UZ Colloid Transport Model Rev 00, ICN 00 ANL-NBS-HS-000028 April 2000 1. PURPOSE The UZ Colloid Transport model development plan (CRWMS M&O 1999a) states that the objective of this Analysis/Model Report (AMR) is to document the development of a model for simulating unsaturated colloid transport. This objective includes the following. 1. Use of a process level model to evaluate the potential mechanisms for colloid transport at Yucca Mountain. 2. Provide ranges of parameters for significant colloid transport processes to Performance Assessment (PA) for the unsaturated zone (UZ). 3. Provide a basis for development of an abstracted model for use in PA calculations. Based on the performance assessment for the viability assessment (DOE 1998), PA determined that the transport of colloids with an actinide irreversibly sorbed or incorporated into the colloid is the first type of colloid to reach the proposed regulatory compliance boundary. Therefore, PA anticipates that the most important colloid-facilitated transport pathway is through the release of waste-form colloids, which are clay colloids formed from the alteration of high-level radioactive waste (HLW) and spent nuclear fuel (SNF). These colloids have actinides incorporated into the clay structure; therefore, the actinides are not available for desorption or reaction with mineral surfaces. Other potential sources of colloids include (1) iron oxyhydroxide colloids formed from corrosion of the waste package or steel components used in the construction of the repository and (2) natural colloids (e.g., clay, silica, and zeolites) that are weathered components of the host rock. Colloid-facilitated transport can occur through the transport of waste-form colloids or sorption of actinides onto other colloids present in the system. Based on the needs of PA, the emphasis of this AMR focuses on the transport of waste-form colloids through the UZ and does not cover the transport of natural colloids to the extent proposed in the development plan. The scope of this AMR is to (1) develop a simplified numerical grid of a two-dimensional discrete fracture, (2) run simulations of colloid transport through the numerical model for waste form colloids, (3) analyze properties that would be used for transport simulations with natural colloids, and (4) assess the important processes for the colloid model that need to be incorporated into the particle tracking algorithm for colloid transport used by PA. The intended use for this process level colloid transport model is to understand the significance of different processes that affect colloid transport such that an abstracted model for PA can be developed. The abstracted model is documented in CRWMS M&O 2000a. This model provides a conservative estimate of physical processes involved in colloid transport that can be derived based on material properties, such as pore size distribution. It is not intended to be representative of the UZ at Yucca Mountain, nor be a site-scale model. At the time of this writing, site-specific data from Busted Butte analyses in the field and laboratory on colloid transport are not available. As a result, these data have not been incorporated as proposed in the development plan. This report is governed by the Office of Civilian Radioactive Waste Management (OCRWM) AMR Development Plan entitled UZ Colloid Transport Model (CRWMS M&O 1999a). ANL-NBS-HS-000028, Rev 00 12 April 2000 INTENTIONALLY LEFT BLANK ANL-NBS-HS-000028, Rev 00 13 April 2000 2. QUALITY ASSURANCE The activities documented in this AMR were evaluated in accordance with QAP-2-0, Conduct of Activities, and were determined to be subject to the requirements of the U.S. DOE Office of Civilian Radioactive Waste Management (OCRWM) Quality Assurance Requirements and Description (QARD) (DOE 2000). This evaluation is documented in CRWMS M&O (1999b and c) and Wemheuer 1999 (activity evaluation for work package WP 1401213UM1). This AMR has been prepared in accordance with procedure AP-3.10Q, Analyses and Models. The conclusions in this AMR do not affect the repository design or permanent items as discussed in QAP-2-3, Classification of Permanent Items. ANL-NBS-HS-000028, Rev 00 14 April 2000 INTENTIONALLY LEFT BLANK ANL-NBS-HS-000028, Rev 00 15 April 2000 3. COMPUTER SOFTWARE AND MODEL USAGE The computer software codes used in this AMR are as follows. The qualification status of the software is indicated in the electronic Document Input Reference System (DIRS) database. 1. Software: FEHM V2.00 (STN: 10031-2.00-00), Sun Ultra Sparc, Unix System Used for: Base model for developing V2.10 FEHM is a finite-element heat and mass transfer numerical code. Version 2.00 of the FEHM application has been tested and verified for a variety of different types of transport problems, including matrix and fracture reactive transport. Detailed information about the verification can be found in the report by Dash et al. (1997). The software was obtained from Configuration Management (CM), is appropriate for the application, and was used only within a range for which it was verified. 2. Software: FEHM V2.10 (STN: 10086-2.10-00), Sun, Unix System Used for: Perform/model colloid size exclusion This version of FEHM is currently being qualified under OCRWM Procedure AP-SI.1Q, Software Management and will be produced under this procedure per Software Activity Number (SAN): LANL-1999-046, Software Tracking Number (STN): 10086-2.10-00. The software is appropriate for the application. 3. Software: TRACRN V1.0 (STN: 10106-1.0-00), Sun Ultra 2, Unix System Used for: Curve fitting The fitting of experimental data on plutonium sorption and desorption onto colloids was done with the Levenberg-Marquardt (LM) based optimization algorithm incorporated into the TRACRN V1.0 code (Travis and Birdsell 1991). The fitting package (TRACR1) is a coupling of TRACRN V1.0 and the LM algorithm as described in Press et al. (1986). The software was obtained from CM, is appropriate for the application, and was used only within a range for which it was verified. ANL-NBS-HS-000028, Rev 00 16 April 2000 INTENTIONALLY LEFT BLANK ANL-NBS-HS-000028, Rev 00 17 April 2000 4. INPUTS 4.1 DATA AND PARAMETERS Locations, brief descriptions, and data tracking numbers (DTN) that were used as input for this AMR are listed in Table 1. The qualification status of the input data is provided in the electronic Document Input Reference System (DIRS) database. All input data are appropriate for the intended use of this AMR, and selection is discussed in Section 6. Table 1. Input Data Sources Data Description Data and Input Tracking Number Location in this Report Unsaturated water retention data from lexan-sealed samples from USW SD-6 measured using a centrifuge GS980908312242.039 Sec. 6.2.1 Moisture retention data from boreholes USW UZ-N27 and UE-25 UZ#16 GS950608312231.008 Sec. 6.2.1 Pore size distribution for TSw4, CH1v, and CH1z LA0002MCG12213.001 Table 3 Hydraulic properties for TSw4, CH1v, and CH1z LB970601233129.001 Sec. 6.1 The radionuclide releases at the edge of the Engineered Barrier System from the expected-value run done for the Total Systems Performance Assessment-Viability Assessment (TSPA-VA). This information is extracted from the 39-radionuclide run, which included all the decay chains. MO9807MWDRIP01.000 Fig. 2 Silica concentration and mass of plutonium in dissolved and colloidal fractions of the leachate from static corrosion tests on defense waste glass at various surface areas of glass-to-solution ratios LL991109751021.094 Sec. 6.2.2.2 Percentages of plutonium and americium released as colloidal, dissolved, and adsorbed forms in leachate of corrosion tests with defense waste glass at 2,000 and 20,000/m LL991109751021.094 Sec. 6.2.2.2 Concentration of silica in solution under which colloids are stable LL991109751021.094 Sec. 6.2.2.2 Concentration of plutonium colloids and total concentration as a function of test duration for defense waste glass at 2,000/m (T = 90°C) LL991109751021.094 Sec. 6.2.2.2 Summary of analyses of glass dissolution filtrates LL000122051021.116 Sec. 6.2.2.2 Experimental data on sorption and desorption amounts for plutonium onto clay colloids LA0003NL831352.001 Secs. 6.2.4, 6.6 Table 6 Model input and output files LA9912MCG12213.001 Figs. 3-12 Forward colloid removal rates in Topopah Spring welded, Calico Hills vitric, and Calico Hills zeolitic tuffs LA0002MCG12213.002 Table 4 ANL-NBS-HS-000028, Rev 00 18 April 2000 Table 1 (Continued). Input Data Sources Data Description Data and Input Tracking Number Location in this Report Colloid removal Kd’s based on varying detachment (reverse) rates in Topopah Spring welded, Calico Hills vitric, and Calico Hills zeolitic tuffs LA0002MCG12213.003 Table 5 Fracture properties for the UZ model grids and uncalibrated fracture and matrix properties for the UZ model layers for AMR U0090, “Analysis of Hydrologic Properties Data.” LB990501233129.001 Sec. 6.1 Forward and reverse filtration rates for microsphere transport through saturated fractured rock for the C-wells experiments LA9912PR831231.006 Sec. 6.2.3.2 Grid flow simulations LB990801233129.003 Sec. 6.1 Scientific Notebooks Used Description of Information Notebook Identifier (Listed by Scientific Notebook Register Number) Location in this Report Accession Number Effect of Organic Material on Plutonium Sorption Kung 1999a, SN-LANL-SCI-177-V1 (LA-CST-NBK-95-001, Volume I) Sec. 6.2.2.2 MOL.19991206.0252 Effect of Organic Material on Plutonium Sorption Kung 1999b, SN-LANL-SCI-177-V2 (LA-CST-NBK-95-001, Volume II) Sec. 6.2.2.2 MOL.19991206.0253 Busted Butte On-Site Logbook #2 Bussod 1999, SN-LANL-SCI-039-V1 (LA-EES-5-NBK-98-020) Sec. 6.2.2.2, 6.4.3 MOL.20000307.0380 4.2 CRITERIA This AMR complies with the DOE interim guidance (Dyer 1999). Subparts of the interim guidance that apply to this analysis or modeling activity are those pertaining to the characterization of the Yucca Mountain site (Subpart B, Section 15), the compilation of information regarding geology of the site in support of the License Application (Subpart B, Section 21(c)(1)(ii)), and the definition of geologic parameters and conceptual models used in performance assessment (Subpart E, Section 114(a)). 4.3 CODES AND STANDARDS No codes or standards are applicable to this AMR. ANL-NBS-HS-000028, Rev 00 19 April 2000 5. ASSUMPTIONS Assumptions made to perform the analysis or develop the model are presented in Table 2 and discussed in Section 6. Table 2. Assumptions Used in the Analysis and Model Development Assumption Number Category Assumption Basis Section Used 1 Pore Size Distribution That the pore size can be estimated from moisture retention curves Capillary bundle theory can approximate the pore size distribution using an equation from Marshall et al. (1996) 6.2.1 2 Percent of Plutonium released as a Waste- Form Colloid That the percent of plutonium that could be released as waste-form colloids can be determined from laboratory experiments That the laboratory observations would serve as a conservative or worst case scenario for the amount of actinides that will be released as waste-form colloids. 