The 35-ton prototype’s second data run has proven quite valuable in the development of the DUNE detector technologies. The detector elements survived the cooling and filling process and the filtration system purified the liquid argon to the required level within a week. Despite observed noise in the TPC readout electronics during data-taking — and an external tubing failure that caused oxygen contamination of the liquid argon, cutting the run short — the collected measurements have already inspired improvements to carry forward.
In the 35-ton cryostat design the inlet and outlet of the filtration/recirculation loop are located at the bottom of the tank and close to each other, which the team found limits the efficiency of the system and thus the achievable electron lifetime.
“Since the purified, returned LAr is colder than the overall body of LAr, it tends to settle at the bottom and resist mixing in. The pump intake, being at the bottom, preferentially picks up this colder, purer layer of LAr to send to the filters, effectively reducing the efficiency,” said Alan Hahn. “A redesign will improve the mixing.”
A second improvement is elimination of the warm plate covering the raised portion of the cryostat top. This change, already incorporated into the ProtoDUNE-SP design, is expected to reduce outgassing and thus also to enhance the purity of the LAr.
Thirdly, to reduce fiducial volume losses in the gaps between the APAs, a deflecting electrode has been proposed for installation in the gaps. The 35-ton has recorded a sample of through-going muon tracks, identified via external scintillation counters, that will help quantify the performance of reconstruction algorithms across the gaps and test the ability to resolve ambiguities introduced by the wrapped wires on the APAs. The 35-ton data will enable tuning of the simulation that models charge loss in the gaps, and will allow comparison with ProtoDUNE-SP data that will be taken with the proposed electrode.
Finally, a combination of frequency filtering and subtraction of coherent noise is being developed that reduces the noise on the TPC readout electronics to manageable levels for offline data analysis, provided that zero suppression is not applied to the data before they are recorded on disk.
“Methods to filter and subtract noise don’t necessarily help us for running the detector with zero suppression, as planned, but this work makes the 35-ton data useful despite the excessive noise levels,” said Michelle Stancari. “We’re learning many things that will help us move forward as we complete the ProtoDUNE-SP design.”
— Anne Heavey