TL;DR
The Light Dark Matter eXperiment (LDMX) is upgrading its electron beam energy from 4 GeV to 8 GeV to enhance its sensitivity to sub-GeV dark matter. This paper demonstrates through detailed simulations that the existing background rejection strategies, developed for the lower energy, are highly effective even at the higher energy, successfully suppressing challenging photon-induced backgrounds. This upgrade promises a significant increase in LDMX's reach into unexplored dark matter parameter space.
What Problem Does This Paper Address?
The search for dark matter (DM) is a major focus in particle physics. A compelling hypothesis is that DM consists of thermal relics from the early universe. For DM particles in the MeV to GeV mass range, this requires a new, light mediator particle to explain their observed abundance without overproduction. The Light Dark Matter eXperiment (LDMX) aims to detect such light dark matter by observing the signature of a deflected electron losing most of its energy and gaining significant transverse momentum after interacting with a target.
A primary challenge for LDMX is distinguishing this signal from various background processes, particularly photon-induced (PN) reactions. These backgrounds can mimic the signal signature, especially when they produce a single neutral hadron or a single charged particle that decays in-flight, carrying away energy invisibly. LDMX is currently designed to operate with a 4 GeV electron beam, and this paper investigates the impact of upgrading to an 8 GeV beam energy on the effectiveness of its background rejection techniques, specifically focusing on PN backgrounds. The goal is to ensure that the established veto methodology remains robust at the higher energy.
Methods
The study employs detailed GEANT4-based simulations to model the detector response and background processes. The core of the analysis involves developing and evaluating a sophisticated veto methodology to distinguish signal events from background. This methodology includes:
- Trigger and Kinematic Selection: An initial selection based on missing energy, requiring a significant energy loss by the electron.
- ECal Shower Shape Veto: This is a crucial component that leverages the high granularity of the electromagnetic calorimeter (ECal). A Boosted Decision Tree (BDT) classifier is trained using various topological features of energy deposits within the ECal. These features capture differences in shower shapes between signal (a clean electron shower) and background (more complex showers from PN reactions).
- HCal Activity Veto: The hadronic calorimeter (HCal) is used as a veto to reject events with significant energy deposition, which are typically associated with hadronic components of PN reactions that might escape ECal rejection.
- MIP Tracking: A fine-grained tracking algorithm is used to identify and reject rare backgrounds that might produce minimum ionizing particles (MIPs) or short tracks, which are not characteristic of the signal.
The paper compares the performance of these vetoes at both 4 GeV and 8 GeV beam energies, using simulated samples for various dark matter mediator masses.
Datasets
The study utilizes simulated datasets generated using GEANT4. These datasets include:
- Signal Events: Simulated dark photon signal events for various dark photon mediator masses ($m_{A'}$).
- Background Events: Simulated events for four main background classes:
- ECal photo-nuclear reactions
- ECal photon conversion to a $\mu^+\mu^-$ pair
- Target photo-nuclear reactions
- Target photon conversion to a $\mu^+\mu^-$ pair
These simulations are based on a customized version of GEANT4 and incorporate specific models for dark photon interactions. The paper also references previous simulation studies performed for the 4 GeV beam energy, providing a baseline for comparison.
Key Insights and Ideas
- Increased Beam Energy Benefits Signal and Background: The 8 GeV beam energy not only increases the expected signal yield but also, importantly, tends to produce more energetic and higher-multiplicity final states for PN backgrounds, making them more detectable and thus easier to reject.
- BDT Robustness: The BDT classifier, trained on features derived from the ECal's high granularity, proves to be a powerful tool for discriminating between signal and PN backgrounds. Its performance remains strong at 8 GeV, demonstrating that the learned topological features are still highly discriminative.
- Complementarity of Vetoes: The combination of ECal shower shape analysis (BDT), HCal activity, and MIP tracking provides a multi-layered defense against various background types. Each veto step is designed to catch different classes of events, and their combined effect is crucial for achieving the desired background suppression.
- HCal Veto Effectiveness: The HCal veto is particularly effective at rejecting PN backgrounds, especially those originating from the target, which might produce single neutrons or other low-energy particles that are hard to detect in the ECal.
- MIP Tracking for Rare Events: The MIP tracking algorithm is essential for vetoing very rare background topologies that might slip through the ECal and HCal vetoes, such as those involving soft muons or decaying kaons.
Key Results
- The existing veto methodology, developed for 4 GeV, successfully rejects photon-induced backgrounds at 8 GeV for integrated luminosities of at least $2 \times 10^{14}$ electrons on target (EoT).
- The fraction of ECal PN events remaining after the BDT veto at 8 GeV is approximately a factor of three smaller than at 4 GeV, while maintaining the same signal efficiency.
- For a sample scaled to $2 \times 10^{14}$ EoT, the full veto sequence (trigger, ECal BDT, HCal veto, MIP tracking) reduces the background events to less than one for both photo-nuclear and muon conversion processes.
- The study demonstrates that no new limiting background event types become relevant at 8 GeV compared to the 4 GeV scenario.
Limitations and Future Work
- Simulation-Based Study: The results are based on detailed GEANT4 simulations. While these simulations are comprehensive, experimental validation will be crucial.
- Specific Dark Matter Model: The analysis focuses on a benchmark dark photon model. While this model captures key features of light dark matter interactions, the sensitivity to other dark matter models might vary.
- Future Work: The paper implicitly suggests that the successful validation of the 8 GeV veto strategy paves the way for the planned data-taking phase, which aims for $4 \times 10^{14}$ EoT. Further experimental studies will be needed to confirm these simulation results and to fully exploit the enhanced sensitivity at 8 GeV. The paper also mentions that the detector design has broader potential for measurements with electron beams in the forward region.
Summary updated