SOFTWARE RELEASE NOTICE

2. Project Title: Radionuclide Transport KTI Project No. 20.06002.01.141

4. Originator/Requestor: Roberto Pabalan Date: 05/27/04

6. Validation Status

Validated

0 Limited Validation

0 Not Validated Explain :

7. Persons Authorized Access II Name I Read Onlv/Read-Write I Addition/Chanfze/Delete R. Pabalan RO Add F.P. Bertetti RO Add E. Pearcy RO Add SOFTWARE SUMMARY FORM

01. Summary Date: 02. Summary prepared by Roberto Pabalan (210)-522-5304 03. Summary Action: NEW

04. Software Date: 05. Short Title: GWB 4.0.2

I 11 06. Software Title: Geochemist’s Workbench Version 4.0.2 I 07. Internal Software ID:

08. Software Type: 09. Processing Mode: 10. Application Area

0 Automated Data System Interactive a. General: ScientifidEngineering Auxiliary Analyses W Computer Program 0 Batch 0 Total System PA 0 Subsystem PA 0 Other 0 SubroutinelModule 0 Combination b. Specific:

11. Submitting Organization and Address: 12. Technical Contact(s) and Phone:

CNWRA/SwRI Roberto Pabalan (210) 522-5304 6220 Culebra Road San Antonio, TX 78228

~~ ~ 13. Software Application: Geochemical Modeling

14. Computer Platform 15. Computer Operating 16. Programming 17. Number of Source Pentium PC System: NT 4.0 Language@):N/A Program Statements: N/A

18. Computer Memory 19. Tape Drives: None 20. Disk Units: 20 Mb 21. Graphics: N/A Requirements: 32 Mb

22. Other Operational Requirements: None

23. Software Availability: 24. Documentation Availability: Available 0 Limited 0 In-House ONLY Available 0 Preliminary 0 In-House ONLY ll 25- 10KE DAT File 11KB DAT File 35943 DAT File 47KB DAT File 31KB DAT File thermo-phrqpitz. dat 40KB DAT File 146KB DAT File 2KE File U V

4Amethyst.in 1KB IN File 6/8/99 2:26 PM 3 CO 3-flood. in 1KB IN File 6/8/99 226 PM 4DiIute.in 2KB IN File 6/8/99 2:26 PM 3 Fe-tracein 1KB IN File 6/8/99 2:26 PM 3 Fortiesh 1KB IN File 6/8/99 2:26 PM Fountain.in 1KB IN File 6/8/99 2:26 PM Fresh.in 1KB IN File 6/8/99 2:26 PM HCl-iniect. in 1KB IN File 6/8/99 226 PM 4H veravik. in 1KB IN File 6/8/99 226 PM 9 Injectimin 1KB IN File 6/8/99 2:26 PM 4 KnQ uar tz. in 1KB IN File 6/8/99 2:26 PM KnS ilica. in 1KB IN File 6/8/99 226 PM lastlab. txt. doc 19KB Microsoft Word Doc... 611 9/99 2: 16 AM 4 Lyons.in 1KB IN File 6/8/99 2:26 PM 4 Lyons100. in 1KB IN File 6/1 8/99 12: 30 AM 4 Metab-in 1KB IN File 6/8/99 226 PM 4 Miller.in 1KB IN File 618199 2:26 PM 4 Morro.in 1KB IN File 618199 2:26 PM 4 NaOH-flood in 1KB IN Fib 618199 2:26 PM 3 NS-seawater.in 1KB IN File 6/8/99 226 PM 4Ostwald.in 1KB IN File 6/8/99 2:26 PM 3 Pecos-scatter. txt 2KB Text Document 6/8/99 2:26 PM a::: 4 React-output. txt 29KB Text Document 6/1 8/99 9:47 PM 3 React-plot.dat 45KB DAT File 6/1 8i99 9:47 PM 4 R eykjanes.in 1KB IN File 6/8/99 2:26 PM 4 Salinein 1KB IN File 6/8/99 2:26 PM 4 Si03-flmd.in 1KB IN File 6/8/99 226 PM 4 S ur f-acid. in 1KB IN File 618/99 2:26 PM 4 S urf-complex. in 1KB IN File 6/8/99 226 PM I CENTER FOR NUCLEAR WASTE REGULATORY ANALYSES QA VERIFICATION REPORT FOR +ACQUIRED SOFTWARE NOT TO BE MODIFIED C

Software Title/Name: &&w&!/ldmcbb~d. ( /y Version: * 0.21 Demonstration workstation: fc Operating System: u; dwJ et- User: '<' PkbA

INOTE: Acquired software may or may not meet all requirements and will be evaluated on a case-by-case basis. I Installation Testing [TOP-018, Section 5.61 Has installation testing been conducted for each intended computer platform and operating system? Yes: 0 NO:^ N/A:O Computer Platforms: P c Operating Systems:%;- 63 +*o*"L* V-B +SC& +s~~-cL;w~Fk-.Lyf+0- wb Location of Acceptance Test Results: - L, $/w(lt*% )lubb& Jb)s;u.L" Comm nts:. U*Nf&J)'j @4 2 h.3. /&4/&,kb )' /t/ll;k 27f'/,sj /U4I\#& 0 Software Output [TOP-018, Section 5.5.41

Is software designed so that individual runs are uniquely identified by date, time, name of software and version? / NO Yes:O No: JX N/A:O Date and Time Displayed: NameNersion Displayed: I Comments: I NOTE: Output identification content and format is typically taken as is. ~~ ~~ IMedium Documentation [TOP-018, Section 5.5.61

The physical labeling of software medium (tapes, disks, etc.) contains: Program Name, Module/Name/Title, Module Revision, File type (ASCII, OBJ, EXE), Recording Date, and Operating System(s)? Yes: 0' No:O N/A:O Comments: kSC

(04/01) Page 1 of 3 CENTER FOR NUCLEAR WASTE REGULATORY ANALYSES QA VERIFICATION REPORT FOR +ACQUIRED SOFTWARE NOT TO BE MODIFIED C

User Documentation [TOP-018, Section 5.5.71

Is there a Users' Manual for the software and is it up-to-date? Yes: Et' No: 0 N/A:O User's Manual Version and Date: $Q $/1A/ 2pp / Comments : -+- Are there basic instructions for the installation and use of the software? Yes: CJ NO: o N/A:O Location of Instructions: 5 &&&% Comments:

Configuration Control [TOP-018, Section 5.7,5.9.3]

Is the Software Summary Form (Form TOP-4-1) completed and signed? Yes: ca/ No:O N/A:O Date of Approval: 6kh@d

Is the list of files attached to the Software Summary Form complete and accurate? Yes: 0 No:O N/A: 0 Comments:

Is the source code available or, is the executable code availgble in the case of (acquired/cpmercial codes)? Yes: W No:O N/A: 0 Location of Source Code: Q/L hu Comments :

Have all the scriptlmake files and executable files been submitted to the Software Custodian? __c_c Only the executable files are being submitted. Yes:* No:O N/A: 0 Location of executable files: Qk-A*s Comments:

Software Release [TOP-018, Section 5.91

(04/01 ) Page2of 3 CENTER FOR NUCLEAR WASTE REGULATORY ANALYSES QA VERIFICATION REPORT FOR +ACQUIRED SOFTWARE NOT TO BE MODZFZED 4-

Upon acceptance of the software as verified above, has a Software Release Notice (SRN), Form TOP-6 been issued and does the version number of the software match the documentation? Yes: d N~:O N/A: o SRN Number: ? 1f Comments:

Software Validation [TOP-018, Section 5.101

Has a Software Validation Test Plan (SVTP) been prepared for the range of application of the software? Yes:La‘ No: 0 N/A:O

Version and Date of SVTP:’; ‘/3/!& Date Reviewed and Approved via QAP-002: ‘a4 3 Comments:

Has a Software Validation Test Report (SVTR) been prepared that documents the results of the validation cases, interpretation of the results, and determination if the software has been validated? Yes:GV NO: o N/A: o Version and Date of SVTR: 0,VZ7b 3 Date Reviewed and Approved via QAP-002: q2%!93 Comments .:

Additional Comments: 1

Software EvaluatorLJserDate

(04/01) Page 3 of 3 V

SOFTWARE VALIDATION TEST PLAN FOR GEOCHEMIST’S WORKBENCH, VERSION 4.0.2

Prepared for

US. Nuclear Regulatory Commission Contract NRC-02-02-012

Prepared by

Lauren Browning

Center for Nuclear Waste Regulatory Analyses San Antonio, Texas

Approved by:

English C. Pearcy Date Manager, Geohydrology and CONTENTS

Section Page

1 SCOPE OF THE VALIDATION ...... 1

2 REFERENCES ...... 1

3 ENVIRONMENT ...... 2 3.1 Software-Introduction ...... 2 3.2 Code Description ...... 3 3.2.1 Input ...... 3 3.2.2 Output ...... 5 3.3 Hardware Requirements and Installation ...... 5

PREREQUISITES ...... 6

ASSUMPTIONS AND CONSTRAINTS ...... 6

TESTCASES ...... 6 6.1 Rxn ...... 7 6.2 Act2 ...... 7 6.3 Tact ...... 8 6.4 React ...... 8

7 TESTINPUTS ...... 9

8 TEST PROCEDURES ...... 9

ii V

TABLES

Table Page

3-1 Basis Species in the Geochemists Workbench Version 4.0.2 Database ...... 4

iii 1 SCOPE OF THE VALIDATION

This document establishes the Software Validation Test Plan for validating the installation and functionality of software tools in the code Geochemist’s Workbench (Bethke, 2002). Geochemist’s Workbench was developed by Craig Bethke at the University of Illinois at Urbana-Champaign, and acquired by the Center for Nuclear Waste Regulatory Analyses (CNWRA) to provide technical assistance to the U.S. Nuclear Regulatory Commission (NRC) in its high-level waste program.

