The effect of the aging of liquid organic scintillators used for gamma-neutron separation

Since the beginning of using liquid scintillators for gamma-neutron separation, there have been many articles dealing with long-term degradation especially due to oxygen presented during scintillator encapsulation. The effect of aging of liquid organic scintillators namely EJ301, EJ309 (both Eljen Technology), and new custom-made cocktails based on 1-Phenyl-3-(2,4,6-trimethylphenyl)-2pyrazoline and 2,5-Bis(5-tert-butyl-benzoxazol-2yl)thiophene fluors were investigated for more than half a year. The research was focused on the Compton edge shifting of gamma particles since the position is proportional to the light yield of the selected scintillator. Furthermore, the gamma-neutron separation was observed and quantified using FOM (Figure Of Merit) for samples prepared and filled under normal and nitrogen atmosphere during the mentioned period. All stated parameters of liquid scintillator NE213 manufactured by Nuclear Enterprises Limited opened more than three decades ago were measured and used for comparison.


I. INTRODUCTION
UENCHING due to dissolved oxygen is an important effect and must be avoided to obtain optimum scintillation efficiency. The explanation is given, that the quenching at aromatic molecules of the singlet state leads to the creation of a triplet state. The ground state of the oxygen molecule is a triplet and the next state is a singlet lying about 0.98 eV over the ground state [1]. Therefore, oxygen molecules surrounding the aromatic molecules can absorb the energy of the singlet state of aromatic molecules and make a spin allowed transition to the triplet state causing a decrease of the fluorescence. Such a transition only occurs in aromatic molecules, which have an energy gap of S1-T1 greater than 0.98 eV [2].
The magnitude of the quenching factor depends on the oxygen solubility in the solvent, which is determined by the external pressure and the temperature, and on the effect of oxygen quenching on the solvent and fluor and possibly other presented compounds (e.g. emulsifiers). At low concentrations, transfer quenching is predominant, but at increased concentrations, it decreases more rapidly than solute quenching. It appears likely that, at practical concentrations, near-ideal concentration for a given system, oxygen quenching of the fluor emission is the main factor.
Recently, few attempts were made to evaluate the basic performances of 18 years old scintillation cocktail. It has been shown that the 18-year storage has less impact on the basic performances (including pulse shape analysis) than batch-tobatch variability in cocktail compositions [3].
This study was focused on the evaluation of the long-term basic characteristics, such as Compton edge stability and g-n PSD, of the EJ301, EJ309, and different custom-made di-isopropylnaphthalene (DIPN) based liquid scintillation cocktails encapsulated under the nitrogen and air atmosphere.
Encapsulation of the liquid scintillators was performed using 20 mL glass vials with a low concentration of K-40 (Wheaton, UK) and the Teflon tape was used as a reflector.
An encapsulated stilbene crystal, with a diameter of 45 mm and a length of 45 mm is used for comparison.
The Cs-137 radionuclide was used for energy calibration in the same way as in [4]. The Cf-252 radionuclide was used to

The effect of the aging of liquid organic scintillators used for gamma-neutron separation
evaluate the neutron-gamma separation. The gamma-neutron discrimination was performed using NGA-01 [6], which is the two-parameter spectrometric system and was used in earlier works [4,7,8].

B. Procedures Preparation of scintillation cocktails
The volume of 14 mL of the selected liquid scintillator was poured inside the 20 ml glass vials with an upper lid containing a Teflon seal to ensure gas tightness and to better reflect photons to the photomultiplier tube (PMT). The glass vial was wrapped with Teflon tape to reflect photons. Two batches of previously mentioned scintillators were prepared this way; the first batch, labeled "N", was prepared under a nitrogen atmosphere and each liquid scintillator was bubbled with nitrogen for several minutes. After the selected period, the vial was properly closed. The second batch was prepared under an ambient atmosphere with no bubbling and was labeled as "A". The batches during the whole observation period were stored in a dry, clean, and dark place with temperatures varying from 19°C to 24 °C.

Measurement setup
The selected scintillator was placed on an RCA 8575 PMT and enclosed in a light-tight box. The radiation source was placed on the holder which was about 4 cm far from the scintillator horizontally. The PMT's anode output was connected to NGA-01. The high voltage of the PMT was set to 1650 V for each experiment. This should guarantee the fixed position of the Compton edge.
The Cs-137 was used for energy calibration, due to the creation of the Compton edge (CE) at 0.477 MeV. This energy belongs to a certain channel, which was determined using the derivative (differentiation) method. Due to the calibration, the term MeV electron equivalent (MeVee) could be introduced to quantify the light yield. The energy calibration was based on a linear function fit of Cs-137 measured data and zero energy channel and zero energy [4].
Gamma-neutron separation was measured using the same setup as was described in the paragraph dealing with the energy calibration, but instead of Cs-137, the neutron source Cf-252 was used.
To guarantee satisfying results, more than a million pulses were acquired in each experiment. The count rate was in the range of from 150 up to 300 cps for both mentioned radiation sources.
III. RESULTS AND DISCUSSION The long-term observation of mentioned parameters started on the 30 th of September 2020 and ended on the 26 th of May 2021. Every month in this interval, the measurements of these two batches were performed using the same adjustment as described above.

