Neutron activation of stable isotopes in soil and groundwater from a radionuclide production facility, South Africa

In the realm of nuclear research facilities and nuclear technologies, the production of radionuclides stands as a crucial endeavour, providing essential tools for medical, industrial, and scientific purposes. However, the inherent risks associated with radionuclide production necessitate a comprehensive understanding of their environmental impact, particularly in the context of neutron irradiation activity. Neutron activation, a phenomenon where stable isotopes absorb neutrons and transform into radioactive isotopes, can significantly alter the isotopic composition of surroundings, such as soil and groundwater. This can be seen in the findings of Pantelias & Volmert (2015) where activation in the concrete walls was produced by the transformation of elements that originally existed as stable isotopes into radionuclides by neutron irradiation during the operation of a nuclear power plant.

The iThemba Laboratory for Accelerator-Based Sciences (iThemba LABS) intends to construct a facility to produce a Low-Energy Radioactive Ion Beam (LERIB). The radioactive ion beam (RIB) from the LERIB facility will be used for nuclear physics research. The Separated Sector Cyclotron (SSC) will be a driver for the isotope separation on-line (ISOL) facility. Beam current up to 50 μA of a 66 MeV proton beam will be delivered by the SSC to produce radioactive beams from a target ion source. The LERIB will be the first stage of the production of radioactive beams at the iThemba LABS. Only low-energy radioactive ion beams will be available from the facility. Stage two will be the post-acceleration of the radioactive beams to energies of 4 to 5 MeV per nucleon with a linear accelerator.

The LERIB project is anticipated to generate significant sources of ionising radiation including neutron radiation. The neutron radiation is associated with the ability to activate matter at which it traverses might pose significant risks to the soil and groundwater underlying the LERIB facility. A study by Mantengu (2016), using a programme called FLUktuierende KAscade (FLUKA) to calculate the optimal shielding design for the LERIBs project at iThemba LABS predicted that the subsurface at a depth of 2–3 m bgl might be prone to activation when containment is insufficient. Therefore, radioactive isotopes resulting from neutron activation underneath the LERIB facility can be expected due to the ability of neutrons to penetrate any matter they are interacting with.

The induced radioactivity in the subsurface can pose potential environmental risks, particularly in the underlying aquifer. The fluctuating water table, during wet seasons (winter), rises close to the surface and might increase the chances for groundwater to interact with neutrons. The groundwater flow might act as a conduit for the migration of the induced radionuclides, resulting in contamination of groundwater bodies on site and down gradient. Activation of the subsurface will deteriorate the groundwater quality and pose negative effects on groundwater users on site and in communities adjacent to iThemba LABS. For these reasons, it might not only be the radiation-producing facility (iTL) that might face contamination risks associated with ionising radiation but also the nearby surrounding communities. Hence, it is important to sufficiently contain and shield ionising radiation to ensure the safety and protection of the public and the environment.

The Witteberg sandstones are feldspar-poor, but the finer-grained lithologies contain alkali-feldspar and are very micaceous (Fourie et al. 2011). Moreover, Rust (1973) as cited in (Fourie et al. 2011) suggested that the TMG is characterised by thick packages of medium- to coarse-grained quartz-arenites, conglomerates and subordinate fine-grained lithologies. Whereas on the other side, the boundary of the Bokkeveld Group with the TMG is represented by a marine-flooding surface with quartz-arenites, but conformably grading into shelf and pro-delta shales and mudstones (Theron 1972 as cited in Fourie et al. 2011). The Witteberg Group, which conformably overlies the Bokkeveld Group, comprises quartz-rich sandstones and mica-rich mudstones with intercalated siltstones.

The Cederberg Formation has a maximum thickness of 120 m (thin formation) but remarkably continuous unit, consisting of black silty shale at the bottom, grading into brownish siltstone and fine sandstone at the top. It is exposed within the southwestern Cape but continues along the whole length of the southern Cape Fold Belt. The Nardouw Subgroup, with its three subdivisions, the Goudini, Skurweberg and Rietvlei (Baviaanskloof in the Eastern Cape) Formations, is another thick (maximum 1,200 m) unit of sandstone that varies between quartz arenite, silty and feldspathic arenites, accompanied by some very minor inter-bedded conglomerate and shale.

The variation in lithological diversity, textural, grain size and bedding thickness differences of these formations lead to pronounced differences in weathering, structural and hydrogeological characteristics (Duah & Xu 2010). Topping the Cape Supergroup are the recent coastal Cenozoic age deposits which consist of quartzose sands, gravel, clay, lignite, calcareous aeolianites, bioclastic-silicilastic aeolianites. According to the Group A Vegter classification, 1995, the Coastal Cenozoic aquifers of South Africa are all classified as primary aquifers.

