.I,.... t .249. .4 .:_, .
OJ :IJ vw : . . ;
.mqu: . .4 $1.. {1r 1%me dig \.,_.
3:. 2!? :- ‘J. 7L1... whim. u.
1.. A :33; 3...; , .. 5t), .. .45
. #35:: J. . .5 ,2 . .01. ,«y. . 5!. v. V
. ‘ a, 3 i...3
J... , . xi) . .

Luau

1.0114135; 32
. 1‘5... .3 v

.
: u»...« .3? a
2 3513.5 .
t, .... .
o: u 1...: f. n :r

Hf ; 2.
, (3 ha. . I , \I .
. fir)... . ‘ ‘ .qu‘mwggw

. 1.12.50 1

. 0..
4313.2. .
)Hr‘rru'shralswa.
C... u: a i .r
L\ .Xu «3..
any . I.)
.£.?.v..15nr
l
.. . 5%.
. f A .-
1qu b. ,
i.

I!
992.13

, €1.1Q u

 

 

a”.
. I... . 11!!
. nagfifii.

3x

a..n.........fl.w.amw..
.313"! z
2.1.
.3253.
.....n..:......
VIA «flux

.1.
. 14.. ...
. $323.... .< 31...}!
p: a 2.4...»
r .5 .5 0 ¢
.3. X9! .
.3..." e. zit-5.1:»...
. 3 !‘

Vivi}? 1.
3‘1...
91! v
A t. 4 3

t) !

 

This is to certify that the
dissertation entitled

CRYSTAL GROWTH MECHANISMS IN NATURAL AND
SYNTHETIC DOLOMITE: INSIGHT INTO DOLOMITIZATION
KINETICS

presented by

STEPHEN EMIL KACZMAREK

has been accepted towards fulfillment
of the requirements for the

Ph. D degree in Geological Sciences

 

 

 

 

MSU is an Aflinnative ActiorVEquaI Opportunity Institution

 

LIBRARY
Michigan State
v University

 

 

 

e‘-I-I-I-c-O-I-I-O-O-l-l-0-0-.-.-0-0-0-.-l-O-I-O-O-l-I---O-O-IQ.‘-O-O-O-l-_!V-V'-I-U-l-l-o-o-o-I-o-0-o-u-o-u-o-n-u-u-o-o-

PLACE IN RETURN BOX to remove this Checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/CIRC/DateDue.tndd—p.1

 

 

CRYSTAL GROWTH MECHANISMS IN NATURAL AND SYNTHETIC

DOLOMITE: INSIGHT INTO DOLOMITIZATION KINETICS

By

Stephen Emil Kaczmarek

A Dissertation

Submitted to
Michigan State University
In partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Geological Sciences

2005

ABSTRACT

CRYSTAL GROWTH MECHANISMS IN NATURAL AND SYNTHETIC
DOLOMITE: INSIGHT INTO DOLOMITIZATION KINETICS

By

Stephen Emil Kaczmarek

The lack of dolomite in modern marine setting is a kinetic problem, yet
relatively few details about dolomitization are understood. Because the
geometries observed on crystal surfaces are dictated by growth and dissolution,
and such mechanisms are theoretically related to kinetics, the current study
utilizes ex situ atomic force microscopy to investigate nanometer-scale features
on synthetic and natural dolomites in order to better understand dolomitization.

Following a relatively long induction period, high-temperature synthetic
dolomite forms very rapidly. Initial dolomite products are poorly ordered (i.e.,
nonideal), whereas stoichiometry (61-50 mole% Ca003) is dependent on the
initial Mgz“:Ca2+ ratio in solution (R2=0.97). Following initial reactant depletion,
dolomite products are stoichiometric and well-ordered (i.e., ideal).

Two distinct nanometer-scale growth features - islands and layers -
characterize dolomite growth surfaces. Islands are rounded positive relief
features. Layers are broad, flat surfaces with steps. Islands occur on nonideal
synthetic dolomite prior to reactant depletion, whereas layers form only after
calcite reactant depletion. Counter to theoretical predictions, dolomite
nanotopography is independent of the Mgz“:Ca2+ in solution. Surface

nanotopography does respond, however, to changing carbonate flux at the

growth interface following reactant depletion, therefore suggesting that carbonate
plays a major role in dolomitization kinetics.

Following chemical etching, ideal synthetic dolomite surfaces exhibit flat
layers with deep euhedral pits, whereas nonideal synthetic dolomite surfaces are
covered with islands identical to the islands observed on growth surfaces.
Chemically etched ideal and nonideal natural dolomites are also characterized by
etch pits and islands, respectively. These features are indistinguishable from
islands and etch pits observed on synthetic dolomite. Based on models of crystal
growth and dissolution, nonideal dolomite surface nanotopography is most
consistent with polynuclear growth. Conversely, ideal dolomite is more
consistent with spiral growth. These observations indicate that dolomite initially
forms by polynuclear growth and is nonideal. Because it is metastable, nonideal
dolomite may later be replaced by ideal dolomite.

Due to the similarities between natural and synthetic dolomites, high-
temperature experimental findings can serve as a model for interpreting
observations from natural low-temperature settings. Therefore, a long induction
period followed by rapid growth is the best model to explain the absence of

dolomite in modem marine environments.

ACKNOWLEDGEMENTS

Many thanks to my advisor, mentor, and friend Duncan Sibley; it has truly
been an honor and a privilege to think about science and life with you over the
years. To my committee members Mike Velbel, Lina Patino, and Dave Long for
helping to make this project worthy of being proud of. I would like to thank Ewa
Danielewicz at the MSU Center for Advanced Microscopy for her help with
mastering AFM and SEM techniques. Much appreciation to Loretta Knutson,
Jackie “the dude” Bennett, Cathy Caswell, and Marsha Walsh for helping me
avoid as many administrative problems as possible. I would also like to
recognize family and friends for remaining supportive and interested even when I

appeared not to be.

iv

TABLE OF CONTENTS

LIST OF TABLES ................................................................................. vii
LIST OF FIGURES ............................................................................... viii
1. INTRODUCTION ............................................................................... 1

1.a. Previous Experimental Work ...................................................... 4

1.b. Ca-Enrichment in Natural Dolomite ........................................... 11

1.c. Crystal Growth Mechanisms .................................................... 16

1.d. Chemical Etching .................................................................. 24

2. METHODS ...................................................................................... 29

2.a. Dolomite Synthesis ................................................................ 29

2.b. X-Ray Diffraction ................................................................... 31

2.c. Atomic Force Microscopy ......................................................... 35

2.d. Scanning Electron Microscopy ................................................. 38

2.e. Crystal Size Distributions ........................................................ 39

3. RESULTS ....................................................................................... 40

3.a. Synthetic Dolomite Reaction .................................................... 4O

3.b. Calcite Reactant Experiments .................................................. 44

3.b.1. Reaction Rate ............................................................ 44

3.b.2. Dolomite Stoichiometry ............................................... 57

3.b.3. Cation Order ............................................................. 65

3.b.4. Solution Chemistry ..................................................... 71

3.c. Aragonite Reactant Experiments ............................................... 73

3.c.1. Reaction Rate ........................................................... 73

3.c.2. Dolomite Stoichiometry ................................................ 77

3.c.3. Cation Order ............................................................. 77

3.d. Synthetic Dolomite Crystal Size Distributions .............................. 80

3.e. Synthetic Dolomite Surface Nanotopography, AFM ...................... 80

3.e.1. Calcite Reactant Experiments ....................................... 80

3.e.2. Aragonite Reactant Experiments ................................... 90

3.f. Synthetic Dolomite Surface Nanotopography, SEM ....................... 96

3.f.1. Calcite Reactant Experiments ....................................... 96

3.f.2. Aragonite Reactant Experiments .................................. 105

3.9. Etched Synthetic Dolomite Surface Nanotopography, AFM ........... 105

3.h. Natural Dolomite Surface Nanotopography, AFM ........................ 112

3.i. Natural Dolomite Surface Nanotopography, SEM ......................... 120

4. DISCUSSION ................................................................................. 124

4a. Islands and Layers Are Growth Features .................................. 124
4.b. Polynuclear Growth .............................................................. 127
4.c. Spiral Growth ...................................................................... 135
4d. Transition From Polynuclear To Spiral Growth ........................... 142
4.d.1. Energy .................................................................... 143

