Gain characterization of LGAD sensors with beta particles and 28-MeV protons

TL;DR

This work extends LGAD characterization to non-MIP beams by comparing 28-MeV protons with $^{90}$Sr beta particles using TCAD simulations and a controlled experimental setup. By measuring pulse amplitude, area, FWHM, and rise time across bias voltages, the authors quantify gain with two independent methods and demonstrate a pronounced gain suppression for 28-MeV protons relative to beta beams. Simulations reveal how electronics shaping affects the observed parameters and corroborate the qualitative trend of reduced multiplication for highly ionizing deposits, while highlighting systematic differences between amplitude- and area-based gain estimates. The results provide guidance for deploying LGAD-based sensors in non-MIP contexts (biology, medical physics, etc.) where charge deposition patterns differ from MIPs. Overall, the study finds that proton-induced signals are significantly larger yet exhibit comparable timing characteristics, with a robust gain suppression effect that must be accounted for in detector design and interpretation.

Abstract

Low Gain Avalanche Diodes, also known as LGADs, are widely considered for fast-timing applications in high energy physics, nuclear physics, space science, medical imaging, and precision measurements of rare processes. Such devices are silicon-based and feature an intrinsic gain due to a $p{^+}$-doped layer that allows the production of a controlled avalanche of carriers, with multiplication on the order of 10-100. This technology can provide time resolution on the order of 20-30 ps, and variants of this technology can provide precision tracking too. The characterization of LGAD performance has so far primarily been focused on the interaction of minimum ionizing particles for high energy and nuclear physics applications. This article expands the study of LGAD performance to highly-ionizing particles, such as 28-MeV protons, which are relevant for several future scientific applications, e.g. in biology and medical physics, among others. These studies were performed with a beam of 28-MeV protons from a tandem Van de Graaff accelerator at Brookhaven National Laboratory and beta particles from a $^{90}{\rm Sr}$ source; these were used to characterize the response and the gain of an LGAD as a function of bias voltage and collected charge. The experimental results are also compared to TCAD simulations.

Figures (18)

• Figure 1: A diode (top-left) and the LGAD (top-right) sensors fabricated by HPK. The photos also shows the wire-bonds connected to the pads and the guard-rings. The FNAL readout board (bottom) is shown with sensors mounted and wire-bonded. (One additional sensor is mounted on the board, but it was found not to be well functioning and was not used for this study.)
• Figure 2: The tandem Van de Graaff (left) and the devices under test inside the vacuum chamber (right).
• Figure 3: Examples of typical pulses generated by Diode-1 in a beam of 28-MeV protons (a), and by the LGAD in a beam of 28-MeV protons (b) and beta-particles (c) at different values of bias voltage.
• Figure 4: Area-normalized distributions of amplitude (a), area (b), FWHM (c) and rise time (d) for Diode-1 in a beam of 28-MeV protons as functions of bias voltage.
• Figure 5: Most probable values of amplitude (a), pulse area (b), FWHM (c) and rise time (d) as functions of the bias voltage with the 28-MeV proton beam for Diode-1. The most probable values and their uncertainties are extracted from Gaussian fits in limited ranges around the peaks of the distributions.