The primary Coastal Cenozoic aquifers include the Sandveld group which is divided into four main hydrogeological units (Meyer 2002) as follows:

• (i) Cape Flats Unit – extending from False Bay to Melkbosstrand.

• (ii) Silwerstroom-Witzand is situated in the Atlantis area.

• (iii) Grootwater Unit is situated in the Yzerfontein area.

• (iv) Berg River Unit situated in the Saldanha area:

• • (a) Adamboerskraal Aquifer Unit

  • (b) Langebaan Road Aquifer Unit

  • (c) Elandsfontyn Aquifer Unit

The study area taps from the Cape Flats unit and is associated with groundwater levels generally shallow (2–5 mbgl) throughout and are also vulnerable to contamination (Conrad 2004). Groundwater within the Cape Flats Aquifer flows in a westerly direction towards Zeekoevlei and a southerly direction towards Monwabisi/Mnandi. The groundwater of the Silwerstroom-Witzand Unit, also known as the Atlantis Aquifer System, flows westwards to south-westwards where it discharges along the coast in areas where the aquifer dips below sea level. The Grootwater Unit basically consists of two aquifer systems, an upper and lower aquifer system. The aquifer systems of the Berg River Unit consist of two aquifers, separated by a clay layer. Borehole yields within the Sandveld Group range between 0.1 and 5 l/s with most boreholes yielding 0.5 l/s or less and about 30% yielding 2 l/s and more (Meyer 2002). Below the primary unconsolidated aquifer, the secondary hard rock aquifer is found and consists of Malmesbury shales and intrusive granites (Saayman & Adams 2002).

Groundwater samples were collected using 50-ml centrifuge tubes. The 50-ml tubes were rinsed in the laboratory using deionised water prior to the collection of groundwater samples to avoid sample cross-contamination. The collected groundwater samples were stored in a temperature-controlled cooler box to avoid exposure to the external environment such as the temperature. The samples were labelled according to borehole numbers, i.e. B1A, B2A, and B3A.

The soil samples were collected at a depth of 2–3 and 3–4 m from the recovered lithological core of B1A, B2A and B3A. The layer at which soil samples were collected is the interface of the subsurface where the shielding material is anticipated to be located with a thickness of 2–4 m stratum. The disposable plastic spoons were used to collect and fill samples into 50 ml centrifuge tubes. A total of 12 soil samples were collected and marked according to the borehole IDs and depths of collection such as B1A (2–3 m), B1A (3–4 m). The Environmental Radioactive Laboratory (ERL) based at iTL was used to store samples at a controlled room temperature of 20 °C until the borehole drilling and sample collection were completed.

The soil and groundwater samples were analysed for background radiation counts for a period of 24 h using the High-Purity Germanium (HPGe) detector (Figure 2). The background radiation measurements were carried out at the iThemba LABS in the Environmental Radiation Laboratory under low-radiation background conditions achieved by shielding the detector (using lead) from ambient gamma radiation. The nuclear electronics were used to process pulses from the detector, through a multi-channel analyser (MCA) and binned in a gamma-ray spectrum (Figure 3).

The protons at 66 MeV energy incident on a Beryllium (⁹Be) target (target current 25 μA) generate a spectrum of neutron energies between 0 and 66 MeV. The bombardment of ⁹Be by protons can provide abundant neutrons which then extend the probability of activating matter in the receiving environment (Osipenko et al. 2013). According to Hughes (1957), neutrons of extremely wide energy range such as 0–66 MeV can interact effectively with nuclei. Moreover, in terms of neutron energy range, a study by Yashima et al. (2004) generated neutrons from a lithium target bombarded by proton energies of 30, 35, 40, 50, 60 and 70 MeV and revealed that neutrons generated from these energies effectively interact with atomic nuclei. Thus, the neutron energy of 0–66 MeV range is deemed adequate to be used in activating samples for the purpose of this study. Samples were placed on a 10 cm thick lucite block serving as a backscatter medium, covered with a 12 mm thick polycaprolactam (nylon 6) build-up material.

The soil and groundwater samples activation were quantified by means of gamma-ray spectroscopy, where directly after irradiation the collective sample dose rate exceeded 1 mSv/h. The neutron activation analysis using gamma-ray spectroscopy instrumentation could not be performed immediately after activation, as the detector dead-time exceeded 90%. Data acquisition for the first soil sample could only be performed 30 min post-activation.

Abundant radioisotopes with short half-lives decayed over the 30-min period to levels quantifiable by the detector and analysing equipment. Gamma-ray spectrum acquisition time was set at 1-h detector live time per sample. The sample acquisition time for samples analysed at time points 24, 72 and 168 h post bombardment was 2 h each to ensure good counting statistics for the low count rates. To distinguish between isotopes resulting from neutron activation and naturally occurring stable isotopes in the study area, the detected isotopes from the background radiation were subtracted from the activated samples. This was to dissociate the current background radiation from the anticipated neutron activity that might be induced by the LERIB project in the subsurface environment.

