George Nibler
Morrison Knudsen Corporation

Morrison Knudsen Corporation (MK) is constructing a disposal cell at the DOE Weldon Spring Site Remedial Action Project (WSSRAP) in Missouri that will contain a variety of low-level radioactive waste, including 160,000 cubic yards of chemically stabilized and solidified (CSS) raffinate sludge. The leachate generated by the cement and fly ash stabilized sludge is alkaline (pH=12), with elevated concentrations of dissolved nitrate, molybdenum, arsenic, uranium, and radium. Although the leachate will be collected and treated, the disposal cell design criteria is based on maximizing the containment of contaminants within the disposal cell, thereby minimizing contaminant concentrations and subsequent treatment requirements of the leachate. To do so, the cell design includes a geochemical barrier overlying the leachate collection system in the area where CSS sludge will be placed.

A study was undertaken to optimize the geochemical barrier design to sequester the contaminants of concern in the disposal cell liner. The base component of the barrier is the locally available clay of the Ferrelview Formation, which is a stiff, plastic silty clay with a hydraulic conductivity of 10^-7 to 10^-9 cm/s. The unmodified clay has a significant cation exchange capacity (0.25 - 0.30 meq/gm), and its low hydraulic conductivity ensures that travel time of the leachate will be sufficiently long to ensure contaminant removal. Several amendments, including zeolite, wood chips, phosphate, and sphagnum peat moss (peat), were evaluated by measuring the sorption capacity for select contaminants by batch tests and miniature simulated disposal cells.

The zeolite and wood chips were ineffective in reducing the contaminants of greatest concern, (uranium, molybdenum, arsenic, and nitrate). These contaminants are present as anionic complexes in the CSS leachate and thus are not amenable to sorption by zeolites, which are generally specific sorbers for simple cationic species. The wood chips were ineffective because they acted essentially as pieces of inert material in the oxidizing conditions of these tests. Phosphate, at 5 wt% added, was effective at reducing uranium concentrations from about 67,000 pCi/L to 400 pCi/L, probably resulting from precipitation of a uranium phosphate phase such as autunite.

Peat was effective in the batch tests at reducing uranium concentrations from 67,000 pCi/L to 1000 pCi/L (or about 1.5 mg/L) at 20 weight percent (wt%) added. This reduction is attributed to sorption by the peat rather than the generation of a reducing environment, because the tests did not control the redox state.

The cell design should virtually eliminate oxygen infiltration and the oxidizable material (such as peat) in the geochemical barrier will consume any oxygen and other oxidants, an anaerobic or reducing environment will result. This condition will have a significant effect because the contaminants of concern (U, Mo, and As) are redox-sensitive and will precipitate under reducing conditions, thereby reducing their concentrations in the leachate. For example, uranium is soluble in the range of only a few mg/L under strongly reducing conditions.

Mass balance calculations show that 1 wt% peat should induce a reducing environment for the 1000 year design life of the disposal cell. Although the Ferrelview clay has 2 to 3 wt% organic carbon, the addition of a few percent peat should guarantee that reducing conditions are maintained through the life of the cell. Furthermore, geochemical modeling indicates that the solubility of most contaminants of concern will be reduced to lower levels by establishing a reducing environment and subsequent precipitation of reduced phases as opposed to simple sorption.

Based on the testing evaluation and modeling, the addition of a minimum of 2 weight % (approximately 20 volume %) sphagnum peat moss was selected as the preferred amendment. The peat will be placed as a 3-inch-thick lift sandwiched between lifts of Ferrelview clay, which will reduce material handling and mixing requirements. Although the peat-amended clay has significant sorption or cation exchange capacity, the design is based on the concept that the peat will guarantee that a reducing environment is maintained in the cell, thereby inducing precipitation of most contaminants of concern and minimizing future leachate treatment requirements.

A low-level radioactive waste disposal cell is currently under construction at the DOE Weldon Spring site in Missouri. The cell will contain a variety of waste, including 160,000 cubic yards of chemically stabilized and solidified (CSS) raffinate sludge. The leachate generated by the cement and fly ash stabilized sludge is alkaline (pH=12), with elevated concentrations of dissolved uranium, radium, molybdenum, arsenic, and nitrate. Although the leachate will be collected and treated, the disposal cell design criteria is based on maximizing the containment of contaminants within the disposal cell, thereby minimizing contaminant concentrations and subsequent treatment requirements of the leachate. It was therefore determined that the cell design should include a geochemical barrier component. The geochemical barrier should have the ability to sequester the contaminants present in leachate within the cell, thus preventing their migration into the leachate collection system.