This Software Validation Test Plan applies to Geochemist’s Workbench Version 4.0.2, and is intended to validate the set of software tools used to evaluate chemical reactions, calculate phase stability and aqueous speciation, establish equilibrium conditions, and trace kinetic reaction pathways. The following software tools are used to perform these geochemical evaluat i o ns : e Rxn e Act2 e Tact e React

Each of these software programs includes a large number of user options, including options to format and constrain data, depict modeling results, or control numerical processing methods. These options increase the diversity of site-specific geochemical applications that can be evaluated by the code, but do not alter the general functionality of the individual software tools. This Software Validation Test Plan has been designed to provide evidence of correct and successful implementation of the underlying theory and algorithms as outlined in the Geochemist’s Workbench Version 4.0.2 User’s Manual and required by the TOP-01 8 (CNWRA, 2001).

2 REFERENCES

The following documents are referenced or used as the basis for this Software Validation Test Plan.

Bethke, C.M. “The Geochemist’s Workbench, Version 4.0: A User’s Guide to Rxn, Act:!, Tact, React, and Gtplot”. University of Illinois: Urbana-Champaign. 2002.

Bethke, C.M. Geochemical Reaction Modeling: Concepts and Applications. New York City, New York: Oxford University Press. 1996.

Bethke, C.M. “The Question of Uniqueness in Geochemical Modeling.” Geochimica et Cosmochimica Acta. Vol. 56. pp. 4,3154320. 1992.

CNWRA. “Technical Operating Procedure (TOP-1 8): Development and Control of Scientific and Engineering Software, Revision 8, Change 0 (October 5, 2001).” San Antonio, Texas: Center for Nuclear Waste Regulatory Analyses. 2001.

1 Drever, J. I. The Geochemistry of Natural Waters, 2nded. Prentice-Hall, Englewood Cliffs, New Jersey. 1988.

Garrels, R.M. and M.E. Thompson. “A Chemical Model for Sea Water at 25 “C and One Atmosphere Total .” American Journal of Science Vol. 260. pp. 57-66. 1962.

Holland, H.D. The Chemistry of the Atmosphere and Oceans.” New York City, New York: John Wiley and Sons. 1978.

Johnson, J.W., E.H. Oelkers, and H.C. Helgeson. “SUPCRT92: A Software Package for Calculating the Standard Molal Thermodynamic Properties of , Gases, Aqueous Species, and Reactions from 1 to 5,000 Bars and 0” to 1,000 “C.” Livermore, California: Lawrence Livermore National Laboratory. 1991.

McDuff, R.E. and F.M.M. Morel. “The Geochemical Control of Seawater (Sillen revisited).” Environmental Science and Technology. Vol. 14. pp.l, 182-1,186. 1980.

Oreskes, N.K., K. Shrader-Frechette, and K. Belitz. “Verification, Validation and Confirmation of Numerical Models in the Earth Sciences.” Science. Vol. 263. pp. 641-646. 1994.

Parkhurst, D.L. and C.A.J. Appelo. “User’s Guide to PHREEQC (Version 2)-A Computer Program for Speciation, Batch-Reaction, One-Dimensional Transport, And Inverse Geochemical Calculations.” Water-Resources Investigations Report 99-4259. Denver, Colorado: U.S. Geological Survey. 1999.

Richardson, S.M. and H.Y. McSween, Jr. Geochemistry: Pathways and Processes. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. 1989.

Rimstidt J.D. and H.L. Barnes. “The Kinetics of Silica-Water Reactions.” Geochimica et Cosmochimica Acta Vol 44. pp. 1,683-1,700. 1980.

Wolery, T.J. “EQ3/6, A Computer Program for Geochemical Aqueous Speciation-Solubility Calculations: Theoretical Manual, User’s Guide, and Related Documentation (Version 7.0).” UCRL-MA-11066 Pt II I. Livermore, California: Lawrence Livermore National Laboratory. 1992.

3 ENVIRONMENT

3.1 Software-Introduction

Geochemist’s Workbench Version 4.0.2 (Bethke, 2002) is a package of interactive software programs that run on IBM-compatible personal computers. The main programs in Geochemist’s Workbench are named Rxn, Act2, React, and Tact. These programs manipulate chemical reactions (e.g., Rxn), evaluate phase stability (e.g., Act2), calculate aqueous speciation, equilibrium conditions, and reaction pathways (e.g., React), and evaluate -activity relationships (e.g., Tact). An additional software tool, called Gtplot, is also included in the Geochemist’s Workbench Version 4.0.2 package to facilitate processing and visualization of the modeling results. This Geochemist’s Workbench suite of programs was developed to support educational, scientific, and practical applications of geochemical modeling. Development of

2 Geochemist’s Workbench began in the mid-l980s, and was based on a simpler batch mode code named GT, which had been developed earlier by Bethke.

The program Rxn balances chemical reactions between dissolved species, minerals, and gases. The program Act2 calculates and plots information on stabilities and predominant aqueous species for a given combination of dissolved components and environmental conditions. Tact is a program that determines species activities and gas at different to show the effect of temperature on mineral stability and the predominance of aqueous species in specific chemical systems. The program React is the geochemical modeling program in the Geochemist’s Workbench software package. React performs aqueous speciation calculations and calculates gas fugacities and fluid saturation states with respect to various minerals. This program allows the user to determine the equilibrium state of a chemical system or the influence of kinetic dissolution/precipitation reactions on species at various times along a reaction pathway.

3.2 Code Description

The following description of Geochemist’s Workbench is based mainly on the user’s manual provided with the Geochemist’s Workbench Version 4.0.2 software (Bethke, 2002). Additional information about the code was obtained from the associated textbook on geochemical reaction modeling (BethkeJ996).

3.2.1 Input

Each of the four geochemical software programs in Geochemist’s Workbench (Le., Rxn, Act2, Tact, and React) are designed for interactive, rather than batch mode, use. This approach allows the user to configure each Geochemist’s Workbench program interactively with a series of commands that modify, run, or save input files. The most current chemical system defined by the user can be saved into a dataset described by commands in the format of the specific program. These datasets can be recalled and used as input scripts later.

All the programs use an independent set of aqueous species, known as the basis, to write reactions. Table 3-1 shows the default list of independent basis species used in Geochemist’s Workbench. Reactions described by basis species may include dissolved species only, or may also involve various combinations of interacting solid phases and gases. Basis species comprise a portion of all thermodynamic databases included with the Geochemist’s Workbench installation package, and can be modified interactively by the user.

The same basic process is used to configure calculations in the Rxn, Act2, Tact, and React programs. Basis species can be “swapped” with other aqueous species, minerals, or gases that are anticipated to play an important role in the modeled system. Swapped species may, for example, constrain conditions or describe minerals co-existing with the system. The next step in the process is to set the system conditions. The user must set a temperature and assign values of , activity, or fugacity to each default and/or swapped basis species in the model system. A variety of options are available to constrain the system. In React, for example, the user can apply a single temperature to the system or define a polythermal reaction pathway. Another option allows the user to linearly adjust gas fugacities or aqueous species activities within a user-defined range. There is a wide variety of different species concentration options, such as molalilty, molarity, or ppm.

3 I Table 3-1. Basis Species in the Geochemist’s Workbench Version 4.0.2 Database*pt I H20 I Eu+++ I Pb++ I Ag+ I F- I PU02++ I AI+++ I H+ I Ra++ I Am+++ I HP04- I Rb+ I Hg++ I Ru+++ I Au+ I I- I Se03- B(OH)3 Fe++ S io2(as) I Ba++ I K+ I S r++ I Br- I Li+ I S04- I Ca++ I Mg++ I Tc04- I HC03- I Mn++ I Th++++ I cs+ I N03- I Sn++++ I CI- I Na+ I U++++ I co++ I Ni++ I V+++ I Cr+++ I Np++++ I Zn++

*Johnson, J.W., E.H. Oelkers, and H.C. Helgeson. “SUPCRT92: A Software Package for Calculating the Standard Molal Thermodynamic Properties of Minerals, Gases, Aqueous Species, and Reactions from 1 to 5000 Bars and 0 to 1,000 “C.” Livermore, California: Lawrence Livermore National Laboratory. 1991. twolery, T.J. “EQ3/6, A Computer Program for Geochemical Aqueous Speciation-SolubilityCalculations: Theoretical Manual, User’s Guide, and Related Documentation (Version 7.0).” UCRL-MA-11066 Pt Ill. Livermore. California: Lawrence Livermore National Laboratory. 1992.

