Influence of in situ Biofilm Coverage on the Radionuclide Adsorption

Environ. Sci. Technol. 2007, 41, 830-836

Influence of in situ Biofilm Coverage on the Radionuclide Adsorption Capacity of Subsurface Granite C R A I G A N D E R S O N , * ,†,§ ANNA-MARIA JAKOBSSON,‡ AND KARSTEN PEDERSEN† Deep Biosphere Laboratory, Department of Cell and Molecular Biology, University of Go¨teborg, PO Box 462, SE-405 30, Go¨teborg, Sweden, and Nuclear Chemistry, Chalmers University of Technology, SE-412 96 Go¨teborg, Sweden

Any migration of radionuclides from nuclear waste repositories is expected to be mitigated by adsorption to the host rocks surrounding hydraulically conductive fractures. Fluid rock interfaces are considered to be important barriers for nuclear waste disposal schemes but their adsorptive capacity can be affected by the growth of microbial biofilms. This study indicates that biofilms growing on fracture surfaces decrease the rocks adsorption capacity for migrating radionuclides except for trivalent species. Potential suppression of adsorption by biofilms should, therefore, be accounted for in performance safety assessment models. In this study, the adsorptive capacity of in situ anaerobic biofilms grown 450 m underground on either glass or granite slides was compared to the capacity of the same surfaces without biofilms. Surfaces were exposed to the radiotracers 60Co(II), 147Pm(III), 241Am(III), 234Th(IV), and 237Np(V) for a period of 660 h in a pH neutral anaerobic synthetic groundwater. Adsorption was investigated at multiple time points over the 660 h using liquid scintillation and ICP-MS. Results indicate that these surfaces adsorb between 0 and 85% of the added tracers under the conditions of the specific experiments. After 660 h, the distribution coefficients, R (ratio between what is sorbed and what is left in the aqueous phase), approached 3 × 104 m for 60Co, 3 × 105 m for 147Pm and 241Am, 1 × 106 m for 234Th, and 1 × 103 m for 237Np. The highest rate of adsorption was during the first 200 h of the adsorption experiments and started to approach equilibrium after 500 h. Adsorption to colloids and precipitates contributed to decreases of up to 20% in the available 60Co, 147Pm, 241Am, and 237Np in the adsorption systems. In the 234Th system 95% of the aqueous 234Th was removed by adsorbing to colloids. Although the range of R values for each surface tested generally overlapped, the biofilms consistently demonstrated lower R values except for the trivalant 147Pm and 241Am adsorption systems. * Corresponding author phone: 503-748-1980; fax: 503-748-1464; e-mail: [email protected]. † University of Go ¨ teborg. ‡ Chalmers University of Technology. § Current address: Oregon Graduate institute, Department of Environmental and Biomolecular Systems, 20000 N.W. Walker Road, Beaverton, Oregon, 97006-8921. 830

