Uranium and Neptunium Desorption from Yucca Mountain Alluvium

~ m 4 b 5 7 aQ C N ~3 / 1 b / o ~ Uranium and Neptunium Desorption from Yucca Mountain Alluvium Cynthia D. Scism, Paul W. Reimus, Mei Ding and Steve J. chipera Los Alamos National Laboratory: P.O. Box 1663, Los Alamos, NM, 87545, [email protected] Abstract - Uranium and neptunium were used as reactive tracers in long-term laboratory desorption studies using saturated alluvium collectedfrom south of Yucca Mountain, Nevada. The objective of these long-term experiments is to make detailed observations of the desorption behavior of uranium and neptunium to provide Yucca Mountain with technical bases for a more realistic and potentially less conservative approach to predicting the transport of adsorbing radionuclides in the saturated alluvium. This paper describes several long-term desorption experiments using a flow-through experimental method and groundwater and alluvium obtainedfrom boreholes along a potential groundwater flow path from the proposed repository site. In the long term desorption experiments, the percentages of uranium and neptunium sorbed as a firnction of time after different durations of sorption was determined. In addition, the desorbed activity as a function of time was Jit using a multi-site, multi-rate model to demonstrate that drfferent desorption rate constants ranging over several orders of magnitude exist for the desorption of uranium from Yucca Mountain saturated alluvium. This information will be used to support the development of a conceptual model that ultimately results,in effective Kd values much larger than those currently in use for predicting radionuclide transport at Yucca Mountain. , I. INTRODUCTION The saturated alluvium south of the proposed highlevel nuclear waste repository at Yucca Mountain, Nevada represents the final feature of the Lower Natural Barrier with characteristics and processes that can substantially reduce radionuclide migration before reaching the regulatory compliance boundary. The objective of this work is to demonstrate that radionuclide retardation in the saturated alluvium is likely to be significantly higher than is currently assumed in Yucca Mountain Project (YMP) models. [I] This work also involves the development of an improved reactive transport modeling approach that is compatible with YMP saturated zone transport process models. [I] Our experimental efforts have focused on the radionuclides of uranium and neptunium because of their high solubility, relatively weak sorption, and their high potential contributions to offsite dose in the Yucca Mountain models. The focus has also been on desorption measurements rather than sorption measurements, as we hypothesize that desorption rates likely control radionuclide fate and transport to a much greater degree than sorption rates. We have developed a flow-through experimental desorption method that provides a nearly continuous measure of desorption rates over a long period of time. [2] Almost all previous experiments conducted by the Yucca Mountain Project [3,4] have focused on batch sorption measurements or very short-duration desorption measurements, which tend to significantly underestimate radionuclide sorption parameters because they do not interrogate the fraction of radionuclide mass that dcsorbs very slowly. Quantitative X-ray diffraction and other methods are being used to characterize the alluvium used in the experiments. To support the interpretation of the experiments and to put the experimental results into a predictive context, we have developed a generalized sorption residence time distribution modeling approach to account for the drastically different desorption rates that have been experimentally observed. [5] This approach is consistent with a conceptual model that involves multiple sorption sites where rates of sorption onto the sites are similar but rates of desorption differ dramatically. The model so far has been able to account qualitatively for desorption behavior observed in the flow-through desorption experiments as well as some anomalous radionuclide transport results in column transport experiments that did not exhibit standard equilibrium or first-order reaction rate behavior. Through continued experimentation, we are developing a mechanistic basis for the model and validating the approach. 