4.d.2. Carbonate ............................................................... 150

4.e. Synthetic Dolomite Stoichiometry ............................................ 159
4.f. Synthetic Dolomite Cation Order .............................................. 161
4.9. Synthetic Dolomite Recrystallization ......................................... 164
4.h. Natural Dolomite Growth Mechanisms ..................................... 167
4.h.1. Nonideal Dolomite And Polynuclear Growth .................. 168

4.h.2. Ideal Dolomite And Spiral Growth ................................ 172

4.i. Surface Nanotopography and lntemal Microstructure .................... 173
4.]. Similarities Between Synthetic and Natural Dolomites .................. 176
4.j.1. Induction Period ........................................................ 176

4.j.2. Reaction Rate ........................................................... 178

4.j.3. Selective Replacement ............................................... 179

4.j.4. Crystal Size Distributions ............................................ 180

4.j.5. Mimetic Replacement ................................................. 181

4.j.6. Stoichiometry ........................................................... 182

4.j.7. Cation Order ............................................................. 185

4.k. Implications for Natural Dolomite ............................................. 187
5. CONCLUSIONS .............................................................................. 207
6. REFERENCES CITED ..................................................................... 211

vi

LIST OF TABLES

Table 1. Natural dolomite sample list ........................................................ 34
Table 2a. Experimental results for high-temperature calcite experiments ......... 41
Table 2b. Experimental results for high-temperature aragonite experiments ..... 43
Table 3. Induction period times for calcite experiments ................................. 56
Table 4. Changes in solution Mgz*:Ca2“ chemistry at 50% products and 100%

products. Column three is the average dolomite stoichiometry observed in the
experiments for the initial solution Mg2":Ca2+ in column one .......................... 72

vii

LIST OF FIGURES

Figure 1. Generalized reaction curve for high-temperature synthetic
dolomitization reactions .......................................................................... 7

Figure 2. Surface expressions of some common crystal growth models .......... 17

Figure 3. Diagram showing theoretical growth rates of spiral and polynuclear
growth as a function of free energy (after Sunagawa, 1984) .......................... 18

Figure 4. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:Ca2+ = 0.429 ..................................................................... 45

Figure 5. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:Ca2+ = 0.50 ....................................................................... 46

Figure 6. Percent product vs. time for calcite reactant experiment with initial
solution Mgz":Ca2* = 0.66 ....................................................................... 47

Figure 7. Percent product vs. time for calcite reactant experiment with initial
solution M92503” = 0.786 ..................................................................... 48

Figure 8. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:Ca2+ = 1.0 ........................................................................ 49

Figure 9. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:Ca2* = 1.14 ....................................................................... 50

Figure 10. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:Ca2* = 1.27 ....................................................................... 51

Figure 11. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:Ca2* = 1.5 ........................................................................ 52

Figure 12a. Induction period (log-scale) vs. initial solution Mgz“:Ca2+ for calcite
reactant experiments ............................................................................ 54

Figure 12b. Induction period (linear-scale) vs. initial solution Mgz*:Ca2+ for calcite
reactant experiments ............................................................................ 55

Figure 13. Dolomite stoichiometry vs. initial solution Mgz*:Ca2" for all dolomites

formed prior to calcite reactant depletion, calcite reactant experiments,
R2=0.9688, N=69 ................................................................................. 58

viii

Figure 14. Dolomite stoichiometry vs. percent product for all dolomites formed
prior to calcite reactant depletion, calcite reactant experiments, R2=0.023,
N=69 ................................................................................................. 59

Figure 15. Dolomite stoichiometry vs. percent product for dolomites formed prior
to calcite reactant depletion, calcite reactant experiment, initial solution Mgz":Ca2+
= 0.79, R2=0.0013 ................................................................................ 60

Figure 16. Dolomite stoichiometry vs. percent product for dolomites formed prior
to calcite reactant depletion, calcite reactant experiment, initial solution Mgz*:Ca2*
= 1.27, R2=0.0005 ................................................................................ 61

Figure 17. Dolomite stoichiometry vs. time for dolomite formed prior to calcite
reactant depletion, calcite reactant experiment, initial solution Mgz":Ca2+ = 0.79,
R2=0.0003 ............................... _ ........................................................... 63

Figure 18. Dolomite stoichiometry vs. time for dolomites formed prior to calcite
reactant depletion, calcite reactant experiment, initial solution Mgz":Ca2+ = 1.27,
R2=0.0097 .......................................................................................... 64

Figure 19. Cation order vs. time for calcite reactant experiment, initial solution
Mgz":Ca2+ = 1.0. Reactant depletion occurs after ~160 hours ........................ 67

Figure 20. Cation order vs. time for calcite reactant experiment, initial solution
Mgz”:Ca2” = 1.5. Reactant depletion occurs between 35 and 77 hours ........... 68

Figure 21. Cation order vs. time for calcite reactant experiment, initial solution
Mgz*:Ca2” = 0.66. Reactant depletion occurs at ~217 hours ......................... 69

Figure 22. Cation order vs. dolomite stoichiometry for all synthetic dolomites
formed prior to calcite reactant depletion. Calcite reactant experiments, N=69,
R2=0.0342 .......................................................................................... 70

Figure 23. Percent product vs. time for aragonite reactant experiment with initial
solution Mgz*:Ca2+ = 1.0 ........................................................................ 74

Figure 24a. SEM photomicrograph showing selective dolomitization of aragonite
when both calcite and aragonite are present .............................................. 75

Figure 24b. SEM photomicrograph showing randomly oriented dolomite rhombs
on an aragonite substrate ...................................................................... 76

Figure 25. Dolomite stoichiometry vs. time for dolomite formed up to 30 hours

after the onset of experimental conditions, aragonite reactant experiment, initial
solution Mgz*:Ca2+ = 1.0, R2=0.841 .......................................................... 78

ix

Figure 26. Cation order vs. time for aragonite reactant experiments, initial
solution Mgz“:Ca2+ = 1.0, R2=0.8325 ......................................................... 79

Figure 27. Coarse skewed CSD, sample 7-22-02-7. Reactant: 140-160 micron
size Iceland spar calcite ......................................................................... 81

Figure 28. Coarse skewed CSD, sample 1-19-87-1. Reactant: reagent grade
calcium carbonate ................................................................................ 82

Figure 29. Coarse skewed CSD, sample 1-24-98-1. Reactant: reagent grade
calcium carbonate ................................................................................ 83

Figure 30. Coarse skewed CSD, sample 1-17-87-2. Reactant: reagent grade
calcium carbonate ................................................................................ 84

Figure 31. Coarse skewed CSD, sample 1-21-87-1. Reactant: reagent grade
calcium carbonate ................................................................................ 85

Figure 32. AFM image of a nonideal synthetic dolomite growth surface showing
pervasive island nanotopography. Calcite reactant experiment, solution
Mgz”:Ca2*= 1.0 .................................................................................... 86

Figure 33. AFM image of a nonideal synthetic dolomite showing cross-section of
island-covered growth surface. Calcite reactant experiment, solution
Mgz*:Ca2* = 0.66 ................................................................................... 87

Figure 34. AFM image of a nonideal synthetic dolomite growth surface showing
variable nanotopography caused by islands. Calcite reactant experiment, solution
M92“:Ca2+ = 1.5 .................................................................................... 89

Figure 35. AFM image of an ideal synthetic dolomite growth surface showing
straight-sided steps. Calcite reactant experiment, solution Mgz*:Ca2+ = 1.0......91

Figure 36. AFM image of an ideal synthetic dolomite growth surface showing
multiple layers with curved steps. Calcite reactant experiment, solution Mgz"‘:Ca2+
= 1.5 .................................................................................................. 92

Figure 37a. AF M image and cross-section of an ideal synthetic dolomite growth
surface showing an emerging growth layer. Calcite reactant experiment, solution
Mgz*:6a2+ = 0.66 .................................................................................. 93

Figure 37b. AFM image and cross-section of an ideal synthetic dolomite showing
the emerging growth layer disappears into the surface ................................. 94