6.2.2.1 3 Natural Colloid Concentration That groundwater colloid concentration can be used as conservative estimates for colloid concentrations in unsaturated pore water Preliminary measurements of water from the UZ heater test indicate that the UZ values are within the range of values measured for saturated zone. 6.2.2.2 4 * Colloid Filtration in the Matrix Matrix filtration of colloids in the UZ can be estimated with filtration theory. In the absence of site-specific data, filtration theory as explained and applied in the paper by Yao et al. (1971) and Harvey and Garabedian (1991) can be used as a conservative estimate. The data used in the equation were based on site data or estimated. 6.2.3.1 5 * Matrix Grain Size Distribution Grain size of the TSw4 was assumed to be comparable to silt, CH1v was assumed to be comparable to coarse sand, and CH1z was assumed to be comparable to fine sand. In the absence of available data, the permeability of the units in the model were compared to the permeability ranges for different types of unconsolidated deposits listed in Table 2.2 in Freeze and Cherry (1979). Based on the type of deposit a range of grain sizes for related soils were used from Fig 1.2 in Marshall et al. (1996). 6.2.3.1 6 Far-Field Geochemistry Far-field geochemistry will not vary significantly from current conditions The chemical conditions in the far field have been stable over extended periods of time, and repository activity is not expected to change the chemistry. 6.6 7 * Reverse Colloids Filtration Rates for Matrix Reverse rates determined for the C-wells experiments for the fracture are conservative for the matrix. Fracture flow in the C-wells experiments were conducted under forced gradient conditions and represent the fastest transport possible. The matrix velocities are slower, and the reverse rates are likely to be even slower, so these serve as a conservative value. 6.2.3.1 ANL-NBS-HS-000028, Rev 00 20 April 2000 Table 2. Assumptions Used in the Analysis and Model Development (Continued) Assumption Number Category Assumption Basis Section Used 8 * Colloid Filtration in the Fracture Fracture filtration of colloids in the UZ can be estimated with field parameters developed for the saturated zone C-wells tracer experiments. This is a conservative assumption because it is expected that less transport of colloids will occur in the UZ relative to the what was observed in the SZ data from C-wells (DTN: LA9912PR831231.006) 6.2.3.2 9 Actinide Oxidation State Oxidation state of actinides can be neglected in the simulations included in this AMR. This AMR is looking for general processes and is not a site-specific model. It is not necessary to include detailed oxidation states at this level of modeling, but it is possible to do so in the future if deemed necessary. 6.6 * Requires confirmation based on field and or laboratory data. ANL-NBS-HS-000028, Rev 00 21 April 2000 6. ANALYSIS/MODEL 6.1 DISCRETE FRACTURE MODEL DESCRIPTION FEHM, V2.10 (STN: 10086-2.10-00), is used for the numerical analysis in this AMR. This analysis examines a two-dimensional discrete fracture model (DFM), in which the fracture and matrix are discretely represented. This approach contains a parallel fracture that is more conservative but consistent with the dual-permeability approach used in the PA model. A simplified stratigraphy (not meant to represent the site) that contains three hydrogeologic units, the Topopah Spring Welded (referred to in this report as TSw4), Calico Hills Vitric (CH1v), and Calico Hills Zeolitic (CH1z), was used in the numerical analysis presented in this AMR (Figure 1). These units are equivalent to those considered in abstraction of transport properties for TSPA-SR (CRWMS M&O 2000a, Section 6.4). The hydraulic properties for these units were derived from DTN: LB970601233129.001 (the properties of tsw34 were assigned to the TSw4 model unit). The grid extends from 630 m to 1150 m and models the half distance between fractures, which was 10 m (i.e., the grid is 5 m wide). This distance corresponds to the fracture spacing for the CH2v to CH5v and CH2z to CH5z units. The lower boundary was selected to be 100 m below the water table to isolate detailed flow conditions related to the boundary from the model domain of interest. The upper elevation boundary represents approximately the top of the TSw4. The fracture width is 1 mm based on estimates from the active fracture model developed in CRWMS M&O (2000a, Section 6.2.1), which used data in DTN: LB990501233129.001 to derive the parameters. The vertical grid spacing is 2.5 m. The horizontal grid spacing is variable and extends from 1 mm to 80 cm. Figure 1. Simplified Stratigraphy Used for Process Level Colloid Model ANL-NBS-HS-000028, Rev 00 22 April 2000 A flux of 5 mm/yr (DTN: LB990801233129.003) averaged over the top surface of the model is injected into the fracture only at the top of the model. This allows the redistribution of flow to occur above the repository level (1087 m) such that the flow field relevant to transport is parallel in the fracture and matrix units and is at steady state conditions for the transport simulations. In general, the flow through the system simulates fracture flow in the TSw4, the transition to predominately matrix flow in the CH1v, and matrix flow in the CH1z. The simulations are unsaturated and do not consider perched water systems. Release of actinides or colloids occurs at the repository level. This model examines important mechanisms for colloid transport that will need to be incorporated into a PA calculation and is not intended to represent the site scale or what happens on the site scale. The parameters and analysis used to develop the inputs for this model are discussed in the next section of this AMR. Specific input parameters for the model are also available in DTN: LA9912MCG12213.001, and a sample input file is included as Attachment I. 6.2 INPUT PARAMETERS AND ANALYSIS FOR THE NUMERICAL MODEL 6.2.1 Pore Size Distributions The pore size distribution for the three units considered in the model were estimated from the moisture retention curves (DTN: GS980908312242.039; GS950608312231.008) judged to be representative of the site. The pore sizes were calculated from the following equation (Marshall et al. 1996, Equation 8.1a). . . . - = cos 2 r (Eq. 1) where r = radius of pores (L) . = surface tension of the water (ML2)/(t2L2) . = contact angle (assumed to be zero) . = matric potential (P). This equation is valid for the range of pore sizes between 100 nm and 1000 µm. Many of the pore sizes in units, such as the TSw4, were estimated to be smaller. A correction for pore sizes less than 100 nm can be calculated if the relative vapor pressure of the soil water is known. These data were not available, so the correction was not made, and a slight error will exist in pore sizes calculated to be less than 100 nm. The change in moisture content between measurements was used to determine the percent of pores within each size range. The percent of colloids that could enter a unit was then determined based on the percent of pore sizes above and below the colloid size of interest for the three units considered in the analysis (Table 3). The percent that can enter other units can also be obtained in the same manner. For example, in the TSw4, only 5% of the 450 nm colloids can enter the matrix versus 65% of the 6- nm colloids. This fact has a couple of implications, depending on the unit through which the colloids are transported. For the case in which fracture flow is dominant, large colloids can ANL-NBS-HS-000028, Rev 00 23 April 2000 experience size exclusion from the matrix, which causes them to stay in the fracture, and smaller colloids can enter the matrix, which will increase their residence time in the system. In contrast, for the case in which matrix flow is dominant, large colloids that cannot enter the matrix are physically removed at the interface between the fracture and matrix or between different matrix units as a filter cake. Table 3. Percent of Colloids Allowed to Enter the Matrix Based on Size Colloid Size Units 450 nm 200 nm 100 nm 6 nm TSw4 5 10 20 65 CH1v 45 50 55 85 CH1z 20 25 25 65 DTN: LA0002MCG12213.001 6.2.2 Source Term Parameters 6.2.2.1 Actinides The actinide source term was based on the 10,000-year radionuclide release at the edge of the Engineered Barrier System (EBS) from an expected value run for TSPA-VA (DOE 1998). Actinide release starts at 1,000 years and is continuous until 10,000 years. Figure 2 shows that the release curve has a double hump that starts around 5,000 years. This release pattern can be misleading, and in Figures 3-12 of this AMR, it should not be interpreted as representing a fracture and matrix breakthrough. The simulations reported in this AMR were only run for the 10,000-year duration of the source. Longer simulations were not considered. The source term used for the waste-form analysis was based on a TSPA-VA (DOE 1998) expected case of plutonium release and did not incorporate the release of the waste-form colloids that will be affected by the pH and ionic strength in the canisters. The chemical conditions of the canisters may delay the release of the waste-form colloids; in which case, the breakthrough of natural colloids carrying actinides could occur sooner than the waste-form colloids. The effect of this on the actinide release will have to be evaluated when the source term for the waste-form colloids and aqueous-phase actinides has been refined. In addition, if the release of waste-form colloids is very low, it is possible that transport of natural colloids could dominate colloidfacilitated transport of actinides. ANL-NBS-HS-000028, Rev 00 24 April 2000 DTN: MO9807MWDRIP01.000 Figure 2. Waste-Form Colloid Source Term 6.2.2.2 Colloids The two most important types of colloids for performance assessment are waste-form and natural colloids. The initial modeling emphasis has been placed on the waste-form colloids because the actinide is incorporated into the structure of the colloid such that desorption or redistribution of the actinide is unlikely to occur. In addition, more site-specific data are available on the wasteform colloids; therefore, calculations that are more defensible can be made. The issue between which of the two types of colloids will be dominant depends on the release times for aqueous actinides and the waste-form colloids. Based on the TSPA-VA analysis (DOE 1998), the wasteform colloid arrives first. It is the relative concentration in the source term between the wasteform and pseudo colloids that will determine which will contribute a higher dose. Where data either on natural colloids or processes important for the transport of pseudo colloids are available, the data are reported with the idea that a future revision of this AMR will include those calculations. For these simulations, a conservative