The Krypton Problem
One of the many advantages of using xenon as a dark matter target is that xenon has no naturally occurring long-lived radioactive isotopes. However, when xenon is distilled from air, about 1 krypton atom per billion xenon atoms is also gathered. A very small fraction of these krypton atoms, only one in one hundred billion, are the radioactive isotope 85-Kr.
The decay of 85-Kr releases an electron which can then scatter in the xenon detector. These electronic recoil events can potentially obscure even rarer signals from interactions with dark matter. Thus, for dark matter detectors using liquid xenon, the krypton needs to be removed. This is done by passing the xenon through a cryogenic distillation column specifically designed for removing krypton.
After going through the krypton column, the xenon is very clean. For XENON100, there are only ~10 krypton atoms per trillion xenon atoms. Finding one of those krypton atoms is like picking out one single star from the entire Milky Way galaxy. XENON1T has 10 times even less krypton in the xenon.
Measuring the Krypton Contamination
Measuring such a tiny amount of krypton is not trivial. One way is to look for the decay signature of 85-Kr using the XENON detector itself. However, due to its relatively long half life (~11 years), it takes many months to get an accurate estimate with this method. So, how do we measure the tiny amount of krypton relatively quickly and accurately?
An atom trapping device has has been developed by the group at Columbia University to do exactly that (see E. Aprile, T. Yoon, A. Loose, L. W. Goetzke, and T. Zelevinsky, “An atom trap trace analysis system for measuring krypton contamination in xenon dark matter detectors”, Rev. Sci. Instrum., 84, 093105 (2013), arXiv:1305.6510). The method, called Atom Trap Trace Analysis (ATTA), was originally developed at Argonne National Lab for the purpose of radioactive dating. It has been adapted to measure samples of xenon gas taken directly from the XENON detectors.
All ATTA devices have the same operating principle: traditional laser cooling and trapping techniques are employed to selectively cool and trap the element of interest present in the sample. The trapped atoms emit light which is detected by a photo detector, in our case an avalanche photodiode. The trapped atoms can thus be counted. The Columbia ATTA device is designed to be sensitive to single trapped atoms, since for clean samples the average number of krypton atoms in the trap at any given time is close to zero.
The rate at which the atoms are loaded into the trap is the number we are after. The device is calibrated carefully in order to find the trapping efficiency, i.e. the fraction of krypton atoms that get trapped and counted successfully. Multiplying the measured loading rate for a given sample by the known trapping efficiency gives the total number of krypton atoms flowing through the system. Finally, measuring how many xenon atoms flow through the system at the same time allows the krypton fraction to be calculated. The entire measurement can be completed in one working day.
The Columbia ATTA device allows the xenon used in XENON1T to be assayed for krypton contamination quickly and accurately, thus ensuring that krypton levels are safe before beginning a dark matter run, and during the run itself. And it looks pretty cool, too!