The cell design under lying the CSS waste includes a three-foot-thick lower liner of the local Ferrelview clay, overlain by a multi-layer geotextile/geomembrane liner, a sand and gravel drain layer for the leachate collection system, and an upper 2.5 foot thick Ferrelview clay layer. It was initially thought that amending the upper one foot of the Ferrelview clay layer with a material to improve the clays sorption capacity would be the preferred option. Several potential amendments were considered, including zeolite, wood chips, phosphate, and sphagnum peat moss (peat). Each amendment was evaluated by mixing with Ferrelview clay and measuring the sorption capacity for select contaminants by batch tests and miniature simulated disposal cells. The wood chips and phosphate were eliminated early in the program due to poor performance (wood chips) and cost (phosphate). The following describes the evaluation of the Ferrelview clay, zeolite, and peat, and the basis of the geochemical barrier design.

The soil component of the geochemical barrier will be composed of low-level radionuclide-contaminated clay soil of the Ferrelview formation. The mineralogy of the Ferrelview clay soil is approximately 50% clay (dominantly smectite) and 50% quartz silt and sand. Consequently the Ferrelview has a capacity to attenuate contaminants in the leachate by virtue of its clay content. Clay minerals have the ability to sorb metal cations by cation exchange, although solution chemistry and the metal species present will determine the whether a particular metal will exchange. Cation exchange capacities of the Ferrelview have been measured at approximately 0.25-0.30 meq/gm (Pacific Northwest Lab [PNL], 1995).The PNL batch tests (unpublished data; Pier et al., 1996) showed that total uranium was reduced from 67,000 pCi/L to 15,000 pCi/L in a control test using clean soil, demonstrating that the Ferrelview has a significant capacity to attenuate uranium.

The Ferrelview contains significant Fe and Al oxides, at 0.8% and 0.2%, respectively, as well as about 3% organic carbon (PNL, 1995). Both the oxides and the organic carbon contribute to the cation exchange capacity of the soil, while also providing an anion exchange capacity (Sposito, 1989). The anion exchange capacity may be significant in controlling alkalinity by sorption of HCO₃ and CO₃ from the leachate solution. PNL 1995 showed that reaction of unmodified borrow and "RAD" soil with simulated leachate reduced pH values to near-neutral. The test pad study results also showed that both peat and zeolite versions of the geochemical barrier reduced the pH of the leachate from up to 13 to near neutral. This effect can probably be attributed to the Fe oxide and carbon content of the unmodified clay, possibly aided by hydrogen exchange from clay. The peat may also play a role in decreasing the pH, as discussed below. Zeolites have a negligible anion-exchange capacity, and are not used for alkalinity control, therefore the pH decrease was likely due to the Ferrelview clay.

The PNL batch tests (unpublished data; Pier et al., 1996) also investigated the consequences of leaching "RAD" soil, containing up to 820 pCi/gm uranium, with CSS leachate. The resulting leachate contained rather high levels of uranium, at up to 67,000 pCi/L. To reduce these high concentrations, the specification for the soil component of the barrier stipulates that "select soil waste" be used, in which U₂₃₈ concentrations are less than 100pCi/gm and Ra₂₂₆ concentrations are less than 5 pCi/gm.

Zeolites have the potential to increase the cation exchange capacity of the clay soil significantly. Clinoptilolite, the zeolite considered for the geochemical barrier, have cation exchange capacities of from about 1.0 to 3.0 meq/gm, compared to the 0.25-0.30 meq/gm measured for the Ferrelview clay.

Clinoptilolite, like other zeolites, is an open framework aluminosilicate mineral whose sorption properties, the mechanism of which is cation exchange, are a function of its crystallography and chemical composition. The framework structure is determined by the specific arrangement of aluminosilicate tetrahedra, which are arranged to form channels and cavities that contain water and exchangeable cations. Clinoptilolite contains channels of two sizes - 7.9 x 3.5D and 4.4 x 3D. Consequently cations larger than these sizes will not fit into the clinoptilolite structure. This property is fundamentally different than cation exchange properties of clays, which are aluminosilicate minerals composed of layers of interconnected tetrahedra. Clays can expand to accommodate cations of varying size between the layers, whereas zeolites cannot and are more selective as to the cation size (the hydrated molecular diameter) as well as charge.