approach is taken that assumes that the diffusion coefficient for the colloids is zero, so the predominant transport mechanism is advective. Waste-Form Colloids The waste-form colloids range in size between 6 and 450 nm (DTN: LL000122051021.116) and are typically negatively charged clay particles, stable in the ambient pore water, and unlikely to react chemically with the host rock. To determine the actinide concentration in the waste-form colloids, the source term provided was divided into a dissolved and waste-form colloid fraction. The waste-form colloid fraction was estimated based on information from waste-form colloid ANL-NBS-HS-000028, Rev 00 25 April 2000 work conducted on the percent of plutonium and americium in the dissolved, colloidal, and sorbed phase for various molar concentrations of silica (CRWMS M&O 2000b; DTNs: LL000122051021.116; LL991109751021.094). In the same study (CRWMS M&O 2000b), it was experimentally determined that the waste-form colloids were only stable when the molar concentration was between 2 x 10-3 and 2 x 10-2 M of silica in solution (DTN: LL991109751021.094). From these two sets of data, an average percent of actinides in the colloidal phase was calculated. This percent was multiplied by the source term provided from TSPA-VA (DOE 1998), and the result was used to represent the amount of plutonium incorporated into the colloid structure. Two additional cases were considered that took the highest and lowest fraction of actinides reported in the colloidal phase from the above-mentioned data (DTNs: LL000122051021.116; LL991109751021.094) and multiplied those values by the TSPA source term to create a high and low case. The release of plutonium in the waste-form colloids over the 10,000-year period for the high, average, and low source term case considered is 1.90 x 10-15, 9.79 x 10-16, and 8.15 x 10-17 M, respectively. The waste-form colloid concentration was held constant at 6 x 10-8 M, which is the high end of the concentration measurement from DTN: LL991109751021.094. The actual release of wasteform colloids is expected to vary based on changes in ionic strength and pH in the near-field environment. The colloid-source-term abstraction being developed for PA will determine a site specific waste-form-colloid source term for their analysis. For the purpose of this analysis, a constant colloid source term is adequate for determining significant processes and the range of parameters for use in the PA model and ignores variations in chemistry that would affect the waste form colloid source term. Natural Colloids Natural colloids are formed as the weathered by-products of the host material and are ubiquitous in nature. Multiple measurements have been done for groundwater samples at Yucca Mountain and other areas on the Nevada Test Site. The measured particle concentrations vary between 1.05 x 106 and 2.72 x 1010 particles/mL, with the lowest being for water from well J-13 (Kung 1999a, pp. L-1-5) and the highest for water from well U19q (Kung 1999b, pp. L-8-3 to L-8-5) on Pahute Mesa. These values are consistent with what has been reported in the literature for various groundwaters around the world (McCarthy and Degueldre 1993). 6.2.3 Colloid Filtration Parameters 6.2.3.1 Matrix Conceptually, the matrix in the UZ will have several different effects on colloid transport, depending on the unit being considered. In general, primary colloid removal in the matrix will be due to physical effects, such as small pore sizes and low volumetric water content. Chemical removal could occur onto mineral surfaces, but it is assumed that the far-field geochemistry will not vary significantly from current conditions, and therefore, chemical removal in the matrix is expected to be small. Specifically, the TSw4 has a very small pore-size distribution and is expected to limit the number of colloids that enter the matrix (see Table 3). Size exclusion will force the majority of the colloids to remain in the fractures in this unit, and the small percent of ANL-NBS-HS-000028, Rev 00 26 April 2000 colloids that do enter the matrix will have a high probability of being retained due to filtration in the matrix and at the interfaces between units. For the CH1v, matrix flow is dominant, the pore sizes are large, and the majority of colloids will be able to enter the matrix due to imbibition. Physical removal of colloids in this unit is likely because the matrix is more tortuous than the fractures and variable pore sizes, but the extent is unknown. The base unit in the model is the CH1z, which has a smaller pore-size distribution than the CH1v. This fact is important because it is expected that a significant quantity of colloids could be filtered at the interface between the two units. The CH1z, will also exhibit size exclusion between the fracture and matrix unit for larger colloids. Although a conceptual picture of the mechanism likely to occur in the UZ exists, there are very limited site-specific data available to support the model for colloid transport through any of the UZ units present at Yucca Mountain. Qualitative visual analysis is available on an overcore sample from borehole #7 from the Phase-1 test at Busted Butte (Bussod 1999), which is an unsaturated field test in the Calico Hills formation near Yucca Mountain. One purpose of the test is to address uncertainties in process-level models. To help understand colloid transport mechanisms, microspheres, an analog for colloids, were injected into the rock with a suite of tracers through a membrane. The injection rate for the sample with visible microspheres was 1 mm/yr, which is comparable to current fluxes at Yucca Mountain. In the sample, the fluorescent dye that was used in the latex microspheres was visible for a few millimeters into the matrix and showed a rind of the microspheres well below the injection location. This observation indicates that the microspheres can move into the matrix, since the concentration required to observe them is high, and implies that they probably migrate further in lower concentrations. The formation of the rind at an interface indicates removal of colloids. Whether the colloids were transported past the interface cannot be determined without further analysis of the sample to determine the concentration distribution above and below the observation. As a result, these interpretations should be considered preliminary. In addition, this infiltration rate was low for the Busted Butte experiments, and higher infiltration rates could lead to additional transport. Unfortunately, overcore samples from higher infiltration rates were too wet to obtain good samples to inspect visually for colloids. Therefore, it is known from preliminary analysis at Busted Butte and studies not related to Yucca Mountain that it is possible to have colloid transport under partially saturated conditions (McGraw 1996; Wan and Wilson 1994a and b). For example, McGraw (1996) showed that for column studies with a coarse sand, which would be equivalent to the CH1v, that over 70% of hydrophilic colloids under 280 nm could be recovered through 10-cm columns at volumetric water contents under 10%. In the same study, recovery was lower for hydrophobic colloids and dependent on colloid size. However, for small hydrophobic colloids (20 nm), greater than 80% recovery was observed even at low volumetric water contents. Mechanisms based on the water film thickness and colloid surface properties have been proposed to explain when colloids are transported in the UZ (Wan and Tokunaga 1997; McGraw 1996). However, these mechanistic approaches are based on sand, and there is not a direct correlation or method to estimate retention and removal of colloids through volcanic tuff. When site data become available from Busted Butte, the results can be used to confirm the mechanism proposed in the conceptual model and through other studies. This information can then be used to enhance the colloid transport model and allow additional credit to be taken for colloid removal. ANL-NBS-HS-000028, Rev 00 27 April 2000 In the absence of data that could defensibly be used, a conservative approach was taken to estimate a colloid removal and resuspension term for the matrix. The removal rate of colloids was estimated using filtration theory, which is an irreversible linear kinetic model. It only considers the removal of colloids and does not account for resuspension. Equation 3 from Yao et al. (1971) can be used to determine the removal rate as C d ) n 1 ( 2 3 dL dC a. - - = (Eq. 2) where C = concentration (M/L3) L = depth of the unit (L) d = diameter of the porous media (L) . = single collector efficiency (dimensionless) a = the collision efficiency factor (dimensionless) n = porosity (dimensionless). Filtration theory was developed for saturated conditions, and the parameters developed are representative of these conditions. Using these values for unsaturated conditions, where more filtration is expected to occur, leads to conservative parameters for UZ modeling. Filtration theory has three basic components that show up in the single-collector efficiency term: Brownian motion (diffusion), interception, and gravitational settling. In the original applications of the theory for waste-water treatment through sand packs, the colloids of interest were typically on the order of 1 µm. In contrast, the colloids of interest in these simulations are much smaller. As a result, the effect of the single-collector efficiency on colloid removal is small and not as significant. The formulation in Equation 2, is the same used by Harvey and Garabedian (1991) in their model (Equations 1 and 2) to represent irreversible adsorption of colloids in their Cape Cod field experiment. What is needed for numerical modeling in FEHM V2.10 is a rate term (1/t). Using the right-hand side of Equation 2, less the concentration, and multiplying by the fluid velocity provides this rate term. FEHM then multiplies the rate parameter by the concentration and the porosity to create the sink or removal term within the model during transport. The forward rates are listed in Table 4. Grain sizes for the three units used in this analysis were not available, so they were determined by comparing the permeability of the units in the model with the permeability ranges for different types of unconsolidated deposits listed in Table 2.2 of Freeze and Cherry (1979). The permeability of the TSw was less then silt, the CH1v was between clean sand and coarse gravel, and the CH1z was clean and silty sand. Based on these types of deposit, a range of grain sizes for related soils were used from Figure 1.2 in Marshall et al. (1996). Consequently, the grain size distribution was assumed for the TSw4, CH1v, and CH1z as silt (2–20 µm), coarse sand (200–2000 µm), and fine sand (20–200 µm), respectively. The mean value was used for the calculation of the removal rate. The collision efficiency factor (a) was also not available; therefore, three values were used that covered the range observed (0.1 x 10-3, 10 x 10-3, and 25 x ANL-NBS-HS-000028, Rev 00 28 April 2000 10-3) in the Cape Cod field experiment for bacteria and microspheres (Harvey and Garabedian 1991). Data on the collision efficiency factor are only available for porous media, and no data are available for unsaturated conditions. Values generated from the field experiments are considered more relevant than laboratory experiments, so variations in the collision efficiency factors were evaluated in a sensitivity analysis. Table 4. Forward Colloid Removal Rates Collision Efficiency Factor Colloid Size (nm) Forward Removal Rate (1/hr) TSw CHv CHz a = 0.1 x 10-3 6 2.68 x 10-2 1.28 x 10-5 5.81 x 10-4 100 4.36 x 10-3 3.93 x 10-6 1.11 x 10-4 200 3.57 x 10-3 9.10 x 10-6 1.42 x 10-4 450 6.44 x 10-3 4.05 x 10-5 4.69 x 10-4 a = 10 x 10-3 6 2.68 x 10-0 1.28 x 10-3 5.81 x 10-2 100 4.36 x 10-1 3.93 x 10-4 1.11 x 10-2 200 3.57 x 10-1 9.10 x 10-4 1.42 x 10-2 450 6.44 x 10-1 4.05 x 10-3 4.69 x 10-2 a = 25 x 10-3 6 6.71 x 10-0 3.21 x 10-3 1.45 x 10-1 100 1.09 x 10-0 9.82 x 10-4 2.76 x 10-2 200 8.91 x 10-1 2.27 x 10-3 3.56 x 10-2 450 1.61 x 10-0 1.01 x 10-2 1.17 x 10-1 DTN: LA0002MCG12213.002 Equations 6, 7, and 8 from Yao et al. (1971) can be used to determine the single-collector efficiency as .