Ultrasensitive Radionuclide Analysis in Water and Sediments for Environmental Radiological Assessment near the Decommissioning Garigliano Nuclear Power Plant (Italy)

Abstract

We report a high sensitivity study of actinides content in water and sediment matrices performed for the first time along the Garigliano river, near the nuclear power plant currently undergoing decommissioning, and in the marine environment surrounding the river mouth up to a depth of 100 m. Ultrasensitive accelerator mass spectrometry measurements were carried out to estimate the absolute values of ²³⁶U, ²³⁸U, ²³⁹Pu, and ²⁴⁰Pu concentrations and ²³⁶U/²³⁸U, and ²³⁹Pu/²⁴⁰Pu ratios. The accuracy of the measurements and the spatial distribution of the radionuclides enable us to discriminate the anthropogenic from the natural radionuclide’s contribution to the environmental radioactivity. The results indicate that the contribution to the anthropogenic contamination of past power plant operations is, in most of the examined environmental compartments, negligible compared to fallout. High resolution γ-ray spectrometry measurements for ¹³⁷Cs and ⁴⁰K show interesting correlations with the actinides results.

1. Introduction

The study of the neighboring environment of a nuclear power plant (NPP) has gathered interest from the scientific viewpoint to address the possible routes of release and diffusion of radionuclides in the environment. For this purpose, environmental campaigns are useful to verify the radiological impact of the activities of the NPP. This is performed using sensitive measurement methods capable of distinguishing the contribution from the fallout, due to nuclear tests in the atmosphere and major accidents at nuclear power plants (Chernobyl, Fukushima), from the possible impact of the NPP. The implementation of such measurement campaigns is necessary to safeguard the health of people and other living organisms. Moreover, it helps to lower the level of risk perception among the population, for which an objective and verifiable scientific investigation can be effective.

Ultrasensitive accelerator mass spectrometry (AMS) turns out to be a powerful method for making such analyses. Indeed, AMS has an unparalleled sensitivity for rare isotope detection by single ion counting [1,2,3,4,5], which brings information about their origin: natural, weapon grade, reactor burn-up, fall out. Its sensitivity makes it possible to measure abundances and isotopic ratios of anthropogenic actinides (²³⁶U, ²³⁹Pu, ²⁴⁰Pu) below the common values due to fallout in the environment. These data may reveal the presence of irradiated U through the ²³⁶U/²³⁸U ratio, while ²⁴⁰Pu/²³⁹Pu gives further information on fuel composition, U enrichment, and burn-up [2,6,7,8,9,10,11].

Over the years, several studies have been carried out near both active and spent NPPs, eventually under decommissioning, using different methods of analysis on various radionuclides. Environmental matrices, such as soil, air, water, vegetables, milk, meat, and fish, have been studied [12,13,14,15,16]. Riverine and sea waters and sediments near nuclear sites are interesting as possible transport routes for radionuclides released into the environment [17,18]. Their study is crucial to assess the radiological status of the seabed and water and to identify any anomalies or variations from the environmental background values, to ascertain impacts caused by the NPP activities.

Sediments are a reservoir of different pollutants and compounds that undergo physical and chemical processes such as absorption, ion exchange, co-precipitation, complexation, and chelation. The type of interactions between chemicals and sediment particles is affected by particle size distribution (content of the finest fraction), clay content, organic carbon content, cation exchange capacity, and pH. Metals and metalloids are preferably removed from surface water by adsorption on oxides or clay minerals or by forming complexes with organic ligands, while organic pollutants tend to absorb and chelate the organic carbon fraction of the sediments. The accumulation of pollutants in sediments constitutes a constant risk of secondary water pollution resulting from redistribution and resuspension processes, which can cause the exchange of pollutants between the lower sediments and the water column. The rate of migration of pollutants depends on the chemical form in which the pollutants are found in the substrate and in the water within the pores, as well as on the physico-chemical conditions at the sediment–water interface.

The chemical characterization of sediments is thus a key point in marine coastal environmental studies and essentially concerns the study of the quality of surface sediments (sandy/pelitic). In particular, the fine sediment, by its very nature (high specific surface), constitutes a preferential environment for the accumulation of contaminants in general and its resuspension could lead to the reintroduction of contaminants in the water column and in the particulate cycle [19].

