JEQ Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 5 July 2005
Published in J Environ Qual 34:1404-1414 (2005)
DOI: 10.2134/jeq2004.0395
© 2005 American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Um, W.
Right arrow Articles by Owen, A. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Um, W.
Right arrow Articles by Owen, A. T.
Agricola
Right arrow Articles by Um, W.
Right arrow Articles by Owen, A. T.
Related Collections
Right arrow Sorption/Exchange
Right arrow Radioactive Waste
Right arrow Other Environmental Contamination
Right arrow Water Pollution
Right arrow Soil Chemistry

TECHNICAL REPORTS

Waste Management

Enhanced Radionuclide Immobilization and Flow Path Modifications by Dissolution and Secondary Precipitates

Wooyong Um*, R. Jeffrey Serne, Steven B. Yabusaki and Antoinette T. Owen

Pacific Northwest National Laboratory, P.O. Box 999, MS P7-22, Richland, WA 99354

* Corresponding author (wooyong.um{at}pnl.gov)

Received for publication October 25, 2004.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Caustic radioactive wastes that have leaked at Hanford Site (Richland, WA) induce mineral dissolution and subsequent secondary precipitation that influence the fate and transport of contaminants present in the waste solutions. The effects of secondary mineral precipitates, formed after contacting solids with simulated caustic wastes, on the flow path changes and radionuclide immobilization were investigated by reacting quartz, a mixture of quartz and biotite, and a Hanford sediment (Warden soil: coarse-silty, mixed, superactive, mesic Xeric Haplocambids) with simulated caustic tank waste solution. Continuous Si dissolution and concomitant secondary mineral precipitation were the principal reactions observed in both batch and flow-through tests. Nitrate–cancrinite was the dominant secondary precipitate on mineral surfaces after 3- to 10-d reaction times in batch experiments. X-ray microtomography images of a reacted quartz column revealed that secondary precipitates cemented quartz grains together and modified pore geometry in the center of the column. Along the circumference of the packed column, however, quartz dissolution continuously occurred, suggesting that wastes that leaked from buried tanks in the past likely did not migrate vertically as modeled in risk assessments but rather the pathways likely changed to be dominantly horizontal on precipitation of secondary precipitate phases in the Hanford vadose zone. Based on batch equilibrium sorption results on the reacted sediments, the dominant secondary precipitates (cancrinites) on the mineral surfaces enhanced the sorption capacity of typical Hanford sediment for radionuclides 129I(–I), 79Se(VI), 99Tc(VII), and 90Sr(II), all of which are of major concern at the Hanford Site.

Abbreviations: FESEM, field emission scanning electron microscopy • FTIR, Fourier transform infrared • XMT, X-ray microtomography • XRD, X-ray diffraction


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
DURING THE Cold War, large volumes of high-level nuclear waste (HLW) were generated from plutonium production and separation processes at the Hanford Site, Washington; this HLW has been stored in 149 single-shell and 28 double-shell underground storage tanks (Mann et al., 2001; Johnston et al., 2002). Because of varying chemical processes (e.g., bismuth phosphate, PUREX, and REDOX) used in Pu extraction from irradiated spent nuclear fuel rods and the decay of short-half-life radionuclides, various forms of waste including sludge, salt cake, and liquid exist in the tanks under conditions of high pH (8 to 14), high ionic strength (e.g., 2–16 M NaNO3), high temperature (60–110°C), and high concentrations of dissolved aluminum (Kaplan, 2000; McKinley et al., 2001; Zachara et al., 2002; Liu et al., 2003). The tank wastes are to be retrieved, separated into two waste streams (HLW and low-level [LLW]), and vitrified. The LLW glass will be disposed in a shallow burial ground at Hanford, while the HLW glass will be sent to the deep geologic repository currently planned at Yucca Mountain, Nevada (Mann et al., 2001).

Leaks from perhaps as many as 67 single-shell tanks at the Hanford Site have been detected in the vadose zone sediments since the 1950s. Additionally, low-level wastes were dumped directly into the ground in the early years; therefore, sediments in the vadose zone are contaminated by radionuclides including 129I, 99Tc, 79Se, 237Np, 90Sr, 137Cs, and 233,235,238U (Kaplan et al., 1998). Because these radionuclides may travel through vadose zone sediments and reach ground water resulting in potential health risks to human beings, predicting the migration and sequestration of radionuclides at the Hanford Site is of critical importance and an understanding of the role of the caustic leachates on the sediments and transport processes is essential to this predictive capability.

