Activation of Mn, Li₂O and LiF in the JSI TRIGA reactor to study potential tritium production monitors for fusion applications

In future fusion reactors, such as ITER or DEMO, based on the fusion reaction of deuterium and tritium nuclei (the D–T reaction), the required quantities of tritium fuel will be produced by bombardment of lithium nuclei with neutrons. Several types of special tritium breeder modules (TBMs) will be installed in the ITER reactor to demonstrate self-sufficiency of tritium production. Experiments will be necessary to confirm adequate rate of tritium production in the reactor. A direct method of measuring the rate of neutron-induced tritium production consists of the irradiation of LiF and subsequent determination of the tritium concentration in the material.

LiF pellets commercially produced as thermoluminescent detectors (TLDs) can be used to measure tritium production. In the scope of the fusion for energy (F4E) project of the European Commission we proposed to investigate, as an alternative, the potential use of manganese as a detector for monitoring tritium production in fusion machines. Important similarities between the sensitivity profiles of the tritium production reaction in ⁶Li and those of the ⁵⁵Mn(n,γ)⁵⁶Mn reaction in the TBMs were observed [1–4], suggesting that the latter reaction could be used to monitor tritium production, at least in the short term (the half-life of ⁵⁶Mn being 2.579 h). However, validation and, possibly, improvements of the ⁵⁵Mn cross-sections are needed in order to meet the required accuracy in some energy ranges. The target accuracy in determination of the tritium production rate that is relevant for ITER is 5% (1 – σ). Figure 1 displays reaction cross-sections for the ⁶Li(n,t)⁴He, ⁵⁵Mn(n,γ)⁵⁶Mn and ¹⁹⁷Au(n,γ)¹⁹⁸Au reactions. Figure 1 also suggests that the epi-thermal contribution to the ⁵⁵Mn(n,γ)⁵⁶Mn reaction has to be excluded, and this can be done by comparing the measurements with/without Cd.

Zoom In Zoom Out Reset image size

Differences in the cross-section evaluations for the ⁵⁵Mn(n,γ)⁵⁶Mn reaction were observed, which resulted in differences in the calculated reaction rates. Calculated-to-experimental (C/E) ratios for the Frascati Neutron Generator (FNG)-bulk shield [5], FNG-tungsten (W) [6, 7] and FNG-helium cooled lithium–lead (HCLL) [3] benchmarks demonstrate that the differences between the results using IRDFF [8] and IRDF-2002 [9] dosimetry nuclear data libraries are up to 5%, up to 10% and around 18%, respectively. Relatively good C/E agreement was found in the FNG-bulk shield experiment which is most sensitive in the 100–1000 eV energy range where the response function uncertainties are low (~5%). On the other hand, large discrepancies between the measured and calculated reaction rates were found for the FNG-W [6, 7] benchmark, in which the ⁵⁵Mn(n,γ)⁵⁶Mn reaction rates are highly sensitive to the response function above 1 keV, and in the FNG-HCLL benchmark. However, some systematic errors may be present in the last two benchmarks. In the FNG-W benchmark, the exact Mn content in the foils needs to be verified and in the FNG-HCLL benchmark, heterogeneity of ⁶Li enrichment in the Li–Pb block was observed. Although the better consistency observed between the ⁵⁵Mn(n,γ)⁵⁶Mn and ¹⁹⁷Au(n,γ)¹⁹⁸Au results seems to indicate better performance of the new data, it is at present still difficult to confirm with certainty the progress achieved.

In the scope of the fusion for energy (F4E) project of the European Commission, a first series of measurements was performed at the JSI TRIGA [10] reactor in 2014 [11]. Aluminium foils containing 1% manganese (Al–1%Mn) or 0.1% gold (Al–0.1%Au) were irradiated in the JSI TRIGA thermal reactor, together with LiF (TLD) and LiPb samples, with the principal objective of studying the energy response of the ⁵⁵Mn(n,γ)⁵⁶Mn reaction as well as the correlation between the ⁵⁵Mn(n,γ)⁵⁶Mn reaction rates and the tritium production rates in LiF and LiPb. Some inconsistencies were observed between the measured and calculated tritium activity for LiF (TLD) and LiPb samples which were encapsulated in Pyrex glass vials; these were probably caused by the presence of boron in the glass [11].

In 2017, a better defined set of experiments was performed at the JSI TRIGA reactor in which foils of certified reference materials Al–1%Mn and Al–0.1%Au and LiF and Li₂O samples were irradiated. The samples were irradiated in different TRIGA irradiation channels, i.e. in the central channel (CC), the pneumatic tube (PT) in the core periphery (F24 position in the outer 'F' ring of the reactor core), in position F19 and in the IC40 irradiation channel in the graphite reflector. Irradiations with different neutron spectra provide complementary information for data validation. All the irradiations were performed at a reactor power of 25 kW. Figure 2 displays the JSI TRIGA reactor core configuration used for the experiments (core no. 220).

