Experimental Studies Concerning the Semipermeable Membrane Separation Efficiency for

The paper presents the experimental tests concerning the treatment by membrane techniques of radioactive aqueous waste. Solutions, which have been treated by using the bench-scale installation, were radioactive simulated secondary wastes from the decontamination process with modified POD. Generally, an increasing of the retention is observed for most of the contaminants in the reverse osmosis experiments with pre-treatment steps. The main reason for taking a chemical treatment approach was to selectively remove soluble contaminants from the waste. In the optimization part of the precipitation step, several precipitation processes were compared. Based on this comparison, mixed [Fe(CN) 6 ] 4-/Al 3+ /Fe 2+ was selected as a precipitation process applicable for precipitation of radionuclides and flocculation of suspended solid. Increased efficiencies for cesium radionuclides removal were obtained in natural zeolite adsorption pre-treatment stages and this was due to the fact that volcanic tuff used has a special affinity for this element. Usually, the addition of powdered active charcoal serves as an advanced purifying method used to remove organic compounds and residual radionuclides; thus by analyzing the experimental data (for POD wastes) one can observe a decreasing of about 50% for cobalt isotopes subsequently to the active charcoal adsorption . . The semipermeable membranes were used, which were prepared by the researchers from the Research Center for Macromolecular Materials and Membranes, Bucharest. The process efficiency was monitored by gamma spectrometry.

Classic filtration is known as separation of two ore more components from a heterogeneous fluid system.Conventionally, the term of filtration denotes the separation of solid unmixable components or particles from a gas or liquid [1,2].The membrane filtration extended these applications by including separation of homogeneous systems [2,3].The main role of the membrane is that of a selective barrier.Thus, the membrane allows passage of some components out from a mixture while retains the others [2,4,5].Consequently, by retention or permeation one can recover the continuous phase, while the discrete components of the systems are being concentrated [2,4].
During the past 5-10 years the membrane technologies were used increasingly in the nuclear field for the treatment of liquid radioactive waste.The main reason was to reach at least one of the following objectives: the recycling of the boric acid, decreasing of the environmental discharge of radioactivity and decreasing of the amount of solid radioactive waste yielded by the existing treatment technologies used for liquid effluents [6a].
The nuclear industry generates a broad spectrum of low and intermediate level liquid radioactive wastes (LRWs).The treatment methods used for liquid radioactive waste are related to the conventional processes used for the treatment of the industrial waste water, like chemical treatment, adsorption, filtration and ionic exchange.The use of such techniques is limited due to their decreased capacity to remove the radioactive contaminants, the increased operational costs (e.g.evaporation) or the yielding potential for significant quantities of secondary solid waste [6b] After development of suitable membrane materials and their long-term verification in conventional water purification fields, these membrane processes have been adopted by the nuclear industry as a viable alternative for the treatment of LRWs.The most used processes in the field of water and wastewater treatment are those utilizing pressure gradient as the process driving force [6b].These processes include reverse osmosis, ultrafiltration and microfiltration, allowing the selective removing both dissolved and particulate contaminants [6,7].
The membrane separation processes can be used alone or as part of complex treatment schemes which combine conventional technologies and membrane techniques [6b].These combined systems appeared to have superior treatment performances and to be capable of producing high quality treated effluents, bearing an acceptable level of residual radioactivity, for discharge.In addition, the volumes of secondary radioactive waste residues are minimized and can be suitably conditioned to meet the waste form criteria for disposal [6c].
Membrane systems are rarely acquired off the shelf but must be designed and then built only after extensive onsite testing for each specific application.Selection of proper membrane materials and the membrane module configuration is a prerequisite for the successful application of a membrane system [6d].
Knowledge of the characteristics of the feed water is mandatory for understanding of any changes that may occur in the plant's performance.It is the non-radioactive components of the feed water that will determine the overall throughput of the plant, with considerations of radioactivity provided for in the design and layout of the plant [6e].
The purpose of the experimental study was to elaborate and evaluate an adequate technology for treatment of low salt liquid radioactive waste, by using semipermeable membrane.The wastes of concern are those produced during the POD, CANDECON, CANDEREM and AP decontamination processes and their treatment should assure: the effluents with activity concentration below the prescribed limits, radioactivity confinement in a volume as low as possible and an acceptable cost [8].
diluted solutions.This is because the heat exchangers have a complex geometry, which requires a frequent and very accurate decontamination.The POD method has three stages.In the first one, a KMnO 4 solution in nitric acid environment at 90 -95 °C oxidizes chrome ions from Cr 3+ to Cr 6+ which is soluble.In the second stage a mixture of oxalic acid and nitric acid is added to the first stage solution, in order to transform KMnO 4 and MnO 2 to Mn 2+ .Finally, the third stage consists in addition of a mixture of citric acid, oxalic acid and KOH to dissolve the porous oxide layer which has been produced in first two stages.Practically, the process follows in a single stage, because all compounds are consecutively added in the same solution [10,11] The final POD solutions have the following characteristics: pH = 2.5 -3; 0.3 -0.4 g/l Mn 2+ ; 0.45 g/L oxalic acid; 0.96 g/L citric acid; corrosion products as chromates, nitrates, citrates, oxalates, complex combinations, in varying quantities according to the composition of the decontaminated material.The POD procedure is used especially for decontamination of stainless steel materials [9 -11].The main radionuclides which are found in the radioactive waste from the decontamination of a CANDU NPP equipment are: 58 Co, 60 Co, 54 Mn, 51 Cr, 95 Zr/Nb, 144 Ce, 140 Ba, 134 Cs, 137 Cs, 90 Sr, 3 H and 99 Mo [8,12,13].
The reverse osmosis, ultrafiltration and microfiltration tests have been done with cellulose acetate (CA) membranes manufactured at Research Center for Macromolecular Materials and Membranes Bucharest.For the reverse osmosis tests the operation pressure was ~4 MPa.
The experimental tests for treatment of radioactive waste were performed by using six solutions prepared to simulate the waste produced by the POD procedure.The POD1 and POD2 waste have the same content of POD solution, namely 10 % (vol.).The POD3, POD4 and POD5 waste contain 25% (vol.) of POD solution and the POD6 waste has the same radioactive composition as POD3, but with 100% content of POD solution (the radioactive contaminants were added directly in the POD solution).The reason to have tests performed with simulated waste, containing various percentages of POD solution, was that it was assumed that the secondary liquid waste from decontamination operations are gathered either separately or together with the washing waters or other decontamination liquids.
Characterization of the solutions as concerns the radioactive composition was performed by measurement of gamma emitting radionuclides concentrations with a high resolution spectrometer comprising: a HPGe detector with 25% relative efficiency, a low background enclosure and a Canberra spectrometric analyzer, model InSpector.The analysis software was GENIE-PC from Canberra.The spectrometer was energy and efficiency calibrated in the range 60 keV -1500 KeV, by considering all counting geometries used in the experiments.The nuclear properties values used for data reduction were gathered