= .D + .I + .G = 0.9 kT µdpdv . . . . . . 2 /3 +1.5 dp d . . . . 2 + .p - . ( )gdp 2 18 µv (Eq. 3) where .D = colloid collection collision caused by Brownian motion (dimensionless) .I = colloid collection collision caused by interception (dimensionless) .G = colloid collection collision caused by settling (dimensionless) k = Boltzman constant ((ML2)/(Tt2)) T = solute temperature (1/T) µ = fluid viscosity (M/Lt) dp = diameter of the colloid (L) . = fluid density (M/L3) .p = colloid density (M/L3) g = gravitational constant (L/t2) v = fluid velocity (L/t). ANL-NBS-HS-000028, Rev 00 29 April 2000 The single-collector efficiency (.) was calculated for colloids of size 450, 200, 100, and 6 nm. Field and laboratory experiments have shown that a forward rate is not adequate for representing colloid transport through the system; a reverse or resuspension rate is required (Reimus et al. 1999; Harvey and Garabedian 1991). There is not a set of equations that can be used to calculate the reverse rate. Therefore, the rate that was determined from the Cape Cod field experiment of 0.4 day–1 was used as a base case (Harvey and Garabedian 1991). A sensitivity study was done to examine the effect of a higher value of 5 day–1 and a lower value of 0.001 day–1. These values are based on the reverse rates for fracture flow calculated from the C-wells experiments, a saturated-zone well complex 2.4 km east of the potential repository in the Bullfrog and Prow Pass units (Reimus et al. 1999). It is assumed that the fracture values from the saturated zone are conservative and represent the minimum amount of filtration that could occur under partially saturated conditions in the matrix. Table 5 shows the distribution coefficients (Kd’s) for the base case, where the collision efficiency factor was 0.01 for the different reverse rates considered. The values presented are highly conservative and are unlikely to lead to significant colloid removal in the matrix. Values for the fracture cannot be calculated with this formulation, so they were not used in the simulations presented in this AMR. In the field, filtration in the fracture is expected to be an important mechanism, but site specific data are necessary prior to inclusion in the numerical model. Table 5. Colloid Removal Kd’s Based on Varying Detachment (Reverse) Rates Colloid Size (nm) TSw CHv CHz 6 2.68 x 10+3 1.28 x 10+0 5.81 x 10+1 100 4.36 x 10+2 3.93 x 10-1 1.11 x 10+1 200 3.57 x 10+2 9.10 x 10-1 1.42 x 10+1 Kd’s assuming Kr = 0.001/day 450 6.44 x 10+2 4.05 x 10+0 4.69 x 10+1 6 6.71 x 10+0 3.21 x 10-3 1.45 x 10-1 100 1.09 x 10+0 9.82 x 10-4 2.76 x 10-2 200 8.91 x 10-1 2.27 x 10-3 3.56 x 10-2 Kd’s assuming Kr = 0.4/day 450 1.61 x 10+0 1.01 x 10-2 1.17 x 10-1 6 5.37 x 10-1 2.56 x 10-4 1.16 x 10-2 100 8.71 x 10-2 7.86 x 10-5 2.21 x 10-3 200 7.13 x 10-2 1.82 x 10-4 2.85 x 10-3 Kd’s assuming Kr = 5/day 450 1.29 x 10-1 8.10 x 10-4 9.39 x 10-3 DTN: LA0002MCG12213.003 6.2.3.2 Fracture Field data are available from the C-wells tracer experiments conducted in saturated, fractured rock for Yucca Mountain for removal of latex microspheres (used to represent colloids) passing through the Bullfrog and Prow Pass units. Microspheres are the only analog currently available to represent colloid transport, and the C-wells study represents the best source of data. No ANL-NBS-HS-000028, Rev 00 30 April 2000 comparison has been made on the transport of microspheres to waste-form or natural colloids. If it is assumed that locally saturated conditions exist during fracture flow in the UZ, whether flow occurs as a film or as a fingered process, then using the forward removal rate calculated from the C-wells experiment is a conservative estimate. Colloids could also be retarded or transported in the UZ at the air-water interface, which is not present in the saturated zone (SZ). Therefore, more information on the effect of the air-water interface is needed to determine if applying these data is a conservative assumption. Regardless, the methodology used to determine the parameters for the matrix is not applicable, and no additional data exist that could be used to estimate a fracture removal rate. For flow in a fracture, an equilibrium relationship between the forward and reverse rate is dependent on the half aperture of the fracture and the bulk density. For example, in the C-wells analysis, the reverse rate is reported as a function of the fracture aperture (Reimus et al. 1999). The values derived for C-wells required that the fracture equilibrium coefficient have consistent units when entered into FEHM for modeling. To account for the correct units in the fracture within FEHM, the relationship is expressed similar to a partition coefficient, as: . = b k k K r f fracture (Eq. 4) where kf = forward rate (1/t) kr = reverse rate (1/t) b = half aperture of the fracture (L) . = bulk rock density (M/L3). Based on the range of forward rates estimated from C-wells for the Bullfrog and Prow Pass units, a range of forward rates that varied between 0.001 and 0.5 hr–1 was considered (Reimus et al. 1999; DTN: LA9912PR831231.006). The reverse rates considered varied between 0.001 and 4 hr–1. Simulation results on the microsphere breakthrough from C-wells that was completed after the analysis reported in this AMR indicate that the reverse rates multiplied by the fracture half aperture vary between 0.0001 to 3.33 hr–1 (Reimus et al. 1999; DTN: LA9912PR831231.006). This indicates that the reverse rate could be an order of magnitude smaller than considered in this analysis. Various combinations of forward and reverse rates were used in the analysis to address any concerns that the equilibrium model is not as conservative as a kinetic model. 6.2.4 Plutonium Sorption/Desorption onto Clay Colloids Laboratory data on the sorption of plutonium (Pu) onto clay colloids is in the process of being fit to parameters that can be used for numerical modeling (DTN: LA0003NL831352.001). The preliminary results are presented in this section. Data are also available on the sorption of Pu onto silica colloids. ANL-NBS-HS-000028, Rev 00 31 April 2000 6.2.4.1 Curve-Fitting Methodology The curve-fitting of the plutonium sorption onto the colloids was accomplished through a coupling of TRACRN V1.0 (STN: 10106-1.0-00) with a widely used LM algorithm (Press et al. 1986), which is the fitting package TRACR1 within TRACRN. The LM algorithm finds parameter values that minimize a target functional. In this case, the target functional is the sum of squared differences (SSD) between a set of observations (the concentration of sorbed Pu versus time) and the values calculated by the code. The TRACRN code solves a set of partial differential equations plus appropriate boundary and initial conditions that approximate the sorbing and desorbing experiments. The parameters used for matching the data are the forward and reverse sorption rates. An initial estimate of each of these was specified at the beginning of the fitting process. Through a series of TRACRN simulations with perturbed parameter values, which provide derivatives of the simulation results with respect to the various parameters, the LM package searches until the SSD is no longer reduced in value with further parameter changes. Coupling the LM algorithm with a simulated annealing process (Press et al. 1986) provides an approximate global SSD minimum rather than only a local one. A kinetic sorption model is used for the preliminary analysis of Pu(IV) and Pu(V) sorption and desorption onto colloidal matter. In TRACRN, the exchange of plutonium between the aqueous and solid phases is governed by (Travis and Birdsell 1991) .S .t = k fC 1 - S Smax . . . . . . - kr 1- C C0 . . . . . . S (Eq. 5) where kf = forward reaction rate (1/t) kr = reverse reaction rate (1/t) S = sorbed concentration (Msolute/Mcolloid) Smax = maximum sorbing capacity of the colloid (Msolute/Mcolloid) C = aqueous phase concentration (M/L3) C0 = the solubility limit (M/L3). The equation describes the kinetic balance between sorption and desorption processes. Under equilibrium conditions, Equation 5 yields a solution of the Langmuir form: S = G(C) = KdC 1 + KLC (Eq. 6) where G(C) = symbolizes the relation between S and C KL = Kd/Smax Kd = kf/kr ANL-NBS-HS-000028, Rev 00 32 April 2000 6.2.4.2 Preliminary Fitting Results for Clay Colloids There are four experimental cases considered: (1) sorption and desorption of Pu(IV) using J-13 water, (2) sorption and desorption of Pu(V) using J-13 water, (3) sorption and desorption of Pu(IV) using synthetic J-13 water, and (4) sorption and desorption of Pu(V) using synthetic J-13 water. Synthetic J-13 is a synthetic groundwater that only contains sodium carbonate and sodium bicarbonate. The sorption experiments covered a 10-day period, whereas the desorption experiments lasted for over nine months (Lu et al. 1998). Values of the sorbed counts per minute (CPM) per mg of Pu on clay colloids were used as the fitting data. These values were normalized by dividing the measured concentration by the initial concentration. Average values versus time and a standard deviation were determined from the duplicate set of experiments. The fitted values of the forward and reverse rates are given in Table 6. The results are reasonably consistent. The forward rate is on the order of one, with case 3 being somewhat larger. The reverse rates are all within an order of magnitude of each other, around 1 x 10-5. The Smax values are in the range of 250 to 450 Msolute/Mcolloid. Sorption is clearly very rapid, whereas desorption is very slow from the clay colloid material. Desorption data may be affected by solubility limits. A small quantity of water was mixed with the colloids for the desorption experiments in contrast