The COBRA detector crystals

There are multiple standard semiconductor materials available, which are often used for semiconductor detectors. For example, Germanium and Silicon can both be relatively easily grown in large homogeneous crystals, which are perfect for the use as a detector.
COBRA, however, decided to try out a relatively new semiconductor material, which is still under development. COBRA uses detector crystals made from CdZnTe. The material is able to produce various interesting double-β-decays.
Because CdZnTe is such a young material, the process of growing homogeneous single crystals is still challenging and under constant development. One cause for the process being so challenging is, that CdZnTe is made up from three different element components. Crystals with multiple such components are always more difficult to grow than ones containing only one type of atom. The different atomic radii of Cd, Zn and Te cause stress within the crystal structure during the growth. This poses the danger of the crystal developing defects in order to relax the crystal structure.
However, since the installation of the first stage of the COBRA experiment in 2011, major improvements have been already made in the growth of CdZnTe. While during the installation of the COBRA demonstrator in 2011 the largest available crystals were of the size of 1 cm³, now crystals of about ten times the size are feasible. That is why in 2018 an additional set of detectors was added to the setup, which are 6 cm³ in size.
The crystals COBRA uses are provided by the manufacturers REDLEN and eV-PRODUCTS/KROMEK.

Different CdZnTe crystals produced by REDLEN. The large crystals have a size of 6 cm³ each and the small crystal of 1 cm³. The small crystals were used for the original COBRA demonstrator array, while the new COBRA XDEM addition to the setup is equipped with the large crystals.
The structure on top of the crystals is the Coplanar Grid used for the read-out of the detector. Different designs of this grid, as visible in the picture, were tested during the installation of XDEM.

Semiconductor detectors function by producing free charge carriers (electrons and holes), when hit by electromagnetic radiation. If the semiconductor is equipped with an anode and a cathode with a voltage applied, those charge carriers can form a current which can then be measured.
Despite a very careful and elaborate production process of the CdZnTe crystals, the resulting semiconductor material ends up having some impurities and defects. Those can form so-called traps, binding charge carriers of the material, which are then missing from the measured current. Especially the holes are affected by this. The further the holes must travel through the detector in order to reach the cathode, the more of them get trapped in the process. Hence, the measured current is dependent on the interaction depth in the crystal. A particle with a certain energy passing the detector would not always cause the exact same current, how it would be preferable.
In order to work around this effect, a special read-out technology is used for COBRA, the coplanar grid. The technology was developed by P. Luke in 1994 and is comparable to a Frisch grid used in ionization chambers.
The underlying idea is, to only measure the signal of the electrons produced in the detector by an ionizing particle. The depth dependent hole signal is discarded. Therefore, a special design of the detector’s read-out electrodes is necessary: The cathode side of the detector is simply completely plated with metal, as it is the easiest way to form such an electrode. But the anode side looks much more complex. Two separate anodes are implanted there, interleaving each other in a comb-like structure. They are called collecting anode, short CA, and non-collecting anode, NCA. Between the cathode and the anodes a bulk voltage of about 1 kV is applied, to drift the charge carriers towards the electrodes. Additionally, there is a small bias voltage of about 50V between the two anodes: The collecting anode lies on ground potential while the non-collecting anode is biased with a slightly negative potential.

Electrode design of the COBRA demonstrator detector crystals.

Free electrons in the detector material are accelerated towards the anodes because of the large bulk voltage. While they move through the main part of the detector, the small difference in voltage between the two anodes does not play a significant role. Only, when the electrons have nearly reached the anodes, their voltage difference starts to gain importance. Because the collecting anode lies on a slightly more positive potential then the non-collecting anode, the collecting anode collects the electrons, while the non-collecting does not. Hence, for the electrons the two anodes see different signals.
The holes, however, usually do not reach the space so close to the anodes, where their difference in voltage starts to get important. The holes just see the large bulk voltage, which accelerates them towards the cathode. Hence, the holes produce the same signal for both collecting and non-collecting anode.
By only working with the signal difference of collecting and non-collecting anode, the influence of the holes on the signal is eliminated and the signal gets depth-independent.

The much smaller effect of electrons being trapped in the detector material as well, can be accounted for by the introduction of a correction factor while calculating the energy from the anode signals at a later stage.