Scintillation in Liquid Nitrogen

• Liquid nitrogen scintillation is a process where ionizing radiation produces approximately 2 photons/MeV, primarily limited by collisional quenching.
• The experiment used a PTFE cup, high-QE PMT, and BiPo tagging for precise pulse shape and timing analysis across LN and GN conditions.
• Results indicate that the LN scintillation background is negligible in rare event detectors, reducing concerns in dark matter searches compared to GN and LXe.

Liquid nitrogen (LN) is a ubiquitous cryogenic medium employed in nuclear and particle physics experiments for both the direct cooling of sensors and as a shielding material. While gaseous nitrogen (GN) possesses well-established scintillation properties, the characteristics of scintillation in the liquid phase have remained largely unquantified until recent investigation. The interaction of ionizing radiation with LN has implications for rare event searches, including dark matter experiments, where even faint backgrounds may become significant. Quantitative measurement of scintillation yield in LN demonstrates a substantially reduced light output relative to both GN and conventional noble liquid scintillators.

1. Experimental Approach and Instrumentation

The paper of LN scintillation was conducted using a detector system designed for low-background measurement and optimized for sensitivity to ultraviolet photons at 337 nm. The active volume was a polytetrafluoroethylene (PTFE) cup chosen for its high ultraviolet reflectivity, optically coupled to a Hamamatsu R11065 photomultiplier tube (PMT) characterized by high quantum efficiency (QE) at the relevant emission wavelength. The assembly was placed within a cryostat, enabling precise control over temperature and pressure.

Alpha decay events, known to produce distinct scintillation signatures in gases, were generated via the introduction of gaseous ^220Rn (thoron) into the chamber. In situ ^212Po decays were tagged with the preceding ^212Bi beta decay, enabling coincidence-time identification ("BiPo tagging") of event pairs and discrimination against background and surface radioluminescence. Measurements were performed sequentially in room temperature GN (RT-GN), cold GN, LN, and vacuum conditions to enable systematic background subtraction.

2. Scintillation Signal Characterization and Measurement Protocol

Detection of scintillation in LN relied on rigorous pulse shape and timing analysis. For each candidate event, the PMT waveform was interrogated for double-pulse structures within a predefined temporal window Δt, corresponding to sequential ^212Bi and ^212Po decays. The photoelectron count associated with the alpha decay pulse (Q₁) provided the primary measurement variable.

Background correction was achieved by acquiring datasets with the chamber evacuated, quantifying radioluminescence intrinsic to PTFE and fused silica detector components. The LN scintillation yield was determined relative to the GN yield, using identical detector geometry and optical response, normalized for QE and corrected for the pressure and temperature dependence of GN scintillation following established literature (AIRFLY collaboration).

3. Quantitative Results: Yield and Comparative Metrics

The experiment revealed a markedly faint scintillation signal from LN: alpha decay events produced only an excess of approximately 5 photoelectrons, contrasted with ∼200 photoelectrons measured in GN under equivalent conditions.

The relative scintillation yield is formalized as:

$$\frac{Y_{\text{LN}}}{Y^{\text{STP}}_{\text{GN}}} = \frac{\mu_{\text{LN}}/QE_{\text{LN}}}{\mu_{\text{GN}}/QE_{\text{GN}}} \cdot \Theta(p,T)$$

where $\mu$ is the mean photoelectron yield per event, $QE$ is PMT quantum efficiency at each phase, and $\Theta(p,T)$ corrects for GN ambient conditions.

Empirically, the measured yield ratio was:

$$Y_{\text{LN}} / Y^{\text{STP}}_{\text{GN}} = 0.014 \pm 0.003$$

Absolute yield was calculated using the canonical GN alpha scintillation benchmark of $125$ photons/MeV:

$$Y_{\text{LN}} = 0.014 \times 125\ \text{photons/MeV} = 1.8 \pm 0.5\ \text{photons/MeV}$$

Accounting for systematics, this result is reported as:

$Y_{\text{LN}} = 2 \pm 0.5\ \text{photons/MeV}$

A comparison with GN and conventional noble liquid scintillators such as liquid xenon (LXe) is shown below:

Medium	Scintillation yield (alphas; photons/MeV)
GN (STP)	∼125
LN	2 ± 0.5
LXe	35,000–70,000

LN scintillation yield is thus approximately 100-fold lower than GN and 10,000-fold lower than noble liquids.

4. Physical Mechanism: Collisional Quenching

The remarkably low yield observed in LN is attributed to collisional quenching inherent to the dense liquid phase. In LN, the spatial proximity of nitrogen molecules (~0.8 g/cm³) facilitates rapid non-radiative energy transfer following excitations by ionizing particles. This process competes strongly with photon emission, suppressing the observable scintillation. In GN, the reduced density mitigates such quenching, allowing more frequent radiative de-excitation.

5. Implications for Low-Background and Rare Event Detectors

The faintness of LN scintillation indicates that direct backgrounds from this process are negligible in most rare-event search contexts, including dark matter detection. Even highly sensitive, low-noise photon detectors registered only a small number of photons from high-energy alpha interactions in LN. Backgrounds resulting from radioluminescence of detector surfaces (PTFE, fused silica) are of comparable or smaller magnitude and can be effectively rejected using time-coincidence techniques.

Nonetheless, in experiments employing extremely sensitive arrays—such as CCDs in the Oscura dark matter search—even a sparse photon flux may trigger spurious signals. This necessitates proactive mitigation, for example, by coating the CCDs with aluminum to render them optically opaque to the scintillation photons.

6. Limitations and Prospects for Further Study

Current results assume the scintillation emission spectrum of LN matches that of GN (centered at 337 nm); further spectral measurements would confirm this presumption. The demonstration of measurable, though extremely faint, LN scintillation constitutes the first quantitative result for this medium. The experimental methodology using BiPo tagging and direct photoelectron counting in a controlled cryogenic environment sets a benchmark for subsequent studies.

A plausible implication is that future detector designs employing LN for cooling or shielding need not implement elaborate optical baffle systems solely to block LN scintillation light, provided the target background rates do not exceed the faint yield measured. However, for applications requiring ultra-low optical backgrounds, additional characterization may be necessary.

7. Summary and Outlook

Scintillation in liquid nitrogen occurs at an absolute yield of $2 \pm 0.5$ photons/MeV for alpha particles, equivalent to $1.4\%$ of the yield in GN. This yield is insignificant compared to noble liquid alternatives and presents minimal concern for most rare event detection backgrounds. The suppression mechanism is identified as collisional quenching due to the dense molecular environment in the liquid phase. This foundational measurement enables more precise modeling and mitigation in low-background experimental efforts and suggests that LN remains advantageous as a cryogen in sensitive detector deployments without introducing significant optical backgrounds via scintillation. Further optical characterization may resolve the emission spectrum with greater fidelity.