BNFL Sellafield Research Articles

Developments in Safety Assessment Over the Last Thirty-Five Years

This paper is based on work carried out while the author was employed by BNFL Sellafield (now Sellafield Ltd) and Vectra in its various guises (and now Arcadis UK) but the views and opinions expressed are solely those of the author.

Changing patterns of radionuclide distribution in Irish Sea subtidal sediments

This paper presents new data on the distribution of long-lived radionuclides in Irish Sea subtidal sediments, contaminated as a result of the BNFL Sellafield discharges. The results from different sampling campaigns in 1999 have been combined to assess the extent of radionuclide mobility relative to earlier surveys, in both the eastern and western Irish Sea areas, and to investigate changes in radionuclide distribution over time. The results appear to confirm the trend of continuing re-distribution and transfer of contamination away from the English coast. West of the Isle of Man, radionuclide concentrations and inventories have remained more or less constant. The inventory of radionuclides in sandy sediments in the eastern Irish Sea is still under-represented by current sampling, but could be improved by deeper and more extensive vibrocoring.

A new model validation database for evaluating AERMOD, NRPB R91 and ADMS using krypton-85 data from BNFL Sellafield

The radioisotope krypton-85 (85Kr) is emitted into the atmosphere from nuclear fuel reprocessing plant. Releases from the BNFL Sellafield complex have been used to evaluate the performance of regulatory atmospheric dispersion models. Field measurements of 85Kr, associated emissions and meteorological data and dispersion model predictions have been collated into a model validation database that also includes routines to automate the execution of the models. This database allows the evaluation of subsets of data; in this paper we have considered local dispersion, to receptors within 3 km of the site, and regional dispersion, to receptors at distances of 12 km and 52 km from the site. Configurations of ADMS and AERMOD that implemented buildings and terrain modules showed a performance advantage. However, a modified configuration of the simpler NRPB R91 model, showed comparable results to the more complex models when all models were configured using effective stack heights. For regional scale dispersion statistically significant differences were not found between the models.

Modelling 99Tc concentrations in Fucus vesiculosus from the north-east Irish Sea

In 1994 there were substantial increases in the quantity of 99Tc discharged into the north-east Irish Sea from BNFL Sellafield (UK), concomitant with improvements in waste treatment procedures. As a consequence, the concentration of 99Tc observed in seawater and biota samples, taken from the Irish Sea coastline, increased significantly. Elevated concentrations were also reported in Dutch, Danish, Norwegian, Swedish and Arctic waters in subsequent years. In the present study a simple numerical model was developed and applied to time-series data of 99Tc concentrations in the brown seaweed Fucus vesiculosus, collected from three UK sites in the vicinity of Sellafield (St. Bees, Heysham, Port William). The model considered site-specific scaling effects, lag times, previous discharge history and potential seasonal variation in uptake. In general, there was a good fit between predicted and observed concentrations, but the degree of uncertainty varied inversely with the frequency of sampling. We did not observe a significant seasonal variation. The modelled lag times to the three sites were consistent with transport times based on observations of the water column distribution of 99Tc. The model was applied to a variety of discharge scenarios, reflecting current discussion on the future management of 99Tc releases. Concentrations in Fucus reached asymptotic values in 3–10 years, depending on the scenario and sampling site under consideration.

MEAD (part II)—Predictions of radioactivity concentrations in the Irish Sea

The predictions from MEAD, a model that simulates the transport of radionuclides in the marine environment, are presented for the Irish Sea. MEAD predictions for 137Cs and Pu( α) are presented following discharges from BNFL Sellafield and the predictions compared to measured data from near the discharge location and further afield in the Irish Sea. The model performs well in most circumstances given the uncertainties involved in both modelling and data collection although some inconsistencies in the predictions are found. MEAD is also compared to other models of radionuclide transport in the Irish Sea.

A model for the bioaccumulation of 99Tc in lobsters ( Homarus gammarus) from the West Cumbrian coast

A biokinetic model is presented that simulates the uptake and release of 99Tc by the European lobster ( Homarus gammarus). This organism is of significant radioecological interest since lobsters, in contrast to most other organisms, have a high affinity for 99Tc. The model is designed to represent annually averaged 99Tc concentrations in lobsters from the Cumbrian coast, where significant levels of 99Tc have been released under authorisation by the nuclear fuel reprocessing plant at BNFL Sellafield. This paper describes the construction of the model, how it was calibrated using data from published literature, and preliminary results indicating that model output agrees well with the available monitoring data. Given that this model successfully combines laboratory and field data, this research could potentially make a significant contribution to the field, as, to date, it has been difficult to predict and explain concentrations of 99Tc in lobsters.