Zoom In Zoom Out Reset image size

Reaction rates for the ⁵⁵Mn(n,γ)⁵⁶Mn and ¹⁹⁷Au(n,γ)¹⁹⁸Au reactions were measured using Al–1%Mn and Al–0.1%Au foils, respectively. The Al–1%Mn foils were prepared by pressing Al–1%Mn material in wire form; the foil masses were approximately 10 mg and their diameter was approximately 5 mm. The Al–0.1%Au foils were punched out of Al–0.1%Au material in foil form; the foil masses were approximately 6 mg and their diameter approximately 5 mm. Tritium production measurements were performed using two different types of sample, LiF and Li₂O, developed and produced in the Institute of Nuclear Physics of the Polish Academy of Science in Krakow.

LiF samples consisted of three TLD pellets wrapped in Al foil. The masses of the single pellets were approximately 34 mg. The Li₂O samples were in powder form, encapsulated in specially designed Al cylindrical containers and closed with a screw cap. The masses of the Li₂O samples were approximately 120 mg each.

A set of one LiF sample (consisting of three TLD pellets), one Al container with Li₂O, one Al–1%Mn foil and one Al–0.1%Au foil were packed together into a cylindrical container, 10 mm in diameter and 22 mm high, with a wall thickness of 1 mm. The containers were either made from Al for bare irradiations or from Cd for Cd-covered irradiations. The containers were inserted into dedicated 3D-printed tubes made from polyethylene. Eight such sets of samples were irradiated in the TRIGA reactor. Figure 3 displays a schematic drawing of one set of samples packaged in the containers and a 3D-printed tube, and photographs of the samples, containers and tubes.

Zoom In Zoom Out Reset image size

2.1. Activation rate measurements in Al–1%Mn and Al–0.1%Au

2.2. Measurements of tritium production in LiF and Li₂O

Calculations of the ⁵⁵Mn(n,γ)⁵⁶Mn, ¹⁹⁷Au(n,γ)¹⁹⁸Au and ⁶Li(n,t) ⁴He reaction rates were performed using the Monte Carlo particle transport code MCNP6.1 [14] in conjunction with the ENDF/B-VII.1 nuclear data library [15] and a detailed computational model of the JSI TRIGA reactor. In the calculations, the core configuration and the control rod positions were reproduced and the irradiation capsules, covers and samples themselves were modelled explicitly. Figure 4 displays the calculated neutron spectra in the Li₂O material in the capsules located in the irradiation channels.

Zoom In Zoom Out Reset image size

The reactor power and control rod positions were kept constant during the irradiations; however, some small perturbation of the neutron flux and the reactor power due to the insertion of Cd covering the capsules was recorded. These detailed power variations have not yet been taken into account in the present calculations. The MCNP calculations resulted in a considerable overestimation of the effective multiplication factor (around 1.05). This is because we did not take temperature feedback and fuel burnup effects explicitly into account; this should be investigated further. However, this overestimation is compensated for in the normalization of the calculated values with regard to the reactor power level [16] according to equation (1):

$$\begin{align} \newcommand{\e}{{\rm e}} \displaystyle {{R}_{{\rm abs}}}={{R}_{{\rm MCNP}}}\frac{P\nu }{w{{k}_{{\rm ef\,\!f}}}},\nonumber \end{align} \tag{ 1 }$$

where ${{R}_{{\rm abs}}}$ is the absolute reaction rate value, ${{R}_{{\rm MCNP}}}$ is the raw calculated result, $P$ is the reactor power level during the irradiation, $\nu$ is the average number of neutrons emitted per fission, $w$ is the recoverable energy per fission and ${{k}_{{\rm ef\,\!f}}}$ is the calculated effective multiplication factor. The MCNP statistical uncertainties were in general of the order of a few per cent, and up to around 10% for the Cd-covered results. Spectral effects due to fuel burnup were seen as a probable cause for the reduction in the thermal flux of the order of about 10% in the central channel and the first fuel element ring [17].

The results of the comparison between the calculated and the measured ⁵⁵Mn(n,γ)⁵⁶Mn and ¹⁹⁷Au(n,γ)¹⁹⁸Au activities are presented in figures 5–8 for the positions CC, PT, F19 and IC40, respectively. The reaction rates per target atom were computed from measurements using both HPGe detectors (denoted as 'D1' and 'D2' in figures 5–8), for the measured peak areas of the 846.8 keV, and 1810.7 keV gamma rays for the ⁵⁵Mn(n,γ)⁵⁶Mn reaction and 411.8 keV gamma rays for the ¹⁹⁷Au(n,γ)¹⁹⁸Au reaction.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The displayed C/E uncertainty bars include the experimental uncertainties (most importantly due to uncertainties in the detection efficiency and counting statistics) and computational uncertainties (statistical uncertainties, 5% power normalization uncertainty). For the Al–1%Mn and Al–0.1%Au foils the agreement between the measurements and the calculations is reasonably good, in most cases within the uncertainty bars. Slightly larger differences were observed for the IC-40 location (in the graphite reflector surrounding the reactor core), which are probably due to deficiencies of the computational model (e.g. uncertainties in the graphite reflector composition).