11. MATERIALS AND METHODS 1I.A. Water and Alluvium The alluvium used in the experiments is from drill cuttings obtained from NC-EWDP-191M I A, NC-EWDP22SA and NC-EWDP-IOSA (Figure 1) at depth intervals of 225.6-227.0 meters (740.0-745.0 feet), 169.93-170.69 meters (557.5-560 feet) and 207.26-208.79 meters (680685 feet) below ground surface, respectively. Alluvium samplcs wcrc sicvctl. ant1 the sirc nngc ktwcen 75-2000 .urn rvas prnportionnlly rect~rnhincrl. Thc hulk mincralagies nl'the t hwc rccomhincd allu~iurnsarnplt%, 3.; tkctcrrn~ncdtry quantitative S-my difliaction, arc I~stctl in T;ihlc I. 1:xpcriments wtth thc 1 9 l M l A and 32SA nlIwium wcrc cnnducrerl i~singn low cnrlionntc I shallow 7onc t'rnrn NC-I'WIIP-It)D tl:~curc 1). rcfcrrc(1 to a s Lnnc 1 ) ant1 a high carhonarc tdct? /one from XC-EWIIP- t OD, rcfitrcd to as %onc 4) ~mtuid\r~ntcr.F-:xperimcnts usinp 1 0 S h nllr~vium wcrc cotrtli~ctcrl with water from rhc sams lir>rchnlc ns lhc Alt~vium. The rmtlntlwarcr was iiltcrcrl wing a 0.2 prn filter lleforu u.ic. T11c tolrll innrganic carbon lcnrhonatc pluq hicnrhnnntc) cnnccntratinns 01' tVic 1g13 Zone 1. 101) Zcrnc 4 nntl IOS ~~ountlwatcrs. as mcasurctl In Ihc I:ihontny. arc. IRq, 21:. and 11111 rn.pl.. rcspcctivull;. ' n c p l Is ot' thc 19U Zonc 1, l9D Zunu 4 nnd to5 ~~oiindwntcrr. a5 mcnsurcd tn ~ h lahontory. c arc X.hn, A.R5 and 7 . X . rccprrtivcty. Inn~cstrengths nf thc 19D Zone t 19D Zone 4 and 105 gmundu atcn arc 0.0044, 0..1104?and 0.OOill mnl x [,-I. rcspccr ivrly Thc nhjccrive of the yraimtl cvpcrimcnts usine gmt~ndwatcr from Zone I 2nd Zonc 4 uT lC)I) is lo ~nvcqtigalethc d T ~ or' t umnyl nntl neptunyl cnrbnnnte ccrmplcx fonn:lrinn on wrptinn ond Imp,-term rtrunrptian, f* l. ~~ c'' LJc VI r 1r:icrr st>lutlr~ns wcrc prcplsct! 1 1 r~l ~ l t III!: ~t - ' W(Vll (aq I?(I:(Y(l,\: 111 ililrltc IIUO,) ctncL r;olul~on Hn n ttc ohtnr ncil f'rnm l~ntnpc I'rodtictc I.;ihrrmtc~r~<*.: rmundivntors ttc%crilrcrl ahavc. *Tllc " ' \ p ~ V ) 1MCCl. .;otr~t~nns were prcpawtl hy rl~lurinp.a ' ' ' N ~\:I ( r a.; NpO? ' in I 1C'I ) hydmchlnnr tit! ~ttick ~nh~tinnohtn~nc~lt'rilln I,os Afarnm Yationill I.nhnr~tnry ! t i rl~~tiltcrlw:!!cr ? r l lnnkc a srock snl~~riun (11' 2 * 1 0 1 tr~crlxl,' ' Np, l,o\vcr ccrnccntratirmu of ncpt~ini~rm tr;lcur snlutionq ~ v c r vm:~dc by cliluticm r i T t)ic ?'In: riinl-I.' Np 5tncL unlt~tivr~ wirh I hc gr~utlw~tc17idwcr~tlcd nhtrvc. 1 - h ~ fillill cancentrntinnc nl'c~chof tlw -"''I)( 1~ 1 an,! -'*N~{ V ) (racer snlutinns nrc Rlvcn In I';iT?le II. 'l'ta sturllcs tlcw-nhrd hcrc~rrwcro pcrformcd i~ntlcr;imb~cnrcontlittuns :lntl t t i t , g~oundwntcr.tnccr wlut ion% ~tscdin thc ckpcrirncntk ;Ire :~ssurnctl 10 he in cqi~ilihririm wrlh tllc nrnhicnl iiirno~phcze((.midifin? corid~ti~nx, O.m?''h('0.) 1 lie \ p ~(ll~lhill!y at tht.\c c~pcmmcnt;tl cnnrl~!ions 1;. ;ippmrin?atcly 1 * l C) 'Z1. [ h ] P w v ~ ~snrpt~~ln ~ ~ * contml crpcrimcnts with qirn~li~r cnnccntmlir~nrnt' NIT i w ill1 1 ~ 1 atluv~umprcscnl 'I slinwctl no cvlttmcc of prcc ~pi~:ltlc~n or sorption tt) contn~ncrw;rlls. 131 ,.? . -1, 5 3 m w m urn wnim I wrm W W rs='ng Imb Fieurc 1 , Map s!iclivin? location of horchnlcs in rcli~tionto tlie rcpocifnty qitc anti a potcnllnl prot~nrlwatcr flow path. \Vc!l< V C ' - E ~ V ~ ~ l ' - lnritl t ~ ~ 3N t + I'\i7LlP- I ~ ~ l \ l larc h loc:~tctl111 S ~ t e19. YC'-F:U't)P-?2SA ill Sitc 2 -:!nclKC'I:\Vnl'- l OSA at thc locar~undcnotcd by 10%. TABLE I. Quantitative X-ray Diffraction results for alluvium used in long-term uranium and neptunium desorption. Mineral Phase Well Well Well NCNCNCEWDPEWDPEWDP191MlA 22SA 1OSA (Wt %) (Wt %) (Wt