Previous laboratory experiments showed simulated caustic tank waste leachates dissolved primary minerals resulting in the release of Si to solution and the secondary precipitation of cancrinite, a zeolite-like mineral at 89°C (Bickmore et al., 2001; Chorover et al., 2003) and cancrinite and zeolite phases at room temperature (Chorover et al., 2003). Nitrate–cancrinite nucleated on and cemented together quartz grains, the primary mineral of the Hanford sediment (Bickmore et al., 2001), and Cs and Sr were taken up within the newly precipitated cage structure (Chorover et al., 2003). There are, however, few studies on the transport behavior and sequestration of radionuclides in sediments after reaction with tank simulants and formation of secondary phases. Zones of altered primary minerals and neo-formed secondary precipitates are likely present beneath the leaked tanks. The migration of radionuclides, driven by natural recharge, past water pipeline leaks, and future tank sluicing activities, from and through these altered zones containing cemented primary grains and new sorption sites from the secondary precipitates, requires more investigation. In this study, batch and column dissolution experiments, X-ray microtomography studies, and radionuclide sorption experiments on the reacted sorbents were conducted to increase our understanding of the effects of these secondary precipitates. In particular we note the flow-path modifications caused by primary mineral dissolution and secondary mineral precipitation and the subsequent sequestration of radionuclides onto the secondary precipitates as two processes that need more recognition in future risk assessments at the Hanford Site.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Materials
Quartz sand (50–70 mesh; Aldrich, Milwaukee, WI) was pretreated with acid (100 g of sand in 100 mL of concentrated H2SO4 for 2 d), rinsed several times with NANOpure water, and oven-dried at 85°C for 48 h. Powdered biotite was prepared from samples obtained from Ward's Scientific (Rochester, NY). The biotite was cleaved, hand-picked to exclude calcium carbonate inclusions and iron oxide coatings, crushed, sieved to obtain the 53- to 105-µm size fraction, and then cleaned of ultra-fine particles by repeated gravity settling in acetone (Optima grade; Fisher Scientific, Hampton, NH). The cleaned biotite particle size, determined by petrographic microscopy, was 28% <50 µm, 63% 50–110 µm, and 9% >110 µm. Powder X-ray diffraction (XRD) analysis of the quartz and mixed quartz–biotite revealed no other mineral phases.

A mixed sample was prepared with quartz (85% by weight) and biotite (15%) as a model for Hanford sediment. The <250-µm size fraction of Warden silt loam (Warden soil) without pretreatment was used to represent the actual Hanford sediment, which is dominated by quartz and feldspar with minor amounts of calcite and mica (Serne et al., 2001). Based on semiquantitative XRD measurements, the clay-sized fraction (<2 µm) of the Warden silt loam consists of smectite (40 wt. %), illite (25 wt. %), chlorite (20 wt. %), and kaolinite (<5 wt. %) with minor amounts of quartz, feldspar, and amphibole (10 wt. %). The Brunauer–Emmett–Teller (BET) specific surface areas of quartz, biotite, the mixed quartz–biotite sample, and Warden soil (0.038, 2.96, 0.289, and 9.94 m2 g–1, respectively), were determined with an ASAP 2010 surface area analyzer (Micromeritics, Norcross, GA) using krypton for quartz and nitrogen for all others. Simulated tank leachate was prepared using reagent grade chemicals and consisted of 2 M NaNO3, 1.04 M NaOH, and 0.01 M Al(NO3)3. This simulant composition was chosen to represent the lower end of the range for bulk tank compositions, which include the compositions of both solids and liquids (Agnew et al., 1996).

Batch Dissolution Experiments
Eight batch dissolution experiments were conducted, two replicates for each of four conditions: simulated tank leachate with (i) quartz, (ii) mixed quartz–biotite, (iii) Warden soil, and (iv) quartz with a tank leachate simulant without dissolved Al. In each experiment, 3 g of solid was mixed with caustic tank simulant (60 mL) in a 60-mL polypropylene bottle. A series of bottles was prepared for each study to provide separate reacted slurries after 12 h and 1, 3, 7, 14, 21, and 30 d of reaction. The bottles were placed in a 90°C oven and shaken gently twice a day to ensure complete mixing. At each collection time, replicate bottles from each test were removed and placed in a water bath maintained at approximately 80°C. The pH was measured at this high temperature using an Accumet pH meter (Fisher Scientific) with a solid state electrode and the measured values were corrected to 90°C using the temperature dependence of the dissociation constant of water (Bickmore et al., 2001). The electrode was calibrated using commercial pH 7.0, 10.0, and 12.0 standards (Fisher Scientific) at room temperature. After measuring pH, the supernatant was filtered through a 0.45-µm PTFE syringe filter, diluted by a factor of two with NANOpure water, and acidified to pH of approximately 2 with concentrated HNO3 (Fisher Scientific). Dissolved Si and Al concentrations were determined using inductively coupled plasma optical emission spectrometry (PerkinElmer, Wellesley, MA).