Kinetic modelling of the dispersion of plutonium in the eastern Irish Sea: two approaches

A new model to simulate the dispersion of plutonium in the eastern Irish Sea is presented. The model solves the 3D hydrodynamic equations using normalized σ coordinates in the vertical simultaneously with the suspended matter equation. Pu can be present in three phases: water, suspended matter and bottom sediments. Reduction and oxidation reactions are also included in the model, in terms of reaction rates, since Pu can be present in the marine environment in principally two different oxidation states. Two kinetic models are presented to describe the transfers of radionuclides between the liquid and solid phases: a one-step model consisting of a single reversible reaction and a two-step model consisting of two consecutive reversible reactions. It has been found that both models can properly simulate the contamination of the waters and sediments from the eastern Irish Sea due to the releases from the BNFL Sellafield nuclear fuel processing plant, since the outputs from both models are very similar and in agreement with observations. Also, both models can simulate the speciation of Pu between the reduced and oxidized forms. However, if the dominant source of radionuclides to the water column is redissolution from a contaminated sediment, a process that is actually occurring in the Irish Sea, it has been found that a two-step model must be used. Indeed, a one-step model predicts the redissolution from the sediment to be unrealistically rapid.

Effect of forest edges on deposition of radioactive aerosols

The possible enhancement of aerosol deposition at forest edges was investigated in a wind tunnel and in the field. The wind tunnel study was carried out using 0.82 μm mass median aerodynamic diameter uranium particles and a composite canopy of rye grass and spruce saplings. The field study was undertaken at a coniferous woodland near to BNFL Sellafield, Cumbria, UK. Two transects were set through the woodland to determine the influence of the forest edge on atmospheric deposition of radionuclides released under authorisation from the Sellafield site. Results from the wind tunnel study showed that the deposition flux of uranium particles decreased with distance downwind from the grass–tree edge towards the interior of the canopy. The deposition flux at the edge was maximal at about 4×10 −7 μg of U cm −2 s −1. This was 3 times higher than that observed over grass where a constant flux of about 1.32×10 −7 μg of U cm −2 s −1 occurred. Results from the field study showed a clear influence of the forest edge on the atmospheric deposition of 241Am and 137Cs. Activity depositions of around 4750 and 230 Bqm −2 for 137Cs and 241Am, respectively, were measured in front of the woodland. Activity deposition inside the forest edge, however, rose to levels of between 20,200 and 50,900 Bq m −2 and 1100 and 3200 Bq m −2 for 137Cs and 241Am, respectively, depending upon the transect. Similar activity concentrations were measured in the pasture to the front and behind Lady Wood. Results from these studies corroborate those obtained from various studies on air pollutants including radionuclides. This underlines the importance of deposition at the edge of forests and its contribution to the overall canopy deposition. The edge effect is therefore an important factor that should be considered in the assessment of fallout impact, whether this is to be made by either direct sampling or by modelling.

Geochemistry of artificial actinide isotopes in west Cumbrian sediments

Authorised low level radioactive waste discharges to sea from the BNFL Sellafield reprocessing plant have led to the accumulation of a wide range of radionuclides in intertidal sediments. The distributions of Pu and Am suggest that these elements are largely immobile in the sediments. Little information can be obtained from most U isotopes but 236U shows a very clear pattern which can be related to changes in reprocessing operations at Sellafield. The profile distribution of 237Np appears not to reflect the likely discharge history, suggesting that this element is mobile. A more detailed study of transuranic element concentrations in sediment pore waters shows that only about 1% of the sediment inventories is in the solution phase. There is an annual cycle in dissolved Pu concentrations although there appears to be no trend in Np or Am.

The effect of chemical speciation on the impact of129I discharges to atmosphere from BNFL Sellafield, Cumbria

lM I is produced through neutron-induced fission of 23i U and neutron capture reactions with the fission products 1!S Te and nn Te in the cores of nuclear reactors. This typically constitutes about 1% of the total fission products, depending on fuel bum-up. Whilst is not released in routine discharges from nuclear power plants it is released during reprocessing. The ability to model '1 behaviour (important for prospective assessments) has been relatively poor. This is due both to variability in the reported behaviour of iodine, which may be related to abundance and physico-chemical form, and to the relative paucity of data for ,29 1. The primary aims of this study are to measure environmental concentrations of 'l around Sellafield; to establish whether the chemical species of iodine in air differs as a function of the source term, and; to determine concomitant values for model parameters defining 1291 transport through the environment. Some differences in the environmental transport of are observed, albeit slight, apparently related to source term. These findings suggest that a source-specific approach may be appropriate. The critical group dose arising from l is estimated to be about 1.4 uSv a1 to adults and 3.2 uSv a1 to infants.