The evaluation of potential radionuclide migration pathways, the consideration of hydrogeological aspects such as the knowledge of the flow rates, the average residence times, the fluvial and marine dynamics, and the knowledge of marine and riverine ecosystems, are important aspects in planning activities for the protection of public health and the environment in the event of accidental releases of radionuclides.

To tackle this issue, we carried out a campaign targeting the Garigliano River (Central Italy) and the marine environment surrounding the river mouth. Next to the river, about 7 km from the coast, there is the Garigliano NPP (GNPP). The GNPP, powered by a 160 MW boiling water reactor, was active between 1964 and 1979, when it was shut down for maintenance. The plant was definitively decommissioned in 1987, when, after the referendum that banned nuclear power plants in Italy, the safety procedure began. The reactor is currently isolated, and the pipes and other parts are sealed. Radioactive wastes of medium and low levels of activity are currently stored.

This is the first extended investigation of actinides in the marine environment in this region, especially for water samples from the area in front of the alluvial Garigliano Plain. Desideri et al. [20] have previously investigated the geochemical partitioning of actinides, ¹³⁷Cs, and ⁴⁰K in sea sediments of the area, while Quinto et al. [9] show ¹³⁷Cs and ²³⁶U in river sediments, upstream and downstream of the NPP, and Roviello et al. [21] have recently shown data along the sandy beaches of Baia Domitia (Campania, Italy). The novelty of this work is that the concentrations of ²³⁶U, ²³⁸U, ²³⁹Pu, and ²⁴⁰Pu and the isotopic ratios ²³⁶U/²³⁸U, and ²³⁹Pu/²⁴⁰Pu are measured by AMS on 160 samples of seawater and marine sediments from the coast to the 100 m bathymetry, and of fluvial waters and sediments, both downstream and upstream of the GNPP. Using high resolution γ analysis on a subset of samples, the activity concentration of ¹³⁷Cs and ⁴⁰K was also measured.

2. Materials and Methods

2.1. Study Area

The GNPP is in the municipality of Sessa Aurunca (Central Italy), about halfway between Rome and Naples. The coastal area of the Garigliano plain is characterized by shallow waters and sandy beaches as the result of the transport and sedimentation processes of alluvial and pyroclastic detrital materials that give the sediments a dark color due to the presence of pyroxenes and volcanic magnetite [22]. The flat area that includes the power plant is mainly made up of the Garigliano floods: conglomerates, sands, and clays formed by the dismantling of the carbonatic and Apennine reliefs that make up, for the most part, the catchment area of the river itself and of volcanic materials from the Roccamonfina complex, crossed by the river just upstream of the plain.

The river environment considered (last 15 km) is characterized by an average flow of around 120 m³·s⁻¹. Floods are not infrequent, even of certain importance in the autumn and winter months (with peaks even over 400 m³·s⁻¹). The full hydrological regime usually runs from November to April, while from May to October, a reduction in flow rates is observed. The reduced section of the riverbed at some points increases the speed of the current, except in the last two kilometers, where the reduced hydrodynamics favors the deposition of finer particles, to which various studies associate the accumulation of contaminants and of radionuclides [23].

2.2. The Sampling Grid

On the basis of previous studies on the distribution of radionuclides in the fluvial and marine-coastal area under examination [9,24], a fluvial sampling grid was created, upstream and downstream of the NPP, and marine-coastal from Massico (about 12 km south of the mouth of the Garigliano) upstream of the promontory of Gaeta, north-west of the mouth itself. In total, 9 sampling points were chosen on the coastline, equally spaced about 2.5 km apart; a sediment sample (SC) and a water sample (AC) were collected from each point. The base point (T0) of this line of points was set on the north side of the mouth of the Garigliano (Figure 1).

