Permeable Environmental LeachingCapsules (PELCAPs) for in SituEvaluation of ContaminantImmobilization in SoilB . P . S P A L D I N G * A N D S . C . B R O O K S

Environmental Sciences Division, Oak RidgeNational Laboratory, P.O. Box 2006,Oak Ridge, Tennessee 37831

We encapsulated radioisotope-spiked soil within awater-permeable polyacrylamide matrix cast in a smallcylindrical geometry (≈5 cm3) to measure the persistenceof immobilized soil contaminants. As a proof-of-principle,soils contained within these permeable environmentalleaching capsules (PELCAPs) were labeled with either 85Sror 134Cs and were leached in both laboratory tests andcontinuously in situ with ground and streamwaters at twofield sites on the Oak Ridge reservation. Groups ofPELCAPs were retrieved, assayed nondestructively forradioisotopes via γ spectroscopy, and then replaced inground and surface water repeatedly over a 6-month period.PELCAPs that contained no soil readily and quantitativelyleached either 85Sr or 134Cs into laboratory extractantsor ground or surface water with effective diffusion coefficients(Deff) of (1.14 ( 0.06) and (4.8 ( 0.2) × 10-6 cm2/s,respectively. PELCAPs containing untreated soil readilyleached >90% of 85Sr but <1% of 134Cs during field leachingat both sites, whereas thermally treated soils quantitativelyretained both isotopes under all conditions. Permeablepolymer encapsulation methods, such as PELCAPs, offerthe potential capability to conveniently test large numbersof soils and soil treatments for contaminant release anduptake under actual field environmental conditions.

IntroductionOne of the challenging problems for any advocate ofcontaminant immobilization in soil is to assess the long-term effectiveness of the immobilization either by a remedialtechnology or by natural attenuation (1). Although manypromising techniques for immobilization of hazardous andradioactive contaminants in soil continue to be developed,comparative field evaluation of several techniques at onesite or of multiple promising techniques over a range ofgroundwater conditions or sites remains a difficult task. Thecostly and destructive nature of contaminated soil andgroundwater field sampling, the number and handling ofsamples required to discern statistically meaningful differ-ences, and the required repetition of sampling over prolongedintervals contribute to this cost and difficulty. In a moregeneral view, the scientific and technical challenges tounderstand and measure contaminant availability in soilshave been identified as a compelling national need for allenvironmental remediation research (1-3).

The objective of this investigation was to develop anddemonstrate a proof-of-principle for an inexpensive, direct,and effective in situ technique to monitor soil contaminantimmobilization nondestructively in the field using radio-isotope-spiked soil contained within a permeable polymermatrix. Soil encapsulated within a polyacrylamide gel in asmall cylinder was coined a permeable environmentalleaching capsule (PELCAP). Polyacrylamide gel has been welldeveloped previously for field applications to measurecontaminant availability in soil, sediment, and water (4-6).However, these previous applications have focused onencapsulating ion-exchange resins or other adsorbents inthe polymer as diffusive gradient thin films (DGT). Afterallowing equilibration of the encapsulated resin with the soil-water phases in the field, the DGT samples are retrieved anddestructively analyzed for the contaminants of interest andinterpreted as measures of contaminant availability. Thediffusive properties of contaminants in polyacrylamide havebeen extensively studied and provided much informationconcerning the relative inertness of polyacrylamide forcontaminant diffusion and for maintenance of ion-exchangeproperties of encapsulated resins over time. Our concept forPELCAP development has taken advantage of this extensiveinformation but proposes three concept-expanding envi-ronmental questions. If the functionality of ion-exchangeresins and other adsorbents can be maintained withinpolyacrylamide gels, then might not a natural ion exchangerand adsorbent like soil also maintain its functionality whenencapsulated in polyacrylamide? If contaminants can diffuseinto adsorbents encapsulated in polyacrylamide, then mightnot contaminants diffuse from encapsulated soil to someequilibrium with natural waters? Because a contaminantassay method like γ spectroscopy can determine residualcontaminant nondestructively, might not the kinetics ofapproach to some residual equilibrium contaminant con-centration through repeated retrieval and replacement be ameasure of its availability?

Methods and MaterialsSoil Preparation. Bulk samples (300 g each) of one well-characterized soil (7), sieved to <100 mesh, were spiked witheither 85Sr or 134Cs to a nominal activity about 0.1 µCi/g usingexcess tap water with stirring. After drying at 110 °C the bulksoils were remixed by sieving to <100 mesh again. Thermaltreatment of the radioisotope-spiked soil was carried out byheating 30 g subsamples in Pt crucibles at 1000 °C for 24 h;following cooling, thermally treated soils (colored red butremaining unsintered and noncohesive powders) were sievedagain to <100 mesh.

