upgraded to the second part of Phase II, where3He was to be used as a buffer gas inside the magnet bores. This change gave the opportunity for several upgrades regarding the detector systems. On the sunset end, a new system was designed for the installation of two new micromegas detectors (Fig. 3). These detectors replaced the Time Projection Chamber that was sitting on that end since 2002. The decision on the replacement of the TPC was taken considering the very good results obtained with the micromegas detector through the CAST data-taking period. On the sunrise end, the system was re-designed in order to implement a new shielding of the detector as well as new controls for several monitoring parameters [9]. The new design has taken into account the possibility to install in the future an X-ray focusing device in front of the detector, which would increase the signal-to-background ratio significantly.
Contents lists available atScienceDirect
journal homepage:www.elsevier.com/locate/nima
Nuclear Instruments and Methods in
Physics Research A
0168-9002/$ - see front matter & 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.nima.2010.06.310
Corresponding author.
E-mail address: Theopisti.Dafni@cern.ch (T. Dafni). 1
Deceased. 2
Present address: CERN, European Organization for Particle Physics and Nuclear Research.
In 2008 CAST started the data-taking with3He, which should enable CAST to probe the masses up to approximately 1 eV/c2.
2. Microbulk: the new micromegas detectors in CAST The micromegas group had invested in efforts to improve the detectors’ performance starting with the construction technique. Two new types of detectors were built using the Bulk and Microbulk technologies. The improvement they presented with respect to the conventional design [10] is that the mesh is no longer suspended above the anode mechanically as in the previous generation but either glued above the active area (Bulk)
[11]or built together with the active area during the construction process (Microbulk) [12]. In this way, they provide better robustness, uniformity and therefore stability of operation. Detectors of both types were constructed, tested and installed in the experiment, achieving very good background rejection factors. In particular the Microbulk micromegas presents advantages of low intrinsic radioactivity and good energy and spatial resolution. The micromegas group decided to use the latest generation, Microbulks, for CAST.
Different types of Microbulks have been manufactured, with different gap sizes, hole diameters and pitches. The ones produced for CAST have typically an amplification gap of 50
m
m, hole diameters of 30m
m, a pitch of 550m
m and a total active area of 36 cm2. They were the first to be constructed with a 2d readout, where the anode consists of square pads interconnected in the diagonal direction through vias (seeFig. 4). More details on their construction can be found in Ref.[12]. The chamber is made out of copper, kapton, plexiglas and aluminum, all being very clean materials from the radiopurity point of view.The shielding used in CAST consists of 2.5 cm of archaeological Pb, surrounded by polyethylene of a thickness of up to 25 cm (limited by space constraints). Between the lead and the Axion energy [keV]
Axion flux [10 10cm -2s -1keV -1] 0 1 2 3 4 5 6 7 0 2 4 6 8 10 12 14
Fig. 2. The Solar axion flux as expected on earth.
Fig. 3. Photographs showing the upgrade with the new sunrise micromegas line (a) and the new micromegas installed on the sunset side (b).
Fig. 4. (a) A photo of a Microbulk detector. (b) Photo of the 2d-readout plane, taken with a microscope, where the square pads of 400mm can be distinguished. Fig. 1. A photograph of the CAST experiment.
polyethylene there is a 2 mm thick Cd foil to absorb thermal neutrons. Inside the lead there is a 5 mm thick Cu layer, also serving as Faraday cage. The front-end electronics are placed outside the shielding. The interior of the shielding is flushed with nitrogen in order to expel radon from the environment. The remaining background seen by the detectors are muons, gammas and neutrons from the environment, which constitute a rate of approximately 1 Hz. The Microbulk detectors, apart from the lower radioactivity, present a better energy resolution and therefore discrimination level than conventional detectors. Another crucial issue is the stability of the detector over time because in such low-background experiments statistics is accu-mulated over long periods. In this point again the Microbulk detectors are more stable than traditional micromegas because of the fact that the mesh is attached to the anode and therefore the amplification gap is very well defined. During the 2009 data-taking the detectors have shown good stability and good energy resolution, as can be seen fromFigs. 5 and 6.