Figure 38. AF M image of a nonideal synthetic dolomite showing pervasive island
nanotopography. Aragonite reactant experiment, solution Mgz":Ca2” = 1.0 ...... 95

Figure 39. SEM photomicrograph showing crystallographically oriented dolomite
rhombs on the edges and corners of the calcite substrate ............................. 97

Figure 40. SEM photomicrograph showing dolomite growth rhombs on a pitted
calcite surface that has been dissolved ..................................................... 98

Figure 41. SEM photomicrograph showing a dolomite encrusted calcite
grain .................................................................................................. 99

Figure 42. SEM photomicrograph of a small dolomite rhomb on a dolomite
surface. .......................................................................................... 100

Figure 43a. SEM photomicrograph of a nonideal synthetic dolomite surface
showing irregular growth nanotopography ............................................... 101

Figure 43b. SEM photomicrograph of the same nonideal synthetic dolomite
growth surface at higher magnification showing rounded topographic highs and
wavy ledges ...................................................................................... 102

Figure 44a. SEM photomicrograph of an ideal synthetic dolomite showing
relatively flat growth surfaces ................................................................ 103

Figure 44b. SEM photomicrograph of the same ideal synthetic dolomite at
higher magnification showing flat growth surfaces separated by straight-sided
steps ................................................................................................ 104

Figure 45. SEM photomicrograph showing gold coating on dolomite grain.....106

Figure 46. AFM image of an etched synthetic nonideal dolomite showing a
pervasive island dissolution surface. Calcite reactant experiment, solution
Mg“’*:Ca2+ = 0.66 ................................................................................ 107

Figure 47. AFM image of an etched ideal synthetic dolomite showing euhedral
etch pits. Calcite reactant experiment, solution Mgz*:Ca2" = 0.66 .................. 109

Figure 48. AFM image of an etched ideal synthetic dolomite showing a high
density of euhedral etch pits. Calcite reactant experiment, solution
Mgz*:0a2* = 1.5 .................................................................................. 110

Figure 49. AFM image of an etched ideal synthetic dolomite showing euhedral
etch pits and relatively flat inter-pit surfaces. Calcite reactant experiment, solution
Mgz“:c;ai2+ = 1.27 ................................................................................ 111

Figure 50. AFM image of an etched natural nonideal dolomite showing pervasive
island-covered surface ........................................................................ 113

xi

Figure 51. AF M image and cross-section of an etched natural nonideal dolomite
showing dissolution island nanotopography ............................................. 114

Figure 52. AFM image of an etched natural nonideal dolomite showing an island-
covered surface with voids and depressions ............................................ 115

Figure 53. AFM image of an etched natural ideal dolomite showing euhedral

etch pits with flat inter-pit layers ............................................................. 116
Figure 54. AFM image of an etched natural ideal dolomite showing
crystallographically oriented rhombic etch pits .......................................... 117
Figure 55. AFM image of an etched natural ideal dolomite showing euhedral
etch pits with rhombic shape ................................................................. 118
Figure 56. AF M image and cross-section of an etched natural ideal dolomite
showing euhedral etch pits with pointed bottoms ....................................... 119
Figure 57. SEM photomicrograph of an unetched natural dolomite surface
showing voids (V), steps (S), and debris (D) ............................................. 121
Figure 58. SEM photomicrograph of an etched natural nonideal dolomite
surface. Arrows mark voids (V), steps (S), and debris (D) ........................... 122
Figure 59. SEM photomicrograph of a natural ideal dolomite showing a highly
irregular and dissolved surface after 10 minutes of etching in dilute acid ........ 124
Figure 60. AFM image from Heraty (1998) showing amorphous growth islands
on a calcite surface in phosphate-doped experiments ................................. 131
Figure 61. Growth mechanism summary diagram for dolomitization ............. 166

Figure 62. 2-theta (i.e. dolomite stoichiometry) vs. percent dolomite for GWP
data from Abu Dhabi, R2=0.3706. Note: 2-theta values are proportional to
dolomite stoichiometry ......................................................................... 199

xii

There is a mask of theory over the whole face of nature

- W. Whewell

xiii

1. INTRODUCTION

The origin of dolomite, [CaMg(C03)2], has been debated extensively for
over a decade. Although most geologists now regard dolomite as a sedimentary
mineral that forms during the replacement reaction between carbonate sediments
and magnesium-rich fluids, no unique geologic environment of dolomitization has
been recognized (e.g., Murray and Pray, 1965; Friedman and Sanders, 1967,
Folk and Land, 1975; Budd, 1997). Dolomite is found in thick, laterally extensive
deposits throughout the rock record, but exists in modern marine environments
only in small abundances and few locations. Because dolomite is the
thermodynamically stable carbonate phase in seawater, its rare occurrence in
modern carbonate-forming marine environments is problematic. Although no
universal explanation has emerged, most workers generally agree that the
“dolomite problem” is one of kinetics (Land, 1985; Hardie, 1987; Sibley, 1990;
Budd, 1997).

Relatively little is known about the reaction details of natural dolomitization
aside from the fact that, for most dolomites, no geologic evidence exists of
unusual temperature or pressure conditions (Krauskopf and Bird, 1995).
However, both the failure to unequivocally synthesize dolomite at temperatures
below 100°C (e.g., Usdowski, 1994; Land, 1998) without the aid of bacteria (see
Vasconcelos et al., 1995; Warthman et al., 2000) and the lack of a modern
analog for massive dolomite deposits have led a number of workers to examine

high-temperature synthesis experiments as a means of acquiring information

about dolomite reaction kinetics (Graf and Goldsmith, 1956; Gaines, 1974; Katz
and Matthews, 1977; Baker and Kastner, 1981; Sibley et al., 1987; Morrow and
Rickets, 1988; Sibley, 1990; Sibley et al., 1994; Kessels et al., 2000). These
high-temperature experiments provide insight into the kinetic details of the
CaCOa to dolomite transformation, yet there exist no theoretical bases for
extrapolating kinetic data over the 25—250°C temperature range between
sedimentary dolomites and temperatures commonly employed in laboratory
experiments. Consequently, applicability of these high-temperature experiments
to natural low-temperature systems remains uncertain.

Dolomitization of CaCO3 requires nucleation and crystal growth, both of
which are dictated by thermodynamic, as well as kinetic factors. The details of
these processes are in turn reflected in the minerals texture and microstructure.
Knowledge of a crystal’s defect microstructure can provide insight to the growth
process (e.g., Reeder, 2000). More specifically, different models of crystal
growth imply different defect structures. This study will test two hypotheses
about the kinetics of dolomitization: 1) high-temperature, synthetic dolomites and
low-temperature, natural dolomites have similar growth mechanisms, and 2) the
growth mechanisms in dolomite are dictated by Mgz“:Ca2“ ratios (i.e., the degree
of supersaturation) of the dolomitizing solutions. Both hypotheses have a great
deal of importance in regard to understanding dolomitization in natural settings.
If the first hypothesis is unsubstantiated, the applicability of high-temperature

experiments to natural dolomite is tenuous. Support for the second hypothesis

suggests that knowledge of the growth mechanism would allow the possibility of
interpreting information about the environment of formation of natural dolomites.

Before proceeding, it is worth noting that both natural and synthetic
dolomites are found in a variety of forms that must be distinguished. Dolomite is
a complex mineral in which considerable variations occur in the composition and
the degree of cation order (Goldsmith and Graf, 1958; Reeder and Wenk, 1979;
Reeder, 1992). The term dolomite is used here to describe Ca-Mg carbonates
with compositions approximating mMg/mCa = 1 and evidence of cation ordering.
For dolomite with non-Mg and non-Ca ions only in trace abundances,
stoichiometry is generally expressed in terms of mole percent CaCOa. Ideal
dolomite, Cao,5oMgo,5o(Coa)2, refers to dolomites with equal molar quantities of
calcium and magnesium that display a relatively high degree of cation order. The
structure of ideal dolomite is characterized by alternating A and B cation planes,
separated by layers of 003, all of which are stacked along the threefold axis
(Reeder, 1981). Both Lippmann (1973) and Reeder (1983) have likened the
structure of ideal dolomite to a composite of alternating calcite and magnesite
layers. When the criteria of a high degree of cation order and stoichiometric
composition are not satisfied, the term nonideal dolomite is appropriate.
Nonideal dolomites are most often associated with Ca-enrichment (Reeder,
1983). Although Mg-rich dolomite does occur in nature, its occurrence is
relatively rare (Rosen et al., 1989) or poorly documented (Reeder, 2000).