PELCAP Preparation and Deployment. A standardpolymer formulation (15.0% acrylamide, 0.3% methylene-bis-acrylamide, 0.2% tetramethylethylenediamine, and 0.18%ammonium persulfate in water) was selected for preparingPELCAPs in a polyacrylamide matrix and tested usinglaboratory sequential extractions for both 134Cs- and 85Sr-spiked soils. Two grams of a standard soil, sieved to <100mesh, were weighed into a cylindrical mold (a cutoff 10-mLplastic syringe using the plunger as the mold’s base) with 3.0mL of the freshly prepared monomer plus catalyst solution;the contents (soil suspended in the solution) were mixedwith a stainless steel spatula intermittently until the mixtureset to a gel, usually within 20 min. Upon expression from thesyringe barrel, a typical PELCAP cylinder was 1.38 cm indiameter × 3.5 cm in height with a volume of 5.23 cm3 anda surface area of 18.17 cm2 (Figure 1). Dimensions and weightsof the PELCAPs did increase as the gels swelled on deployment

* Corresponding author phone: (865)574-7265; fax: (865)576-8543;e-mail: spaldingbp@ornl.gov.

Environ. Sci. Technol. 2005, 39, 8912-8918

8912 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005 10.1021/es051379y CCC: $30.25 2005 American Chemical SocietyPublished on Web 10/15/2005

in ground or surface water or during laboratory extractions,increasing up to 80% in fresh weight.

Six replicate PELCAPs of each treatment/isotope com-bination were prepared so that statistically valid estimatesof sample isotope retention variance in the field could beobtained. Two suites of 54 PELCAPs each (six replicates of9 types: PELCAPs spiked with either or no isotope [3 types]and with either no, untreated, or thermally stabilized soils[3 types]) were deployed in both streamwater and ground-water. Each PELCAP was retained in a perforated (18 drillholes, 1/8-in. diameter, on walls, top, and bottom) 20-mLsnap-top polyethylene vial which, in turn, was restrained ina plastic test tube rack for submersion in either a streampool or groundwater collection sump on the Oak Ridge site.

Between August 10, 2003 and February 6, 2004 one suiteof 54 PELCAPs was deployed on the DOE Oak Ridgereservation in the northwest tributary (NWT) of White OakCreek (8) approximately 0.25 miles above its confluence withthe main channel of White Oak Creek. The initial four retrievalintervals (after 1, 3, 8, and 17 days) were actually carried outin the laboratory by immersion of each PELCAP in itsperforated container in a polyethylene bottle of 200 mL offreshly collected NWT streamwater; once the leachingbehavior and stability of the of the PELCAPs had beenestablished, all further leaching intervals were carried outdirectly in the stream pool after cumulative times of 24, 40,59, 87, 135, and 180 days. A grab sample of streamwater wascollected with each retrieval of PELCAPs, its temperaturetaken, and analyses of electrical conductance, pH, alkalinity,and hardness (9) carried out within 24 h. Additional aliquotsof the streamwater were stabilized by acidifying with nitricacid to pH <2 for later analyses of dissolved cations (Ca, Mg,Na, Fe, and Sr) via inductively coupled plasma massspectroscopy (Perkin-Elmer Elan 6000) and 90Sr via Cerenkovradiation emissions (10) after allowing 3 weeks for ingrowthof 90Y to seqular equilibrium. At each PELCAP retrieval, therack of perforated containers was carried to the laboratorywhere each PELCAP was removed and transferred to anothertared snap-cap sealed container within an hour such that

significant drying did not occur for weighing and γ activityassay. The γ assays of all 54 PELCAPs were completed within10 h of retrieval. Samples were transferred back into theirspecific perforated containers and test tube racks andreplaced in the streamwater within 12 h of their retrieval.

A second suite of 54 PELCAPs was immersed in ground-water in a collection sump for the core hole no. 8 (CH8)between September 9, 2003 and March 31, 2004. Thisautomatic sump collects 90Sr-contaminated groundwaterfrom two wells and an interceptor drain into an approximately2-ft deep by 10-ft diameter pool in a tank from which it ispumped via level-actuated control to the ORNL radiologicalwastewater treatment plant (11, 12). The CH8 groundwatersite offered a conveniently large, accessible, and rapidlyflowing “well” for testing PELCAPs by immersion in ground-water already contaminated with 90Sr with a well-documentedrecord of contamination and flow. Retrievals were carriedout after 1, 2, 6, 13, 28, 54, 89, and 180 days of cumulativeimmersion in the groundwater. Samples of groundwater werecollected initially and at each PELCAP retrieval date andanalyzed for the same parameters as the NWT streamwatersite.