yo) Quartz 15.3 10.1 8.7 Plagioclase K-Feldspar Clinoptilolite Mica Kaolinite Cristobalite Tridymite Opal-CT Hematite Smectite 4.6 19.4 8.O Total 100.6 99.4 97.4 I1.C. Data Interpretation Methods Results of long-term desorption of uranium after fourteen days of sorption were analyzed using a mathematical model written in FORTRAN that utilizes multiple sorption sites with different first-order forward and reverse reaction rate constants. The adjustable parameters are sorption rate constant, desorption rate constant, number of different types of reaction sites, and maximum sorption capacity for each type of site. The model was used to fit the concentrations in samples collected from desorption columns as a function of time, allowing for flow rate changes and flow interruptions. Equations (1) and (2) below are used in the model to fit the experimental column desorption data as activity desorbing as a function of time. The experimental data were fit as closely as possible using one type of sorption site. Additional site types were added as necessary to. improve the fit, while balancing the amount of uranium on the alluvium after the sorption step and the cumulative amount of uranium desorbed from the alluvium. II.B. Long-term Desorption Experiments Desorption rates were determined by long-term desorption of uranium and neptunium from the alluvium samples. First the alluvium was brought into contact with a tracer solution containing a single radionuclide of interest. The radionuclide was batch sorbed to the alluvium for one to fourteen days and the supernatant was decanted and either centrifuged or filtered to remove fine particles before being analyzed for radionuclide concentration to determine the partition coefficient (ratio of sorbed to nonsorbed radionuclide, or Kd value, mug). A partition coefficient implies a linear sorption isotherm, which should provide a good approximation of the sorption isotherm at the relatively low concentrations used in the experiments and the even lower concentrations that would likely occur in the saturated zone. Also, isotherm nonlinearity is expected to be a second-order effect compared to sorption heterogeneity due to variability in mineralogy and water chemistry in the alluvium. The radionuclide-bearing alluvium and tracer free groundwater were added to a small column and placed on a rocking shaker to maximize alluvium-solution contact. Tracer-free groundwater was eluted through the column and collected in fractions. The activity in the samples was measured by a Packard 2500TR liquid scintillation counter. The groundwater zone, sorption period, tracer solution concentration and liquid to solid ratio used for the experiments are listed in Table 11. I where, C = concentration out of column, CPMImL Cin= cone. in solution flowing into column, CPMImL Si = amount sorbed to site i, CPMIg Si,, = maximum sorption capacity of site i, CPMIg V = volume of solution in column, mL Q = flow rate through column, mL1hr M = mass of solid, g ki = sorption rate constant for site i, mLhr kri = desorption rate constant for site i, gfhr CPM = counts per minute. t = time, hr 11. RESULTS To date, we have conducted 18 long-term uranium desorption experiments and 12 long-term neptunium desorption experiments (Table 11). These experiments have been conducted with 3 different alluvium samples and 3 different waters collected from the alluvium south of Yucca Mountain. Each experiment has consisted of a TABLE 11. Experimental parameters used in uranium and neptunium batch sorption and long-term desorption column experiments. Alluvium/ Sorption RadioTracer Liquid: 19D Ground Period nuclide Solution Solid Water Zone (days) Conc. Ratio . - . (M-~) (mug) 19IMIN1 1 U-233 3 2.22:l 19IMlN1 3 U-233 3 2.24: 1 19IMIN1 14#1 U-233 1 7.55:1 19IMIN1 14#2 U-233 1 7.53:'l 191MlN4 1 U-233 3 2.38:l 19IMlN4 3 U-233 3 2.33:l 19IMlN4 14#1 U-233 1 7.52: 1 19IMlN4 14#2 U-233 1 7.58:l 19IMlNl 2 Np-237 3 2.38:l 191MlN1 4 Np-237 3 2.22: 1 191MlN1 14 Np-237 3 1.76:1 191MlN4 2 Np-237 3 2.36:l 191MlN4 4 Np-237 3 2.38:l 191MlN4 14 Np-237 3 1.76:1 22SN1 1 U-233 3 1.95:l 22SA I1 3 U-233 3 1.78:1 22SAl1 14#1 U-233 1 7.38:l 22SA 11 14#2 U-233 1 7.34: 1 22SAl4 1 U-233 3 1.92:1 22SAl4 3 U-233 3 1.93:l 22SAl4 14#1 U-233 1 7.47:l 22SAl4 14#2 U-233 1 7.39:l 22SAl1 1 Np-237 3 2.00:l 22SA I1 3 Np-237 3 1.47:1 22SAl1 14 Np-237 3 1.73:l 22SAl4 1 Np-237 3 1.96:1 22SAl4 3 Np-237 3 2.00: 1 . 