A review of radionuclide behaviour in the primary system of a very-high-temperature reactor

During normal operation of (V)HTRs radiologically-significant contamination of the primary system will occur this being of prime importance in relation to depressurization accidents. This paper reviews information relevant to radiocontaminant transport in (V)HTR primary systems paying particular attention to chemical forms, interactions with dust and overall distribution. The primary-system environment comprises nuclear graphites, alloys, dust and high-purity helium into which low releases of the isotopes Cs-134, Cs-137, Sr-90, Ag-110m, I-131, Xe-135 and Kr-85 can be anticipated. Additionally, proper treatment of radiological risk requires accounting for tritium. A likely gas-phase speciation of the chemically-active fission products is proposed: - for caesium and strontium, hydroxides would be dominant with iodides as minor species if a relatively low iodine concentration can be assumed; - for iodine, a split between Csl and HI are likely to dominate with atomic iodine as a minor species. Strong sorption of radionuclides onto carbonaceous dust can be expected. Such dust is likely to cover all surfaces in a pebble-bed (V)HTR so radionuclides will principally associate with this dust rather than underlying alloys. This is unlikely in prismatic (V)HTRs with lower and uneven dust deposits. Where caesium interacts with alloys strong implanting of a large fraction can occur via adsorption and reaction with low-concentration silicon. Silver shows no special affinity for carbonaceous dust but may interact preferentially with nickel-rich alloys, i.e., in the IHX and/or the gas turbine. Quantitative evaluations of radionuclide distribution are hampered by a lack of data regarding sorption onto the graphites, alloys and carbonaceous dust of modern (V)HTR systems; a long time will elapse before sufficient data are forthcoming. In the meantime, some form of best-estimate distribution and upper-bound concentration for contamination is needed if deterministic safety evaluations are to begin. This distribution will be different for pebble-bed and prismatic designs. (C) 2009 Elsevier B.V. All rights reserved.