The idea behind CDMS is to observe a WIMP coming into contact with another atom. The WIMP will interact very slightly with the atom and the atom will recoil and release an exact amount of energy. This recoil action is measured and if the energy is an exact value predicted, it is clear that the atom interacted with a WIMP. (4, 6)
To create this experiment, physicists fill a vat full of semi-conducting materials, such as silicon and germanium, and cool the materials to a fraction of a degree Kelvin above absolute zero. Currently, the setup for CDMS II is located in an old mine in Minnesota at a depth of 780 meters. Figure 7 shows the setup of CDMS II. It is important to have the detector so low in the ground because it allows other particles, such as neutrinos to interact with the rocks and therefore they would not make it into the vat. If they do make it into the vat, they could produce a signal similar to that of a WIMP, so putting the experiment deep into the Earth creates a better chance of detecting WIMPs. This means that the detector must be extremely exact and precise to analyze such a small energy change.
Another problem with the detector having to be so precise is that the radiation from the walls of the mine needs to be removed completely. Radiation produces gamma rays that could easily interact with the germanium or silicon to produce a fake signal. In order to remove the possibility of contamination from radiation, the detector is surrounded by liquid elements that will remove all possible background noise. (4, 6)
Another problem with trying to detect WIMPs through a recoil of the atom is the fact that the WIMP must come into contact with the WIMP-nucleon cross section, which is extremely small. An analogy to this cross section is trying to throw a ball at a target. The larger the target, the better the chance of hitting it. When it comes to a WIMP hitting the “target” atom, the cross section is extremely small about \(10^{-42} cm^2\)! This means, not only is there an extremely low chance of a WIMP entering the container, but it needs to interact with a target of only \(10^{-42} cm^2\). The probability seems impossible! So, even after operating for over 10 years, CDMS still has seen no evidence of WIMPs. (4, 6)