with the sorption experiments, which had a much larger volume of water. Table 6. Preliminary Parameters for the Sorption of Pu onto Clay Colloids kf (1/sec) kr (1/sec) Smax Msolute/Mcolloid Case 1 (J13, Pu(IV)) 1.8 1.56 x 10-5 252 Case 2 (J13, Pu(V)) 1.5 4.39 x 10-5 340 Case 3 (syn. J13, Pu(IV)) 111.5 8.34 x 10-6 457 Case 4 (syn. J13, Pu(V)) 0.45 4.46 x 10-5 430 DTN: LA0003NL831352.001 Other fits to the data based on a different sorption isotherm (e.g., Freundlich) or a two-site-type model, that is, one with a strong and a weakly binding group of sites, may or may not improve the fit. Further, a better fit might be found by letting the search algorithm continue with perturbed parameters. The LM algorithm is a local optimum finder; it is coupled with a global optimum search, which can occasionally find a better fit through a random search. The chisquares for each case were 27,313, 34,997, 349, and 168,510 for cases 1, 2, 3, and 4, respectively. Clearly, case 3 is fit quite well, whereas case 4 is not. 6.3 CONSERVATIVE TRACER To help understand the system through which colloids will be transported, a simulation was conducted with a conservative tracer. The tracer was released at the repository level into a steady-state flow field at time t = 0. The initial breakthrough at the water table for this system takes approximately 1,000 years without diffusion and 1,100 years with a diffusion coefficient representative of tritium transport based on laboratory experiments (Figure 3). This is because the breakthrough time is dominated by matrix flow in the CH1v, such that having or not having ANL-NBS-HS-000028, Rev 00 33 April 2000 diffusion will not significantly affect transport in this model. This case is a conservative one because the fracture is continuous in the system in the discrete model used to represent the system. In the natural system, there would be multiple fractures of differing degrees of connections that will influence transport. The small variation between the non-diffusing and diffusing case exists because the flow in this model is fracture dominated. It is only in the CHv1, which is the thinnest unit in the model, that matrix flow is dominant (Figure 1). DTN: LA9912MCG12213.001 Figure 3. Normalized Breakthrough for a Conservative Tracer 6.4 COLLOID RETARDATION BASED ON PORE SIZE DISTRIBUTION 6.4.1 Modifications to FEHM to Include Size Exclusion In preliminary simulations that examined colloid transport through unsaturated units at Yucca Mountain, it was obvious that for matrix-flow-dominated units, there was a strong flow component that was due to imbibition. This resulted in excess colloids entering the matrix based on colloid size and the pore size distribution. Therefore, the computer code FEHM V2.0 (STN: 10031-2.00-00; Zyvoloski et al. 1997) was modified to accommodate size exclusion. The code was modified so that the convective and dispersive terms of the solute transport equation could be disabled along specified interfaces. This option restricts colloids from flowing into designated regions but allows the colloids to enter other areas in the model domain. It is important to note that, as currently represented in FEHM V2.10 (STN: 10086-2.10-00), this modification only excludes colloids that are too large to enter from the matrix. The colloids are not removed in the fracture or the matrix as a result of this modification. For multiple species or ANL-NBS-HS-000028, Rev 00 34 April 2000 reactive transport, an additional option is available that allows size exclusion to be applied to a single species while allowing other species to be transported freely. A constraint on this method is that the colloid distribution must be broken down into a number of subgroups that have a uniform distribution, and the mean colloid size is used in the model to represent the entire subgroup. Specifically, colloids were only allowed to flow freely when the pore size was larger than the colloid size. As discussed in Section 6.2.1, this constraint was based on a percentage of pore sizes that were larger than the colloid size of interest. To accommodate different percentages of colloids that could enter the matrix based on size or different matrix units, simulations were conducted as separate events and combined using the principle of superposition. The principle of superposition can be applied to solutions of the convection-dispersion equation describing the transport of a conservative solute during steadystate flow conditions. The majority of simulations illustrated in this AMR concerning the transport of polydispersive colloids are linear processes. For reactive transport, the idea of superposition is generally not valid. However, under the condition that the colloid concentration is in excess, which is usually the case in natural field situations, the principle of superposition can be used to approximate the true behavior of the system. If the simulations can be represented as Qin(t) = aQ1in(t) + ßQ2in(t), (Eq. 7) then Qout(t) = aQ1out(t) + ßQ2out(t), (Eq. 8) where Q1out and Q2out are the solute mass outflow rate corresponding to the input rate Q1in(t) and Q2in(t), respectively. For example, the results from a distribution of colloids divided into discrete sizes are simply a linear combination of the results from each size group. For each case examined in this AMR, four separate simulations were run and superimposed to capture size exclusion for the three matrix units of interest. 6.4.2 Demonstration of Size Exclusion for a Pulse Input As incorporated into FEHM V2.10 (STN: 10086-2.10-00), the size exclusion routine does not remove mass from the system. Therefore, whatever enters the system must leave the system. To verify that the method was working correctly, a constant-source-pulse input was simulated for 500 years during steady-state flow conditions. Figure 4 shows the normalized cumulative colloid mass released at the groundwater table while controlling the units that colloids are allowed to enter. For the case in which the colloid was only allowed into the fracture, a fast arrival of the colloid to the groundwater is observed. On the other hand, when colloids can flow freely into all the matrix units, the longest travel time is observed. Colloid transport through the Calico Hills vitric (middle unit) increases the travel time, and when the colloids are allowed to enter the zeolitic unit, the initial arrival time doubles. The mean colloid travel time in these simulations is proportional to the area of the units the colloids were allowed to enter. Figure 4 also illustrates ANL-NBS-HS-000028, Rev 00 35 April 2000 that complete recovery of the colloid through the system is obtained during the 15,000-year simulation period. DTN: LA9912MCG12213.001 Figure 4. Normalized Breakthrough for a 500-Year Pulse Input Figure 5 illustrates the relative behavior for colloid transport in the fracture and in the matrix. In the simulation, five observation points were selected along the center horizontal transect of the CH1v. Colloids were allowed to enter the fracture and CH1v unit but were completely excluded from the TSw4 and CH1z. The results indicate that the colloids were transported into the CH1v unit when the pulse of solute passed through the fracture domain. However, the farther into the matrix that the observation was made, the lower the relative concentration of colloids. This result is partly due to the time it takes for the colloids to be transported advectively into the matrix, but it is also an indication of a model limitation. That is, the model is a finite domain of 5 m with a discrete fracture on one side of the model and a no-flow boundary on the other. In the natural system, the distance used in the model would represent the distance between fractures, and colloids that entered the matrix would be just as likely to diffuse into a different fracture as they would be to reenter the fracture where they started. This process would enhance the travel time of colloids in the matrix and enhance retardation of the colloids. However, given these constraints, whatever enters the matrix will eventually be transported back into the fracture. ANL-NBS-HS-000028, Rev 00 36 April 2000 DTN: LA9912MCG12213.001 NOTE: These curves show normalized breakthroughs along the center horizontal transect of the CH1v at 865 m where colloids were allowed to enter the fracture and CH1v unit only. Figure 5. Normalized Breakthroughs Along the Center of CH1v 6.5 WASTE-FORM COLLOID SIMULATIONS The waste-form colloid is considered the largest source of actinide release associated with colloids from the EBS based on analysis for TSPA-VA (DOE 1998). Therefore, it is considered in detail in this analysis. Although it is possible for actinides present in the dissolved phase to react with the waste-form colloid, thereby increasing the actinide loading, that scenario is not considered in this analysis. This AMR focuses on the transport of waste-form colloids through the UZ and the effects of size exclusion and colloidal removal on the transport. 6.5.1 Size Exclusion With and Without Colloid Removal The first of two cases compares the significance of size exclusion from the matrix with colloid retardation/removal based on a linear kinetic model, which is described in the FEHM V2.0 manual (Zyvoloski et al. 1997). The parameters used are the Kd’s and forward rates based on filtration theory discussed in Section 6.2.3.1. Figure 6 shows the results for the four different colloid sizes considered. The initial breakthrough for the case with and without removal was around 500 years after actinide release at 1,000 years, which is shorter than the 1,000 years after release observed for the conservative tracer in Figure 3. This result indicates that the sizeexclusion mechanism can increase the transport rate of colloids through the UZ. In addition, these results indicate an insignificant difference between the simulation with and without ANL-NBS-HS-000028, Rev 00 37 April 2000 removal. This is due in large part to the rate parameters used in the simulations, which are based on filtration theory and are highly conservative. The rates do not completely account for physical or chemical removal, nor do they account for the effect of reduced volumetric water content because these are not part of filtration theory. All of these factors would enhance the colloid removal rate in the system. The sensitivity of the rate parameters will be examined further in a sensitivity analysis later in this section. Figure 7 compares the relative importance of size exclusion and colloid removal on the simulation results for a 100-nm colloid. The figure shows that, for the cases with size exclusion, the results are the same with or without the colloid removal rate, as seen in Figure 6. However, if size exclusion is not incorporated, imbibition into the matrix helps retard the colloids, and the removal rate enhances the retardation effect further. This result indicates that colloid removal can be a significant process, but based on the conservative estimates made, it is not a significant process in these simulations. 