Weir 2004 argued that ionizing radiation can be mutagenic, and exposure increases the risk of cancer. The gamma radiation, which the activated isotopes in the study area emit when decaying, is associated with penetrating ability and can cause severe cell damage and mutagenesis even when the source is not taken internally (Campbell 2009). It is for this reason that the proper shielding of neutron radiation is recommended.

When comparing the ²⁴Na counts from both soil and groundwater samples, it was observed that the soil samples contained high concentrations compared to groundwater samples (Figure 6). The variation in concentrations can be attributed to the fact that water being a product of two hydrogens and one oxygen to form H₂O, with ¹H from H₂O considered to be a good neutron moderator might have retarded neutrons and lowered the energy to reduce the radiation activity (Bacchi et al. 2002).

Based on the presence of ²⁴Na from the activated soil and groundwater samples, it can be argued that the soil and groundwater samples contained a stable isotope of ²³Na which when interacting with the anticipated neutron flux from the LERIBS facility will form the radioactive ²⁴Na. According to Firestone (1996), ²⁴Na has a half-life of 15 h. The ²⁴Na peaks in groundwater samples decreased drastically with time, as a result when measuring the B3A water sample, the concentration decreased from >1 count per second (Cps) (sample B1A and B2A water) to <1 Cps (Figure 6). The measured groundwater velocity suggests that ²⁴Na will undergo full decay before migrating a distance exiting the facility of iTL. Therefore, minimum environmental and health effects from ²⁴Na within the study area and down gradient can be expected.

The spectra analysed 1-week post activation revealed the presence of isotopes with longer half-lives (Figure 8). The radionuclides included manganese-54(⁵⁴Mn), manganese-52(⁵²Mn), sodium-22(²²Na) and scandium-46(⁴⁶Sc) (Figure 8). The radionuclides were only measured from soil samples.

The isotope ⁵⁴Mn was measured at gamma energy 834 keV and can be formed through activating a stable isotope of ⁵⁴Fe, by a nuclear reaction ⁵⁴Fe(n,p)⁵⁴Mn (Gritzay et al. 2002) (Figure 8). Iron is among the abundant major elements in the earth's crust and consists of numerous natural isotopes such as ⁵⁶Fe, ⁵⁷Fe, and ⁵⁸Fe. It can be noted that the iron isotopes will undergo activation at different conditions. The evidence is in the fact that only ⁵⁴Fe underwent activation when bombarded with neutron flux. It is therefore expected that, among the iron isotopes, only ⁵⁴Fe will be prone to activation once the anticipated neutrons from the LERIBs interact with the soil in the subsurface. The ⁵⁴Mn has a half-life of 312.3 days (Firestone 1996) and has a likelihood of exiting the study.

The isotope of ⁵²Mn was measured at 744, 935 and 1,434 keV gamma energies (Figure 7). Through neutron activation, ⁵²Mn is formed by a nuclear reaction ⁵⁴Fe (n, t)⁵²Mn (Sublet et al. 1990). A study showed that bombarding a ⁹Be target with 40 MeV (for 2.0 h of irradiation) produced neutrons that activated ⁵⁴Fe to form ⁵²Mn in soil. The findings of the study suggest that soil samples containing ⁵²Fe might be activated when interacting with neutrons to form a radioactive isotope ⁵²Mn with tritium being released in the reaction (Sublet et al. 1990). For this reason, the formation of ⁵²Mn due to neutron radiation from the LERIBs facility can be expected in soil material (Figure 8). The ⁵²Mn has a half-life of 5.6 days (Firestone 1996). According to the measured groundwater velocity, ⁵²Mn will undergo full decay before migrating a distance to any of the boreholes down the gradient of the LERIB facility.

The unstable isotope, ²²Na was measured from soil samples at energy 1,274 keV (Figure 8). The radioactive ²²Na, through the neutron activation process, is formed by activating the stable isotope ²³Na following a nuclear reaction ²³ Na(n, 2n)²²Na and has a half-life of 2.6 years (Firestone 1996; Gritzay et al. 2002). Due to the high dissolution nature of sodium, it is anticipated that ²²Na might dissolve with water and migrate at the speed of groundwater. The half-life of ²²Na suggests that the isotope will migrate a distance exiting the study area.