Isotope Quantification. The γ activity of all PELCAPs,extracts, waters samples, soils, and extract filter materialswas carried out using γ-ray spectroscopy with a 3-in. diameter×3-in. high NaI well-type crystal with a multichannel analyzeremploying Accuspec (Canberra Industries) γ spectroscopysoftware and an automatic sample changer as describedpreviously (13). Assays were routinely performed for 10-minintervals which, for samples spiked with or containing thenominal 0.1 µCi (220 000 disintegration per min) of either85Sr or 134Cs, yielded whole-spectrum counting rates of about50 000 and 150 000 counts per min, respectively, abovebackground; initial uncertainties ((2σ) in the observed 10-min counts were, thus, 0.28% and 0.16%, respectively.Subsequent observed counts were corrected for isotope decayto a uniform reference date (i.e., the starting date of adeployment or extraction sequence). Because the half-livesof 85Sr and 134Cs are 65 and 753 days, respectively, final

FIGURE 1. Replicate PELCAPs after 6 months of deployment in groundwater including neat PELCAPs (top row), PELCAPs containingunheated soil (middle row), and PELCAPs containing thermally stabilized soil (bottom row).

VOL. 39, NO. 22, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8913

counting rates of PELCAPs fell to approximately 11 and 83%,respectively, of their initial counting rates. Rather than convertobserved counts into activities which would involve dailycorrections for efficiency and possible shifts in photopeakgain, all PELCAP sample and extract counts were referencedto those of an average of a triplicate set of undeployed PELCAPstandards for 85Sr or 134Cs or unspiked backgrounds sealedin snap-cap counting containers and prepared at the sametime as the suite of deployed PELCAPs. Each observedPELCAP sample count was simply divided by the averagecounts in the sealed standard PELCAPs to yield directly thefraction of starting activity present at each assay date.Corrections for small daily shifts in counting efficiency and/or gain were thus not necessary. The γ spectrum was dividedinto three regions-of-interest to allow independent analyses(spectral resolution) of γ-ray photopeaks if necessitated bycross-contamination of PELCAPs (particularly of unspikedPELCAPs) within a group or uptake of other γ-emittingisotopes from the stream or groundwater. However, no cross-contamination or uptake of other γ-emitting isotopes wasobserved in either suite of PELCAPs and, thus, γ spectralresolution was not required. Counts in all 2048 channels ofthe whole spectrum (0.05-2.0 MeV) were simply totaled,and average whole-spectrum backgrounds in daily assayswere subtracted to obtain the net counts. To correct forpotential assay geometry differences, a group of triplicatestandards for both 85Sr and 134Cs were prepared in each ofthe following configurations: 4 g of neat PELCAP, 4 g ofPELCAP plus 2 g soil, 2 g of soil alone, 20 mL of 0.1 N CaCl2,and adsorbed on a 0.06 g (5 cm diameter) nitrocellulose filter(used for extract filtering of any suspended matter). Geometryfactors were used to correct counts collected under differinggeometries (e.g., extracts vs PELCAPs) to a common geometryfor direct comparison as fractions of starting activities.

Sequential Extraction Procedures. Additional suites ofPELCAPs were prepared for laboratory sequential extractions(13). Sequential extractions were also carried out on equiva-lent amounts of soil without encapsulation in polyacrylamide.A suite of PELCAPs, containing triplicate samples of eachtreatment (neat, unheated, and heated soil) and each isotope(85Sr or 134Cs) and reagent blanks (no soil or PELCAPs), werecarried through a previously developed sequential extractionprotocol employed for minerals and soils (13). In summary,this sequence encompasses five successive extractions withwater (each at 10 mL of extract per gram of soil) to determinethe water soluble fraction, followed by five extractions with0.1 N CaCl2 to determine the cation exchangeable fraction,followed by five extractions with 0.2 N HCl to determine theacid soluble fraction, followed by γ activity assay of theresidual soil to determine the residual fraction. Followingcompletion of the sequential extraction, the nitrocellulosefilter was cut from the disposable filter unit, dried, weighed,and assayed for isotope activity as above and reported as the“suspended fraction”. The initial extraction with each solutionwas routinely carried out for 16 h (overnight) with thefollowing four extractions for 1 h each when applied tounencapsulated soil; however, for direct comparison ofencapsulated versus free soil, all extraction intervals werelengthened to 24 h each to allow sufficient time for isotopesto diffuse from the gel and equilibrate with the extractingsolutions. Triplicate samples of each of the 9 types of PELCAPsdeployed for 6 months at the CH8 site were subjected to thisextraction sequence to determine if any changes in thedistribution of their residual 85Sr or 134Cs had occurred. Inaddition to the usual assay for γ activity as described above,each extract was assayed for 90Sr activity using its Cerenkovemission (10, 14) and standard solutions of 90Sr in 0.1 N HClin 20-mL polypropylene scintillation vials. Additional stan-dards, containing either 85Sr or 134Cs with γ activities in excess