Solids were removed from the remaining solution by filtration, washed several times with ethanol (3 g of solid in 300 mL of ethanol) to remove any precipitated sodium salts, and then dried at room temperature. The secondary precipitates were identified by XRD on a Philips PW3040/00 X'pert MPD system (PANalytical, Natick, MA) with Cu K{alpha} radiation. Fourier transform infrared (FTIR) spectra were obtained using a Bruker FTIR spectrometer (IFS66) (Bruker Optics, Billerica, MA) in transmission mode between 400 and 4000 cm–1. Characterizations of morphology and chemical composition were made by field emission scanning electron microscopy (FESEM) and energy dispersive X-ray spectroscopy (EDS) using a LEO982 FESEM system (JEOL, Peabody, MA) with an Oxford Instruments (Scotts Valley, CA) ISIS energy dispersive X-ray detector.

Column Experiments
Three column experiments were performed: one each with quartz, the mixed quartz–biotite, and Warden soil. The columns, made of poly-ether-ether-ketone (PEEK), were 1.91 cm in diameter and 7.62 cm in height, yielding a total volume of 21.7 cm3. Column ends were covered by a small piece of 5-µm Spectra/Mesh to retain the particles. Each column was packed using a continuous particle stream while vibrating the column on a Vortex-Genie 2 mixer (Fisher Scientific) to achieve uniform packing. The columns were saturated with uncontaminated Hanford ground water from well 699-S3-25 (I = 0.0094 M and pH = 7.5) that was allowed to flow through each column for 2 d at room temperature to stabilize the flow and to remove any easily dispersible particles. A constant upward flow rate was maintained using a syringe pump to minimize depositional effects due to gravity and potential preferential flow along the column walls. Bromide-traced ground water was used as a conservative tracer to determine the pore water velocity. After obtaining a bromide breakthrough curve, the bromide solution was immediately replaced by the caustic tank simulant. Each column was then placed inside a 90°C oven while the pump and influent solution were outside the oven. Influent solution was delivered through Tygon tubing threaded through a hole at the top of the oven. Effluent was collected in a centrifuge tube inside the oven; the top of the tube was sealed with parafilm to decrease possible evaporation during sample collection. Concentrations of Si and Al were corrected using an average evaporation factor obtained from a separate experiment measuring volume decrease in parafilm sealed tubes as a function of time inside the 90°C oven. The pH measurements and chemical analyses of the effluent solutions were performed using the same procedures described above.

X-Ray Microtomography (XMT)
Scanned images of the packed columns before and after contact with the caustic solution were generated using the XMT system (ACTIS 200/160 KXR; Bioimaging Research, Chicago, IL). The X-ray source used a tungsten target and X-rays were emitted through a 0.25-mm-thick beryllium window. The RLS-2048/100 CCD detector (Invitrogen, Carlsbad, CA) is a cesium iodide scintillation phosphor screen optically coupled to a linear photodiode array through a fiber optic taper. The diode array is an EG&G Reticon (Sunnyvale, CA) RL2048S, monolithic self-scanning linear photodiode array with 2048 photodiode sensor elements having 25-µm center-to-center spacing. Each column was scanned using a spot size of 65 µm at 160 keV and a current of 0.3 mA. The images are displayed using 5122 pixels, each at 12 bits with a spatial resolution of 50 µm. The computed tomography (CT) scans per column were generated at the bottom, middle, and top of the column (20, 40, and 60 mm from the inlet end of the column) and the reconstructed images of a two-dimensional cross-sectional slice were obtained using Clemex Vision image analysis software (Clemex Technologies, Longueuil, QC, Canada). Manual gray threshold with calibration of each phase to the pixel number ranges was programmed and used for each slice to separate each phase (air, water, or solid) and assign three bitplane colors.

Sorption Experiments
Batch sorption experiments using both unreacted and reacted sorbents in contact with radionuclide-spiked ground water were performed. The radiotracers chosen were those of major concern at the Hanford Site: 99Tc(VII), 129I(–I), 79Se(VI), and 90Sr(II). Quartz and Warden soil were reacted with caustic tank simulant at 90°C for 14 d, after which the solids were washed with ethanol, air-dried, and used as reacted sorbents. Characterization of these samples revealed the presence of secondary precipitates on the mineral surfaces. Pure single crystals of nitrate–cancrinite synthesized at high temperature (Buhl et al., 2000) were also used as sorbents in batch sorption experiments.

