TRT Trophies, Cobras and BBQs

22 September 2010

This article in fact is all about the performance of the detector with the best transition radiation detection capabilities and possibly best barbeques in ATLAS. Not to mention the only ATLAS detector that has ‘knighted’ its project leader.

The Transition Radiation Tracker (TRT) is the outermost charged particle tracking device of the ATLAS Inner Detector. The TRT has about 300,000 straws, each of which is a proportional drift tube with a diameter of 4 mm providing a position resolution of about 130 microns. Charged particles typically cross more than 30 straws, and as a result the TRT significantly contributes to the momentum resolution of the Inner Detector. In addition to being a tracker, the TRT also has particle identification capabilities, in particular for electron identification. The space between straws is filled with materials with different dielectric constants, CO2 gas and plastic fibers in the barrel and plastic foils in the end-cap region. When particles of sufficiently high Lorentz gamma factor cross a boundary between the different materials, they emit transition radiation (TR) photons. When absorbed in the xenon gas mixture in the straws, these photons give rise to a much higher signal of about 9-10 keV, compared to 2.2 keV deposited on average by a more generic ionizing particle such as a pion. The detection of many TR photons on the track is used to differentiate between electrons and pions for particle momenta of up to 200 GeV, where pions are still below the threshold to produce TR.

Thanks to the hard work of our online detector experts, the TRT continues to run with 100% operational efficiency. This is particularly impressive as the TRT is one of the most sensitive ATLAS detectors to changes between internal ATLAS and external LHC clocks. Resynchronization procedures can automatically recover the TRT readout in less than 10 seconds without the need to stop the run. Most of the other rare readout problems are recovered with stop-less removal and recovery procedures, which also do not require ATLAS to stop and restart a run. Nevertheless, TRT shifters in the ATLAS Control Room are ready for anything – even for cobras that appear in the shifter's control window as shown here.

The “Cobra panel” is available at the TRT shifter's desk in the ATLAS Control Room. It shows the status of the TRT back-end electronics. If a problem appears, the corresponding region is lit in red and the shifter can navigate through the directory structure to identify where exactly the problem is coming from.

The large number of channels in the TRT and the expected high detector occupancy at higher luminosities present additional challenges to the detector readout. To cope with large data throughput, lossless data compression is implemented in the back-end electronics; validation stress tests have confirmed that the TRT can run stably with full compression at readout rates up to 80 kHz and at occupancies approaching 100%. To help diagnose and solve any future problems, a test setup of a slice of the TRT has been built in the SR1 facility. The duplicate setup runs the full readout chain and has the capability to record standalone cosmic-ray data. If you would like to see a real TRT end-cap wheel and barrel modules, it is definitely worth a visit to SR1.

Stable operation requirements and the high cost of xenon impose strict requirements on the TRT gas system, which is the tightest gas system in ATLAS, with typical losses of less than 0.5 liters per hour. The TRT operates with a gas mixture of 70% xenon, 27% CO2 and 3% O2. It is important to keep these concentrations stable, as variations affect drift velocities and can worsen the position resolution. Active control of the gas mixture ensures that the CO2 concentration is within 0.3% of the target value, and O2 within 0.1%.

Of course, such a good detector deserves an equally good offline counterpart. The TRT measures time, but what is needed during track reconstruction is a position measurement. For a precise measurement of the trajectory of a charged particle, the relation between the measured time of the start of the signal and the distance of closest approach between the track and the anode wire needs to be calibrated. As a part of the ATLAS prompt calibration loop, TRT calibration constants are validated for every run with collision data, and updated when appropriate. The calibration is very stable over time; the main reason to update constants has been to compensate for slow drifts of the LHC clock. On a longer timescale, the main offline efforts focus on understanding and improving the detector performance as much as possible. Recent improvements to tracking include straw-level alignment, which improves the position resolution, in particular in the end-cap region, and time-walk corrections based on signal size, which correct for the fact that hits with larger energy deposit tend to exceed the threshold sooner.

Studies of the transition radiation performance have shown that the amount of TR in data slightly exceeds expectations from Monte Carlo as seen on the plot below. New results from data are being used to improve the Monte Carlo description. Dedicated studies have confirmed that the high threshold setting is very close to its optimal value, and that the performance does not significantly depend on the threshold in this range. A requirement on the fraction of TRT hits on track that have TR hits is part of the standard ATLAS tight electron selection, which was used in the first W->eν observation. In addition, this fraction was used in an early inclusive electron analysis to help distinguish real electrons from hadron fakes. Typical hadron rejection factors of over an order of magnitude, independent of the calorimeter, can be obtained with 90% efficiency to keep real electrons.

The probability of a TRT high-threshold hit as a function of the Lorentz factor gamma in the end-cap region, measured in 7 TeV collisions data. For gamma above 1000, a pure sample of electrons is obtained from photon conversions. For more information, see the description on the publicly available plots page.

Much progress has also been made in using the signal size to differentiate between electrons, pions, kaons and deuterons at lower energies, based on differences in ionization loss. Studies on how to use this information for SUSY searches are also well advanced. In addition, the capability to measure the track drift time with better than one nanosecond accuracy will help identify slow stable massive particles, as well as help reject muons from cosmic-ray background.

Work also continues offline in understanding how best to use the TRT at larger instantaneous luminosities and higher detector occupancies. The TRT can and will run during the heavy ion run at the end of this year at the nominal settings. After all, the detector and the team around it are always ready and eager to record collision data.

And what about the trophies? Well, there was more than one. To learn more about the TRT, our BBQs and trophies, or to have a chance of being knighted, subscribe to atlas-trt-bbq at cern.ch where everybody is welcome.



 

 

Jahred Adelman, Yale University

Sasa Fratina, University of Pennsylvania