Calculated solution-solid relations in the low temperature system Ca-Mg0-Fe0-C0â‚‚-Hâ‚‚0

Citation
Calculated solution-solid relations in the low temperature system Ca-Mg0-Fe0-C0â‚‚-Hâ‚‚0

Material Information

Title:
Calculated solution-solid relations in the low temperature system Ca-Mg0-Fe0-C0â‚‚-Hâ‚‚0
Creator:
Woods, Terri L.
Place of Publication:
Tampa, Florida
Publisher:
University of South Florida
Publication Date:
Language:
English
Physical Description:
xvii, 121 leaves : ill. ; 29 cm

Subjects

Subjects / Keywords:
Carbonate minerals -- Analysis -- Australia ( lcsh )
Geology, Stratigraphic -- Precambrian ( lcsh )
Dissertations, Academic -- Marine science -- Doctoral -- USF ( FTS )

Notes

General Note:
Thesis (Ph. D.)--University of South Florida, 1988. Bibliography: leaves 111-118.

Record Information

Source Institution:
University of South Florida
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
025128764 ( ALEPH )
19842700 ( OCLC )
F51-00173 ( USFLDC DOI )
f51.173 ( USFLDC Handle )

Postcard Information

Format:
Book

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

CALCULATED SOLUTION-SOLID RELATIONS IN THE LOW TEMPERATURE SYSTEM CaO-MgO-FeO-CO -H 0 2 2 by Terri L. Woods A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Marine Science in the University of South Florida August, 1988 Major Professor: Robert M. Garrels, Ph D

PAGE 2

Graduate Council University of South Florida Tampa, Florida CERTIFICATE OF APPROVAL Ph. D. Dissertation This is to certify that the Ph. D. Dissertation of Terri L. Woods with a major in the Department of Marine Science has been approved by the Examining Committee on June 23, 1988 as satisfactory for the dissertation requirement for the Ph D. degree. Examining Committee: Major Professor: Robert M. Garrels Hember:IRbbert H Bvrne] Jf

PAGE 3

Terri L Woods 1988 All Rights Reserved

PAGE 4

ACKNOWLEDGEMENTS It i s impossible to acknowledge adequately the contribution that Robert Garrels has made to this r esearch, my thinking processes, and my career. The six years during which I was associated with him a s a research assistant a nd student are the most valuable I have ever spent, and probably will ever spend. His conduct as a person and as a scientist serves as a model to whic h I will aspire for the remainder of my life. Some of the fun of being a geologist has gone ou t of this world with him. I would like to thank the members of my dissertation committee, Professors R. H. Byrne, J S. Compton, A. Rosenzwe ig, and R N Strom for their helpful s uggestions and discussions about all facets of this research a nd f o r their valuable criticisms of this manuscript. I am especially grateful for their extra efforts under the difficult circumstances of the last year. Dr. Roland Wollast improved the results of my research and this manuscript by contributing his time and his expertise with thermodynamics. I would a lso like to tha n k Cynthia Gar r els for facilitating communication between Bob, myself and my other committee members and for making various other arrangements without which this work could no t have been co m p leted. She was exceedingl y helpful to me during a time that was particularly difficult for her. Finally, I would like to thank m y husband for his seemingly endless ii

PAGE 5

patience and interest during the stressful and time-consuming process of aspiring to a Ph. D. degree. Financial support for my various research projects during the past six years has come from numerous sources, including the National Science Foundation, the Petroleum Research Fund, the Department of Marine Science at the University of South Florida, and the Graduate Research Council of the University of South Florida. iii

PAGE 6

TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES LIST OF ABBREVIATIONS ABSTRACT INTRODUCTION AVAILABLE INFORMATION ON CARBONATES IN THE SYSTEM Ca0-Mg0-Fe0 -C02 -H 2 0 Nature of the rhombohedral carbonates Available experimental work Basics of the binary phase relations Basics of the ternary phase relations Available compositional data DATA ON THE SYSTEM FROM THE MARRA MA}ffiA BANDED IRON-FORMATION General introduction to the banded iron-formations Description of the Marra Mamba Iron-formation DERIVATION OF THE MODEL Available thermodynamic data for the system Derivation of the low temperature phase relations Calculation of the Gibbs free energy of formation o f ankerite Calculation o f the aqueous solution--solidsolution relations based on the composition of the solid phase Ef fect of elevated temperature on the calculated relations PARAGENESIS AND DIAGENESIS OF THE CARBONATES OF THE l'lARRA MAMBA Paragenesis Siderite-magnesite Dolomite-ankerite i v vi viii xiii XV 1 4 4 10 10 12 20 27 27 33 40 40 45 45 60 75 82 8 2 8 2 84

PAGE 7

Calcite Conditions and timing of the diagenesis DISCUSSION CONCLUSIONS REFERENCES APPENDIX v 86 88 97 108 111 119

PAGE 8

LIST OF TABLES TABLE 1. Rhombohedral carbonates with the calcite and dolomite structures [Radii from Prewitt (1969)]. The reference value for the 0 ion is 140 A. 5 TABLE 2. Comparison of the degree of miscibility exhibited by a solid-solution at temperatures greater than 300C, the 2!fferences between the ionic radii of the M and N cations in the MC03 -Nco 3 solid-solution. From Reeder (1983). 8 TABLE 3. Comparison of the differences in cation radii between the carbonate pairs that form stable dolomite stuctures and those that do not. From Reeder (1983). 9 TABLE 4. Average compositional ranges of banded iron-formations (Klein, 1983) and averages for the upper and lower Marra Mamba (Klein and Gole, 1981). 28 TABLE 5. Key characteristics of the world's major banded iron-formations. 29 TABLE 6. The most common iron minerals in the four facies of banded iron-formations.(Compiled from various sources). 32 TABLE 7. Representative microprobe analyses of carbonates in the upper part of the Marra Mamba Iron-formation [From Klein and Gole (1981)]. 35 TABLE 8 Representative microprobe analyses of carbonates in the lower part of the Mar r a Mamba Iron-formation [After Klein and Gale, 1981]. 38 TABLE 9. Available thermodynamic properties of the Ca0 Mg0-Fe0co2-H20 system. 41 TABLE 10. Thermodynamic data used for the calculations in this work (From Robie et al. (1978) unless otherwise noted). TABLE 11. Values of pairs of the pair product. equilibrium constants for reactions between carbonate minerals. The first mineral of is the reacting species and the second is the In the reactions, ankerite and dolomite are vi 43

PAGE 9

always written as the half formula (e.g., Ca0 5Mg0 5 -CO ) The following abbreviations a r e used: ank= ankerite, dol=dolomite, mag=magnesite, sid=siderite, cal=calcite. 52 TABLE 12. Values of 6G0(ank) calculated from the tie-lines on the composittonal diagram of Figure 14. (These values are for the half formula. ) 55 TABLE 13. Activity coefficients calculated for the 13 tie-lines of Figure 12 assuming a regular solid-solution model to represent the activity-composition relations of the carbonates. The resultant values for the free energy of formation of ankerite are compared to those calculated assuming an ideal solid-solution model (These values are for the half formula. ) 57 TABLE 14. Composition of the aqueous solution in equilibrium with various members of the ankerite-dolomite solidsolution and calcite. The difference between the numeric and algebraic ratios is explained in the text. 64 TABLE 15. Values of the ratios of the aqueous ions in solution in equilibrium with various pairs of ankerite-dolomite and siderite-magnesite, calculated from equation 21 (algebraic solution) and the numeric method outlined in the text (numeric solution). The ratio of Mg/Fe is calculated from the other ratios using the values for the numeric solution. TABLE 16. Comparison of the values of the dissociation constants for the Ca-Fe-Mg carbonates at 25C and 150C. TABLE 17. Comparison of the equilibrium constants for some carbonate exchange reactions at 25C and 150C. TABLE 18. Comparative mineral assemblages in diagenetic and low-grade metamorphic iron-formation. (Pressure and temperature conditions are estimates from French, 1973 ) TABLE 19. Temperatures of formation of the two-carbonate assemblages calculated from the geothermometer of Talant-69 79 80 89 sev and Sazonov (1979). 91 TABLE 20. Chemical analysis of the average basalt and ultramafic groundwater. From White, Hem, and Waring (1963). 104 TABLE 21. Proposed Proterozoic iron-basin feed-water from Garrels (1987) 105 vii

PAGE 10

LIST OF FIGURES FIGURE 1. Comparison between the ionic radius of the divalent cation in the calcite-structure carbonates with the a) unit cell volume; b) metal ion-oxygen bond length; c) octahedral distortion parameter, quadratic elonga-tion of the various carbonates. From Reeder (1983). 7 FIGURE 2 Phase diagram of the system, Caco 3 -MgC03 from 500 to 1200C. C = calcitic phase, D = dolomitic phase, M = magnesitic phase. Bulk compositions that fall in the regions labelled "C + D" or "D + M" crystallize as two phases; the compositions of the two phases in those regions are found by drawing a horizontal line through the point in question to the extremes of the region. Bulk compositions falling in the regions labelled "C" or "D11 crystallize a single phase with the composition indicated directly below it on the abscissa. From Goldsmith and Heard (1961, Fig. 4). 11 FIGURE 3. Experimental phase diagram of the system, Caco 3 -MnC03 from 325 to 625C. The region under the curve is that in which a two-carbonate assemblage was observed. Elsewhere homogeneous solid-solutions were obtained. From Goldsmith (1983). 13 FIGURE 4 Phase diagram of the system, Caco 3-Feco3 from 300 to 900C Bulk compositions falling under the curve crystallize as two-carbonate assemblages. Elsewhere a single homogeneous solid-solution was obtained. From (1962, Fig. 6). 14 FIGURE 5. Thermodynamically calculated phase diagram (with tielines) for the system, Ca0-Mg0-Fe0-C0 2H 20, at 250C The three phase field of calcite, dolomife-ankerite solid-solution, and magnesite-siderite solid-solution is shaded. The regions without tie-lines, along the base and at the very top, are one-phase fields. In the regions with tie-lines a two-phase assemblage crystallizes. From Anovitz and Essene (1987, Fig.l2). 16 FIGURE 6 Thermodynamically calculated phase diagram (with tielines) for the system, Ca0-Mg0-Fe0-C0 2-H 20, at 400C The three phase field of iron-bearing calcite, dolom-viii

PAGE 11

FIGURE 7. FIGURE 8. FIGURE 9. FIGURE 10. ite-ankerite solid-solution, and magnesite-siderite solid-solution is shaded. The regions without tielines, along the base and at the very top, are onephase fields. In the regions with tie-lines a twophase assemblage crystallizes. From Anovitz and Essene (1987, Fig. 13). Thermodynamically calculated phase diagram with (tielines for the system, Ca0-Mg0-Feo-co 2 -H 20, at 550C. The three phase field of iron-rich calcite, dolomiteankerite solid-solution and magnesite-siderite solidsolution is shaded. The white regions without tielines, along the base and at the very top, are onephase fields. In the regions with tie-lines a twophase assemblage crystallizes. From Anovitz and Essene (1987, Fig. 14). Carbonate compositions (mol%) in the system, MnC03 Mgco3-Feco3. Carbonates with more than 5 mol % Caco 3 are excluded. The complete solid-solution between MgC03 and Feco3 and between Feco3 and Mnco3 is shown by tfiese data. The carbonate compositions 1n the figure are taken from hydrothermal, sedimentary, and metamorphic occurrences described in the literature. From Essene (1983, Fig. 2). Carbonate compositions (mol%) in the system, CaMn(C03)2-CaMg(C03)2-CaFe(C03)2. Carbonates with mor e than 60 mol% or less than 40 mol % CaCO in solidsolution are excluded. The solid-solution between CaMg(co 3 ) 2 CaFe(C0 3 )2 and CaMn(CO ) 2 is extensive for magnesium-rich compositions. carbonate compositions in the figure are taken f rom hydrothermal, sedimentary and metamorphic occurrences described in the literature. From Essene (1983 Fig. 3). Triangular compositional diagram for natural carbonates in the system, Ca0-Mg0-Fe0-C02-H 20, at 250C. The field boundaries were determined graphically in agreement with the results for 400C and 550C (Figures 11 and 12). The three-phase triangular field is shaded. The carbonate analyses were renormalized to Ca + Mg + Fe = 1. The temperature estimate is believed to have a 50C error. From Anovitz and Essene (1987, Fig. 6). FIGURE 11. Triangular compositional diagram for natural carbonates in the system, Ca0 -Mg0-Feo-co2 -H 20, at 400C. The field boundaries were calculated from the phase equilibria model of Anovitz and Essene (1987). The three-phase triangular field is shaded. The carbonate analyses were renormalized to Ca + Mg + Fe = 1. The temperature estimate is believed to have a ix 17 18 21 22 24

PAGE 12

50C error. From Anovitz and Essene (1987, Fig. 7). 25 FIGURE 12. Triangular compositional diagram for natural carbonates in the system, Ca0-Mg0-Fe0-C0 2 -H 0 at 550C. The field boundaries were calculated from the phase equilibria model of Anovitz and Essene (1987). The three-phase triangular field i s stippled. The carbonate analyses were renormalized to Ca + Mg + Fe = 1. The temperature estimate is believed to have a 50C error. From Anovitz and Essene (1987, Fig. 9). 2 6 FIGURE 13. Carbonate compositions determined by Klein and Gole (198 1 ) for the upper portion of the Marra Mamba Ironformation. The apices of the triangle represent 100 mol% of the indicated component. Dolomite-ankerites ranging from 25-64 mol% of the CaFe(C0 3 ) 2 component, and containing minor amounts of excess Caco 3 and Mgco3-Feco3 are shown in the stippled region. These dolomite-ankerites were found to be in equilibrium with calcite compositions in the hatched region. 36 FIGURE 14. Carbonate compositions determined by Klein and Gole (1981) for the lower portion of the l1arra Mamba Ironformation. The apices of the triangle represent 100 mol% of the indicated component. Dolomite-ankerites ranging from 26-71 mol % of the CaFe(C0 3 ) 2 component, and containing minor amounts of excess Caco 3 and Mgco3-Feco3 are shown in the stippled region. These dolomite-ankerites were found to be in equilibrium with magnesite-siderite compositions ranging from 50 97 mol % of the FeC01 component (shown in the irregular stippled region at tfie base of the triangle). The tie-lines connect coexisting (i.e. physically touching) pairs of carbonates, and are separated into three approximately parallel sets; I, II and III. These correspond to the most gently dipping, those of intermediate slopes, and the most steeply dipping, respectively. 39 FIGURE 15. Diagram comparing tie-lines in the lower half of the compositional triangle derived from various sources. The light solid lines represent those calculated for 25 C from the model described in the text. The heavy solid lines represent those calculated for 250C from Anovitz and Essene (1987). The dashed lines are the assemblages (set III) observed by Klein and Gole (1981). 61 FIGURE 16. Aqueous solution--solid-solution phase diagram for the CaO-FeO-MgO-co 2 -H 2o system. The diagram is calculated from the compositional data of Klein and Gale (1981) for the upper portion of the Marra Mamba. The term indicates aqueous iron in its stable valence stafe (divalent in the ancient ocean and the proposed iron-basin feed-water, and trivalent in the X

PAGE 13

modern ocean). The stability fields of the various minerals and solid-solutions are indicated. The points on the curved phase boundary mark the indicated mole fraction of ankerite in the solid phase in equilibrium with an aqueous solution of the indicated composition. FIGURE 1 7 Calculated composition of the aqueous solution in equilibrium with the range of ankerite-dolomite compositions observed in the upper portion of the Marra Mamba. The term indicates aqueous iron in its stable valence (divalent in the ancient ocean and the proposed iron-basin feed-water, and trivalent in the modern ocean). The solid bar on the curves indicates the range of ankerite-dolomite compositions, observed by Klein and Gole (1981). The filled circle on the ordinate indicates the composition of present-day seawater, calculated using the iron concentrations and activity coefficient given by Byrne and Kester (1976). The double filled circles indicate the composition calculated for Proterozoic seawater. The activity ratio of magnesium to calcium is postulated to be the same for ancient and modern seawater. The asterisk indicates the composition of the proposed Proterozoic iron-basin feed-water from Garrels (1987). (The point for this value plots off 65 the central diagram at 0.2.) 67 FIGURE 18 Aqueous solution--solid-solution phase diagram for the CaO-FeO-MgO-co 2-H2 0 system, calculated from the compositional data of Klein and Gole (1981) for the upper and lower portions of the Marra Mamba. The term indicates aqueous iron in its stable valence sfate (divalent in the ancient ocean and the proposed iron-basin feed-water, and trivalent in the modern ocean). The stability fields of the various minerals and solid-solutions are indicated. The points on the curved phase boundary indicate the mole fractions of ankerite and siderite in the solid phase, in equilibrium with the aqueous solution of the indicated composition. The two dashed lines indicate the shift caused in the phase boundaries (for the range of mineral compositions observed in the Marra Mamba) by assuming a regular, instead of an ideal solid-solution model for the carbonates. 70 FIGURE 19. Diagram showing the sensitivity of the phase boundaries shown in Figure 8, to changes in the free energies of formation used in the calculations. The two curves indicated with dots were generated using a value for the free energy of ankerite of about 8 kJ/mol less stable than the value chosen in this paper. The phase boundaries indicated with xi

PAGE 14

dashed lines were calculated using only the thermodynamic data of Robie et al. (1978) and the value for the free energy of-ankerite calculated in this paper. See the text for a description of the values used to calculate this boundary. 72 FIGURE 20. Calculated composition of the aqueous solution in equilibrium with the range of ankerite-dolomite compositions, observed in the lower portion of the Marra Mamba. The term indicates aqueous iron in its stable valence (divalent in the ancient ocean and the proposed iron-basin feed-water, and trivalent in the modern ocean). The solid bar on the curves indicates the range of ankerite-dolomite compositions observed by Klein and Gole (1981). The filled circle on the ordinate indicates the composition of presentday seawater, calculated using the iron concentrations and activity coefficient given by Byrne and Kester (1976). The double filled circles indicate the composition calculated for Proterozoic seawater. The activity ratio of magnesium to calcium is postulated to be the same for ancient and modern seawater. The asterisk indicates the composition of the proposed Proterozoic iron-basin feed-water from Garrels (1987). 74 FIGURE 21. Diagram showing the shift in the aqueous solution-solid-solution boundary, calculated for the lower portion of the Marra Mamba, due to assuming a temperature of precipitation of the carbonates of 150C, instead of 25C. The position of the phase boundary with respect to the logarithm of the ratio of magnesium to calcium, decreases by approximately one log unit. 81 FIGURE 2 2 Relationship between the value of Anovitz and Essene (1987, defined in text) and temperature, for natural assemblages of coexisting dolomite-ankerites and magnesite-siderites. The points represent the values for these assemblages including the data from Klein and Gole (1981). The calculated values from the data of Anovitz and Essene (1987), Goldsmith et al. (1962), and Rosenberg (1967) are shown with the dashed and solid lines. 93 FIGURE 23. Diagram o f the proposed paragenetic and diagenetic history of the carbonate and silica phases, found in the Marra Mamba. The sequence of events, temperatures, and pressures are estimated from French (1973), Dimroth and Chauvel (1973), and Klein and Gole (1981). 99 xii

PAGE 15

A ank aq arg byo c oc cal car deg dol !:. vi r g /:,.Go f !:.GR grey !:.Ho f J kJ liq mag me tam. mol n ox QE R LIST OF ABBREVIATIONS angstrom ankerite aqueous argillite billions years old crystalline solid degrees Celsius calorie carbonate degree dolomite difference between the ionic radii of two cations gas standard molal Gibbs free energy of formation Gibbs free energy of reaction greywacke standard molal enthalpy of formation joule degrees Kelvin kilojoule liquid magnesite metamorphic mole number of crystallographic sites in a mineral oxide quadratic elongation Universal gas constant (8.31470 joules/deg/mol) standard molal entropy xiii

PAGE 16

standard molal entropy of reaction R sh shale sid siderite sil silicate sul sulfide T temperature molal volume change of a reaction V0 molar volume of a phase X mole fraction xiv

PAGE 17

CALCULATED SOLUTION-SOLID RELATIONS IN THE LOW TEMPERATURE SYSTEM by Terri L. Woods An Abstract Of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Marine Science in the University of South Florida August, 1988 Major Professor: Robert M. Garrels, Ph.D. XV

PAGE 18

The slowness with which some reactions reach equilibrium, or the formation of metastable minerals, requires the development of nonempirical models of phase equilibria. A thermodynamic model was derived relating the composition of coexisting carbonate minerals to the composition of the equilibrated aqueous solution. The model illustrates a theoretical framework to describe physical and chemical conditions of formation of systems that can not be handled experimentally or b y observation of natural occurrences. The model provides the only available description of phase relations of the Ca-Mg-Fe carbonates for 25C. The model was constructed with data from the Marra Mamba Banded Iron-formation (Hamersley Basin, Western Australia). Carbonates in these rocks show extensive Fe/Mg solid-solution permitting determination of the ionic activity ratios of iron, magnesium, and calcium in the equilibrated solution. The model requires knowledge o f the iron, magnesium, and calcium contents of the carbonates and thermodynamic data for the minerals and aqueous species. It was derived by algebraically combining expressions for the equilibrium constants of the following reactions: 1) dissolution of the carbonates, 2) Ca-Mg-Fe exchange between pairs of carbonates and an aqueous solution, and 3) equilibrium between the end-members of the solid-solutions. The methods used can be applied to other mineral assemblages in which extensive solid-solution is observed, and for which compositional and thermody namic data are available. Depending on the quality of this data, many ion ratios in the depositing solutions can be constrained.