Clinoptilolite has a generally well-defined selectivity sequence for cations that approximates: Cs > Ca > Cd > Cu > Zn > Al > Mg > Al > Na. This sequence will vary depending on relative concentrations and solution chemistry, especially the presence of complexing ligands. Radionuclides that are present as simple cations, such as Cs and Sr are effectively sorbed by zeolites. However, large complex cations such UO₂ ²⁺are not affected (Vaniman and Bush, 1993).

Clinoptilolite has a negligible anion exchange capacity, and therefore will not sorb anions, such as NO₃, or elements that are present in solution as anionic complexes, such as As, likely present in the WSSRAP leachate as H₂AsO₄^- or HAsO₄^2-, or U, present as UO₂(CO₃)₃^4- or a related species.

The PNL Batch tests (Pier et al., 1996) and the WSSRAP Test Pad Study (MK-Ferguson, 1996) confirmed that clinoptilolite has little to no capacity to attenuate uranium. Correspondingly, and according to the characteristics described above, it would also be expected that clinoptilolite would not be effective at reducing the concentrations of other contaminants of concern, such as molybdenum, arsenic, or nitrate, which will also be present as anions. Although radium may be present as a simple cation and therefore amenable to sorption by a zeolite, no data is available to demonstrate as much. For these reasons zeolite was considered inappropriate and eliminated from the geochemical barrier.

Sphagnum peat moss, or peat, is partially carbonized organic matter formed by the partial decomposition of various wetland or marsh vegetation, especially mosses of the genus Sphagnum. The material generally has a high specific surface area, >200 m²/gm, similar to that of activated carbon, which contributes to a high sorption capacity for organic compounds and, to a lesser extent, metals. Peat has an intricate pore structure with a wide variety of pore sizes, ranging from 10s to 100s of D, and a wide variety of pore shapes. Consequently, adsorption by peat is not subject to the molecular size and charge constraints that apply to clays or zeolites.

Peat also has the ability to promote reducing conditions by consuming oxygen and other oxidants as its organic matter content is oxidized and further decomposed. Many metals readily precipitate under reducing conditions. Uranium, in particular, is known to precipitate rapidly under reducing conditions, with equilibrium solution concentrations on the order of a few mg/L (Drever, 1988; Langmuir, 1978). Molybdenum and arsenic also precipitate under reducing conditions generated by organic material (WSSRAP unpublished data; Thompson and Associates, 1988). Redox-sensitive metals concentrations were observed to be much greater than that predicted under reducing conditions in both the PNL Batch Tests and the WSS Test Pad Study (MK Ferguson, 1996), no doubt because oxygen infiltration was not controlled during those tests.

The WSS batch test results (Pier et al., 1996) did show marked reduction in total uranium, which was reduced from 67,000 pCi/L to 4,000 pCi/L with 90 wt% soil and 10 wt% peat, and further reduced to 1000 pCi/L with 20 wt% peat. This reduction is attributed to sorption by the peat rather than the generation of a reducing environment. As noted above, the tests did not control the redox state, and uranium levels would be reduced to much lower levels if redox-controlled precipitation were the operating mechanism. Batch and column tests of geochemical modifiers to acid leachate from UMTRA sites have shown that peat in concentrations as low as 1% removes greater than 90% of uranium, arsenic, lead and sulphate from solution (Thompson and Associates, 1988), but again are attributed to sorption rather than precipitation.

It is not practical to incorporate more than a few weight percent peat into a clay/peat mix because of the low density of the peat. For example, 5 wt% peat is equivalent to about 50 volume percent in a clay/peat mix. The compressibility and low strength of the peat may lead to settlement of the cell, among other problems.

For the reasons stated above it was decided to use peat as the sole amendment, at a percentage based on its ability to consume oxidants and promote a reducing environment. Mass balance calculations show that only about 1 wt% peat is required to induce a reducing environment for the 1000 year life of the disposal cell.