Assessment of prospective foodchain doses from radioactive discharges from BNFL Sellafield

This paper presents the method used by the UK Food Standards Agency (FSA) to assess the potential impact of proposed radioactive discharges from the Sellafield nuclear site on food and determine their acceptability. It explains aspects of a cautious method that has been adopted to reflect the UK government policy and uncertainties related to people's habits with regard to food production and consumption. Two types of ingestion doses are considered in this method: ‘possible’ and ‘probable’ doses. The method is specifically applied to Sellafield discharge limits and calculated possible and probable ingestion doses are presented and discussed. Estimated critical group ingestion doses are below the dose limit and constraint set for members of the public. The method may be subject to future amendments to take account of changes in government policy and the outcome of a recent Consultative Exercise on Dose Assessments carried out by FSA. Uncertainties inherent in dose assessments are discussed and quantified wherever possible.

SRP Scientific Meeting: Recent Legislation - Achievements and Future Challenges London, 9 October 2001

SRP Scientific Meeting: Recent Legislation - Achievements and Future Challenges London, 9 October 2001

SRP Meeting: Social and Political Implications of CommunicatingRadiation Risk Daresbury, Warrington, 20 June 2001

The SRP held a very interesting meeting in June at the Daresbury Laboratory in Warrington on the social and political implications of communicating radiation risk. In today's risk-aware society, effective communication is just as important as the control measures introduced to prevent or restrict exposure.

Comparison between field measurements of <SUP align=left>85</SUP>Kr around the BNFL Sellafield reprocessing plant and the predictions of the NRPB R-91 and UK-ADMS atmospheric dispersion models

Air concentrations of the radioisotope krypton-85 (85Kr), released from the BNFL Sellafield reprocessing site, were compared with the predictions of two atmospheric dispersion models NRPB R-91 and the next generation UK-ADMS model (version 3). Data on stack emissions and local meteorology were used as model input, whilst building effects were included using a simple "effective stack height" approach. Both models tended to make conservative estimates of air concentrations, with slightly higher overestimations being predicted by the UK-ADMS model. Both models had similar fractions of predictions within a factor of two (35% for UK-ADMS and 34% for R-91) and five (56% for UK-ADMS and 62% for R-91) of the measurements. It was concluded that, when using the simple "effective stack height" approach, there was little discernible advantage of applying the UKADMS model in preference to the R-91 model at the BNFL Sellafield site. Further research will investigate the influence of the UK-ADMS "complex effects" modules on the dispersion predictions.

Alpha-emitting `hot particles' in the vicinity ofBNFL Sellafield, Cumbria

In a survey of environmental samples in the vicinity of BNFL Sellafield, twoalpha-emitting radioactive particles were found in samples of grass. Oneparticle appears to be of mineral origin, the other was not definitivelyidentified, but may be a fragment of fuel cladding. Conservative estimates ofthe activities of these particles are very low. The abundance of radioactiveparticles in the terrestrial food chain appears to be low, with no evidencefor any alpha-emitting hot particles in foodstuffs for consumption by humans.Results suggest that there is no significant dose to man through inhalation oringestion pathways.

Doses to Terrestrial Biota in the Vicinity of BNFL Sellafield, Cumbria, UK

Source terms and corresponding radionuclide activity concentrations in biota for 134 Cs, 137 Cs, 238 Pu, 239+240 Pu and 241 Am have been assessed for three semi-natural ecosystems in the vicinity of BNFL Sellafield, Cumbria, UK. Estimates of absorbed doses (mGy.d - 1 ) have been calculated. Doses to key indicator species, Oniscus asellus (detritivorous invertebrate), Carabus violaceous (predatory invertebrate) and Apodemus sylvaticus (granivorous wood mouse) are discussed with reference to the 1 mGy.d -1 level, below which it is postulated that no observable effects on populations in a terrestrial ecosystem occur. Implications for the 'critical group' and reference model' approaches for a framework of radiological environmental protection are discussed. The need to assess the most highly exposed species is advanced. New research focused on the application of biomarker techniques as a mechanism for determining the interactions and effects of environmental contaminants on ecosystem structure and functioning is presented.

Conceptual design of an underground radioactive waste repository

Between 1989 and 1997, Britain's nuclear waste management agency Nirex developed a comprehensive design for an underground radioactive waste repository next to the BNFL Sellafield works in Cumbria. Using a network of rock tunnels up to 1 km deep, the facility was intended to accommodate up to 275 000 m3 of intermediate and low-level nuclear waste. However, the project was halted when a planning application for an underground rock characterization laboratory—to confirm the suitability of the site—was rejected. This paper explains the historical Sellafield design—further development of generic design solutions for alternative sites will build on this work.

CONCEPTUAL DESIGN OF AN UNDERGROUND RADIOACTIVE WASTE REPOSITORY.