2.1. Ultra-low background periods
As mentioned above, background reduction in micromegas detectors is achieved by building the detectors using low-radioactivity materials and by optimizing the shielding, but also by using the signal information provided by the readout for
rejecting non-X-ray events. Two signals are extracted from the detector, the charge collected on the strips/pixels and the one induced in the mesh by the ions. With the help of regular calibrations with a 55Fe source the pattern of the X-ray and background events is studied and the background-rejection criteria defined. In the case of the strips, characteristics like the number of strips forming a cluster, the number of clusters or the total energy collected are examined. The mesh pulse allows for pulse-shape analysis and the combination of the two allows to reach high levels of background-rejection efficiency. With the help of these tools, in CAST these detectors have achieved background rates as low as 4 10 6counts keV 1s 2cm 2 over
long periods of operation, a background level already lower than the best achieved previously.
Beyond this nominal operation, there have been some periods where the background has been reduced even down to a level of 2 10 7s 1cm 2keV 1
. This reduction has been observed in different Bulk and Microbulk detectors at different periods of data-taking, but so far only on the sunrise side. An example of these periods, with a Microbulk detector on the sunrise end can be seen inFig. 7. The possibility of unexpected systematic effects
0 100 200 300 400 500 600 700 800 900
12/Sep 26/Sep 10/Oct 24/Oct 07/Nov 21/Nov 05/Dec
ADC Gain [mV]
Date
Strips Gain Amplitude gain
Fig. 5. An example of the gain history for one of the sunset Microbulks. One can see that twice there were changes in the operating voltage.
X [mm] Y [mm] 0 10 20 30 40 50 0 50 100 150 200 250 300 350 Entries 3923221 Integral 3.923e+006 Calibration_X_Y_plot_15015-16674
Total Energy [keV]
dN/dE [s -1 cm -2 keV -1] 0 0.2 0.4 0.6 0.8 1 1.2
Calibration 55Fe Strips
16% (FWHM)
0 2 4 6 8 10 0 10 20 30 40 50
Fig. 6. (a) An example of the energy resolution of the sunrise detector during data taking in 2009, of the order of 16% FWHM for the 5.9 keV line of the55Fe source. (b) Two-dimensional plot of the charge collected by the strips. One can see that there are two strips (out of the 106 in each direction) missing from the readout.
Date 16/06/09 02:00:00 16/07/09 02:00:00 15/08/09 02:00:00 14/09/09 02:00:00 14/10/09 02:00:00 13/11/09 01:00:00 13/12/09 01:00:00 [counts / hour] 0 0.5 1 1.5 2 2.5 3 3.5 4
Fig. 7. An example of the rate history of the sunrise Microbulk where two periods of very low background can be seen.
is reduced by the fact that calibration runs are taken daily between two consecutive background runs and were acquired at different times during the day. Although not conclusive, and while
the reasons for this reduction on the background level are under investigation, it is thought that it could be partially linked to variations in the radon concentration inside the shielding.
Fig. 8. A picture of the duplicated setup in the Zaragoza laboratory where the detector can be seen mounted inside the copper and lead shielding and the electronics housed in a Faraday cage. In (b) the Faraday cage surrounded by the polyethylene shielding.
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 ADC signal [mV] Time [ns] RC parameters wC = 8.15 MHz wL = 2.41 MHz Data PilUp Fit PilUp Event 1 Event 2 -300 -250 -200 -150 -100 -50 0 0 1000 2000 3000 4000 Generator gaussian amplitude [mV] Mean Time : 1084.7 ns Sigma : 31.9 ns Amplitude : 3417.1 "mV" Mean Time : 758.5 ns Sigma : 34.3 ns Amplitude : 3074.6 "mV" f(t) g(t)
3. How low can it go?
There are still puzzling aspects on the periods of very low background levels that have been observed on the sunrise side. The micromegas group working for CAST is devoting a substantial effort to understand and quantify these levels. Two complemen-tary working lines have been followed, focusing on simulations and experimental measurements.
3.1. Experimental measurements
A first area of study focuses on experimental background measurements in an environment with very well controlled parameters. A new acquisition system is operative in Zaragoza together with a Microbulk micromegas identical to those that are giving very low background levels in the CAST zone. The setup is housed inside a Faraday cage, with an inner shielding of copper and lead and an outer shielding of polyethylene, and comple-mented by a nitrogen flux in the vicinity of the chamber, imitating the setup of the CAST detectors (Fig. 8). A computer-controlled calibration system has been installed which interrupts the continuous background-taking periodically for the calibrations to take place. A minimal slow control system registers continuously the detector’s pressure and temperature.