of any observed in any PELCAP extract, failed to exhibitCerenkov emissions sufficient to interfere with the 90Sr assay.

Cation Exchange Capacity. Cation exchange capacity(CEC) of both the untreated and thermally treated soil wasmeasured using a standard 85Sr-labeled 0.1 N SrCl2 (15) forboth unencapsulated and encapsulated soil. The net amountof Sr in mequivs (+) adsorbed from this saturating solution,after correcting for residual solution retained in a soil or ina soil plus polymer matrix, was computed as the CEC. Furtherdetails of this method and calculations are presented in theSupporting Information.

Data Modeling. The equations describing diffusion fromregular geometries are well developed, and solutions to theequations have been reported elsewhere (16-19). Thesolution presented by Anders et al. (18, eq 14, see detailedsolution in the Supporting Information) describing thefraction of solute remaining in a finite right cylinder as afunction of time was fit to the observed PELCAP data byadjusting the value of the diffusion coefficient, D (L2/T). Theeffective diffusion coefficient obtained (Deff) incorporates theeffects of tortuosity through the polyacrylamide/soil matrixas well as sorption onto the polyacrylamide and encapsulatedsoil. A partitioning coefficient describing solute partitioningbetween sorbed and aqueous phases can be obtained from

Rs ) partitioning coefficient; Dref ) reference value for thediffusion coefficient using neat PELCAPS; Deff ) effectivediffusion coefficient (16) for soil-containing PELCAPS. Valuesof Rs were calculated for 134Cs and 85Sr using data from lableaching of neat PELCAPS to establish the reference value,Dref. The partitioning coefficients determined from thisprocedure are analogous to equilibrium distribution coef-ficients (Kd’s) of sorption isotherms but should not beconsidered true equilibrium Kd’s.

Results and DiscussionBoth 134Cs and 85Sr were quantitatively leached (>99%removed) from the PELCAPs within a period of 1-4 daysduring repeated leaching into 0.1 N CaCl2 solution (Figure2). Although each radionuclide was added to PELCAPs prior

FIGURE 2. Retention of 134Cs and 85Sr within PELCAPs without soilfollowing repeated leaching with 0.1 N CaCl2. Symbols representthe average of replicates (n ) 2); error bars indicating the rangeof measurements are smaller than the symbols. Lines indicate themodel-calculated leaching behavior from a right cylinder using theestimated effective diffusion coefficients indicated.

Rs )Dref

Deff- 1 (1)

8914 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005

to polymerization, subsequent leaching behavior of eachradionuclide indicated minimal reaction with the finalpolymer gel and continued freedom of the radionuclides todiffuse from the gel. 134Cs and 85Sr were added as monovalentand divalent cation chlorides, respectively, and, thus, wouldnot be expected to react with the neutral hydrophilicpolyacrylamide matrix. The kinetics of leaching for eachradioisotope from PELCAPs are described well by thediffusion model for small right cylinders with best-fit valuesfor the effective diffusion coefficients only slightly lower thanthose exhibited in water. The average values of the diffusioncoefficients for Cs and Sr ions were found to be (4.8 ( 0.2)× 10-6 cm2/s and (1.14 ( 0.06) × 10-6 cm2/s, respectively(Figure 2). Chang et al. (5) reported similar values of 1.92 ×10-5 and 7.72 × 10-6 cm2/s, respectively, for Cs and Sr ionsin similar polyacrylamide gel materials; diffusion coefficientsfor Cs and Sr ions in water at 25 °C have been measured at2.07 × 10-5 and 7.97 × 10-6 cm2/s, respectively (20). Thus,the present estimated values for the diffusion coefficients ofCs and Sr ions in PELCAPs were sufficiently large thatcontinuous leaching from such a matrix of such cylindricaldimensions could be expected to be nearly complete withina period of 24 h. Differences in gel preparation andexperimental objectives between this study and that of Changet al. (5) may account for our smaller estimated value for Deff.Chang et al. (5) have focused on the uptake of ions by ionexchangers encapsulated in the gels. Therefore, prior to theiruse, they soaked the polyacrylamide gels in either ultrapurewater, dilute NaNO3 solution, or synthetic lake water tominimize ion uptake artifacts caused by residual reagents inthe gels. They noted that this washing decreased the uptakeof Sr by the polyacrylamide. As this study was focused on therelease of ions from spiked gels and their encapsulated soilsrather than the uptake of ions from solution, such apreconditioning step would have the undesirable effect ofremoving the ions of interest prior to the study period.