The sorption experiments were conducted in 15-mL polypropylene centrifuge test tubes by mixing sorbents and Hanford ground water spiked with 3.7 x 105 Bq L–1 of each individual radionuclide and placing the tubes on a slow-moving platform shaker for 7 d. Solids concentrations of 100 and 5 g L–1 were used for anionic (99Tc, 125I, 75Se) and cationic (90Sr) radionuclides, respectively. Nitric acid (HNO3) and sodium hydroxide (NaOH) were used to adjust the pH of the radiotraced ground water. 125I(–I) and 75Se(VI) were used as analogs for 129I(–I) and 79Se(VI), which are not commercially available and also more difficult to monitor using radio-counting. The higher solid-to-solution concentration for anion radionuclides was used to compensate for the low sorption affinity of anions at high pH. A blank sample of radionuclide-spiked ground water containing no solid was also prepared in an identical manner to determine the initial activity of each radionuclide and to quantify any radionuclide sorption to the test tubes. After 7 d of contact, the suspensions were centrifuged and the supernatants were filtered through 0.45-µm filters. Sorption of radionuclides on the test tubes and filters was negligible. The final activities in the filtered effluent solutions were determined using Wallac 1415 liquid scintillation counting (LSC; PerkinElmer) for 125I(–I), 99Tc(VII), and 90Sr(II) and a Wallac 1480 Wizard 3-inch NaI automatic gamma detector (PerkinElmer) for 75Se(VI). All radiological measurements were performed long enough to obtain less than a 3% combined error with systematic error accounting for <1% of the total error. Partition distribution coefficients, Kd (mL g–1), for each radionuclide on both reacted (caustic-solution treated) and unreacted sorbents were calculated using the difference between initial and final activities in solution to represent the activity sorbed to the solids.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Batch Dissolution and Precipitates
When the caustic simulated tank waste solution, both with and without dissolved Al, contacted the different solid samples at high temperature, the dissolved silica concentration increased due to the dissolution of Si-bearing minerals. Except for the experiment reacting pure quartz with Al-free solution, dissolved silica increased until the solutions reached supersaturation with respect to an aluminosilicate phase and precipitation occurred. The initial pH of the solutions was approximately 13.0, while the pH values of the batch dissolution supernatants ranged from 12.0 to 13.0. For the tests with quartz and mixed quartz–biotite (Fig. 1a) , the dissolved Al concentration was constant at the initial value of 0.01 M for up to one day. After one day of reaction, the dissolved Al concentration decreased, dropping to near zero (detection limit) after 14 d. The decreased concentration of dissolved Al could also be caused by Al sorption onto the solids present. However, because the Al-rich deposits on remaining primary minerals completely disappeared after 14 d (see Characteristics of Secondary Precipitates, below), secondary precipitates were considered to consume most of the dissolved Al present in the beginning solution. In contrast, dissolved Si concentration increased with time. The Si release rate (the slope of Si concentration with time) for the quartz experiment with added Al increased during the first 3 d, decreased slightly from Day 3 to 7, and then increased again after 7 d of contact. The period of reduced Si dissolution rate during the 3- to 7-d time period was considered to be the main period of secondary mineral precipitation. The pattern of Si dissolution and cancrinite precipitation was similar to that observed by Bickmore et al. (Bickmore et al., 2001).



View larger version (27K):
[in this window]
[in a new window]
  		Fig. 1. Dissolved Si and Al concentrations with time. (a) Quartz samples with and without 0.01 M Al, and the mixed sample (qz and biotite); (b) Warden soil sample with 0.01 M Al.

 
Dissolved Si from quartz in the caustic solution without added Al was higher in concentration than in the experiment with added Al. Although the presence of alkali metal cations such as Na increases the dissolution rate of quartz (Berger et al., 1994; Dove and Nix, 1997; Dove, 1999), the inhibition of dissolution by Al species has also been widely reported in quartz and other aluminosilicate mineral dissolution experiments (Van Bennekom et al., 1991; Gautier et al., 1994; Oelkers and Schott, 1994; Oelkers and Gislason, 2001).

Caustic solution attack on the mixed sample (85% quartz and 15% biotite) exhibited a shorter reaction time for secondary precipitation and led to generally higher dissolved Si concentrations than pure quartz samples. Higher total surface area in the mixed sample due to the presence of biotite (0.289 vs. 0.038 m2 g–1 for pure quartz) likely explains the higher release of Si. High dissolved Si concentrations were also evident in the Warden soil effluent (Fig. 1b), probably due to the presence of clay minerals and basalt fragments having higher total surface area and dissolution rate, respectively, compared with quartz. In addition, because of the greater mineralogical complexity of the Warden soil, the solution composition also showed complicated time history for the release of Fe, Mg, and K (data not shown), and several plateaus were present in the dissolved Si data indicating the precipitation of more than one mineral species in this system.