A. Light yield
Since the magnitude of light collected at the photocathode of the PMT is proportional to the height of the electrical pulse, the light yield can be measured indirectly using the position of the Compton edge. As was stated in the introduction, the oxygen prevents excitation of the luminophore, thus causing quenching. Therefore, there should be some difference in the position of CE between samples prepared under the nitrogen and ambient atmosphere.
Since the diameter of the used PMT was 51 mm, the variance in CE position of the used scintillators was performed. This variance of the CE was caused due to differently placed vials on the face of the PMT, different orientations of the vial, and different states of optical contact. It was observed, that the deviation was up to 7% for all scintillators based on 10 repetitions for each cocktail.
As can be seen in Table I, the differences in the position of the CE of samples prepared under nitrogen and ambient atmosphere on the first day corresponded with those previously published [9,10], and the difference in quenching, i.e. position of CE between samples prepared under nitrogen/ambient atmosphere and lower and higher (optimal concentration C O) is visible. Furthermore, the values of the standard deviation of the CE means of investigated scintillation cocktails and concentrations expressed in percentage are within the range of variation, when the samples are changed. The obtained data support previously published data, that the dissolved oxygen causes only quenching at some degree (based on the composition of the solvent-solute system), as was stated above. HV of PMT was set to 1650 V. N and A stand for nitrogen and air atmosphere respectively.

B. Gamma-neutron discrimination
In addition to measuring the stability of the CE, the efficiency of gamma-neutron discrimination was also observed. To characterize the discrimination of two peaks originated from gamma and neutron radiation, the Figure of Merit (FOM) was used in the same way as in [4]. It was observed (Fig. 1) that the discrimination, especially at low energies, was improving with elapsing time. This improvement was noticeable for all cocktails in the energy range up to 0.4 MeVee. The highest improvement was observed in cocktails marked THIO. Since this luminophore is very problematic to dissolve in DIPN solvent, it is believed, that after the dissolution process, undissolved microscopic particles remained in the cocktail and they were gradually dissolving during the first weeks. On the other hand, the most stable cocktails were based on DIPN solvent and PPO and PYR luminophore prepared under a nitrogen atmosphere. The influence of the nitrogen atmosphere was especially noticeable for the xylene + PPO and DIPN + THIO systems. Systems composed of DIPN + PPO and DIPN + PYR did not show clear differences.

C. Comparison with NE213
To verify and compare our results, the NE213 scintillator was measured. This scintillator has a similar composition as EJ301, which means it is based on xylene solvent with PPO as a luminophore. This bottle was opened in 1986 meaning that the effect of oxygen deterioration should be visible after more than 35 years.
To perform energy calibration, the Compton edge of Cs-137 had to be investigated. However, the data obtained when NE213 was measured was completely different and as can be seen in Fig. 2, the shape or even the evidence of CE was almost missing. The position of the CE was estimated to be at the 180 th channel at HV=1750 V. Despite the worse energy threshold for good discrimination, the gamma-neutron separation was still possible even after 35 years from the opening of the bottle and exposure to the cocktail to atmospheric oxygen. As can be seen (Fig. 3) in comparison with other scintillation cocktails (same geometry of measurement) the light output was lower and the energy resolution was bad.

IV. CONCLUSION
This study aimed to evaluate the long-term performance of different scintillation cocktails prepared under nitrogen and ambient atmosphere. Especially, the focus was aimed to observe the stability of the Compton edge due to the light yield and, ability to discriminate between gamma rays and neutrons.
The effect of nitrogen and ambient atmosphere on CE position was following literature, where the higher shift in the position of the CE was observed at low concentration of luminophore and for higher concentration, the quenching was negligible in comparison of the variance in the position of CE due to sample changing, which was 7 %. It was proved, that the position of the CE in time is not stable with a negative slope at almost all cocktails meaning, that the light yield was constantly decreasing.
To compare observed results, the 35 years old and opened NE213 scintillation cocktail was measured and evaluated. It was demonstrated, that the cocktail didn't exhibit any visible Compton edge and therefore the energy calibration was very difficult to perform. Despite this difficulty, the cocktail was able to separate pulses originated from gamma rays and neutrons. The explanation of this worsening probably lies in chemical changes in the phosphor molecules caused by oxygen (e.g. oxidation) rather than in simple interrupting energy transfer from solvent to luminophore.