D–D Neutron Scattering Kinematics

• D–D neutron scattering kinematics is defined by using nearly monoenergetic 2.45 MeV neutrons from D–D fusion to precisely calibrate nuclear recoils in dual-phase TPCs.
• The technique employs elastic scattering formulas to reconstruct recoil energy, thereby reducing systematic uncertainties and enabling calibration down to sub-keV levels.
• Experimental setups like LUX utilize double-scatter events, time-of-flight measurements, and refined NEST modeling to achieve accurate charge and light yield calibration for dark matter searches.

Deuterium-deuterium (D–D) neutron scattering kinematics underpins the absolute in situ calibration of low-energy nuclear recoils in large dual-phase noble element time projection chambers (TPCs), particularly relevant for dark matter direct detection efforts. The D–D fusion reaction produces nearly monoenergetic 2.45 MeV neutrons, which facilitate precise nuclear recoil energy assignment in the TPC by exploiting the geometry of multiple neutron scatter events. This approach achieves reduced systematic uncertainties and extends calibration sensitivity to sub-keV nuclear recoils, essential for probing weakly interacting massive particle (WIMP) dark matter, as demonstrated in LUX and other leading experiments (Verbus et al., 2016, Collaboration et al., 2016).

1. Theoretical Basis: D–D Fusion and Neutron Scattering Kinematics

The D–D fusion process,

$$^2{\rm H} + ^2{\rm H} \rightarrow ^3{\rm He}\,(0.82\,\mathrm{MeV}) + n\,(2.45\,\mathrm{MeV}),$$

generates neutrons with energy $E_n \approx 2.45\,\mathrm{MeV}$ nearly isotropically in $4\pi$ solid angle. When a neutron of mass $m_n$ with kinetic energy $E_n$ elastically scatters off a nucleus of mass $m_A$ (initially at rest), the recoil energy $E_r$ of the nucleus is determined by two-body kinematics: $$E_{r} = E_n \frac{2 m_n m_A}{(m_n + m_A)^2} [1 - \cos\theta_{\rm lab}],$$ or equivalently, \begin{equation} \label{eq:recoil_vs_angle_simple} E_{r} = E_n \frac{4 m_n m_A}{(m_n + m_A)^2} \sin^2\left( \frac{\theta_{\rm CM}}{2} \right) \approx E_n \frac{4 m_n m_A}{(m_n + m_A)^2} \sin^2\left( \frac{\theta_{\rm lab}}{2} \right), \end{equation} where $\theta_{\rm lab}$ is the neutron's laboratory-frame scattering angle, and $\theta_{\rm CM}$ is the center-of-mass scattering angle. For heavy targets ($m_A \gg m_n$), the approximation $\theta_{\rm lab} \approx \theta_{\rm CM}$ is valid to within a few percent. The relationship between $\theta_{\rm lab}$ and $\theta_{\rm CM}$ is

$$\tan\theta_{\rm lab} = \frac{\sin\theta_{\rm CM}}{(m_n/m_A) + \cos\theta_{\rm CM}}.$$

These relations form the mathematical basis for reconstructing $E_r$ from observed scatter geometries in noble TPCs (Verbus et al., 2016, Collaboration et al., 2016).

2. Derivation of Key Kinematic Formulae

The energy and momentum conservation equations for elastic neutron–nucleus scattering, starting from a stationary target, yield

$$E_n = E_n' + E_r, \qquad E_r = \frac{1}{2} m_A V^2,$$

$$p_n \hat{\mathbf{z}} = p_n' \hat{\mathbf{k}} + m_A \mathbf{V},$$

where $E_n'$ and $p_n'$ are the neutron's post-scatter energy and momentum, and $V$ is the recoil velocity of the target. Eliminating $p_n'$, one obtains the standard recoil energy–scattering angle relationship,

$$E_{r} = E_n \frac{2 m_n m_A}{(m_n + m_A)^2} [1 - \cos\theta_{\rm lab}],$$

consistent with the equations above (Verbus et al., 2016, Collaboration et al., 2016).

3. Experimental Methodology in Dual-Phase Noble TPCs

A collimated D–D neutron beam generated by, for example, the Adelphi DD108 source, is incident on the TPC. Neutrons scatter predominantly twice within the active noble volume. Each interaction produces:

• S1: Prompt scintillation signal (photons)
• S2: Delayed ionization-induced proportional scintillation (electrons drift to the gas phase and are extracted)

Event reconstruction employs:

• 3D event localization: $(x, y)$ from S2 hit pattern on the top PMT array (cm-scale resolution), $z$ from drift time ($\sim$mm resolution).
• Angle extraction: With reconstructed scatter vertices $\mathbf{r}_1$, $\mathbf{r}_2$ and known beam direction $\hat{\mathbf{n}}$,

$$\cos\theta_{\rm lab} = \frac{(\mathbf{r}_2 - \mathbf{r}_1) \cdot \hat{\mathbf{n}}}{\|\mathbf{r}_2 - \mathbf{r}_1\|}.$$

The measured angle is inserted into the kinematic formula to assign absolute $E_r$ to the first vertex, enabling absolute calibration of S1 and S2 yields as a function of $E_r$ in electrons/keV and photons/keV (Verbus et al., 2016, Collaboration et al., 2016).