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 3, 2007

Introduction Spent nuclear fuel (SNF) from electricity generation must be safely stored for over 100 000 years before radioactivity is reduced to the levels emitted by natural uranium. In Sweden, the Swedish nuclear fuel and waste management company (SKB) has proposed the KBS-3 multi-barrier model for the safe disposal of SNF. This model involves encasing the SNF in steel and copper canisters and then surrounding these canisters with bentonite clay 500 m below ground level in a granitic rock repository. In the unlikely scenario of storage canister failure, it is necessary to investigate how the radionuclides may migrate from the repository and whether their movement will be retarded by the surrounding matrixes. Performance assessment (PA) models assess the reliability of the engineered barriers and include predictions about migration rates and retardation effects (1, 2). In nonporous crystalline rock, the dominant transport medium for radionuclides is groundwater flowing through subsurface fractures. The fracture surfaces then become an important interface within the multi-barrier concept as they are involved in metal adsorption processes, ion exchange, precipitate formation and mineral dissolution. Fracture surfaces also support biological growth and the microbial consortia can significantly affect subsurface geochemical interactions (3-5). Geochemically, metal adsorption can be assessed by comparing the concentration of the metal in the aqueous phase with the concentration of metal that appears on a solid phase in contact with the aqueous phase. In the case of near-field nuclear waste disposal investigations, solid phases such as clay and cement are investigated as these are expected in the immediate repository environment. For farfield adsorption investigations the common host rock minerals constitute the most extensive solid phase along with fracture filling minerals such as carbonates, iron oxides, and aluminum oxides. All these minerals are ubiquitous in the environment and have good surface adsorption characteristics in both aerobic and anaerobic conditions (6-8). In addition to solid surfaces, the influence of the groundwater chemistry must also be investigated. For example, competing groundwater ions and inorganic carbon (CO2) and can increase the concentration of some actinides in solution due to carbonate complexation (9, 10). Along these lines, the presence of natural organic matter (NOM) can affect the adsorption capacity of host rocks both positively and negatively. Research on NOM has focused primarily on humic acids (HA), fulvic acids (FA), and fractionated humic acids (11, 12). If NOM is in solution it can aid complexation of radionuclides thereby increasing the aqueous phase concentration (11). Conversely, coatings of organic macromolecules (such as HA or FA) on solid surfaces like hematite can effectively block the adsorptive functional groups on the mineral surface and decreasing the adsorption capacity for lanthanides such as Eu (III) (13). In a subsurface repository environment, fracture surfaces are covered with NOM in the form of biofilms. Subsurface biofilms range in activity and composition depending on the pervading geochemical conditions (14, 15). Currently there are only a few studies investigating bacterial radionuclide adsorption directly connected with SNF disposal in the subsurface (5, 16, 17). It has recently become apparent that subsurface biofilms have a significant effect on the adsorption capacity of host rock formations by forming a barrier between the rock surface and the groundwater (18). This implies that adsorption and desorption chemistry in a 10.1021/es0608702 CCC: $37.00

 2007 American Chemical Society Published on Web 12/22/2006

TABLE 1. Overview of Radionuclide Distribution Coefficients (R) for Scenariosa Each Component is Described in the Footnote to the Table

a (1) Including colloidsswhere nonspecific adsorption to experimental vessel walls is accounted for and adsorption to colloids is included and (2) excluding colloidsswhere nonspecific adsorption to experimental vessel walls is accounted for and colloidal effects are removed. Rwc and Rwc′ ) radionuclide distribution coefficient between the vessel wall and free radionuclides combined with colloidal radionuclides for the blank adsorption systems and treatment adsorption systems respectively. Rs/l ) radionuclide distribution coefficient between the surface area of the treatment (biofilms, rock, or glass) and the free radionuclide in aqueous phase combined with colloidal radionuclides. Rwall and Rwall′ ) radionuclide distribution coefficient between the vessel wall and free radionuclides excluding adsorption to colloids for the blank adsorption systems and treatment adsorption systems respectively. Ra ) radionuclide distribution coefficient between the surface area of the treatment (biofilms, rock or glass) and the free radionuclide in aqueous phase excluding adsorption to colloids. Rcolloid ) radionuclide distribution coefficient between the free aqueous phase radionuclides and the colloidal radionuclides in either the blank or the treatment adsorption systems. Atot ) total activity at zero time in the adsorption systems. Awall ) adsorption to the experimental vessel walls. Aaq ) total aqueous phase activity for the adsorption systems at specific sampling occasion (includes activity adsorbed to colloids). Acent ) aqueous phase activity after centrifugation at specific sampling occasion.

natural system is affected by both fracture minerals and biofilms. To date, no comprehensive study has assessed the radionuclide adsorption capacity of in situ biofilms from anaerobic granitic rock environments. Since plans have been made to bury SNF in subsurface granite hosted repositories it is now important to quantify the biofilm effect for performance assessment modeling purposes. In the experiments presented here, biofilms were pregrown on glass and rock slides in situ, 450 m below sea level at the A¨ spo¨ Hard Rock Laboratory (HRL) in south-eastern Sweden. The HRL was used because it acts as an analogue for studying the effects of subsurface SNF disposal systems. The radionuclide adsorption capacity of these in situ biofilms was then directly compared to clean glass and granite rock surfaces by monitoring the aqueous phase radionuclide concentration. This study investigates 60Co(II), 147Pm(III), 241Am(III), 234Th(IV) and 237Np(V) or 60Co2+, 147Pm3+, 241Am3+, 234Th4+, and 237NpO +. 2