The Geochemist’s Workbench Version 4.0.2 package includes the following thermodynamic databases: thermo.dat (employs Debye-Huckel equation for calculating activity coefficients) thermo-pitzer.data (contains coefficients needed for Pitzer activity model) thermo-hmw.dat (contains coefficients needed for Harvie-Moller-Weare activity model) thermo-hdata.dat (contains coefficients needed to calculate activities for saline ) thermo.phrqpitz (contains coefficients needed for modified Harvie-Moller-Weare activity model),

4 V W which can be accessed by any of the Geochemist’s Workbench software programs. The thermodynamic datasets contain properties of aqueous species, minerals, and gases. The datasets also include equilibrium constants for reactions to form these species, and information required to calculate activity coefficients. When available, data are provided for the temperature range 0-300 “C: at one atm pressure below 100 “C, and along the vapor pressure curve for water at higher temperatures. If information on equilibrium constants is missing at specific temperatures, the omissions are generally denoted by a 500 in the datasets. All thermodynamic databases in the Geochemist’s Workbench software package can be modified by the user, and new thermodynamic databases can be created.

At this time, it is expected that Geochemist’s Workbench calculations performed in support of the NRC high-level waste program will primarily access the thermo.dat database. The thermo.dat database is the default database in the Geochemist’s Workbench software package, because it is the one most commonly employed by users. This dataset was compiled at Lawrence Livermore National Laboratory for use with the geochemical code EQ3/6 (Wolery, 1992), and is based mainly on the SUPCRT data compilation (Johnson, et al., 1991). Other databases included in the Geochemist’s Workbench installation package were largely designed to support calculations involving high ionic strength solutions, which are applicable to a smaller subset of geologic problems commonly addressed by thermodynamic and kinetic models. The thermo-hdata.dat and thermo-pitzer.dat datasets are included in the GWB 4.0 software package to maintain compatibility with earlier releases of the software, and are neither supported or recommenced for use with GWB 4.0.2 (Bethke, 2002). All validation exercises described in Section 6 of this report will therefore be performed by the thermo.dat database. If a decision is made to use thermo-pitzer.dat, thermo-hdata.dat, thermo-hmw.dat or thermo.phrqpitz databases in the future in support of the NRC high-level waste program, then additional validation exercises should be performed.

3.2.2 output

The results from the software packages React, Act2, and Tact are written to output files React. Output, Act2.0utput and Tact.Output, respectively. In addition, the programs Act2 and Tact display a diagram on the user’s workstation screen, which can be modified interactively with standard pull-down menus. The program Rxn is usually run interactively, with the output written on the workstation screen. It is possible, however, to divert Rxn’s output to a file, by using the “output” command in place of “go”.

3.3 Hardware Requirements and Installation

The Geochemist’s Workbench is designed for IBM-compatible personal computers. Successful installation requires the following system features:

Windows 98/ME/NT/2000/XP Penti u m-level processor 32 MB RAM 20 MB free disk space.

In order to install the Geochemist’s Workbench programs and supporting data, it is first necessary to log in as the system administrator. The “Geochemist’s Workbench” CD is then

5 inserted into the CDROM drive. After double-clicking on setup.exe, an interactive installation program provides the necessary instructions and guidance for completing the installation.

4 PREREQUISITES

Prerequisites for successful installation and application of the Geochemist’s Workbench software include an appropriate level of hardware capabilities and compliance with Geochemist’s Workbench installation process, as described in Section 3.3. Input files for all validation tests will be created.

5 ASSUMPTIONS AND CONSTRAINTS

It is difficult, or perhaps even impossible (Oreskes, et al., 1994), to produce a quantitative representation of a complex earth science system that is unique. Assumptions, constraints, and limitations are unavoidable in quantitative geochemical models. This is, in part, because there are different, and yet often equally valid, equations to describe geochemical processes, such as reaction rates or nonideal speciation behavior. Two different modelers may conceptualize the same chemical system in different ways, electing to include different minerals and aqueous species in his or her model or to evaluate the same data with somewhat different equations. In addition, thermodynamic data for a given phase often vary from dataset to dataset. Bethke (1992) demonstrates how modelers may obtain different results due to the presence of multiple mathematical roots to a single set of governing equations by employing slightly different sets of initial assumptions and constraints to a model. There are thus assumptions and constraints involved in the algorithms, the mathematical formulations, the data, and the conceptual models generated by Geochemist’s Workbench. Other geochemical codes, such as EQ3/6 (Wolery, 1992) or PHREEQC (Parkhurst and Appelo, 1999) have similar limitations. It is thus important to clarify what aspect of Geochemist’s Workbench Version 4.0.2 is being validated here. This software validation test plan is designed to provide evidence only that the Geochemist’s Workbench Version 4.0.2 software package successfully implements the underlying theory and algorithms described in the Geochemist’s Workbench user’s manual, as required by TOP-01 8 (CNWRA, 2001).

6 TESTCASES

In some cases, it is appropriate to validate software functionality by direct comparison between modeling results with analytical solutions (Richardson and McSween, 1989). Comparison with analytical results is a useful test of whether or not the software is reproducing anticipated results for a specific problem evaluated under relatively simple conditions. If all geochemical applications could be treated analytically, however, it would not be necessary to write codes to obtain solutions numerically. Geochemical systems generally include a large number of species, gases, and solid phases that interact in complex ways that are difficult to represent analytically. To test the ability of the underlying theory and algorithms in the Geochemist’s Workbench software package to represent accurately complex geochemical processes more representative of natural systems, validation of the Geochemist’s Workbench software package is mainly based on comparisons with published geochemical data from field studies, experiments, and modeling exercises. One test for the program React will be performed by comparison with an analytical calculation. Validation tests for the programs Rxn and React will be more detailed than those for Act2 and Tact, because it is anticipated that Rxn and React will

6 be the main Geochemist’s Workbench programs used in support of the NRC high-level waste program. The following tests should, however, provide a adequate test of the basic underlying theory and algorithms of the full set of geochemical programs included in the Geochemist’s Workbench package. Installation of the Geochemist’s Workbench software package will be considered to have been successful if the individual software packages are determined to function as expected.

Rxn

The software program Rxn balances chemical reactions among minerals, dissolved species, and gases, and provides thermodynamic data describing equilibrium conditions for that reaction at different temperatures. Through an iterative process of swapping basis species for other species, minerals, or gases, the user instructs the program to construct balanced reactions and present associated thermodynamic data for that reaction. To obtain the desired results, species stoichiometries may be manipulated by the program to calculate a transformation matrix that gives the desired balanced reaction and associated equilibrium constant. Alternatively, the RXN code may obtain this information by addition, subtraction, and multiplication of the balanced reactions described in the thermodynamic database until the desired reaction appears on the computer screen. Data used to balance reactions are thus sourced from the thermodynamic dataset that was selected by the user. A more detailed description of the underlying equations used to perform reaction balancing in Rxn are described in Chapter 9 of Bethke (1996).

Several tests will be performed to validate the primary capabilities of the Rxn program. These tests and anticipated results are: e accurate reproduction of Analcime dissolution reaction described in thermo.dat thermodynamic database e accurate reproduction of equilibrium constants for the Analcime dissolution reaction described in the thermo.dat thermodynamic database e construction of a reaction that is stoichiometrically and electrically balanced for the mineral pyrite using aqueous species other than the default basis species (e.g., swapped species).

Act2

In support of the NRC high-level waste program, it is expected that the main function of the software program Act2 will be to develop activity-activity plots To test the ability of Geochemist’s Workbench to calculate activity-activity diagrams, a test will be performed to determine if Act2 is capable of reproducing a series of activity-activity diagrams published on page 46-47 in Bethke (2002).

Successful validation of the Act2 program will require accurate reproduction (Le., less than 5 percent difference) of the following -pH diagrams for uranium drawn at 25 “C assuming a species activity of IO-’’.

7 e uranium speciation as a function of dissolved oxygen and pH e uranium speciation as a function of oxygen fugacity and pH e Eh versus pH e uranium speciation as a function of the activity ratio of Fe+++ to Fe++ and pH

Tact

The program Tact calculates and plots mineral stabilities and the predominance of specific species or gases at different temperatures. To demonstrate the basic capabilities of Tact, the following tests will be performed: e a diagram will be constructed depicting stability regions for different phosphate species as a function of temperature and pH, and compared with the published temperature-pH relations for phosphate species show on page 78 in Bethke (1998) e a diagram depicting the solubility of quartz as a function of temperature will be constructed and the results will be compared with solubility data for quartz described in the thermo.dat database at discrete temperatures.