6.5.2 Effect of Source Term High, average, and low source terms were considered in the analysis as discussed in Section 6.2.2.1. Figures 8 and 9 show the results for a 6- and 100-nm colloid with and without colloid removal. As shown in Figure 2, the shape of the curve reflects the input data, and the distance between the curves represents the relative differences in mass input. The results for these figures are nearly identical, which indicates that the size exclusion is more important than colloid removal. The release rate of waste-form colloids can vary several orders of magnitude, depending on the initial source term, as shown in these figures. Therefore, having a good model of colloid release rates that incorporates chemical effects, such as pH and ionic strength, is a key component of developing realistic colloid breakthrough scenarios. 6.5.3 Effect of Colloid Removal in CH1v The most permeable unit in these simulations is the vitric Calico Hills (CH1v), which also has the largest pore size distribution as calculated in Section 6.2.1. As a result, it is expected that a significant quantity of colloids will enter this unit. In addition, this unit will have the greatest potential to remove colloids physically due to removal in small pores and removal at the interface with the lower unit (CH1z) that has a smaller pore size distribution. Using the data in Table 5, a set of simulations was done to compare a relatively small and large distribution coefficient (Kd), which is the ratio of the forward removal rate to the reverse rate. For the case considered, the forward rate varied by three orders of magnitude, and the reverse rates were comparable between the simulations. The results show that, for the case in which size exclusion is included, the colloid removal rate is still sufficiently low that the breakthrough is the same for the high and low case (Figure 10) and fairly similar to the case where only fracture transport is considered. In contrast, when the colloids are not excluded from entering the matrix, the breakthrough for both the high and low Kd values breaks through over 4000 years later. The difference is due to whether or not the colloids enter the matrix because the Kd values used are sufficiently small that they do affect the results for this simulation. This simulation clearly indicates that, when fracture flow is possible, transport is controlled by colloids that remain in the fracture because they are restricted from entering the matrix based on size. ANL-NBS-HS-000028, Rev 00 38 April 2000 DTN: LA9912MCG12213.001 Figure 6. Colloid Breakthrough for Different Colloid Sizes with and without Removal DTN: LA9912MCG12213.001 Figure 7. Effect of Size Exclusion Versus Colloid Removal for 100-nm Colloids ANL-NBS-HS-000028, Rev 00 39 April 2000 DTN: LA9912MCG12213.001 Figure 8. Effect of Source Term on Colloid Breakthrough without Removal DTN: LA9912MCG12213.001 Figure 9. Effect of Source Term on Colloid Breakthrough with Colloid Removal ANL-NBS-HS-000028, Rev 00 40 April 2000 If size exclusion is not considered, the difference is, of course, significant. For example, the colloids with a low removal rate break through around 6,000 years, whereas the case with a high removal rate does not break through during the 10,000 years of simulation. A Kd value that was six orders of magnitude larger than the low retardation case was also tested (not shown). This simulation indicates that large Kd values retard the same amount of colloids that enter the matrix. The limitation is that, even for the CH1v unit where the flow is matrix dominated, the amount of colloids in the matrix relative to fracture is relatively small. This is why no difference is observed for the case in which size exclusion is included. The result that size exclusion is a more significant process than colloid removal holds when a continuous fracture or fracture network is present. The model in this AMR is conservative in the assumption that a continuous fracture exists. Size exclusion only occurs at the interface between the fracture and the matrix in this model. So, when a particle is excluded, it is not removed from the simulation as might be expected in a natural system where there is no longer flow in the fracture. Instead, it continues to be transported in the fracture. For cases where the fracture ends or the predominant flow direction is from the fracture to matrix, the size exclusion represents one form of filtration. Another type of filtration not considered in this analysis is filtration at interfaces between matrix units. This type of filtration can lead to the formation of a filter cake at interfaces, which will help remove a significant quantity of colloids. Unfortunately, even though this mechanism is known to occur and visual observations from the Busted Butte experiments show the formation of a colloid rind at such an interface (Bussod 1999), this mechanism is not incorporated into this model. It has been incorporated to the abstracted colloid model discussed in CRWMS M&O (2000a) by comparing the pore sizes to the colloid sizes. When the pore size is too small the colloids are filtered at the interface. This is a conservative approach taken for the TSPA calculations, but not available at the time of the analysis for this AMR. However, without more quantitative data the extent to which this mechanism can help reduce colloid transport is unknown. This is especially true because of how significant colloid transport through fractures will be for the UZ. ANL-NBS-HS-000028, Rev 00 41 April 2000 DTN: LA9912MCG12213.001 Figure 10. Colloid Breakthrough as a Function of Removal Rates in the CH1v DTN: LA9912MCG12213.001 Figure 11. Colloid Breakthrough as a Function of Fracture Removal Rates ANL-NBS-HS-000028, Rev 00 42 April 2000 6.5.4 Effect of Colloid Removal in Fracture Given the significance of colloid transport in the fracture, a final sensitivity study examined the range of possible colloid removal rate in the fracture based on data available from C-wells and discussed in Section 6.2.3.1. The results follow the expected pattern, that is, the higher the removal rate, the later breakthrough occurs (Figure 11). The ranges considered in these simulations are limited, and the data available from C-wells indicate that the removal rate could be significantly higher than the values used. In addition, this model is of the UZ versus the Cwells tests that were conducted under saturated conditions. The reduction in water content could lead to additional colloid removal and retardation, but data are needed to support higher removal rates. A limitation of the model is that the fracture size considered is consistent throughout the model although this is unlikely to be the case in the field. It is known from the C-wells tests in the Bullfrog and Prow Pass units that, for the less-permeable Prow Pass unit, the colloid removal rate was higher; a corollary is expected in the UZ. 6.6 NATURAL COLLOID SIMULATION For natural and waste-form colloids, it is possible for aqueous phase actinides to sorb and desorb onto colloids and be transported through the system. The analysis of experimental data presented in Section 6.2.4 is preliminary and has not been completed. Therefore, a preliminary calculation was done to test the relative significance of actinide transport on natural colloids. Based on the experimental results for Pu sorption onto different types of colloids, it is known that this process is kinetic with a fast forward rate and a slow reverse rate (DTN: LA0003NL831352.001). In the most simplistic form, the reaction can be written as Puaq + Colloidaq .. PuColloidaq (Eq. 9) For this case, only the sorption and desorption of aqueous plutonium onto colloids that are stable in suspension are considered. Other reactions, such as aqueous plutonium sorption onto the matrix or immobile colloids, colloid removal, and reactions with fracture minerals are not included in this preliminary analysis. As written, this reaction requires several simplifying assumptions, which include neglecting the oxidation state and changes in water chemistry that will determine the species and complexes that can form. Although these factors are important, there are limited data on the changes that would occur in the far field, and the simplification is sufficient for understanding the significance of the mechanisms on actinide transport through the UZ. Figure 12 shows the results of the preliminary simulation. The rate terms used for this simulation (Table 6) were based on a very preliminary analysis of the experimental data on Pu sorption to clay colloids from the results presented in Section 6.2.4.2. In addition, the early breakthrough of aqueous Pu observed in the figure is due to the source term used and the fact that sorption onto the matrix was not considered However, the results still illustrate two points. First, using the average case of plutonium released in the dissolved phase, the aqueous phase of Pu breaks through after 6,000 years, even without considering sorption to the matrix. This result is important because, unlike the colloids, the Pu can diffuse into the matrix, which provides additional retardation. The second point is that breakthrough of the Pu-colloid complex, even ANL-NBS-HS-000028, Rev 00 43 April 2000 DTN: LA9912MCG12213.001. Figure 12. Colloid Breakthrough for Kinetic Reaction of Puaq and Clay Colloidsaq considering size exclusion for the colloids, does not occur until after 5,000 years. This result is significantly greater than the 1,500 years observed for the waste-form colloids (Figure 6), even for the cases that considered colloid removal, which was not done in this simulation. In addition, the concentration that breaks through is lower than for the waste-form colloids over the same breakthrough period. The point is that, for natural colloids or other degradation products (such as iron that may form colloids in the repository area), the reaction of these colloids with aqueousphase actinide species and, then, the transport of these complexes is much lower than for the waste-form colloids. In addition, the high sorption capacity of the surrounding rock and minerals will further limit the availability of aqueous-phase actinides for participation in these reactions. This is not to say that it will not occur, just that, relative to the waste-form colloids, this mechanism is likely to play a minor role in releases from the repository. 