In a number of instances, the term protodolomite (Graf and Goldsmith,

1956) has been used in the literature to describe Ca-Mg carbonates with very

poor or non-existent ordering, the inference being that protodolomite might
represent a metastable precursor of well-ordered stoichiometric dolomite for
high-temperature experimental dolomites described by Graf and Goldsmith
(1956) and natural dolomites as well (Lippmann, 1973; Reeder, 1981). This
latter view is consistent with the observation that many recent dolomites are Ca-
rich and poorly ordered. Land (1980) has argued against the term protodolomite
on the grounds that natural Mg-Ca-carbonate phases often display some degree
of cation order and should therefore be considered dolomite. Following this
reasoning, the term protodolomite is avoided when describing poorly ordered Ca-
Mg-carbonate phases, unless making reference to previously published

observations.

1.a. Previous Experimental Work

Several experimental investigations have attempted to identify the
thermodynamic and kinetic controls on dolomite formation (Graf and Goldsmith,
1956; Land, 1967; Lippmann, 1973; Gaines, 1974, 1980; Katz and Matthews,
1977; Baker and Kastner, 1981; Sibley and Bartlett, 1987; Sibley et al., 1987a,b;
Morrow and Rickets, 1988; Sibley, 1990; Nordeng and Sibley, 1993; Sibley et al.,
1994; Morrow et al., 1994; Kessels et al., 2000). These investigators employed
temperatures in excess of 100°C because dolomite has a very low reactivity at
lower temperatures (Usdowski, 1994). Moreover, scientists have largely been
unsuccessful in synthesizing dolomite inorganically at low temperatures even

over very long times (6.9. 30 years, of. Land, 1998) other than poorly ordered

Ca-rich (nonideal) dolomite (Deer et al., 1992; Deelman, 1981; Busenberg and
Plummer, 1989). Katz and Matthews (1977) examined the effects of
temperature on the dolomitization rate. They reported that complete
dolomitization of calcite took 112 hours at 252°C, 38 hours at 263°C, 14 hours at
274°C, 6 hours at 285°C, and 4 hours at 295°C. From this data, Katz and
Matthews (1977) calculated the activation energy for the synthetic dolomite to be
48-50kcallmol, which they used to invoke the high temperatures needed to
synthesize dolomite over reasonable time-scales. Based on experimental data
from Gaines (1968), Arvidson and Mackenzie (1997) recalculated the activation
energy of ordered (ideal) dolomite to be 42.1 kcallmol. Arvidson and Mackenzie
(1999) later calculated the activation energy for poorly ordered calcium-rich
dolomite (protodolomite) to be 31 .9kcaI/mol, approximately 10 kcallmol less than
the value for ideal dolomite. Arvidson and Mackenzie (2000) postulated that the
activation energy of dolomite is a function of 1) the energy associated with the
dehydration of magnesium ions, and 2) the energy associated with cation
ordering. They plotted activation energy (E,) vs. composition for Ca-Mg
carbonates based on a published value for calcite from Kazmierczak et al. (1982)
and an extrapolated value for magnesite. Although the experimentally
determined activation energy for the nonideal dolomite (protodolomite) used by
Arvidson and Mackenzie (1999) was in nearly perfect accord with the value
predicted from the extrapolated line between the calcite and magnesite values,
the experimentally derived E. for ideal dolomite was approximately 5.5 kcallmol

greater than the value predicted by the line. Arvidson and Mackenzie (2000)

interpreted the additional energy to represent the effects of cation order in ideal
dolomite. The additional energy reported by Arvidson and Mackenzie (2000) for
ideal dolomite is in agreement with the E, of 5.5kcal/mol calculated earlier by
Malone et al. (1996) which was determined for the rate of recrystallization of
disordered Ca-Mg carbonate phases to ordered dolomite between 149-197°C.

Graf and Goldsmith (1956) formed dolomite by reacting combinations of
various Mg-bearing and carbonate-bearing species (calcareous algae, echinoids,
Ca-Mg-carbonate gel, aragonite and Mg-carbonate) in hydrothermal bombs at
various temperatures. In lower temperature and shorter duration experiments,
they observed that initial products (called protodolomite) were compositionally
different from the ideal dolomite that was formed in higher temperature and
longer duration experiments. Powder X-ray diffraction (XRD) patterns of the
initial products (protodolomite) showed expanded unit cell dimensions, which
Graf and Goldsmith (1956) interpreted as substitution of larger calcium cations
into magnesium sites. The initial products (protodolomite) also exhibited missing
superstructure reflections, which they interpreted to indicate an absence of the
ideal succession of alternating magnesium and calcium cation planes.

Results of subsequent high-temperature studies demonstrated that the
overall reaction curve for the dolomitization reaction is sigmoidal-shaped (Figure
1; Katz and Matthews, 1977; Sibley and Bartlett, 1987; Sibley et al., 1987; Sibley,

1990; Nordeng and Sibley, 1993; Sibley et al., 1994; Kessels et al., 2000), the

 

I100

Percent Product

20

 

I
l

Generalized Dolorritization Reaction Cave

 

f

 

 

 

/
Rapid Replacement, Ideal Dolomite
Nonideal Dolomite Forms Forms

 

 

 

Induction Period

 

 
   

 

Tlrre

Figure 1. Generalized reaction curve for high-temperature synthetic
dolomitization reactions.

 

slope at any point being a function of growth and nucleation rates (Avrami, 1939).
The dolomitization reaction has a relatively long induction period followed by
rapid replacement of reactants (Katz and Matthews, 1977; Sibley and Bartlett,
1987; Sibley et al., 1987; Sibley, 1990; Nordeng and Sibley, 1993; Sibley et al.,
1994; Kessels et al., 2000). The induction period is the time after the onset of
experimental conditions when no dolomite products are detected (Walton, 1969).
Sibley et al. (1987) showed that the induction period, as determined by XRD,
could be 20% to >80% of the total duration of the reaction. Nordeng and Sibley
(1993) conducted thermal cycling experiments to show that nucleation occurs
relatively early in the induction period. The length of the induction period is
therefore dependent, in part, on the ability to detect small percentages of
product. Kessels et al. (2000) showed that the length of the induction period was
lowered, but not eliminated when atomic force microscopy was used to detect the
presence of reaction products.

The induction period is followed by a rapid replacement stage where
calcite dissolves and metastable, nonideal dolomite products form (refer to
Figure 1). Following the rapid replacement stage, when all reactants have been
consumed, ordered, stoichiometric dolomite forms (Gaines, 1974, 1980; Katz and
Matthews, 1977; Baker and Kastner, 1981; Sibley and Bartlett, 1987; Sibley et
al., 1987; Morrow and Rickets, 1988; Sibley, 1990; Nordeng and Sibley, 1993;
Sibley et al., 1994; Kessels et al., 2000).

A number of factors have been shown to affect the overall rate at which

dolomitization of a carbonate precursor takes place in high-temperature

experiments. These include: Mgz":Ca2* ratio of the dolomitizing solution,
available surface area of the reactant, mineralogy of the reactant, and the
presence of inhibitors or catalysts (Graf and Goldsmith, 1956; Land, 1967;
Lippmann, 1973; Gaines, 1974, 1980; Katz and Matthews, 1977; Baker and
Kastner, 1981; Sibley et al., 1987; Morrow and Rickets, 1988; Sibley, 1990;
Sibley et al., 1994; Kessels et al., 2000). Gaines (1974) reported that
temperature, the solution Mgz*:Ca2” ratio, and the type of reactant (calcite vs.
aragonite) all have an effect on the rate of dolomitization. Land (1967), Katz and
Matthews (1977), Gaines (1980), Sibley et al. (1987), Sibley (1990), and Kessels
et al. (2000) showed that increasing solution Mgz“:Ca2+ ratios caused the reaction
rate to increase. Gaines (1980) showed that the reaction rate increased when
the Mgz*:Ca2* was increase from 3 to 5, but decreased when the ratio was
increased from 5 to 7. Sibley et al. (1987) argued that the decreasing reaction
rate for Mgz*:Ca2+ ratios greater than 5 was likely due to the precipitation of
magnesite.