Characteristics of Secondary Precipitates
X-ray diffraction patterns of two samples (quartz and mixed quartz–biotite) after 14 d of reaction reveal a secondary precipitate identified as cancrinite (Fig. 2a) . Cancrinite is a feldspathoid and has a hexagonal crystal system with space group P63 (Barnes et al., 1999a, 1999b; Buhl et al., 2000); its framework structure is similar to zeolites (Buhl et al., 2000). Another sodium aluminosilicate phase that may have formed is sodalite, which has a cubic crystal structure with space group P43n (Barnes et al., 1999b; Chorover et al., 2003). Because cancrinite and sodalite consist of AB and ABC sequences of 6- and 12-membered ring planes, respectively, the structural frameworks are similar. Consequently, they have many common XRD peaks and only a few differences at d = 4.69 Å (18.9 degrees 2{theta}) and a 211 reflection at d = 3.24 Å (27.5 degrees 2{theta}) (Barnes et al., 1999b; Buhl et al., 2000; Chorover et al., 2003).



View larger version (46K):
[in this window]
[in a new window]
  		Fig. 2. (a) X-ray diffraction (XRD) patterns of quartz and the mixed sample after 14 d of reaction (Q, quartz; B, biotite; C, cancrinite; S, sodalite); (b) XRD patterns of reacted quartz as a function of time. Arrows indicate typical peaks of cancrinite. Single mineral nitrate cancrinite was separately synthesized according to Buhl et al. (2000).

 
Because both quartz and mixed quartz–biotite samples reacted with caustic tank simulant for 14 d showed typical cancrinite XRD peaks at d = 18.9 degrees 2{theta} and 27.5 degrees 2{theta}, cancrinite was considered to be the dominant species in the secondary precipitates. The time series of XRD patterns for a quartz sample also showed cancrinite peaks (Fig. 2b). Peaks identifying the presence of cancrinites appeared after 3 d of reaction and the most intense peak was found after 7 d of reaction, consistent with previous batch dissolution results where the main precipitation phase occurred between 3 and 10 d. Cancrinite peaks for reacted Warden soil were not distinguishable (patterns not shown), because of multiple peaks from the various primary and secondary minerals originally present in Warden soil.

Because of the high concentration of NO–3 in the tank simulant, nitrate–, rather than carbonate–, cancrinite was considered to dominate and FTIR spectra supported this hypothesis. The FTIR spectra of the 30-d reacted quartz sample (Fig. 3) showed a band split for nitrate absorption at 1380 and 1422 cm–1 (Buhl and Lons, 1996). The FTIR CO2–3 absorbance peaks at 1450 to 1470 cm–1 normally found in carbonate cancrinite and sodalite were not detected (Barnes et al., 1999c). Although a high concentration of NaOH was used, because a typical OH-stretching band for NaOH was also not found around 3640 cm–1 (Zhao et al., 2004), nitrate cancrinite was considered to be the dominant secondary precipitate formed in the experiments.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 3. Fourier transform infrared (FTIR) spectra of 30-d reacted quartz.

 
The FESEM images revealed ball-shaped aggregates of cancrinite with radiating needle-shaped hexagonal prisms on the quartz surfaces and on aluminum-rich deposits (or aluminosilicate gels) on quartz surfaces after 3 d of reaction (Fig. 4a) similar to previous work (Bickmore et al., 2001). The aggregates coalesced after 14 d (Fig. 4b). Elemental mapping and energy dispersive X-ray spectroscopy (EDS) analysis of precipitates on quartz surfaces after 7 d showed a dominant composition of Na, Al, and Si (data not shown). Cancrinite precipitates formed on thin sheets or flakes of biotite from the mixed sample are also visible in Fig. 4c. Column-packed Warden soil after 22 d of reaction showed ball-of-string shaped cancrinites and euhedral analcime on the mineral surfaces (Fig. 4d). The continuous supply of Al in the feed solution in the column experiment allowed the cancrinite crystals to develop more fully compared with the batch experiments where the Al supply was finite. The FESEM images suggest qualitatively that leaked caustic tank wastes have likely modified the mineral properties of sediments immediately underneath the tanks at the Hanford Site.