6.7 MODEL VALIDATION There are two parts to validation of FEHM V2.1 for colloid transport modeling. The first part is to run a standard suite of problems to make sure that the flow and solute transport parts of the code are working correctly after modifications. This will be done using the test problems documented in Dash et al. (1997). The second part is to validate the subroutines and modifications specific to the transport of colloids. Unfortunately, there is not a standard test problem, analytical solution, field, and/or laboratory data that could be used for validation of the model. As a result, the approach taken has been to develop a conservative model that is defensible based on the physical properties of the fracture and matrix that could transport colloids given the lack of site-specific data. The model considers the physical removal that ANL-NBS-HS-000028, Rev 00 44 April 2000 would result from the colloids being smaller than the distribution of pore sizes. This is a defensible approach because if the pore sizes were smaller than the colloids of interest, no one would expect for the colloids to be transported through that path. Similarly, by using the conservative values for filtration in the matrix derived from the well established theory of filtration for saturated systems, the additional filtration that would result from the unsaturated conditions are under estimated rather than ignored. Thus, the input parameters are judged reasonable for the model use. The model also accurately reproduces the input data. Data and parameters used for modeling were either site specific or were derived from field and laboratory measurements made at other sites. Given the limited range of data available, many process know or expected to occur were not considered, including filtration at the interface between matrix units, the effect of volumetric water content on colloid removal, surface chemical properties of the colloids, etc. As a result, the output from this model would over estimate the amount of colloids likely to be transported. This conservative analysis provides a method for addressing the colloid issue without having to make the assumption that they can all be transported. The model and its assumptions and approximations were intended to be realistically conservative, that is to capture the essence of the major processes in a context that would not underestimate the efficiency of colloidal transport given the limitations on availability of data to support the model. The model is assessed to be valid for its use as a tool to examine the impact of different processes that could then be incorporated into the particle-tracking code used by PA to model colloids (CRWMS M&O 2000a). The availability of site-specific data would permit further validation of both models and enhancement of the processes considered. ANL-NBS-HS-000028, Rev 00 45 April 2000 7. CONCLUSIONS The UZ colloid transport numerical model indicates that waste-form colloids can be transported through the UZ. The size-exclusion mechanism that forces colloids to remain in the fractures enhances transport in rock units with fracture flow and retards transport in units where matrix flow dominates because there is no advective flux in the fracture to move the colloids. This modeling is not intended to represent the site and does not consider many of the mechanisms likely to occur in the real system. Instead, it describes conservatively the transport possibilities given our assumptions. The modeling and analysis to date indicate that some colloids, particularly waste-form colloids or natural colloids with kinetically controlled actinide sorption/desorption, could enhance actinide transport. These results depend on many assumptions, including the colloid and actinide source term used, which in turn will depend on the design of the engineered barriers themselves. Additional site-specific data on the colloid filtration parameters through the different UZ geologic units, physical and chemical properties of the colloids, local chemical conditions around the waste canisters, grain-size distributions, and pore-size distributions are necessary to enhance the current model. In addition, site-specific data on the removal of colloids, particularly at the interfaces between units and the interface between fractures and matrix material, would help to better constrain the amount of colloids that can be transported through the UZ. Although the assumptions made for this analysis were adequate to develop a colloid model for PA that is better than what was used for the viability assessment (DOE 1998), there is still considerable uncertainty that needs to be addressed. The addition of other mechanistic processes into the numerical model is not the limitation. The limitation is not having the data to support implementation of the processes, which leads to assumptions made on top of assumptions, making it more difficult to defend the conservative approach taken in this analysis. Given the caveats and assumptions of the model, it can be concluded that the waste-form colloids will play a more significant role than natural colloids because the actinides are already incorporated into the structure and do not have to compete with matrix and fracture minerals. For the waste-form colloids, the most significant mechanism is size exclusion of colloids from the matrix, which leads to fracture-dominated transport of these colloids. Although this effect occurs for natural colloids, the reaction of actinides with the colloids is not restricted to the fracture, and transport of the colloids without actinides do not affect repository performance. Based on the priorities set by the PA analysis, modeling of the natural colloids was very limited. Additional analysis is necessary to incorporate the available data for sorption of Pu and Am onto different types of colloids and to evaluate the relative importance of the different colloids on transport. This study is particularly important if the aqueous release of actinides is greater than the portion incorporated into the waste-form colloids. A better source term will be developed from current PA analysis, which provides a feedback loop for the colloid modeling. The data and model developed by this analysis are included in DTN: LA9912MCG12213.001. ANL-NBS-HS-000028, Rev 00 46 April 2000 This document may be affected by technical product input information that requires confirmation. Any changes to the document that may occur as a result of completing the confirmation activities will be reflected in subsequent revisions. The status of the input information quality may be confirmed by review of the Document Input Reference System database. ANL-NBS-HS-000028, Rev 00 47 April 2000 8. INPUTS AND REFERENCES 8.1 DOCUMENTS CITED Bussod, G.Y. 1999. LA-EES-5-NBK-98-020. Busted Butte On-Site Logbook #2 UZ Transport Field Test. SN-LANL-SCI-039-V1. ACC: MOL.20000307.0380. CRWMS M&O (Civilian Radioactive Waste Management System Management and Operating Contractor). 1999a. UZ Colloid Transport Model, Rev 00. TDP-NBS-HS-000052.. Las Vegas, Nevada: CRWMS M&O. ACC: MOL.19990831.0050. CRWMS M&O 1999b. M&O Site Investigations. Activity Evaluation, January 23, 1999. Las Vegas, Nevada: CRWMS M&O. ACC: MOL.19990317.0330. CRWMS M&O 1999c. M&O Site Investigations. Activity Evaluation, September 28, 1999. Las Vegas, Nevada: CRWMS M&O. ACC: MOL.19990928.0224. CRWMS M&O 2000a. Particle Tracking Model and Abstraction of Transport Processes. ANLNBS- HS-000026, Rev 00. Las Vegas, Nevada: CRWMS M&O. URN-0037. CRWMS M&O 2000b. Water Form Colloid-Associated Radionuclide Concentration Limits: Abstraction and Summary. ANL-WIS-MD-000020, Rev 00. Las Vegas, Nevada: CRWMS M&O. ACC: Draft AMR. Submit to RPC. URN-0209. Dash, Z.V.; Robinson, B.A.; and Zyvoloski, G.A. 1997. Software Requirements, Design, and Verification and Validation for the FEHM Application: A Finite Element Heat and Mass Transfer Code. LA-13305-MS. Los Alamos, New Mexico: Los Alamos National Laboratory. TIC: 235234. DOE (Department of Energy) 1998. Total System Performance Assessment. Volume 3 of Viability Assessment of a Repository at Yucca Mountain. DOE/RW-0508. Washington, D.C.: U.S. Department of Energy, Office of Civilian Radioactive Waste Management. ACC: MOL.19981007.0030. Dyer, J.R. 1999. “Revised Interim Guidance Pending Issuance of New U.S. Nuclear Regulatory Commission (NRC) Regulations (Revision 01, July 22, 1999), for Yucca Mountain, Nevada.” Letter from J.R. Dyer (DOE) to D.R. Wilkins (CRWMS M&O), September 9, 1999, OL&RC:SB-1714, with enclosure, “Interim Guidance Pending Issuance of New U.S. Nuclear Regulatory Commission (NRC) Regulations (Revision 01).” ACC: MOL.19990910.0079. Freeze, R.A. and Cherry, J.A. 1979. Groundwater. Englewood Cliffs, New Jersey: Prentice- Hall Inc. TIC: 217571. Harvey, R.W. and Garabedian, S.P. 1991. “Use of Colloid Filtration Theory in Modeling Movement of Bacteria through a Contaminated Sandy Aquifer.” Environmental Science and Technology, 25 (1), 178–185. Washington, D.C.: American Chemical Society. TIC: 245733. ANL-NBS-HS-000028, Rev 00 48 April 2000 Kung, S.K. 1999a. Batch Sorption Studies, Geochemistry, Site Characterization. 8.3.1.3.4.1, R0. LA-CST-NBK-95-001, Volume I. ACC: MOL.19991206.0252. Kung, S.K. 1999b. Research and Development Notebook for Colloid Study Notebook. LACST- NBK-95-001, Volume II. ACC: MOL.19991206.0253. Lu, N.; Triay, I.R.; Cotter, C.R.; Kitten, H.D.; and Bentley, J.. 1998. Reversibility of Sorption of Plutonium-239 onto Colloids of Hematite, Goethite, Smectite, and Silica. LANL Report LA-UR- 98-3057. Los Alamos, New Mexico: Los Alamos National Laboratory. ACC: MOL.19981030.0202. Marshall, T.J., Holmes, J.W., and Rose, C.W. 1996. Soil Physics, Third Edition, 207–212. New York, New York: Cambridge University Press. TIC: 246638. McCarthy, J.F. and Degueldre, C. 1993. “Sampling and Characterization of Colloids and Particles in Groundwater for Studying their Role in Contaminant Transport.” Chapter 6 of Environmental Particles, Buffle, J. and van Leeuwen, H.P., editors. Boca Raton, Florida: Lewis Publishers. TIC: 245905. McGraw, M.A. 1996. The Effect of Colloid Size, Colloid Hydrophobicity, and Matrix Saturation on Colloid Transport in the Subsurface. Ph.D. dissertation. Berkeley, California: University of California. TIC: 245722. Press, W.H.; Flannery, B.P.; Teukolsky, S.A.; and Vetterling, W.T. 1986. Numerical Recipes: The Art of Scientific Computing. New York, New York: Cambridge University Press. TIC: 234187. Reimus, P.W.; Adams, A.; Hagga, M.J.; Humphrey, A.; Callahan, T.; Anghel, I.; and Counce, D. (with contributions from USGS staff). 1999. Results and Interpretation of Hydraulic and Tracer Testing in the Prow Pass Tuff at the C-Holes. Milestone SP32E7M4. Los Alamos, New Mexico: Los Alamos National Laboratory. TIC: 246377. Travis, B.J. and Birdsell, K.H. 1991. TRACR3D: A Model of Flow and Transport in Porous Media: Model Description and User’s Manual. LA-11798-M. Los Alamos, New Mexico: Los Alamos National Laboratory. TIC: 201398. Wan, J. and Tokunaga, T.K. 1997. “Film Straining of Colloids in Unsaturated Porous Media: Conceptual Model and Experimental Testing.” Environmental Science and Technology, 31 (8), 2413–2420. Washington, D.C.: American Chemical Society. TIC: 234804. Wan, J. and Wilson, J.L. 1994a. “Colloid Transport in Unsaturated Porous Media.” Water Resources Research, 30 (4), 857–864. Washington, D.C.: American Geophysical Union. TIC: 222359. ANL-NBS-HS-000028, Rev 00 49 April 2000 Wan, J. and Wilson, J.L. 1994b. “Visualization of the Role of the Gas-Water Interface on the Fate and Transport of Colloids in Porous Media.” Water Resources Research, 30 (1), 11–23. Washington, D.C.: American Geophysical Union. TIC: 246300. Wemheuer, R.F. 1999. “First Issue of FY00 NEPO QAP-2-0 Activity Evaluations.” Interoffice correspondence from R.F. Wemheuer (CRWMS M&O) to R.A. Morgan (CRWMS M&O), October 1, 1999. LV.NEPO.RTPS.TAG.10/99-155, with attachments, Activity Evaluations. ACC: MOL.19991028.0162. Yao, K.