Astronomical observations give strong evidence for the existence of non-luminous and non-baryonic matter, presumably composed of a new type of elementary particle.  A leading candidate is the Weakly Interacting Massive Particle (WIMP). If they exist, WIMPs should form a cold thermal relic gas, which could be detected via elastic collisions with nuclei of ordinary matter. The detection of these WIMPs is based on the capability of measuring the recoils of target nuclei with kinetic energy in the range of 10-100 keV. The signal is therefore quite elusive and is expected to be a rare event given the weak coupling between WIMPs and ordinary matter.

A better understanding of the DAMA result or alternatively the coverage of large fractions of the remaining theoretical parameter space of theories accommodating WIMPs (e.g. supersymmetric extensions of the Standard Model) can only be achieved at the cost of improved detector technologies that are compatible with large target masses. Current experimental results from CDMS, ZEPLIN, and EDELWEISS exclude cross-sections above a few
10–42 cm2. Sensitivities down to 10-46 cm2 might be needed to exclude completely or find the existence of WIMPs.  This is four orders of magnitude lower than current experimental results.  Achieving this sensitivity is a currently very exciting challenge for experimental particle physics. Ton-scale targets must be contemplated in order to study potential signals from dark matter with high statistical power.

Noble liquid detectors using Xenon or Argon can efficiently act as targets for Weakly Interacting Massive Particles (WIMP) detection. Xenon or Argon provide a high event rate because of their high density and high atomic number and large target masses are readily conceivable.  They have high scintillation and ionization yields because of their low ionization potentials.  Both scintillation and ionization are measurable and can be used to very effectively discriminate between nuclear recoils and gamma/electron backgrounds.

The use of noble liquid gases to detect WIMP dark matter is currently the subject of intense R&D carried out by a number of groups worldwide. In these detectors, one relies on the simultaneous detection of the ionization charge and of the scintillation light produced during a nuclear recoil event.  A main subject for any such detector is the method of the readout for the ionization and scintillation.  Currently, the XENON ZEPLIN and WARP designs rely exclusively on photomultipliers (PMTs) for their readout.  The possibility to directly detect the ionization charge is less well developed although it might provide alternative and potentially large benefits.  Given the low energy thresholds necessary to efficiently detect WIMP signals, this method however requires the charge to be amplified before it is read out.  While amplification is not possible in the liquid Argon phase, it can be achieved in the vapor in equilibrium on top of the liquid, although operation in this context precludes the inclusion of common avalanche quenchers, since they will condense in the liquid phase. 

 

 

Figure 1 Cross-section normalized to nucleon versus WIMP mass.  The expected event rates for a true recoil energy threshold of 30 keVr are indicated by horizontal thick lines.  With such a threshold a WIMP-nucleon cross-section of 10^-44 cm² yields one event per ton per day.

 

In 2004 we have initiated the Argon Dark Matter experiment. The goal of this project is to design, assemble and operate a bi-phase ≅1 ton Argon detector with independent ionization and scintillation readout, to demonstrate the feasibility of a noble gas ton-scale experiment with the required performance to efficiently detect and sufficiently discriminate backgrounds for a successful WIMP detection. 

The choice of natural Argon for the initial ton-scale target instead of Xenon can be motivated by three arguments:

(1) The detection energy threshold depends on the achievable performance of the light and ionization detection systems.  The event rate in Argon is less sensitive to the threshold on the recoil energy than for Xenon because of form factors.  For a threshold of ≅30 keVr, the rates on Xenon and Argon per mass are similar. With such a threshold a WIMP-nucleon cross-section of 10^-44 cm² yields about one event per ton per day (See Figure 1).

(2) Argon is much cheaper than other noble gases, and we have acquired sizeable experience in the handling of massive liquid Argon detectors within the ICARUS program.  A ton-scale Argon detector is hence readily conceivable, safe and economically affordable.

(3) The scientific relevance of obtaining data on Argon and Xenon is given by the fact that recoil spectra in Xenon and Argon are different (due to kinematics), providing an important crosscheck in case of a positive signal.

