HOME > Latest News >Topics > April 2017

Development of software towards Hyper-Kamiokande (HK)

Development of software is absolutely needed to investigate physics potential of HK. Software plays important role to determine detailed design of the detector. For example, it is planned that HK uses about 40,000 photo sensors. There are some candidates of photo sensor, and we can measure their properties at laboratory one by one. However, it is difficult to imagine how different in neutrino observation caused by different photo sensor candidates with 40,000 installed in the detector. In such a case, we often use a method called “simulation” in which a detector is virtually built on computer and examine physics phenomena.

In this virtual experiment, there are two simulations: simulation for physics process, like how neutrino interacts with matter and produces particles, and simulation of detector which denotes how particles are detected and transformed to signal. As for physics simulation, we can use programs used in SK or T2K because physics phenomena are universal. We have to establish the detector simulation by our own including tank size and shape, optical property of water, performance of photo sensor, readout of signal, and so on. We (core members from Duke university, US) are developing water Cherenkov detector simulator for wide use based on GEANT4, a software package of CERN. This simulator can easily change tank size, number of photo sensor, and type of photo sensor and read out by modularize them. Figure 1 shows an event sample of HK simulation.

Figure 1; Event display of HK simulation in which muon is emitted from center of the tank toward side wall. Yellow dots correspond to photo sensor which detect Cherenkov light.

To understand detector performance, software which reconstructs physics variables from information of photo sensors is also needed. The conventional method used in SK determines physics variables one by one sequentially. We are developing a new method which determines all physics variables at once by fitting expected distributions to photo sensor information. The new method uses information of un-hitted photo sensors which are not used in the conventional method, and we can expect to reconstruct physics variables more precisely.

Figure 2 show momentum resolution calculated by the new method for electron and muon which are generated by HK simulation with various momentum. The new photo sensor for HK has two times better efficiency for detecting one photon than SK and if the photo sensors cover 40 % of the detector wall as SK, this figure shows the momentum resolution improves much better than SK.

Figure 2; Momentum (horizontal axis) vs momentum resolution (vertical) for electron (left) and muon (right) reconstructed by the new method. Green corresponds to SK performance, red is HK tank which new photo sensors cover 14 % of detector wall, blue is 40 % coverage case.