The CRESST detectors are operated at about 10 mK where superconducting
phase transition thermometers (tungsten films) become superconducting. To obtain such
low temperatures we use 3He/4He dilution refrigerators.
The 3He/4He dilution refrigerator operation
principle was originally proposed by
Heinz London in 1951, but the first working systems were not built until more than
ten years later. Since that time, the performance of these systems has steadily
improved, and the physical processes involved have become much better understood.
These days, commercial 3He/4He dilution
refrigerators are available for temperatures
down to about 4mK, however, they are expensive and demand substantial experience,
time and manpower. The minimum temperature obtained by the method to be discussed below
is about 2mK, achieved by Frossati and co-workers.
The usual method of maintaining a low temperature is by evaporation of liquids with
a low boiling point . The essence of the dilution refrigerator method is to use a mixture of 3He (boiling Temperature 3.19K) and 4He
(4.2K) which is separated into two phases (see Figure 1).
Inside the "mixing chamber", where the mixture is condensed, the two distinct phases are formed. One of them contains only 3He while the other mainly consists of 4He with only a small fraction of 3He.
During operation, one removes 3He from the 4He-rich phase by pumping on it. As a consequence, 3He atoms are forced to move from the 3He-rich phase into the 4He-rich phase, a process which is endothermal and therefore cools the system. For continuous operation, one keeps this process alive using a closed circulation of 3He, so that it is possible to run a built-in experiment at low temperatures.
Figure 2 shows a 3He/4He mixing cryostat
at the Max Planck Institut München.
Fig. 1: Lambda line for the superfluid transition of 4He.
Fig. 2: A 3He/4He dilution refrigerator at the Max Planck Institut München.