One non-negligible drawback of natural Argon liquefied from the atmosphere is the existence of the radioactive isotope ³⁹Ar which is a beta-emitter with a lifetime of 269 years and a value Q=565 keV. Its concentration in atmospheric Argon is well known since the 1980's and will induce a background decay rate of ≅1 kHz in a 1 ton detector.  In principle, the intrinsic electron/nuclear recoil rejection provided by the ratio of the scintillation to the ionization yields, which is extremely high for nuclear recoils (i.e.  WIMP events), is sufficient to suppress this background, provided this ratio can be measured precisely. This fact needs to be experimentally further understood since rejection factors exceeding >10⁹ are needed.  We intend to fully address it with our proposed 1 ton prototype, i.e. a detector of the relevant size[1]. We are also studying other ways to obtain ³⁹Ar-depleted targets, by using Argon extracted from well gases (extracted from underground natural gas) rather than from the atmosphere.  This would provide a reduction of this background although its cost is to be estimated.  On the other hand, the ³⁹Ar decays, evenly distributed in the target, provide a precise calibration and monitoring of the detector response as a function of time and position.

The conceptual layout of the detector has been defined (See Figure 2). A main feature is the possibility to independently detect the ionization charge and scintillation light.  Following an ionizing event, ionization charges will be drifted towards the top of the detector where they will be extracted from the liquid to the gas phase.  There, a Large Electron Multiplier (LEM) system will amplify the electrons in order to produce a detectable signal.  By segmenting the LEMs, an image of the event will be obtained, retaining the salient features of the ICARUS imaging technology, although with a much lower energy threshold.

 

 

 

Figure 2 The conceptual design of the ArDM experiment

 

Because background discrimination requires the ratio of the scintillation to the ionization yields, the primary VUV scintillation light of argon (128 nm) will be reflected by specially conceived high reflectivity mirrors made of Al-MgF₂ coated Mylar foils and located on the field shaping electrodes.  The photons are detected via a light readout system located behind the transparent HV cathode.  R&D efforts are under way to improve on the light collection efficiency (about 5% with PMTs), and hence on the threshold and background discrimination, by using wavelength shifters and alternative light readout systems such as avalanche photodiodes.

Charge imaging and time correlation between scintillation and charge will provide a precise localization of the event vertex (in space), hence a good fiducial volume definition, important for γ-ray and neutrons background rejection from surrounding elements. 

The time dependence of scintillation light can be used to further discriminate between heavy recoils and other backgrounds (in addition to primary versus secondary signal).

A second feature of the experiment is the possibility to reach very high drift fields up to 5 kV/cm in order to detect an ionization signal even in the presence of highly quenched nuclear recoils as in the case of a WIMP interaction.

Given the challenging nature of the experiment which requires innovations both at the level of the detection methods and at the level of background rejection, our immediate plan is to fully design and acquire the needed equipments to setup and operate the 1 ton prototype at CERN. The operation of the prototype will involve cryogenic, LAr purification, HV system, drift volume, charge amplification plus readout, and light readout.  It will allow us to define and set up all the necessary equipment and infrastructure for a safe operation of the detector.

Our first milestone is a proof of principle and stability studies, and further optimization of the design for a highly efficient γ-ray and beta electron (³⁹Ar) rejection vs.  nuclear recoils.  Strong neutron shielding and stringent requirements on detector radio-purity will be fully addressed in a second phase. Assuming the successful operation of the prototype, we will consider a deep underground operation. With the assumed recoil energy threshold of 30 keVr, a WIMP-nucleon cross-section of 10^-42 cm² would yield 100 events per day per ton (See Figure  Figure 1). The sensitivity expectation of the ArDM 1 ton prototype would therefore be around ∼10^-6 pb or better.  The discovery region of the ArDM 1 ton detector, assuming that sufficiently low gamma and neutron backgrounds can be reached, would be ∼10^-8 pb.  Its ultimate sensitivity for a year of operation would be ∼10^-10 pb.  Scaling linearly with mass, a ≅10 ton detector would reach <10^-11 pb in a year of operation.