PAGE 19

The diagenetic and metamorphic history of the carbonates of the Marra Mamba indicate that they have retained their original compositions and are, therefore, useful indicators of the composition of the depositing solution. Based on the calculated ion activity ratios, the solution that deposited the carbonates in portions of the Marra Mamba appears not to have been seawater, neither present-day nor that estimated for the Proterozoic. Abstract approved: Major Professor: Robert M. Garrels, Ph.D Professor, Department of Marine Science August, 1988 xvii

PAGE 20

1 INTRODUCTION minerals known to be stable thermodynamically at Earth' s surface, are extremely difficult to synthesize in the laboratory. At such low temperatures equilibrium is often attained very slowly. Also, intermediate, metastable phases may form during dissolution and precipitation. Dolomite, quartz, and iron oxides are good examples of stable minerals that are notoriously difficult to precipitate directly from laboratory solutions at ambient temperatures. Although, these minerals occur widely in sedimentary rocks, natural aqueous environments in which they are currently forming, are not common (if they are known at all). Therefore, empirical data are inadequate to describe the chemical and physical conditions of formation of such minerals. Thermodynamic models can often be developed for systems for which empirical data do not exist. When accurate thermodynamic and chemical data are available for the minerals of interest, models of the relationship between mineral and solution composition can often be derived. The concept of the equilibrium constant for a reaction and its relationship to the free energy of a reaction are central to the development of such models. The Ca-Fe-Mg carbonates are common minerals at Earth's surface, but data on their low-temperature phase relationships are rare. Information on their phase equilibria with an aqueous solution would provide the means to determine the major element chemistry of the fluids in which

PAGE 21

2 they equilibrated. The thermodynamic models that had been developed previously were based on experiments done at temperatures above 250C and at pressures higher than those at Earth's surface. Therefore, it was necessary to develop a thermodynamic framework to model the compositions of natural low-temperature assemblages. In order to develop the model it was necessary to find a natural system in which a wide range of carbonate compositions is observed, and for which extensive data exist. Detailed, well-documented compositional data for the Caco 3-MgC03 -FeC0 3 system are available from the Marra Mamba Banded Iron-formations of Western Australia (Klein and Gole, 1981). Therefore, these data were used to develop a thermodynamic method of calculating the chemistry of the aqueous solution in which the mineral assemblages equilibrated. The Marra Mamba exhibits a wide range in iron to magnesium ratios, extensive solid-solution among the carbonate minerals, and the highest iron content found in do lomite-ankerites. The major thrus t of the research was to develop a method to elucidate the chemistry of the depositing solution from the observed compositions of minerals showing extensive solid-solution. The method developed can be applied to any system for which compositional and thermodynamic data are available. Although it can provide useful additional information, it is not necessary to study the stratigraphic section, or t o determine the mineralogical compositions and petrographic relations firsthand. As long as accurate information is available in the literature, the chemistry of the minerals can be used to determine solution compositions. Familiarity with the rocks would make interpretation of the results more meaningful, but it was not the aim of this study to investigate the banded iron-formations extensively. However,

PAGE 22

in the process of developing this method, it was possible to shed some light on the origin of the Marra Mamba banded iron-formations. A brief introduction to the nature of the rhombohedral carbonates and the banded iron-formations will be given before the derivation of the thermodynamic model is described. 3

PAGE 23

AVAILABLE INFORMATION ON CARBONATES IN THE SYSTEM, CaC03-MgC03 -FeC0 3 Nature of the rhombohedral carbonates 4 In this paper the terms ankerite (CaFe(C0 3 ) 2), dolomite (CaMg(C03)2), siderite (Feco3), and magnesite (MgC03 ) will be used to refer to the pure iron and magnesium end-members of the dolomite-ankerite and magnesite-siderite solid-solutions. They do not refer to any intermediate compositions, as has been the case in many previous papers. As will be discussed later, the pure iron end-member, ankerite, has never been found naturally or produced in the laboratory. Ankerite is a hypothetical end-member useful in carrying out the thermodynamic modelling. The intermediate members of the series will be referred co as dolomite-ankerite and magnesite-siderite, or dolomitic, ankeritic, sideritic, and magnesitic carbonates. The rhombohedral carbonates are divided into two groups. One has the calcite structure (space group R3c) and includes calcite, siderite, magnesite, rhodochrosite, otavite, etc. The other has the dolomite structure (R3) and includes dolomite, ankerite, and kutnahorite (Table 1) The two structures are very similar with alternating layers of cations and carbonate ions along the c-axis, and the unit cell sizes are comparable. The major difference is that in the dolomite structure, the cation layers are alternately populated by calcium and the other major cation. The orientation of the planar carbonate group also changes.

PAGE 24

TABLE 1 Rhombohedra l carbonates with the calcite and dolomite structures2lRadii from Shannon and Prewitt (1969)]. The reference value for the 0 ion is 140 A. Name Chemical formula Radius of the divalent cation in 6-fold coordination (A) The R3c carbonates Calcite Caco 3 1.00 Magnesite MgC03 0. 72 Siderite FeC0 3 0 .78 Rhodochrosite MnC03 0 83 Otavite CdC03 0 95 Smithsonite ZnC03 0.75 Sphaerocobaltite CoC03 0 74 Gaspeite NiC0 3 0 69 The R3 carbonates with the dolomite structure Dolomite CaMg(C03 ) 2 Ankerite CaFe(C0 3 ) 2 Kutnahorite CaMn(C03 ) 2 Minrecordite CaZn(C0 3 ) 2 Synthetic Cd11g c co3 ) 2 5

PAGE 25

6 The minerals with the calcite structure form an isostructural group, whose differing characteristics are mainly a function of cationic radius (Reeder, 1983). Figures la and lb show the variation of unit cell volume and the length of the metal-oxygen bond with the radius of the cation. The calcium ion, which has the largest radius of any in the group (l.OOA), is near the upper limit of ion size for 6-fold coordina-tion. 2+ The faces of the Ca o 6 octahedra are considerably larger in calcite than in the other members of the group, except for otavite, Cdco 3 (cadmium is the next largest ion to calcium). In fact, cation size is one of the factors influencing the change to the orthorhombic, 2 + aragonite structure for cations larger than Ca (Reeder, 1983). To a large extent, the ideality of the solid-solutions that occur within this group, is determined by the differences in ionic radii of the cations. (The ideality of a solid-solution is measured by the extent to which the activity of the end-member component in the solid-solution is equal to its mole fraction.) This can be seen in Table 2 which compares MC03 -Nco 3 miscibility, at greater than 300C with the difference in the H 2 + and N 2 + radii ( !:::.VIr). Cation pairs showing a difference of less than O.llA in their ionic radii, demon-strate complete miscibility except for the NiC0 3 -Mgco 3 system. The reasons for the miscibility gap in this system are not well understood. Cation radius is by no means the only factor influencing either miscibility between the various end-member carbonates, or their lattice parameters. This can be seen from examination of Figure lc. The figure shows the relation between cation radius and a distortion parameter, quadratic elongation (QE). Quadratic elongation is a measure of the distortion of the M 2+-0 octahedra from an ideal octahedron. 2+ The Ca o 6

PAGE 26

1.0 (a) 2.4 (b) 0.8 2.3 -II( -0.6 2.2 0 I (.'f) 2 2 0.4 2 1 270 350 0.7 o.8 0 9 3 RM2+ Un i t Cell Vo l ume M FIGURE 1. Comparison between the ionic radius of the divalent cation in t h e calcite-structure carbonates with the a) unit cell volume; b) metal ion-oxygen bond length; and c) octahedral distortion parameter, quadratic elongation, of the various carbonates. From Reeder (1983). 7

PAGE 27

2+ octahedron shows the greatest distortion followed by Fe o 6 octahedron shows the least distortion, although the divalent cadmium ion is the largest in radius next to calcium. Obviously, there is no well-developed correlation between cation radius and octahedral distortion. TABLE 2. Comparison of the degree of miscibility exhibited at temperatures greater than 300C by a with the differences between the ionic radii of the M and N cations in the MC03 -Nco 3 solid-solution. From Reeder (1983). ComElete miscibility Limited miscibility 2+ 2+ M N 6vir (A) 2+ 2+ M N VI l:l r (A) Fe-Mg 0.06 Ca-Mg 0 28 Ca-Cd 0.05 Ni-Mg 0 03 Mg-Co 0.02 Ca-Fe 0 22 Fe-Mn 0.05 Ca-Mn 0 .17 Ng-Mn 0.11 Cd-Mg 0 .23 Ca-Co 0.26 Ca-Ni 0 31 Octahedral distortion plays a large part in the phase relations of the dolomite group, as explained by Reeder (1983). In dolomite the Cao 6 octahedron is less distorted than in calcite, and the Mgo6 octahedron is slightly less distorted than in magnesite. This decreased distortion may increase the stability of dolomite (Althoff, 1977) cao6 octahedra in ankeritic carbonates are noticeably more distorted than in dolomite, while the (Mg,Fe)o 6 octahedra are distorted similarly to dolomite. Apparently, increased substitution of Fe for Mg leads to a greater distortion of Cao 6 octahedra, than of the (Mg,Fe)o6 octahedra. This excess distortion may be responsible for the instability of CaFe(co3 ) 2 and its limited solubility in dolomite (Rosenberg and Foit, 1979). The pure iron end-member of the series is not found, but dolomite-ankerites, with up to about 70 mol% of the iron end-member, are known to occur.

PAGE 28

9 Strong site preferences caused by fundamental differences in cations, are the main reason for the existence of intermediate carbon-ates with t h e dolomite structure (Goldsmith, 1959). In most cases where the radii are similar, a solid-solution will form rather than an inter-mediate phase. Although ionic radius is a primary factor influencing the existence of these minerals, octahedral distortion must be consid-ered as well. Differences in ionic radii do not e xplain the instability of pairs such as Ca-Fe, Ca-Co, Ca-Ni, and Ca-Cu (Table 3). TABLE 3 Comparison of the differences in cation radii for the carbonate pairs that form stable dolomite structures with those that do not. From Reeder (1983). Stable pairs Unstable Eairs M2+_N2+ 6VI r (A) 2+ 2+ 6VI (A) M -N r Ca-Mg 0.28 Ca-Fe 0.22 Cd-Mg 0.23 Ca-Co 0.26 0.17 Ca-Ni 0.31 Ca-Zn 0.26 Ca-Cu 0.27 On the basis of the degree of octahedral distortion and estimated values for the free energies of formation of the transition metal dolomites, Rosenberg and Fait (1979) predicted the following relative stabilities 2 + 2+ for the pure, ordered Ca R (C03 ) 2 compounds; Mg>> > Zn > Fe > Co > Ni >>Cu. The first three compounds occur naturally (or have been synthesized in the laboratory) while the last four do not, thereby s upporting the concept that octahedral distortion may determine the stability of the transition metal dolomites (Rosenberg and Foit, 1979).

PAGE 29

10 Available experimental work Most of the data on phase relations in the Ca0 Mg0-Fe0 -C02 -H 2 0 system are for temperatures and pressures well above earth-surface conditions. They are derived from studies of both the binary and ternary phase relations. Data on the phase relations of other divalent metal carbonates are much more limited. Basics of the binary phase relations The most extensive work has been done on the Caco 3-MgC03 Caco 3 MnC03, Mgco3-Feco3 and Caco 3 -FeC0 3 joins, as these are the most commonly occurring natural systems. The earliest work by Harker and Tuttle (1955) on Caco 3-MgC03 was for the subsolidus relations between 500 and 900C at approximately 2 kilobars. Graf and Goldsmith (1955) conducted similar investigations at moderate temperatures and pressures. Other high-temperature work on this system was done by Goldsmith and Heard (1961), and Goldsmith and Graf (1958). Goldsmith and Heard (1961) produced the first diagram of the calcite-dolomite solvus at high temperature (Figure 2). More recent work by Fanelli et al. (1983), does not indicate significant changes in the position of these curves. The ultimate configuration of the Caco 3 MgC03 phase diagram depends on the high-temperature disordering conversion of dolomite (Goldsmith, 1983), which occurs at temperatures far above those with which the present study is concerned. The Caco 3 -Mnco 3 system has also been investigated at high temperatures and pressures (Goldsmith and Graf, 1957; de Capitani and Peters,

PAGE 30

0 0 Q) as '-Q) Q, E Q) f-0 11 c D+M C+O 20 40 60 80 100 Mole 'Wa MgC03 i n (Ca,Mg)C03 solid solutions FIGURE 2 Phase diagram of the system, CaC03-MgC03, from 500 to 1200C C =calcitic phase, D = dolomitic phase, M = magnesitic phase. Bulk compositions that fall in the regions labelled "C + D or "D + M crystallize as two phases; the compositions of the two phases in those regions are found by drawing a horizontal line through the point in question to the extremes of t h e region. Bulk compositions falling in the regions labelled "C" or "D" crystallize a single phase with the composition indicated directly below it on the abscissa. From Goldsmith and Heard (1961, Fig. 4).

PAGE 31

12 1981). The two sets of experimental data produce essentially the same solvus (Figure 3). There is no experimental evidence for a solvus in the calcium-rich side of the diagram, although some natural occurrences suggest its existence (Goldsmith, 1983). The Caco 3-Feco3 join (Figure 4) does not have a stable intermediate compound (i.e. ankerite). Instead it exhibits a single asymmetri-cal solvus at elevated temperatures [Goldsmith et al. (1962); Rosenberg (1963)]. Three other joins have been investigated that show a single solvus and incomplete solid-solution between the end-members [Caco3 NiC03, Caco 3 -coco3 and MgC03 -NiC0 3 ; Goldsmith and Northrop (1965)]. Goldsmith (1983) notes that all of the solvi considered are asymmetrical and that the greater amount of solid-solution is found on the limb representing the larger cation. This observation is not restricted to the carbonates. Complete miscibility is seen on the joins, MgC03-Feco3 (295-500C, Rosenberg, 1963), MnC03 -Feco3 (450C, Rosenberg, 1963), }fuC03 -Mgco 3 (450-500C, Goldsmith and Graf, 1960), Coco 3 -Mgco 3 (600C, Goldsmith and Northrop, 1965), and Caco 3-cdco3 (530C, Chang and Brice, 1971). Basics of the ternary phase relations The Caco 3-MgC03-Feco3 system is the most extensively investigated ternary of the carbonates and the only one to be studied experimentally by more than one laboratory (Goldsmith, 1983). Rosenberg (1967) studied the system between 350C and 550C at 2-3 kilobars, and Goldsmith et al. (1962) at 600 C to 800C and 15 kilobars. The two sets of data are not consistent (Reeder, 1983; Anovitz and Essene, 1987). The three-phase

PAGE 32

I 600 C ale ian Rhodochrosite a Solid Solution I I o sooManganoan Calcite o Solid Solutionl Q) .. .. (G .. Q) Q. E Q) 0 20 40 Two-phase l Carbonate Assemblage l 60 80 Mole CJ6 MnCO in (Ca,Mn)CO s o II d 3 s o I u t i o n s 3 100 13 FIGURE 3 Experimental phase diagram of the system, CaC03-MnC03 from 325 to 625C. The region under the curve is that in which a two-carbonate assemblage was observed. Elsewhere homogeneous solid-solutions were obtained. From Goldsmith (1983).

PAGE 33

(J 0 700 ... :1 ... 500 ... Ferroan Calcite Solid Solution + Calcian Siderite Solid Solution 14 __ _. __ _. __ 0 20 40 eo ao 100 Mole !. FeCO in (Ca,Fe)CO s o ll d 3 s o I u t i o n s 3 FIGURE 4 Phase diagram of the system, CaC03FeC0 3 from 300 to 900C. Bulk compositions fallin g under the curve crystallize as two-carbonate assemblages. Elsewhere a single h omogeneous solid-solution was obtained. From Goldsmith et al. (1962, Fig. 6). ---

PAGE 34

15 field for coexisting calcite-siderite-ankerite from Goldsmith al . (1962) is significantly wider than that suggested by Rosenberg. Also, neither of these experiments correctly predicts the maximum amount of CaFe(C0 3 ) 2 possible in natural dolomite-ankerites. Anovitz and Essene (1987) combined a careful evaluation of the earlier experimental work with data on the compositions of naturally occurring carbonates to derive the approximate ternary phase diagrams at 250, 400, 550, and 700C. They used a ternary subregular two-phase model that considers macroscopic thermodynamic properties without incorporating microstructural considerations. They determined that the compositional limit of Rosenberg (1967) at 500C represents more closely the compositional limit seen in natural assemblages, and that the tie-lines of Goldsmith al. (1962) more correctly predict the Mg/Fe partitioning observed. Figures 5 through 7 show the ternary diagrams of Anovitz and Essene (1987) for 250, 400, and 550C The apices of the triangular diagrams r epresent a mineral containing 100 mol % of the indicated component The stippled triangular field on the right of the diagram demarcates the region of bulk compositions of the system in which three phases would make up the stable assemblage. These three phases would be a Caco 3 phase, containing minor amounts of iron, a member of the dolomite-ankerite solid-solution series and a member of the magnesite-siderite solid-solution series, containing minor amounts of excess Caco 3 The three regions with subparallel lines in them demarcate two-phase fields In the field in the upper left, the two phases are a Caco 3 phase, con-taining minor amounts of iron, and a member of the dolomite-ankerite solid-solution series. In the field to the lower left, the two phases

PAGE 35

16 FIGURE 5. Thermodynamically calculated phase diagram (with tie-lines) for the system, CaOMgO-FeO-C02-H20, at 250C The three phase fiel d of calcite, dolomite-ankerite solid-solution, and magnesit e -siderite solid-solution is shaded. The regions without tie-lines, along t h e bas e and at the very top, are one-phase fields. In the regions with tie-lines a two -phase assemblage crystallizes. From Anovitz and Essene (1987, Fig. 12).

PAGE 36

MgCO 3 17 FIGURE 6. Thermodynamically calculated phase diagram (with tie-lines) for the system, CaO-MgO-FeO-COz-HzO, at 400C. The three phase field of iron-bearing calcite, dolomite-ankerite solid solution, and magnesite-siderite solid-solution is shaded. The regions without tie-lines, along the base and at the very top, are one-phase fields. In the regions with tielines a two-phase assemblage crystallizes. From Anovitz and Essene (1987, Fig. 13).

PAGE 37

FIGURE 7 18 CaCO 3 Thermodynamically calculated phase diagram (with tie-lines) for the system, CaO-MgO-FeO-COz-HzO, at 550C. The three phase field of iron-rich calcite, dolomite-ankerite solidsolution and magnesite-siderite solid-solution is shaded. The white regions without tie-lines, along the base and at the very top, are one-phase fields. In the regions with tie-lines a two-phase assemblage crystallizes. From Anovitz and Essene (1987, Fig. 14).

PAGE 38

19 are a member of the dolomite-ankerite solid-solution-series and a member of the magnesite-siderite solid-solution-series containing minor amounts of excess Caco 3 In the field to the far right, the two phases are a member of the magnesite-siderite solid-solution-series containing minor amounts of Caco 3 and a Caco 3 phase containing minor amounts of iron. The subparallel lines connect the compositions of the coprecipitating carbonates. Unmarked fields a t the top and bottom of the diagrams are one-phase fields in which a Caco 3 phase, with minor amounts of iron, and a member of the magnesite-siderite solid-solution series containing minor amounts of excess Caco 3 respectively, would crystallize. Anovitz and Essene (1987) compared these figures with the experi-mental work of al. (1962) and Rosenberg (1967), and pointed out a significant difference; "The calcium content of natural siderites in equilibrium with ankerite is much less than that predicted by the experiments (Klein and Gale, 1981). The calcium content of natural siderites either represents the true equilibrium position of this boundary or the effects of retrograde compositional resetting." There is only a small amount of experimental data available for the carbonate-containing assemblages of the banded iron-formations at low temperatures. Although Grubb (1971) did some experimental gel runs with the ferroan silicate minerals at 110-450C, he did not investigate the carbonate minerals. Ricketts (1980) investigated Mg-Ca-carbonate precipitation in hydrated silica gel s at 25C by adding a supernatant solution of Ca-and Mg-chlorides to a sodium metasilicate gel containing NaHC03 Mg-calcites with up to 16 mol % Mgco3 and aragonite precipitated 2+ 2+ by diffusion of Mg and Ca into the gel. Ricketts suggested that the formation of high-Mg-calcites (and possibly even dolomites) may be

PAGE 39

controlled more by the rates of diffusion of components, than by their relative concentrations in the solution. 20 Finally, Goldsmith and Graf (1957) precipitated a complete series of solid-solutions along the Caco 3 -Mnc o 3 join at room temperature. The precipitates were poorly crystallized, but their composition remained unchanged for up to six months in contact with their supernatant liquids. They reported that Vegard (1947), however, precipitated calcites containing only up to 15 mol% MnC03 and rhodochrosites with only up to 25 mol % Caco 3 from concentrated solutions of CaC12 -MnC12 and NaHco3 They concluded that the compositions of the precipitates resulting from such experiments could be expected to vary according to the experimental conditions. In preliminary experiments done in our own laboratory a wellcrystallized, buff-colored siderite, which gave an easily recognizable X-ray pattern, was derived from mixing powdered Caco 3 with a solution of Feso4 7H2 0 at pH = 2.5. Also, a similar buff-colored precipitate presumed to be siderite, which unfortunately has not been successfully X-rayed due to oxidation p roblems, resulted from the combination of solutions of Feso4 .7H2 o and NaHco3 It appears that the iron-rich carbonates precipitate more readily than the magnesium-or manganeserich phases. Available compositional data Essene (1983) compiled most of the available data for carbonates in the system of interest. They appear on figures 8 and 9 (Essene (1983), Figures 2 and 3). The data were taken from hydrothermal, sedimentary,

PAGE 40

.\ -t. ... \ .. ... .. \ ... ........ . . ... \ . . :.. ... . .... . . . . ::::i\ .. : ., .. , w .. :,. I ,jl .;,\ , .. .,. MgCO, Fe CO, 21 FIGURE 8. Carbonate compositions (mol%) in the system, MnC03-MgC03 FeC03. Carbonates with more than 5 mol% CaC03 are excluded. The complete solid-solution between MgC03 .and FeC03, and between FeC03 and MnC03, is shown by these data. The carbon ate compositions in the figure are taken from hydrothermal, sedimentary, and metamorphic occurrences described in the literature. From Essene (1983, Fig. 2).

PAGE 41

. CaMn(C05 ) 1 .. .. f\ . .. . . . . .. . . . .. . . .. . .. . . . . . . .. . .. ... . . . . . ..... . .. . .:. \ : .. .. i..!! ... 'u .. : tr'; ',',y: Co Mg !COsl2 Co Fe (C01 )2 22 FIGURE 9. Carbonate compositions (mol % ) in the system, CaMn(C0 3 ) 2 CaMg(C03)2-CaFe(C03) 2 Carbonates with more than 60 mol % or less than 40 mol% CaC03 in solid-solution are excluded. The solid-solution between CaMg(C03)2, CaFe(C03)2, and CaMn(C03)2 is extensive for magnesium-rich compositions. The carbonate compositions in the figure are taken from hydrothermal, sedimentary and metamorphic occurrences described in the literature. From Essene (1983, Fig. 3).

PAGE 42

23 and metamorphic occurrences. The continuous ranges in composition indicate the extensive solid-solution observed in this system. Composi tional joins (such as magnesite-rhodochrosite) for which carbonate compositions are missing, is taken as permissive evidence of a solvus gap. (Essene (1983) points out, however, that there may be geochemical or petrological reasons for the lack of intermediates.) The solidsolution between dolomite and ankerite does not extend through the entire range of iron-rich compositions, although it appears to be continuous up to about 70 mol% of the CaFe(C0 3 ) 2 component. This is consistent with the experimental work of Goldsmith et al. (1962) and Rosenberg (1960, 1963). Figures 10, 11, and 12 show the data compiled from natural assemblages by Anovitz and Essene (1987) for the CaC03 -MgC03-FeC03 system at the specific temperatures 250, 400, and 550C. Temperatures of formation were estimated f rom observed mineral assemblages, or other data available for the rocks, and are considered to have an error of C. Coexisting pairs of carbonates are connected with tie-lines. The indicated field boundaries were calculated from their solution model.