Encouraged that the polyacrylamide gel matrix exhibitedlittle reactivity with 134Cs or 85Sr ions either during or after

polymerization, the reactivity of these radionuclides withsoil was then examined in the presence of in situ polymerizedacrylamide and compared to their reactivity with soil in theabsence of polyacrylamide. The sequential extraction be-haviors of these radioisotopes from radionuclide-spiked andencapsulated soils and from the same soils without polymerencapsulation were found to be quite similar (Figure 3 andthe Supporting Information). This sequential extractionprocedure (13) was originally developed using 1-h extractionintervals for each of four successive extractions with each ofthree solutions (water, 0.1 N CaCl2, and 0.2 N HCl) followingan initial extraction with each solution for 16 h. For the presentinvestigation, a 24-h equilibration interval for each of the 15total successive extractions was found necessary to allowsufficient time for diffusive equilibration from the PELCAPsinto the leaching solutions (Figure 2). With the use of 24-hextraction intervals, sequential extraction profiles for bothradioisotopes from polymer-encapsulated soils were foundto be quantitatively similar to those of the nonencapsulatedsoils. First, radioisotope-spiked PELCAPs without soil ex-hibited quantitative removal from the polymer after com-pleting the five leachings with 0.1 N CaCl2 as did extractionreagent blanks (no polymer and no soil). Second, whetheror not encapsulated in PELCAPs, 85Sr exhibited maximalextractability from soil into the 0.1 N CaCl2 leachates as wouldbe expected for exchangeable soil cations. Nevertheless, thetest soil also exhibited a small acid soluble fraction (about5.8-8.9%) of the added 85Sr that required leaching by thefinal 0.2 N HCl extractions. In contrast to 85Sr, 134Cs exhibitedstrong and specific adsorption to the test soil as would beexpected for soils containing micaceous minerals (21); 134Cswas quantitatively retained by the test soil through the entiresequential extraction (Figure 3 and the Supporting Informa-tion) whether or not the soil was encapsulated in polyacryl-amide. This similarity of soil extractive behavior providesevidence for the inertness of the polymer itself to radioisotopeadsorption as well as stability of the polymer to the extractionreagents. In addition, the polyacrylamide matrix did not

FIGURE 3. Cumulative sequential extraction of 85Sr and 134Cs from PELCAPs in the presence and absence of an Oak Ridge test soil.

VOL. 39, NO. 22, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8915

interfere with the expected soil adsorption mechanisms, i.e.,cation exchange for 85Sr and specific adsorption for 134Cs.

Although it was encouraging that both 134Cs and 85Sr werereleased from either neat or soil-containing PELCAPs similarlyto nonencapsulated soil, the uptake of 85Sr by soil previouslyencapsulated in PELCAPs was also determined quantitativelyby measuring each soil’s cation exchange capacity (CEC)using a SrCl2 solution labeled with 85Sr (15) and comparingto the CEC measured with soil alone. Measurement of theunheated soil CEC, with or without encapsulation in PEL-CAPs, were found to be similar, i.e., (13.3 ( 0.2) versus (12.7( 0.3) mequiv (+)/100 g as well as similar to the CEC ofunheated soil in PELCAPs subjected to 6 months of in situleaching in streamwater (13.0 ( 0.2 mequiv (+)/100 g). Theheated test soil exhibited a markedly reduced initial CEC of1.9 mequiv (+)/100 g which increased to 3.8 mequiv (+)/100g when measured in week-old “fresh” PELCAPs. After 6months of deployment in surface water, the heated soil inPELCAPs apparently recovered to a CEC of 7.9 mequiv (+)/100 g. The apparent calculated CEC of neat PELCAPs was 0.9

( 0.3 mequiv (+)/PELCAP (no soil basis) supporting theconclusion that the polymer gel remained relatively inert yetvery permeable to these ionic species and did not interferewith the cation exchange properties of the soil eitherimmediately after encapsulation in polyacrylamide or afterlong-term in situ field leaching.