6.4 React

React is the primary equilibrium and kinetic reaction pathway modeling program in the Geochemist’s Workbench Version 4.0.2 software package. The program calculates equilibrium aqueous speciation relationships, a fluid’s saturation state with respect to solid phases, and gas fucacities. The program is extremely flexible, allowing the user to select, among other options, isothermal or polythermal, open or closed, and reversible or irreversible reaction pathways. The underlying theory and algorithms used in React are described in Bethke (1996), predominantly in Chapters 5, 7, and 14. React calculations have two different components: characterization of the initial system and the system’s evolution after user-defined processes have altered it. Two tests of the React program will be performed to validate the underlying theory and algorithms for calculation of the equilibrium chemical system and a kinetic reaction pathway. The following tests were designed to evaluate the two fundamental components of React’s calculations, as described above: e activity coefficients and fluid saturation state with respect to various minerals in seawater will be compared to those obtained from chemical models of seawater described in Bethke (1996). The seawater model described in Bethke (1996) generally follows the approach first undertaken by Garrels and Thompson (1962), and is constrained by the major element composition of seawater as determined by chemical analysis and by modeling assumptions described in Drever (1988), Holland (1978), McDuff and Morel (1980) and others. Successful completion of these tests will validate Reacts calculation of the initial equilibrium component. e the reaction rate of quartz sand at 100 “C will be calculated according to the rate law described in Rimstidt and Barnes (1980), assuming a specific surface area of 1,000 cm2/g. Analytical calculations will be performed to determine the amount of quartz expected to dissolve under these conditions, and the results will be compared with the calculated volume losses for quartz.

8 7 TESTINPUTS

In all cases, test inputs will consist of a series of interactive commands described in the Geochemist’s Workbench Version 4.0.2 User’s Manual. Chemical conditions in the Geochemist’s Workbench Version 4.0.2 input files such as temperature, component concentration and mineral reaction rates will be set at values consistent with the problems described in Section 6 of this software validation test plan. In all test cases, thermodynamic data will be accessed from the thermo.dat database.

8 TEST PROCEDURES

The results of all Geochemist’s Workbench validation test cases will be compared to analytical solutions and published data and analyses, as described in Section 6. Results of Act2 and Tact test cases will be presented as plots to facilitate comparisons with published plots. Output from the interactive program Rxn will be copied directly from the screen printouts, and output files from React will be used to perform validation exercises. The test will be documented in Scientific Notebook 331.

9 SOFTWARE VALIDATION REPORT FOR GEOCHEMIST’S WORKBENCH, VERSION 4.0.2

Lauren Browning

Center for Nuclear Waste Regulatory Analyses San Antonio, Texas

English C. Pearcy, Manager Geohydrology and Geochemistry

Date 1 TABLE OF CONTENTS

Section Page

TABLE OF CONTENTS ...... ii FIGURES ...... iii TABLES ...... iv

1 SCOPE OF THE VALIDATION ...... 1

2 THERMODYNAMIC DATA USED IN TESTING ...... 1

3 TESTCASES ...... 2

3.1 Installation Check-Results ...... 4

3.2 Validation Check-Results ...... 4 3.2.1 Rxn ...... 5 3.2.2 Act2 ...... 8 3.2.3 Tact ...... 12 3.2.4 Reaction ...... 16 4 SUMMARY AND CONCLUSIONS ...... 20

5 REFERENCES ...... 20

APPENDIX

ii Figure Page

3-1 Uranium Speciation Diagrams reported in Bethke (1998), Assuming a Total Uranium Activity of IO-’’ ...... 9 3-2 Uranium Speciation Diagrams Calculated Using the Act2 Program, Assuming a Total Uranium Activity of IO-” ...... 11 3-3 pH Versus Temperature Diagram Generated by the Tact Program Showing Stability Regions for Different Phosphate Minerals...... 13 3-4 Quartz Solubility as a Function of Temperature ...... 15

iii TABLES

Table Page

3-1 Basis Species in the Geochemist’s Workbench. Version 4.0.2 Database ...... 3 3-2 Major Element Composition of Seawater ...... 16 3-3 Log Activity Coefficients (a) of the Most Abundant Species in Seawater ...... 17 3-4 Saturation Indices for Various Minerals in Seawater ...... 17

iv I SCOPE OF THE VALIDATION

The purpose of this document is to report the results of validation testing for the geochemical code Geochemist’s Workbench, Version 4.0.2 (Bethke, 2002) performed under the Geochemist’s Workbench Software Validation Test Plan (Browning, 2003). The Geochemist’s Workbench code was developed by Craig Bethke at the University of Illinois at Urbana-Champaign, and acquired by the Center for Nuclear Waste Regulatory Analyses (CNWRA) to provide technical assistance to the U.S. Nuclear Regulatory Commission (NRC) in its high-level waste program.

The following software programs are used to perform geochemical evaluations in the Geochemist’s Workbench, Version 4.0.2 software package:

* Rxn * Act2 * Tact * React

These programs include a large number of user options, including options to format and constrain data, depict modeling results, or control numerical processing methods. These options increase the range of site-specific geochemical applications that can be evaluated by the code, but do not modify the general functionality of the individual software tools. This software validation test has been designed to provide evidence of correct and successful implementation of the underlying theory and algorithms as outlined in the Geochemist’s Workbench Version 4.0.2 User’s Manual and required by the Technical Operating Procedure (TOP)-018 (CNWRA, 2001).

Geochemist’s Workbench, Version 4.0.2 includes a number of capabilities that are not planned for CNWRA applications related to the NRC high-level waste program, and were not tested at this time. These include:

Ion exchange Sorption onto mineral surfaces Microbial metabolism and growth Virial methods of calculating species activities

If a decision is made to use these code capabilities, or others for which the underlying theory and algorithms have not been tested, then the Geochemist’s Workbench, Version 4.0.2 software validation test plan and report will be modified, as necessary.

2 THERMODYNAMIC DATA USED IN TESTING

The results from geochemical codes are dependent on the type and quality of the data used in the simulation. These are typically contained in databases that are searched and read by the code based on the input provided by the user. Databases are considered separately from the input file that defines the geochemical problem, and are not modified within the context of a simulation that is running.

1 The Geochemist’s Workbench, Version 4.0.2 package includes the following thermodynamic databases:

thermo.dat (employs Debye-Huckel equation for calculating activity coefficients) thermo-pitzer.data (contains coefficients needed for Pitzer activity model) thermo-hmw.dat (contains coefficients needed for Harvie-Moller-Weare activity model) thermo-hdata.dat (contains coefficients needed to calculate activities for saline solutions) thermo.phrqpitz (contains coefficients needed for modified Harvie-Molter-Weare activity model), which can be accessed by any of the Geochemist’s Workbench software programs. These thermodynamic databases contain properties of aqueous species, minerals, and gases. The databases also include equilibrium constants for reactions to form these species, and information required to calculate activity coefficients. When available, data are provided for the temperature range 0-300 “C: at one atm pressure below 100 “C, and along the vapor pressure curve for water at higher temperatures. If information on equilibrium constants is not available at specific temperatures, the omissions are generally denoted by a 500 in the databases. All thermodynamic databases in the Geochemist’s Workbench software package can be modified by the user, and new thermodynamic databases can be created.

The most commonly used database in the Geochemist’s Workbench, Version 4.0.2 software package is the thermo.dat database, which calculates activity coefficients using the B dot equation. The thermo.dat database was compiled at Lawrence Livermore National Laboratory for use with the geochemical code EQ3/6 (Wolery, 1992), and is based mainly on the SUPCRT data compilation (Johnson, et al., 1991). The thermo.dat database is the default database for all Geochemist’s Workbench software programs. Other databases included in the Geochemist’s Workbench installation package were largely designed to support calculations involving high ionic strength solutions, which are applicable to a smaller subset of geologic problems that may be addressed by thermodynamic and kinetic models. The thermo-hdata.dat and thermo-pitzer.dat databases are included in the Geochemist’s Workbench, Version 4.0.2 software package to maintain compatibility with earlier releases of the software, and are neither supported or recommended for use with Geochemist’s Workbench, Version 4.0 (Bethke, 2002).

Validation exercises described in this report are therefore performed using the thermo.dat database. If a decision is made to use thermo-pitzer.dat, thermo-hdata.dat, thermo-hmw.dat or thermo.phrqpitz databases in support of the NRC high-level waste program, then the validation test plan (Browning, 2003) will be modified and additional validation exercises will be performed.

3 TESTCASES

The test cases to be used in the validation testing have been identified previously (Browning, 2003). The following sections are intended to report the results of the testing.

Validation tests are performed for each of the four geochemical software programs in Geochemist’s Workbench (i.e., Rxn, Act2, Tact, and React). These programs are designed for interactive, rather than batch mode, use. This approach allows the user to configure each Geochemist’s Workbench program interactively with a series of commands that modify, run, or

2 save input files. While input scripts for each of the programs can be written and saved, these do not include the history of interactive commands used to generate the files. The sequence of interactive commands for each validation test is thus documented below. Output files can be written for all programs, and are automatically generated with the use of the Act2, Tact, and React programs.