PAGE 43

24 FIGURE 10. Triangular compositional diagram for natural carbonates in the system, CaO-MgO-FeO-C02-H20, at 250C. The field boundaries were determined graphically in agreement with the results for 400C and 550C (Figures 11 and 12). The three-phase triangular field is shaded. The carbonate analyses were renormalized to Ca + Mg + Fe = 1. The temperature estimate is believed to have a 50C error. From Anovitz & Essene (1987, Fig. 6).

PAGE 44

25 FIGURE 11. Triangular compositional diagram for natural carbonates in the system, CaO-MgO-FeO-C02-H20, at 400C The field boundaries were calculated from the phase equilibria model of Anovitz and Essene (1987). The three-phase triangular field is shaded. The carbonate analyses were renormalized to Ca + Mg + Fe = 1. The temperature estimate is believed t o have a 50C error. From Anovitz and Essene (1987, Fig.

PAGE 45

26 coco. FIGURE 12. Triangular compositional diagram for natural carbonates in the system, CaO-MgO-FeO-C02-H20, at 550C. The field bound aries were calculated from the phase equilibria model of Anovitz and Essene (1987). The three-phase triangular field is stippled. The carbonate analyses were renormalized to Ca + Mg + Fe = 1. The temperature estimate is believed to have a 50C error. From Anovitz and Essene (1987, Fig. 9).

PAGE 46

27 DATA ON THE SYSTEM FROM THE MARRA MAMBA BANDED IRON-FORMATION General introduction to the banded iron-formations The best guides to the low temperature carbonate system are probably the data collected from the banded iron-formations. James (1966) defined banded iron-formations as iron-rich rocks of largely Precambrian age, exhibiting alternating bands of iron minerals and chert. Table 4 gives the average compositional ranges of the world's banded ironformations. The banded iron-formations are one of the two major types of deposits, which together make-up more than 90% of the world's iron production. (Oolitic ironstones are the other major type.) Table 5 lists the key characteristics of the world's major banded iron-formations, giving their approximate age, initial extent, lithologic associations, sedimentary structures, facies, and degree of metamorphism. Probably 90% or more of all known iron deposits are contained in the five great districts indicated with an asterisk (James and Trendall, 1982). (The deposit from which the data modelled in this report are taken is one of these districts, the Hamersley Range of Western Australia.) The largest deposits are of Archean and Earl y Proterozoic age, and all the very largest deposits are between 2 and 2.5 byo. This clustering of ages suggests that a tectonic, environmental, and/or climatic regime, peculiar to this time in earth history, led to the extensive deposition of these rocks (Eugster and Chou 1973; James and Trendall, 1982; Holland, 1984).

PAGE 47

TABLE 4 Average composi tional ranges of banded iron-formations (Klein, 1983) and averages for the u pper and lowe r portions of the Marra Mamba (Klein and Gole 1981). Component Content (In weight percent) Average b anded Harra Mamba iron-formation Upper Lower Si02 4 3 3 50 6 42 4 45 9 Al 2o3 0 1 1.8 0 4 2 2 Fe 2o3 12 9 -26.9 22. 5 3.4 FeO l7 .5 -25. 5 16 9 27.2 MgO 2.8 6 2 3 8 4 0 CaO 1.8 6 6 5 8 1.5 MnO 0.05 -1.2 0 1 0.1 Na2o 0 04 0 5 0.4 0 3 K 2o 0.02 -1.2 0 2 0 5 P205 0.04 0 3 0 1 0 .05 Ti02 0.01 0 2 0 2 0 1 s 0 .00 0.8 0 1 0 5 c 0 00 0 5 0 1 0 7 28

PAGE 48

TABLE 5. Key characteristics of the world's major band e d iron-formations Locality (Deposits) (Age) Lake Superior Mich./Minn. (Biwabik, Gunflint, Negaunee)(2. 0byo) Labrador Trough Eastern Canada (Sokoman) (2.2byo) Nabberu Basin Western Australia (Frere)(2.2byo) Krivoy Rog European USSR (2. 3byo) Transvaal Supergroup, South Africa (Kuruman, Penge Griquatown)(2.3byo) Ninas Gerais Brazil (Itabarite) (2 4byo) Lateral extent (kilometers) Along strike for up to 300 km. Continuous narrow belt for 800 km 80,000 km2 Discontinuously along strike for 3 000 km Discontinous but possibly correlative for 950 km 80,000 km2 Thickness (meters) 30-600 100-38 0 2000 1300 700-1000 50-600 Stratigraphic sequence in order o f decreasing abundance Grey., slate Mafic Volcanics Arg., sh., grey. Limestone, dolomite Sandstone, quartzite Basic volcanics Sandstone, quartzite Limestone, dolomite Arg sh. grey. Conglomerate Arkose, phyllite Schist, quartzite Dolomite Dolomite, limestone Arg., sh., grey. Conglomerate Quartzite Basic volcanics Volcanics Quartzite Phyllite Dolomite Sedimentary Structures Oolites Granules Banding Oolites Granules Banding Oolites Granules Banding Banding Oolites Banding Banding Facies Alteration Ox/car/sil Diagenetic to medium grade metam Car/ox/sil Diagenetic to highgrade metam Ox/car Diagenetic to highgrade metam O x/(car) Car/sil/sul Diagenetic to low-grade me tam Ox/car Low-to highgrade metam. Source Bayley & James ( 1973) Dimroth & Chauvel (1973) Hall & Goode (19 78) Alexandrov (1973) Beukes (1983) Dorr (1973) N \0

PAGE 49

TABLE 5. (cont'd). Locality Lateral extent Thickness Stratigraphic sequence Sedimentary (Deposits) (kilometers) (meters) in order of decreasing Structures (Age) abundance Hamersley Basin 100,000 km2 900 Basic & acid volcanic Banding Western Australia Arg sh., grey. Diagenetic (Marra Mamba, Dolomite, limestone to low-grade Brockman,Boolgeeda) Sandstone me tam (2.5byo) Yilgarn & Pilbara 30 Ultra-mafic Banding Blocks, Western km and up to mafic volcanics Low-to high-Australia (2 6byo) 300 km long Granite grade metam. Arg sh. grey. Canadian Shield Isolated cirUp to Mafic and ultra-Banding Ontario/Quebec cular basins 300 mafic volcanics (Michipicoten, up to 600 km Turbidites Abitibi)(2.8byo) in diameter Abbreviations: Arg .-argillite, byo-billion years old, car-carbonate, grey-greywacke, metam .-metamorphic, ox-oxide, sh.-shale, sil-silicate, sul-sulfide, facies listed in parentheses are minor. Facies Alteration All facies Ox Ox/sul/ (car) Source Tren-dall (1973) Gole & Klein (1981) Goodwin (1973) w 0

PAGE 50

31 The bands in banded iron-formations occur in various thicknesses. Microbands, a term to be used later, are defined as regularly spaced, millimeter-scale concentrations of iron and silica minerals. Mesobands are centimeter-scale bands of chert-dominated and iron-dominated microbands. Macrobands are 1-10m thick bands dominated by groups of ironformation beds and less-iron-rich beds (Trendall and Blockley, 1970). Banded iron-formations are believed to have been deposited in deep-water, low-energy environments, either in large inland basins, nearshore bays, or on quiet continental shelves (James and Trendall, 1982; Garrels, 1987; Holland, 1984). With regards to such deposits, James and Trendall (1982) stated; "The regional continuity of microbands precludes any depositional mechanism other than chemical precipitation." James (1954) described four major facies of iron-formation in which the iron minerals alternating with the chert bands consist mainly of oxides, silicates, carbonates, or sulfides. Table 6 lists the most common minerals making up each facies. The carbonate facies usually has the simplest mineralogy, and is most often found as thin beds or laminae alternating with chert layers in approximately equal proportions (James, 1954; Garrels, 1987). The conditions of chemical stability and the observed natural occurrences of the sulfide, carbonate, and oxide facies, do not overlap to any large extent. Those of the silicate facies overlap all three (Maynard, 1983; James, 1954; Krumbein and Garrels, 1952) The major control on the deposition of the facies, at least for the sulfide, carbonate, and oxide facies, was probably the oxidation potential (James, 1954; Maynard, 1983). The iron bands vary from a single phase to a complex mixture of iron minerals (carbonates, silicates and sulfides) while the chert bands are monomineralic.

PAGE 51

TABLE 6. The most common iron minerals in the four facies of banded iron-formations. (Compiled from various sources.) Hematite Magnetite Greenalite Minnesotaite Stilpnomelane Riebeckite Oxide-rich Silicate-rich 2 + 3+ (Fe ,Fe ) 2 3si2 05(0H)4 2+ (Fe ,Mg)3S14o10 (0H)2 2+ 3+ K(Fe ,Fe ,Al)10s11 2o30 (0H) 12 2 + 3+ Na2(Fe ,Mg) 3 Fe 2si8o22 (0H) 2 Chlorite 2+ 3+ 3+ (Fe ,Fe ,Al)4 6(Al,Fe ,S1)4o10(0H,0)8 Carbonate-rich Dolomite-Ankerite 2 + Ca(Mg, Fe )(C03 ) 2 Magnesite-Siderite 2 + (Mg, Fe )C0 3 Calcite Caco 3 Sulfide-rich Pyrite Pyrrhotite 32

PAGE 52

33 Description of the Marra Mamba Iron-Formation A meticulous study by Klein and Gole (1981) on the mineralogy and petrology of the Marra Mamba Iron-formation of the Hamersley Basin, Western Australia, provided detailed mineralogic and compositional data on the coexisting carbonate phases in a relatively unmetamorphosed (i.e., subjected to less than 300C after deposition) iron-formation. The iron-formations of the Hamersley Basin are some of the least altered and best-exposed iron-formations in the world (Trendall and Blockley, 1970). The modelling of these compositionally diverse, well-laminated rocks was simplified by considering precipitation of the carbonates to have occurred independently of all other minerals observed. This is a reasonable assumption since bands of pure Ca-Fe-Mg carbonate are frequently observed in the Marra Mamba between bands of silicates and chert (Klein and Gole, 1981). The following brief description of the deposit is condensed from the work of Klein and Gole (1981). The Marra Mamba Iron-formation is of early Proterozoic age (probably 2.5 byo). It varies from approximately 15 to 230 meters in thickness and is found in outcrop over about 90,000 km2 in Western Australia (Trendall and Blockley, 1970). It was originally deposited in a basin the size of Pennsylvania plus the western extension of New York. I t shows well-developed mesobanding and abundant laminations. Bulk chemical analyses (Table 4), petrographic studies (textural analyses) and electron microprobe analyses (mineral compositions) were conducted by Klein and Gole (1981) The minerals analyzed included the carbonates (calcite, dolomite-ankerite, and magnesite-siderite), sulfides (pyrite a nd pyrrhotite), oxides (hematite a nd magnetite), and silicates

PAGE 53

(minnesotaite, riebeckite, stilpnomelane). The carbonates, stilpnome lane, and minnesotaite show significant solid-solution and riebeckite, greenalite, and chlorite show limited variations in composition. 34 The uppermost and lowermost parts of the Marra Mamba Iron-formation show significant differences in mineralogy. The uppermost part, called the magnetite-rich assemblage by Klein and Gole (1981) includes quartz, magnetite, minnesotaite, riebeckite, stilpnomelane, dolomite-ankerite, calcite, and minor pyrite. The lowermost part contains significantly more pyrite than the upper, plus pyrrhotite, quartz, magnesite-siderite, dolomite-ankerite, stilpnomelane, minnesotaite, and minor carbon ( up to 2 wt.%). These differences in mineralogy are reflected in the differences of bulk composition shown in Table 4. The Upper Marra Mamba is richer in oxidized iron and calcium while the Lower Mar r a Mamba is richer in reduced iron, magnesium, sulfur and carbon. Many carbonates in the Marra Mamba are more magnesian than in other iron-formations. In the uppermost, magnetite-rich section, carbonates are often major constituent s and pairs of coexisting calcite and dolomite-ankerite are abundant. These carbonates tend to be medium-to coarse-grained with almost euhedral outlines interpreted to indicate considerable recrystallization (Klein and Gole, 1981). The carbonates often form relatively continuous, thin bands between bands of magnetite and silicates. Compositions of the carbonates in this portion of the Marra Mamba are given in Table 7. The calcites are nearly pure Caco 3 (usually with less than 2 wt.% FeO or MnO and less than 1 wt.% MgO). Members of the dolomite-ankerite series contain about 25-64 mol % of the CaFe(co3 ) 2 component and usually less than 1 wt.% MnO. These compositions are displayed graphically in Figure 13.

PAGE 54

35 TABLE 7. Representative microprobe analyses of carbonates in the upper part of the Marra Mamba Iron Formation [After Klein and Gole (1981)]. Weight percent oxides Mole fraction of the cations on the basis of 2(Fe,Mn,Mg,Ca) FeO MnO MgO CaO Total Fe Mn Mg Ca Total Calcite 0.94 0.25 0 .24 53 02 54.45 0 027 0 .007 0.012 1.954 2 000 0 .97 0 .99 0.49 54.03 55.73 0.027 0.007 0 024 1. 921 2.000 1.00 1.14 0 42 55 49 58.05 0.028 0.032 0.021 1.969 2.000 1.08 0.42 0.51 54.41 56.42 0 030 0.012 0.025 1.933 2.000 1.38 1.01 0.56 53.46 56.41 0 039 0 029 0.028 1.904 2 000 1.53 0.43 0.49 55.88 58 33 0.041 0 012 0.023 1.924 2.000 1.65 1.86 0.67 52 .45 56 .63 0 046 0.052 0.033 1.869 2.000 2.03 0.41 0 .47 52.27 55.18 0.058 0.012 0.024 1.906 2.000 Dolomite-ankerite 12.97 0.92 10.44 26.12 50.45 0.393 0.028 0.564 1.015 2.000 16.82 0.98 9.30 26.43 53.53 0.493 0.029 0.486 0.992 2.000 18.77 0.78 7 .46 26.73 53.74 0 560 0 023 0.396 1.021 2.000 20.31 0.82 7.74 26.47 55 34 0.590 0 024 0.401 0.985 2.000

PAGE 55

MgCO 3 Caco3 Calcite 58 analyses Molecular 4lft 50 FeCO+MnCO 3 3 36 FIGURE 13. Carbonate compositions determined by Klein and Gale (1981) for the upper portion of the Marra Mamba Iron-formation. The apices of the triangle represent 100 mol% of the indicated component. Dolomite-ankerites ranging from 25-64 mol% of the CaFe(C03)2 component, and containing minor amounts of excess CaC03 and MgC03-FeC03, are shown in the stippled region. These dolomite-ankerites were found to be in equilibrium with calcite compositions in the hatched region.

PAGE 56

37 Members of the dolomite-ankerite series occur throughout the Marra Mamba, but magnesite-siderite/dolomite-ankerite pairs occur only in the lower part. Like the carbonates in the upper section, the carbonates of the lower section are usually medium-to coarse-grained and tend to have euhedral outlines. This texture is interpreted by Klein and Gole (1981) to indicate recrystallization. Fine banding and laminations are common, but regionally continuous microbanding is generally poorly developed. Alternating bands tend to have distinctly different mineralogies. The continuity of fine laminations is commonly interrupted by well-crystallized, medium-grained carbonates (Klein and Gole, 1981). Compositions of carbonates in this portion of the Marra Mamba are given in Table 8, which is adapted from Klein and Gole (1981). The members of the dolomite-ankerite series contain about 26-71 mol% of the CaFe(C0 3 ) 2 component, which appears to be the maximum possible iron content for any temperature (Anovitz and Essene, 1987). They usually contain less than 0.5 wt.% MoO. The compositions of the magnesite-siderites, ranging from Fe 0 5Mg0 5co3 to nearly pure iron siderite (97 mol %), are displayed in Figure 14 The iron-rich siderites are also the most calcium-rich (Klein and Gole, 1981).

PAGE 57

38 TABLE 8. Representative microprobe analyses of carbonates in the lower part of the Marra Mamba Iron-Formation (After Klein and Gole, 1981). Weight Eercent oxides Mole fraction of the cations on the basis of 2(Fe,Mn,Mg,Ca) FeO MnO MgO CaO Total Fe Mn Mg Ca Total Dolomite-ankerite 10.61 0.30 14.82 27.90 53.63 0 290 0.008 0 723 0.979 2 000 11.77 0 52 13.91 27.27 53.47 0.327 0 015 0 688 0.970 2.000 12.46 0.19 12.14 27.31 52.10 0.360 0.005 0.625 1.010 2.000 15.93 0 .46 11.15 26.06 53.60 0.457 0.013 0.571 0.959 2.000 16. 85 0 .42 11.18 26.73 55 .18 0.472 0.012 0.558 0 958 2.000 16.88 0 .45 10.19 27.45 54 .97 0.478 0.013 0 514 0 995 2.000 18.21 0.25 10.19 26.41 55.06 0.517 0 007 0 516 0.960 2.000 21.15 0.02 7.57 25.92 54 .66 0.623 0 001 0.397 0 979 2.000 23.11 0.18 6 34 25.85 55.48 0.683 0 005 0.334 0.978 2.000 Magnesite-siderite 36.09 1.41 19.68 0 16 57 34 0 992 0.039 0.963 0.006 2.000 43.87 0 .90 12.08 0.16 57.01 1.318 0 028 0.647 0.007 2.000 44.02 0.84 10.06 0.17 55.09 1.397 0 027 0.569 0.007 2.000 47.00 0.35 9.50 0.53 57.38 1.447 O.Oll 0.521 0.021 2.000 50.14 0.27 9.45 0.30 60.16 1.482 0 008 0.498 0.012 2.000 49.96 0.21 7.04 0.48 57 .69 1.578 0.007 0.396 0.019 2.000 51.98 0.05 5.35 0.19 57.57 1.682 0 002 0.308 0 008 2.000 50.16 0.84 2 .27 2.71 55.98 1.714 0 029 0.138 O .ll9 2.000 53 .07 0.07 0.78 2.43 56 35 1.842 0 002 0.048 0.108 2.000 54.55 0.05 0.90 1.87 57.37 1.861 0.002 0.055 0.082 2 000 56.06 0.00 1.08 1.03 58.17 1.891 0.000 0.065 0.044 2.000

PAGE 58

MgCO 3 Molecular% Siderite; FeCO + MnCO 159 analyses 3 3 39 FIGURE 14. Carbonate compositions determined by Klein and Gole (1981) for the lower portion of the Marra Mamba Iron-formation. The apices of the triangle represent 100 mol % of the indicated component. Dolomite-ankerites ranging from 26-71 mol% of the CaFe(COs)z component and containing minor amounts of excess CaCOs and MgCOs-FeCOs, are shown in the stippled region. These dolomite-ankerites were found to be in equilibrium with magnesite-siderite compositions ranging from 50-97 mol% of the FeCOs component (shown in the irregular stippled region at the base of the The tie-lines connect coexisting (i. e. physically touching) pairs of carbonates, and are separated into three approximately parallel sets; I, II and III. These correspond to the most gently dipping those of intermediate slopes, and the most steeply dipping, respectively.