Experimental part
Experimental determinations concerning the treatment of liquid radioactive waste on semipermeable membranes were performed by using a plane filtration device at the bench-scale, which have been entirely designed and manufactured at the Institute for Nuclear Research (ICN) Pitesti -Romania.
The figure 1 shows schematically the filtration device that is used for tests [9].
The figure 2 shows the design scheme of the experimental set-up used for filtration tests [9].
The solutions treated on semipermeable membrane, at the laboratory scale, were simulated radioactive waste which were fed within the hydraulic buffer as shown in figure 2.
In order to characterize the secondary waste originated from decontamination operations one has to establish the procedures and methods to be used.Therefore, it was recommended the usage of POD method for the decontamination of fine components for which an aggressive corrosion is not acceptable.For the decontamination of the heat exchangers there was recommended the usage of a soft procedure based on  from the original generic library of the software GENIE PC v. 1.4.The radioactive compositions of simulated solutions which were treated with semipermeable solutions are presented in table 1 [11].To express the difficulties faced by the radioactive waste treatment as concern the work at nanoconcentrations level, the table 2 presents the radionuclides concentrations in pg/l corresponding to the activity concentrations contained in table 1.
The relationship between the activity and the mass of the radioactive elements is [14]: Since the purpose of the study was not to establish the optimum quantities of sorbent material which should be added during the pre-treatment stage, but to prove the possibilities for treatment of radioactive waste by semipermeable membrane, during the experiments there were used excess quantities of activated charcoal and volcanic tuff.Thus, the pretreatment based on adsorption on activated charcoal was performed by contacting 500 mLof initial waste with 15 g activated charcoal.After 24 h of static contact, the suspension was filtrated with quantitative filter paper, then the filtrate was treated by microfiltration, ultrafiltration and reverse osmosis.The pretreatment based on adsorption on natural zeolite was performed by contacting 500 mL of initial waste with 15 g of Mârºid volcanic tuff (particle size smaller than 0.1 mm).After 24 h of static contact the suspension was filtrated with quantitative filter paper, then the filtrate was treated by microfiltration, ultrafiltration and reverse osmosis.Two of the simulated waste (POD2 and POD3) were used to perform also, experiments of reverse osmosis and ultrafiltration precedeed by coprecipitation with [Fe(CN) 6 ] 4- /Al 3+ /Fe 2+ [15].The selection of optimum doses of coagulants was done by using the Jar test (ASTM D2035-80).