View larger version (165K):
[in this window]
[in a new window]
  		Fig. 4. Field emission scanning electron microscopy (FESEM) images of secondary precipitates: (a) after 3 d of reaction on quartz surfaces; (b) after 14 d of reaction on quartz surfaces; (c) after 14 d of reaction on mixed sample; (d) cancrinites and analcimes on Warden soil after 22 d of reaction in the middle of the column.

 
Column Experiments and X-Ray Microtomography (XMT)
Dissolved concentrations of Si and Al released from three columns packed with quartz, mixed quartz–biotite, and Warden soil are given in Fig. 5 . The relatively high concentrations of Si released in the column experiments compared with those of the batch experiments were attributed to the high solid-to-solution ratio in the columns. The quartz column showed higher Si release compared with the mixed quartz–biotite and Warden soil columns, especially between 3 and 15 d (Fig. 5a). The mixed quartz–biotite column showed high Si releases in both early (<3 d) and late (>17 d) stages similar to those of the quartz column. Low Si release in the intermediate stage (3–15 d) in the mixed column was attributed to the precipitation of a secondary mineral in addition to cancrinite. Although other secondary precipitates were not detected in the mixed column, more Si was consumed by the formation of secondary precipitates during the main precipitation phase than in the quartz column, likely because elements such as K, Mg, and Fe released from the dissolving biotite were available for the formation of other secondary phases. This hypothesis is supported by the similar pattern of Si release in the Warden soil column and the evidence of secondary precipitates other than cancrinite such as analcime observed by FESEM (Fig. 4d).



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 5. Changes of (a) Si concentration and (b) Al concentration versus time in flow-through column experiments.

 
The concentrations of Al in the effluents from the three different columns provided results consistent with our interpretations of the causes for the changes in the rate of Si release. Aluminum concentrations in the eluted solutions did not reach the initial Al concentration (0.01 M) in any of the three columns, suggesting that most of the Al in the caustic input solution was consumed to form secondary precipitates (Fig. 5b). Aluminum consumption was fastest in the Warden soil column, indicating of the most extensive precipitation reactions occurred in this column relative to the other two columns. Because continuous dissolution occurred in the packed columns, the dissolution resulted in increasing concentrations of dissolved Si, K, Mg, Fe, and other elements in the pore solutions that scavenged most of the Al from the input solution to precipitate secondary minerals. Initial caustic solution pH was about 13.0 and quartz column effluent after 0.7 d of reaction was pH 8.5. The pH increased with reaction time and was relatively constant at about 12.1 after 5.8 d of reaction. The mixed and Warden soil columns showed similar pH changes versus time during experiments.

Since dissolution and secondary precipitates were expected to have modified pore structures and flow paths in the columns, the XMT was used to investigate these potential changes. Computed tomography (CT) scans of the middle of the quartz column (40 mm from the inlet end) taken at different reaction times are shown in Fig. 6 . Because denser materials are represented as lighter colors, the individual quartz grains are white or very light gray, while the surrounding water is gray and the air-filled void spaces are black. Due to resolution limitations in our XMT system, the XMT results can be used only for qualitative interpretations on the effects of mineral dissolution and secondary precipitation on changes in pore structure and fluid flow paths. Air-filled (dry) and water-saturated pores were evenly distributed in the quartz column before contact with the caustic solution (Fig. 6a and 6b). After 8 d of reaction, the population of white spots (green in the digitized image) at the center of the column (square field with 212.1-mm2 area) (Fig. 6c) increased slightly compared with the initial water-saturated column before caustic solution injection, suggesting an increased number of newly formed secondary precipitates around quartz sand grains, especially in the center of the column. Several air-filled pores (black or blue spots) also started to appear both around the center and along the circumference of the packed column (Fig. 6c). As reaction time increased to 22 d, the relative area of the solid phase (white or green spots) in the limited square field area around the center of the column increased some more (Fig. 6d). This is attributed to the increased formation of secondary precipitates cementing quartz sand grains in the center of the column. An increased population of air-filled pores (black or blue spots) was also found around the edges of the column. However, because air-filled pores along the edges (outside circumference of the column) after 8 d reaction were few, we suggest that a relatively evenly distributed flow path and continuous dissolution–precipitation was occurring throughout the column during the first 8 d of reaction. After 22 d of caustic solution reaction, it was more apparent that selected dissolution around the edges had occurred in the middle section of the column. We suggest that Si dissolved from the bottom of the column (inlet end) and it precipitated at the center of the middle of the column increasing cementing processes that led to changes in flow paths. The changes in flow path were from an evenly distributed flow throughout the whole cross-section of the inlet portion of column to a preferential flow mainly along the edges of the column farther from the inlet (such as the middle section shown in Fig. 6d). In addition, even though we did not directly measure the hydraulic conductivity, the flow rate dropped from 5.66 mL d–1 (t = 0 d) to 4.30 mL d–1 (t = 22 d) likely from back-pressure in the column due to the increased cementing process.