-M.; Habibian, M.T.; and O’Melia, C.R. 1971. “Water and Waste Water Filtration: Concepts and Applications.” Environmental Science and Technology, 5 (11), 1105–1112. Washington, D.C.: American Chemical Society. TIC: 239214. Zyvoloski, G.A.; Robinson, B.A.; Dash, Z.V.; and Trease, L.L. 1997. Summary of Models and Methods for the FEHM Application—A Finite-Element Heat- and Mass-Transfer Code. LA- 13307-MS. Los Alamos, New Mexico: Los Alamos National Laboratory. TIC: 235587. 8.2 CODES, STANDARDS, REGULATIONS, AND PROCEDURES AP-3.10Q, Rev 2, ICN 0. Analyses and Models. Washington, D.C.: U.S. Department of Energy, Office of Civilian Radioactive Waste Management. ACC: MOL.200000217.0246. AP-SI.2Q, Rev. 2, ICN 4. Software Management. OCRWM Procedure. Washington, D.C.: U.S. Department of Energy, Office of Civilian Radioactive Waste Management. ACC: MOL.200000223.0508. DOE (U.S. Department of Energy) 2000. Quality Assurance Requirements and Description. DOE/RW-0333P, Rev. 9. Washington, D.C.: U.S. Department of Energy, Office of Civilian Radioactive Waste Management. ACC: MOL.19991028.0012. QAP-2-0, Rev. 5. Conduct of Activities. Las Vegas, Nevada: CRWMS M&O. ACC: MOL.19980826.0209. QAP-2-3, Rev. 10. Classification of Permanent Items, Revision 10 (C). Las Vegas, Nevada: CRWMS M&O. ACC: MOL.19990316.006. 8.3 SOFTWARE Los Alamos National Laboratory. FEHM V2.00. 1999. V2.00. SUN Ultra Sparc. 10031-2.00- 00. Los Alamos National Laboratory. FEHM V2.10. 2000. V2.10. SUN Ultra Sparc. 10086-2.10- 00. Los Alamos National Laboratory. TRACRN V1.0. 1999. V2.10. SUN Ultra 2. 10106-1.0-00. ANL-NBS-HS-000028, Rev 00 50 April 2000 8.4 SOURCE DATA, LISTED BY DATA TRACKING NUMBER GS950608312231.008. Moisture Retention Data from Boreholes USW UZ-N27 and UE-25 UZ#16. Submittal date: 06/06/1995 GS980908312242.039. Unsaturated Water Retention Data for Lexan-Sealed Samples from USW SD-6 Measured Using a Centrifuge. Submittal date: 09/22/1998 LA0002MCG12213.001. Pore Size Distribution for TSw4, CH1v, and CH1z. Submittal date: 02/02/2000 LA0002MCG12213.002. Forward Colloid Removal Rates in Topopah Spring Welded, Calico Hills Vitric, and Calico Hills Zeolitic Tuffs. Submittal date: 02/18/2000 LA0002MCG12213.003. Colloid Removal Kd’s Based on Varying Detachment (Reverse) Rates in Topopah Spring Welded, Calico Hills Vitric, and Calico Hills Zeolitic Tuffs. Submittal date: 02/18/2000 LA0003NL831352.001. Experimental Data on Sorption and Desorption Amounts for Plutonium onto Clay Colloids. Submittal date: 03/16/2000 LA9912MCG12213.001. UZ Colloid Transport Model Input and Output Files. Submittal date: 02/24/2000 LA9912PR831231.006. Simulations of Microsphere Tailing in Bullfrog and Prow Pass Tests. Submittal date: 12/14/1999 LB970601233129.001. The Site-Scale Unsaturated Zone Model of Yucca Mountain, Nevada, for the Viability Assessment. Submittal date: 06/09/1997 LB990501233129.001. Fracture Properties for the UZ Model Grids and Uncalibrated Fracture and Matrix Properties for the UZ Model Layers for AMR U0090, “Analysis of Hydrologic Properties Data.” Submittal date: 08/25/1999 LL000122051021.116. Summary of Analyses of Glass Dissolution Filtrates. Submittal date: 01/27/2000 LB990801233129.003. TSPA Grid Flow Simulations for AMR U0050, “UZ Flow Models and Submodels.” Submittal date: 11/29/99 LL991109751021.094. Data Associated with the Detection and Measurement of Colloids in Scientific Notebook SN 1644. Submittal date: 01/10/2000 MO9807MWDRIP01.000. Chapter 6, TSPA-VA Technical Basis Document Repository Integration Program, Saturated Zone. Submittal date: 08/06/1998. Submit to RPC. ANL-NBS-HS-000028, Rev 00 51 April 2000 8.5 OUTPUT DATA LA9912MCG12213.001. UZ Colloid Transport Model Input and Output Files. Submittal date: 02/24/2000 ANL-NBS-HS-000028, Rev 00 I-1 April 2000 ATTACHMENT I. SAMPLE INPUT FILE FROM FEHM V2.1 This data file was used for clay Waste Form Colloid Simulation where Colloid size = 100 nm Alpha in filtration equation = 0.01 Fracture Filtration Kd = 5, Kf = 0.5 Initial run of discrete fracture model #************************************************************************75 cont avs 1000000 1.e30 liquid concentration formatted endavs flxo 11 52 1 2 1 53 52 104 103 155 154 206 205 1430 1429 3470 3469 5510 5509 7550 7549 8570 8569 # Zone 1 is matrix, zone 2 is fracture zone 1 -1. 6. 6. -1. 600. 600. 1160. 1160. 2 -1. .0001 .0001 -1. 600. 600. 1160. 1160. 3 .0001 6. 6. .0001 833.5 833.5 898. 898. 4 .0001 6. 6. .0001 600. 600. 833.4 833.4 perm -1 0 0 1.46e-17 1.46e-17 1.46e-17 -2 0 0 3.44e-9 3.44e-9 3.44e-9 -3 0 0 1.73e-12 1.73e-12 1.73e-12 -4 0 0 1.47e-17 1.47e-17 1.47e-17 #************************************************************************75 # TSW4 properties used for this problem rlp 3 0.18 1.00 0.008 1.524 2 0.181 3 0.01 1.0 0.78 1.969 20 0.011 3 0.04 1.00 0.597 1.233 2 0.041 ANL-NBS-HS-000028, Rev 00 I-2 April 2000 3 0.36 1.00 0.002 1.567 2 0.361 -1 0 0 1 -2 0 0 2 -3 0 0 3 -4 0 0 4 #************************************************************************75 rock -1 0 0 2000. 1.e20 0.089 tsw4 -2 0 0 2000. 1.e20 0.9999 tsw4 -3 0 0 2000. 1.e20 0.265 ch1v -4 0 0 2000. 1.e20 0.193 ch1z #************************************************************************75 #************************************************************************75 # Flux is 5 mm/y average over the entire reach of the # model domain, but all put into the fracture node flow 1 1 1 0.1 0.15 1.e-4 10609 10609 1 -7.922021e-7 .999 0. #************************************************************************75 text Time Parameters time 0.5 3.6525e6 1000 1 1995 5 3.6525e5 7.304e5 -2 1 1 100000. 1.826e6 -2 1 1 100000. 2.922e6 -2 1 1 100000. text Numerics ctrl -10 1.e-4 40 1 0 0 3 0 1.0 2.0 1. 10 2. 1.e-10 1.e20 1 1 iter 1.e-5 1.e-5 1.e-5 -1.e-6 1.2 0 0 0 0 14400. sol 1 -1 air -1 20.0 0.1 #************************************************************************75 node 1 2092 #************************************************************************75 itfc ANL-NBS-HS-000028, Rev 00 I-3 April 2000 1 1 2 1 0. 1 3 0. trac userc 0.0 1.0 1.e-6 1.0 3.6525e5 1e30 3.66e5 1e30 10 1.2 .01 500. 5 2 1 0 0. 0. 1. 1.00E-30 1. 1e-34 1e-34 1 0 0 1 1 0 0 1e-30 9334 9334 1 -9876 0.0 3.6525e6 0 1 0 0 1e-30 #************************************************************************ rxn ** NCPLX, NUMRXN 0, 4 ** Coupling of the aqueous components (dRi/dUj) 1 1 ** IDCPNT(IC),CPNTNAM(IC),IFXCONC(IC),CPNTPRT(IC) (NCPNT rows) 1 WFC[aq] 0 0 1e-9 ** IDCPLX(IX), CPLXNAM(IX),CPLXPRT(IX) (ID # and name of complex, NCPLX rows) ** IDIMM(IM), IMMNAM(IM),IMMPRT(IM)(ID # and name of immoblie spec, NIMM rows) 1 WFC[s] 0 ** IDVAP(IV), VAPNAM(IM), VAPPRT(IV) (ID # and name of vapor spec, NVAP rows) ** skip nodes? 0 ** RSDMAX 1.0e-10 ****** Chemical reaction information ******** ** LOGKEQ (=0 if stability constants are given as K, =1 if given as log(K)) ** CKEQ(IX) (Stability constants, NCPLX rows) ** STOIC(IX,IC) (Stoichiometric coeff: NCPLX rows, NCPNT columns) ** LINEAR KINETIC REACTION for TSW4 ** 1 ** Where does the reaction take place? ** -1 0 0 ** Aqueous Component/Complex #, Solid Component # 1 1 ** Distribution coeffienct kg water/ kg rock ** 1.09E+00 ** Mass transfer coefficient (1/hr) ** 4.36E-01 ** LINEAR KINETIC REACTION for fracture** 1 ** Where does the reaction take place? ** ANL-NBS-HS-000028, Rev 00 I-4 April 2000 -2 0 0 ** Aqueous Component/Complex #, Solid Component # 1 1 ** Distribution coeffienct kg water/ kg rock ** 2.50E-06 ** Mass transfer coefficient (1/hr) ** 1.00E-03 ** LINEAR KINETIC REACTION for CHV ** 1 ** Where does the reaction take place? ** -3 0 0 ** Aqueous Component/Complex #, Solid Component # 1 1 ** Distribution coeffienct kg water/ kg rock ** 9.82E-04 ** Mass transfer coefficient (1/hr) ** 3.93E-04 ** LINEAR KINETIC REACTION for CHZ ** 1 ** Where does the reaction take place? ** -4 0 0 ** Aqueous Component/Complex #, Solid Component # 1 1 ** Distribution coeffienct kg water/ kg rock ** 2.76E-02 ** Mass transfer coefficient (1/hr) ** 1.11E-02 stop This is the userc data file referenced in the main data file. 2 90 Average Pu(aq) colloid (mol/s) 1.27E-28 0.00 1.06E-26 0.00 1.42E-25 0.00 6.03E-25 0.00 1.61E-24 0.00 3.38E-24 0.00 6.12E-24 0.00 1.00E-23 0.00 1.54E-23 0.00 2.23E-23 0.00 3.11E-23 0.00 7.97E-23 0.00 1.23E-22 0.00 1.81E-22 0.00 2.54E-22 0.00 3.44E-22 0.00 4.52E-22 0.00 5.82E-22 0.00 7.34E-22 0.00 ANL-NBS-HS-000028, Rev 00 I-5 April 2000 9.11E-22 0.00 1.11E-21 0.00 1.34E-21 0.00 1.60E-21 0.00 1.89E-21 0.00 2.22E-21 0.00 2.57E-21 0.00 2.97E-21 0.00 3.40E-21 0.00 3.87E-21 0.00 4.38E-21 0.00 4.93E-21 0.00 5.15E-21 0.00 5.71E-21 0.00 6.31E-21 0.00 6.95E-21 0.00 7.63E-21 0.00 8.35E-21 0.00 9.11E-21 0.00 9.92E-21 0.00 1.08E-20 0.00 1.17E-20 0.00 8.86E-20 0.00 1.44E-19 0.00 2.18E-19 0.00 3.14E-19 0.00 4.32E-19 0.00 5.77E-19 0.00 7.51E-19 0.00 9.55E-19 0.00 1.19E-18 0.00 1.47E-18 0.00 1.78E-18 0.00 2.13E-18 0.00 2.53E-18 0.00 2.97E-18 0.00 3.46E-18 0.00 4.00E-18 0.00 4.59E-18 0.00 5.24E-18 0.00 5.94E-18 0.00 6.71E-18 0.00 7.54E-18 0.00 8.44E-18 0.00 9.40E-18 0.00 1.04E-17 0.00 1.15E-17 0.00 1.27E-17 0.00 1.40E-17 0.00 1.53E-17 0.00 1.68E-17 0.00 1.83E-17 0.00 2.00E-17 0.00 2.17E-17 0.00 2.35E-17 0.00 2.55E-17 0.00 2.76E-17 0.00 ANL-NBS-HS-000028, Rev 00 I-6 April 2000 2.98E-17 0.00 3.21E-17 0.00 3.46E-17 0.00 3.73E-17 0.00 4.01E-17 0.00 4.30E-17 0.00 4.62E-17 0.00 4.95E-17 0.00 5.30E-17 0.00 5.67E-17 0.00 6.06E-17 0.00 6.47E-17 0.00 6.90E-17 0.00 7.36E-17 0.00 time(Seconds) 3.4713E+10 3.7869E+10 4.1025E+10 4.4181E+10 4.7336E+10 5.0492E+10 5.3648E+10 5.6804E+10 5.9959E+10 6.3115E+10 6.6271E+10 6.9427E+10 7.2582E+10 7.5738E+10 7.8894E+10 8.2050E+10 8.5206E+10 8.8361E+10 9.1517E+10 9.4673E+10 9.7829E+10 1.0098E+11 1.0414E+11 1.0730E+11 1.1045E+11 1.1361E+11 1.1676E+11 1.1992E+11 1.2307E+11 1.2623E+11 1.2939E+11 1.3254E+11 1.3570E+11 1.3885E+11 1.4201E+11 1.4516E+11 1.4832E+11 1.5148E+11 1.5463E+11 1.5779E+11 1.6094E+11 1.6410E+11 ANL-NBS-HS-000028, Rev 00 I-7 April 2000 1.6726E+11 1.7041E+11 1.7357E+11 1.7672E+11 1.7988E+11 1.8303E+11 1.8619E+11 1.8935E+11 1.9250E+11 1.9566E+11 1.9881E+11 2.0197E+11 2.0512E+11 2.0828E+11 2.1144E+11 2.1459E+11 2.1775E+11 2.2090E+11 2.2406E+11 2.2721E+11 2.3037E+11 2.3353E+11 2.3668E+11 2.3984E+11 2.4299E+11 2.4615E+11 2.4931E+11 2.5246E+11 2.5562E+11 2.5877E+11 2.6193E+11 2.6508E+11 2.6824E+11 2.7140E+11 2.7455E+11 2.7771E+11 2.8086E+11 2.8402E+11 2.8717E+11 2.9033E+11 2.9349E+11 2.9664E+11 2.9980E+11 3.0295E+11 3.0611E+11 3.0926E+11 3.1242E+11 3.1558E+11