[Paper Review] Crystal Defects: A Portal To Dark Matter Detection

Fedja Kadribasic, N. Mirabolfathi+2 more|arXiv (Cornell University)|Feb 10, 2020

Dark Matter and Cosmic Phenomena4 citations

TL;DR

This paper proposes using defect creation energy loss in crystal detectors as a statistical signature to distinguish dark matter nuclear recoils from electron recoil backgrounds in low-threshold experiments. By simulating defect formation via density functional theory and molecular dynamics, the authors show that carbon-based detectors exhibit strong, resolvable spectral features due to energy loss, making diamond a highly promising material for next-generation dark matter searches with sub-10 eV thresholds.

ABSTRACT

We propose to use the defect creation energy loss in commonly used high energy physics solid state detectors as a tool to statistically identify dark matter signal from background. We simulate the energy loss in the process of defect creation using density functional theory and molecular dynamics methods and calculate the corresponding expected dark matter spectra. We show that in phonon-mediated solid state detectors, the energy loss due to defect creation convolved with the expected dark matter interaction signal results in a significant change in the expected spectra for common detector materials. With recent progress towards $\sim$10 eV threshold low-mass dark matter searches, this variation in expected dark matter spectrum can be used as a direct signature of dark matter interactions with atomic nuclei.

Motivation & Objective

• To develop a new method for discriminating dark matter nuclear recoils from electron recoil backgrounds in low-threshold solid-state detectors.
• To quantify the energy loss due to defect creation in crystal lattices as a function of recoil energy and direction.
• To evaluate the detectability of defect-induced spectral features in common detector materials like carbon, silicon, and germanium.
• To identify the most promising detector materials for realizing this signature in next-generation dark matter experiments.
• To assess the potential of this method to enhance sensitivity to sub-10 GeV/c² dark matter candidates.

Proposed method

• Simulating nuclear recoil events in diamond-cubic crystals using molecular dynamics with classical potentials (Stillinger-Weber for Si/Ge, Tersoff-Nordlund and Erhart for C), calibrated against density functional theory (DFT) calculations.
• Initializing 4096-atom unit cells thermalized to 40 mK, then displacing a central atom with energies from 1 to 200 eV in multiple directions to probe defect formation thresholds.
• Measuring the difference between initial and final system potential energy after 10 ps relaxation to isolate the energy lost to defect creation.
• Calculating the differential recoil rate with and without energy loss effects, using a 3 eV resolution and 10 eV threshold to model realistic detector response.
• Applying a normalized root-mean-square (RMS) statistic to quantify the distinguishability of the defect-induced spectral features from the background.
• Comparing results across four models (C-Erhart, C-Tersoff-Nordlund, Ge, Si) to assess material-dependent sensitivity and spectral resolution.

Experimental results

Research questions

• RQ1 Can defect creation energy loss produce a statistically resolvable spectral feature in the recoil energy spectrum of solid-state dark matter detectors?
• RQ2 How does the defect creation energy loss vary with recoil energy and direction in elemental semiconductors like C, Si, and Ge?
• RQ3 Which detector material—carbon, silicon, or germanium—exhibits the strongest and most resolvable defect-induced spectral features for low-mass dark matter detection?
• RQ4 To what extent can the normalized RMS statistic quantify the detectability of defect signatures above the noise floor in a realistic detector with 3 eV resolution and 10 eV threshold?
• RQ5 Can the location and shape of defect-induced spectral peaks be used to infer the mass of a dark matter particle?

Key findings

• Carbon-based detectors exhibit the strongest defect creation energy loss features, with signal strength far exceeding those in silicon and germanium, regardless of the potential model used.
• The defect creation energy loss leads to a measurable spectral shift in the recoil energy distribution, particularly above the defect threshold, which is most prominent in carbon.
• For a detector with 3 eV resolution and 10 eV threshold, the normalized RMS statistic shows that the defect signature is most distinguishable for dark matter masses between 1 and 10 GeV/c².
• Spectral features in germanium become increasingly smeared and approach a linear regime above 100 eV, while carbon shows sharp, resolvable features even at low energies.
• The integrated dark matter detection rate increases when energy loss is included, due to the redistribution of energy into defect formation rather than phonons.
• Diamond detectors are identified as the most promising candidate for realizing this method experimentally, due to their strong, resolvable spectral signatures and favorable material properties.