A comparison of the calculated and measured tritium activities expressed as C/E ratios is shown in figure 9. The abscissa denotes the irradiation position and cover. The results indicate a systematic overestimation of the calculations for the LiF (TLD) and Li₂O samples by roughly 20%, and 40%, respectively.

Zoom In Zoom Out Reset image size

Several reasons for the observed overestimation were identified. Preliminary analysis of measured values for LiF (TLD) indicated differences occurring within a single package of three LiF samples. Typically, the TLD pellet in the middle of the package shows systematically smaller activity (around two standard deviations), most likely due to neutron absorption in the outer pellets. However, this effect does not represent an important source of uncertainty or bias in the comparison between the experimental and calculated values since (a) the experimental tritium production rates in LiF are in fact averages over the three LiF samples and (b) in the calculations, tallies for the ⁶Li(n,t)⁴He reaction were defined in the region comprising all three LiF samples.

For the Li₂O samples, the measured ³H activities are subject to an additional systematic error coming from the chemical preparation, and caused by the incomplete recovery of the irradiated Li₂O samples (up to 10% of the initial mass). This explains the outlying measurement points and irregular C/E values for Li₂O (figure 9). Another source of the observed differences in tritium activity in LiF and Li₂O samples is the abundance of ⁶Li in the compounds used, which is known with an uncertainty no better than  ±2%. Deviations from the certified Li abundances are not uncommon and have already been experienced in some previous studies (e.g. [3]). A possibly lower ⁶Li abundance than assumed for natural Li may also contribute to the observed difference. The effects of tritium self-absorption in the irradiated material and tritium escape were not taken into account.

Table 1 summarizes the main sources of uncertainty in the experimental and calculated ⁵⁵Mn(n,γ)⁵⁶Mn, ¹⁹⁷Au(n,γ)¹⁹⁸Au and ⁶Li(n,t)⁴He reaction rates.

In order to study the potential use of the ⁵⁵Mn(n,γ)⁵⁶Mn reaction as a tritium production monitor, a series of foils of certified reference materials Al–1%Mn and Al–0.1%Au as well as LiF (TLD) and Li₂O samples were irradiated in different irradiation channels in the JSI TRIGA research reactor, both bare and under Cd. The irradiations were performed in different neutron spectra, i.e. in the CC, the PT in position F24 in the outer 'F' ring of the reactor core, in position F19 and in the IC-40 irradiation channel in the graphite reflector. A detailed MCNP6.1 computational model was prepared including the irradiated samples, containers, covers and tubes modelled explicitly.

The experimental and calculated results and their associated uncertainties were studied. Good consistency between the experimental and calculated values was observed for the ⁵⁵Mn(n,γ)⁵⁶Mn and ¹⁹⁷Au(n,γ)¹⁹⁸Au reaction rates, in most cases within their respective 1 – σ uncertainties. The experimental and calculated values agreed to within 10% for the CC, PT and F19 locations (except ¹⁹⁷Au(n,γ)¹⁹⁸Au in the F19 location) and to within 20% for the IC40 location.

In the case of LiF (TLD) and Li₂O samples, the model calculations overestimate the measurements results of tritium activity by roughly 20%, and 40%, respectively. Systematically lower experimental results obtained for both materials can be attributed in part to incomplete recovery of the irradiated material from the capsule (up to ~10%), abundance and/or isotopic inhomogeneity of ⁶Li.

The difficulty of sufficiently accurate radiometric assessment of tritium activity produced in the reaction with ⁶Li is a major reason for the search for monitors that can measure tritium production by easier procedures. Gamma spectrometric measurements present a promising alternative, and the ⁵⁵Mn(n,γ)⁵⁶Mn reaction is a suitable candidate for monitoring neutron flux at the energies relevant for tritium production. It was demonstrated in this paper that, although not originally planned, Mn foil irradiation can serve to verify and validate other direct measurements of tritium production.

Further studies and laboratory experiments leading to improved assessment of correlation (with reduced uncertainty) between induced tritium activity in ⁶Li and neutron fluence measured by use of the ⁵⁵Mn(n,γ)⁵⁶Mn reaction can deliver a practical and easy-to-use tool for measuring tritium in the TBM (ITER) project.

The work leading to this publication has been funded partially by Fusion for Energy under the specific grant agreement F4E-FPA-395-01. This publication reflects the views only of the authors, and Fusion for Energy cannot be held responsible for any use which may be made of the information contained therein.