PAGE 59

40 DERIVATION OF THE MODEL Available thermodynamic data for the system The thermodynamic data available for the Caco 3 -MgC03 FeC0 3 system are listed in Table 9. The values used in the calculations are listed in Table 10. There is significant variation in the values for both the 2+ carbonate minerals and some of the aqueous cations, notably Fe The values used were carefully chosen from these data based on best agree-ment with observed natural assemblages or the best experimental work available. For example, the free energy of formation of calcite was chosen to yield a solubility product for the dissolution of calcite of 1 0 8 5 [in agreement with the work of al. (1974) and Plummer a nd Busenberg (1982)]. 2+ The numbers for Fe and Feco3 were taken from the compilation of Naumov al. (1974), because they yield the experimental value for comm.). His preferred K siderite preferred by R. M Garrels (pers. sp value for dolomite, from Helgeson et al. (1978), was also used. The magnesite and Mg2+ values of Robie et al. (1978) were used, because the magnesite value of Naumov is based on experi-mental data in which serious discrepancies have been shown to exist (R. M. Garrels, pers. comm. ) Thermodynamic data for ankerite are notably absent from the compilation and the necessary values will be derived in a later section. The accepted value for the free energy of formation of dolomite is from al. (1978) and is for a fully ordered

PAGE 60

41 TABLE 9. Available thermodynamic properties system. of the CaO-HgO-FeO-CO H 0 2 2 Phase Description State b. Ho b. Go so Source f f kJ/mol kJ/mol J/deg/mol 2+ 543 0 553 0 55 2 1 Ca aq -542.7 -552.7 -55.2 2 -553.1 3 -542.8 -553.5 -53.1 4 -542.8 -553 6 53 1 5 -543.0 -552.8 6 -542.7 -552. 7 -55.2 7 Caco 3 calcite c -1206.9 -1128.8 92. 9 1 -1206.8 -1128.3 91.7 2 -1207.4 -1128.8 91.7 4 -1206.9 -1128.8 92 9 5 -1207 4 -1128.8 6 -1207 7 92.9 8 -1208 2 -1130.1 92. 7 9 -1129. 3 10 -1129.7 11 -1209.0 -1130 6 91.8 1 2 -1129.8 1 7 CaMg(C03 ) 2 dolomite c -2331.7 2 169. 3 1 2314 6 -2151.9 155 2 2 2324 5 -2161.7 155 2 4 -2326. 3 2163.4 155 2 5 2332 7 2 170. 0 155 2 8 -2329. 9 -2167.2 155 2 9 -2177.8 13 Fe 2 + aq -87.9 -84.9 -113. 4 1 -92.6 -92. 2 -104.6 2 -89.1 -78. 9 -138 0 4 -89.1 -78.9 137 7 5 89.1 91.2 6 -92.6 -92.2 104.6 7 -92. 6 90.1 -111.7 14 -92.0 -91.3 105 8 15 Feco 3 siderite c -747.7 673 9 92.9 1 753. 1 -680.3 96. 1 2 737 0 666 7 105 0 4 -740.6 666. 7 92 9 5 -737.0 679 4 6 -749.6 679 4 105.0 9 -752.2 679 5 96.1 14 -749.6 679 5 15

PAGE 61

TABLE 9 (cont'd). phase Description State LlH0 f Fe Mn M .956c0.042 g.002 3 Mg 2 + Mgco3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 kJ/mol natural c siderite aq -462.0 -461.7 -466. 8 -466.8 -461.7 magnesite c -1113. -1095.8 -1113.3 -1095.8 -1113.2 -1111.4 -1113.1 Rossini et al. (1952) Naumov Tardy and Garrels (1974) Robie et al. (1978) Wagman et al. (1982) Sangameshwar and Barnes (1983) Babushkin et al. (1985) Karpov et al.-cl971) Helgeson-er-81. (1978) Harvie et-al:-(1984) Christ et al. (1974) Robinson-er-81. (1982) Garrels Mel 'nik 0972) Helgeson (1983 and 1984) Robie et al. (1984) Plummer-and Busenberg (1982) Abbreviations: aq-aqueous, c-crystalline solid. 42 LlG0 so Source f kJ/mol J/deg/mol 95.5 16 -456.0 -118.0 1 -455.3 -119.7 2 -454.8 -138.0 4 -454.8 -138.1 5 -455. 2 -119.7 14 1029. 65.7 1 -1012.3 65.7 2 -1029. 5 65.1 4 -1012.1 65.7 5 -1029.6 65.8 8 -1027.8 65.7 9 -1027.3 10 -1029.7 65.7 14

PAGE 62

43 TABLE 10. Thermodynamic data used for the calculations in this manuscript (From Robie et al. (1978) unless otherwise noted). Phase CaC03 (c) calcite MgC03 (c) magnesite Feco 3 (c) siderite CaMg(C03 ) 2 (c) dolomite CaFe(C0 3 ) 2 (c) ankerite ca2 + (aq) Mg 2+ (aq) Fe 2+ (aq) C02 3 (aq) FeO (c) MgO (c) H 2 0 (1) co 2 (g) 1 Christ et al. (1974) 2 Naumov et al. (1974) 3 Helgeson al. (1978) 4 This paper /:). Ho f kJ/mol -1113.3 -753.12 2329.93 -542.8 -466.8 -92.62 -677.1 -272.0 -601.5 285.8 -393.5 /:).Go f kJ/mol -1129.71 -1029.5 -680.32 -2167.03 -1815.24 -553.5 -454.8 -92.22 -527.9 -251.2 -569. 2 237.1 -394.4 Abbreviations: aq-aqueous, c-crystalline solid, g-gas, 1-liquid. so J/mol/deg 91.7 65.1 96.12 155.23 189.64 -53.1 -138.0 -104.62 -56.9 59.8 26.9 70.0 213.8

PAGE 63

44 (RJ) dolomite. Therefore, the value of the free energy of formation of ankerite derived from the calculations in the following section, is also that for a fully ordered ankerite (RJ). It is assumed that an ideal solid-solution exists between the two minerals requiring that they both have the same structure and degree of ordering. Helgeson et al. (1978) give a 6. Gf(dol) for a fully disordered dolomite (R3c). This value is 8.8 kJ/mol less negative (i.e., less stable) than that of the ordered dolomite. If this value of AGf(dol) were used in the calculations, the resultant value for ankerite would also be that of the disordered mineral. al. (1978) also give a value for 6.Gf(dol) in its stable ordering configuration for 25C. The value given is identical to that of a fully ordered dolomite. They believe that a typical sedimentary dolomite is probably metastable, and that it has an ordering parameter (s) of 0. 7 ("s" is defined to equal 2XCa,A -1. "Xca,A" equals the mole fraction of calcium atoms on the larger "A" sites in dolomite.). They caution that metastable disorder in dolomite can have a significant effect on geochemical processes. Reeder (1983), however, questions the assumption that most sedimentary dolomites are disordered. He cites the fact that there are no published determinations of the ordering state of sedimentary dolomites, except for that of the Lake Arthur dolomite (Reeder, 1983). That dolomite was found to be ideally ordered. In conclusion, disorder in dolomite would have a significant effect on the calculated value of 6.Gf(ank) There is still, however, considerable controversy over the extent of disordering in sedimentary dolom -ites. Therefore, the effects on these calculations will not be discussed further, except to say that the same problems will presumably

PAGE 64

45 apply to both dolomite and ankerite. Whichever value is used for dolomite will produce a value for an ankerite with the same degree of order. Derivation of the low temperature phase relations Calculation of the Gibbs free energy of formation of ankerite Figure 14 shows the range in composition of dolomite-ankerites and magnesite-siderites determined by microprobe analysis for the lower portion of the Marra Mamba (Klein and Gale, 1981). The stippled field in the center of the diagram represents the range of compositions observed for the dolomite-ankerites. The stippled region near the base of the triangle shows the observed range in magnesite-siderite compositions. The tie-lines in the lower half of the triangle connect pairs of coexisting carbonates [i.e., physically touching according to Klein and Gale (1981)]. (The relative stratigraphic positions and complete analyses of the minerals connected b y these tie-lines are unknown to this author.) The tie-lines are divided into three subparallel sets designated I, II, and II, for the purpose of. the following calculations. Crossing tie-lines indicate apparent inconsistencies in the equilibrium relations of the coexisting dolomite-ankerite/magnesitesiderite pairs (Figure 14). There are several explanations for crossing tie-lines, including analytical inaccuracies, changes in stability caused by minor, non-stoichiometric amounts of CaO or MnO, differences in the water equilibrated with the various pairs of coexisting carbona tes, differences in the temperatures of met amorphism of coexisting pairs, a nd disequilibrium between the phases.

PAGE 65

46 The carbonates analyzed by Klein and Gole (1981) were medium-to coarse-grained indicating that there was little likelihood of significant analytical errors. [Very fine-grained minerals of less than a few microns in diameter (i. e less than the beam diameter of most probes) would have made accurate analysis of individual grains ver y difficult.] The effects of small amounts of excess CaO or MnO on the stabilities of the minerals will not be considered in this simplified treatment of the system. The effect of the minor amount of Mnco3 (less than 2 mol%) is probably negligible, as indicated in the recent modelling study of Anovitz and Essene (1987). They concluded that greater than 6 mol % of Mnco3 (or any other component) is required to limit the precision of their derived phase relations significantly. As an example, 2 mol % MnC03 in siderite makes less than a 0 .5% change in the free energy of formation of siderite (-683.3 versus -680.3 kJ/mol, assuming an ideal solid-solution model). Such a change has only a minor effect on the phase relations, considering that all the carbonates contain about the same amount of manganese (Tables 7 and 8). It is likely that the composition of the solution remained constant for long periods of time during precipitation and equilibration of the lower and then of the upper portions of these deposits. This is a result of two factors. First, the climatic and tectonic regime postulated for this period of deposition of banded iron-formations was unusually stable (James and Trendall, 1982). Therefore, the composition of the source water flowing into the depositional basin was undoubtedly quite constant. Second, only small amounts of diagenesis appear to take place in carbonate sediments that remain under the sea (Blatt et al. 1980; Berner, 1971) The former authors state; "Because diagenesis is a

PAGE 66

47 response of carbonate minerals to the water within the pore space, it is logical that minerals formed within a given water will undergo relatively little change until such time as the composition of the water changes." Isotopic and other evidence cited in the Discussion section of this paper indicate that diagenesis of these deposits was largely isochemical with virtually no influx of external fluids (i.e., those other than the source fluid) into the sediments. Differences in temperatures of metamorphism, perhaps due to slightly different positions of the minerals in the stratigraphic column or isolated chemical and physical environments within the sediment, are another possible explanation for the crossing tie-lines. However the data from Anovitz and Essene (1987) shown in Figures 5, 6, and 7, indicates that the slopes of the tie-lines change very slightly with temperature. If the relationships observed at higher temperatures hold for lower temperatures, a change of approximately 300C would be necessary to explain the changes in slope indicated on Figure 14. The limited vertical extent (less than 230 meters of thickness) and the lack of nearby intrusives precludes such large local temperature differences. Klein and Gole (1981) who used extensive petrographic and microprobe studies for their research, mention that all the assemblages may not be equilibrium assemblages. They state; "It has been impossible for us, at the very low temperatures these assemblages have undergone, to differentiate equilibrium and disequilibrium coexistences." Therefore, it appears that disequilibrium is the most likely explanation for the crossing tie-lines observed on Figure 14. So, it will be necessary to determine which set of tie-lines most closely approaches equilibrium after a value for6Gf(ank) is determined for each tie-line.

PAGE 67

48 Three sets of approximately parallel tie-lines were separated out of the group of 13 tie-lines shown on Figure 14. The tie-lines belonging to each of the 3 sets are indicated by the Roman numerals I, II, and III corresponding to the most gently dipping, those of intermediate slopes, and the most steeply dipping, respectively. For modelling purposes, the tie-lines were e xtended to intersect t h e dolomite-ankerite and magnesite-siderite joins and the compositions of the coexisting carbonates were taken off the diagram. This procedure involves a certain amount of simplification, because the actual phases analyzed were not stoichiometric. Also, errors arose from the mechanics of picking compositions off the diagram. In order to calculate6 Gf(ank)' as described in this section, it is assumed that the two carbonate s6lid-solutions are ideal. All available evidence indicates that the magnesite-siderite solid-solution is ideal. The experimental work of Rosenberg (1963), conducted down to temperatures of 295C, indicated no evidence of a miscibility gap anywhere along the magnesite-siderite join. The other piece of important evidence is the occurrence of natural Mg-Fe carbonates containing the entire range of Fe/Mg ratios (Figure 8). Also, Lippman (1982), in his theoretical derivation of carbonate solub ility diagrams, determined that this solid-solution is virtually ideal. The case for ideality of the dolomite-ankerite solid-solution, is not as strong as that for the magnesite-siderite solid-solution. The pure iron end-member has neither been synthesized nor found to occur naturally. Also, high temperature experimental work (295600C) indicates that any bulk compositions containing more than about 71 mol % of the CaFe(co3 ) 2 end-member, occur as two-or three-phase assemblages

PAGE 68

49 rather than a single-phase solid-solution. However, in the range of compositions containing less iron than this, a continuous range of ratios has been observed in iron-formations and other CaFe-Mg carbon-ates (Figure 9). For the purposes of this study, therefore, the carbonates of the dolomite-ankerite series will also be considered as an ideal solid-solution. An approximation of the effects of non-ideality on the calculated values for 6Gf(ank) and on the aqueous solution compositions will be given below. A value for 6Gf(ank) can be calculated from each of the tie-lines on the figure using existing thermodynamic data for magnesite, siderite, dolomite, and the aqueous species and the compositional information on the diagram. The derivation of the equations that describe equilibrium between a solid-solution and an aqueous phase was first developed by Gibbs (1876, 1878). Thorstenson and Plummer (1977) and Mackenzie al. (1983) describe more recent applications of the criteria for such equilibria. The first step is to write the equations representing equi-librium between the two end member components of the solid-solution and a dilute, aqueous solution. For siderite the equation is Feco3(c) 2+ 2-= Fe (aq) + co3 (aq) (1) and analogously for magnesite MgC03(c) 2+ Mg (aq) + 2-co3 (aq) (2) The equilibrium constants for these reactions can be written. Because the solid-solution is considered to be ideal, no activity coefficient is required for the FeC0 3 and MgC03 components in the solid-solution. (The activity of the end-member in the solid-solution is considered to equal its mole fraction.) The activity coefficients for the divalent cations ca2+, Mg2+, and Fe 2 + are similar under any

PAGE 69

50 weakly complexing conditions. These coefficients are significant, even at low ionic strengths, but in the final expressions used in this model they cancel each other, and can be ignored in the following derivation. Therefore, the equilibrium constants for these reactions (which are also called solubility products) can be written as follows, where the symbol "[ ]" is used to indicate the mole fraction of the enclosed species in either the aqueous or solid phase; [Fe 2+][co;-] K mag = = [FeC03 ] [Mg2+J[co;-J (3) (4) The values of the equilibrium constants for these reactions can be cal-culated from the equality D. G0 -RTlnK f and the values of the free energies of formation for the species in reactions 1 and 2 (Tables 9 and 10). Because the mineral containing these end-members is in equilibrium with the aqueous solution, both these equilibrium constants must be satisfied and the following distribution coefficient can be defined; K sid/mag which reduces to = K .d K mag Ksid/mag [FeC03 ] = 2+ [Fe ][NgC0 3 ] ---------------2+ [FeC0 3)[Mg ) (5)

PAGE 70

51 This does not imply that both the end-members are in equilibrium with the aqueous solution. For example, as explained by Thorstenson and Plummer (1977), a state of "metastable equilibrium" can exist in which the solid-solution is in equilibrium with an aqueous solution that is supersaturated with respect to one of the pure endmembers. This distribution coefficient also represents the following ex-change reaction of iron and magnesium between the two carbonates FeC03(c) + 2 + Mg (aq) = l1gC03 (c) + 2+ Fe (aq). (6) Values of this and other necessary equilibrium constants are given in Table 11. An equation like 6 can also be written for the Fe/Mg exchange between dolomite and ankerite in the aqueous solution, and takes the form [To simplify some of the algebra involved in these and other calcula-tions half formulas for ankerite and dolomite are used, but the mineral represented is still the ordered carbonate with the dolomite structure (R3), and not the disordered mineral with the calcite structure (R3c) ] The distribution coefficient (Kank/dol) for this reaction is 2 2+ [Ca 0 5 MgO.S(C0 3)] [Fe ] K ank/dol __________________ 2 ___ 2+--[Ca0.5Fe0.5(C03)] [Mg ] (8) To determine a value for the free energy of formation of ankerite, it is necessary to determine a value for Kank/dol and to combine it with the values for the free energies of formation of dolomite and the aqueous ions (Tables 9 and 10). Compositional data from the diagram of Klein a nd Gole (1981) are used t o calculate Kank/dol (Figure 14). The two solid-solutions (ankerite-dolomite and siderite-magnesite), coexist in equilibrium with the same aqueous solution and with each

PAGE 71

TABLE 11. Values of equilibrium constants for reactions between pairs of carbonate minerals. The first mineral of the pair is the reacting species and the second is the product. In the reactions ankerite and dolomite are always written as the half formula (e.g., ca0 5Mg0 5co3). The following abbreviations are used: ank=ankerite, dol=dolomite, mag= magnesite, sid=siderite, cal=calcite. Carbonate pair ank/dol sid/mag dol/mag ank/sid dol/cal ank/cal sid/mag ank/dol K expression [dol]2[Fe2+] _____ 2 ___ 2+[ank] [Mg ] 2+ [mag] [F e ] --------2+[sid][Mg ] [mag][Ca 2+]0.S [sid][Ca2 +JO.S [cal][Mg2 +JO.S [cal][Fe2 +JO.S [ank]2 [mag] Value 10-2.35 1 0 0 .81 10-0. 55 10-0.46 52

PAGE 72

53 other. Therefore, both Ksid/mag and Kank/dol must be satisfied at the same time, so from a procedure analogous to that used to derive equation 5 the following expression results; K sm/ad Kank/dol 2 [Ca 0 5 Fe 0 5(co3)] [MgC03 ] __________________ 2 _______ [Ca0.5Mg0.5(C03)] [Feco3 ] (9) The symbol K I d designates the combined equilibrium constant from the sm a quotient of the equilibrium constants of reactions 5 and 8, and actually represents the distribution coefficient for an exchange reaction involv-ing the four endmembers. Using compositional data for the first tie-line to the left in Set I (Figure 14) as an example, a value for Ksm/ad and, therefore, can be derived. The concentrations of the four end-members are their mole fractions in the two solid-solutions and upon substitution into equation 9 they yield = [0.37]2[0.16] [0.02] Kank/dol = 0.06. [0.33] Substitution of the value for K .d/ from Table 11 yields s1 mag 10-2.35 K ank/dol 0.06 and K ank/dol The calculation of (ank) for each tie-line, from the exchange reaction between ankerite and dolomite (Equation 7), is as follows -5.707 log Kank/dol -5.707 ( -1.128) 6 44 kJ. (10) Substituting free energy of formation values for the species gives; 6.44 kJ = 2/.'.Go f (dol) + f (Fe 2 +) UG0 f (ank) (Mg2+) f = 2(-1083.5) + ( -92.2) f (ank) ( -454.8)

PAGE 73

-1804.4 kJ 21:{; 0 f (ank) -905.4 kJ = fl Go f (ank) The values ofll G (ank) (for the half formula) calculated for all the tie-lines on Figure 14 are given in Table 12 These values vary by less than 4 kJ/mol indicating that the free energy of formation of ankerite is well-constrained by this method. 54 The effect of non-ideality of the two carbonate solid-solutions on the calculated phase relations can be roughly According to Garrels and Christ (1965), a regular solid-solution model provides a close approximation to the behavior of many binary nonelectrolyte solutions of various kinds." For regular, binary solid-solutions the activity coefficient is represented by the following equation (Garrels and Christ, 1965) ln Y 1 = where B is a constant, independent of composition, and X is the mole fraction of the indicated end-member. Lippman ( 1982) calculated values of B for a number of binary carbonate solid-solutions by using a Madelung-Vegard approach in calculating electrostatic excess energies of the mixed crystals. The constants he obtained are 0.2521, 0 .4393, 8.7627, and 6.4357 for the MgC03 -FeC0 3 FeC0 3 -Mnco 3 Caco 3 -MgC03 and Caco 3-Feco3 solid-solutions, respectively. Such constants are very rare in the geological literature because there is virtually no experimental data for 25C and (until this paper) no thermodynamic description of the phase relations at 25C. Therefore, these constants are approximate. The B value calculated for the Caco 3 -HgC0 3 solid-solution (i.e., 8 7627) was used to describe the dolomite-ankerite solid-solution. Therefore,

PAGE 74

55 TABLE 12. Values of the free energy of formation of ankerite calculated from the tie-lines on the compositional diagram of Figure 12. (These values are for the half formula.) Mole fractions of end member s Ankerite Dolomite Siderite Set I 0.37 0.63 0.84 0 .41 0.59 0.91 0 46 0 54 0.96 Set II 0.32 0.68 0.69 0.39 0.61 0.74 0.42 0.58 0.76 0.44 0.56 0 80 0.47 0.53 0.81 0.51 0 49 0 87 Set III 0.33 0.67 0.51 0 35 0 .65 0.54 0.41 0.59 0.63 0 60 0.40 0 .76 Magnesite 0.16 0 .09 0.04 Average 0 .31 0.26 0 .24 0.20 0.19 0 13 Average 0.49 0 .46 0.37 0 .24 Average b. G{(ank) kJ/mol -905. 4 -905. 2 -904.7 -905.1 -906.1 -906.6 -906.8 -906.7 -906.9 -906.7 -906.6 907 2 -907.3 -907.4 -908.6 -907.6

PAGE 75

the resultant free energy values and solution boundary should show the maximum likely effect of non-ideality on these calculations. 56 The activity coefficients and the values of (for the half formula) calculated assuming a regular solid-solution model to describe the activity-composition relations of the two carbonates found in the Marra Mamba are shown in Table 13. The activity coefficients for ankerite, dolomite, siderite, and magnesite in the compositional range observed in the Marra Mamba vary from 4.1-57.5, 2.4-23.4, 1.0-1.1, and 1.1-1.3, respectively. The magnesite-siderite is virtually an ideal solid-solution while the dolomite-ankerite solid-solution shows consid-erably more non-ideality. However, the effect on the calculated value of the free energy of formation of ankerite is minimal. The differences are the largest at the compositional extremes. The average of the 13 values found in the last column of Table 13 is -910.1. This varies from the value used in these calculations by only 2 5 kJ/mol. The previous calculations yield 1 3 different values G (ank). To determine which values most closely represent the true value for ankerite at 25C, the concept that the free energy of a reaction at equilibrium equals 0, is used. (When the equals 0 the product and reactant assemblages are equally stable. ) For each tie-line it is possible to write a reaction representing equilibrium between the coexisting carbonates and the aqueous solution. Then, using the values of the free energy of formation at 25C for all the products and reactants, including the calculated value for (ank)' a free energy of reaction can be calculated and compared to 0 The tie-line reactions that yield a closest to 0 will be the ones that are calculated using the value for (ank) closest to the true value at 25C

PAGE 76

57 TABLE 13. Activity coefficients calculated for the 13 tie-lines of Figure 12 assuming a regular solid-solution model to represent the activity-composition relations of the carbonates. The resultant values for the free energy of formation of ankeriLe are compared to those calculated assuming an ideal solid-solution model. (These values are for the half formula.) Ankerite Dolomite Siderite Magnesite t.G5 (ank) X y X y X y X y k /mol Ideal Non-ideal Set I 0.37 32 4 0.63 3.3 0.84 1.0 0.16 1.2 905.4 -911.4 0 .41 21.1 0 59 4.4 0.91 1.0 0.09 1.2 -905.2 -909.0 0.46 12.9 0.54 6.4 0.96 1.0 0.04 1.3 -904.7 -906.6 Set II 0.32 57. 5 0.68 2 4 0.69 1.0 0.31 1.1 -906.1 914.0 0.39 26.1 0.61 3.8 0.74 1.0 0.26 1.1 -906.6 -911.4 0.42 19. 1 0.58 4.7 0.76 1.0 0.24 1.2 -906.8 -910. 4 0.44 15.6 0.56 5 4 0.80 1. 0 0 .20 1.2 -906.7 -909.4 0.47 11.7 0.53 6 9 0.81 1.0 0.19 1.2 -906. 9 -908. 3 0.51 8 2 0.49 9 8 0.87 1.0 0 .13 1.2 -906 7 906 4 Set III 0 33 51.1 0 .67 2.6 0.51 1.1 0 49 1.1 -907 2 -914.5 0.35 40.5 0.65 2.9 0.54 1.0 0 .46 1.1 -907.3 -913.8 0.41 21.1 0.59 4 4 0.63 1.0 0.37 1.1 -907.4 -911.3 0.60 4.1 0.40 23.4 0 .76 1.0 0.24 1.2 908 6 -904.4

PAGE 77

58 First, the free energies of formation for each of the 26 carbonate minerals (13 dolomite-ankerites and 13 coexisting magnesite-siderites) connected by the 13 tie-lines must be calculated. Each value derived for b. G (ank), plus the values of the free energies of formation of pure dolomite, magnesite and siderite at 25C are used. A term for the ideal free energy of mixing is also included. For an ideal solid-solution, made up of end-members A and B, the free energy of formation of the component minerals is given by the following equation 6 G (ss) = X b. G0 A f (A) + + + (11) where XA and XB are the mole fractions of the A and B end-members (Nordstrom and Munoz, 1986). "n" is the number of sites on which mixing occurs. (In this case "n" is 1, because the Mg and Fe atoms both occupy the same site in the carbonates and the chemical formulas of the miner-als are written so that the total number of Mg and Fe atoms in the mineral adds up to one mole). "R" is the universal gas constant (8.3143 JrK/mol). "T" is the temperature in degrees Kelvin (298.15K). The last term on the right in equation 11, is the contribution of the ideal entropy of mixing Fe and Mg atoms, to the free energy of formation of the solid-solution. This term is always negative, since for all com-positions between the pure end-members the mole fractions are less than 1 (and the logarithms are less than 0). Therefore, substitution of one atom for another in the intermediate members of a solid-solution serves to stabilize that composition with respect to the pure end-members. Using values of the free energies of formation calculated for the 26 carbonates, as well as values for the aqueous species, an exchange reaction can be written representing equilibrium between the two