The in situ retention of both isotopes by PELCAPs in fieldground and surface water displayed a broad range ofbehaviors but were generally well described by the diffusionmodel (Figure 4 and the Supporting Information). Correlationcoefficients (Table 1) between observed and model-fitretained fractions often indicated that the model accountedfor more than 90% (r2 > 0.90) of the temporal variance. Threedistinct ranges of diffusion coefficients were encounteredduring modeling and are depicted by the selected treatmentsin Figure 4. First, all of the radioisotope-spiked neat PELCAPtreatments without soil rapidly lost their radionuclides, oftenwith <1% retention within 2 days of immersion in groundor surface water. The modeled diffusion coefficients for theseneat PELCAPs without soil typically fell in the 10-6 cm2/smagnitude. All of these soil-free PELCAP models alsoexhibited high correlation coefficients likely resulting fromthe large range of retained fraction values observed (from 1initially to near 0 for the several later intervals). The secondtype of leaching response was exhibited by the PELCAPtreatments containing 85Sr-spiked soil (unheated) deployedat either the NWT or CH8 sites. For these types of PELCAPs,85Sr leaching was slower, requiring 90 days or more to reacha relatively stable and low retained fraction (5-10%). Best-fit diffusion coefficients with values of 10-7 to 10-8 cm2/smagnitude resulted in models with high correlation coef-ficients (Table 1). Such relatively high coefficients likely alsoresulted from the wide range of observed retained fractionsthroughout the testing interval. This type of retention functionis consistent with the mechanism of 85Sr retention by soilcation exchange in equilibrium with a low ionic strengthleachate like ground or surface water (see the SupportingInformation). The third type of PELCAP retention responsewas observed for treatments composed of 134Cs-spiked soil(heated or unheated) or of thermally stabilized (heated) 85Sr-spiked soil. These treatments exhibited near quantitativeretention of radioisotope over the entire 6-month interval ofin situ leaching by natural waters at either Oak Ridge site.Best-fit model diffusion coefficients fell in the range of 10-13

to less than 10-15 cm2/s. The associated correlation coef-ficients were generally much smaller than for the PELCAP

TABLE 1. Best-Fit Effective Diffusion Coefficients (Deff ( 1 Standard Error) Used to Describe Retained Isotope Fractions inPELCAPs during Cumulative Leaching at Two Oak Ridge Field Sites (NWT and CH8) and in Laboratory (LAB) Tests

location isotope soil Deff (cm2/s) R2data

pointsfinal fraction

retained Rs

LAB 85Sr none (1.14 ( 0.06) × 10-6 0.9933 23 0.004LAB 134Cs none (4.8 ( 0.2) × 10-6 0.9970 23 0.01

NWT 85Sr none (1.02 ( 0.03) × 10-6 0.9973 64 0.002 0.12CH8 85Sr none (2.39 ( 0.02) × 10-6 0.9999 54 0.002 -0.52

NWT 134Cs none (2.50 ( 0.04) × 10-6 0.9998 64 0.0003 0.92CH8 134Cs none (5.5 ( 0.3) × 10-6 0.9999 53 0.0003 -0.13

NWT 85Sr unheated (2.5 ( 0.1) × 10-8 0.9691 65 0.086 45CH8 85Sr unheated (1.14 ( 0.05) × 10-7 0.9824 54 0.055 9

NWT 134Cs unheated (1.9 ( 1.5) × 10-14 0.081 64 0.993 250 000 000CH8 134Cs unheated (1.8 ( 0.5) × 10-13 0.3163 54 0.993 27 000 000

NWT 85Sr heated a a 65 0.993 aCH8 85Sr heated a a 54 1.005 a

NWT 134Cs heated (1.8 ( 0.9) × 10-15 0.2293 56 0.996 3 000 000 000CH8 134Cs heated (1.0 ( 0.3) × 10-13 0.1524 53 0.996 50 000 000

a No meaningful values for Deff or its dependent variable, Rs, could be calculated for this group because isotope retention fractions neitherdeclined with time nor were significantly different from 1 at any interval, whereas the diffusion model is constrained to decrease with time.

FIGURE 4. Retention of 134Cs and 85Sr by selected PELCAPs during6 months of in situ leaching by stream and groundwaters. Symbolsand error bars represent the average ( one standard deviation ofreplicate deployed cylinders (n ) 6). Lines indicate the model-calculated leaching behavior from a right cylinder using theestimated effective diffusion coefficients indicated.

