Pergamon Geochimica et Cosmochimica Acta,Vol. 59, No. 18, pp. 3853-3856, 1995 Copyright© 1995Elsevier ScienceLtd Printed in the USA.All rightsreserved 0016-7037/95 $9.50 + .00 0016-7037( 95 ) 00259-6 REPLY Reply to the Comment by R. M. Kettler, D. A. Palmer, and D. J. Wesolowski on "Predictions of diagenetic reactions in the presence of organic acids" WENDY J. HARRISON j and GEOFFREYD. THYNE 2 ~ Department of Geology and Geological Engineering, Colorado School of Mines, Golden, CO 80401, USA Department of Physics and Geology, California State University at Bakersfield, Bakersfield, CA 93311, USA (Received October 10, 1994; accepted in revised form June 4, 1995) 1. INTRODUCTION Kettler et al. (1995; hereinafter KPW95) object to the com- pilation of cation-organic acid anion (OAA) equilibrium con- stants made by Harrison and Thyne ( 1992; hereinafter HT92) and suggest that geochemical models constructed by HT92 "may generate misleading implications of the role of these compounds in geologic processes." The objections raised by KPW95 fall into three general categories: (1) Table 2 of HT92 contains errors; (2) Assumptions made by HT92 result in inadequate or inappropriate corrections to experimentally measured equilibrium constants and to enthalpies and entro- pies of reaction; and (3) HT92 made no evaluation of internal consistency. We address each of these three categories below. 2. ERRORS IN TABLE 2--HT92 Comments raised by KPW95 in this category reflect real errors which we correct in this reply. In HI92, we compiled ionic strength, measured values of log K, enthalpies, and en- tropies of reaction from well known, highly regarded publications, without prejudice. The three principal sources, noted as secondary sources by KPW95, are Martell and Sillen (1964), Martell and Smith (1977), and Christensen and Izatt (1983). We did not cross check the sources of the data with the primary sources, as KPW95 have subsequently done. We find the effort of KPW95 commendable and agree that in cer- tain instances errors are present in the secondary sources and these propagate into HT92--Tables 2 and 3. Specifically, sta- bility constants for AI succinate and AI malonate are indeed for monoprotonated ligands, not fully deprotonated ligands as reported in the cited sources. The formation reaction for A1 salicylate (comment P) was incorrectly read by HT92 from Christensen and Izatt (1983). Additionally, several errors were made by HT92 in translating ionic strengths, tempera- tures, AH~, and AS~ from the cited sources into HT92--Ta- ble 2 (Comments A, D, F-H, J-V: KPW95; all comments are listed in the footnote of Table 1 in KPW95). These cor- rections are noted and are presented in Table 1 of this reply. We also agree that we made a sign convention error for the pKa values in HT92 Tables 2 and 3, but this was not present in our compilation of the EQ3/6 database. Comments M and S require no reply. 3. ASSUMPTIONS MADE BY HT92 In this section we with deal specifically with comments B, C, E, I, and discussion of the similar slope method. Stability constants that fall into this category of criticism by KPW95 are not in error, rather the reported values reflect assumptions and judgements made by H'I92 which are thus open to alter- native interpretation. In our opinion, H192 gathered data without prejudice and listed assumptions very clearly. We did not correct log K values for infinite dilution when, in our judgement, this correction could not readily be done using available data (Comment I: KPW95). We agree with KPW95 that error results when the infinite dilution extrapolation is not performed; however, we assert that as much error would be introduced by making the infinite dilution correction without adequate supporting experimental data. We deliberately chose not to introduce additional uncertainty into our analysis. When infinite dilution corrections were possible, HT92 did these graphically, by plotting log K values vs. ionic strength and then using linear regression analysis of these data (Com- ment B: KPW95). One such example is shown in Fig. 1. In the case of these three datasets, zero ionic strength measure- ments were made and regression analysis is not necessary. Each of the three datasets in Fig. 1 shows a different func- tional relationship between I and log K despite the fact that there is a common ligand and geochemically similar cations. Figure 1 illustrates quite clearly the dilemma facing HT92 when making zero ionic strength corrections where few ex- perimental data were available: each dataset has a unique slope! Where two or more data points were present at non- zero ionic strengths, H'I92 used linear regression analysis of these data to derive a zero ionic strength stability constant. Where only one datum was available, HT92 made no correc- tion to zero ionic strength. Figure 1 shows that in the case of Sr oxalate the ionic strength dependence is small and for Ca oxalate the dependence is negligible at ionic strengths less than 0.2 molal. HT92 decided that when experimental mea- surements were made at nonzero ionic strengths and no sup- porting data were available for zero ionic strength corrections, that the higher ionic strength stability constant could be used uncorrected. Although KPW95 point out the errors in our ap- proach, they do not suggest an alternative correction method that would be feasible for the available data. A similar line of reasoning was used by HT92 when as- sessing the feasibility of ionic strength corrections to enthal- 3853