Materials and Methods Field Site, Apparatus, Biofilm Growth and Slide Transfer. These features are described in detail in Anderson et al. (18). Briefly, biofilms were grown on glass or granite slides under in situ repository conditions 450 m underground at the A¨ spo¨ Hard Rock Laboratory (HRL) in south-eastern Sweden for up to 480 days. In situ conditions were maintained until the slides were transferred to the sorption vessels as outlined below. Biofilm parameters have been described before by Anderson et al. (18). Preparation of A 2 spo1 Synthetic Groundwater and the Radionuclide Stock Solutions. This was based on a synthetic water composition calculated for borehole SA2600A which is 400 m below sea level in the A¨ spo¨ HRL tunnel (19). The preparation protocol was slightly altered from Vuorinen and Snellman (20) where 4 different salt solutions were prepared and then mixed. The water has an ionic strength of 0.3 M VOL. 41, NO. 3, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

831

with the major cations being Na and Ca and minor cations are Mg, Sr, K, and Li. The major anion is Cl- with minor silicate, sulfate, and carbonate. The actual composition of the water can be found in the Supporting Information section, and the water preparation protocol has been described previously by Anderson et al. (18). Three separate experiments (E1, E2, and E3) were conducted where the water was sterilized by autoclaving in experiment E1 but was filtersterilized through 0.2 µm filters in experiments E2 and E3. Filtration was chosen in experiments E2 and E3 as precipitates were formed during autoclaving. The preparation of 60Co, 147 Pm, 241Am, 234Th, and 237Np stock solutions has previously been described in Anderson et al. (18). Sorption Experiments. Three separate sorption experiments (E1, E2, and E3) were undertaken and are outlined in tabular form in the Supporting Information section. Adsorption experiments were performed in 50 mL sterile plastic (polypropylene) tubes where 50 mL of A¨ spo¨ synthetic groundwater was added to each tube along with 0.5 or 1 mL of the appropriate radionuclide stock solution. All additions were recorded gravimetrically and then the different sorption surfaces were immersed in the solutions. The final radionuclide concentrations before the start of each experiment were calculated based on the specific activity for each radionuclide and are presented in Anderson et al. (18) and in the Supporting Information section. Up to seven sorption surfaces were tested for each radionuclide with three replicates where applicable (see the Supporting Information section). The sorption surfaces were as follows: biofilms attached to glass slides (Biofilm), biofilms attached to glass slides stripped of metals (Stripped biofilm), loose biofilm with no glass slides (Loose biofilm) (i.e., biofilm was scraped off from glass surface and resuspendedssee section “TOC measurements” in Anderson et al. (18) for a complete protocol), biofilms attached to rock slides (Biofilm on rock), plain glass slides (Glass), polished granite rock slides (Polished rock), and non-polished granite rock samples (Granite rock). Three replicates were used for each treatment except for glass and blank systems where only two replicates were used. Experiment E1 was not performed in triplicate. The pH of each experimental vessel was measured before and after each experiment. For a detailed description of the sorption surfaces and how they were prepared refer to Anderson et al. (18). Samples were extracted at 0, 12, and 24 h, 3 , 7, 14, 21, and 28 days. At each time point, 0.5 mL was extracted for radioactivity and/or metal concentration measurements. In one set of replicates 2.0 mL was extracted, 0.5 mL for immediate radioactivity measurement and the remaining 1.5 mL from each sample was centrifuged at 15 000 g for 30 min before 0.5 mL of the supernatant was extracted for measurement. Pipet tips were equilibrated three times with the solution before sampling. Radioactivity Measurements. The activity of 147Pm, 241Am, 234Th, and 237Np was determined in the aqueous phase using liquid scintillation (1219 RackBeta counter from Wallac, Finland in experiment E1 and a 1414 Perkin-Elmer WinSpectral counter from Wallac, Finland in experiments E2 and E3). Samples of 0.5 mL were mixed with 5 mL of HionicFluor LSC-cocktail from Packard BioScience (Netherlands). The 60Co activity was determined in aqueous phase using a NaI(Tl) γ counter (Intertechnique CG 4000, France) in experiment E1 and a HPGe detector (GEM 23195, Canberra, Stockholm, Sweden) in experiments E2 and E3. ICP-MS Measurements. An Elan 600 inductively coupled plasma mass spectrometer (ICP-MS) (Perkin-Elmer, Stockholm, Sweden) was also used to measure the aqueous concentrations of 237Np. A 0.5 mL sample volume was mixed with 4.5 mL of suprapure HNO3 with a 2.5 ppb internal standard of Bi (LGC Promochem AB, Gothenburg, Sweden). 832