8916 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005

treatments where much smaller radioisotope retentions wereobserved. Several of the retention fractions for this class ofbehavior appeared to be independent of leaching time, i.e.,the retained fraction was statistically indistinguishable from1 after the 6-month testing interval. Diffusion coefficients ofsuch magnitude would require leaching intervals in the 100-1000 year range to attain the lower retained fractions observedfor the other two types of PELCAP treatments discussedabove. Although the present data has too few observationsto merit any sophisticated chemical modeling of the soil-solution equilibria, a simple quantitative descriptive inter-pretation is possible using the partition coefficient, Rs,described by eq 1.

Regardless of the success in fitting PELCAP leachingbehavior to diffusive retention models or interpreting ra-dionuclide releases with solid-solution equilibria descriptivestatistics such as Rs, the field performance of the PELCAPsoffered some encouraging empirical characteristics for theirpotential generic application to evaluate soil contaminantimmobilization in situ. First, the visual appearance of thePELCAPs after 6 months of exposure in ground or surfacewater (Figure 1) was virtually indistinguishable from theirinitial appearance. In particular, the PELCAPs deployed inthe northwest tributuary of White Oak Creek were subject toa considerable range of streamflow rates and associatedsuspended particulates and leaf litter fall frequently withopaque streamwater and intermittent burial in litter and bedgravel. The PELCAPs in Figure 1 were simply rinsed for a fewseconds in a water flow while gently rubbing their surfaceswith vinyl-gloved fingers prior to photographing. Thiswashing technique removed a minor surface film of par-ticulates yielding clear neat PELCAP gels. There was noclouding or discoloration of the interior of the PELCAPs withmicrobial growth or entrapped particulates. These PELCAPsendured all the mechanical and physicochemical stresses ofthe stream and groundwater sump environment withoutapparent damage. Second, the quantitative retention of 134Csby soil in PELCAPs and of 85Sr in thermally stabilized soil inPELCAPs (Table 1) also indicates that the soil solids werequantitatively retained by the polyacrylamide matrix during6 months of field deployment. In only one of 108 PELCAPsdeployed in the field was a physical loss of soil/gel materialobserved (an approximately 1% loss of 134Cs from one PELCAPpresumably lost from a visible defect at a wall-base border).PELCAPs did swell somewhat when initially sampled fromstream or groundwater but, when weighed regularly in theirassay containers, maintained constant weight thereafter.Third, there was no cross-contamination of PELCAPs al-though each suite was deployed within in a test tube rackwith approximately 1 in. between adjacent samples; blankneat PELCAPs and PELCAPs with unspiked soils weremaintained with each suite of deployed samples, and neither134Cs nor 85Sr activity was observed in these blanks at anysampling interval. Fourth, variations within replicate PELCAPtreatments were extremely small (i.e., the overall standarddeviation within PELCAP treatment groups was 0.0044 ofthe fraction retained). Such reproducibility bodes well forthe capability to observe small statistical differences overtime when comparing several contaminant immobilizationmethods at one field site or differences among field sites forone contaminant immobilization method.

Although we have presented only evidence for theapplication of PELCAPs to assess the in situ leaching ofradioactively labeled contaminants from soils treated priorto encapsulation, potential direct applications seem possiblefor assessment of other stable contaminants using isotopiclabeling as well as contaminant uptake applications withother radionuclides. Other analytical methods, includingdestructive methods, may expand applications for PELCAPsto other contaminants provided potential contaminant

interactions with the polymer matrix are understood as wellas potential indirect effects through possible changes in thechemical, particularly the redox, state within the polymermatrix. However circumscribed our assessment of the presentfindings, potential applications abound for encapsulated soilsamples to assess contaminant uptake and release in thesoil-groundwater environment.

AcknowledgmentsThis research was sponsored by the Laboratory DirectedResearch and Development Program of the Oak RidgeNational Laboratory (ORNL) managed by UT-Battelle, LLC,for the U.S. Department of Energy under Contract DE-AC05-00OR22725.

Supporting Information AvailableDetails of the cation exchange measurement methods asadapted for polymer-encapsulated soils, the detailed solutionto the diffusion equation used to model isotope retention inPELCAPs, and data on the simultaneous uptake of 90Sr byPELCAPs from 90Sr-contminated groundwater during fieldtesting; also contained are detailed experimental measure-ments of all isotope retentions by soil and PELCAPs in bothlaboratory extractions and in situ field leaching by groundand streamwater as well as temporal stream and groundwaterchemical characteristics. This material is available free ofcharge via the Internet at http://pubs.acs.org.