Dolomitization rates in high-temperature experiments increase with
reactant surface area (Katz and Matthews, 1977; Gaines, 1980; Sibley and
Bartlett, 1987). More reactant surface area corresponds to more potential
nucleation sites per unit mass of reactant. A greater reactant surface area also
enhances the rate of dissolution of calcium carbonate, the source of carbonate
anions for dolomite formation. Sibley and Bartlett (1987) showed that the rate of
dolomitization was nearly three times faster in experiments where 1-10um size

calcite crystals were used as the reactant compared to experiments where 10-

40pm size calcite crystals were used. In their aragonite experiments, Sibley and
Bartlett (1987) showed complete dolomitization took approximately 25 hours for
1-10um calcite crystals and approximately 35 hours for 10-40pm calcite crystals.

Katz and Matthews (1977), Baker and Kastner (1981), and Sibley and
Bartlett (1987) showed that aragonite is dolomitized much faster than calcite.
Baker and Kastner (1981) showed that aragonite was completely converted to
poorly ordered, Ca-rich dolomite in 48 hours whereas calcite experiments
showed only minor amounts of nonideal dolomite after 48 hours at 200°C.
Furthermore, aragonite was completely dolomitized in 112 hours, but calcite was
only partially converted to nonideal dolomite after 112 hours. Sibley and Bartlett
(1987) observed that aragonite was completely dolomitized in approximately 25
hours at 175°C, whereas calcite required nearly 65 hours.

Other experiments have indicated that certain additives may alter the way
dolomitization takes place at high temperature. Baker and Kastner (1981)
showed that dolomitization was suppressed in high-temperature experiments by
the addition of sulfate. Gaines (1980) showed that the addition of 0.05 Li
increased the reaction rate. Sibley (1990) observed that when N2 and CO; gas
were added to hydrothermal bombs, reaction rates increased. Morrow and
Rickets (1988) added HCOa’ through NaHC03 to high-temperature experiments
and recorded an increased reaction rate. Because calcite equilibrates relatively
fast at high temperatures, Sibley et al. (1994) argued that it was unlikely that

addition of NaCO3 significantly increased the ion activity product of dolomitizing

10

solutions. Lippmann (1973) suggested that 0032' might help dehydrate Mg”
ions in solution, thus explaining the increased rate of M9003 precipitation.

Sibley (1990) showed that the first products to form during dolomitization
experiments exhibited a compositional dependence on the Mgz“20a2+ ratio in
solution. When a solution Mgz*:0a2“ ratio of 1 was used, the average composition
of the initial products was approximately 60 mole% Ca003. When the solution
Mgz*:0a2* ratio was 0.66, initial products had an average composition of 65
mole% 0a003. Sibley et al. (1994) also reported that solutions with higher initial
Mgz”:0a2" ratios lead to more stoichiometric products. They reported the
formation of very high magnesium calcite (VHMC, 35-40 mole% M9003 with no
ordering) when initial solutions had a Mgz*:0a2+ ratio of 0.66 and
nonstoichiometric dolomite (43-45 mole% M9003) when initial solutions had a
Mgz“:Ca2+ ratio of 1.2. Despite highly variable compositions and poor ordering in
initial dolomite products, and the fact that no experiment has produced well-
ordered stoichiometric (ideal) dolomite as an initial reaction product, all products
tend towards well-ordered, stoichiometric (ideal) dolomite after all reactants have

been consumed.

1.b. Car-Enrichment in Natural Dolomite

Although ideal dolomite is the thermodynamically stable carbonate phase
in seawater, a considerable fraction of sedimentary dolomite contains Ca in
excess of the ideal 1:1 Mg:Ca ratio. The excess calcium content in natural

dolomite is quite variable and can reach 9-10 mole% 03003 (Graf and

11

Goldsmith, 1956; Fllchtbauer and Goldschmidt, 1965; Lippmann, 1973;
Lumsden, 1979; Lumsden and Chimahusky, 1980; Reeder, 1981, 1983; Wenk et
al., 1983; Sperber et al., 1984; Gregg et al., 1992; Malone et al., 1994; Budd,
1997; Bettcher et al., 1998; Jones et al., 2001). In addition, calcium-rich
dolomites exhibit a lower degree of cation order than their ideal counterparts.
Although Ca-enrichment in natural dolomite is a common phenomenon,
the factors responsible for deviation from ideality are poorly understood.
Goldsmith and Graf (1958) using X-ray diffraction were the first to report on the
structural and compositional variations in natural Ca-rich and stoichiometric
dolomites. Powder-mount and single crystal X-ray diffraction (XRD) data
indicate expanded unit cell dimensions, attenuated superlattice reflections, and
diffuse or bimodal basal plane reflections in Ca-rich dolomite. Dolomite
diffraction peak intensities were shown to indicate differences in occupancies and
scattering factors of the Ca and Mg atoms in the structurally nonequivalent A and
B cation sites, respectively (Graf and Goldsmith, 1956). Therefore the
occurrence of attenuated superlattice reflections suggested to Goldsmith and
Graf (1958) that there was some degree of Ca-Mg disorder in the Ca-rich
samples. They argued that some defect mechanism or alternate crystal structure
was required in order to accommodate excess calcium. Graf et al. (1967)
reported that cation substitution and/or mixed-layer disorder were necessary in
order to produce the reflections reported by Goldsmith and Graf (1958). Recent
refined single crystal X-ray diffraction studies have produced similar results

(Reeder, 2000). Complete cation order is characteristic of most ancient

12

stoichiometric dolomites, whereas Ca-rich dolomite is most commonly associated
with highly variable lattice parameters (Reeder, 1983; Miser et al., 1987; Reeder,
2000).

Navrotsky and Copobianco (1987) and Chai et al. (1995) studied the
energetics of Ca-enrichment in natural sedimentary dolomites. Solution
calorimetry experiments for a suite of sedimentary dolomites with 0-6 mole%
excess calcium were used to show that enthalpies of formation increase in a
near-linear manner with Ca-enrichment. This supports the widely recognized
view that nonideal dolomite forms more easily than ideal dolomite because of
kinetic reasons, as they are metastable and more soluble.

Although recent dolomites are characterized by being Ca-rich and having
expanded unit cell dimensions, the structural description of these dolomites
varies significantly from the ancient Ca-rich dolomites (Lippmann, 1973; Gregg et
al., 1992; Mazzullo, 1992; Wenk et al., 1993). The XRD patterns of the recent
deposits show that superlattice reflections are weak or absent. This is most likely
because the Ca and Mg cations are distributed more randomly between the A
and B sites. Other reflections are also broad and diffuse, which further suggests
a low degree of structural order in the dolomite. Wenk et al. (1993), for example,
observed a rhombohedral Ca-Mg carbonate with weak or no ordering reflections
that contained approximately 30 mole% M9003 in recent sediments from Abu
Dhabi. In contrast, ancient nonideal dolomites, although typically calcium-rich,
display a much greater degree of cation ordering than their Holocene

counterparts.