PAGE 78

59 minerals and the aqueous solution. For the first tie-line on Figure 14 the reaction is 2+ 2+ 2+ 2Fe 0 84Mg0 16(co3 ) + 0.31Mg + Ca = CaFe 0 37Mg0 63(co3 ) 2 + 1.31Fe Using the value of 6Gf(ank) calculated from the tie-line, the free energy of formation of the ankerite-dolomite solid-solution, and thereby 6GR, are calculated. For the above equation the result is 6GR = -2037.0 + 1.31(-92.2) -2(-737.3) -0.31(-454.8) -(-553.5) 11.31 kJ. For a reaction at equilibrium the 6GR should be 0. Therefore, calculation of the free energies of reaction for all the assemblages represented by the 13 tie-lines on Figure 14, will reveal which set of tie-lines yields free energies of reaction closest to 0 Such a cal-culation indicates free energies of reaction of about 11, 8, and 3 kJ/mol, respectively, for sets I, II, and III. Therefore, the average value of the free energy of formation for ankerite, calculated from the tie-lines of set III, -907.6 kJ/mol, is most likely to represent the true 6Gf(ank) at 25C and is adopted for use in these calculations. The preceding method of choosing the set of 25 C tie-lines has been challenged on the basis that at higher temperatures ( 250C) the tie-lines projected for this system steepen with increasing temperature. If differing temperatures of equilibration caused the crossing tie-lines plotted on Figure 14, this would indicate that Set III represents the highest temperature assemblage of carbonates analyzed by Klein and Gole (1981). Since it is unlikely that the carbonates represented by Sets I and II were deposited at lower than 25C, an apparent contradiction arises. However, the choice of the set of 25C tie-lines (and there-fore, the value used for 6Gf(ank)) has an insignificant effect on the

PAGE 79

60 overall results. Using Set I instead of Set III would change the value for 6Gf(ank) by 2.4 kJ/mol. This is a difference of approximately 0.25%, which would not make an appreciable difference in the calculated ion ratios for the aqueous solutions. I could have just as easily calculated a value for ankerite by considering it to be an ideal mixture of calcite and siderite end-members in a 50/50 solid-solution. (A similar calculation for dolomite yields a value ver y close to the 6 Gf (dol) accepted in this paper.) This "ideal-mixture method" results in a6Gf(ank) of -910 kJ/mol which varies from my chosen value by 0 25 % All these values are well within the standard error of 2 kilocalories (or about kJ) associated with most thermodynamic calculations. Comparison of the real and calculated tie-lines, plotted in Figure 15, demonstrates that the predictions of the model are quite good Calculation of the aqueous solution--solid-solution relations based on the composition of the solid phase Now that a value for the free energy of formation of ankerite has been derived, it is possibl e to calculate the composition of the aqueous solution in equilibrium with the mineral. For the upper part of the compositional triangle (Figure 13), calcite is in equilibrium with members of the ankerite-do lomite solid-solution. The calcite is not pure stoichiometric Caco 3 but it contains less than 2 wt.% FeO and (MnO + MgO) (Table 7). Therefore, in the calculations that follow it will be considered to be pure. 2 + 2+ In order to calculate the ratios of Fe /Ca and M g 2+/ca2 + in the equilibrium solution, two equations are necessary. The first describes equilibrium between calcite and either end-member of

PAGE 80

Tie-lines from Anovitz & Essene ( 1987) Calculated t i e-lines Actual tie-lines Calcite Dolomite 0.35 Magnesite 61 Ankerite Siderite FIGURE 15. Diagram comparing tie-lines in the lower half of the compositional triangle derived from various sources. The light solid lines represent those calculated for 25C from the model described in the text. The heavy solid lines represent those calculated for 250C from Anovitz and Essene (1987). The dashed lines are the assemblages (set III) observed by Klein and Gale (1981).

PAGE 81

62 the ankerite-dolomite solid-solution. The other describes the Fe-Mg exchange equilibrium between dolomite and ankerite. The reaction between dolomite and calcite in equilibrium with the solution is 2+ 2+ ca0 5Mg0 5co3(c) + 0.5Ca (aq) = Caco 3(c) + 0 .5Mg (aq) (12) The expression for the distribution coefficient for this reaction is K dol/cal = [Caco 3 ][Mg 2+]0.5 [ca0.5Mg0.5co3J[ca ] The values for all such equilibrium constants are listed in Table 11. Calcite is eliminated from the expression, because it is a pure phase and its activity (and therefore its mole fraction) equals unity. Solving for the mole fraction of dolomite yields [Mg2+]0 5 dol/cal (13) The other equation necessary, is that representing equilibrium between ankerite and dolomite, (14) The expression for the distribution coefficient for this reaction is 2+ 0 5 [ca0 5 Fe 0 5co3 ](Mg ] K dol/ank [ca0.5Mg0.5co3JcFe 1 Recognizing that the mole fraction of ankerite equals 1 minus the mole fraction of dolomite, yields K = dol/ank Substituting the expression for the mole fraction of dolomite from equation 13 into equation 15, yields (15)

PAGE 82

K dol/ank [1 ][Mg2+]0 5 [Fe2+]0.5 [Ca ] Kdol/cal Rearranging and simplifying, yields [ M 2+]0.5[Fe2+]0.5K g dol/ank [Ca ] Kdol/cal = Multiplying equation 17 through by 1 simplifying a nd rearranging, yields [Mg2+]0.5 K dol/cal [Fe2+1o 5 K -----------d ol/ank [ Ca2+]0.5 63 ( 16 ) (18) Using equation 18, the boundary representing equilibrium between calcite and ankerite-dolomite in an aqueous solution can be calculated when values for Kdol/cal and Kdol/ank from Table 11 are s ub stituted. The ion ratios calculated by this equation a r e listed in Table 14 and are designated as the algebraic solution to the problem. The boundary is shown on Figure 16 The resultant phase diagram shows a rapid change in 2+ 2+ the Fe /Ca r a tio a t the dolomite e nd of the boundary. T h e boundary has a modera t e slope in the middle. 2+ 2 + Like the F e /Ca ratio a t the 2+ 2 + dolomit e end, the Mg /Ca ratio changes rapidly a t the ankerite end of the boundary To calculat e the ion ratios in equilibrium with particular members of the solid-solution, the following e quations are required [Mg2+]0.5 Kdol/cal = 1 0 0.55 o r

PAGE 83

64 TABLE 14 Composition of the aqueous solution in equilibrium with various members of the ankerite-dolomite solid solution and calcite. T h e difference between the numeric and algebraic ratios is explained in the text. Mole fractions 2 + 2+ log Mg /Ca Fe 2 + Mg2+ log------ca2 + log------Fe 2 + Anke rite Dolomite Numeric Algebraic 0.99 0 .01 -5.10 -4.05 -3.01 -1.04 0.90 0.10 3 10 -2.93 -3.09 0.16 0.80 0 20 -2. 50 -2. 42 3 .19 o. 77 0.70 0 30 -2.14 -2.10 -3.31 1. 21 0 .60 0 40 -1.90 -1.86 -3. 44 1.58 0.50 0.50 -1.70 -1.68 3 .60 1.92 0.40 0.60 -1.54 -1.53 -3.80 2.27 0 30 0.70 -1.41 -1.40 4 .05 2.65 0 .20 0 80 -1.29 -1.29 -4. 40 3 .11 0.10 0.90 -1.19 -1.19 5 00 3 81 0 .01 0.99 -1.11 -1.11 -7.00 5.89

PAGE 84

1.0 ANKERITE -3.0 0. 1 -5.0 CALCITE DOLOMITE -7.0 -5.0 -3.0 1 0 1 0 LOG A 2+ I A 2+ Mg Ca FIGURE 16. Aqueous solution--solid-solution phase diagram for the CaO-FeO-MgO-C02-H20 system. The diagram is calculated from the compositional data of Klein and Gale (1981) for the upper portion of the Marra Mamba. The term AFe indicates aqueous iron in its stable valence state (divalent in the ancient ocean and the proposed iron-basin feedwater, and trivalent in the modern ocean). The stability fields of the various minerals and solid-solutions are indicated. The points on the curved phase boundary mark the indicated mole fraction of ankerite in the solid phase in equilibrium with an aqueous solution of the indicated composition. 65

PAGE 85

66 [Mg2+]0.5 and (19) Kank/cal = [Fe2+1o s ---------------2+-5:5-[ca0. 5Fe0 .5co3J[ca ] [Fe2+10.5 = 10 1 .50 or (20) The boundary between calcite and ankerite-dolomite can also be derived from equations 1 9 and 20, by plotting the ratios calculated for corres-ponding values of the mole fraction of ankerite and dolomite together (e.g. XANK= 0.7, 0 .3). This calculation produces the values listed in Table 14, which are designated as the numeric solution. The algebraic and numeric approaches give slightly different results for compositions near the ankerite end-member (due to imprecision in the thermodynamic data). 2+ 2 + Figure 17 displays the variation of Mg /Fe 2+ 2 + 2+ 2+ Fe /Ca and Mg /Ca with the mole fraction of the ankerite end-member. The solid bars on the curves represent the range in ankerite compositions seen in the Upper Marra Mamba deposit. The filled circle on the o rdinate represents the composition of present-day seawater. Values for the concentration and activity coefficient of trivalent iron -8 in the modern ocean were taken from Byrne and Kester (1976) and are 10 and 0 .04, respectively. 2+ (Although Fe is used to represent aqueous iron throughout this paper, the iron in present-day seawater is actually trivalent.) The double filled circles show the ratios calculated for Proterozoic seawater. The asterisk represents the value for the pro-posed iron-basin feed-water of Garrels (1987). (See discussion section for details). 2 + 2+ 2+ 2+ Variations of the ion ratios of Fe /Ca and Mg /Ca

PAGE 86

+ C\1 Cll 0 c 1.0 -3.0 CJ 0 .... .... 0 0.5 X ANK 1.0 0.5 X ANK 1. 0 8 0 8.0 0 0.5 X ANK 1 0 67 FIGURE 17. Calculated composition of the aqueous solution in equilie rium with the range of ankerite-dolomite compositions observed in the upper portion of the Marra Mamba. The term AFe indicates aqueous iron in its.stable valence state (divalent in the ancient ocean and the proposed iron-basin feed-water, and trivalent in the modern ocean). The solid bar on the curves indicates the range of ankerite-dolomite compositions, observed b y Klein and Gole (1981). The filled circle on the ordinate indicates the composition of presentday seawater, calculated using the iron concentrations and activity coefficient given by Byrne and Kester (1976). The double filled circles indicate the composition calculated for Proterozoic seawater. The activity ratio of magnesium to calcium is postulated to be the same for ancient and modern seawater. The asterisk indicates the composition of the proposed Proterozoic iron-basin feed-water from Garrels (1987). (The point for this value plots off the central diagram at 0.2.)

PAGE 87

68 are rapid at one of the compositional extremes of the solid-solution and gradual at the other. The analogous calculation for carbonates in the lower portion of the Marra Mamba, is similar to, but slightly more complicated than, that for the carbonates in the upper half of the Harra Mamba. The algebra for developing the equation representing equilibrium between the two solid-solutions is more complicated. It is necessary to combine three equations. Two represent exchange reactions between the end-members of the solid-solutions. The third represents equilibrium between the two solid-solutions and the aqueous solution (either an ankerite/siderite equation or a dolomite/magnesite equation). The algebra for developing the equation, is given in the Appendix. The equation that results is 0 2+ 0.59[Fe ] -10 ------[ca2+] [ Fe2+10.5 + ----------[Ca2+1o s 2+ -1.76 (Mg ] 10 ------+ [Ca 2+] [ M 2+]0.5 0 .94 g 10 -------[Ca2+]0 5 (21) This equation can be solved by the quadratic equation. The last two terms on the right side become the "c" term of the quadratic equation, 2+ 2+ when a value for Mg /Ca is substituted. The values calculated from this equation are in Table 15. The boundary is shown on Figure 18 The composition of the aqueous solution in equilibrium with the double solid-solution can also be calculated by what is designated here as a numeric model. The calculation begins with an arbitrary choice of the mole fraction of siderite in the siderite-magnesite solid-solution. The equilibrium constant for the ankerite-dolomite/magnesite-siderite reaction (Table 11) yields a value for the mole fraction of ankerite in equilibrium with this siderite-magnesite. Then, using the equilibrium constant for the ankerite/siderite exchange reaction, the ratio of 2+ 2+ 2+ 2 + Fe /Ca is derived. Similarly, the ratio of Mg /Ca for the

PAGE 88

TABLE 15. Values of the ratios of the aqueous ions in solution in equilibrium with various pairs of ankerite-dolomite and sideritemagnesite, calculated from equation 21 (algebraic solution) and the numeric method outlined in the text (numeric solution). Mole fractions 2+ 2+ log Mg /Ca 2+ 2 + log Fe /Ca 1 M 2+/F 2 + og g e 69 [sid] [ank] Numeric Algebraic Numeric Algebraic 0.99999 0.995 -3.836 -1.186 -1.175 -2.650 -2.660 0.9999 0.983 -2.826 -1.176 -1.165 -1.650 -1.660 0 999 0.949 -1.795 -1.146 -1.135 -0.650 -0.660 0 .99 0.854 -0.708 -1.062 -1.052 0.354 0.344 0.9 0.638 0.504 -0.892 -0. 880 1.396 1 384 0.8 0.541 0 898 -0. 850 -0.836 1. 748 1 734 0.7 0.474 1.131 -0.851 -0.834 1.982 1. 965 0 6 0 419 1.296 -0.878 -0.859 2 174 2.155 0 5 0.371 1.420 -0.930 -0. 905 2 350 2 325 0 4 0.325 1.517 -1.009 -0.976 2.526 2 493 0.3 0.278 1.593 -1.125 -1.076 2.718 2.669 0 2 0.227 1 650 -1.302 -1.216 2.952 2.866 0.1 0.164 1.684 -1.620 1.384 3.304 3.068 0 .01 0.056 1. 661 -2.684 -1.258 4.345 2.919 0.001 0.018 1.635 -3.714 -1.169 5.349 2.804 0.0001 0.006 1.625 -4.725 -1.143 6 350 2.768 0 00001 0.002 1.622 -5. 728 -1.089 7.350 2 711

PAGE 89

--, \ SIDERITE \ \ \ \ 0 9 1 0 0 .995 0 .949 ANKERITE ---, \ \ 0.16 0 1 7 0 A LOG FE 0 9 ACA2+ 0 0 r--5. 0 CALCITE 0 1 0 > G) z -i m m (/) -i m -7.0 -5.0 -3.0 -1. 0 1. 0 3.0 LOG A 2 + I A 2+ MG CA FIGURE 18. Aqueous solutien--solid-solution phase diagram for the CaO-FeO-MgOC02-H20 system, calculated from the compositional data of Klein and Gale (1981) for the upper and lower portions of the Marra Mamba. The term indicates aqueous iron in its stable valence state (divalent in the ancient ocean and the proposed iron-basin feed-water, and trivalent in the modern ocean). The stability fields of the various minerals and solid-solutions are indicated. The points on the curved phase bo undary indicate the mole frac t i o n s of ankerite and siderite i n the solid phase, in equilibriu m with the aqueous solution of the indicated composition. The two dashed lines indicate the shift caused in the phase boundaries (for the range of mi neral compositions observed in the Marra Mamba) by assuming a regular instead of an ideal, solid-solution model for the carbonates.

PAGE 90

71 dolomite/ankerite exchange reaction is calculated. The value of 2+ 2+ Mg /Fe is easily derived from the tw o previous ratios. The results of this numeric calculation are listed in Table 15 This second calculation serves as a check on the algebraic solution, discussed above. As was the case for the composition of the aqueous solution in equilibrium with carbonates of the upper portion of the Marra Mamba, the values calculated from the algebraic and numeric methods diverge at one compositional extreme. The effect of imprecision in the thermodynamic data is magnified, in the case of the Lower Marra Mamba, because three equations, instead of two, are used. The effect of assuming a regular solid-solution model (instead of an ideal solid-solution model) to describe the activity-composition relations of the two carbonates is significant. The two partial bound-aries plotted with dashed lines on Figure 18 represent the solution boundaries calculated for a regular solid-solution model. They were calculated using activity coefficients of 57.5, 2 .4, 1.0, and 1.1 for ankerite, dolomite, siderite, and magnesite, respectively. Although these activity coefficients actually vary with composition for a regular solid-solution model, these values represent maximum deviations from ideality for the compositional range observed in the Marra Therefore, the dashed curves on Figure 18 show the maximum likely effect of non-ideality on the calculated solution boundaries. The sensitivity of these models, to changes in the values of the free energies of formation of the component species, is demonstrated in Figure 19. The phase boundaries, indicated with the dashed lines, were calculated using only the thermodynamic data of al. (1978). This involves an increase of approximately 13 kJ/mol in the values of

PAGE 91

A LOG Fe 1 0 ----------------------, --" ..--..... . l I I I \ I I I \ -5.0 -1.0 1.0 L 0 G A 2+-/ A 2+ Mg Ca '\ \ \ I I 72 FIGURE 19. Diagram showing the sensitivity of the phase boundaries shown in Figure 8, to changes in the free energies of forma tion used in the calculations. The two curves indicated with dots were generated using a value for the free energy of ankerite of about 8 kJ/mol less stable than the value chosen in this paper. The phase boundaries indicated with dashed lines were calculated using only the thermodynamic data of Robie et al. (1978) and the value for the free energy of ankerite-calculated in this paper. See the text for a .description of the values used to calculate this boundary.

PAGE 92

73 2+ the free energies of formation of siderite and Fe (aq), an increase of 5 kJ/mol in the value for dolomite, and an increase of about 1 kJ/mol in the value for calcite. 2+ 2+ Using these values, the Fe /Ca ratio of the calcite-ankerite boundary decreases by about 2.5 log units. That of the ankerite-siderite boundary increases by about the same amount, due to the decreased stability of siderite. Such changes yield an unreasonably large field of stability for the ankerite-dolomite phase, considering its limited distribution in nature. The two curves, indicated with dots, were generated using a value for 6Gf(ank)' about 8 kJ (2 kcal)/mol less stable than the value chosen in this paper. 2+ 2+ For th1s value, the Fe /Ca ratio of the calcite-ankerite boundary is increased by about 2 log units. That of the ankerite-siderite boundary is decreased by about the same amount, due to the decreased stability of ankerite. A less-stable value for 6Gf(ank) accentuates the hump in the boundary between the two solidsolutions. It also limits the field of the ankerite-dolomite phase to the magnesium-rich side of the diagram. Figure 20 displays the composition of the aqueous solution, in equilibrium with various mole fractions of ankerite in the solid phase. As in Figure 17, the solid bars on the curves represent the range in ankerite-dolomite compositions seen in the Marra Mamba deposit. The filled circle on the ordinate represents the composition of present-day seawater, calculated as described for Figure 17. The double filled circles represent the composition of Proterozoic seawater. The asterisk indicates the composition of the proposed Proterozoic iron-basin feed-water from Garrels (1987). The shapes of the curves are very similar to those on Figure 17, but the absolute values of the ratios are

PAGE 93

0 + ('I lL. CG 0 c( ...... + ...... ...... N + ('I 0) Q Q) -2.0 c( 0 (!J
PAGE 94

75 different. The solution that deposited the carbonates of the Upper Marra Mamba was much less iron-rich than that depositing the Lower Marra Mamba. However, as was the case for the solution compositions in equilibrium with minerals in the upper portion of the Marra Mamba, present-day seawater does not appear t o be a likely depositing solution. Effects of elevated temperature on the calculated relations One way to estimate the effect of elevated temperature on the calculated phase boundaries is to assume that the carbonate minerals reequilibrated with a pore fluid, assumed to be the same as the original, basinal fluid from which they precipitated, at a higher temperature during burial diagenesis of the iron-formations. First, it is necessary to calculate the values of the equilibrium constants for the dissolution reactions at elevated temperature. The van't Hoff equation was used to calculate the equilibrium constants at 150C (423K) [which is taken as an average temperature of diagenesis for these deposits (see discussion below)] ln b.H0 [ 1 1 ] ------------. R T 2 T 1 (22) K (T1 ) The standard derivation for this relationship is given by Moore (1964, p. 180-181). Before calculating the enthalpy of the ankerite dissolution reaction, it is necessary to derive a value for the entropy of ankerite. The method used was adapted from al. (1978). The standard molal entropy of ankerite can be estimated using the reaction

PAGE 95

76 The entropy of this reaction is S0dol + SFe0 soMgO where6 SR is the entropy of reaction and S0 is the absolute entropy of the phase at 25C. A term S0S(ank) is defined as follows by Helgeson et al. (1978) so S(ank) Similarly, = = v0dol s0dol + = 155.18J/mol + SFe0 59.8J/mol = 188.04J/mol/deg. + VFe0 voMgO vo ank SoMgO 26 .94J/mol (where is the volume change for the reaction and V0 is the molar volume of the phase at 25C), and V0S(ank) is vo S(ank) = vo ank = v0dol 3 64.34cm 3 65.092cm + VFe0 3 l2.00cm voMgO 3 ll.248cm (S0 and V0 values are from Robie et al. (1978) unless otherwise noted) Close estimates of the standard molal entropies of iron-containing minerals at 25C and 1 bar can be computed from the following equation (after Helgeson et al., 1978) So (Vo + Vo ) So = ____ ___ (23) ank(P,T) 8.368m. 2V0S(ank)P,T The last term is a correction for ferrous iron, and is the same as that used for ferrous iron in silicates. It equals -8.368J/mol per mole of ferrous iron in one mole of the mineral (Helgeson et al. 1978, p. 51). Dana's Manual of Mineralogy (Hurlbut, 1971) lists the specific gravity of ankerite (CaFe(co 3 ) 2 ) as 2.95-3.00, so the molar volume was

PAGE 96

77 estimated from the following equation vo ank molecular weight of ankerite -------------------------------= 215.94gm/mol -----------3--= 71.98cm 3/mol. specific gravity of ankerite 3 00gm/cm Upon substitution into equation 23 this yields so ank(P,T) 188.04 (65.092 + 71.98) = --------------------------8.368 2 (65.092) 189.62 J/mol/deg For simple salts, the method of Latimer (1952) also permits the estimation of the entropy from the sum of the entropies of the consti-tuent elements. Using his tables (87 and 90) of the entropies of the elements and negative ions a value of 166.10 J/mol/deg is derived for dolomite. This varies from the currently accepted value of 155.18 J/mol/deg by 10.92 J/mol/deg. Applying this same correction to an estimate of the entropy of ankerite by Latimer's method yields S0 = 177.82 J/mol/deg ank 10.92 J/mol/deg = 166.9 J/mol/deg. 166.9 J/mol/deg is quite different from the calculated value of 189.62 J/mol/deg. Apparently, the ferrous iron in ankerite really "throws a monkey wrench in the works." Analogies with dolomite are not neces-sarily valid, because of the excessive distortion of the Cao 6 octa-hedron, caused by increased substitution of iron for magnesium. The following analogy does seem to work. In it the ferrous iron discrepancy is probably accounted for by using a siderite component, which takes into account the structural distortions caused by substi-tuting iron for magnesium. Assuming that the entropy of mixing, and therefore of the following reaction is zero yields a value very close to that previously calculated.