Results and discussions
The tested solutions are based on the following series as concern the radioactivity concentrations: The ratios of decontaminant solution added in the tested solutions are based on the following series: POD1= POD2 = 10% (vol.)< POD3 = POD4 = POD5 = 25%(vol.)< POD6 = 100%(vol.) The results obtained at the treatment of simulated pairs of solutions POD1-POD2 and POD4-POD5 allows the study of the retention efficiency as a function of the initial radioactivity concentration, for a given ratio of POD decontaminant (10%).In the mean time the experiments involving POD3 and POD6 simulated solutions allow the study of the retention efficiency as a function of decontaminant content at a given initial radioactivity concentration.
Rejection, or retention, is a measure of the fraction of solute or solid that is retained or does not pass through the membrane.It is calculated using the following equation: (2) where: R -retention (rejection efficiency) C iniþial -the concentration of a specific component in the feed solution to the membrane process C permeat -the concentration of the same specific component in the cleaned discharge stream leaving the membrane system.
Figures 3÷8 present the treatment diagrams for each type of the waste and the obtained retention efficiencies.Differences between the retention for the isotopes of the same element are due to the fact that the concentrations  for 134 Cs, 57 Co, 58 Co are laying in the vicinity of the measurement method's detection limit, thus the variations being within the limits of the measurement uncertainties.
The divalent transitional metals, Co 2+ and Mn 2+ which are present in the tested radioactive waste, form moderate stability complex compounds [16a] ML (M -metal, Lligand) with oxalic acid (H 2 L) which is present in the POD solution.Co 2+ can form, also, with oxalic acid the species: ML 2 , ML 3 , MHL and M(HL) 2 [16b], and Mn 2+ can form, also, the complex compound ML It is difficult to anticipate the adsorption of uncomplexed radioactive ions, uncomplexed chelating agents and complex compounds chelate-radionuclide, due to the multiple interfering factors like: pH, concentration of challenging ions, concentration of chelating agents, etc.
[16h].By analyzing the results obtained at the treatment of POD1 and POD2 solutions through adsorption on natural zeolite followed by reverse osmosis, as well as those obtained at treatment on activated charcoal followed by reverse osmosis, it can be observed that the retention efficiencies for all radionuclides were greater for the simulated waste with smaller initial activity concentration (namely POD2).This might be explained by the fact that in reverse osmosis the salt flux across the membrane is due to effects coupled to water transport, usually negligible, and diffusion across the membrane.Eq.( 3) describes the basic diffusion equation for solute passage [17a].∆Π=Π f -Π p ∆P=P f -P p Π -osmotic pressure , Pa, (indexes f and p refers to feed and permeate, respectively) P -applied pressure, Pa, (indexes f and p refers to feed and permeate, respectively) The diffusion is independent of pressure, so as ∆P-∆Π→ 0, rejection → 0. This important factor is due to the kinetic nature of the separation.Salt passage through the membrane is concentration dependent.Water passage is dependent on P-Π [17a].
Considering the results obtained at the treatment of POD1 solution by adsorption on natural zeolite followed by ultrafiltration or reverse osmosis it can be observed that better results were obtained in combination which includes reverse osmosis as compared to the one that includes ultrafiltration, in which case good retentions were found just for 134,137 Cs (89,4%, and 96,4%, respectively) due to the fact that this mineral is known as having a good affinity for cesium [18].
The results obtained at the treatment of POD1 waste on activated charcoal, followed by reverse osmosis or ultrafiltration indicate the presence of the critical radioactive species in treated solution as soluble compounds, as a conclusion of the fact that increased retention efficiencies were obtained both for reverse osmosis and for ultrafiltration.It is also known that reverse osmosis is used to retain the dissolved ions and smaller molecules which contaminates water solutions [6i], while the microfiltration is used to retain macromolecules like proteins and small dimensions colloids but not for ionic species [6j].Adsorption of metals on activated charcoal, from aqueous solutions, is determined by both electrostatic (coulombian) and non-electrostatic interactions.The electrostatic interactions are assigned to the surface charge which is generated on the charcoal after its contact with water and dissolved ions.The non-electrostatic interactions can vary in nature with a prevalence of the Van-der-Waals one [19,20,21a].The series of affinity Mn 2+ <Co 2+ <Ni 2+ <Zn<Cu 2+ is the same with the one of metallic complexes (Irving-Williams series) [21b, 22].Thus, one can observe that the treatment of the POD1 waste by adsorption on activated charcoal and ultrafiltration showed a slighter retention of 54 Mn (35,1%), while for cobalt isotopes the retention efficiencies were in the range 50%÷65%.
The treatment of POD2 waste by precipitationcoagulation and reverse osmosis showed very good results both for cesium and for activation products ( 54 Mn, 57,58,60 Co), since most of these radionuclides can be precipitated, coprecipitated and adsorbed by the insoluble compounds and thus, removed from the aqueous solution by various mechanisms, like: co-precipitation, isomorphic precipitation with carrier, radionuclide sorption on macroparticles which are present in the waste or which form during the precipitation process or retention on flocs through ionic exchange or chemical adsorption [23].
The best treatment efficiency was obtained for POD2 waste, by adsorption on natural zeolite and reverse osmosis.One can observe that there is a relationship between the ionic exchange capacity and the ionic radius.Thus, the smaller the ions radius is, the greater are the values of the exchange velocity and exchange capacity.We can assume that this behaviour is primarily determined by the diffusion of the ions through the pores of the solids.It is also possible for the greater ions not to have access in certain zones from the inner structure of the zeolite [24a].
The aspects presented are characteristic for samples obtained by exchanges with solutions (citrates, oxalates, etc.) at the laboratory scale.When speaking about the recovery of the ions from the spent waters, the problem might simplify or complicate according to the nature of the counter-ions which are found in the liquid environment [24b].
As in the case of POD1 and POD2 solutions, one observe that at the treatment of POD4 and POD5 solutions by adsorption on natural zeolite and ultrafiltration, better retention efficiencies were obtained for all radionuclides in the case of simulated waste with smaller initial radioactive contamination (POD 5).
Considering the treatment of POD4 and POD 5 by ultrafiltration and reverse osmosis, can be observed that better results were obtained for POD4 solution and this is due to the fact that prior the ultrafiltration, the pH was adjusted to the alkaline range, thus inducing formation of flocs.In this set of experiments the osmotic membrane used, yielded unsatisfactory results.
Poor cesium retention efficiencies were obtained at treatment of POD4 and POD5 solutions by adsorption on activated charcoal and ultrafiltration or microfiltration.By comparing these results with those obtained in experiments of ultrafiltration or microfiltration preceded by sorption on natural zeolite, one can state that the retention of cesium was done mainly in the pretreatment stage.In this set of experiments good results were obtained for retention of activation products (isotopes of cobalt and manganese).Related to the retention of metallic ions from aqueous solutions by adsorption on activated charcoal, it was demonstrated [21b, 25,26] that the immobilization of the ions exclusively by the ionic exchange mechanism takes place just for the alkaline metals.In case of the heavy metals sorption the process of immobilization is realized simultaneously by ionic exchange and formation of coordinative compounds [21b].
By analyzing the results obtained at the treatment of POD3 and POD6 solutions by coprecipitation and microfiltration or ultrafiltration one observed that the retention efficiencies are close together, which shows that over a certain value of concentration of oxalate and citrate (25% of POD solution) the results of coprecipitation experiments did not vary significantly.