l3

Selected area electron diffraction (SAED) and transmission electron
microscopy (T EM) have been used to demonstrate that Ca-rich dolomites are
both structurally and compositionally heterogeneous in comparison to ideal
dolomites (Barber, 1977; Reeder and Wenk, 1979; Reeder, 1981, 1983, 1992,
2000; Wenk et al., 1983; Van Tendello et al., 1985; Frisia, 1994; Schubel et al.,
2000). These workers observed the occurrence of a heterogeneous
microstructure in almost all pre-Holocene Ca-rich dolomites, including laminar-
Iike structural modulations and superstructures. Modulation occurs as a regular,
alternating dark/light wavy pattern in TEM images (Barber, 1977; Reeder and
Wenk, 1979; Reeder, 1981, 1992,2000; Schubel et al., 2000). Modulations can
have sharp to diffuse boundaries and are observed in both brightfield and
darkfield imaging (Barber, 1977; Reeder and Wenk, 1979; Reeder, 1981, 1992,
2000; Schubel et al., 2000). However, the greatest contrast occurs with the {104}
reflection beam in darkfield imaging, which is concurrent with indices of the
cleavage faces in dolomite (Reeder, 1981, 2000). More detailed work has
shown that modulations are oriented subparallel to the growth direction and are
therefore perpendicular to the dominant {104} faces (Reeder, 2000).
Modulations are pervasive features throughout the dolomite crystal lattice
(Reeder, 1981, 1992, 2000). Laminar-like modulations typically have
wavelengths between ten and several tens of nanometers (Barber, 1977; Reeder
and Wenk, 1979; Reeder, 1981, 1992, 2000; Schubel, 2000). Wenk et al. (1983)
reported that although modulations were absent in some very recent nonideal

dolomites, microstructures were heterogeneous.

l4

Most ancient sedimentary dolomites with stoichiometric compositions and
a high degree of cation order lack modulated microstructures and SAED
superstructures (Wenk et al., 1983; Reeder, 1983, 1992). However, the most
common structural defects in ideal dolomites are dislocations, which are
observed in variable densities (Reeder, 1992). Although modulated
microstructures have been observed in stoichiometric dolomite, the occurrence is
relatively rare. Miser et al. (1987) reported on TEM observations of modulation-
like features in a well-ordered, stoichiometric twinned dolomite crystal.
Modulations were reported to be weaker and coarser than commonly reported for
ancient calcian dolomites. Reeder (1992) reasoned that because modulations in
stoichiometric dolomite are typically localized and volumetrically insignificant, the
phenomenon seems to be the exception. In the only study of its kind, Rosen et
al. (1989) reported observing modulations in a Mg-rich dolomite that exhibited
poor ordering and a structure that was expanded in the co direction and
contracted in the ao direction relative to ideal dolomite.

SAED patterns used in conjunction with TEM imaging reveal the presence
of weak and/or diffuse extra spots that coincide with small elongate domains that
are seen with dark field imaging. High-resolution TEM (HRTEM) confirms that
the domains have a different cation distribution from that of ideal dolomite.
Modulated microstructures were therefore interpreted as strain contrasts within
the crystal lattice that result from distinct chemical and/or structural domains
(Barber, 1977; Reeder and Wenk, 1979; Reeder, 1981, 1992,2000; Frisia, 1994;

Schubel et al., 2000). TEM and SAED observations led a number of workers to

15

suggest that the relationship between stoichiometry and dolomite microstructure
may be the result of distinct growth processes (Wenk et al., 1983; Reeder and
Sheppard 1984; Reeder and Prosky, 1986; Miser et al, 1987). The occurrence
of modulations that change orientation according to different growth sectors
supports this hypothesis (of. Reeder, 2000). Fouke and Reeder (1992) reported
that various levels of Ca-enrichment in dolomite were a function of surface
structure and suggested that “the nature of [the] growth mechanism may be a

important factor in controlling the range of CazMg ratios observed in dolomite.”

1.6. Crystal Growth Mechanisms

Mononuclear, birth and spread, polynuclear, and spiral crystal growth
models (Figure 2), all assume that crystal growth occurs by the addition of
structural units (ions or molecules) to energetically favorable attachment sites on
a crystal surface (Nielsen, 1964; Ohara and Reid, 1973). Because the rate of
crystal growth is related to the crystal growth mechanism (Sunagawa, 1982, see
Figure 3), crystal growth models are characterized according to differences in the
rate of formation of surface nuclei versus the rate of lateral attachment to an
already existing surface feature. Nucleation dependent models, like
mononuclear growth, birth and spread, and polynuclear growth represent
successively higher rates of nucleation relative to lateral growth. Because
nucleation and crystal growth are energetically driven processes, theoretical
models emphasize that different growth mechanisms dominate under different

thermodynamic conditions. At conditions near equilibrium, no growth can occur

16

 

 

I

 

m

 

Spiral Growth

 

 

 

 

 

Birth and Spread

 

 

 

 

 

Mononuclear Growth

 

 

Polynuclear Growth

 

 

 

 

Figure 2. Surface expressions of some common crystal growth models.

17

 

 

 

growth rate

  
 

polynuclear
growth

 

W
Free Energy Drive

Figure 3. Diagram showing theoretical growth rates of spiral and polynuclear
growth as a function of free energy (after Sunagawa, 1984).

18

on a crystal surface free from defects because a certain degree of superaturation
is required to form a critical nucleus. The critical nucleus is the minimum size a
particle must attain to be stable in a supersaturated solution. Particles smaller
than the critical nucleus dissolve back into solution. The size of the critical
nucleus is directly related to the activation energy for nucleation. Nucleation on
a crystal surface is difficult because there is an energy barrier associated with
increasing the surface area and therefore the surface free energy of the crystal
(Nielsen, 1964; Ohara and Reid, 1973; Kirkpatrick, 1981; Lasaga, 1990). Once
this has occurred, however, crystal growth can proceed by attaching to the
nucleus because there is not such a large increase in surface area.
Mononuclear crystal growth is characterized by the attachment of a single
nucleus to a flat crystal surface followed by lateral growth along the periphery of
the nucleus. Each additional growth layer requires a new nucleation event to
occur. The nucleation rate is relatively slow compared to the rate of lateral
growth in the mononuclear model. Because each successive growth layer
requires the energetically difficult step of surface nucleation, mononuclear growth
is considered to be an inefficient growth mechanism (Nielsen, 1964). At higher
free energy drive, the birth and spread model is predicted by crystal growth
theory. In the birth and spread model, crystal layers are generated in a manner
similar to mononuclear growth, but instead of lateral growth occurring by
attachment at a single nucleus per layer, growth proceeds by attachment along
the periphery of multiple surface nuclei per layer. Similar to mononuclear growth,

the birth and spread model is characterized by a lateral growth rate that is higher

19

than the nucleation rate. In conditions where the degree of supersaturation is
sufficiently high, growth is predicted to occur in a manner consistent with the
polynuclear growth model. Polynuclear growth proceeds as multiple nucleation
events occur simultaneously on the crystal surface (Nielsen, 1964; Ohara and
Reid, 1973; Sunagawa, 1984). Polynuclear growth differs from the birth and
spread model in that the rate of surface nucleation is more rapid than the rate of
lateral growth.

As opposed to the former models, crystal growth by a spiral mechanism
does not require the energetically difficult step of surface nucleation. Instead,
spiral growth takes advantage of crystalline defects, like screw dislocations,
which create offsets on the crystal surface and in turn provide a continuous
supply of energetically favorable growth sites (Burton et al., 1951). As a result,
crystal growth can occur at conditions near equilibrium (Burton et al., 1951;
Sunagawa, 1984). According to Sunagawa (1984), spiral growth is the most
common crystal growth mechanism in nature.

Because different mechanisms add material to a growing crystal by
distinctly different means (e.g., by lateral attachment to a spiral dislocation arm or
surface nucleation), surface nanotopography is indicative of the dominant crystal
growth mechanism. In fact, studies investigating crystal surface nanotopography
have established a close relationship between the dominant crystal growth
mechanism and the resulting expression of the crystal surface nanotopography.
The most commonly observed surface nanotopographies are consistent with

models that describe 1) spiral growth centered on a dislocation and 2) surface

20

nucleation followed by lateral growth (Sunagawa, 1984; Hochella et al., 1990;
Hillner et al., 1992; Gratz et al., 1993; Hochella, 1990; Dove and Hochella, 1993;
Bosbach and Rammensee, 1994; Pina et al., 1998; Bosbach et al., 1998; Maiwa
et al., 1998; Kessels et al., 2000; Pina et al., 2000; Jiang et al., 2001a; Pina et
al., 2004). Moreover, observations of crystal surface nanotopography are
seemingly consistent with theoretical crystal growth models in that different
growth mechanisms operate under the relative thermodynamic conditions for
which they are predicted.