22SAl4 14 Np-237 3 1.31:1 1OSA 14#1 U-233 1 7.35~1 1OSA 14#2 U-233 1 7.40: 1 1- to 14-day sorption period (with sorption times being varied) foilowed by several months of desorption in the flow-through experimental apparatus. The experiments have consistently indicated that both uranium and neptunium initially desorb from the alluvium rather quickly, but their desorption rates eventually slow down and the ratio of sorption to desorption rates (i.e., effective Kd value) for the last 10-50% of the sorbed radionuclide is as much as two orders of magnitude higher than the initial ratio. Figure 2 shows the long-term desorption of uranium from the I9IMlA alluvium sample. For the Zone 1 groundwater (left plot) the long-term desorption results after one and three days of sorption show little difference; it is only after 14 days of sorption that there is a significant change in the amount of desorption. The Zone 4 groundwater (right plot) shows noticeable differences in the amount of desorption after one, three and fourteen days of desorption. Figure 3 shows thc longterm desorption of neptunium from the 191M1A alluvium sample. The long-term desorption results indicate that for both the Zone 1 groundwater (left plot) and the Zonc 4 groundwater (right plot) there is a significant difference in the amount of desorption after two, four and fourteen days of sorption. Figure 4 shows fits of uranium desorption data with a multiple sitelmultiple reaction rate model that is consistent with the generalized sorption residence time distribution modeling approach mentioned above. Thc model fit shown in Figure 4 yielded desorption ratc constant estimates that varied over nearly five orders of magnitude for the experiments with the 19D Zonc 1 (lower carbonate) groundwater (upper plot) and over 2 orders of magnitude for the experiments with the 19D Zone 4 (higher carbonate) groundwater (lower plot), with the rates decreasing over time. The jumps in the data and model curves in Figure 4 correspond to flow ratc changes in the desorption apparatus, including flow interruptions (both planned and unplanned) as indicated on the plots. The model fit for the experiment with 19D Zonc 1 groundwater (red line, upper plot) was obtained using a four-site desorption model with rate constants ranging from .07 g/hr (early) to .000001 d h r (late). Approximately 13% of the uranium initially in contact with the solid remained sorbed to the alluvium at the end of the long-term desorption. The model fit for the experiment with 19D Zone 4 groundwater (red line, lower plot) was obtained using a three-site desorption model with rate constants ranging from .04 g/hr (early) to .0001 g/hr (late). Approximately 6% of the uranium initially in contact with the solid remained sorbed to the alluvium at the end of the long-term desorption. The 191M 1A and 22SA paired column runs using the 19D Zone 1 water consistently result in a higher percentage of initial uranium sorption and a lower percentage of uranium desorbed as a function of time than the experiments utilizing the 19D Zone 4 water in all of the long-term desorption experiments conducted to date (Table Ill). These results are most likely due to the formation of negatively-charged uranyl carbonate and ncptunyl carbonate complexes in solution, which should tcnd to decrease the tendency for the uranium and neptunium to sorb to the alluvium surfaces. In addition, the 22SA alluvium sample contains a higher weight percent of smectite (Table I) and results in greater partitioning of the radionuclide