View larger version (191K):
[in this window]
[in a new window]
  		Fig. 6. Computed tomography (CT) scans for quartz columns collected at the middle of the column (40 mm from the inlet). Digitized images (c and d) are shown with different colors (blue, red, and green for black, gray, and white spots, respectively) in a limited square field area (212.1 mm2). (a) Dry column; (b) initial water saturated column; (c) after 8 d of reaction; (d) after 22 d of reaction.

 
The XMT images collected at the bottom and top of the column (20 and 60 mm from the inlet, respectively) after 8 and 22 d of reaction are shown in Fig. 7 . Dissolution processes started in both the bottom and top of the column after 8 d of reaction. Noticeable preferential flow along the column outside edges was not found after 8 d of reaction. However, after 22 d of reaction, dissolution along the outside edges of the column was easily found at both the bottom and the top of the column. This suggested that continuous input of caustic solution started to form precipitates at the center of the bottom section (close to inlet) of the column resulting in the increased cementing processes at the center of the bottom column and the beginning of the preferential flow along the edges of the entire column.



View larger version (201K):
[in this window]
[in a new window]
  		Fig. 7. Computed tomography (CT) scans for quartz column collected from different locations (bottom and top) after 8 and 22 d of reaction. (a) Bottom (20 mm from the inlet) after 8 d of reaction; (b) top (60 mm from the inlet) after 8 d of reaction; (c) bottom (20 mm from the inlet) after 22 d of reaction; (d) top (60 mm from the inlet) after 22 d of reaction.

 
A larger population of the solid phase (white spots) was also found in the middle of the mixed and Warden columns (40 mm from the inlet) after 22 d of reaction compared with 8 d of reaction (Fig. 8) . However, because of the presence of finer grained particles in these columns, the resolution of XMT image was low. Even though quantitative image analysis was not conducted because of low resolution, the difference was visually observed and the interpretations were similar to those of the quartz column, suggesting that increased cementing processes between grains were dominant in the center cross-sections of all the columns and that the probability of preferential flow was low until 8 d of reaction (Fig. 8a and 8c). The diversion of the flow path of the caustic solution from the center to the edges of the column suggests a possible alteration of the fluid pathway for tank waste leachates underneath tanks that leaked in the past at the Hanford Site. Field observations around and below single-shell tanks that leaked significant quantities of caustic waste showed evidence of significant horizontal migration, which was somewhat unexpected (Knepp, 2002a, 2002b). The XMT results also suggest that cementing of grains may be a major process to increase the horizontal flow component (as opposed to gravity driven vertical flow) of highly caustic fluids intentionally disposed or inadvertently leaked into the vadose zone sediments.



View larger version (153K):
[in this window]
[in a new window]
  		Fig. 8. Computed tomography (CT) images for mixed (a and b) and Warden soil (c and d) columns collected at the middle of the column after 8 and 22 d of reaction, respectively.

 
Sorption Experiments
The presence of secondary minerals precipitated on the surfaces of mineral grains after contact with caustic waste fluids might have a significant role in the immobilization and ultimate fate of key radionuclides such as 129I(–I), 99Tc(VII), 79Se(VI), and 90Sr(II) in the vadose zone sediments at the Hanford Site. Newly formed precipitates tend to possess higher specific surface areas than the original untreated minerals and therefore may provide additional sorption sites for removal of key radionuclides. Because cancrinites have a feldspathoid structure similar to the "cage structure" in zeolites, cancrinites are expected to sequester quite effectively radioactive and other toxic metal contaminants (Chorover et al., 2003; Um and Papelis, 2004). Uptake of Cs, Sr, and Se in the structure of cancrinite showed nearly complete Sr sorption for low concentration (Sr ≤ 10–4 M) and increasing Cs uptake with reaction time during secondary mineral precipitation (Chorover et al., 2003), and incorporation of Se with homogeneous distribution in cancrinite (Lindner et al., 1996).