Dove and Hochella (1993) investigated in situ growth of calcite under
varying degrees of supersaturation with scanning force microscopy (SF M) at
25°C and 0.96atm. They observed the formation of surface nuclei ~20-30nm
wide and ~6-9 nanometers high in solutions with supersaturation Q>2. At lower
degrees of supersaturation, flat layers hundreds of nanometers across with 0.3
nanometer high steps were observed. Surface nuclei were reported to be
rhombohedral in shape, except in experiments where orthophosphate (a growth
inhibitor) was added, in which case nuclei were rounded in shape. After 10
minutes of continued growth in the solutions with orthophospate, calcite nuclei
coalesce into larger islands that were nearly 200 nm in diameter and up to 31 nm
high. The overall growth surface was highly irregular and covered with rounded
islands and small voids. After 1 hour of growth, Dove and Hochella (1993)
observed that the calcite surface became flatter, which was interpreted as the
result of coalescing islands (their Figure 3). Bosbach and Rammensee (1994)

made in situ observations of the surfaces of gypsum with the atomic force

21

microscope (AFM). They observed layers with steps >1000pm long and one to
nine monolayers (~0.7-6.1nm) high as well as a lack of surface nuclei for
supersaturations less than 1.1. At higher degrees of gypsum supersaturation
(>8), Bosbach et al. (1998) observed crystal growth proceeding by the formation
of islands up to 300nm long, <100nm wide and generally 0.7nm (1 monolayer)
high. Bosbach et al. (1998) demonstrated that barite surface nanotopography
was dominated by layers and spiral hillocks during crystal growth at low to
moderate supersaturations (0<5). These features were interpreted to result from
crystal growth by a spiral mechanism. At higher degrees of barite
supersaturation, surface nucleation was observed to dominate during crystal
growth. Island heights were equal to 0.35nm, consistent with the height of a
BaSO4 monolayer. Islands were euhedral in shape (their Figures 8, 10). After 5
minutes of growth islands spread and covered the entire surface resulting in a flat
layer. With the addition of a crystal growth inhibitor (NTMP), Bosbach et al.
(1996) observed that individual islands became rounded. Pina et al. (1998)
demonstrated that surface nucleation and growth spirals could occur
simultaneously during crystal growth in barite. Maiwa et al. (1998) observed
growth islands, which were interpreted as 2-D nuclei on the surfaces of well-
developed {111} and {100} faces of barium nitrate crystals. The islands were
interpreted to form when crystals were exposed to brief periods of elevated
supersaturation upon removal from the experimental solutions. The islands were
circular to slightly elongated. Islands were 0.41-0.5nm high, consistent with the

height of unit layers formed during spiral growth (their Figure 5). Upon continued

22

growth, nuclei coalesced with the adjacent layers formed earlier during spiral
growth. Several layers of different height were observed and were interpreted as
evidence for a birth and spread crystal growth mechanism. Pina et al. (2000)
showed that barite growth occurred by the advancement of steps generated on
screw dislocations when the degree of supersaturation was relatively low. In
contrast, they observed that surface nucleation was the dominant growth
mechanism at higher degrees of barite supersaturation. The degree of
supersaturation was shown to be a function of the solid composition in the
(Ba,Sr)SO4 solid solution series.

Kessels et al. (2000) reported a change in surface nanotopography on
dolomite crystals when all calcite was depleted at the end of their high-
temperature experiments. They observed that island nanotopography was
present throughout most of the reaction, but after reactant depletion dolomite
surface topography changed to broad flat layers. Round growth islands were
reported to be 25 to 200nm wide and 2-50nm high. Jiang et al. (2001a) reported
the occurrence of elliptical or irregular 2-D nuclei on the growth surfaces of
Cadmium Mercury Thiocyanate Crystals (0MTC). The 2-D nuclei were
described as having a height equal to the interplanar spacing for CMTC (0.81
nm). In a later paper, Jiang et al. (2001b) observed CMTC growth by layer
advancement at low concentrations, whereas two-dimensional nuclei readily
formed during experiments with relatively high degrees of supersaturation.
These observations were interpreted as being consistent with the spiral and

polynuclear growth models, respectively.

23

1.d. Chemical Etching

During the crystal growth process, surface features are continually
covered as new material is deposited. As a result, growth features ultimately
become part of a mineral’s internal crystalline microstructure. Because solid-
state diffusion is a relatively slow process, defect microstructures formed during
growth are likely to remain unmodified long after the mineral initially forms
(Reeder, 1992). It should therefore be possible to acquire information about the
details of crystal growth by examining a mineral’s microstructure. If surface
material can be removed in a manner reverse to growth, it seems likely that
imperfections created and utilized during growth can be detected (Sangwal,
1987a). In fact, workers were using chemical etching to detect defects in
crystalline solids as early as the late 1800’s (e.g., Becke, 1890). Since then, a
number of workers have employed chemical etching as a means to reveal
structural information concerning defects in crystalline materials (Horn, 1952;
Vogel et al., 1953; Sangwal, 1987a; Brantley et al., 1986; Kirkland et al, 1999).
Preferential dissolution occurs at imperfections because they are often
associated with strain and increased reactivity. Because a variety of defects may
occur during growth, the chemical reactivity of the crystal and the type of
dissolution that occurs at the defect are highly dependent on the nature of the
defect (Lasaga and Blum, 1986; Brantley et al., 1986; Sangwal, 1987a; Gratz et
al., 1991 ).

Etch-pits, the most commonly documented dissolution feature, result from

selective dissolution at the intersection of a defect and the crystal surface

24

(Johnston, 1962; Berner and Morse, 1974; Berner and Holdren, 1977; Berner,
1978, 1981; Lasaga and Blum, 1986; Brantley et al., 1986; Sunagawa, 1987;
Pande and Vaderbade, 1990; Stipp et al., 1994). Etch pits are commonly
polyhedral with a shape and orientation that is constrained by the structure of the
crystalline material (Sunagawa, 1987; Blum et al., 1990). The geometry of the pit
depends heavily on the nature of the lattice defect (Johnston 1962; Brantley et
al., 1986, Sangwal, 1987a). In the case of dissolution at an extended line defect
(e.g. screw dislocation), thermodynamic barriers to pit nucleation are overcome
more easily because of the excess strain energy that is associated with a defect
(Frank, 1951; Cabrera and Levine, 1956, Heimann 1982; Brantley et al, 1986;
Lasaga and Blum, 1986; Sangwal, 1987a). The linear nature of a dislocation
promotes continued dissolution at an angle to the crystal surface and the
resulting etch pit will be deep (Joshi et al., 1970; Maclnnis and Brantley, 1993).
According to several reviews (e.g., Heimann, 1982; Brantley et al., 1986; Lasaga
and Blum, 1986; Sangwal, 1987) the driving force for etch pit formation at
dislocation outcrops is contingent on: 1) the strain energy associated with the
dislocation, 2) surface free energy of the crystal, 3) impurities associated with the
line defect, and 4) the degree of undersaturation of the solution.

In contrast, shallow etch pits generally form at point defects. Once pit
nucleation at a point defect occurs, and the defect is eliminated, the energetic
drive for dissolution in a direction normal to the crystal surface is removed
(Brantley et al., 1986; Sangwal, 1987a). Most dissolution at point defects

therefore takes place in the lateral direction, which results in a broad shallow pit

25

(perhaps only a few monolayers deep). Upon continued dissolution complete
annihilation of the pit is likely (Johnston, 1962; Brantley et al., 1986; Maclnnis
and Brantley, 1993).

Berner and Holdren (1977) observed the formation of deep etch pits on
natural feldspar crystals when etched in hydrofluoric acid (HF). They suggested
that dissolution of the crystal occurred at etch pits that formed along dislocations.
Lasaga and Blum (1986) used Monte Carlo simulations to model etch pit
formation at dislocation outcrops. They noted that etch pit geometry was highly
dependent on the degree of undersaturation, the degree of surface diffusion, and
a number of other mineral-dependent parameters. Although the shape and size
of etch pits in their simulations were related to the degree of undersaturation, all
pits were relatively deep with relatively flat bottoms. Using TEM, Lin and Shen
(1993) observed dislocations in willemite that corresponded to {0001} direction,
consistent with the orientation of the observed etch pits that developed along the
c-axis. The pits they observed were tens of microns in size after 120 minutes of
dissolution. Etch pit density was about 10‘-105cm'2, which was consistent with
the relatively low dislocation density observed in transmission electron
microscopy (T EM) analysis. They interpreted these observations to indicate that
etch pits formed primarily at dislocation outcrops.