The marine grid was chosen by identifying six transects (Figure 1), originating from six of the coastal points mentioned above (T0, T1, T2, T5, T6, and T7), and fanning out towards the open sea to better cover the gulf in front of the mouth. For each transect, 7 points with fixed bathymetry of 5, 10, 15, 20, 40, 70, and 100 m have been identified. Surface water (AS), depth water (AP), immediately above the sea floor, and sediment (SM), taken directly from the sea floor, were sampled at each of these points.

Finally, eight sampling points were identified along the Garigliano River: two upstream of the power plant, representing an undisturbed reference area, and six downstream, along the river course. The points were chosen considering the sediment accumulation areas, so they were identified near the banks of the meanders where sediments accumulate. Near the mouth, there are two points that complete, with T0, the transect transverse to the outflow of the river, where the current slows down and helps sedimentation. One sediment sample and one water sample were collected at each point.

The study area with the location of all sampling points in the river and the sea is shown in Figure 1.

Figure 2 shows the bathymetric profile of a typical transect in the area under study (T0). After a slight descent in the first 5–6 km, there is a steeper slope. A similar behavior can be seen in the other transects.

2.3. Sampling Typology

The sampled environmental matrices are marine water, river water, and sediments. For the sake of brevity, abbreviations of the names of the samples reported in brackets have been used: marine sediments (SM), deep sea waters (AP), superficial marine waters (AS), coastal sediments (SC), coastal waters (AC), river sediments (SF), and river waters (AF).

The total number of expected samples is 160, divided as follows:

SF + AF: 16 samples (8 + 8).

SM + SC: 51 samples (42 + 9).

AP + AS + AC: 93 samples (42 + 42 + 9).

The labeling of the samples was performed according to the following scheme:

• The marine coastal sampling points (AC, AS, AP, SC, SM) are identified by an identification code of the transect: “TX_”, where the X indicates the identification number of the transect. For river samples, the code becomes “FX_”, with X the identifier of the point.
• The three digits of the bathymetric depth in meters (unsigned) are added to this prefix: from “000” for the points on the coast and along the river, to “100” for the bathymetric depth 100 m.
• Finally, the second underscore “_” and two letters indicate the type of sample.

For example, T0_040_AP indicates the deep seawater sample taken a few meters from the seabed of the bathymetry point 40 m along transect 0.

2.4. Sampling Operations

Marine water and sediment sampling along the identified transects (see Figure 1) was carried out from December 2019 to February 2020. River water and sediment sampling were carried out in spring 2020. Both sampling campaigns were performed by Hydrolab s.r.l., which fully provided all logistics, from collection to shipping to the analytical laboratory. Marine and river waters were collected using Niskin (or Van Dorn) bottles for a total volume of 5 L. The samples, stored in clean plastic containers, were delivered, shortly after sampling, to the analytical laboratory. The sampling depth is a few meters below the surface for AS samples, a few meters from the bottom for AP samples, and about halfway for the water column for river water samples, avoiding sampling stagnant areas.

The sediment samples were taken using a Van Veen type bucket and stored in decontaminated plastic containers. Each sample delivered to the laboratory has a mass of 4 kg. The coastal sediment samples (T0–T8) were sampled in an area smaller than 0.5 m², up to a maximum depth of 0.2 m.

2.5. Samples Preparation and Measurements

Samples intended for measurements of concentrations of ²³⁶U, ²³⁸U, ²⁴⁰Pu, and ²³⁹Pu and respective isotopic ratios must be reduced to the form of pure oxides of the species to be analyzed, bound with iron oxides (U_xO_y + Fe₂O₃ and Pu_xO_y + Fe₂O₃); therefore, the samples undergo a preparation, which consists of the following steps: 1. pre-treatment; 2. chemical extraction; 3. purification on the chromatographic column; 4. precipitation; 5. calcination; 6. pressing cathodes.

Sediment samples are first dried in an oven at 70–80 °C for at least 2 days. Once water is removed, coarse sieving (>0.5 mm) removes any unwanted coarse debris. An aliquot of the dry and sieved sample (10–30 g each) is taken and weighed for reference in the final concentration measurements and burned in a muffle furnace at 550 °C through the night to produce ashes. Aqueous samples are filtered and acidified to eliminate suspended solids, to stop biological processes, and to keep the analytes in solution as soon as they are delivered to the laboratory. Subsequently, an aliquot is taken in clean containers of an almost uniform and known volume of 2.5 L, and the mass is weighed.