Individual input files used in the Geochemist’s Workbench, Version 4.0.2 software validation tests are included in Appendix A. It is not feasible to include the associated output files in an Appendix to this report, however, because the size of individual output files is typically large. The electronic files for all input and output files used in this validation report are thus provided as a CD attachment.

Unless otherwise specified, all geochemical calculations will be conducted using an independent set of aqueous species, known as the basis species. Table 3-1 shows the default list of independent basis species used in Geochemist’s Workbench. Reactions described by basis species may include dissolved species only, or may also involve various combinations of interacting solid phases and gases. Basis species comprise a portion of all thermodynamic databases included with the Geochemist’s Workbench installation package, and can be modified interactively by the user.

H2O Eu+++ Pb++ Ag+ F- PU02++ AI+++ H+ Ra++ Am+++ I HP04- I Rb+ AS(0 H)4 - I Hg++ I Ru+++ Au+ I I- I Se03- ~~ ~~~~~ B(OH)3 Fe++ S i02(aq ) Ba++ K+ Sr++

~~ ~~ Br- 1 Li+ I S04- Ca++ Mg++ Tc04- HC03- Mn++ Th++++ cs+ I N03- I Sn++++ CI- I Na+ I U++++ co++ I Ni++ I V+++ Cr+++ I Np++++ I Zn++

*Johnson, J.W., E.H. Oelkers, and H.C. Helgeson. “SUPCRT92: A Software Package for Calculating the Standard Molal Thermodynamic Properties of Minerals, Gases, Aqueous Species, and Reactions from 1 to 5,000 Bars and 0” to 1,000 “C.” Livermore, California: Lawrence Livermore National Laboratory. 1991. tWolery, T.J. “EQ3/6, A Computer Program for Geochemical Aqueous Speciation-Solubility Calculations: Theoretical Manual, User’s Guide, and Related Documentation (Version 7.0).”UCRL-MA-11066 Pt Ill. Livermore, California: Lawrence Livermore National Laboratory. 1992.

3 3.1 InstalI at ion Ch ec k-Res uI ts

The Geochemist’s Workbench is designed for IBM-compatible personal computers. Successful installation of the code on an IBM-compatible personal computer requires the following system attributes:

Windows 98/ME/NT/2000/XP Pentium-level processor 32 MB RAM 20 MB free disk space.

Installation of the Geochemist’s Workbench software package was performed by a computer information systems administrator in accordance with TOP-018 for the CNWRA (2001). An interactive installation program was provided with the software, which provides the necessary instructions, guidance, and tests for completing the installation. Installation of the Geochemist’s Workbench software package will be considered to have been successful if the individual software packages are determined to function as expected.

3.2 Va Ii da t ion C heck-Resu Its

TOP-01 8 for the CNWRA (2001) describes acceptable methods for validating commercially available software. These methods were reviewed by staff to determine the most appropriate tests for the Geochemist’s Workbench software package, and the review results were reported in Browning (2003).

There are inevitably assumptions and constraints involved in the algorithms, the mathematical formulations, the data, and conceptual models that are incorporated into simulations generated by most commonly used geochemical codes, such as Geochemist’s Workbench Version, 4.0.2 (Bethke, 2002), EQ3/6 (Wolery, 1992), and PHREEQC (Parkhurst and Appelo, 1999). It is therefore difficult, or perhaps even impossible (Oreskes, et al., 1994; Bethke, 1992), to produce a quantitative representation of a complex earth science system that is unique. Validation exercises under TOP-018 thus emphasize the importance of testing the ability of the code to apply underlying theories accurately, generally by comparison with independent applications of those same theories.

Validation of the Geochemist’s Workbench software package is based mainly on comparisons with published geochemical data from field studies, experiments, and modeling exercises, rather than by direct comparison between simplistic modeling results with analytical solutions (Richardson and McSween, 1989). As described in Browning (2003)’ the approach adopted in this report tests the ability of the underlying theory and algorithms in the Geochemist’s Workbench software package to represent accurately complex geochemical processes, such as those that might result from the emplacement of high level nuclear waste at Yucca Mountain, Nevada. However, results of Geochemist’s Workbench, Version 4.0.2 calculations are also compared with those from analytical expressions, when appropriate.

4 3.2.1 Rxn

Fundamental capabilities of the Rxn program are evaluated here by three different tests, as described in Browning (2003). The software program Rxn balances chemical reactions among minerals, dissolved species, and gases, and provides thermodynamic data describing equilibrium conditions for that reaction at different temperatures. Through an iterative process of swapping basis species for other species, minerals, or gases, the user instructs the program to construct balanced reactions and present associated thermodynamic data for that reaction. To obtain results in the desired forms, species stoichiometries may be manipulated by the program to calculate a transformation matrix that gives the desired balanced reaction and associated equilibrium constant. Alternatively, the Rxn code may obtain this information by addition, subtraction, and multiplication of balanced reactions described in the thermodynamic database until the desired reaction appears on the computer screen. A more detailed description of the underlying equations used to perform reaction balancing in Rxn are described in Chapter 9 of Bethke (1996).

Rxn Test 1: Analcime Dissolution Reaction

This test determines if the Rxn program can access the thermo.dat database and correctly read the formatted data in it to construct a reaction that is stoichiometrically and electrically balanced for the mineral analcime.

Geochemist’s Workbench Input (also see Rxn-analcimetestl-input.wpd): The input file shows the scripted geochemical constraints used by the Rxn program, and the commands listed below were used to develop that script.

Commands: > react Analcime > output > go

Geochemist’s Workbench Output (also see Rxn-analcimetestl-input.wpd):

Analcime + 4 H+ = 3 H20 + Na+ + AI+++ + 2 Si02(aq)

Geochemist’s Workbench thermodvnamic database: The following information about analcime is included in the thermo.dat database, and will be compared with output from the Rxn program:

Analcime type= zeolite formula= NaAISi206:H20 mole vol.= 97.100 cc mole wt.= 220.1539 g 5 species in reaction 3.000 H20 1.000 Na+ 1.OOO AI+++ 2.000 Si02(aq) -4.000 H+ Log K’s (T = 0, 25, 60, 100, 150, 200, 250, and 300 “C, from left to right): 8.6163 7.2800 5.6735 4.2314 2.8879 1.8680 0.9663 -0.1 538

5 By default, the thermo.dat database represents reactions in terms of mineral dissolution, rather than precipitation. A negative sign for species in the reactions indicates it’s position on the left side of the reaction.

Rxn validation test 1 results: Comparison between the basis species listed in Table 3-1 and the Geochemist’s Workbench output for Rxn test 1 shows that the Rxn program correctly accessed basis species from the thermo.dat database to write a stoichiometrically and electrically balanced reaction for the analcime. There are a total of 4 positive charges on both the left (e.g., 4 H”s) and right sides (e.g., 1 from Na’ and 3 from AI+++)of the reaction, demonstrating electrical balance. Stoichiometric balance is obtained by comparing the number of elemental components on each side of the reaction. There are 6 H, 1 Na, IAl, 2 Si, and 4 Si on each side of the reaction, demonstrating that the Rxn program has resourced the thermo.dat database correctI y to e nsure appropriate stoic hi omet ric ba Ia nce .

Rxn Test 2: Equilibrium Constants for Analcime

This test determines if the Rxn program can reproduce temperature dependent equilibrium constants (K) for the analcime dissolution reaction from the thermo.dat thermodynamic database (see Rxn test 1).

Geochemist’s Workbench Input (also see Rxn-analcime-test2-input.wpd):

Commands: > react Analcime > output > long >go

Geochemist’s Workbench Output (also see Rxn~analcime~test2~output.wpd):

Rxn validation test 2 results:

A comparison between the equilibrium constants (K) displayed in the Rxn test 2 output file and those provided in the thermodynamic database illustrates the successful reproduction of log K values for the analcime dissolution reaction. The expression labeled polynomial fit provides an approach for interpolating between log K values for different temperatures.

Rxn Test 3: Basis Species Swap for the Pyrite Dissolution Reaction

This test determines if the Rxn program can calculate a stoichiometrically and electrically balanced dissolution reaction for the mineral pyrite that is not defined in terms of the default basis species used in Geochemist’s Workbench, Version 4.0.2 (Table 3-1). The test will be performed by swapping HS- for S04- in the pyrite dissolution reaction, and then checking the output to determine whether or not the resultant reaction is stoichiometrically and electrically balanced. Geochemist's Workbench Input (also see Rxn~analcime~test3~input.wpd)

Commands: > react pyrite > swap HS- for S04- > 90

Geochemist's Workbench Output (also see Rxn analcime test3 output.wpd)

+ + Pyrite + H, 0 = Fe" 2 HS- .5 O,(aq)

Geochemist's Workbench thermodvnamic database: The following information about pyrite is included in the thermo.dat database, and will be compared with output from the Rxn program:

Pyrite type= sulfide formula= FeS2 mole voI.= 23.940 cc mole wt.= 119.9670 g 5 species in reaction -1.000 H20 1.000 Fe" 1.750 HS- 0.250 S04-- 0.250 H' Log K's (T = 0, 25, 60, 100, 150, 200, 250, and 300 'C, from left to right): -26.5279 -24.7025 -22.8643 -21.4327 -20.3362 - 19.8398 -19.91 68 -20.8284

As discussed in Rxn test 1, the thermo.dat describes mineral dissolution reactions in terms of the default basis species Table 3-1). The thermo.dat representation of the pyrite dissolution

reaction is thus:

+ + + Pyrite + H20 3.5 02(aq) = Fe" 2 S04-- 2 H'

Rxn validation test 3 results:

The Rxn program generated the following reaction for pyrite dissolution:

Pyrite + H20 = Fe" + 2 HS- + .5 02(aq)

Comparison between the pyrite dissolution reaction defined by the thermo.dat database and the pyrite dissolution reaction defined by Rxn test 3 demonstrates that HS- was successfully swapped into the reaction for the default basis species, S04-.