to the solid phase in the initial sorption, although not always the case after long term cicsorption (Table 111). TABLE 111. Results of uranium and neptunium long-term desorption column experiments. The 191M1A and 22SA paired column runs using the 19D Zone 1 water consistently result in a higher percentage of initial uranium sorption and less desorption than the experiments utilizing the 19D Zone 4 water in all of the long-term desorption experiments conducteci to date. Experimental parameters are given in Table 11. Desorption Kd % remaining Alluvium/ RadioSorption Sorption & 19D nuclide Period (mwa Period (mllg) at end of sorbedc Ground (days) (hours) long-term Water Zone desorptionb 19IMIN1 U-233 1 3.39 1209 1077 13.94 19IMlA/1 U-233 3 5.20 1257 617 12.40 19IMlA/I U-233 14#1 5.21 3542 2896 12.74 19IMIA/1 U-233 14#2 5.21 3542 303 1 13.04 19IMlN4 U-233 1 1.06 1209 773 6.93 19IM 1A/4 U-233 3 2.33 1 0.76 1257 362 19IM 1A/4 U-233 14#1 1.34 2806 983 6.00 19IM 1A/4 U-233 14#2 1.34 2806 400 2.73 19IMlNI Np-237 2 5.32 1149 603 14.96 19IMINI Np-237 4 9.24 1197 926 33.50 19IMIN1 Np-237 14 10.12 1507 2370 47.80 19IM 1A/4 Np-237 2 3.41 1149 666 1 1.43 19IM 1A/4 Np-237 4 4.05 1197 793 17.36 19IMlA/4 Np-237 14 5.71 1507 1000 36.30 22SN1 U-233 1 27.04 1364 205 10.96 22SA/1 U-233 3 9.57 1412 2 17 12.73 22SA/1 U-233 14#1 9.92 679 197 19.95 22SN1 U-233 14#2 9.92 679 204 20.93 22SN4 U-233 1 1.97 5.15 1364 20 1 22SN4 U-233 3 1.87 1412 155 5.74 22SN4 U-233 14#1 3.13 2800 948 2.56 22SAI4 U-233 14#2 3.13 2800 140 0.9 1 22SN1 Np-237 1 16.8 30.14 1276 61 1 22SN1 Np-237 3 16.65 1324 659 30.96 22SA/1 Np-237 14 23.31 1340 332 4 1 .OO 22SA/4 Np-237 1 14.26 1276 585 27.00 22SN4 Np-237 3 15.27 1324 409 32.80 22SN4 Np-237 14 18.69 1340 576 43.10 1OSA U-233 14#1 5.14 664 158 20.44 1OSA U-233 14#2 5.14 1 1.66 664 128 "Reference is made to Table I1 for radionuclide tracer concentrations and liquid to solid ratios which effect sorption Kd values. Kd values are ratios of activity per gram on the solid to activity per milliliter in solution at the end of the long-term desorption. 'Percent of radionuclide initially sorbed to solid phase remaining sorbed at the end of the long-term desorption experiment. 111. DISCUSSION AND CONCLUSIONS The results shown in Figures 2, 3 and 4 indicate that there is a very wide range of desorption rates from the alluvium, undoubtedly the result of many different reaction sites in the heterogeneous material. The slower rates (i.e., stronger sorption sites) are not normally apparent from batch sorption experimental results or from short-duration batch desorption experiments. The radionuclide & distributions used for YMP saturated zone transport modeling [I] are almost cxclusivcly derived from such experiments. By not investigating these slow rates, previous YMP experiments havc yielded radionuclide Kd values that our experiments indicate may underestimate sorptionlretardation in the alluvium. Specifically, the alluvium & values used for U and Np in Yucca Mountain models are at least 1-2 orders of magnitude lower than what our experiments would indicate over long time and distance scales. The same probably holds true for other radionuclides as wcll. I 1 I 8 4 1 1 - r - . - . L : - " . - pthi* elpmamd m e m of tbe UaIMpH .~*a&%akoflQOOmllgwlllhB .r : ~ p r m o f * - e r s l l a p o r t ~ ~ m r w ~ ~ r c n d ~ & i f t b ~ o r # l r ~ m . ~ r # t k ~ t I , ~0~ 1. Bsc CeSFhpel ' ! - sluc ampany) 200Q. SIlaLSerrk to umkki~n%wpa ---=,a,--. - - -3. U m 1 3 4 1 . . a . -237 ia jl ~ f i r r m ~ 19IMIA. G ~ 10% W Im d 2 z s A U n d s r A m b i ~subiud ha!wt112W)3. 41341.OW. l h n b S n t p h Q A h v b n fmm NGgWW W d b 10Ikfl& l Q S k :-.7 l u p d 2 2 8 h U & h m b i s s t ~ . submm ?' lhw42111lm3. " 5 . EElMlJS* P., SCISM.C, AND DMO, M. 2004. ~qfw~-rdPIfN~ dW&JfnW--U-UR-0s- r 7797,prmd&a-~of* ~ ~ ~ ~ C Q , 6. I3. W Q M JOB C. MAR, D A m 1 JmIEKY* P. w, P b t d m ~ 2 mm ' T r n ~ u r d & ~ , i a r b .-' l k h t a v ,32,3893 (IW. SElmEs d ~