At the end of the pre-treatment phase, known quantities of “spike” are added to the selected aliquots (standard of ²⁴²Pu and standard of ²³³U), to normalize the yields of the preparation and the measurement process. Blanks and standards for the quality check are also produced at this stage to accompany the samples along all subsequent phases, up to the final analysis of the data. Chemical extraction of soils and sediments is essentially an acid digestion of the ashes resulting from the pre-treatment phase. The ashes are transferred into glasses of borosilicate and dissolved into adequate quantities of 8 M nitric acid (approximately 100 mL per 10 g of ash). The acid is added gradually and cold to allow the carbonates to dissolve and release CO₂. When the effervescence produced by the dissolution of the carbonates stops, the actual phase of acid leaching/mineralization begins, letting the solutions react at about 250 °C for 3 h, placing them on heated plates under the hood. Undigested inert material is removed by filtration before the samples are dried and fumed at around 80 °C with concentrated HNO₃ and H₂O₂ (in a ratio of about 5:1 in volume) to remove the digestive acid and the chemical and organic contaminant, before transferring the samples into suitable solutions, for the chromatographic purification.

For aqueous samples, the extraction procedure is co-precipitation with iron III hydroxides around pH 8–10, which is achieved by gradually adding adequate volumes of ammonia to an aqueous solution. Once a thick reddish deposit has formed, the supernatant is removed, until to 1/10 of the initial volume, using a pump. The residual part is transferred into polypropylene tubes and centrifuged in successive stages, each time removing the supernatant and accumulating the deposit in the tube. Finally, two or three rinses are carried out with Milli-Q water; each addition of water is followed by centrifugation and discharge to remove residual salts and bring the pH back to neutrality. At the end of the chemical extraction process, all types of samples are dissolved in a volume of 40/50 mL of HNO₃ 8 M, to obtain suitable eluent solutions. At the end of this phase, two different sets of columns with different ion exchange resins, Pu and U, are extracted and concentrated separately. First, Pu is retained and then released in columns with Dowex ion exchange resin, or Bio-Rad 1 × 8. NaNO₂ is added aforehand to the samples before elution, and converts the Pu bared in the sample into Pu (IV) oxidation state to ensure complete affinity with resins to increase the retention efficiency of the Pu. Uranium is concentrated using UTEVA resin. The eluate produced by each chromatographic process (U and Pu) is precipitated with Fe(OH)₃, centrifuged, and freed of the supernatant. After at least three rinses and centrifugations with Milli-Q water, the precipitate is dried at 80 °C for at least one night. Each sample baring U and Pu separately is then converted into U and Pu oxides and iron oxides (U_xO_y + Fe₂O₃ and Pu_xO_y + Fe₂O₃) by calcination in a muffle furnace at 550 °C for at least 6 h, before transferring the samples in aluminum cathodes for the AMS measurements. Chemical yields are not reported here because they are the subject of a detailed optimization procedure that will be discussed in a future study.

Measurements were performed at the CIRCE AMS facility based on an NEC 9SDH-2 Pelletron accelerator operated at 3 MV, according to the method in [25]. Thirty-nine cathodes form each measurement batch, including cathodes of U, Pu, chemical blanks, chemical standard blanks, and measuring standards.

Some samples, along chosen transects (T0, T1, and T5), were measured with a high-resolution germanium hyperpure γ-ray detector (1.9 keV resolution at 1.332 MeV and 70% efficiency) properly shielded. The scope was to measure ¹³⁷Cs, produced in human activities connected to power plant operations and nuclear accidents and experiments, and ⁴⁰K, a naturally occurring radioisotope chemically related to Caesium, widespread in the environment. For this purpose, samples were dehydrated, sieved (2 mm), sealed in Marinelli vessels, and analyzed with a high-resolution germanium hyperpure γ-ray detector (the details of the preparation and measurement are the same as [15]).

On some marine sediment samples, the particle size distribution was determined by the sieve method. Sieves with mesh sizes of 2, 1, 0.5, 0.250, 0.125, and 0.063 mm were used to determine the percentage of the various types of sand and silt.