PAGE 97

78 CaC03 + calcite siderite ankerite 91.7 J/mol/deg + 96.1 J/mol/deg* 187.8 J/mol/deg Analogously, this assumption gives a good approximation of the S0 for dolomite (155.18 J/mol/deg). + = calcite magnesite dolomite 91.7 J/mol/deg + 65.09 J/mol/deg 156.79 J/mol/deg From a strictly empirical standpoint, including the term for the ideal entropy o f mixing [i.e., fiSmix = -nR(XAlnXA + does not give as good an approximation of the entropy of these phases as does the assump-tion that the entropy of mixing is zero. This is probably because the mixing of these two calcite-structure components to form a carbonate with the dolomite structure is not ideal. Therefore, a value of 189 6 J/mol/deg, will be used for the absolute entropy of ankerite at 25C and 1 bar. The fiHR at 25C was calculated as follows 2+ Ca (aq) + 2+ Fe (aq) + 2 co2 -(aq) 3 CaFe(C0 3 ) 2 (c) 6 SR S0 2 + + Ca sFe2+ + s;nk 3 = (-53.1) + (-104.6) + 2(-56.9) (189.6) = -461.1 J/mol = 113.7 + 298.15(-0. 4611) -23.81 kJ/mol *Siderite value i s from Naumov et al. (1974)

PAGE 98

79 The enthal py of reaction is substituted into equation 22 and the equation is solved for ln K (150C) yielding -48.68. This is equivalent to a log K (150C) value of 2 1 .14. In a similar fashion, the equilib-rium constants at 150C, for the dissolution of siderite, dolomite, and magnesite, were calculated, yielding the values in Table 16. TABLE 16 Comparison of the values of the dissociation constant s for the Ca-MgFe carbonates at 25C and 150C. Mineral log K (150C) log K (25C) Difference Ankerite -21.14 -19.93 1. 2 1 Siderite -11.37 -10.55 0 82 Magnesite -9.78 8 20 1.58 Dolomite -19.85 -18.03 1.82 These dissolution constants were used to calculate the equilibrium constants at 150C for the various solution-solid reactions, such as those in Table 11 (which lists the values for 25C). The equilibrium constants that result are given in Table 17. These equilibrium con-stants are substituted into equation 9 in the Appendix They result in the algebraic solution for the phase bound a r y between ankerite-dolomite and siderite-magnesite. This boundary i s represented by the following equation and is compared with the boundary for 25C on Figure 21. 0 2 + 0 .54[Fe ] -10 ------[ca2+] [Fe2+]0.5 + ---------[Ca2+]0.5 2 + -1.05 [Hg ) 10 ------+ [ca2+] [ M 2+]0. 5 -0. 90 g 10 -------[Ca2+]0. 5 The result is a difference of approximately one log unit in the ratio of Mg2+/Ca2+. The change is seen mainly in this ratio, because it is multiplied by K in the above equatio n K is the only sid/mag sid/mag constant that changes significantly with temperature.

PAGE 99

TABLE 17. Comparison of the equilibrium constants for some carbonate exchange reactions at 25C and 150C. Carbonate pair K expression ank/dol sid/mag ank/sid 2 2+ [ank] [Mg ] 2 + [mag][Fe ] 2 + [sid][Mg ] 100.59 80

PAGE 100

SIDERITE -tot=1 50 oc I 12 s c A LOG FE -30 ANKERIT E A 2+ CA I I > C> z m f-DOLOMITE I I en -t m -5.oL I I I I I l J I I 1 5 0 3 0 -co 1 0 LOG A 2 + / A 2 + MG CA FIGURE 21. Diagram showing the shift in the aqueous solution--solid-solution boundary, calculated for the lowerportionof the Marra Mamba, due to assuming a temperature of precipitation of the carbonates of 150C, instead of 25C. The position of the phase boundary with respect to the logarithm of the ratio of magnesium to calcium, decreases by approximately one log unit. CX> t-'

PAGE 101

82 PARAGENESIS AND DIAGENESIS OF THE CARBONATES OF THE MARRA MAMBA The preceeding model is useful for determining the composition of the solution from which assemblages containing solid-solutions were precipitated, providing there exists adequate compositional and thermo-dynamic data. It requires, however, that the assemblages currently observed in the rocks are the ones initially deposited, that they are in equilibrium, and that their compositions have not been altered significantly by diagenesis or metamorphism. The following section discusses the data available on the banded iron-formations, particularly the Marra Mamba, which pertain to this problem. Paragenesis Siderite-magnesite The following phases are generally considered to be primary precip-itates (i.e., exhibiting no detectable evidence of replacement of any pre-existing phase) in banded iron-formations; amorphous silica gels, 3+ 2+ ferric/ferrous hydroxides (e.g. Fe(OH) 3 or Fe 2 Fe (OH)8), iron-bearing silicate gels with a composition resembling that of greenalite, and siderite-magnesite (Ayres, 1972; Dimroth and Chauvel, 1973; French, 1973; Govett, 1968; Maynard, 1983; Perry et al., 1973). These minerals were either a part of the original precipitate or equilibrated with it during early diagenesis (French, 1973). Siderite-magnesite is the only

PAGE 102

83 phase generally believed to have precipitated in microcrystalline form (Maynard, 1983; Ayres, 1972; LaBerge, 1964; Dimroth and Chauvel, 1973). In places the silica gel is believed to have contained significant amounts of iron hydroxides, which coprecipitated with the silica. [Laboratory experiments have verified that such a process could have occurred. For example, Harder (1965) and Harder and Flehmig (1970) successfully coprecipitated amorphous silica and hydroxides of iron, aluminum, manganese, magnesium and other elements at temperatures below 80C. The y used solutions with Si02 concentrations comparable to those of natural waters.] In some cases, this precipitate is believed to have later segregated into discrete iron-rich and silica-rich bands, which eventually recrystallized to hematite, or magnetite, and to quartz. In other cases, it remained as mixed silica and iron bands, which became the mixed chert/iron oxide bands observed today. The coprecipitated phase may also have altered to bands of iron-silicate minerals during diagenesis and metamorphism. The siderite-magnesite seems to have been initially precipitated as uniformly fine-grained, rounded grains which averaged 5-35 microns in diameter (Ayres, 1 972 ; Dimroth and Chauvel, 1973; Floran and Papike, 1975; LaBerge, 1964). Dimroth and Chauvel (1973) made a detailed study of the textures and phase relations of the iron-formations of the Sokoman of Quebec. Analogously to micritic limestones, they visualize the deposition of siderite-magnesite as having occurred in a crystalline ooze. They cite the occurrence of cross-bedding i n carbonate femicrite, as evidence of the lack of cohesiveness of the femicrites at the time of deposition, and as an indication that the femicrites were likely deposited as oozes of crystalline minerals, not as gels. The term

PAGE 103

84 femicrite (describing the equivalent of limestone micrite) is used to describe the matrices composed predominantly of iron carbonate and iron silicate (Dimroth, 1968). Such matrices are found in iron-rich rocks, like banded iron-formations. The matrices of these rocks have a much larger grain size than that of a typical limestone micrite (2 microns). Floran and Papike (1975) believe this is due to a combination of two effects. First, extensive grain coarsening due to recrystallization could have occurred during the extended time for which these deposits have been lithified and buried. Second, certain components in these iron-rich matrices have a higher chemical reactivity than those of a limestone micrite. Many siderite-magnesite occurrences show indications of significant recrystallization. Recrystallization textures include large variations in grain size in a single sample; coarse, euhedrally-shaped, interlocking grains; and rhombic overgrowths on smaller spheroidal grains (Floran and Papike, 1975). They point out that in some samples several stages of postdepositional growth and recrystallization occurred. Dolomite-ankerite The dolomite-ankerite phase occurs in many forms in the banded iron-formations. It is found most commonly as a replacement of other carbonates or as a recrystallization product. However, it has also been found showing presumably primary textures. Ankeritic carbonates are seen to replace siderite-magnesite in many banded iron-formations, such as those observed in the Superior Iron Ore province by LaBerge (1964) In places, Dimroth and Chauvel (1973)

PAGE 104

85 observed siderite-magnesite and ankerite-dolomite porphyroblasts replacing the microgranular sideritic fabric. Floran and Papike ( 1 975 p. 1177) noted, "Within the thicker beds, relict granules suggest that these laminae have, for the most part, been converted to ankerite after deposition. Ankerite also occurs as vein fillings associated with soft sediment deformation." They also describe recrystallization textures. "In both cherty and slaty iron-formation, ankerite is present typically as coarse, zoned rhombs generally containing abundant inclusions. Most ankerite is coarsely recrystallized Ayres (1972, p. 1230) also notes the presence of replacement and recrystallization textures; "The volume-for-volume replacement of other minerals by ankerite is indicated by the absence of disruption of adjacent minerals or siderite microbands and by the occurrence of linear arrays of mineral inclusions that lie parallel to the banding and pass uninterrupted through the ankerite rhombs and" crystallization of many carbonate mesobands to coarser-grained ankerite or siderite has taken place indicated by residual traces of the original bedding laminae. A feature of some ankerite rhombs which may indicate more than one period of crystal growth is that the cores are finer grained and contain minute chert blebs, whereas the rims are coarser and clearer." Ankerite-dolomite is also observed as what may be a primary precipitate (Becker and Clayton, 1972). Floran and Papike (1975) note that fine-grained anhedral siderite-magnesite (the type that is usually designated as primary) is commonly associated with other carbonates in small mottles or concretions. They observed that most dolomite-ankerite is coarsely recrystallized. However, one sample was found to contain thin monomineralic laminae of cryptocrystalline ankerite-dolomite. They

PAGE 105

86 decided that this occurrence was probably primary, although a variable grain size suggested that some recrystallization had taken place. Dimroth and Chauvel (1973) identified only microgranular sideritemagnesite (i.e., probably primary) in their thin sections. They cautioned, however, that the other carbonates may have been present in that form. They may not have been detected due to the extremely fine grain size of this microgranular component (less than 5 microns) and to the small number of thin sections, in which they studied the mineralogy of fine-grained carbonates. Trendall and Blockley (1970) state; . the precipitate was particulate on the colloidal scale and settled to the basin floor as a soft gelatinous precipitate which formed a "primitive chert" with even, regular 0.5 mm thick microbanding of equal proportions of chert and ankerite." Ayres (1972) believes that the original carbonate was either siderite-magnesite (containing minor Ca or Mg) or ankerite-dolomite, because both these minerals are observed defining the microbanding in the cherts. Perry et al. (1973) believe that the ankerite-dolomite found in concordant layers may be primary. Calcite Calcite appears to be the last carbonate to crystallize in banded iron-formations. It often exhibits poikiloblastic or replacement textures (Floran and Papike, 1975). Klein and Gole (1981, p.511) describe the pairs of coexisting calcite and dolomite-ankerite of the Upper Marra :t-lamba this way; Such carbonates tend to be medium-to coarsegrained, commonly with almost euhedral outlines, reflecting considerable recrystallization." Floran and Papike (1975, p. 1177) give the

PAGE 106

87 following description of carbonate occurrences in the Gunflint Ironformation of Ontario-Minnesota; At Kakabeka Falls, thin laminae of the chert-carbonate facies consist of spheroidal siderite with rhombic overgrowths disseminated throughout calcite poikiloblasts. In some of the lamina, the cores have been replaced by calcite." Calcite in these deposits also occurs as a coarse-grained cement and as a replacement mineral. Ayres (1972, p. 1230) described the calcite in the Dales Gorge Member of the Hamersley Iron-formations as follows; "Calcite has clearly formed by replacing pre-existing fine-grained carbonate." Another similarit y between the calcite occurrences in banded ironformations is their stratigraphic location. In the rocks studied by Perry et al. (1973) and Floran and Papike (1975), as well as in the Marra Mamba, the calcite-rich beds seem to occur at the top of the ironformation. In s u mmary, most students of the banded iron-formations believe that crystalline siderite-magnesite was a primary precipitate from the original depositing solution, which suffered varying degrees of recrystallization during diagenesis and metamorphism. The origin of ankerite-dolomite is more ambiguous Most textures indicate significant recrystallization, or even development by replacement of other minerals, usually carbonates. Some occurrences, however, are believed to be primary. Finally, calcit e i s usually the last carbonate to crystallize and often occurs as a replacement of other carbonates. Calcite-rich beds are also often the youngest carbonate-containing beds in banded iron-formations. This is the case for the Marra Mamba, where the

PAGE 107

calcite-rich beds are stratigraphically discontinuous with the older iron-formation beds. Conditions and timing of the diagenesis 88 Attempts have been made to estimate the temperatures of metamorphism and diagenesis from the mineral assemblages observed in these rocks, from isotopic analyses, and from the Fe/Mg distribution between coexisting carbonate minerals. Table 18 gives the temperature estimates of French (1973) for various assemblages characteristic of banded iron-formations. He states that; "The primary minerals [i.e., quartz, siderite, greenalite, and hematite) of iron-formation formed from the original precipitate near the sediment-water interface at pressures of less than a few hundred bars and temperatures less than 100-150C." Klein (1983) believes that diagenesis and very low-grade metamorphism of the Precambrian iron-formations occurred at 100-300C. Based on isotopic and experimental data for the observed assemblages of minerals, he considers the diagenetic reactions to have occurred below 180C. Ayres (1972) also considers 300C to be the maximum temperature of metamorphism of the Dales Gorge Member of the Hamersley iron-formations. The oxygen isotope data of Becker and Clayton (1976) indicate a maximum temperature of 310C for the Dales Gorge Member at maximum burial. Use of the iron and magnesium contents of siderite-magnesite and ankerite-dolomite as a geothermometer has been discussed by Talautsev and Sazonov (1979), and Anovitz and Essene (1987). The former observed that for iron-free dolomite and magnesite the coefficient of distribution of Mgco3 (in a very wide range of pressure/temperature conditions

PAGE 108

TABLE 18. Comparative Mineral Assemblages in Diagenetic and Low-Grade Metamorphic Iron-Formation. (Pressure and temperature conditions are estimates. From French, 1973). Original Material Pressure 100 b Temperature 100C Na-Si gel Fe(OH) 3 Fe(OH) 2 Fe CO (gels) Fe-Si gel Greenalite (?) Volcanic glass clay minerals (Na,K) Fe-Si gel Diagenetic (Primary) 1 kb 100-150C Chert Hematite Magnetite (?) Siderite Greenalite Chamosite (?) Montmorillonite Illite glass Diagenetic (Secondary) 1-2 kb 150-200C Quartz Hematite Magnetite Siderite Ankerite Greenalite (?) Chamosite Chlorite Montmorillonite Illite Chlorite Stilpnomelane (?) Low Grade Hetamorphic 2-5 kb 200-350C Quartz Hematite Magnetite Siderite Ankerite Magnetite Minnesotaite Stilpnomelane Riebeckite Chamosite (?) Chlorite Magnetite Minnesotaite Stilpnomelane Riebeckite Chlorite Stilpnomelane Riebeckite CX> \0

PAGE 109

90 of mineral formation) remains practically constant at 0.5. As the iron content of these phases increases, the values of this coefficient gradually increase, reaching nearly 2.0. They concluded that the value of this coefficient is adjusted mainly to the level of iron in the mineral-forming medium, and not to temperature or pressure. Talantsev and Sazonov (1979) do believe, however, that the distri-bution of Feco3 between these minerals is a function of temperature and pressure. Based on studies of the compositions of natural assemblages (for which reliable temperatures of formation can be estimated), they define the following relationship between temperature, and pressure; log = 7.541 [log T (K)] 0.873 [log P] -19.111. (24) K_ is the ratio of the mole fraction of the iron component in dolomite ankerite to that of the mole fraction of the iron component in magnes-ite-siderite. Table 19 compares the values of temperature calculated for the thirteen tie-lines of Figure 14 for pressures of 1, 100, and 1000 bars using equation 24. For 1 bar the temperatures range from 34-59C. For 100 bars they range from 250-292 C For 1000 bars they range from 683-738C. Equation 24 suggests an extremely large depen-dency of the Feco3 distribution between the two carbonates on pressure. In order to achieve the indicated temperatures at 100 bars (1 km, assuming hydrostatic pressure prevails in the saturated sediment), a geothermal gradient of about 2 00 C/km is required. This is unlikely to be the case, even for Proterozoic times, when a much higher rate of radiogenic heat flux prevailed. [It is believed that the Archean-Proterozoic continental geotherms were similar to those of the present (Davies, 1979), or at most three times greater than those of today

PAGE 110

91 TABLE 19. Tempera t u res of formation of the two-carbonat e assemblages, calculated from the geothermometer of Talantsev and Sazonov (1979) Mole fraction of end members TemEerature (C) at various pressures (bars) Ankerite Siderite 1 100 1000 Set I 0.37 0.84 0.44 34 250 683 0.41 0 .91 0.45 35 252 685 0.46 0.96 0.48 37 256 690 Set II 0.32 0 .69 0.46 36 253 688 0 39 0 .74 0.53 42 263 699 0 42 0.76 0.55 43 266 704 0.44 0.80 0.55 43 266 703 0.47 0 81 0.58 45 270 708 0.51 0.87 0.59 46 2 7 1 709 Set III 0.33 0.51 0.65 50 278 719 0.35 0.54 0 .65 50 278 719 0 .41 0 .63 0.65 50 278 719 0.60 0 .76 0.79 59 292 738

PAGE 111

92 (Ernst, 1983)]. Also, the temperatures predicted for 1000 bars [a pressure conceivably reached during late diagenesis (French, 1973; Klein and Gole, 1981)] are about 400C higher than the maximum temperature predicted for these rocks by other methods (Becker and Clayton, 1972; French, 1973; Klein, 1983) It appears that equation 24 is not valid at these low temperatures, probably because it was developed using data for assemblages with estimated temperatures of 400C or more. Anovitz and Essene (1987) also discuss the use of Fe/Mg partition-ing between siderite-magnesite and ankerite-dolomite, as a geothermo-meter. Figure 22 is a reproduction of their Figure 15, showing the relationship between (25) and temperature for their compiled analyses of natural assemblages. Even though the natural assemblages formed at varying pressures, this does not explain such a large scatter in the natural data. The data from the Marra Mamba, which were included in the compilation of Anovitz and Essene (1987), show considerable scatter to 0.47). The low pressures postulated for these rocks are not likely to have a signifi-cant effect on the degree of solid-solution (Barron, 1974; Harker and Tuttle, 1955). This scatter indicates that partitioning of iron and magnesium between the two carbonates of the Marra Mamba is more a function of the composition of the precipitating solution than of the temperature and pressure. Figure 22 shows the relationship between and temperature (solid and dashed lines) calculated b y Anovitz and Essene (1987 Figure 15) They used their own data, as well as that o f Rosenberg (1967) and

PAGE 112

'0 CD Cii ...... ...... G) c La. < 0 0.8 0 6 0 4 0.2 / 0 / / Temperature C / / / 800 93 FIGURE 22. Relationship between the Kn value of Anovitz and Essene (1987 defined in the text) and the temperature, for natural assemblages of coexisting dolomite-ankerites and magnesitesiderites. The points represent the values for these assemblages including the data from Klein and Gale (1981). The calculated values from the data of Anovitz and Essene (1987), Goldsmith et al. (1962), and Rosenberg (1967) are shown with the dashed-and solid lines.

PAGE 113

94 Goldsmith et al. (1962). Extrapolation of these curves down to 25C yields negative values for which are obviously meaningless. It appears that the high temperature phase relations are not a very good guide to the relationship between temperature and Fe/Mg partitioning at Earth's surface. Also, at temperatures greater than 250C their data suggest that the tie-lines between siderite-magnesite and ankeritedolomite, will steepen with increasing temperature. The data in Table 12 calculated for the Marra Mamba suggest just the opposite. The steepest set of tie-lines is believed to represent equilibrium between the two carbonate solid-solutions at 25C and it is unlikely that the other sets of tie-lines represent temperatures less than 25C. (There may be a problem with this choice of tie-lines to represent equilibrium at 25C (R N. Strom, pers. comm.), but, quantitatively, it makes only a small difference to the final results calculated by the model.) Anovitz and Essene (1987), in fact, warn against using their model at temperatures below 250-300C, since they did not fit low-temperature data. Diagenesis and metamorphism of the iron-formations is widely believed to have been virtually isochemical (except for H 2 0 and co2), on a district-wide scale, although some exchange of elements may have occurred between bands (Ewers and Morris, 1981; Klein, 1983; Perry et al. 1973). Ayres (1972, p. 1 227) makes this statement about the Dales Gorge Member of the Hamersley Iron-formation; "There is no petrographic evidence to suggest that large volumes of any of these constituents [i.e., silicon, iron, aluminum, calcium, magnesium or oxygen] have been added since deposition." Also Floran and Papike (1975) point out, It should be emphasized that many of the carbonate replacement textures are of local origin and due, in part to recrystallization.

PAGE 114

95 Although the textures indicate replacement, little or no evidence exists in many rocks t o s uggest that carbonate was introduced metasomatically." Even oxygen is not believed to have been very mobile. Klein (1983) notes, that the ferric to ferrous ratios in Precambrian iron-formations, do not change very much from very low grades to higher grades of metamorphism. James and Trendall (1982) commented; "The sheer bulk of banded iron-formations laid down in the basin makes it unlikely that gross chemical modification has taken place." Becker and Clayton (1972) argue against post-depositional exchange with ground water having affected the oxygen and carbon isotope ratios in the carbonates of the Brockman Iron-formation. They argue that (p.585): vertical variations of 2-5 per mil exist over distances of less than 1 em, between adjacent mesoband s even though lateral variations in these same mesobands, over distances of kilometers, are fractions of one per mil. Such variations would be expected to have been wiped out if large scale exchange had occurred. Similarly, the lack of consistent differences between coexisting carbonate indicates that exchange between them has not occurred to any great extent, as might be expected if a major amount of exchange with a fluid phase has taken place . These points make it seem unlikely that significant exchange of carbon has occurred in the iron-formation since its burial at depth. The Marra Mamba Iron-formation is stratigraphically only a few hundred meters below the Brockman Iron-formation, so this same argument can be applied to the exchange of iron and magnesium between the carbonates of the Marra Mamba. Talantsev and Sazonov (1979) point out that the dolomite-ankerite and magnesite-siderite lattices are very stable. They do not easily

PAGE 115

96 transform during post-crystallization alteration, by comparison to those of aragonite and calcite. Klein (1974) notes that the compositions of both the fine-grained, and the coarse-grained, siderite-magnesites in the Sokoman Iron-formation are essentially the same. He considers the coarse-grained siderite-magnesite to be a recrystallization product of earlier, much finer-grained, material of sideritic composition. There is abundant evidence indicating that recrystallization of the carbonates occurred early in diagenesis. Dimroth and Chauvel (1973) believe that the originally precipitated carbonates of the Sokoman Iron-formations, all crystallized and recrystallized during early diagenesis. Therefore, by analogy, many iron-formations showing fine laminations may have done the same. Ewers and Morris (1981) described the nature of the fracturing in the Dales Gorge Member of the Brockman Iron-formation. They believe it indicates that crystallization from the original precipitate occurred very early in the diagenetic process. French (1973) believes chat crystallization of the primary minerals took place near the sediment-water interface. Experiments conducted by Harder and Flehmig (1970) revealed that, even at cemperatures below 80C, quartz can form in a matter of a few weeks from X-ray amorphous, hydroxide-silica precipitates. This indicates that the initial assemblage could have crystallized instantaneously, relative to geologic time.

PAGE 116

97 DISCUSSION Details of the conditions and processes of precipitation of banded iron-formations are highly controversial (Garrels, 1987) An analysis of the various proposed mechanisms is beyond the scope of this thesis. A model similar to that developed by Garrels (1987) is envisioned to explain the precipitation of carbonates of the Marra Mamba in discrete layers. Garrels (1987) postulated influx into a large, low-energy basin of a solution similar to modern groundwaters found in equilibrium with basaltic rocks at earth-surface conditions. Saturation due to evaporation is the mechanism proposed to initiate precipitation of the carbonates. He showed that an annual cycle of evaporation and recharge of this solution can account for the observed thickness (0. 15 mm per annual pair) and purity of the alternating bands of chert and siderite. Garrels (1987) proposed an ambient temperature of 40C and a partial pressure of co2 in the atmosphere of approximately 1 bar. These two physical factors enhance separation of the amorphous silica and siderite when they precipitate as saturation is achieved. He postulates evaporation down to one fifth of the original volume input each y e a r to account for the thickness of chert and siderite observed. The large amounts of iron in solution (Tables 20 and 21) required by the model are made possible by two conditions. First, the increased co2 content of the atmosphere enhances siderite solubility, and second, a much lower atmospheric content of oxygen enhances maintenance of iron in the more

PAGE 117

98 soluble divalent state. Higher temperature permits separation of amorphous silica and siderite, due to their differing solubilities, and enhances the degree of evaporation. The calculations for this research were done assuming 25C in an effort to keep things as simple as possible. However, an increase in temperature to 40C would have only a minor effect on the calculated ion activity ratios because the rate of change with temperature of the equilibrium constants of the carbonates is very similar. Figure 23 shows the composite paragenetic and diagenetic sequence postulated for the Marra Mamba Iron-formation. This sequence was compiled from this research, and from that of the papers discussed in the previous section [particularly French (1973) ; Dimroth and Chauvel ( 1973) ; Klein and Gole (1981)]. It is proposed that siderite-magnesite initially precipitated from the solution as a crystalline phase during formation of the Lower Marra Mamba. The origin of the ankerite-dolomite is more problematical. As noted previously, preliminary experiments in our laboratory indicate that siderite precipitates readily from aqueous solutions, whereas the difficulty of precipitating a dolomitic phase is well-known (Althoff, 1977; Halla et al., 1962; Morrow and Ricketts, 1986; Morse, 1983; Ricketts, 1980). A few samples of ankerite-dolomite exhibit what appear to be primary textures and may, in fact, be primary precipitates. These are not very common, however The more likely origin for the dolomite-ankerite involves replacement or recrystallization of an initially precipitated phase. There are a number of viable mechanisms for generating the abundant ankeritedolomite found in these rocks. First, in cases where ankerite-dolomite

PAGE 118

ORIGINAL MATERIAL P < 100 bars -T< 40 C u H E-1 H ll:l
PAGE 119

100 appears to be a replacement of siderite-magnesite, it probably formed as an early replacement, soon after deposition of the siderite-magnesite (which may have precipitated first, even though dolomite-ankerite was also stable). Once siderite-magnesite formed, dolomite-ankerite could have replaced it immediately by a dissolution-reprecipitation mechanism. This all could have occurred while the phases were still in contact with the original depositional fluid. Second, although pure dolomite seems notoriously difficult to precipitate, it is conceivable that an ironrich ankerite-dolomite may precipitate more readily and that such a mineral could have precipitated initially and then recrystallized immediately thereafter into a two-carbonate assemblage. Third, the originally precipitated minerals may have been calcite and a magnesitesiderite. The calcium required for the formation of dolomite-ankerite could, thereby, have been incorporated into the sediment instead of remaining in the solution. Then, by a reaction such as the following, the two-carbonate assemblage could have recrystallized into dolomiteankerite: Caco 3 (s) + (Mg,Fe)C0 3 (s) = Ca(Mg,Fe)(C0 3 ) 2 (s) During diagenesis, the two carbonates may have recrystallized into coarse, interlocking, euhedral grains, without further chang e in their Fe/Mg ratios. Recrystallization of nearby hydrous amorphous silica would have released considerable amounts of water, thereby enhancing dissolution-reprecipitation of the carbonate phases to a stable twocarbonate assemblage. The original precipitate in iron-formations is believed by some (Ewers and Morris, 1981) to have contained up to 85 % water. These reactions probably occurred at less than 150C, near to

PAGE 120

101 the sediment-water interface, in equilibrium with the depositional fluid. It is not believed that the low grades of metamorphism observed in these rocks caused any significant alteration of the Fe/Mg ratios in the carbonates. Figure 15 compares the calculated tie-lines of Anovitz and Essene (1987) for 250C (bold solid lines) with the actual tie-lines [Klein and Gole (1981), Set III, dashed lines), and with those calculated by this method for 25C (light solid lines). The tie-lines for 250C have a shallower slope than those determined by these calculations to represent 25C, but the slopes of the tie-lines approach each other for the most iron-rich compositions. At higher temperatures a more iron-rich siderite-magnesite is in equilibrium with a given dolomite-ankerite than at lower temperatures. This comparison shows, that if the compositions of the carbonates were to alter due to higher temperatures during metamorphism, their iron contents could shift by up to about 30 mol % for the more magnesium-rich assemblages. Klein and Fink (1976, p.486) describe the formation of the early assemblages of the Sokoman Iron-formation in this way; "It is noteworthy that the assemblages of this study appear to have reached a state of equilibrium as deduced from the generally consistent major elements fractionation pattern among the coexisting phases. The period over which lithogenesis took place must have been relatively long and gradual and many, and complete chemical reactions must have taken place (close to equilibrium conditions) between the more solid sedimentary gels (silicates and iron-hydroxides) and the finely crystalline carbonate and iron oxide precipitates and the interstitial water-rich solutions." Similar assemblages, and consistent fractionation patterns, are observed

PAGE 121

for the Marra so equilibrium is also assumed to have been achieved there during diagenesis. 102 As in the Marra Mamba, the calcite-bearing assemblages occur late in the deposition of many iron-formations. They are sometimes stratigraphically separated from the siderite-magnesite bearing rocks (Klein and Gole, 1981) It appears that following a hiatus in iron-formation deposition, a new set of chemical, and possibly physical, conditions prevailed in some of the basins. After this hiatus, calcite/dolomiteankerite assemblages precipitated from an entirely different solution than that which deposited the dolomite-ankerite/magnesite-siderite assemblages (Garrels, 1987) The "double-solid-solution" assemblage in the lower portion of the Marra Mamba is a more precise indicator of the compositional variation of the depositing solution than the calcite/dolomite-ankerite assemblage above it. This is because the iron and magnesium contents of the two solid-solutions vary more widely than do those of calcite. The increased variation mean s that a small error in the determination of the composition of the solid phase has a less significant effect on the composition determined for the aqueous solution. In contrast with this, minor iron in calcite can cause a large error in the configuration of the tie-lines and, thus, in the calculated solution composition. Crystals of the size of the initial siderite precipitate (5-35 microns) show only an insignificant effect of surface energy on the free energy of formation of the mineral. A grain size of the initial precipitate of around one micron is usually considered the maximum size which can still generate sufficient surface free energy to affect the free energy of formation of the mineral (Stumm a nd Morgan, 1981) Therefore,

PAGE 122

103 it is reasonable to consider the free energy of formation of this material to be indistinguishable from the value for well-crystallized, macrocrystalline siderite used in this model. Holland (1973) estimated the iron content of the ocean during the time of deposition of the Hamersley Basin iron deposits to be about 3 ppm for a pH of the ancient ocean of 7.0, and about 30 ppm for a pH of 8.0. He assumed that seawater was saturated with respect to siderite and calcite, that the calcium content of the ocean was the same as 2+ today, and that Fe was reasonably strongly complexed by OH-. Con-sidering that the co2 content of the ancient atmosphere was undoubtedly significantly larger than at present (Holland, 1984), a lower pH than at the present time is likely. Garrels (1987), for example, postulates a pH of about 6. These iron contents are calculated assuming a substan-tially lower content of oxygen in the atmosphere. This permits siderite to remain unoxidized in the depths of the ocean. 2 + So, assuming a concentration of 3 ppm Fe (as opposed to the 3+ present value of about 3 ppb Fe ) for the aqueous iron content of the 2 + ancient ocean and a total ion activity coefficient of 0 2 for Fe (Whitfield, 1975), yields values for the logarithms of the ion activity 2 + 2+ 2 + 2 + ratios of Fe /Ca and Hg /Fe of 2 3 and +3.1, respectively. Examination of Figure 20 reveals that these compositions do not fall within the range of the solutions predicted to have deposited the carbonates of the lower portion of the Marra Hamba Iron-formation. The data do not preclude P roterozoic seawater as the depositional medium for the Upper Harra Mamba. The Upper Harra Mamba is characterized by abundant iron oxide minerals. The Lower Marra Mamba, however, commonly has significant iron sulfides. Perhaps, the major difference between

PAGE 123

104 the depositional environments of the upper and lower portions of the Marra Mamba lvas the operation of some oxidative mechanism during "Upper2+ Marra-Hamba-time" that significantly decreased the Fe content of the depositing solution. Alternatively, a change in local topography could have led t o the influx of an entirely different solution following deposition of the Lower Marra Mamba. A more likely source of the depositing solution is river water, or perhaps, a ground water. The composition of a likely water, one in equilibrium with an average basalt, is given in Table 20. Modern "basalt-water" analyses contain virtually no iron, because under a modern, oxygenated atmosphere aqueous iron is oxidized to Fe 3+. Fe 3 + TABLE 20 Chemical analysis of the average basalt and ultramafic groundwater. From \fuite, Hem, and Waring (1963). Constituent Parts per million Holality or mg/liter 40.7 0.000678 Ca2 26.3 0 00065 7 Mg2+ 26.3 0 00108 Na+ 1 8.4 0 00080 K+ 2.9 0 0000 74 Cl 23.0 0 00065 so2 4 17 3 0.00018 HC03 205.2 0.003364 Totals 360 1 precipitates almost instantaneously as ferric hydroxide. Equivalents per liter 0.00131 0.00216 0 00080 0.000074 0.00065 0 00036 0 003364 0.00872 In the low-oxygen atmosphere of the Proterozoic, aqueou s iron was carried in solution in the divalent state and would have behaved much like divalent calcium and magnesium Garrels (198 7) adjusted the

PAGE 124

105 composition in Table 20 t o contain enough iron and s ilica to yield the typical annual thickness of iron minerals and chert deposited in banded iron-formations. The ferrous iron was balanced by bicarbonate. The resultant iron concentration lies halfway between calcium and magnesium which is in keeping wit h t he mole proportions of calcium, magnesium and iron in the average basalt. (A typical basalt has a mole ratio of iron t o calcium almost equal to one. MgO is usually 1/3 to 1/2 greater than CaO or FeO.) The new "Proterozoic iron-basin feed-water" from Garrels (1987) is given in Table 21. The logarithm of the resultant activities 2+ 2+ 2+ 2+ 2+ 2+ of Mg /Ca Fe /Ca and Mg /Fe are 0.34, 0.21, and 0.13, respec-tively. Comparison of these values with Figures 17 and 20 indicates that a solution similar to a "basalt-water" is a more likely medium of TABLE 21. Proposed Proterozoic iron-basin feed-water from Garrels (1987). Constituent Parts or Ca2 Mg2+ Fe 2 + Na+ K+ -Cl HC03 Totals Ionic s trength 0 00743 per million mg/liter 63.0 18.0 24.3 41.0 18. 7 3 0 35 5 259 1 462.6 Molality 0.00105 0.00045 0.0010 0.000735 0.00080 0.000077 0. 0010 0.004247 Equivalents per liter 0.00090 0 0020 0 00147 0 00080 0 000077 0 0010 0 004245 0.010492 Ca-MgFe carbonate precipitation, because it contains much more iron than does seawater.

PAGE 125

106 From Figure 17 it appears that the Mg/Ca ratio (as well as the Fe/Ca ratio) of modern seawater is too high for the equilibrium deposition of the Upper Marra l>lamba. It seems that a much higher ratio of Mg/Ca than that determined for equilibrium is required to precipitate dolomite. Even the seawater value, which is much greater than the critical ratio as estimated from either laboratory determinations of dolomite solubility or groundwater equilibrium data, is insufficient to foster precipitation of dolomite at 25C (Blatt et al. 1980). Unlike the Lower Marra Mamba, no siderite-magnesite is found in the Upper Marra I have suggested, therefore, that the precipitation of sideritemagnesite may be a neccesary precursor to the deposition of Mg-rich dolomite-ankerite and a "higher-than-equilibrium" value of Mg/Ca may have been required to precipitate dolomite-ankerite. It is also possible that the ratio of Mg/Ca was higher in the ancient ocean than it is in the modern ocean (Chilingar, 1956). (Although, I assumed that the modern ratio applies to the ancient ocean for the figures in this paper.) Perhaps, the increased proportion of mafic rocks in the Proterozoic led to an increased value of Mg/Ca. The method which has been described in this paper can be applied to other minerals. For example, the system Caco 3 -MgC03 -HnC03 describes a common mineral assemblage found in ore deposits such as the Creede, Colorado Pb-Zn-Ag epithermal veins al., 1977) or the Ryujima Mine of Japan (Tsusue, 1967). In these deposits veins containing magnesite-siderite, rhodochrosite, kutnahorite and/or calcite are common. For the Ryujima Mine compositional data are available for coexisting carbonates along the joins dolomite-kutnahorite and magnesite-rhodochrosite. Such information used in conjunction with the

PAGE 126

107 appropriate equations, such as chose developed here for the system Caco 3-MgC03 -FeC0 3 could yield valuable information on the composition of the ore-forming fluid. Similarly, the system CaSi03 -MgSi0 3-FeSi03 characterized by the pyroxenoid, wollastonite; the clinopyroxenes, diopside-hedenbergite; and the orthopyroxenes, enstatite-ferrosilite can be treated with the appropriately-formulated equations to constrain the composition of the equilibrating solution. The amphiboles, tremoliteactinolite and anthophyllite, could probably be treated in the same way The equations could possibly be adapted to a sulfide system such as CuS-AsS-SbS in which an intermediate solid-solution [cennantite (Cu 12As4 s 13)-tetrahedrite (Cu 12sb4 s 1 3)] exists.

PAGE 127

108 CONCLUSIONS Comparison of real and calculated tie-lines indicates that the proposed model adequately describes the low-temperature phase relations of the Ca-l1g-Fe carbonates. This is especially true for the more magnesium-rich compositions found in the Marra Mamba. The equation representing equilibrium at 25C between calcite and the dolomite-ankerite solid-solution series is [Mg2+]0.5 = K dol/cal [Fe2+]0.5 K -----------dol/ank (Ca2+10.5 The equation representing equilibrium at 25C between dolomite-ankerite and magnesite-siderite is 0 2+ 0.59(Fe ] -10 ------(ca2+J (Fe2+1o.s + ---------[Ca2+10.5 2+ -1. 76(Mg ] 10 ------+ (Ca 2+] [ M 2+]0.5 -0.94 g 10 --------[Ca2+]0.5 The free energy of formation of the hypothetical, pure iron end-member of the ankerite-dolomite solid-solution series has been well-constrained. The free energy of formation of ankerite is equal to -907.6 kJ/mol. The model permits comparison of the solutions from which the carbonates of the Upper and Lower Marra Mamba Iron-formation precipi-tated. This comparison indicates that the solution depositing the Lower Marra Mamba had much higher iron to calcium and magnesium to calcium ratios than did that which deposited the Upper Marra Mamba. For the carbonates of the Upper Marra Mamba, the logarithm of the ion activity

PAGE 128

109 2+ 2+ ratio in the depositing solution ranged from -1.2 to 2 0 for Mg /Ca 2+ 2+ 2+ 2+ from -3.35 to -4.25 for Fe /Ca and from 2.8 to 1 4 for Mg /Fe For the Lower Marra Mamba, the logarithm of the ion activity ratios ranged from 1.6 to 0.2, from -1.2 to -0.9, and from 3.0 to 1.25, for 2+ 2+ 2+ 2+ 2+ 2+ Mg /Ca Fe /Ca and Mg /Fe respectively. A particular ankerite-dolomite in the Upper Marra Mamba would be in equilibrium with 2+ 2+ 2+ 2+ a solution having lower Fe /Ca and Mg /Ca ratios than the same ankerite-dolomite in the Lower Marra Mamba. Therefore, it is essential to know the nature of the entire carbonate assemblage to use the method. The evidence presented indicates that the compositions of these two carbonate solid-solutions are much more responsive to changes in solu-tion composition than to temperature. Therefore, they should be good indicators of the composition of the depositing aqueous solution. It does not appear that diagenesis of these phases involved reequilibration of the initial minerals with a solution other than the depositing solution that was buried with them although recrystallization of the original phase did occur. It also does not appear that the compositions of the carbonates varied significantly with low degrees of metamorphism. The model developed is quite sensitive to the choice of thermo-dynamic values used in the calculations. However, it is believed that the other available sets of thermodynamic data yield unlikely phase diagrams for the carbonate system. The results are also fairly sensitive to the activity-composition model used to describe the behavior of the two carbonate solid-solutions. Therefore, a knowledge of the activity coefficients of the end-members in the solid-solutions would significantly improve the results.

PAGE 129

llO Preliminary calculations using the method described in this paper indicate that neither present-day, nor Proterozoic, seawater is likely to have deposited the banded iron-formations of the lower portion of the Marra Mamba. However, a Proterozoic seawater may have contained enough iron to precipitate the iron-formations of the Upper Marra Mamba. A solution resembling a modern river water in equilibrium with a basalt is a more likely depositional medium for these deposits. The system modelled here is a very simple one involving equilibrium precipitation of only the Ca-Mg-Fe carbonates from the solution at any given time. This treatment of the deposition of banded iron-formations is justified by two observations. First, Klein and Gole (1981) state that the carbonates of the Marra Mamba often form relatively continuous, thin bands, between bands of magnetite and silicates. Second, the calculations of Garrels (1987) show that the precipitation of the carbonate and the silica phases are separated in time due to differing solubilities. The model becomes much more complicated if the precipitation of an iron-bearing silicate is occurring at the same time as that of the iron-bearing carbonates. Consideration of the effects of Fe/Mg partitioning between silicate and carbonate phases is bey ond the scope of this paper. When adequate compositional and thermodynamic data are available for coexisting solid-solutions, many ion activity ratios for the depositing solution can be constrained. Such data permit description of the physical and chemical conditions of formation of minerals that can not be empirically determined. This method could be applied to mineral assemblages including other carbonates, pyroxenes, amphiboles, sulfides, and many other phases.

