Studying a particle beam
PHYS291 project


The goal

The aim of this project is to study a beam of carbon-ions detected using
the ALPIDE detector. The data is a part of a bigger project with the goal
of using protons for imaging the human anatomy.
Carbon-ions are used here so that the results can be compared with protons,
since they are both heavy ions, and as a measure to reduce scattering.
For a more detailed explanation of the project see [1]

The detection

The ALPIDE chip is divided into 1024 columns and 512 rows that are linked.
The ALPIDE was shot from the side which means that the particles would remain
in the epitaxial layer of the detector for many columns, with the longest tracks
in order of 700 columns.
The source data is divided by 4960 frames, each 485 micro seconds long.
All the hits the detector has registered during each frame are organised by columns
and rows.

Paticle tracks

The data is visualized in histograms with all the source data unchanged, but
later two attempts at removing some of the noise from the particle tracks was made.

Data structure

The source file (availible in [3]) has the following structure:
TTree frames
Figure 1 - shows the tree (Frames) structure in the root-file
The second line in each box is the data type.

where frames is a root TTree object with ru_id, frame_id, stave_id,
chip_id, abs_time, bunch_counter, column and row.
ru_id, stave_id and chip_id are the same for all the data, while column and row
contain the number of hits each column/row in the ALPIDE detector has detected.

Visualization of data

It is possible to show the unaltered data from the detector in a 2D histogram
such as the one below.

Histogram 1
Figure 2 - 2D histogram made by track.C (availible in [3]), which shows the hits from exactly one frame

The source of the data is the root file 151218_04237.root in [3]

A histogram for each frame has and can be made using the program track.C.
track.C is the main program that reads the root file, sorts the frames by number
of entries, from the largest to smallest.
It shows each frame of the original data, noise reduced and smoothened data and
distribution of tracks up to that frame. An example of each one of these histograms
is shown below.
The images are availible in [3].

Isolation of tracks

As an attempt to isolate the track, first the isolated pixels in each line of the x direction was removed.
As explained in the background section, the particles are expected
to remain in the detector, and therefore give a signal, for many
consecutive columns. The noise is then the signal without neighbors.
In the y direction, in an attempt to make the counting of tracks easier, the average value of the pixels before
was calculated and if it was more than a certain value, the current one was assumed to be a part of the track.
And therefore it was filled.

Here is an example of how a frame can look like when the main program, track.C is run.
Histogram x
Figure 3 - The same frame as the last image, but the data along the x-axis are noise reduced

Histogram y
Figure 4 - The same frame as the last figure, after noise reduction in x direction and smoothening along the y direction
Here the registered tracked are drawn over the histogram.
The average width of the registered tracks is shown in the title of the histogram.

Distribution of tracks
Figure 4 - Distribution of tracks in the 100 frames with the largest number of entries after smoothening in y

Results and discussion

The results show many from zero to close to 10 tracks in each frame.
The tracks are mostly in the middle rows. As it can be seen from figure 4.
The width of the tracks varies between 2 and 8 when registered correctly,
but the average value is usually between 3 and 4 pixels or rows.

Most of the tracks make it about half way through the columns,
but the maximum number of tracks are registered at the highest columns in the histograms.
That is to say along y=900.
The carbon-ions don't usually follow a straight path and they give a much stronger signal
near the end of the track, usually around y=500. This is probably caused by the fact that
heavy particles, such as carbon-ions, lose most of their kinetic energy right before they are stopped
according to the Bethe-Bloch equation (see [4]).

There were many misregistrations of tracks along y=500 in some frames. This is probably due to
a lot of noise at and near the column 400.
The cause of this is the electronic equipment used, and no special measures were taken to remove
or reduce these after the attempts at noise reduction in x and smoothening in y direction..

Last update 09/06/2021 - Alfred Abbasi