9


All volumes were recorded gravimetrically on an analytical balance to correct final concentrations. Solutions of Np (10 µg mL-1) (LGC Promochem AB, Gothenburg, Sweden) were used to prepare 0.1, 1, and 10 ppb standards for the calibration curves. All samples were analyzed, with the plasma gas flow set at 15 L min-1, the nebulizer gas flow at 0.95 L min-1 and the radio frequency (RF) power at 1150 W. Distribution Coefficients and Calculations. The adsorption processes are represented by the distribution coefficient K (as defined by Davis and Kent in ref 21) where K denotes the distribution of the ion of interest between the solid phase and the aqueous phase. Since we do not assume equilibrium in our experiments we prefer to use the denotation R instead of K. As the system gets closer to equilibrium, R approaches K. Radionuclide (RNn+) adsorption was calculated based on surface area and accounted for nonspecific sorption to the experimental vessel walls along with adsorption to colloids. The surface area of each slide was measured using a digital metric vernier calliper. Rs/l accounts for adsorption to both the solid phase and to colloids, Ra denotes adsorption which excludes colloidal effects. Table 1 outlines the relationships between the different adsorption processes in the adsorption systems. A representation of the equations used for the R coefficient calculations can be found in the Supporting Information section.

Results and Discussion In these experiments biofilms were grown in situ at the A ¨ spo¨ hard rock laboratory to produce an analogue of the natural adsorption surfaces expected in a repository environment. The effect of biofilm coverage on radionuclide adsorption and desorption behavior was compared with granitic host rock from the same environment and quantified using distribution coefficients. Far-field repository conditions were reproduced in the laboratory by using a reduced saline groundwater of near-neutral pH, anaerobic conditions, and in situ temperature. The total organic carbon (TOC) that subsurface biofilm coatings contain is approximately 3.80 mg/m2 (18). Adsorption of radionuclides to biofilm and rock surfaces was calculated using mass balance calculations. Assessment of adsorption in this manner was problematic as the surface areas of the solid phases investigated were either 1.49 × 10-3 or 2.90 × 10-3 m2 which is extremely small relative to other studies where oxide powders such as TiO2 and Al2O3 are used giving surface areas for sorption >1 m2 (22, 23). Low detection limits and measurement accuracy become extremely important in order to note small changes in the aqueous phase activity as each radionuclide adsorbs. Assessment of adsorption to each treatment surface also relies heavily on the accuracy of the blank systems as it is these systems that are used to estimate sorption to the vessel wall. Calculation of negative adsorption coefficients in this study is the result of inaccuracies in all the above-mentioned areas but is predominantly associated with overestimation of wall sorption associated with the blank systems in comparison to the treatment systems. Assumptions were made that the localized differences between the nature and distribution of functional groups on the rock, biofilm, and vessel walls could be averaged across the entire surface area, and treatments could be directly compared. Although these assumptions were made, mineralogical homogeneity and biofilm homogeneity of in situ materials is impossible thus the experimental variation increases. Except for the trivalent species 147Pm3+ and 241Am3+, biofilms generally reduce the adsorption capacity of the natural rock for radionuclides of varying oxidation states expected from the decay of radioactive waste. If these biofilms are removed from the rock surface then their effect is minimized, as in these experiments, all the “loose biofilm” treatments indicated no detectable adsorption. The rate of