3. Results and Discussion

3.1. River Water

The results of the river water samples are shown in Figure 3. Recall that F1 and F2 refer to samples taken upstream of the GNPP, F3–F8 downstream, and F8 at the mouth (see Figure 3). Error bars correspond to 1 standard deviation. Values below the detection limit are not shown.

The concentrations of ²³⁸U are smaller than 2.5 ng/g. ²³⁶U/²³⁸U isotopic ratio is below the detection limit (<10⁻⁹), corresponding to a maximum concentration of ²³⁶U of 2 ag/g. Moreover, there is no notable trend with distance from the NPP nor between upstream and downstream samples.

Concentrations of ²³⁹Pu and ²⁴⁰Pu in river water samples are below the detection limit (<0.3 fg/g).

3.2. River Sediments

Results relative to river sediments are shown in Figure 4. The concentrations of ²³⁸U are all smaller than 1.2 μg/g; this value corresponds to the typical natural abundance of uranium in the environment. It is almost constant along the stream. The concentration of ²³⁶U is at the detection limit upstream of the NPP and slightly increasing downstream. This can also be seen in the ²³⁶U/²³⁸U ratio.

The increase in the ²³⁶U/²³⁸U ratio observed downstream of the NPP indicates detectable contamination, likely due to NPP releases accumulated during reactor operation, whereas no significant increase in U concentration C is observed. In fact, due to the uncertainties affecting the measurements, a possible DC variation would only have been detectable if DC/C > 15%. Using a mass balance approach, we calculated the expected percentage increase as a function of the assumed isotope ratio of the contaminant, using the measured values of the upstream and downstream isotope ratios (2.0 × 10⁻¹⁰ and 1.5 × 10⁻⁹, respectively). The curve thus obtained (Figure 5) shows that the minimum value of the isotope ratio that would produce an increase in a concentration lower than its uncertainty is 10⁻⁸, which is hardly attributable to any other source than NPP releases.

The ²⁴⁰Pu/²³⁹Pu isotopic ratios are nearly constant, considering the uncertainties, and fluctuate around the value of 0.18 (Figure 6). This agrees with global fallout values in the northern hemisphere. In fact, the distribution of ^240/239Pu values in marine sediments (Figure 7) shows a peak at about 0.18 [26,27].

We obtain an identical distribution for ²³⁶U/²³⁸U ratio, peaked at 2.8 × 10⁻⁹, compatible with global fallout values.

Concentration values of ²³⁹Pu are ≤20 fg/g and increase slightly downstream, by a factor ~2, except at the river mouth (sample F8). In this case, mass balance indicates that the isotopic ratio of the contaminant is expected to be the same as the upstream value within ±0.04, in agreement with observation.

3.3. Sea Water

Figure 8 shows the concentration and isotopic ratio of plutonium for coastal, shallow, and deep seawaters. Concentrations of ²³⁹Pu were measurable and, in many samples, greater than river ones. In coastal water, they also fluctuate around 10.0 ag/g, while in deep water, their mean value is 5.0 ag/g. The ²⁴⁰Pu/²³⁹Pu isotopic ratio values also fluctuate around the 0.18 value.

The concentration and isotopic ratio of uranium in coastal, shallow, and deep seawater are reported in Figure 9. Concentration values of ²³⁸U are greater than river ones and below 6.0 ng/g, with an average of 3.0 ng/g. The ²³⁶U/²³⁸U isotopic ratio values fluctuate around the value 2.0 × 10⁻⁹, corresponding to an average concentration of ²³⁶U of 6 ag/g.

3.4. Sea Sediments (Coastal, Marine)

Figure 10 shows the concentrations of ²³⁸U and ²³⁶U/²³⁸U isotopic ratio for marine sediments along the transects. Note that coastal sediments are also included as zero water depth sediments. Concentration values of ²³⁸U vary in the range of 0.40–0.75 mg/g, except for a few outliers, up to a depth of 15 m, and in the range of 1.0–1.5 mg/g for deeper samples. Slightly similar behavior is seen in the ²³⁶U/²³⁸U ratio values, with an increase in ²³⁶U for depths greater than 75 m.