Although the design originally assumed that the peat would be mixed in a 12-inch thick clay/peat layer, it was determined that it would be as effective and less costly to apply the peat as a discrete lift within the Ferrelview clay. 1 wt% peat is about 1.5 inches thick. It was thought that it would be difficult to assure consistent coverage at this thickness, and the peat specification was increased to 2 wt%, or a 3 inch thick layer, which provides an additional margin of safety and is still less expensive than mixing. Note that the calculations do not include oxidizing potential of the 2 to 3 wt% organic carbon content of the Ferrelview clay organic carbon, therefore the addition of 2 wt% peat is an even more conservative measure. Note also that the addition of 2% peat to Ferrelview clay increases the overall cation exchange or sorption capacity by about 15%.

The following summarizes the logic and calculations used to estimate the effective lifespan of 2 wt % peat (20% by volume) used as a geochemical barrier in the disposal cell. Detailed calculations are included in WSSRAP Technical Memorandum 3840TM-7225A.

The peat specified is of pH 3 to 4.5 and > 90% natural organic matter. The acidity of the peat is provided largely by humic acids, that may in part neutralize the alkalinity present in the leachate, as observed in the Test Pad Study (MK-Ferguson 1996). Additionally, the decomposition and mineralization of the organic matter of the peat will release NO₃ and SO₄, which will also contribute to neutralization of alkalinity in the leachate as nitric and sulfuric acids.

A total of 25.6 gallons of seepage/ft² will enter the cell during its 1000-year life. This estimate is based on the conservative assumption that the cell will remain uncovered for three months after the geochemical barrier is completed, during which time a maximum of 2500 gallons/acre-day will infiltrate the barrier. It is assumed that this water will be fully oxygenated precipitation, and during this period dissolved oxygen will oxidize the peat. Following the capping of the cell, the seepage rate through the barrier drops to an estimated 10.1 gallons/acre-day, and continues to decrease, reaching 1.8 gallons/acre-day for years 200 to 1000 (leachate volume estimates are from WSSRAP calculation no. DC-7036).

Oxidation of the peat utilizes dissolved O₂ as the oxidant as long as it is available. When the available oxygen is consumed, oxidation continues by a series of bacterially-mediated reactions, the most important of which, in order of decreasing Eh values, include denitrification (NO₃6N₂), sulfate reduction (SO₄^2-6HS^-), and methane generation (C_organic or CO₂6CH₄). These decay reactions proceed stepwise, as each species is consumed, to the reaction of the next lower redox state (Schwarzenbach et al., 1993, p. 407). For example, sulfate reduction will not occur until all nitrate is consumed.

Oxidation by dissolved oxygen (at an estimated concentration of 9 mg/L) will only consume about 150 mg/ft³ (0.005 ounce/ft³) of peat in the initial 90 days. This amount is negligible compared to oxidation by NO₃ in leachate following placement of the CSS waste, as the CSS leachate is expected to contain up to 4000 mg/L nitrate (PNL, 1995) . Therefore oxidation by dissolved O₂ will be ignored and it will be assumed that all infiltrating water will contain 4000 mg/L NO₃. This is a conservative assumption as it considers that all precipitation that infiltrates the geochemical barrier will first percolate through CSS waste. The CSS leachate contains only 10-20 mg/L sulfate, therefore oxidation by sulfate can also be ignored for purposes of estimating peat oxidation. However, the oxidation of HS^- may be important in the coupled reductive precipitation of uranium, as described below.

It is assumed here that oxidation of the peat by NO₃^- will proceed until all NO₃^- is consumed. Note that denitrification occurs at Eh's of 0.4 to 0.8 Volts, which is too high to precipitate uranium. Oxidation by NO₃^- will consume only about 17% of the available organic carbon in the peat. Following oxidation of the peat by NO₃^-, the next significant decomposition process will be methane generation. Methane generation occurs at a Eh of -0.2 to -0.5 Volts, which is sufficiently low to promote uranium precipitation (Langmuir, 1978). The reductive precipitation of uranium is the result of coupled oxidation and reduction reactions, as all redox reactions are, and therefore requires that a corresponding oxidation occurs. The oxidation of methane to CO₂ in itself may be kinetically unable to precipitate uranium, although there are a variety of other reactions that will effectively serve as the coupled oxidation reaction. Potential oxidation reactions include the oxidation of Fe²⁺ or HS^-, both of which are present in sufficient amounts to provide the corresponding oxidation reaction to the reduction of uranium.