The test 3 reaction has a sum zero charge on both the left (e.g., no charged species) and right sides (e.g., 2 positive charges from Fe" and 2 negative charges from HS-) of the reaction, demonstrating electrical balance. Stoichiometric balance is obtained by comparing the number of elemental components on each side of the reaction. There are 2 S, 8 0, 2 H, and 1 Fe on each side of the reaction, demonstrating that the Rxn program has resourced the thermo.dat database correctly to ensure appropriate stoichiometric balance.

7 3.2.2 Act2

In support of the NRC high-level waste program, it is expected that the main function of the software program Act2 will be to develop activity-activity plots. To test the ability of Geochemist's Workbench to accurately calculate activity-activity diagrams, calculations are performed to determine if the Act2 program is capable of reproducing the series of activity- activity diagrams published on page 44 in Bethke (1998). The diagrams in Bethke (1998) are modified in Figure 3-1 to facilitate comparison with the diagrams calculated using Geochemist's Workbench, Version 4.0.2 (Figure 3-2).

Figure 3-1 shows uranium speciation diagrams for a system at 25 "C and a total uranium species activity of 10-lo, after Bethke (1998). The individual diagrams show:

* uranium speciation as a function of dissolved oxygen and pH (Figure 3-la) * uranium speciation as a function of oxygen fugacity and pH (Figure 3-1b) * Eh versus pH (Figure 3-IC) * oxygen fugacity as a function of the activity ratio of Fe"' to Fe" and pH (Figure 3-Id)

Geochemist's Workbench input (also see Act2a-input.wpd; Act2b-input.wpd, Act2c-input.wpd, Act2d-input.wpd, where the geochemical content of the input files Act2a-d corresponds with that of Figures3-I a-d, respectively). The input files show the scripted geochemical constraints used by the Act2 program, and the commands listed below were used to develop that script.

Commands: Act2a-i nput. wpd : > diagram U"" on 02(aq) vs pH > log a U"" = -10 > x from 0 to 14; y from -90 to 0 > go

Act2 b-i npu t .wpd : (in te ractive cont in ua t ion of previous run ) > swap 02(g) for 02 (as) > go Act2c-input.wpd: (interactive continuation of previous run) > swap e- for 02(aq) > diagram on Eh > y from -0.75 to 1.25 > go

Act2d-i nput. wpd : (interactive continuation of previous run) > swap Fe"'/Fe" for 02(aq) > y from -25 to 5 > go

8 0 11111111rlrlll

-20

n A U UJ + 40 v 0 0"c (D 8 -8 -60 -86

-80 -80

0 2 4 6 8101214 PH

5

0

+ -5 +' +' tf x -10 s Q -15

-20 t 1 -25-- I I I I I I I I I I1 1-4 I I 0246810121 0 2 4 6 8101214 PH PH (c) (d)

Figure 3-1. Uranium Speciation Diagrams Modified After Bethke (1998), Assuming a Total Uranium Activity of IO-''

9 -1 0

-20

-30 h 0-

% -40 -50 -0 -60

-70

-80

-90 0 2 4 6 8101214 PH PH

.5 + -5 - +a) v) LL -w +' >0 W +a) -10 c LL W CQ 0 m -15 -.-.. Uraninite

-20 +5 --. I25iC -_ I I I1 I I I I I I I r-,1 0 2 4 6 8101214 0 2 4 6 8101214 PH PH

Figure 3-2. Uranium Speciation Diagrams Calculated Using the Act2 Program, Assuming a Total Uranium Activity of IO-''

10 Geo c he mi st's Work bench out I> ut ( Nu mer i ca I va Iu es : Act 2a-d-o ut pu t .w pd ; Di a g rams: Act2a-d-output .eps)

Act2 validation test results:

Figure 3-2 shows the uranium activity diagrams generated by the Act2 program for comparison against diagrams from Bethke (1998) that employ the same geochemical constraints (Figure 3-1). Accurate reproduction of the uranium activity diagrams from Bethke (1998) would constitute adequate validation of the Act2 program, based on the expected usage of this program for the NRC client. The underlying Act2 program capabilities that are being tested in this report are the ability to perform geochemical speciation calculations and generate accurate activity-activity diagrams using this information. Numerical output used to generate the diagrams is provided by the Geochemist's Workbench validation tests, but is not provided in Bethke (1998). Requisite visual comparisons between Figures 3-1 and 3-2 are accomplished by examining the numerical intersection points of stability lines on X-Y axes, and the relative orientation of the various stability fields.

Comparison between Figures 3-1 and 3-2 demonstrates excellent agreement between the uranium activity diagrams calculated by the Geochemist's Workbench, Version 4.0.2 Act2 program and those calculated by Bethke (1998). Uranium speciation calculations performed by Act2 predicted that the dominant dissolved species for all 4 validation test cases are UO,", UO, UO,', U(OH),- OH' (UO,), (OH)-7, U"", UOH"', and that the stable solids are U,O,(c) and Uraninite. These results (Figure 3-2) are consistent with those determined by Bethke (1998), as shown in Figure 3-1. A comparison between Figures 3-1 and 3-2 also shows that the relative orientation of all solid and dissolved uranium species calculated by Geochemist's Workbench, Version 4.0.2 matches well with those calculated by Bethke (1 998). Geochemist's Workbench, Version 4.0.2 calculated pH dependent field boundaries between U02" and UOZOH', U020H' and (U0,),(OH)-7, and uraninite and U(OH)5-at values of 5, 8, and 1I, consistent with those calculated by Bethke (1998). Comparisons between other calculated stability fields are also favorable. For example, the Eh condition between the UO," and the U"" stability fields in Figures 3-ICand 3-2both occur at a value of 0.25 volts.

3.2.3 Tact

The program Tact calculates and plots mineral stabilities and the predominance of specific species or gases as a function of temperature. In support of the NRC high-level waste program, it is expected that the main function of the software program Tact2 will be to construct diagrams showing the stability fields for minerals and/or aqueous species at different temperatures. Two validation exercises were developed by Browning (2003) to test the basic capabilities of the Tact software, and the results are described below.

Tact Validation Test 1: Phosphate Speciation as a Function of pH and Temperature

This test determines if the Tact program can accurately calculate and plot the stability regions for different phosphate species on a diagram of pH versus temperature. The results of these GWB Version 4.0.2 calculations will be compared with the published temperature-pH relations for phosphate species shown on page 80 in Bethke (1998), and in Figure 3-3a.

11 Geochemist's Workbench input (Tactl-input.wpd): The input files show the scripted geochemical constraints used by the Tact program, and the commands listed below were used to develop that script.

Commands: > diagram HP04-- vs pH > a HP04-- = 10A-3 > x from 0 to 14 > go

Geochemist's Workbench output (Numerical values: Tactl-output.wpd; Diagram: Tact 1-output .e ps)

Tact validation test 1 results:

Figure 3-3 shows the results of the Tact calculations performed for this validation test. Comparison between Figures 3-3a and 3-3b illustrate excellent agreement between the two sets of tem pera t ure-de pen dent ca Icu lat ions. Both d iagra ms iI I ust rate that the predo m i na nt d issolved phosphate species are H3P20;, HP20;, H2P20;-,H2PO;-, and PO,-. The numerical value of intersection points for stability field lines in both diagrams match very well. For example, the H,PO,- and HPO,- fields in both diagrams are separated at a pH of approximately 7.2 at a temperature of 0 "C.

Tact Validation Test 2: Quartz Solubility as a Function of Temperature

This test determines if the Tact program can accurately calculate and plot the solubility of quartz as a function of temperature. The results of these Geochemist's Workbench, Version 4.0.2 calculations will be compared with solubility data for quartz described in the thermo.dat data base at d iscrete t em pe rat ures.

Geochemist's Workbench input (Tact2 input.wpd): The input files show the scripted geochemical constraints used by the Tact program, and the commands listed below were used to develop that script.