Conclusions
Even if there are somewhat more complicated in operation as compared with the conventional treatment systems, the membrane based separation systems mark an important step forward by providing an improved treatment capability for the liquid radioactive waste.The success in designing and operation of a membrane separation system needs collective involving and a devoted workgroup which shall include designers, operators and membrane suppliers [6h].

Fig. 3 .
Fig. 3.The treatment diagram for POD1 waste and the obtained retention efficiencies 2 [16c].At this time there are no data available concerning the oxalic acid complex compounds formed with Cs [16d].Co 2+ and Mn 2+ form, also with citric acid (H 4 L), moderate strength complex combinations.Co 2+ known complex compounds are: ML, MHL, MH 2 L and MH 3 L.The stability of these compounds decrease with the increase of citrate's degree of protonation [16e].Mn 2+ can form with citric acid the following species: ML 2 , MHL, MH 2 L ºi MH 3 L [16f] and Cs + can form the compound MHL [16g].

Fig. 4 .Fig. 5 .Fig. 6 .Fig. 7 .
Fig. 4. The treatment diagram for POD 2 waste and the obtained retention efficiencies Ni -the mass of component i, kmol/m 2 s z -thickness of the active layer of membrane, m.C f ºi C p -concentrations in feed and permeate, kmol/ m 3 Di -diffusivity within membrane, m 2 /s [17b]

Fig. 8 .
Fig. 8.The treatment diagram for POD6 waste and the obtained retention efficiencies

Table 1
THE RADIOACTIVE COMPOSITIONS OF SIMULATED SOLUTIONS

Table 2
RADIONUCLIDE CONCENTRATIONS IN SIMULATED SOLUTIONS