The Net Advance of Physics: The Nature of Dark Matter, by Kim Griest -- Section 6F.

Direct Detection

The most exciting result would be direct detection of the Wimp

particles. Since we roughly know the speed ( km/s) and the

density ( GeV cm ), we can say that for a Wimp of mass of

order 10-100 GeV, roughly 100,000 dark matter particles a second

pass through every square centimeter of the Earth, including our

bodies. If they exist, these are very weakly interacting particles, so it

is quite rare that one of them will interact at all. In addition, if one

does elastically scatter off a nucleus, the deposited energy is usually

in the keV to 100 keV range, so extremely sensitive devices must be

used. These difficulties, however, have not stopped many groups

throughout the world from developing devices capable of detecting

Wimps. See references [22,23] for details.

In deciding the size, sensitivity, and energy threshold of a detector,

the experimentalist would like to know what event rate one expects

in the case that the dark halo consists entirely of Wimps. For an

unspecified Wimp, only rough estimates can be made using general

arguments [24], but for neutralino Wimps, the elastic scattering

cross section and the event rate per kilogram detector per year can

be calculated, once the supersymmetric model parameters are

chosen. Figure 3 ([fig], [captions]) shows a scatter plot of the rate in

a Ge 73 detector, for all the models that pass the accelerator

constraints and have relic abundances in the range .

The ``stripes" in the plot are due to the finite grid we sampled in

parameter space, and so the spaces between the stripes should be

mentally filled in. A kilogram of germanium was chosen since this

is roughly the material and size of one of the most advanced

experimental efforts (see below). We see that if neutralinos of

around 50 GeV mass make up the dark matter, the expected event

rate is probably between and 1 event/kg/year.

When a Wimp scatters off a nucleus, the nucleus recoils, causing

dislocation in the crystal structure, vibrations of the crystal lattice

(i.e. phonons or heat), and also ionization. The main difficulties in

these experiments come from the fact that the events are rare and

that there are many backgrounds which deposit similar amounts of

energy on much more frequent time-scales. So in the past few years

the main experimental efforts have gone toward increasing the

mass of the detectors and discriminating the nuclear recoil signal

from the background. Generally the detectors must be operated

deep underground at milli-Kelvin temperatures, and be heavily

shielded.

An illustration of the problem is shown in Figure 4, ([fig],

[captions]) which shows the background in a germanium detector

built by the Berkeley, LBL, UCSB group and operated under the

Oroville dam [26, 27]. One sees many background processes

including lines from radioactive elements and tritium, electron

noise at low energy deposited, and a roughly constant background at

about one event/kg/day/keV. Comparing this to a typical expected

signal in Figure 5 ([fig] , [captions]), one sees the problem. However,

the vast majority of the background comes from gamma rays,

while the Wimp signal would be nuclear recoils, and it has been

established that gamma rays deposit a much larger fraction of their

energy in ionization than in phonons or heat. So the

experimentalists measure simultaneously the energy deposited

in heat and the energy deposited in ionization and are therefore

able to reject perhaps 99% of the background gamma rays [25].

This kind of discrimination is possible only in materials such as

germanium and silicon which can be used as ionization detectors,

but for other materials such as NaI and Xenon other effects such as

pulse shape or scintillation light may be used to separate the

gamma-rays from the nuclear recoils [28]. Using the CDMS

(Berkeley/LBL/UCSB/Stanford/Baksan) collaboration as an

example [25], the sensitivity of experiments starting to run this year

is in the 0.1 to 1 event/kg/year range, and upgraded versions hope to

reach 0.01 event/kg/year within a few years.

Returning to Figure 3 ([fig], [captions]), one sees that there are

viable supersymmetric models which will be explored and that a

discovery is possible. However, one also sees that a definitive

experiment will not be possible within the next few years, since rates

below the expected experimental sensitivity are common. However,

it is remarkable to realize that these small underground

experiments are competing directly with CERN in the race to

discover supersymmetry. And the enormous increases in sensitivity

these experiments have accomplished in the past few years leads

one to expect further such advances in the future.

BIBLIOGRAPHY