Commands: > diagram Si02(aq) vs Si02(aq) > x from -5 to -1 > go

Geochemist's Workbench thermodvnamic database: The following information about quartz is included in the thermo.dat database, and will be compared with output from the Tact program:

Quartz type= silica formula= Si02 mole voI.= 22.688 cc mole wt.= 60.0843 g 1 species in reaction 1.OOO Si02(aq) Log K's (T = 0, 25, 60, 100, 150, 200, 250, and 300 "C, from left to right):

12 Figure 3-3a. pH Versus Temperature Diagram Generated by the Tact Program Showing Stability Regions for Different Phosphate Minerals

300

250

200

n 2 150 c

100

50

0 0 2 4 6 8 TO 12 14 PH Figure 3-3b. pH Versus Temperature Diagram for Different Phosphate Minerals Modified After Bethke(l998)

13 -4.5021 -3.9993 -3.5026 -3.0951 -2.71 76 -2.4272 -2.1 943 -2.01 06

The thermo.dat representation of the quartz dissolution reaction is thus:

Quartz = Si02(aq).

Quartz solubility is thus measured directly by the value of K at a prescribed temperature, because K = a Si02(aq)/a Quartz, where the activity of quartz, a pure mineralogical phase, is defined as one.

Geochem ist’s Workbench out D ut (N u me ri ca I va Iues : Tact 2-ou t put .w pd ; Diag ram : Tact 1-0 ut pu t .e ps)

Tact validation test 2 results:

Figure 3-4 shows the results of the Geochemist’s Workbench, Version 4.0.2 calculations of quartz solubility as a function of temperature. The solubility of quartz increases with temperature. Log K values for quartz at discrete temperatures defined in the database agree with the log a Si02(aq) values for those same temperatures that are plotted in Figure 3-5. The interpolation technique used by the Geochemist’s Workbench, Version 4.0.2 program has defined a smooth and continuous solubility curve between the discrete Log K values identified in the database, as would be required by a user simulating quartz solubilities at alternative temperatures.

3.2.4 React

React is the primary equilibrium and kinetic reaction pathway modeling program in the Geochemist‘s Workbench, Version 4.0.2 software package. The program calculates equilibrium aqueous speciation relationships, a fluid’s saturation state with respect to solid phases, and gas fugacities, and can simulate the rate dependence of specified mineral dissolution/precipitation reactions. The underlying theory and algorithms used in React are described in Chapters 5, 7, and 14 of Bethke (1996). Two tests of the React program will be performed to validate the underlying theory and algorithms for calculation of the equilibrium chemical system and a kinetic reaction pathway.

React Test 1 : Calculation of Activity Coefficients and Fluid Saturation States

Garrells and Thompson (1 962) first calculated the distribution of species in surface seawater, and much work has since been done to refine our understanding of the ocean’s major element composition, species concentrations, and fluid saturation states for co-existing solids (Drever, 1988; Holland, 1978; McDuff and Morel, 1980).

Using the major element composition of seawater in Table 3-2 as input, the React program was used to calculate the activities of the predominant dissolved species and the corresponding fluid saturation states with respect to various solid precipitates. Tables 3-3 and Table 3-4 compare calculated species activities and fluid saturation states, respectively, with those published in Bethke (1996).

14 300

250

200

150 I-

100

50

0+5 +4.5 f4 *3.5 f3 f2.5 f2 f1.5 fl log a Siq(aq)

Figure 3-4. Quartz Solubility as a Function of Temperatu re

Chemical component Concentration (mg/kg) CI- 19,350 Na+ 10,760 S04- 2.71 0 Mg" 1,290 Ca" 41 1 K' 399 HC03- 142 Si02(aq) 0.5-1 0 Waq) 0.1-6

15 ~ Table 3-3. Log Activity Coefficients (a) of the Most Abundant Species in Seawater* Species I Log A* I Log A, Calculated For This Report

CI- I -0.462 I - 0.4619

Na' I -0.50 I - 0.4958

Mg" I - 1.90 I - 1.goo9 S04- I -2.57 I -2.5657 K' I -2.19 I -2.1881 MgCI' I -2.21 I -2.2125 ~ ~ ~ ~ NaS04- -2.3 -2.3676

Ca" -2.83 - 2.8334 MgS04 I -2.24 1 -2.2391

CaCI' I - 2.60 I - 2.60 *Bethke, C.M. Geochemical Reaction Modeling: Concepts and Applications. New York City, New York: Oxford University Press. 1996.

Table 3-4. Saturation Indices for Various Minerals in Seawater Mineral I SI (log Q/K)* SI (log WK); this report Antigorite I 44.16 44.16 Tremolite I 7.73 I 7.73 Talc I 6.68 I 6.68 Ch ryso t ile I 4.72 4.72 Sepiolite I 3.93 I 3.93 Anthop hyllite I 3.48 3.48 Dolomite I 3.46 3.46 Dolomite-ord I 3.46 I 3.46 Magnesite I 1.02 1.02 Calcite I 0.81 I 0.81 ~ ~ *Bethke, C.M. Geochemical Reaction Modeling: Concepts and Applications. New York City, New York: Oxford University Press. 1996.

16 Geochemist's Workbench input (React1 input.wDd): The input files show the scripted geochemical constraints used by the Tact program, and the commands listed below were used to develop that script.

Commands: > swap CO2(g) for H+ >swap 02(g) for 02(aq) >log f C02(g) = -3.5 >f 02(g) = 0.2 >TDS = 35080 %I- = 19350 mg/kg %a++ = 41 1 mglkg >Mg" = 1290 mg/kg >Na+ = 10760 mg/kg >K+ = 399 mglkg >S04-- = 2710 mg/kg >HC03- = 142 mg/kg >Si02(aq) = 6 mg/kg >print species = long > go

Geochemist's Workbench outDut (Numerical values only: Reactl-output.wpd). Some values extracted from the Reactl-output.wpd file are also shown in Tables 3-3 and 3-4.

React validation test 1 results:

The React program within the Geochemist's Workbench, Version 4.0.2 software package succeeded in reproducing the geochemical characteristics of seawater published in Bethke (1996). Tables 3-3 and 3-4 list activity coefficients and fluid saturated states for predominant phases in sea water, respectively. Comparison between the data in the second and third columns of Tables 3-3 and 3-4 shows that values calculated using the React program in Geochemist's Workbench, Version 4.0.2 are equivalent to those reproduced from Bethke (1996). In Table 3-3, the favorable comparison demonstrates that the React program calculated reasonable activity coefficients for different dissolved species in seawater, and then correctly predicted which of the dissolved species should be most abundant. These calculations validate the ability of the React program to correctly apply the B dot form of the Debye Huckel equation to complex compositions. Similarly, the excellent match between data in the second and third columns of Table 3-4 indicate that the React program is correctly reading log K values from the thermo.dat database, utilizing calculated activity coefficients, and calculating ion activity products for the various minerals based on these data.

React Test 2: Calculation of Reaction Rates in An Unequilibrated Solution

Many mineral phases and geochemical conditions associated with the emplacement of nuclear waste in a potential repository at Yucca Mountain, Nevada, will likely be controlled by kinetic processes. As such, validation of the React program's ability to correctly apply kinetic rate laws is important. In this test, the React program is used to calculate the reaction rate of quartz sand at 100 "C according to the rate law described in Rimstidt and Barnes (1980). A specific surface area of 1,000 cm2/g is assumed. Analytical calculations are then performed to determine the

17 amount of quartz expected to dissolve under these conditions, and the results are compared with the calculated volume losses for quartz.

Geochemist's Workbench input (React2 input.wpd1: The input files show the scripted geochemical constraints used by the Tact program, and the commands listed below were used to develop that script.

Commands: > time begin = 0 days, end = 5 days >T=100 > Si02(aq) = 1 umolal > react 5000 g Quartz > kinetic Quartz rate-con = 2.e-15 surface = 1000 > go

Geochemist's Workbench output (Numerical values only: React2-output.wpd).

React validation test 2 results: (analytical calculations: React2-calcs.xls); estimated number of moles of quartz dissolved in 5 days.

The React program within the Geochemist's Workbench, Version 4.0.2 software package was used to calculate the number of moles of quartz that would dissolve in 5 days under a specific set of geochemical conditions. The simulations depict the kinetically-controlleddissolution of 5,000 grams of quartz in deionized water spiked with negligible silica at 100 "C, and assume a fixed specific surface area of 1,000 cm2/g and an intrinsic rate constant for quartz of 2 x mol/cm2s. To validate these simulations, the total number of moles of dissolved quartz calculated by React is compared with the results of analytical calculations. The total number of moles of dissolved quartz is estimated analytically by employing the ion activity product values calculated by React and the following rate law for quartz dissolution given by Rimstidt and Barnes (1980):

where rqtzis the rate of quartz dissolution in mol/s, A, is the surface area (cm2), K' is the intrinsic rate constant for quartz, QIK is the saturation state for quartz. This is a simplified form of the rate law used in the React program of the Geochemist's Workbench, Version 4.0.2 software package.

