Choose timezone
Your profile timezone:
Powered by Indico
v3.3.5
Online meeting. Please join at the following Zoom link.
https://anu.zoom.us/j/83173169338?pwd=azBJWXN0bU42MGlwazVBSy85S0I5Zz09
Introduction and plan for the meeting
Overview of the Sheffield CYGNUS program
Japan/Kobe program overview
Italian/CYGNO program overview
The MIGDAL project and applications to directional DM and neutrino searches
Hawaii program overview
Australian CYGNUS program overview
I will discuss a few theoretical issues of relevance to directional detection. Firstly, how confident are we that there is a DM wind pointing back towards Cygnus? Then, I will explain what the “neutrino fog” is - a term increasingly being used in place of “neutrino floor” by the community. I will then describe how using the neutrino fog as a way to rank techniques for directional detection naturally leads to gas TPCs as the optimal strategy.
Molecular sieve (MS) is expected to remove impurities such as radon from gases that require low radioactivity in dark matter search detectors. For practical use, it is necessary to reduce the amount of radioactive impurities released from itself to the utmost limit. In this talk, I report on the development of zeolite with radon adsorption capacity by ion exchange of low-radioactive MS.
A new molecular sieve-based gas recycling system is presented that provides for simultaneous removal of both radon and common impurities from SF6:CF4:He gases in TPCs, hence minimising the total amount of gas required. Removal of internally-produced radon and associated progeny is important for background suppression whilst removal of outgassing and leaked-in contaminants such as water, oxygen and nitrogen is required to suppress capture of interaction-produced electrons which causes gain suppression. The system utilises a Vacuum Swing Adsorption (VSA) technique, allowing continuous long-term operation. Studies are presented of a new low radioactive molecular sieve, developed for this work and found to emanate radon up to 98% less per radon captured than commercial material.
After the detection of the first light of a radiation source using C/N-1.0, we solved various remaining issues; e.g. gas leak, spark and detector readout problems. We will introduce about these update and also report the development of our circular system.
The basis for Garfield ++ simulation for negative ion gas was developed last year. Two types of detach models (Threshold model & Cross-section model) were prepared and we simulate the gas gain for each. We confirmed that the measured gain in GEM and the results of simulation are consistent.
We present the status of ongoing studies comparing the electron background rejection performance of multivariate combinations of physically motivated observables tuned to optimize electron rejection performance with a convolutional neural network (CNN) event classifier trained directly on reconstructed 3D charge distributions of recoil events. Using samples of simulated charge distributions containing O(1e7) electron recoils and O(2e5) fluorine recoils after 25cm of drift in an 80:10:10 mixture of He:CF4:CHF3 at 60 torr, binned into 100um^3 voxels and ranging in energies between 0.5 and 10.5 keVee, we first show that combining nine predefined observables using either a boosted decision tree (BDT) or a feed forward neural network (NN) to classify recoils in this detector, leads to between a 1.5 and 5 fold increase in the number of rejected electrons as a function of energy at 50% F-recoil efficiency compared to a previously published multivariate combination of these same observables that doesn't use machine learning. We further show that training a CNN with fluorine and electron recoil charge distributions re-binned into a 32 x 32 x 32 voxel grid leads to a noticeable improvement in electron rejection over both the BDT and NN combinations of predefined observables at nearly all energies. Finally, we briefly summarize the status of extending this work to measurements from miniature TPCs with pixel ASIC readout.
Discussion of the Snowmass process and potential contributions to the white paper