It follows from the above discussion that about 80% of the organic carbon of the peat will be available for fermentation reactions to generate methane, after all oxygen, nitrate, and sulfate is consumed. It is not possible to accurately estimate the reaction rate for fermentation in a system such as this for several reasons, including the fact that fermentation involves numerous different microbiologically-catalyzed reactions, each with different rates, and toxic substances are likely to be present that will impede the growth of microorganisms. The fact that organic matter persists in sediments for geologic time, such as coal and petroleum, indicates that such reactions often occur very slowly or incompletely. Note that this discussion and these calculations do not include the effects of the organic carbon content initially present in the Ferrelview, which has been measured at up to 3% (PNL, 1995), and which will provide additional capacity to ensure an anaerobic environment. As discussed below, the infiltration rate through the barrier is sufficiently slow to allow methane to diffuse through the disposal cell faster than it will be generated, eliminating the potential for methane accumulation.

The infiltration rate at which the leachate moves through the geochemical barrier will be low enough to allow sufficient residence time for sorption and/or precipitation reactions to occur. A maximum of 5.25 gallons/ft² of leachate will infiltrate the barrier in the first 90 days. The field capacity of the peat-amended clay will be on the order of 3-4 gallons/ft³. Although under saturated conditions the leachate might move through the 12-inch thick (30.5 cm) barrier in as little as 100 days, because the amount of infiltration is limited, the barrier will be under non-saturated conditions. Therefore travel times are likely to be orders of magnitude greater. Furthermore, following the capping of the disposal cell, infiltration will be drastically reduced, to an average of 0.02 gallons/ft²-year for the 1000 year life of the cell. This is equivalent to a few drops per square foot per day. At these rates decades to hundreds of years will be required for water to percolate through the barrier. Consequently sufficient time will be available for all bacterially-mediated reactions to proceed to their conclusion, ensuring that an anaerobic environment exists in the geochemical barrier through the life of the disposal cell.

The discussion above shows that the simple mixing of 2 wt % peat to 98% Ferrelview clay will ensure that reducing conditions exist at the base of the disposal cell fro the 1000-year life of the cell. Other additives that were considered, such as zeolites, were shown to be ineffective and were eliminated from the barrier design.

CTL, 1997a. Soils Laboratory Testing for Portion of Test Request No. 5, [Partial] Summary of Laboratory Testing. MKES Doc. No. 3840-D:EN-L-09-13158-00. September.

CTL, 1997b. Consolidated-Undrained Triaxial Test Data (GCS-1, GCS-2, and GCS-3).

MKES Doc. No. 3840-D:EN-N-02-13205-00. October.

Pier, J.; Hodges S.; and Bailey, R.; 1996; Investigation of Leachates and a Geochemical Barrier for the WSSRAP Disposal Cell. Unpublished memorandum.

Langmuir, D.; 1978; Uranium Solution-Mineral Equilibria at Low Temperatures with Applications to Sedimentary Ore Deposits. Geochemica et Cosmochemica Acta V. 42 p.547.

MK-Ferguson, 1996; CSS Test Pads and Related Tests, Test Execution Plan, Final Report. Report No. 3840R-7211-00

Pacific Northwest Laboratory; 1995; Waste Form Leachate Interaction with Soils from Weldon Spring.

Schwarzenbach, R.P.; Gschwend, P.M.; and Imboden, D.M.; 1993; Environmental Organic Chemistry. John Wiley and Sons.

Thompson and Associates, 1988; Neutralization and Chemical Reduction for Immobilizing Inorganic Contaminants in Uranium Mill Tailings, Gunnison, Colorado. Thompson and Associates, Albuquerque, NM, 505-268-6003.

Vaniman, D.T; and Bush, D.L.; 1993; The Importance of Zeolites in the Potential High-Level Radioactive Waste Repository at Yucca Mountain, Nevada. In Natural Zeolites '93, Occurrence, Properties, Use. Internation Committee on Natural Zeolites, Brockport, New York.

BACK