FIGURE 1. Representative graphs of Rs/l values over time for 60Co (II) (A), 147Pm (III) (B). Adsorption is to polished rock slides (filled circle), biofilm on rock slides (square with x), biofilm on glass slides (filled upright triangle) and stripped biofilm slides (filled upside down triangle). Values represented are combined from all experiments (E1, E2, and E3) and do not include zero sorption or negative Rs/l values. The range of values collected from each treatment surface is shaded differently to distinguish each treatment. The final Rs/l range from where the adsorption is approaching equilibrium is represented by the double ended arrows at the right of the figures with a dotted extension arrow to zero if some data points plot below zero. The dotted extensions within the biofilm Rs/l range arrows represent the stripped biofilm data. The fastest adsorption occurs within the first 200 h and equilibrium is approached after 400-500 h. The profiles presented in these graphs are similar to the adsorption profiles for 241Am, 234Th, 237Np which are shown in the Supporting Information section. adsorption to all surfaces in these experiments occurs on the scale of days with the maximum sorption rate occurring between 0 and 200 h and adsorption approaching equilibration after 500 h (Figure 1). Oxidation State and Effects on Adsorption Capacity. Oxidation states of II and III (60Co2+, 147Pm3+, and 241Am3+). Observations indicate that 60Co(II) does not readily adsorb to biofilm surfaces, but 147Pm(III) and 241Am(III) do. For 60Co, Rs/l and Ra data in all experiments indicated that sorption to granite rock samples and the polished rock slides was up to 28 times higher than the other treatments (Table 2, Figure 1A). The stripped biofilm had a comparable Rs/l to the biofilm on rock and both these surfaces had adsorption capacities up to 200 Rs/l units higher sorption than non-stripped biofilm. For 147Pm the stripped biofilms adsorbed at least 25% more than non-stripped biofilms and polished rock slides (Table 2, Figure 1B). Ra data was generally higher with nonpolished granite rock samples approaching a maximum of 5.25 × 105 m (Table 2). The data for 241Am was similar to 147Pm with all R ranges overlapping and the stripped biofilms s/l presenting the highest adsorption coefficients approaching 3 × 105 m (Table 2). Biofilm surfaces had the largest spread of distribution coefficients (10-100-fold variation) whereas the polished rock and glass presented more defined Rs/l ranges (data not shown). The Ra data range and spread was generally lower than the Rs/l data (Table 2).

In terms of percentages, biofilms will adsorb approximately 5% of the available 60Co but more than 80% of the available 241Am and 147Pm (includes sorption to the experimental vessel wall and colloids) (Table 3). In comparison, the rock absorbs 40% of the available 60Co and 70% of 147Pm and 241Am. These biofilm figures approach the levels reported in other published reports where bacterial surfaces (planktonic cells with a density of 1011 cell l-1) can adsorb up to 90% of the available 147Pm (24) or 95% of the available 241Am (5). In comparing Cd2+ and Fe3+adsorption, Daughney et al. (25) present clear data that 10% more Fe adsorbs to bacterial surfaces than Cd (depending on growth phase). For the study presented here, a similar relationship exists but with 90% more trivalent versus divalent ions adsorbing when comparing similar radionuclide starting concentrations (1.53 × 10-10 M for 60Co and 3.77 × 10-10 M for 241Am). Daughney et al. (25) suggest that adsorption differences are related to surface stability constants for divalent metals. For example, surface stability constants for the binding of Co, Cd, Ni, and Zn to carboxyl groups on Bacillus subtilis are at least 10-fold lower than for the trivalent ions Al3+ and Nd3+ (26). Furthermore, to compare biological surfaces with mineral surfaces, the bacterial surface complex constants for Co, Cd, and Ni are also slightly weaker than surface complex constants to hydrous ferric oxides (HFO) (27). Weaker surface complexation for 60Co by biofilms could explain why biofilm sorption is lower than the rock. Higher adsorption of 147Pm and 241Am is also expected based on these observations as the biofilm is more likely to retain trivalent ions as the surface complexation constants are higher. Samples from one replicate from each experiment were also centrifuged before measurement. This was to assess the potential of radionuclide adsorption to precipitates and colloids. The tendency for colloid formation for actinides and rare earth elements (REE) according to Geckies et al. (28) is Np (V) < U(VI) < REE/Am(III)/Cm(III) < Th(IV). Ultracentrifugation was not used so colloids smaller than 7 nm could not be centrifuged or analyzed in these experiments (based on a density of 9 kg/L for colloids and 1 kg/L for water). Experiment E1 had precipitates due to autoclaving the synthetic groundwater used in experiment E1. Based on the chemistry of the synthetic groundwater, this precipitate is most probably calcite as the solubility of calcite is exceeded at the temperature of autoclaving (120 °C). In experiment E1, approximately 40 and 15% of the 60Co and 147Pm adsorption respectively was attributed to precipitates (241Am was not investigated in experiment E1). The activity lost to precipitates decreased from an Rcolloid value of 0.7-0 over about 20 days. This implies that the precipitates either dissolve and/or settle out of solution over time. In experiments E2 and E3 the adsorption of 60Co, and 241Am to colloids was minimal as the Rs/l and Ra values were almost identical. This was confirmed by Rcolloid values which were close to zero for the duration of the experiments (data not shown). For 147Pm in the presence of precipitates, the adsorption capacity of granite rock samples and biofilm samples was 5-10% lower than the adsorption capacity of polished rock slides and biofilm slides from experiments E2 and E3. Although the adsorption was higher in experiments E2 and E3, divergence of Rs/l and Ra values over time indicated that there was still some colloid formation in these experiments. Rcolloid values were also above zero throughout the 660 h experiments supporting some adsorption of 147Pm to colloids (data not shown). Oxidation states of IV, V (234Th4+, 237NpO2+). For the radionuclides of higher oxidation state, the situation is similar to that for 60Co where the biofilms have a lower sorption capacity than the rock surfaces. The differences, however, are less pronounced as the upper range of Rs/l for biofilm VOL. 41, NO. 3, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