The same behavior can be seen in Pu. Indeed, a sharp increase in ²³⁹Pu concentration values, from 10 fg/g to 400 fg/g, is evident for sediment samples from the shoreline to 100 m depth (Figure 11), while ²⁴⁰Pu/²³⁹Pu isotopic ratio values are, as usual, about constant with an average value of 0.18. The spatial distribution of concentrations of ²³⁶U and ²³⁹Pu is shown in the maps in Figure 12 and Figure 13.

3.5. Gamma and Particle Size Analysis of Marine Sediments

Gamma analysis on sediment samples was carried out to determine the activity concentration of ¹³⁷Cs, a fission product both from the nuclear power industry and nuclear weapons, and of the widespread naturally occurring ⁴⁰K isotope. The scope was to compare their specific activity with the abundance of actinides in the spirit of other studies shown in [28,29,30], and to find confirmation of the previously seen behavior of actinides in marine sediments.

Studies in the Garigliano area over the last decades on soil, river sediments, and vegetable and animal matrices found values of ¹³⁷Cs compatible with fallout from the Chernobyl accident [15,31]. A recent analysis of the temporal behavior over 4 decades of the specific activity of ¹³⁷Cs, confirmed this hypothesis [32].

Marine sediment samples from the three inner transects (T0, T1, and T5), where actinides have the highest values of specific activity and distinct trends, were chosen and analyzed for the γ study.

The ¹³⁷Cs specific activity vs. the sample depth (Figure 14) shows the same behavior found for actinides: low values for shallow sediments, up to 20 m depth, then a neat increase at 40 m and a softer increment at 70/100 m, suggesting a possible plateau. Similar behavior can be seen for the naturally occurring ⁴⁰K specific activity, in the left axis, in red. The main difference consists of a relatively higher value for shallower samples.

To highlight any effects due to particle size differences in the marine sediments of the study area, measurements were made on the sediment samples from transect T0 and all samples at 100 m bathymetry. The results show an increase in the percentage of finer particle size from the coast toward 100 m depth along T0 and a fairly uniform composition of samples from all transects at 100 m. The activity concentrations for the measured radionuclides agree with this result, since as the grain size decreases, the absorption surface area increases. The activity concentration of ⁴⁰K is quite uniform, and thus the Cs/K ratio follows the trend of ¹³⁷Cs, showing that the ¹³⁷Cs uptake was not particularly affected by possible large differences in potassium in the different sediment samples.

4. Conclusions

This work concerns the study of the possible environmental impact of the Garigliano nuclear power plant (Caserta, Italy), currently being dismantled, located about 7 km from the coast in a bend of the Garigliano river. We carried out a survey, which is the first in this environment, targeting sediments and waters of the Garigliano river and of the sea surrounding the river mouth, up to a depth of 100 m. In the sea, samples were taken along transects perpendicular to the coast to obtain an areal map of the radionuclides analyzed. High sensitivity mass spectrometry measurements were performed with the accelerator of the CIRCE laboratory to determine the absolute values of the concentrations of ²³⁶U, ²³⁸U, ²³⁹Pu, and ²⁴⁰Pu and ²³⁶U/²³⁸U, and ²³⁹Pu/²⁴⁰Pu ratios.

The precision of the measurements and the spatial distribution of the radionuclides allowed us to discriminate the contribution of anthropogenic radionuclides from natural radionuclides and estimate the possible impact of NPP.

The results show no evidence of NPP impact on water and marine sediments. A very slight contribution appears in river sediments downstream of the NPP, significantly different beyond 3σ from the upstream values, which could be due to its discharges. Trace amounts of Pu from weapons test fallout and ²³⁶U from the Chernobyl accident were detected. At sea, actinide abundance is greatest in deep water sediments (more than 20 m deep), as is the case for anthropogenic ¹³⁷Cs. This, together with no clear asymmetry in the direction perpendicular to the transects, suggests a geochemical enrichment mechanism. Overall, the results indicate that the contribution of past power plant operations to contamination is zero or negligible compared to fallout.