The soil samples were measured to contain a radionuclide of scandium – 46(⁴⁶Sc) at gamma energies 889 and 1,120 keV (Figure 8). The ⁴⁶Sc is produced by activating a stable isotope of ⁴⁶Ti with a nuclear reaction ⁴⁶Ti (n, p)⁴⁶Sc and has a half-life of 83.8 days (Gritzay et al. 2002; Firestone 1996). The formation of radioactive ⁴⁶Sc indicates the presence of a titanium element in the soil samples. Titanium has 5 naturally occurring isotopes (Audi & Wapstra 1993). The formation of ⁴⁶Sc distinguishes which of the titanium isotopes is present in the study area and prone to neutron activation. The results suggest that the neutrons from the envisioned LERIBs demonstrator might activate a stable isotope of ⁴⁶Ti to form ⁴⁶Sc.

The soil and groundwater containing radioactive material from iThemba LABS can be eroded and migrate to the nearby Kuilsriver and Eersterivier basins. A study by Valković (2019) suggests that water serves as the major transport medium of the U-series nuclides. The measured short half-life and long half-life radionuclides in the study area can be expected to be transported through migration in groundwater as is the case with the U-series nuclides. The physical soil properties such as porosity and permeability might allow activated matter to seep through and travel distances with flowing groundwater. The groundwater flow can be explained using Darcy's law. Darcy's law states that the velocity is directly proportional to the permeability of the aquifer and the hydraulic gradient (Todd et al. 1976). Typical groundwater flow velocities fall in the range of 1.5 m per year to 1.5 m a day depending on the permeability of the aquifer (Todd et al. 1976).

Darcy's Law was used to describe the flow of groundwater flow as follows:

The following step was to determine the flow velocity of groundwater in the area. The groundwater velocity will assist in anticipating the distance of the activated matter likely to migrate in the subsurface. A geological report conducted by Kantey & Templer 2008 suggested that the study area is composed of coarse sand material. According to the order of magnitude of K (hydraulic conductivity) for different kinds of rocks (Bouwer 1978), the coarse sand material has a K with an order of 2 × 10¹–10²m/day.

Based on the calculations using Equation (4), it can be concluded that the velocity of groundwater in the study area is around 0.8 m/day. The measured average distance of the monitoring boreholes located down gradient from the envisioned LERIB facility was measured to be about 167 m. Based on these measurements, it is anticipated that the radionuclide formed in water, which is ²⁴Na, will not migrate a distance exiting the study area. This is because ²⁴Na has a half-life of 15 h and therefore will decay before exiting the study area. Therefore, the facility should only put measures to protect the groundwater from irradiation within the site of iThemba LABS in compliance with section 24 of the Constitution and the National Environmental Management Act 108 of 1998. All radionuclides measured 1-week post activation were within the background radiation concentrations.

The soil and groundwater samples were activated in the radiotherapy vault at iThemba LABS with neutrons at an energy of 0–66 MeV. The HPGe detector was used to measure the gamma spectroscopy of the activated isotopes in the soil and groundwater samples. The results revealed that some of the stable isotopes will undergo activation when interacting with neutron radiation from the envisioned LERIB facility. However, only the isotopes of major elements were activated from both soil and groundwater samples. The activated isotopes in soil samples included ²²Na, ²⁴Na ⁵²Mn, ⁵⁴Mn, and ⁴⁶Sc. Ony ²⁴Na occurred in groundwater.

It is therefore anticipated that isotopes of major elements in the subsurface are prone to activation from which short and long-lived radionuclides can be expected. The coarse sand material in the subsurface reduces the hydraulic flux resulting in slow groundwater velocity of 0.8 m/day. Based on the groundwater velocity, it is anticipated that the only radionuclide formed in water, which is ²⁴Na, will not migrate a distance exiting the study area. The ²⁴Na radionuclide will decay before exiting the study area. Therefore, the facility should put measures to protect the groundwater from irradiation within the site of iThemba LABS facility, as no impacts on down gradient can be expected. The long-lived isotopes in soil samples have shown background concentration readings. Due to the long half-lives of such radionuclides, they can be expected to migrate and exit the facility. However, these radionuclides might pose minimal threats to the soil and groundwater as they are occurring at background concentration levels. Due to the dangers of exposure to ionising radiation, it is recommended that proper shielding from neutron radiation be considered. The results of this study contributed valuable information to the understanding of the environmental impact of radionuclide production facilities. The findings have implications for risk assessment, environmental monitoring, and the development of remediation strategies. For environmental safety practises it is recommended that a thick shielding concrete should be built to minimise the amount of neutron radiation that passes through the shielding material.

The authors would like to thank P. Maleka and the Environmental Radioactivity Laboratory group for making the laboratory available for this research. Furthermore, the authors would like to acknowledge the effort made by the iThemba LABS medical radiation department to allow access to the activation of samples in the radiotherapy vault.

This work was supported by the National Research Foundation: iThemba LABS.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.