Literature Cited(1) MacDonald, J. A. Evaluating natural attenuation for groundwater

cleanup. Environ. Sci. Technol. 2000, 34, 346A-353A.(2) National Research Council (NRC). Natural Attenuation for

Groundwater Remediation; National Academy Press: Wash-ington, DC, 2000.

(3) Brady, P. V.; Spalding, B. P.; Krupka, K. M.; Waters, R. D.; Zhang,P.; Borns, D. J.; Brady, W. D. Site Screening and TechnicalGuidance for Monitored Natural Attenuation at DOE Sites;SAND99-0464; Sandia National Laboratories: Albuquerque, NM,1999.

(4) Zhang, H.; Davison, W. Performance characteristics of diffusiongradients in thin films for the in situ measurement of tracemetals in aqueous solution. Anal. Chem. 1995, 67, 3391-3400.

(5) Chang, L.-Y.; Davison, W.; Zhang, H.; Kelly, M. Performancecharacteristics for the measurement of Cs and Sr by diffusivegradients in thin films (DGT). Anal. Chim. Acta 1998, 368, 243-253.

(6) Murdock, C.; Kelly, M.; Chang, L.-Y.; Davison, W.; Zhang, H.DGT as an in situ tool for measuring radiocesium in naturalwaters. Environ. Sci. Technol. 2001, 35, 4530-4535.

(7) Spalding, B. P.; Jacobs, G. K. Evaluation of an in Situ VitrificationField Demonstration of a Simulated Radioactive Liquid WasteDisposal Trench; ORNL/TM-10992; Oak Ridge National Labora-tory: Oak Ridge, TN, 1989; 1-158.

(8) Cerling, T. E.; Spalding, B. P. Distribution and relationship ofradionuclides to streambed gravels in a small watershed.Environ. Geol. 1982, 4, 99-116.

(9) American Public Health Association (APHA). Standard Methodsfor the Examination of Water and Wastewater; APHA: Wash-ington, DC, 1995.

(10) Haberer, K. Measurement of beta activities in aqueous samplesutilizing Cerenkov radiation. Atomwirtschaft 1965, 10, 36-43.

(11) Foster Wheeler Environmental Corporation (FWEC). RemovalAction Plan for Corehole 8 in Waste Area Grouping 1 at OakRidge National Laboratory, Oak Ridge, Tennessee; ORNL/ER-283; Martin Marietta Energy Systems Inc.: Oak Ridge, TN, 1994;1-15.

(12) Bechtel Jacobs Company (BJC). Removal Action Report for CoreHole 8 Plume Source (Tank W-1A) at Oak Ridge NationalLaboratory, Oak Ridge, Tennessee; DOE/OR/01-1969&D2; Bech-tel Jacobs Company LLC: Oak Ridge, TN, 2002; 1-20.

(13) Spalding, B. P. Fixation of radionuclides in soil and minerals byheating. Environ. Sci. Technol. 2001, 35, 4327-4333.

(14) Spalding, B. P.; Munro, I. L. Determination of the ArealDistribution of 90Sr in Groundwater via Single-Use Boreholes;

VOL. 39, NO. 22, 2005 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 8917

ORNL/TM-8991; Oak Ridge National Laboratory: Oak Ridge,TN, 1984; 1-42.

(15) Francis, C. W.; Grigal, D. F. A rapid and simple procedure using85Sr for determining cation exchange capacities of soils andclays. Soil Sci. 1971, 112, 17-21.

(16) Carslaw, H. S.; Jaeger, J. C. Conduction of Heat in Solids, 2nded.; Oxford University Press: Oxford, U.K., 1959.

(17) Crank, J. The Mathematics of Diffusion, 2nd ed.; OxfordUniversity Press: Oxford, U.K. 1975.

(18) Anders, O. U.; Bartel, J. F.; Altschular, S. J. Determination of theleachability of solids. Anal. Chem. 1978, 50, 564-569.

(19) American Nuclear Society (ANS). Measurement of the Leach-ability of Solidified Low-Level Radioactive Wastes by Short-Term

Test Procedure; ANSI/ANS-16.1-2003; American Nuclear Soci-ety: La Grange Park, IL, 2003; 1-33.

(20) Li, Y. H.; Gregory, S. Diffusion of ions in seawater and in deep-sea sediments. Geochim. Cosmochim. Acta 1974, 38, 703-714.

(21) Francis, C. W.; Brinkley, F. S. Preferential adsorption of 137Cs tomicaceous minerals in contaminated freshwater sediment.Nature 1976, 260, 511-513.

Received for review July 15, 2005. Revised manuscript re-ceived September 7, 2005. Accepted September 9, 2005.

ES051379Y

8918 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 39, NO. 22, 2005