The effects of secondary mineral precipitates on radionuclide mobility were investigated by sorption affinity of each radionuclide onto quartz and Warden soil, which were previously reacted with caustic solution for 14 d. Distribution coefficients (Kd) for key radionuclides including 125I(–I), 75Se(VI), 99Tc(VII), and 90Sr(II) on various sorbents are shown in Table 1. Because reacted sediments have secondary precipitates and a minor amount of NaOH salt still remained at the surfaces, higher pH (approximately 10.0) was found in the reacted sediment than in unreacted sediment without any pH adjustment (7.4–8.5). The pH was adjusted to approximately 10 for sorption on the unreacted sediments to allow comparison with the Kd values for the reacted sediments at the same pH condition. Reacted Warden soil and quartz showed higher sorption selectivities for radionuclides than unreacted Warden soil and quartz, and the increased sorption capacity was attributed to uptake by the secondary mineral precipitates. Anionic radionuclides (125I as iodide, 75Se as selenate, and 99Tc as pertechnetate) normally showed low sorption affinities on quartz and Warden soil at high pH conditions resulting in high mobility in the subsurface environment at the Hanford Site. However, caustic-reacted sediment samples showed significantly enhanced sorption capabilities for these anionic radionuclides, even at high pH. Selenate showed one order of magnitude higher sorption affinity onto reacted Warden soil compared with untreated Warden soil. Anion exchange or substitution for nitrate or hydroxide in the cancrinite structure might be a dominant process for increased 75Se(VI) oxyanion sequestration while there was still a minor possibility of surface complexation reactions for the anions on the protonated cancrinite edges or surfaces. However, further spectroscopic investigation would be necessary to differentiate the specific sorption mechanisms for anionic radionuclides.


View this table:
[in this window]
[in a new window]
  		Table 1. Distribution coefficients (Kd) of key radionuclides on the reacted sorbents.

 
Cationic 90Sr(II) showed a tremendous increase in sorption affinity on reacted Warden soil when compared with unreacted Warden soil. The increased sorption uptake of strontium was attributed to the high cation exchange capacity of cancrinite-like zeolites (Um and Papelis, 2003). Synthesized nitrate–cancrinites showed much higher sorption capacities for all key radionuclides, suggesting that potential secondary precipitates such as cancrinite on natural sediment surfaces can sequester radionuclides and thus result in reduced mobility of these radionuclides in the Hanford vadose zone.


   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Results from batch and column experiments indicated that secondary mineral precipitation processes occurred after contact with simulated caustic Hanford tank leachates, which impacted both the porosity of packed solids and sorption tendencies of the reacted solids. Dissolved Si from the primary solids reacted with simulated caustic leachates and formed secondary precipitates on the mineral surfaces. Dominant secondary precipitates were identified as nitrate–cancrinite and the precipitation of secondary phases including cancrinite on mineral surfaces cemented individual grains enough to change the flow path of fluids through packed columns. In addition, the secondary mineral coatings increased the sorption capacities of quartz and Warden soil in comparison with untreated sorbents, especially at high pH. It is likely that both phenomena occurred in the sediments directly beneath Hanford single-shell tanks that leaked caustic fluids in the past. These findings provide an explanation for the observed pronounced lateral migration of contaminant plumes and the concentration of many contaminants in the sediments within 10 m from the tank bottoms and sides (Serne et al., 2002).


   ACKNOWLEDGMENTS
 
The authors appreciate the comments of three anonymous reviewers and the associate editor on the initial version of the manuscript. This work was supported by the U.S. Department of Energy (DOE) Environmental Management Science Program Grant DE-FG07-99ER-15009. Pacific Northwest National Laboratory (PNNL) is operated for the DOE by Battelle Memorial Institute under Contract DE-AC06-76RLO 1830. The authors wish to thank Dr. Kathy Nagy and Dr. Sherry Samson currently at University of Illinois at Chicago for collaboration on the test design, supplying the well-characterized biotite, and review efforts on early drafts of this manuscript. Parts of this work were conducted in the Environmental Molecular Sciences Laboratory, a national scientific user facility sponsored by the U.S. Department of Energy's Office of Biological and Environmental Research and located at Pacific Northwest National Laboratory (PNNL). The authors thank James S. Young, Mark H. Engelhard, Paul Gassman, Igor Kutnyakov, and Robert D. Orr for various sample analyses.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Vadose Zone JHome page
J. M. Zachara, J. Serne, M. Freshley, F. Mann, F. Anderson, M. Wood, T. Jones, and D. Myers
Geochemical Processes Controlling Migration of Tank Wastes in Hanford's Vadose Zone
Vadose Zone J., November 20, 2007; 6(4): 985 - 1003.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (5)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Um, W.
Right arrow Articles by Owen, A. T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Um, W.
Right arrow Articles by Owen, A. T.
Agricola
Right arrow Articles by Um, W.
Right arrow Articles by Owen, A. T.
Related Collections
Right arrow Sorption/Exchange
Right arrow Radioactive Waste
Right arrow Other Environmental Contamination
Right arrow Water Pollution
Right arrow Soil Chemistry

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Vadose Zone Journal Journal of Plant Registrations
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal