Treatment of PWR evaporator concentrates19 August 2013
A process to decontaminate borates has been demonstrated at the pilot scale. By Bernard Rottner
Evaporator concentrates from PWRs are harmful to the environment, both because of their radioactivity and their borate content (borates have been classified as toxic for human reproduction). Cementation may be difficult because their main constituents are highly soluble salts. Borates also interact with cement.
We have designed a process named SOGEBOR which enables the separation of borated compounds from radioactive waste, and immobilization of the radioactive non-soluble solid into a stable matrix. The separation of borated compounds from radwaste has two advantages: first, borated compounds may be reused, or disposed of in a repository adapted for chemical risks (not all radwaste repositories can handle toxic chemicals), and second, the final volume of the radioactive waste is minimized. This paper presents the favourable results of this process from pilot-scale experiments.
SOGEBOR is a two-step process for treatment of the evaporator concentrates (see Figure 1). In the first step, we extract borates from the concentrates. There are several known extraction processes, based on dissolution/precipitation, steam stripping, or membranes . While other processes extract boric acid (H3BO3), we chose dissolution/precipitation because it extracts borates as mainly the relatively heavy borax, Na2B4O7). Therefore the quantity of extracted dry matter is higher with this dissolution/precipitation process, and so the volume of final waste is minimized. The activity level of extracted borax is very low, and allows reuse of the boric acid in the power plant for reactivity control (after processing to produce boric acid), or disposal as non-radioactive waste (or very low level radioactive waste in France).
After extraction of borates, the residual waste contains mainly alkali salts: nitrates, sulfates, chlorides, and some residual lithium, sodium and potassium borates. The residual waste contains also non-soluble compounds (sludge) like metallic salts, concrete powder and organics. The composition of this residual waste depends on the history of the unit (for instance past decontamination operations), and also on the PWR technology; for instance, VVERs use potash for neutralization of boric acid, when French PWRs use lithia. Therefore there is more lithium and less potassium in the French PWRs than in the VVERs.
The target of the second step is to produce a non-soluble solid with highly-soluble alkali salts, together with keeping the volume of the final waste as low as possible. Therefore we heat the residual waste up to melting, with appropriate additives. Heating reduces the quantity of waste through the evaporation of water, incineration of organics, and thermal dissociation of nitrates and carbonates. The additives were selected in order to produce a non-soluble solid. Melting occurs at below 1000°C. The melt is cast into 200 L drums.
The final product may be a glass or a synthetic (polycrystalline) rock, depending on the selection and the quantity of additives. Glass has the advantage of requiring fewer additives, but has the drawback of splitting into small pieces during cooling. A waste in the form of split pieces of glass is not directly compliant with acceptance criteria of surface repositories (for LLW and ILW waste), which generally require that the waste be immobilized into a solid block.
The solubility of borax varies dramatically with the temperature: from 1.23% (ratio of the weight of Na2B4O7 to the weight of solution) at 0°C to 28% at 100°C . Therefore, it is possible to extract borax by concentrating the initial waste at 100°C and precipitating it at low temperature. Borax is purified during precipitation.
Boric acid forms many different borate species with soda, depending on the B/Na ratio: H3BO3, NaB3O5, Na2B4O7 (borax), NaB4O7, NaB5O8, NaBO2, and so on. As the solubility of these species are different, the solubility of borates depends strongly on pH (Figure 2). A minimum of solubility is reached at around pH 9.2 (point 2 on Figure 2) and corresponds to borax. In order to have the greatest ratio between solubility at low temperature and solubility at high temperature, we have to adjust the pH in the concentrate so that borates form borax.
According to the level of residual radioactivity allowed in the purified borax, one or several steps of dissolution/precipitation are necessary (see Figure 3). Figure 3 represents the time sequence of operations, but does not require three precipitation tanks and two dissolution tanks. Indeed, used sequentially, only a single precipitation tank and a single dissolution tank are required.
Recycling of supernatant increases the proportion of extracted borate, which is equal to the proportion of the last precipitation step. The theoretical extraction proportion Pth is related to the ratio of the low-temperature solubility SLT to the high-temperature solubility SHT:
Pth = 1 - SLT/SHT = 95.6%
In fact practical factors such as solubility margins to avoid precipitation in the pipes limit the extraction proportion to 80-90%. This is why there is an additional evaporation step after filtration (Figure 3): the solution is filtered at a concentration lower than the maximum solubility, so that no precipitation occurs during filtration. Then the solution is concentrated to maximum solubility, before cooling down.
Decontamination factors per precipitation/dissolution step range from 10-50 for I to 40-150 for Cs. The DF for Fe is very sensitive to filtration quality. Organic molecules may chelate non-alkali metals (Co, Ni, Fe, Sr): when chelated, the metal is in solution and is separated when the borax precipitates, and when not chelated, the metal is separated during filtration.
We carried out numerous laboratory tests to select a target residual waste composition (with additives) that presents these characteristics:
¦ Non-soluble final product: our criterion was a loss of weight during the first few days of less than 5%, and an additional loss of weight of less than 0.5% over six months.
¦ As high as possible incorporation ratio: This ratio is defined as the amount of initial waste divided by the volume of the final product (for instance 50g/L concentrate).
¦ Processing temperature less than 1050°C: at higher temperatures, borates and alkali oxides (including caesium) volatilize. Lower temperatures enable the use of cheaper technologies for the furnace and for the off-gas treatment.
¦ If possible, the final product should be in polycrystalline form (synthetic rock), so it does not split into small pieces during cooling.
In the end, we developed two compositions, a glass and a synthetic rock. Na2O, K2O, B2O3, others (oxides of multivalent metals) and a part of CaO and MgO come from the raw waste, and the other compounds are additives. One 200 L drum of final product of composition 1 contains the residual waste arising from the treatment of 26.3 m3 of 50 g/L concentrate; one 200 L drum of final product of composition 2 contains the residual waste arising from the treatment of 22.8 m3 of 50 g/L concentrate.
Composition 1 is obtained by melting at 975°C, and composition 2 at 950°C. In both cases samples de-gas and produce foam during heating. Degassing arises from water evaporation, combustion of organics, and thermal dissociation of nitrates and carbonates. Degassing also generates foaming.
We built a pilot-scale system to test the process, firstly on non-radioactive simulated waste and secondly on radioactive waste, on batches of about 20 kg of initial dry raw waste (Figure 4). This pilot unit has been successfully tested. Results presented in table 2 were obtained with this unit.
We selected microwave heating technology because it allows quick melting (less than 1 hour for a 450kg/200L batch). Microwaves are injected into a rotating crucible, whose rotation stirs the heated product. Stirring during heating is very important because it allows an easy degassing and thus avoids foaming. We successfully tested the melting process at a pilot scale (10kg of final product).
At an industrial scale, we would first melt a small amount of compound, about 20kg, and then add the raw compound continuously, until reaching 450kg/200L melt. This procedure avoids foaming and facilitates off-gas treatment. Then we would cast the melt into a 200L drum. Once cooled, the drum may be disposed of into a final repository, directly or after over-packing, depending on its activity and the local waste specifications. In some cases, we would produce a glass, which, after splitting during cooling, would be grouted for final disposal.
The off-gas treatment system uses proven and commercially- available technologies:
¦ Cooling by mixing with cold air, or by water injection
¦ Decloggable bag filters; dust collected is re-injected into the crucible
¦ HEPA filters are protected by bag filters and do not require frequent replacement
¦ Scrubber, for acid gas capture. The scrubber produces a non-radioactive wastewater.
¦ 'Denox' unit reduces NOx into nitrogen
¦ Controls dust, CO, volatile organic compounds, NOx, SOx, radioactivity.
SOGEBOR is a process treatment of evaporator concentrates from PWRs that incorporates recycling of a part of the waste, a high-quality final product, and very high incorporation ratios: the final waste arising from the treatment of 26m3 of 50g/L concentrate is conditioned into one single 200L drum. The process has been successfully tested at pilot scale. The detailed design of the industrial scale unit is in progress. It is designed to treat 250kg/h of raw waste (as dry content; equivalent to an input of 5m3 of 50g/L concentrate), and producing 125kg/h of non-radioactive borax, and 150kg/h of final end product.
Bernard Rottner, technical director, Onet Technologies, France
This article was first presented at 'Decommissioning Challenges: International Reality and Prospects' conference organized by the French nuclear society SFEN, 7-11 April 2013, Avignon, France