4. Calibration Implementation and Yield Measurements

For the Large Underground Xenon (LUX) detector, double-scatter D–D neutron events were selected with stringent criteria (raw S2 thresholds, spatial separation, background suppression). Key event-selection cuts include: S2 area $>$ 36 phd, S2 width $<$ 775 ns, site separation $>$ 5 cm, geometric beam-purity volume, and S1/S2 maxima to reject inelastic backgrounds, ultimately yielding over 1,000 double-scatter candidates (Collaboration et al., 2016).

Charge yield ($Q_{y}$, electrons/keV) is computed event-wise with

$$Q_{y} = \frac{S2_{\rm corr}}{g_2 \times E_{\rm nr}},$$

where $S2_{\rm corr}$ is corrected for position and electron lifetime, and $g_2$ is the gain (11.5 $\pm$ 0.9 phd/e$^-$). Binning in $E_r$ provides the following (selected values):

$E_r$ [keV]	$Q_y$ [e$^-$/keV]	$\Delta E_r/E_r$ [\%]
0.7 $\pm$ 0.13	8.2$^{+2.4}_{-2.1}$	$+8$/$-2$
2.0 $\pm$ 0.10	8.0$^{+0.9}_{-0.6}$	$+2$/$-1.3$
24.2 $\pm$ 0.2	4.62$^{+0.13}_{-0.10}$	$+0.4$/$-1.0$
74 (endpoint)	3.06$^{+0.05}_{-0.06}$ (stat) $^{+0.2}_{-0.4}$ (sys)	—

Light yield ($L_{y}$, photons/keV) is similarly determined via single-scatter events, mapped by S2-binned S1-shape fitting and NEST MC modeling with measured $Q_{y}$. Results include:

$E_r$ [keV]	$L_y$ [$\gamma$/keV]	$\Delta E_r/E_r$ [\%]
1.08 $\pm$ 0.13	4.9$^{+1.2}_{-1.0}$	19
10.9 $\pm$ 0.3	8.1$^{+0.4}_{-0.5}$	2
74 (endpoint)	14.0$^{+0.3}_{-0.5}$ (stat),$^{+1.1}_{-2.7}$ (sys)	—

Systematic uncertainties on $Q_y$ and $L_y$ derive from extraction efficiency (8\%), gain calibrations (7.8\% for $g_2$, 3.5\% for $g_1$), S2 and S1 3D corrections, non-uniform electric field, and neutron energy ($\pm 2$\%). These direct, event-by-event calibrations enable refined NEST parameterizations for nuclear recoil response (Collaboration et al., 2016).

5. Time-of-Flight Characterization and Neutron Energy Resolution

To ensure calibration integrity, the neutron energy spectrum is characterized via time-of-flight (ToF) using an NaI(Tl) "tag" and a BC501A "stop" detector: $$E_n = \frac{1}{2} m_n v^2, \quad v = \frac{d}{\mathrm{ToF}},$$ with corrections for energy lost in the NaI(Tl) (via the factor $\zeta$). The fitted neutron ToF peak with a "Crystal Ball" function yields a mean ToF $\bar{t} \approx 148$ ns over $d = 3.09$ m, corresponding to $\bar{E}_n \approx 2.40$ MeV with $\sigma_E / E \approx 4\%$ (orientation A), validating the D–D source for this application. Systematic uncertainty from detector placements and acceptance is subdominant at $\sim$2.5\% (Verbus et al., 2016).

6. Systematic Uncertainties and In Situ Advantage

Traditional ex situ calibrations are susceptible to neutron backgrounds from passive material, ambiguous single-phase detector topology, and angular acceptance biases, necessitating Monte Carlo corrections and leading to a floating energy scale. The in situ D–D double-scatter technique, employing a collimated beam, water shielding, and self-shielding from the TPC, achieves

• 95% event acceptance from neutrons within 6% of initial $E_n$,
• angular reconstruction to $\mathcal{O}(1^\circ)$ (1–3 cm in $(x, y)$, sub-mm in $z$),
• resulting in $\sigma_{E_r} / E_r < 5\%$ down to $\sim$1 keV$_{\rm nr}$ for xenon.

Short-pulse neutron generator operation provides a clean time-zero and suppresses accidental backgrounds by over 99\%, extending calibration down to S2-only (charge) yields at $\sim$100 eV$_{\rm nr}$. Further shortening to sub-$\mu$s bunches allows event-by-event ToF tagging for a broader energy spectrum, enhancing calibration range. This approach dramatically lowers systematic uncertainty compared to earlier methodologies (Verbus et al., 2016).

7. Theoretical Modeling and Impact on Dark Matter Searches

Experimental data from these calibrations directly inform the NEST (Noble Element Simulation Technique) framework. In LUX, fits to the combined charge and light yields as a function of $E_r$ use Lindhard-based (with $k = 0.1735 \pm 0.0060$) and Bezrukov-based (Ziegler stopping, $\alpha = 2.31 \pm 0.27$) parameterizations, accounting for biexcitonic and Penning effects ($\eta = 13.2 \pm 2.3$). The updated NEST model, incorporating these effects, provides agreement to within a few percent over the full $1$–$74$ keV$_{\rm nr}$ range and is preferred for LUX WIMP analyses, significantly improving sensitivity to low-mass WIMPs and quantifying coherent neutrino-nucleus scattering backgrounds (Collaboration et al., 2016).

---

References:

(Verbus et al., 2016) Proposed low-energy absolute calibration of nuclear recoils in a dual-phase noble element TPC using D-D neutron scattering kinematics (Collaboration et al., 2016) Low-energy (0.7-74 keV) nuclear recoil calibration of the LUX dark matter experiment using D-D neutron scattering kinematics