TL;DR
Researchers have developed a novel internal calibration method for the nEXO experiment using the radioisotope $^{127}$Xe. By injecting small amounts of this gas directly into the detector’s liquid xenon, they can continuously map the detector's response in real-time without interfering with the search for rare physics events.
What Problem Does This Paper Address?
The nEXO experiment is a massive, 5-tonne liquid xenon (LXe) time projection chamber (TPC) designed to search for neutrinoless double beta decay ($0\nu\beta\beta$). If discovered, this decay would prove that neutrinos are their own antiparticles and help explain the matter-antimatter asymmetry of the universe.
To succeed, nEXO needs incredible energy resolution (better than 1% at the signal peak). However, a detector this large has non-uniform responses:
- Lightmap: Scintillation light is absorbed or reflected differently depending on where an event occurs.
- Electron Lifetime: As ionized electrons drift through the xenon, they can be captured by impurities, "dimming" the signal over time.
Standard external calibration (placing radioactive sources outside the detector) fails because 5 tonnes of xenon act as a powerful shield, preventing radiation from reaching the center. Existing internal sources like $^{220}$Rn are useful but have short half-lives or produce backgrounds that can leak into the signal region.
Methods
The authors propose using $^{127}$Xe, an isotope that decays via electron capture with a 36.3-day half-life.
- Production: They produced the source by exposing natural xenon gas to neutrons at the McClellan Nuclear Research Center (MNRC). This activates the $^{126}$Xe already present in the gas.
- Radioassay: To ensure purity, they used High-Purity Germanium (HPGe) detectors to identify the resulting isotopes. They confirmed the presence of $^{127}$Xe and checked for long-lived contaminants like $^{137}$Cs.
- Experimental Demonstration: The team injected the activated gas into a small-scale prototype TPC at Stanford. They monitored the trigger rate to confirm mixing and used the resulting data to measure the "electron lifetime" ($\tau_e$) of the system.
- Neural Network Reconstruction: For the full-scale nEXO projections, they moved beyond simple data binning. They trained a 5-layer feed-forward Neural Network (NN) to learn the 3-dimensional "lightmap" of the detector, which allows for smoother and more accurate corrections in a massive volume.
Datasets
- Experimental Data: Collected from the Stanford LXe TPC prototype to validate injection and mixing.
- Simulated Datasets: Generated using the
nexo-offlinesoftware package (based on Geant4 and NEST). These datasets included up to $10^7$ simulated $^{127}$Xe decays to project how the source would perform in the full 5-tonne nEXO detector.
Key Insights and Ideas
- The "Goldilocks" Isotope: $^{127}$Xe is ideal because its 36.3-day half-life is long enough to allow the gas to mix uniformly throughout the 5-tonne volume, but short enough that it doesn't become a permanent background.
- Low Energy Footprint: The $Q$-value of $^{127}$Xe is 662 keV, which is far below the $0\nu\beta\beta$ signal region of 2,457 keV. This allows for "always-on" calibration (at ~1 Bq activity) that doesn't interfere with the primary physics search.
- No-Threshold Analysis: To avoid bias in measuring electron lifetime caused by charge diffusion, the authors developed a "no-threshold" reconstruction. By summing signals from the five nearest channels regardless of whether they hit a trigger threshold, they recovered "lost" charge and improved the accuracy of the lifetime measurement.
- NN-Driven Calibration: Using a Neural Network to model the lightmap is a significant shift from traditional voxel-based maps. The NN can interpolate between data points and handle the complex geometry of a massive TPC more efficiently.
Key Results
- Successful Production: The team successfully produced ~6.5 kBq of $^{127}$Xe, demonstrating a viable supply chain for the experiment.
- High Precision: Simulations show that with just 1 Bq of continuous activity, nEXO can calibrate its electron lifetime daily with negligible impact on energy resolution.
- Lightmap Accuracy: After 16 days of data collection, the NN-based lightmap reconstruction achieved an error of less than 1% across the entire detector volume, and less than 0.5% in the inner 2 tonnes (the "sweet spot" for physics).
Limitations and Future Work
- Contamination: While $^{137}$Cs production was lower than predicted, any long-lived cesium would need to be removed via standard purification (e.g., getters or distillation) before full deployment to prevent background buildup.
- Trigger Thresholds: The analysis assumed the 236 keV peak of $^{127}$Xe might fall below nEXO's trigger threshold, so they relied primarily on the 408 keV peak. If the threshold is higher than expected, calibration times might increase.
- Hardware Scaling: The injection system worked for a small prototype, but metering precise, sub-Bq amounts of gas into a 5-tonne system will require further engineering of the gas handling system.
Summary updated