833

TABLE 2. Maxima and Minima Rs/l Values for all Experiments (E1, E2, and E3) Combined with Maximum Ra Valuesa degree of adsorption (× 102 m)e (Rs/l and Ra values after 660 h) nuclide

60Co

minimum Rs/l

maximum Rs/l 280 9.33 9.74 3.66 5.46

maximum Ra

6.5, 7.5, 7.5

33.0 1.17 2.94 -39.9 2.35

6.5, 7.1, 7.1

(III)

rock biofilm on rock stripped biofilm biofilm glass

78.5 256 982 459 10.9

3640 863 2870 2270 52.9

n/a, n/a, 6.5

(III)


850 n/a 924 228 140

1810 n/a 2900 1860 258

752 n/a n/d 694 144

6.8, 6.8, 6.8

(IV)b


103 141 -98.9 44.9 -150

1240 299 118 109 566

21800 4950 n/d 214 81.3

7.2, n/a, 6.5

(V)

rock biofilm on Rock stripped biofilm biofilm glass

147Pm

241Am

237Np

pHd throughout E1, E2, E3


(II)

234Th

treatmentc

-1.92 n/a -5.88 -7.18 0.88

9.35 n/a 0.28 0.71 6.32

281 13.3 n/d 27.6 5.36 5250 882 n/d 2500 28.2

9.5 n/a n/d 0.96 3.95

a R includes the radionuclide adsorption to both colloids and the treatment surfaces. R excludes the radionuclide adsorption to colloids and s/l a is closer to the actual adsorption occurring on the treatment surfaces alone. b Data is from after 324 h as this experiment was stopped after 324 h. c Values for “Rock” include data points from both polished and unpolished granite rock sections. “Loose biofilm” data was excluded as there was no adsorption. “Biofilm on Rock” was not investigated for 241Am and 237Np. d pH was measured before and after each experiment. No pH differences were observed so data is presented as “pH throughout”. The values are presented in order E1, E2, and E3. 241Am was only investigated in experiment E3. 237Np was only investigated in experiments E1 and E3. e Ra values are calculated for only one replicate set in each experiment as only one replicate set was centrifuged. The maxima and minima data represents the highest and lowest data points for all experiments combined.

sorption capacity always overlap with the lower range of Rs/l for the rock. Broad trends were evident in the Rs/l data for 234Th adsorption in all experiments but results were highly variable due to colloids. Biofilm slides, stripped biofilm slides, and blank samples had very similar levels of 234Th adsorption when compared directly and had a maximum Rs/l sorption of 1.09 × 104 m (Table 2). Biofilm Rs/l values are 5-10-fold lower than polished rock slides. The adsorption curve for biofilm on rock lay within the same range of 234Th sorption to polished rock (data not shown). In terms of percentages, almost all of the 234Th adsorbs. When the amount of 234Th adsorbed to colloids is removed there is 5% “free” radionuclide remaining. Approximately 10-25% of this remaining 234Th adsorbs to the treatment surface and 70-85% adsorbs to the vessel walls (Table 3). The competing surface area of the vessel wall in these experiments obviously has a significant effect as figures reported by Francis et al. (5) suggest that 74% of Th can potentially associate with bacterial surfaces (although their studies used planktonic cells). The pH of the system used in the current experiments would have also affected adsorption as Sar et al., (29) show the maxima for Th sorption to bacteria at pH 4, beyond which the sorption decreases. The same authors also indicate that the presence of Fe2+, Cu2+, Al3+, and K+ ions can decrease the biosorption of Th up to 43% by competing for binding sites by other ions. Along these lines, the 234Th daughter product 234Pa, can also decrease adsorption by competing for binding sites. 234Pa hydrolyzes extensively even in acidic solution and the protactinyl (V) 834

