Next Contents Previous


The direct detection of WIMPs became a very dynamic field of research during the last years. There are about 20 experiments running or being prepared worldwide and even more planned for the near future (for collections of contributions from most of these, see [21], [22]). Various techniques and detector systems are applied. They can be classified by the applied detectors [7]. Here they are classified by their ability to discriminate nuclear recoils to some extent, i.e. to use recoil-specific observables (compare Fig. 2). That separation of experiments might give the impression that the most advanced experiments will use recoil discrimination techniques. On the other hand, note that this particular ability also adds complexity and therefore there are in fact experiments which do not apply any recoil discrimination but, nevertheless, give competitive perspectives (see below).

As shown in Fig. 2, the trick is for all no-discrimination detectors, like germanium semiconductor detectors (HD-Moscow [23], HDMS [23], GENIUS [24]), the cryogenic bolometers (CUORE [25], Tokyo [26], CRESST [12], Rosebud [27]), the superheated superconducting grains detector (ORPHEUS [28]) or the NaI scintillator ELEGANT V [29], to use materials and shieldings for the setup as radio-pure as possible. Since by now this kind of experiments have already collected a large amount of experience in using clean materials it has been thought that their sensitivity might have reached a saturation and no further breakthroughs could be expected. As it turned out, this assumption is not true at least for germanium detectors (see the GENIUS expectation in the next section).

The lowest background and therefore the best limit from raw data is still obtained by the Heidelberg-Moscow experiment. It uses an enriched 76Ge detector of 2.758 kg active mass and reached a background near threshold of 5.7 10-2 counts/(day kg keV). Due to the rather high threshold of 9 keV its limit for lower WIMP-masses can be combined with another germanium experiment [30] to give a combined Ge-diode limit (see Fig. 5) close to the currently best limits on spin-independent WIMP-nucleon interactions. HDMS is a specialised WIMP-detection setup from the same collaboration using a germanium well-type detector as active veto for a small inner germanium detector. After exchanging the inner natural germanium crystal by a 73Ge enriched crystal the prospects are to improve the existing limit by about a factor of 5-10, thereby challenging current limits.

Other more complex techniques used for raw data experiments are cryogenic detectors. Two collaborations published first results recently, the LiF-crystal experiment in Tokyo [26] and Rosebud [27] using sapphire (Al2O3) crystals. While the sapphire setup does not give competitive results so far it gives important insight into background contributions for cryogenic experiments in general. The LiF-experiment, although operating still at a shallow depth (15 m.w.e.), now improved the limit for light WIMP masses below 5 GeV (see Fig. 3). Due to their target materials both experiments are most sensitive for light WIMPs and the spin-dependent interaction channel. The Tokyo experiment will soon move to a deep underground laboratory and tries to remove identified background sources close to the detectors so that their estimate is to improve the current limit by an order of magnitude. Similar considerations have been put forward for Rosebud.

Figure 3

Figure 3. Collection of spin-dependent cross section limits for several direct detection experiments as function of the WIMP-mass (from [26]). Note the improved limit below 5 GeV and the large distance of limits to expectations (for neutralinos) for this particular interaction channel.

The pulse-shape discrimination technique for NaI-scintillator detectors (DAMA [31], UKDMC [32]) has been the first applied discrimination technique and turned out to be quite effective for increasing energies. Still the best limit on WIMP-nucleon cross sections come from the DAMA collaboration (for the scalar channel above 40 GeV, compare Fig. 5). The calibration of this method by the production of nuclear recoils from neutrons showed that these pulses are significantly faster than electron recoil pulses from photons or electrons. Recently, there emerged the feature of a class of pulses even faster than nuclear recoil pulses in two different experiments using NaI detectors (UKDMC and Modane [33]). These are considered to belong to an unknown background source and might even give a systematic limit of sensitivity for this technique. However, this effect is still under investigation and might as well be removed in the near future.

Figure 4

Figure 4. Short summary of the most intriguing result from the DAMA NaI experiment (from [31]). Note that the upgrade to 250 kg has been approved.

A highlight of NaI detectors is not only their discrimination ability and thereby the high sensitivity for WIMP detection but also the possibility to setup high target masses like the DAMA-experiment (87.3 kg active volume, see Fig. 4). This puts the DAMA-experiment into the unique position of having the ability to even use WIMP-specific observables like the WIMP-signature of an annual modulation. Since that effect is just of the order of a few percent one has to collect a sufficient statistic to filter out the tiny modulation [34]. In fact, this collaboration announced an evidence for the detection of that WIMP-signature. As shown in Fig. 4, they see an annual modulation in their data consistent with the expectation for a WIMP at

Equation 2

where xi sigmascalarnucleon is the local halo density normalised (to 0.3 GeV/cc) WIMP-nucleon scalar cross section. The consistency requirements are the proper kinematic modulation (phase, amplitude and period), single hits in detectors and the proper signal shape (maximum signal in the lowest energy bin and subsequent decrease). However, the excitement about this evidence is accompanied by a similarly engaged criticism [35] inside the dark matter community. Fortunately, this is a matter of experiment, i.e. it will be possible already in the near future to test the evidence by experiments rather than arguments.

The first competitors of the DAMA experiment which are expected to be able to test their result in the near future are the cryogenic detector experiments CDMS and EDELWEISS [37]. They use a combined signal readout of phonon signals and ionization signals from germanium (and silicon in case of CDMS) crystals. The clue of this kind of readout scheme is that in germanium crystals the ionization efficiency of nuclear recoils is just about 25% (energy dependent, see [38] and references therein) compared to an electron recoil event. So far, they both suffered from the effect of incomplete charge collection of surface electron events which could mimic nuclear recoil events. Recently the CDMS experiment got rid of this problem (see Fig. 5 and [36]) and now they already test the DAMA result below about mW = 40 GeV. The EDELWEISS collaboration is expected to follow that development soon and release a new improved limit comparable to or even better than the CDMS result.

Figure 5

Figure 5. Summary of current WIMP-nucleon cross section limits for spin-independent interactions from the CDMS collaboration [36]. The best limit for WIMP-masses above 40 GeV stem from the DAMA collaboration, below CDMS now tests the DAMA evidence contour. The combined Ge-diode limit is shown as dash-dotted line and dashed the UKDMC NaI result (limited by the unknown fast pulse shape component, see text).

Apart from that actual status report it is worth mentioning rather mid-term projects which show the variety of applied detection techniques in order to reduce background. The CASPAR proposal [39] uses small grains of CaF2 scintillators (of the order of a few hundred nanometer diameter) dissolved in an organic scintillator. Calcium or fluorine nuclear recoils would only produce a scintillation signal from their grain whereas electron recoils would have a much larger range and would give signals from the organic scintillator as well which then could be discriminated. Another discriminating detector concept using ionisation and scintillation signal readout from liquid Xenon (or two-phase Xenon, gas and liquid), the ZEPLIN [40] project, is still in its R&{D-phase but first tests are very promising.

The superheated droplet detectors PICASSO [41] and SIMPLE [42] also use the specific energy loss of nuclear recoils to discriminate against minimum ionising particles. They use a well known technique for neutron dosimetry, namely droplets of a slightly superheated refrigerant liquid embedded in a gel. The droplets would then act like tiny bubble chambers, exploding due to a nuclear recoil event. By tuning the relevant parameters, pressure and temperature, the bubbles can be made insensitive to nuclear radiation so that practically only recoils from fission processes and neutrons remain background sources. Both experiments are currently build up and first results from PICASSO and SIMPLE have been reported recently.

Next Contents Previous