Reviews in Mineralogy & Geochemistry Vol. 70 pp. 1-46, 2009 Copyright © Mineralogical Society of America 1 Thermodynamic Databases for Water-Rock Interaction Eric H. Oelkers, Pascale Benezeth, and Gleb S. Pokrovski Geochimie et Biogeochime Experimentale LMTG, CNRS-UPS-OMP UMR5563 14 Avenue Edouard Belin 31400 Toulouse, FRANCE oelkers @ Imtg. obs-mip.fr INTRODUCTION The creation of thermodynamic databases may be one of the greatest advances in the field of geochemistry of the past century. These databases facilitate creation of phase diagrams de- scribing which mineral phases are stable as a function of temperature and pressure, enabling detailed interpretation of metamorphic systems (e.g., Essene 1982; Spear and Cheney 1989; Zaho et al. 2000). The versatility of these databases provide insight into the fate and conse- quences of subsurface storage of radioactive waste (e.g., van der Lee and De Windt 2001; Lichtner et al. 2004; Zhang et al. 2008), toxic waste (e.g., Glynn and Brown 1996; Steefel et al. 2005), and C0 2 (e.g., Knauss et al. 2005; Oelkers and Schott 2005; Oelkers and Cole 2008; Oelkers et al. 2008). The impressive utility of thermodynamic databases has lead to their incorporation into 'user-friendly' chemical speciation, reactive path, and reactive transport computer codes including EQ3 (Wolery 1983), PHREEQC (Parkhurst and Appelo 1999), and CHESS (van der Lee et al. 2002) allowing rapid calculation of mineral solubility and solute speciation in a variety of geochemical systems. A selected list of chemical speciation codes is provided in Table 1. These codes differ is ease of use, but all accurately solve for the equilibrium assemblages of minerals and aqueous species, and mineral solubilities within the limits of their thermodynamics databases. The quality of the results of each of these codes is directly related to the quality of these databases. Geochemical modeling codes such as those listed in Table 1 have advantages and disadvantages. On one hand these computer algorithms allow calculation or prediction of the equilibrium state and/or the evolution of geochemical systems as a function of reaction progress with the press of a few buttons. Such calculations appear to have an amazing accuracy; results of these computer codes are commonly reported to 4 or more significant digits. Therein lines the disadvantage of these computer codes as they give the appearance that the thermodynamics databases on which they are based are perfect and accurate beyond all imagination. This, however, is a misunderstanding. The raw data on which common thermodynamic databases are based are sparse; for many species or minerals few data exist, and in many other cases, no experimental data exist at all. In such cases the thermodynamic 'data' were 'created' using correlations. The intent of such 'guesses' was not malicious, but rather to provide reasonable approximations until sufficient new experimental data became available to replace these estimates. Unfortunately, the success and widespread use of chemical speciation, reactive path, and reactive transport computer codes has obscured the provisional nature of existing thermodynamic databases. We are left with limited computational models that relatively few are currently improving and updating. This current state of affairs is bewildering considering that predictions made using currently available chemical speciation, reactive path, and reactive 1529-6466/09/0070-0001 $05.00 DOI: 10.2138/rmg.2009.70.1 Brought to you by | Chalmers University of Technology Authenticated Download Date | 11/18/19 6:40 AM