PAGE 130

REFERENCES Alexandrov, E. A., 1973, The Precambrian banded iron-formation of the Soviet Union: Econ. Geol., v. 68, p. 1035-1063. 111 Althoff, P. L., 1977, Structural refinements of dolomit e and a magnesian calcite and implications for dolomite formation in the marine environment : Amer. Mineral., v. 62, p. 772-783 Anovi tz, L M., and Essene, E. J., 1987, Phase equilibria in the system Caco 3-MgC03 -FeC03 : Jour. of Pe trology, v. 28, p 389 414 Ayres, D. E., 1972, Genesis of iron-bearing minerals in banded ironformation mesobands in The Dales Gorge Member, Hamersley Group, Western A ustralia: Econ. Geol., v. 67, p 1214-1233. Babushkin, V. I., Matveyev, G. M., and Mched lov-Petrossyan, 0 P., 1985, Thermod ynamics of Silicates: Berlin, Springer-Verlag, 459p. Barron, B. J 1974, The use of coexisting calcite-ankerite solid solutions as a geothermometer: Contrib. }1ineral. Petrol., v 47, p 77-80. Barton, P. B., Jr. Bethke, P M and Roedder, E 1977, Environment of ore deposition in the Creede mining district, San Juan Mountains, Colorado: Part III. Progress toward the interpretation of the chemistry of the ore-forming fluid for the OH vein: Econ. Geol., v. 72, p. 1 24 Bayley, R. W and James, H L., 1973, Precambrian iron-formations of the United States: Econ. Geol., v. 68, p. 934-959. Becker, R H., and Clayton, R N., 1972, Carbon isotopic evidence for the origin of a banded iron-formation in Western Australia: Geochim. et Cosmochim Acta, v. 36, p. 5 77-595. Becker, R H and Clayton, R. N 1976, Oxygen isotope study of Precambrian banded iron-formation, Hamersley Range, Western Australia: Geochim. et Cosmochim. Acta, v. 40, p 115 3-1165. Berner, R A 1971, Principles of Chemical Sedimentology: New York, McGraw-Hill Book Co. 240p. Beukes, N. J., 1983, Palaeoenvironmental setting of iron-formations in the depositional basin of the Transvaal Supergroup, South Africa,

PAGE 131

112 in Trendall, A F., and Morris, R. C., eds. Iron-formation; Facts and Problems: Developments in Precambrian Geology #6, Amsterdam, Elsevier, p 131-209. Blatt, H Middleton, G., and Murray, Raymond, 1980, Origin of Sedimentary Rocks: Englewood Cliffs, New Jersey, Prentice-Hall, Inc., 782p. Byrne, R. H and Kester, D. R., 1976, Solubility of hydrous ferric oxide and iron speciation in Marine Chemistry, v. 4, p. 255-274. Chang, L. L Y and Brice, W. R., 1971, Subsolidus phase relations in aragonite-type carbonates: .II. The systems Caco3-srco3 -PbC0 3 and Caco 3 -BaC0 3 -PbC03 : Amer. M1neral., v 57, p. 155-168. Chilingar, G. V., 1956, Relationship between Ca/Mg ratio and geologic age: Bull. Amer. Assoc. Petrol. Geologists, v. 40, p. 2256 2266. Christ, C L Hostetler, P. B and Siebert, R M., 1974, Stabilities of calcite and aragonite: Jour. Research U. S. Geol. Survey, v. 2, p. 175-184. Davies, F. G 1979, Thickness and thermal history of continental crust and root zones: Earth Planet. Sci. Lett. v 44, p. 231-238 deCapitani, C., and Peters, T., 1981, The solvus i n the system Mnco3 Caco3: Contrib. Mineral. Petrology, v. 76, p 394-400. Dimroth, E., 1968, Sedimentary textures, diagenesis, and sedimentary environment of certain Precambrian ironstones: Neues Jahrb. Geologie u. Palaontologie Abh., v. 130, p 247-274. Dimroth, E., and Chauvel, J. J., 1973, Petrography of the Sokoman ironformation in part of the Central Labrador Trough, Quebec, Canada: Geol. Soc. Americ a Bull. v 84, p. 111-134. Dorr, J V. N 1973, Iron-formation of South America: Econ. Geol., v 68, p. 1005-1023 Ernst, W. G., 1983, The early Earth and the Archean rock record, in Schopf, J W., ed., Earth's Earliest Biosphere: Princeton, New Jersey, Princeton University Press, p 41-52. Essene, E J., 1983, Solid solutions and solvi among metamorphic carbonates with applications to geologic thermobarometry, in Reeder, R J., ed., Carbonates: Mineralogy and Chemistry: Reviews in Mineralogy, v 11, Washington, D.C., Mineral. Soc Amer., p .77-96. Eugster, H. P., and Chou, I-M, 1973, The depositional environments of Precambrian banded iron-formations: Econ. Geol v. 68, p. 1144-1168

PAGE 132

113 Ewers, W. E., and Morris, R C., 1981, Studies of the Dales Gorge Member of the Brockman iron-formation, Western Australia, Econ Geol., v. 76, p 1929-1953. Fanelli, M. R., Cava, N., and Wyllie, P. J., 1983, Calcite and dolomite without portlandite at a new eutectic in Ca0 -MgO-co2 -H 20, with applications to carbonatites: Int'l Mineral. Assoc., Proc. 13th General Meeting, Varna, Bulgaria, 1982, (in press). Floran, R. J., and Papike, J J., 1975 Petrology of the low-grade rocks of the Gunflint iron-formation, Ontario-Minnesota: Geol. Soc. America Bull., v. 86, p. 1169-1190. French, B. M., 1973, Mineral assemblages in diagenetic and low-grade metamorphic iron-formation: Econ Geol., v. 68 p. 1063 1074 Garrels, R. M., 1987, A model for the deposition of the microbanded Precambrian iron-formations: Am. Jour. Sci., v. 287 p 81-106. Garrels, R. M., and Christ, C L., 1965, Solutions, Minerals, and Equilibria: San Francisco, Freeman, Cooper and Company, 450p. Garrels, R. M Thompson, M. E., a nd Siever, R., 1960, Stability of some carbonates at 25C and 1 atmosphere total pressure: Amer. Jour. Sci., v 258, p. 402-418. Gibbs, J. W., 1 876 and 1878, On the equilibriu m of heterogeneous substances: Conn. Acad Trans., III, 108 -248, 343-524. Goldsmith, J. R., 1959, Some aspects of the geochemistry of carbonates, in Abelson, P. H., ed., Researches in Geochemistry: New York, John Wiley and Sons, p. 336-358. Goldsmith, J. R., 1983, Phase relations of rhombohedral carbonates, in Reeder, R. J., ed., Carbonates: Mine ralogy and Chemistry: Reviews in Mineralogy, v. 11, Washington, D C., Mineral. Soc. Amer., p 49 -76. Goldsmith, J R., and Graf, D. L., 1957, The system solid solution and decomposition relations: Geoc him. et Cosmoch1m Acta, v. 11, p. 310-334. Goldsmith, J. R., and Graf, D L., 1958, Structural and compositional variations in some natural dolomites: Jour. Geology, v 66, p. 678 -693. Goldsmith, J. R., and Graf, D. L., 1960, Subsolidus relations in the system Caco 3 -Mgco3 }mco 3 : Jour. Geology, v 68, p. 324 335 Goldsmith, J. R., and Heard, H. C., 1961, Subsolidus phase relations in the system Caco 3 -MgC03 : Jour. Geology v. 69, p. 45-74. Goldsmith, J. R., and Northrop, D., 1965, Subsolidus phase relations in

PAGE 133

114 the system CaC03-MgC03 -coco3 and Caco 3 -Ngco 3 NiC0 3 : Jour. Geology, v 73, p. 817-829. Goldsmith, J. R., Graf, D. L., Witters, J., and Northrop, D., 1962, Studies in the system Caco 3 -Mgco 3-Feco3 : 1. Phase 2 A method for major-element spectrochemical analysis; 3 Compositions of some ferroan dolomites: Jour. Geology v. 70, p. 659-688. Gole, J., and Klein, C., 1981, Banded through much of Precambrian time: Jour. Geology, v. 89, p. 169 183 Goodwin, A.M., 1973, Archean iron-formations and tectonic basins of the Canadian Shield: Econ. Geol., v 68, p. 915-934. Govett, G. J. S., 1968, Origin of banded iron-formations: Geol. Soc America Bull., v 77, p. 1191-1211. Graf, D L., and Goldsmith, J. R., 1955, Dolomite-magnesian calcite relations at elevated temperatures and co2 pressures: Geochim. et Cosmochim. Acta, v. 7, p 109-128. Grubb, P. L. C. 1971, Silicates and their paragenesis in the Brockman Iron-formation of Wittenoom Gorge, Western Australia: Econ. Geol., v. 66, p. 281 292. Hull, W D. M., and Goode, A. D T., 1978, The early Proterozoic Nabberu Basin and associated iron-formations of Western Australia: Precambrian Research, v. 7, p. 129-184. Halla, F., Chilingar, G V., and Bissell, H. J., 1962, Thermodynamic studies on dolomite formation and their geologic implications: An interim report: Sedimentology, v 1, p 296 -303. Harder, H., 1965, Experimente zur "Ausfallung" der Kieselsaure, Geochim. et Cosmochim Acta, v. 29, p. 429-442. Harder, H., and Flehmig, W., 1970, Quarzsynthese bei tiefen Temperaturen: Geochim et Cosmochim. Acta, v 34, p 295-305 Harker, R. I., and Tuttle, 0. F., 1955, Studies in the system Ca0-Mg0co2, II. Limits of solid solutions along the binary join Caco 3 -Mgco3: Am. Jour. Sci., v. 253, p 274-282 Harvie, C E., Moller, N., and Weare, J. H., 1984, The prediction of mineral solubilities in natural waters: The NaK-Mg-Ca-H-Cl -so4 0H-HC03-co1-co2-H2o system to high ionic strengths at 25C : Geochim. et Cosmochim. Acta, v. 48, p 723 -751. Helgeson, H. C., 1983 and 1984, Supcrt Update Notices, Univ. of California at Berkeley. Helgeson, H. C., Delany, J. M., Nesbitt, W. H., and Bird, D. K., 1978, Summary and critique of the thermodynamic properties of

PAGE 134

rock-forming minerals: Am. Jour. Sci., v. 278-A, 229p. Holland, H. D., 1973, The oceans: A possible source of iron in ironformations: Econ. Geol., v. 68, p. 1169-1173. Holland, H. D., 1984, The Chemical Evolution of the Atmosphere and Oceans: Princeton, N. J., Princeton University Press, 582p. Hurlbut, C. S., 1971, Dana's Hanual of Hineralogy: 18th edition, New York, John Wiley and Sons, 579p. 115 James, H. L., 1954, Sedimentary facies of iron-formation: Econ. Geol., 49, p. 235-293. James, H. L., 1966, Chemistry of the iron-rich sedimentary rocks: U. S. Geol. Survey Prof. Paper 440-W, 6lp. James, H. L., and Trendall, A. F., 1982, Banded iron-formation: distribution in time and paleoenvironmental significance, in Holland, H. D., and Schidlowski, H. eds., Mineral Deposits and the Evolution o f the Biosphere: New York, Springer-Verlag, p. 199-218. Karpov, I. K., Kiselev, A. I., and Letnikov, F. A., 1971, Chemical Thermodynamics in Petrology and Geochemistry (In Russian): Irkutsk, Akademia Nauka, 385p. Klein, C., 1974, Greenalite, stilpnomelane, minnesotaite, crocidolite and carbonates in a very low-grade metamorphic Precambrian ironformation: Can. Mineral., v. 12, p. 475-498. Klein, C., 198 3 Diagenesis and metamorphism of Precambrian banded ironformations, in Trendall, A. F., and Morris, R. C., eds., Ironformation; Facts a nd Problems: Developments in Precambrian Geology #6, Amsterdam, Elsevier, p. 417-469. Klein, C., and Fink, R. P., 1976 Petrology of the Sokoman ironformation in the Howells River area, at the western edge of the Labrador Trough: Econ. Geol., v. 71, p 453-487. Klein, C., and Gale, M. J., 1981, Mineralogy and petrology of parts of the Marra Mamba iron-formation, Hamer sley Basin, Western Australia: Amer. Mineral., v. 66 p. 507-525. Krumbein, W. C., and Garrels, R. M., 1952, Origin and classification of chemical sediments in terms of pH a nd oxidation-reduction potentials: Jour. Geology, v. 60, p. 1-33. LaBerge, G., 1964, Development of magnetite in iron-formations of the Lake Superior Region: Econ. Geol., v. 59, p. 1313-1342. Latimer, w. M., 1952, Oxidation Potentials: New York, Prentice-Hall Inc., 392p

PAGE 135

116 Lippman, F., 1982, Stable and metastable solubility diagrams for the system Caco 3-MgC03 -H 2 o at ordinary temperatures: Bull. Mineral. v. 105, p. 273-279. Mackenzie, F. T., Bischoff, W. D., Bishop, F. C., Loijens, M Schoonmake r J and Wollast, R 1983, Magnesian Calcites: Lowtemperature occurrence, solubility and solid solution behavior, in Reeder, R J., ed., Carbonates: Mineralogy and Chemistry: Reviews in Mineralogy, v. 11, Washington, D.C., Mineral. Soc Amer. p.97-144. Maynard, J. B., 1983, Geochemistry of Sedimentary Ore Deposits: New York, Springer-Verlag, 305p. Mel'nik, Y P., 1972, Thermodynamic Constants for the Analysis of Conditions of Formation of Iron Ores (In Russian): Kiev, Naukova Dumka, 196p. Moore, W J., 1964, Physical Chemistry: Englewood Cliffs, N.J., Prentice-Hall, Inc., 844p. Morrow, D. W., and Ricketts, B. D., 1986, Chemical controls on the precipitation of mineral analogues of dolomite: The sulfate enigma: Geology, v. 14, p 408-410. Morse J. W., 1983 The kinetics of calcium carbonate dissolution and precipitation, in Reeder, R. J., ed., Carbonates: Mineralogy and Chemistry: Reviews in Mineralogy, v. 11, Washington, D C., Mineral. Soc Amer., p 227-264. Naumov, G B Ryzhenko, B. N., and Khodakovsky, I L., 1974, Handbook of Thermodynamic Data: Nat'l. Tech. Inf. Service, Pb-226, 722/7GA, U. S. Dept. Commerce, 328p Nordstrom, D. K., and Munoz, J. L., 1986, Geochemical Thermodynamics: Palo Alto, California, Blackwell Scientific Publications, 477p. Perry, E C., Jr., Tan, F C., and Morey, G. B 1973, Geology and s table isotope geochemistry of the Biwabik iron-formation, Northern Minnesota: Econ Geol., v 68 p 1110-1125. Plummer, L N., and Busenberg, E., 1982, The solubilities of calcite, aragonite and vaterite in C02 -H 2 o solutions between 0 and 90C, and an evaluation of the aqueous model for the system Caco 3-co2 H 2o : Geochim et Cosmochim Acta, v. 46, p. 1011-1040. Reeder, R J., 1983, Crystal chemistry of the rhom b ohedral carbonates, in Reeder, R. J ed., Carbonates: Mineralogy and C hemistry: Reviews in Mineralogy, v 11, Washington, D.C., Mineral. Soc Amer., p 1-47 Ricketts, B D., 1980, Experimental investigation of carbonate precipitation in hydrated silica ge ls: Journ. of Sedim Petrol., v 50,

PAGE 136

117 p. 963-970. Robie, R. A., Hemingway, B. S., and Fisher, J R., 1978, Thermodynamic of minerals and related substances at 298.15K and 1 bar (10 Pascals) pressure and at higher temperatures: U. S. Geol. Survey Bull., v. 1452, 456p. Robie, R. A., Haselton, H. T., Jr., and Hemingway, B. S., 1984, Heat capacities and entropies of rhodochrosite (MnC03 ) and siderite (Feco3 ) between 5 and 600 K: Amer. Mineral., v 69, p. 349-357. Robinson, G. R., Jr., Haas, J. L., Jr., Schafer, c. M., and Haselton, H. T., Jr., 1982, Thermodynamic and thermophysical properties of selected phases in the Mg0-Si0 2 -H 2o-co2 CaO-Al 2 o 3-Si02 -H 2 0-C0 2 Fe-Fe0-Fe2 o 3-Si02 chemical systems, with special emphasis on tfie properties of basalts and their mineral components: U. S Geol. Survey Open-file Report 83-79, 429p. Rosenberg, P E., 1960, Subsolidus studies in the system Caco 3-MgC03 FeC03-Mnco3: Ph.D. thesis, Pennsylvania State Univ., University Park, Pennsylvania, 146p. Rosenberg, P. E., 1963, Subsolidus relations in the system Caco 3 -FeC0 3 : Am. Jour. Sci., v. 261, p. 683-690. Rosenberg, P. E., 1967, Subsolidus relations in the system CaC03-MgC03 FeC03 between 350C and 550C : Amer. Mineral., v 52, p. 787-796 Rosenberg, P. E., and Foit, F. F., 1979, The stability of transition metal dolomites in carbonate systems: a discussion: Geochim. et Cosmochim. Acta, v. 43, p. 951-955. Rossini, F. D., Wagman, D. D., Evans, W. H., Levine, S., and Jaffe, I., 1952, Selected Values of Chemical Thermodynamic Properties: National Bureau of Standards Circ. 500, U.S. Dept. Commerce, Washington, D.C. Sangameshwar, S. R., and Barnes, H. L., 1983, Supergene processes in zinc-lead-silver sulfide ores in carbonates: Econ. Geol., v. 78, p. 1379-1397. Shannon, R. D., and Prewitt, C. T., 1969, Effective ionic radii in oxides and fluorides: Acta Crystallographica (B), v 25, p. 925-946. Stumm, W., and Morgan, J J., 1981, Aquatic Chemistry: New York, John Wiley and Sons, Inc., 780p. Talantsev, A. S., and Sazonov, V. N., 1979, Variatzzii sostavov sosushchestvuyuschich dolomit-ankerit i magnezit-siderit kak pokazatel P-T-uslovii mineraloobrazovaniya: Akad Nauk SSSR Ural'skii Nauchaii Tzentr. 95-103

PAGE 137

Tardy, Y., and Garrels, R. M., 1974, A method of estimating the Gibbs energies of formation of layer silicates: Geochim. et Cosmochirn. Acta, v. 38, p. 1101-1116. 118 Thorstenson, D. C., and Plummer, L. N., 1977, Equilibrium criteria for two component solids reacting with fixed composition in an aqueous phase--example: the magnesian calcites: Amer. Jour. Sci., v. 277, p. 1203-1223. Trendall, A F., 1973, Precambrian Iron-formations of Australia: Econ. Geol., v 68, p. 1023-1034. Trendall, A. F., and Blackley, J. G., 1970, The iron-formations of the Precambrian Hamersley Group, Western Australia: Geol. Survey Western Australia Bull., v. 119, p. 3-346. Tsusue, A., 1967, Magnesian kutnahorite from Ryujima Mine, Japan: Amer. Mineral., v. 52, p. 1751-1761. Vegard, L., 1947, Investigations into the structure and properties of solid matter with the help of X-rays: Skrifter Utgitt av Det Norske Videnskaps-Akademi Oslo: I. Mat.-Naturw. Klasse, No. 2 Wagman, D. D., Evans, W. H., Parker, V. B., Schumm, R. H., Halow, I., Bailey, S. M., Churney, K L., and Nuttall, R. L., 1982, The NBS Tables of Chemical Thermodynamic Properties: Selected Values for Inorganic and c1 and c2 Organic Substances in SI units: Jour. Physical and Chemical Reference Data ll Supplement #2, 392p White, D. E., Hem, J.D., and Waring, G A., 1963, Chemical composition of subsurface waters: U S Geol Survey Professional Paper 440-F, 67p. Whitfield, M., 1975, The extension of chemical models for sea water to include trace components at 25C and 1 atm pressure: Geochim. et Cosmochim. Acta, v. 39, p. 1545-1557.

PAGE 138

119 APPENDIX Algebra of the derivation of the phase boundary between ankeritedolomite and siderite-magnesite in equilibrium with an aqueous solution In the derivation these abbreviations are used: [mag] = [HgC0 3], [sid] = [Feco3], [ank] = The reaction between the end members of the siderite-magnesite solid solution i s Feco3 (c) 2+ + Mg (aq) = MgC03 (c) + 2+ Fe (aq) The expression for the equilibrium constant for this reaction is 2+ [mag][Fe ] ---------:z:;-[sid][Mg ] K = sid/mag S ub stituting (1 -[sid]) for [mag] yields Ksid/mag Crossmultiplying, rearranging and simplifying yields 2 + (1 -[sid])(Fe ] 2+ [sid][Fe ] [sid] [ Fe2+ ] 2+ [sid]([Fe ] 2+ K .d/ [sid][Mg ] mag 2+ Kid/ s mag 2+ + K .d/ [sid][Mg ] mag K [M 2+]) + sid/mag g [Fe2+] --:z:;-----------------:z:;--(Fe ] + K .d/ [Mg ] mag [sid] (l) (2) (3) Analogously, the reaction between end members of the ankerite-dolomite

PAGE 139

120 solid solution is 2+ 2+ Ca0 5 Fe 0 5co3(c) + 0.5Mg (aq) = ca0 5Mg0 5co3(c) + 0.5Fe (aq) (4) The expression for the equilibrium constant for this reaction is [dol][Fe2+J0.5 Kank/dol = Substituting (1 -[ank}) for [dol] yields (1-[ ank])[Fe2+J0.5 Kank/dol = Crossmultiplying, rearranging and simplifying yields 2+ 0.5 2+ 0.5 (1-[ank])[Fe ] = Kank/dol[ank][Mg ] [Fe2+]0.5-[ank][Fe2+]0.5 = Kank/dol[ank][Mg2+]0.5 [Fe2+]0.5 = [ank][Fe2+]0.5 + Kank/dol[ank][Mg2+]0.5 [ank]([Fe2+]0.5 + Kank/dol[Mg2+]0.5) [Fe2+1o.5 ----------------------------------= [ank] [Fe2+]0.5 + K [M 2+]0.5 ank/dol g Finally, the reaction relating ankerite and siderite via the aqueous solution is (5) (6) 2+ 2+ Ca0 5 Fe 0 5co3(c) + 0.5Fe (aq) = Feco 3(c) + 0.5Ca (aq) (7) The expression for the equilibrium constant for this reaction is K ank/sid [ sid][Ca2+]0.5 Now substitute equations 3 and 6 into equation 8 [Fe2+][Ca2+]0.5 ----2;-----------------2;-[Fe ] + Ksid/ma [Mg ] = [Fe2+]0.5[Fe2+]0.5 Kank/sid Crossmultiplying and rearranging yields (8)

PAGE 140

121 K = ank/sid 2+ 2+ Kank/sid([Fe ] + Ksid/mag[Mg ]) = [Ca2+]0.5[Fe2+]0.5 + Kank/dol[Mg2+]0.5[Ca2+]0.5 Dividing through by [Ca 2 +] yields [Fe2+] [Mg2+] K ------ank/sid(Ca2+] + K K ------ank/sid sid/mag[Ca2+] Now rearranging and substituting values for the equilibrium constants 2+ -1.76 (Mg ] 10 ------+ [Ca 2+] [l-f 2+]0.5 -0.94 g 10 --------[Ca2+]0.5 (10) This equation can be solved b y the quadratic equation where the last two terms on the right side are combined into the "c" term of the quadratic 2+ 2+ equation, and a value for [Mg ]/[Ca ] is substituted. The results of solving this equation are given in Table 1 4 a nd shown on Figure 16. (9)


printinsert_linkshareget_appmore_horiz

Download Options

close
No images are available for this item.
Cite this item close

APA

Cras ut cursus ante, a fringilla nunc. Mauris lorem nunc, cursus sit amet enim ac, vehicula vestibulum mi. Mauris viverra nisl vel enim faucibus porta. Praesent sit amet ornare diam, non finibus nulla.

MLA

Cras efficitur magna et sapien varius, luctus ullamcorper dolor convallis. Orci varius natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Fusce sit amet justo ut erat laoreet congue sed a ante.

CHICAGO

Phasellus ornare in augue eu imperdiet. Donec malesuada sapien ante, at vehicula orci tempor molestie. Proin vitae urna elit. Pellentesque vitae nisi et diam euismod malesuada aliquet non erat.

WIKIPEDIA

Nunc fringilla dolor ut dictum placerat. Proin ac neque rutrum, consectetur ligula id, laoreet ligula. Nulla lorem massa, consectetur vitae consequat in, lobortis at dolor. Nunc sed leo odio.