9


ion can readily take up protons and is more accurately represented as PaO(OH)2+ rather than PaO2+ (9). With only one strongly covalent M-O bond compared to other actinides it hydrolyzes and binds to surface functional groups more readily than other actinides including hydrated Th4+ ions (9). Since bacterial and polysaccharide surface sites are generally more susceptible to ionic strength (30, 31), the 0.3 M synthetic groundwater used in the current experiments coupled with a stronger sorbing actinide like 234Pa could cause a decrease in the amount of 234Th adsorption. The Rcolloid values in the 234Th experiments decreased from 25 to between 5 and 15 over the duration of the experiments suggesting that the colloids settled out of solution over time (data not shown). 234Th (IV) readily hydrolyzes in water and as a consequence easily forms colloids and polynuclear species (32). In experiment E1, there was a relative increase in the rate of sorption after the settling or dissolution of the precipitates pointing to precipitate interference with remaining 234Th or some kind of equilibrium between three components: ions-colloids-precipitates. This has been observed before with Pu (IV) colloids and precipitates (33). In experiments E2 and E3, where there were no visible precipitates, up to 95% of the available 234Th still disappeared to form Th eigen colloids despite the fact that the experiments were performed well below the solubility limit for 234Th. There is also an unknown effect of carbonate in the groundwater with Th(OH)3CO3- potentially being one of the predominant soluble species at pH 7 (34). Rs/l data for 237Np indicated very low adsorption with all values below 103 m representing an overall adsorption of

TABLE 3. Theoretical Adsorption (%) of Radionuclides to the Treatment Surface, the Experimental Vessel Wall, and Colloids after 660 ha theoretical “% adsorption” at 660 h based on adsorption coefficient calculationsa treatment

% adsorbed to vessel walls


45.17 5.65 5.80 1.88 2.78

12.43 16.09 21.27 22.15 22.06

0.02 0.99 0.51 1.21 1.22

0.66 -3.47 -1.15 -0.18 -0.34

(III)


28.71 38.04 66.95 52.41 1.55

47.18 41.49 23.71 34.13 72.85

3.20 -0.23 0.64 1.63 2.04

3.03 0.42 -2.22 3.99 2.12

(III)


57.28 n/a 79.22 71.83 32.41

19.82 n/a 9.22 13.05 31.33

-2.10 n/a 1.21 0.26 -0.12

0.30 n/a 10.44 0.64 0.88

(IV)c


11.88 22.17 12.41 11.61 13.93

83.26 73.20 81.36 83.54 79.03

(V)


1.81 n/a 1.98 2.09 1.88

20.56 n/a 20.12 20.12 19.72

nuclide

60Co

(II)

147Pm

241Am

234Th

237Np

% adsorbed to colloids

% adsorption to colloids based on sample [Aaq] - [Acent]b

% adsorbed to treatment

93.89 94.31 72.68 93.22 91.83 1.89 n/a 1.39 1.49 1.70

-2.12 n/a -2.19 -1.72 -2.29

a The contribution of colloids is calculated from the difference between the radionuclide left in solution (centrifuged versus noncentrifuged) and as a percentage difference between the activity of the centrifuged and noncentrifuged samples taken throughout the experiments. These figures are not representative of all possible distribution scenarios as they only represent one replicate set. b Calculated only from experiments 2 and 3 where no precipitates were formed. As only one replicate set was centrifuged, these results only represent one of the possible distributions for the experiments performed and are therefore not a true representation of the dominant trends in the collated experiments. c Colloids were removed from this calculation to more accurately represent proportion adsorbed to the wall. Numbers represent sorption of non-colloid bound 234 Th to wall and surface treatment.

Influence of in situ Biofilm Coverage on the Radionuclide Adsorption

Recommend Documents