2024 International Nuclear Atlantic Conference - INAC 2024

Escola de Guerra Naval, Rio de Janeiro, RJ, May 6 - 10, 2024

Python software for data processing and quality control in

HPGe detectors
BERTACO, G. P.1∗, SILVA, P. S. C.1†, ZAHN, G. S.1 , SILVA, B. F.2 and SEMMLER, R.1
1
Instituto de Pesquisas Energéticas e Nucleares (IPEN / CNEN -SP)
Av. Prof. Lineu Prestes, 2242, Cidade Universitária 05508-000 São Paulo, SP
2
Center of Nuclear Energy in Agriculture-CENA USP, Piracicaba, SP

1. Introduction

Python is an extremely popular high-level programming language that can be used for data
analysis, scientific programming, and many other purposes. This popularity is due to the read-
ability of Python code. Many of its expressions and functions are very similar to the English
language, which greatly facilitates learning, understanding and executing programs and codes.
The language aims at high productivity and a high degree of readability, in addition to adding
excellent libraries and resources.

Radiation detectors made of High Purity Germa- nium (HPGe) are in the front line of fundamen-
tal research since their first development, mainly used as gamma-ray detector [1].

2. Methodology

2.1. Get Calibration Measurements

Daily calibration measurements are performed by the user places an uncalibrated detector-
specific stacked 57+60 Co source at the specified position and performs a 600 s real-time acquisi-
tion. Spectra are analyzed using suitable software [2] (called CAX) that calculates peak position,
resolution, area (in counts per second of real-time acquisition) and uncertainty[4].

2.1.1. Old Method

The program saves this verification data in a .csv file which, using the previous method, was
manually transcribed onto a paper spreadsheet, along with the date and time of the measurement
and who carried out this verification, which were later typed into an electronic spreadsheet in
order to store these data.

2.1.2. Method GPB

The program written in python named Gamma Precision Balance (GPB), takes this .csv file,
imports/extracts its necessary data, and using repetition structures it finds the determined peak
of the 2 elements with a margin of variation that is commented in section 2.2. When it finds
∗
[email protected]
†
[email protected]
 Bertaco, G. P. et al.

a value between these limiting parameters, it takes that value, which in this case is the energy,
extracts the position of this data in the variable, with this position it takes the other relevant
data in the other variables, such as the resolution.

The collected data are stored in the same digital spreadsheet used by the old method, but the
first difference and advantage is that it is done automatically and in an extremely short period
of time. After storing the calibration data for this day of 57+60 Co source measured , the program
imports/extracts all the calibration data for that particular month present in this worksheet to
perform the quality control plots, on Appendix Pictures 5, with certain parameters, section 2.2,
which are later automatically saved as a .png file in the directory.

And other benefits will be discussed later in the section 2.3.

The GPB was divided into two parts, in figure 1), in purple and pink, they are the parameters
and the blue part, which is the final plot of the quality control graph, which are executed through
the subprocess module in Python which allows you to execute other parts of the code.

Figure 1: GPB Method

2.2. Quality Parameters

For the Energy subplot, the acceptable limits were added (green and red lines, that is, the
maximum and minimum parameters where the calibration is considered good, and a variation
of 2 keV up and down was used, that is 120.06 and 124.06 for 57 Co and 1330.5 and 1334.5 for
60 Co.

In the subplots of the resolution, it is possible to notice a blue line, which represents the average
of the resolution of the previous month, using this average the acceptable limits are plotted, of
which a variation of 20% was used for 57 Co and 30% for 60 Co.

These subplots are very important since they show the existence of factors that can affect the
efficiency and/or resolution of HPG detectors, and include problems in the detector’s crystal
(such as defects induced by neutrons or contact migration, for example), in the assembly of the
detector (loss of vacuum is usually the dominant factor) or in the associated electronics[3].

2.3. Errors

The computer spreadsheets, one for each detector, were first checked for obvious typing errors,
as problems with dates and/or the decimal separator; in this process some of the data had to be
discarded as the results were completely incompatible with the whole, indicating some form of

2
 Bertaco, G. P. et al.

unidentified experimental mistake[4], errors that in the chart plotting process can cause errors
of the ”ValueError” type, which indicates an incompatibility between the amount of data on the
X and Y axes, which occurs when the data that were imported from these spreadsheets have
sizes/quantity different from the expected.

It is also possible to comment on the TypeError, which occurs when we have an incompatibility
of data types, which basically originates when it contains a word instead of a numerical value,
such as observations about the equipment, which prevents the graph from being generated. By
automating the old process, 2.1.1, these errors can be eliminated.

3. Results and Discussion

The quality control chart tests carried out, in addition to allowing the testing of the code for
quality control, showed the efficiency of the equipment over the years, allowing the identification
of days on which the calibration carried out was outside the ideal usage parameters or even.

In Figure 2 we can see a large dispersion of data for the resolution of Co57 . Looking at Figure 4
possible to see the energy rising and the resolution falling in the image at the end of the month,
when the equipment vacuum was redone, leading to an increase in energy and a significant drop
in resolution. A good calibration will maintain a certain constancy in the data as seen in Figure
5.

4. Conclusions

Automation ends up bringing benefits, considering that many users of detectors do not calibrate
because of the old process and precisely Automating eliminates human errors like ValueError
and TypeError that were described in the 2.3 section.

In addition, the GPB quality control in a visual way, taking the time to analyze and even some
prior knowledge, considering that it is not a fixed parameter, but rather a constancy of the
data. In addition, it avoids calibration if there are errors or unusual parameters using this data
constancy, thus providing data of greater credibility and confidence.

5. Appendix Pictures

Here you can find the images mentioned in this article

Figure 2: April 2009. Figure 3: January 2017.

3
 Bertaco, G. P. et al.

Figure 4: January 2018. Figure 5: September 2018.

Appendix extra information

The detector using at IPEN’s Neutron Activation Analysis Laboratory, which used the data to
carry out the tests, was manufactured by Canberra Industries with a nominal resolution of 2.0
keV.

The Open Source code discussed in this document can already be found in the repository avail-
able on GitHub and further information is available upon reasonable request to the corresponding
author.

Acknowledgments

Instituto de Pesquisas Energéticas e Nucleares-IPEN

Comissão Nacional de Energia Nuclear- CNEN
Conselho Nacional de Desenvolvimento Cientı́fico e Tecnológico-CNPq
The LATEX format has been prepared by Professor Helio Pedro Amaral Souto, Instituto Politécnico
(IPRJ), Universidade do Estado do Rio de Janeiro (UERJ), Brazil.

References

[1] Nikolay Abrosimov, M Czupalla, N Dropka, J Fischer, A Gybin, K Irmscher, J Janicskó-

Csáthy, U Juda, S Kayser, W Miller, et al. Technology development of high purity germanium
crystals for radiation detectors. Journal of Crystal Growth, 532:125396, 2020.

[2] Canberra. Genie-2000 Spectroscopy System – operations manual, 1999.

[3] Glenn F Knoll. Radiation detection and measurement. John Wiley & Sons, 2010.

[4] GS Zahn, FA Genezini, RB Ticianelli, and M Saiki. Long-term performance assessment of

hpge detectors used in the neutron activation analysis laboratory of ipen-cnen/sp (brazil).
Applied Radiation and Isotopes, 125:108–112, 2017.