Thus equipped, the set-up will systematically explore the various parameters and configurations (including installation in the Canfranc Underground Laboratory in the Spanish Pyrenees) in search of a correlation with backgrounds levels.
3.2. Simulations
The second strategy consists in a simulation of the detector. Geant4 simulations are being prepared with a detailed description of the geometry of the detector and its shielding (Fig. 9(a)). These simulations will be executed in different configurations including information about radiopurity measurements that have been performed at the Canfranc Underground Laboratory, using the materials that compose the detector. The simulations are being developed implementing diffusion as well as the electronic readout that is actually used in CAST (Fig. 9(b)). An energy calibration with a 55Fe source was first simulated: the photoelectric interaction of 5.9 and 6.5 keV photons and their escape peaks are quite well reproduced, as shown inFig. 10(a). The two peaks (as well as the corresponding escape peaks) are indistinguishable after including the energy resolution of the
detector. This spectrum is compared to the corresponding experimental spectrum recorded in the Zaragoza set-up in
Fig. 10(b) and shows good agreement.
The results obtained from these simulations will allow us to determine the relative reduction due to the actual shielding configuration and to study possibilities for optimization in the future.
4. Conclusions
In the 2009 data-taking period, for the first time three new micromegas detectors, built with the Microbulk technology, were taking data in the CAST experiment. The background levels achieved are very low, of the order of 4 10 6counts keV 1s 2cm 2, as a result of parameters such as their stable
operation in time, good energy resolution, the radiopure materials they are made of and the external shielding that covers them. There have been periods for the sunrise detector when the background level would reach very low levels, of the order of 2 10 7s 1cm 2keV 1
. This behaviour is still not well under-stood and several studies of the possible intervening parameters have been done or are underway. Within this context, the micromegas group of CAST has launched two lines of studies: on the one hand a complete simulation is underway describing the detector, its shielding and all the setup-components up to the signal creation, while on the other hand a replica of the CAST system has been mounted in the Zaragoza laboratory under controlled conditions in the hope of gaining insight into the parameters affecting the signal.
References
[1] K. Zioutas, et al., A decomissioned LHC model magnet as an axion telescope, Nucl. Instr. and Meth. Phys. Res. Sect. A 425 (1999) 480.
[2] K. Zioutas, et al., First results from the CERN Axion Solar Telescope, Phys. Rev. Lett. 94 (2005) 121301.
[3] S. Adriamonje, et al., J. Cosmol. Astropart. Phys. 04 (2007) 010. [4] M. Kuster, et al., New J. Phys. 9 (2007) 169.
[5] D. Autiero, et al., New J. Phys. 9 (2007) 171. [6] P. Abbon, et al., New J. Phys. 9 (2007) 170.
[7] E. Arik, et al., J. Cosmol. Astropart. Phys. 02 (2009) 008. [8] S. Aune, et al., Nucl. Instr. and Meth. A 604 (2009) 15. [9] S. Aune, et al., J. Phys. Conf. Ser. 179 (2009) 012015.
[10] Y. Giomataris, P. Rebourgard, J.P. Robert, G. Charpak, Nucl. Instr. and Meth. Phys. Res. A 376 (1996) 29.
[11] Y. Giomataris, et al., Nucl. Instr. and Meth. 560 (2006) 405. [12] S. Andriamonje, et al., J. Instr. 5 (2010) P02001.
energy (keV) 0 5000 10000 15000 20000 25000 30000 TRestDaq: microMegasGain energy (keV) 0 50 100 150 200 250 300 350 Experimental (M13)
Simulated (avalanche charge)
0 1 2 3 4 5 6 7 8 9 10 0 1 2 3 4 5 6 7 8 9 10
Fig. 10. (a) The energy spectrum in consecutive steps: original charge deposition in the chamber (black lines), charge creation (red) and including the energy resolution of the detector (yellow). (b) A comparison of the energy spectra with an55
Fe source: the blue line indicates the spectrum acquired from the complete simulation where the energy resolution of the detector was introduced while the black spectrum is real data taken with the Microbulk detector in the Zaragoza setup. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