Between 1989 and 1997, Britain's nuclear waste management agency Nirex developed a comprehensive design for an underground radioactive waste repository next to the BNFL Sellafield works in Cumbria. Using a network of rock tunnels up to 1 km deep, the facility was intended to accommodate up to 275 000 m3 of intermediate and low-level nuclear waste. However, the project was halted when a planning application for an underground rock characterization laboratory—to confirm the suitability of the site—was rejected. This paper explains the historical Sellafield design—further development of generic design solutions for alternative sites will build on this work.

The role of near-shore industrial waste releases in the dispersion of radionuclides in the NE Irish Sea

Over the past 27 years, through the use of autoradiographic methods combined with field observations and laboratory studies, I have concluded that the behaviour and distribution of the α-active actinide radionuclides in the estuarine and marine sediments of the NE Irish Sea are significantly influenced by the releases of other non-radioactive industrial wastes. Since the 1700s, the various industrial activities in the Cumbrian coastal region have included: haematite mining, diverse non-ferric metal extraction industries, coal mining and a large number of blast furnaces for the manufacture of iron. More recently (1954–92), the Albright and Wilson phosphoric acid factory at Whitehaven, Cumbria, has discharged large quantities of phosphogypsum slurries into the NE Irish Sea. Iron wastes and slag products, together with phosphogypsum and its associated by-products containing the rare earth elements, are extremely reactive towards the actinides. These wastes are now slowly being removed from the region by natural processes following the rapid decline of heavy industry in the area. These wastes have been present since BNFL first started to discharge radionuclides into the NE Irish Sea and have not, so far, been considered in any models for the dispersion of radionuclides in the region. It is shown that sediments of the NE Irish Sea and local estuaries contain a significant part of the actinide content as coatings on two iron minerals, magnetite and haematite; there is also a significant diffuse distribution associated with hydrated iron oxides attached to quartz grains. However, not all magnetite and haematite grains from a given site show α-activity. Relative to the intensity of the α-activity of the constituent minerals in sediments, the two iron minerals can be regarded as hot particles and are associated with a further set of far more intense hot particles that either may be totally derived from BNFL Sellafield or may also include a contribution from the Albright and Wilson effluents. The retention of actinides in sediments of the NE Irish Sea is believed to be influenced by debris derived from redox recycling. Redox debris can be distributed to depth in sediments through bioturbation. For surface sediments, the distribution is likely to be controlled by sediment mobility. Actinide-enriched magnetite and haematite placer deposits are common in the region, especially in sands. Fractionation and differentiation of iron minerals takes place within the sands as a consequence of energy inputs associated, for example, with tides and storms. Historic data for NE Irish Sea and estuarine sediments provide evidence for a change in facies from muds to sandy silts and a decrease in the concentration of iron over the past 30 years. This will influence the potential behaviour and distribution of actinides in the sediments irrespective of any reductions in discharges. The worst scenario would be continued loss of fine-grained material from the NE Irish Sea and the selective retention of iron minerals and related redox debris, such that the inventory of actinides present in the NE Irish Sea (∼95% of that discharged) becomes associated with surficial sediments. At present, evidence to support this hypothesis is meagre, but it should not be ignored as the iron mineral re-concentration processes are already well-developed.

High-temperature melter tests for vitrificationof bnfl high-level nuclear wastes

AbstractVitrification has become the international method of choice for the immobilization of highlevel nuclear wastes (HLW) from both commercial and defense fuel reprocessing activities. However, there are clear economic incentives for "next-generation" melters that might achieve increased waste loadings and higher throughput rates. In addition, BNFL is investigating potential immobilization processes for projected high-zirconium waste streams that would be generated from advanced nuclear fuel reprocessing processes. One approach towards addressing both of these challenges is waste vitrification at temperatures above the limits that are imposed on all currently operating HLW melters (around 1100 -1200°C). A unique high-temperature joule-heated melter that has been operating at the Vitreous State Laboratory for several years was used to conduct a matrix of tests designed to address these issues. The tests used simulants of two reprocessing waste streams: (i) A blend of "Magnox" and "Oxide" wastes, similar to those currently being vitrified at BNFL's Sellafield site; and (ii) The projected high-zirconium wastes. Each type of waste was investigated both as a wet slurry feed using chemical additives and as a dry calcined waste feed mixed with glass frit. Three target glass compositions were designed for these tests in order to permit processing over a wide range of temperatures. Data were collected on processing rates, off-gas composition, glass product composition and product quality, as well as on general processing behavior. In all cases, glass production rates increased rapidly with processing temperature, typically doubling for every 100°C, and decreased with increasing water content of the feed. The glass production rates per unit melt surface area were significantly higher than those typically observed for hot-wall induction melters operating under comparable conditions. Interestingly, off-gas emissions were much greater for the slurry feeds than for the comparable solid feeds. The resulting glass products were characterized for melt viscosity and electrical conductivity, leachability by both soxhlet and PCT procedures, glass transition temperature, and the amounts and types of any secondary phases that were present.