Scintillation Pulse Shape Discrimination in a Two-Phase Xenon Time Projection Chamber

The energy and electric field dependence of pulse shape discrimination in liquid xenon have been measured in a 10 gm two-phase xenon time projection chamber. We have demonstrated the use of the pulse shape and charge-to-light ratio simultaneously to obtain a leakage below that achievable by either discriminant alone. A Monte Carlo is used to show that the dominant fluctuation in the pulse shape quantity is statistical in nature, and project the performance of these techniques in larger detectors. Although the performance is generally weak at low energies relevant to elastic WIMP recoil searches, the pulse shape can be used in probing for higher energy inelastic WIMP recoils.

💡 Research Summary

**
The paper presents a systematic study of pulse‑shape discrimination (PSD) in a two‑phase xenon time‑projection chamber (TPC) and demonstrates how combining PSD with the charge‑to‑light ratio (S2/S1) improves background rejection beyond what either observable can achieve alone. The experimental apparatus is a compact 10‑gram liquid xenon (LXe) TPC equipped with a variable drift field ranging from 0.1 kV cm⁻¹ to 2 kV cm⁻¹. Nuclear recoils (NR) are generated with a neutron source, while electronic recoils (ER) are produced with γ‑ray sources. Both recoil types emit vacuum‑ultraviolet scintillation at 175 nm that can be described by a fast component (≈ 2 ns decay) and a slow component (≈ 27 ns decay). The relative contribution of the fast component is larger for NR, giving rise to a steeper early rise in the scintillation waveform.

To quantify the early‑time behavior the authors define a “prompt fraction” (PF) as the fraction of the total integrated light that arrives within the first 30 ns of the pulse. PF is measured simultaneously with the ionization signal (S2) and the primary scintillation (S1) for each event, allowing the construction of a two‑dimensional discriminant space (PF versus S2/S1). The data set spans recoil energies from 5 keVₑᵣ up to 100 keVₑᵣ, covering the region of interest for dark‑matter searches.

Key experimental findings include: (1) The PF distribution is essentially independent of the applied electric field, whereas S2/S1 decreases with increasing field due to reduced recombination. (2) At low energies (≤ 10 keVₑᵣ) the PF distributions for NR and ER overlap strongly, limiting discrimination. (3) Above ≈ 20 keVₑᵣ the two‑dimensional cut yields a leakage (ER events misidentified as NR) that is an order of magnitude lower than using PF or S2/S1 alone. For example, at 30 keVₑᵣ a leakage of ~2 × 10⁻³ obtained with S2/S1 alone is reduced to < 1 × 10⁻⁴ when PF is added.

To understand the origin of the PF fluctuations, a detailed Monte Carlo simulation is constructed. The model incorporates the measured photon detection efficiency (~15 %), PMT gain, the statistical (Poisson) nature of photon counting, and the intrinsic spread of the S2/S1 ratio. The simulated PF distributions match the data across all fields and energies, confirming that the dominant source of PF variance is the finite number of detected photons rather than any intrinsic physics variation. Consequently, the authors project that larger detectors, which collect many more photons per event, will experience a substantial reduction in PF statistical uncertainty and thus improved PSD performance.

The paper then extrapolates these results to ton‑scale xenon detectors, which are the current frontier for weakly interacting massive particle (WIMP) searches. In such detectors the typical recoil energy of interest lies below 10 keVₑᵣ, where the PF separation is minimal. Even when combined with S2/S1, the projected leakage remains above 10⁻⁴, indicating that PSD alone cannot provide the level of background rejection required for elastic WIMP searches at low energy. However, the authors point out that for higher‑energy inelastic WIMP scattering (30–100 keVₑᵣ) the PF discrimination becomes robust, offering a complementary handle to identify such events.

The discussion also addresses practical considerations for implementing PSD in large experiments. Increasing the electric field improves S2/S1 but does not affect PF; therefore, field optimization must balance ionization yield against detector stability. Enhancing photon collection—through higher‑reflectivity PTFE walls, larger or more sensitive photomultiplier tubes, or silicon photomultipliers—directly reduces the Poisson term that dominates PF fluctuations. The authors suggest that achieving a photon detection efficiency of 20 % or higher would halve the PF statistical spread, making PSD a more viable tool even at modest energies.

In conclusion, the study provides the first quantitative demonstration that simultaneous use of pulse‑shape information and charge‑to‑light ratio yields superior background rejection in a two‑phase xenon TPC. The Monte Carlo analysis clarifies that the limiting factor for PSD is statistical photon counting, implying that detector designs focused on maximizing light collection will naturally improve PSD performance. While the technique remains weak for low‑energy elastic WIMP recoils, it holds promise for probing higher‑energy inelastic interactions and could be incorporated as an auxiliary discriminant in next‑generation dark‑matter experiments. Future work is recommended to explore advanced optical designs, high‑voltage stability, and the integration of PSD into real‑time event selection pipelines.