Results of the React simulations indicate that the dissolution rate of quartz slows over time as the reaction approaches equilibrium. The solution becomes saturated with quartz after about 3.5 days, at which time the dissolution rate for quartz becomes negligibly slow. At all times, the saturation state for quartz at 100 "C is defined by the ion activity product for quartz, Q, divided by the equilibrium constant for quartz dissolution at I00 "C (e.g., log K = -3.0951, or 8.03 x that is recorded in the thermo.dat database (also, see section above on Tact Validation Test 2). React calculations indicate that a total of 0.0007985 moles of quartz have been dissolved after 5 days.

18 To calculate the total number of moles of quartz dissolved over 5 days analytically using Eq 3-1, continuous changes in the average saturation states for quartz calculated by the React program are approximated as 5 constant saturation states corresponding to each of the simulated days of dissolution. As shown in the React2-calcs.xls file, ion activity values calculated by the React program for 0.5, 1.5, 2.5, 3.5, and 4.5 days were adopted as representative values for days 1 through 5. These saturation states, plus the geochemical input constraints used to define the React input files, are applied to the Rimstidt and Barnes (1980) rate law (Eq. 3-1). Specific surface area and initial mass of quartz in the system utilized by the React calculations are converted to surface area (cm2) in the following manner:

As = specific surface area of quartz * total number of grams quartz = = 1000 cm2/g* 5000 g = 5e6 cm2.

Analytical calculations are provided electronically in the file React2-calcs.xls. React calculated a total loss of 7.985 moles of quartz in 5 days, whereas a loss of 7.64 moles was estimated analytically. This represents less than a 5 percent difference between the total numbers of moles of dissolved quartz calculated analytically and by the React program. Given that the time dependent saturation states for quartz calculated by React were simplified as 5 constant values to support the analytical estimates, this difference is expected and insignificant. This validation test demonstrates that the React program is able to correctly apply kinetic rate laws.

4 SUMMARY AND CONCLUSIONS

Comparisons for a variety of geochemical problems show that Geochemist’s Workbench, Version 4.0.2 produces results that are consistent with those produced by other methods such as hand calculations and other computer codes. Validation tests performed for this report demonstrate that the Geochemist’s Workbench, Version 4.0.2 software package will correctly execute a variety of fundamental geochemical calculations, including the correct calculation of activity coefficients, mineral solubilities, activity-activity diagrams, temperature dependent aqueous speciation, and time-dependent mineral dissolution. A less than 5 percent difference between the results of analytical calculations and those performed by the Geochemist’s Workbench, Version 4.0.2 was judged to be satisfactory, because the differences resulted from known simplifications in the analytical treatment of the problem.

Overall, the agreement of trends and results between Geochemist’s Workbench, Version 4.0.2 and the various calculational methods used in the model validation exercise indicates that the code is correctly and consistently applying the underlying geochemical theorems and algorithms needed to evaluate the equilibrium and kinetic behavior of geochemical systems.

5 REFERENCES

Bethke, C.M. “The Geochemist’s Workbench, Version 4.0: A User’s Guide to Rxn, Act2, Tact, React, and Gtplot.” University of Illinois: Urbana-Champaign. 2002.

Bethke, C.M. Geochemical Reaction Modeling: Concepts and Applications. New York City, New York: Oxford University Press. 1996.

Bethke, C.M. “The Question of Uniqueness in Geochemical Modeling.” Geochimica et Cosmochimica Acta. Vol. 56. pp. 4,3154320. 1992.

19 Browning, L. “Software Validation Test Plan for Geochemist’s Workbench, Version 4.0.2.” San Antonio, Texas: CNWRA. 2003.

CNWRA. “Technical Operating Procedure (TOP-I 8): Development and Control of Scientific and Engineering Software, Revision 8, Change 0 (October 5, 2001).” San Antonio, Texas: CNWRA. 2001.

Drever, J. I. The Geochemistry of Natural Waters, 2nded. Prentice-Hall, Englewood Cliffs, New Jersey. 1988.

Garrels, R.M. and M.E. Thompson. “A Chemical Model for Sea Water at 25 “C and One Atmosphere Total Pressure.” American Journal of Science Vol. 260. pp. 57-66. 1962.

Holland, H.D. The Chemistry offhe Atmosphere and Oceans.” New York City, New York: John Wiley and Sons. 1978.

Johnson, J.W., E.H. Oelkers, and H.C. Helgeson. “SUPCRT92: A Software Package for Calculating the Standard Molal Thermodynamic Properties of Minerals, Gases, Aqueous Species, and Reactions from 1 to 5000 Bars and 0” to 1000 “C.” Livermore, California: Lawrence Livermore National Laboratory. 1991.

McDuff, R.E. and F.M.M. Morel. “The Geochemical Control of Seawater (Sillen revisited).” Environmental Science and Technology. Vol. 14. pp.1,182-1,186. 1980.

Oreskes, N.K., K. Shrader-Frechette, and K. Belitz. ‘Verification, Validation and Confirmation of Numerical Models in the Earth Sciences.” Science. Vol. 263. pp. 641-646. 1994.

Parkhurst, D.L. and C.A.J. Appelo. “User’s Guide to PHREEQC (Version 2)-A Computer Program for Speciation, Batch-reaction, One-Dimensional Transport, And Inverse Geochemical Calculations.” Water-Resources Investigations Report 99-4259. Denver, Colorado: U.S. Geological Survey. 1999.

Richardson, S.M. and H.Y. McSween, Jr. Geochemistry; Pathways and Processes. Englewood Cliffs, New Jersey: Prentice-Hall, Inc. 1989.

Rimstedt J.D. and H.L. Barnes. “The Kinetics of Silica-Water Reactions.” Geochimica et Cosmochimica Acta Vol44. pp.1,683-1,700. 1980.

Wolery, T. J. “EQ3/6, A Computer Program for Geochemical Aqueous Speciation-Solubility Calculations: Theoretical Manual, User’s Guide, and Related Documentation (Version 7.0).” UCRL-MA-11066 Pt Ill. Livermore, California: Lawrence Livermore National Laboratory. 1992.

20 V

APPENDIX A Rxn Test 1: # Rxn script, saved Wed Jut 16 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify react Analcime activity H20 = ? activity Na' = ? activity AI"' = ? activity Si02(aq) = ? activity H' = ?

Rxn Test 2: # Rxn script, saved Wed Jul 16 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify react Analcime activity H20 = ? activity Na' = ? activity AI"' = ? activity Si02(aq) = ? activity H' = ? long

Rxn Test 3: # Rxn script, saved Wed Jul 16 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify react Pyrite swap HS- for S04-- activity H20 = ? activity Fe" = ? activity HS- = ? activity H' = ? activity 02(aq) = ?

Act2: Figure ??a: # Act2 script, saved Thu Jul 17 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify diagram U"" on 02(aq) vs pH log activity main = -10 x-axis from 0 to 14 increment 1 y-axis from -90 to 0 increment 5 Figure ?b: # Act2 script, saved Thu Jul 17 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify swap 02(g) for 02(aq) diagram U"" on 02(g) vs pH log activity main = -10 x-axis from 0 to 14 increment 1 y-axis from -90 to 0 increment 5 Figure ?c:

A- 1 # Act2 script, saved Thu Jul 17 2003 by lbrowning data = "c:\p rogram f iles\gw b\gtda ta\t hermo .da t" verify swap e- for 02(aq) diagram U"" on Eh vs pH log activity main = -10 x-axis from 0 to 14 increment 1 y-axis from -.75 to 1.25 increment .25 Figure ?d: # Act2 script, saved Thu Jut 17 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify swap Fe"'/Fe" for 02(aq) diagram U"" on FeTFe" vs pH log activity main = -10 x-axis from 0 to 14 increment 1 y-axis from -25 to 5 increment 2.5

Tact test 1: Tact script, saved Thu Jul 17 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify diagram HP04-- vs pH log activity main = -3 x-axis from 0 to 14 increment 1 y-axis from 0 to 300 increment 25

Tact test 2: # Tact script, saved Thu Jul 17 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify diagram Si02(aq) vs Si02(aq) x-axis from -5 to -1 increment .25 y-axis from 0 to 300 increment 25

React test 1: # React script, saved Mon Jul 21 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify temperature = 25 swap C02(g) for H' swap 02(g) for 02(aq) 1 kg free H20 fugacity C02(g) = .000316227766 fugacity 02(g) = .2 balance on CI- total mg/kg CI- = 19350 total mg/kg Ca" = 41 1 total mg/kg Mg" = 1290 total mg/kg Na' = 10760 total mg/kg K' = 399 total mg/kg S04-- = 2710 total mg/kg HC03- = 142 total mg/kg Si02(aq) = 6

A-2 TDS = 35080 printout species = long

React test 2: # React script, saved Tue Jut 22 2003 by lbrowning data = "c:\program files\gwb\gtdata\thermo.dat" verify time start = 0 days, end = 5 days temperature = 100 1 kg free H20 total umolal Si02(aq) = 1 react 5000 gram of Quartz kinetic Quartz rate-con = 2e-15 surface = 1000

A-3