Hillner et al. (1992) studied in situ calcite dissolution with AFM. They
observed the formation of deep and shallow euhedral etch pits on calcite
surfaces after exposure to undersaturated solutions. They interpreted the deep

etch pits (40-90nm) to be dissolution at dislocations and the shallow pits (<5nm)

26

to be from dissolution at point defects. Gratz et al. (1991) used ex situ AFM
observations to study etch pit formation in quartz crystals. Pits were relatively
broad (1500nm wide, 20nm deep), but were interpreted as dissolution at
dislocation outcrops. They reasoned that dissolution at point defects would not
result in such deep pits (i.e. ~20nm) because the driving force for dissolution into
the crystal would be eliminated after a point defect was dissolved. Using in situ
AFM techniques, Liang et al. (1996) observed the dissolution surfaces of Iceland
Spar calcite. They observed the formation of two distinct types of etch pits. Type
1 pits were shallow (~0.3nm, 1 atomic layer) and interpreted to form at point
defects. Type 2 pits were deep (15nm, 40 atomic layers) and interpreted to
indicate dissolution at dislocations. The orientation and shape of the type 2
(deep) pits were crystallographically controlled.

Empirical observations suggest that, although not all dislocations form
deep etch-pits, there is good correlation between deep euhedral etch pits and
dislocations. (Vogel et al., 1953; Gilman et al., 1958; Johnston, 1962; Patel et al.,
1965; Joshi et al., 1970; Joshi and Paul, 1973; Heimann, 1975; McCIay, 1977;
Grandstaff, 1978; Reeder, 1982; Brantley et al., 1986; Lasaga and Blum, 1986;
Sangwal, 1987a,b; Murphy, 1989; Meike, 1990a,b; Pande and Vaderbade, 1990;
Blum et al., 1990; Gratz et al., 1991; Lin and Shen, 1993; Maclnnis and Brantley,
1992; Maclnnis and Brantley, 1993; Lin and Shen, 1995; Liang et al., 1996;
Bosbach et al., 1998). An extended discussion on the energetics of etch pit
formation can be found in reviews by Lasaga and Blum (1986), Brantley et al.

(1986), and Sangwal (1987a).

27

Although less commonly observed than etch pits, etch hillocks
(topographically high surface features) have also been identified as dissolution
features during etching experiments (Lin and Sheri, 1993; Zbik and Smart, 1998;
Kirkland et al., 1999). The formation of etch hillocks has been attributed to many
things, such as surface contaminations (Batterman, 1957), etch resistant point
defects (Stadler, 1963), reaction products (Tuck, 1975), and gas bubbles
attached to the crystal surface (Weyher and Van Enckevort, 1983). Although
their interpretation is more ambiguous than etch pits, it is generally accepted that
etch hillocks form due to suppressed dissolution on the crystal surface (Tuck,

1975).

28

2. METHODS

2.a. Dolomite Synthesis
Dolomite was synthesized according to equation 1 by combining a
calcium carbonate reactant with M9012 - 0a0l2 solutions in pre-weighed, Teflon-

lined stainless steel reaction vessels (bombs).

2CaC03 +Mg2+ <:> CaMg(C03)2 +Ca2+ (Equation 1)

All bombs initially contained 15 ml of a 1M MgClz - Ca0l2 solution mixture and
were heated to 218°C in a muffle furnace for various durations. Synthesis
experiments started with solid reactants consisting of either 1) 0.1 9 ground
Iceland spar calcite (40-60um size-fraction), or 2) a mixture of 0.059 Iceland spar
calcite (40—60um size-fraction) and 0.059 ground aragonite (12-22um size-
fraction). These experiments will be referred to herein as calcite reactant
experiments and aragonite reactant experiments, respectively. Because the
aragonite reactant experiments were originally designed to provide information
only regarding dolomite surface nanotopgraphy, a different size-fraction of
aragonite was used than in the calcite reactant experiments. Reactant surface
area was not determined for either reactant. Eight loaded reaction vessels were
placed in the oven at a time and remained there for times that ranged from hours
to weeks depending on the initial solution M92*:Ca2+ ratio. Reaction times were

estimated based on findings from previous experimental work done by the

29

current author as well as other published high-temperature studies. Individual
bombs were periodically removed from the oven and allowed to cool to room
temperature. Teflon bombs were re-weighed and the contents were filtered,
rinsed with distilled water, dried, and stored in a vacuum desiccator until
analyses. Reaction vessels that lost more than 5% of the initial weight were not
used in the analyses.

According to equation 1, the free energy drive with respect to dolomite is a
function of the Ca”: M92” in solution (activity of products/activity of reactants).
For the sake of simplicity, experimental solutions will described herein by the
initial M92*:Ca2+ ratio. This convention states that solutions with higher M92":Ca2+
ratios have a higher free energy relative to solutions with lower M92*:Ca2+ ratios.
The range of initial Mgz“:Ca2* ratios used in the experiments (0.43, 0.50. 0.66,
0.79, 1.0, 1.14, 1.27, 1.50) were obtained by mixing different proportions of 1M
0a0l2-2H20 and 1M MgCl2-6H20 solutions. Although exact M92*:0a2” activity
ratios cannot be directly measured in the bombs, the assumption is made that
aMg""’:aCa2+ ratios are directly proportional to the M92“:0a2+ molar ratios in
solution. Plotting the activity coefficients of calcium and magnesium using
Pitzer's Equation demonstrates that the behavior of these ions in solution is
similar at high ionic strengths under standard temperatures and pressures.
However, there is not an accepted way to accurately extrapolate Ca2+ and Mg2+

activity coefficients for the high temperatures used in these experiments.

30

2.b. X-Ray Diffraction

The percent products, as well as product compositions (dolomite
stoichiometry) and degree of cation order were determined by powder X-ray
diffraction (XRD). X-ray diffraction analyses were conducted using CuK,I
radiation on a Rigaku Instruments Diffractometer. Powdered fluorite (Can) was
added as an internal standard to calibrate diffraction peak positions. Data were
collected between 25-45° 29 using a step size of 0.004° with a 1 second count
time. Percent products provide a measure of reaction progress and were
determined according to the peak-height ratio of the dolomite dam) peak divided
by the dolomite dam) plus the calcite dam) peak (Royce et al., 1971). Both
tetragonal and hexagonal (i.e. three and four Miller-indices) conventions are
commonly used in the literature when describing crystal planes in minerals, such
as dolomite. Although both are equivalent nomenclatures, this paper will use the
tetragonal system (i.e. three-Miller index) for internal consistency.

Dolomite stoichiometry was calculated according to the relative corrected
position of the dam) peak, consistent with the methods of Goldsmith and Graf
(1958) and Lumsden and Chimahusky (1980). The 2-theta peak positions are a
function of the interplanar d-spacing in carbonates. Smaller 2-theta diffraction
angles correspond to larger interplanar d-spacings (measured in Angstroms).
Therefore larger d-spacings indicate calcium enrichment in dolomite because
larger calcium ions sustitute for smaller magnesium ions. The relationship
between dolomite stoichiometry and d-spacing is illustrated in equation 2 from

Lumsden and Chimahusky (1980).

31

Mole% CaCO3 = [(333.33 x d — spacing) - 91 1.99] (Equation 2)

This method assumes a linear relationship between dolomite composition
and the position of the dolomite dam) peak. Reeder and Sheppard (1984) argued
that this could lead to inaccuracies of up to 2-3 mole% 03003, particularly at
higher M9003 contents. Therefore Ca003 mole% can vary up to 1% and 3% for
Ca-rich and stoichiometric dolomites, respectively. According to Equation 2,
dolomite samples with relatively small d-spacings (i.e. large 2-thetas) yield
dolomite compositions enriched in magnesium. For example, a number of
dolomite samples are characterized by a 31 .0° 2-theta measurement, which
corresponds to a d-spacing of 2.8824 Angstroms. This d-spacing corresponds to
48.8 mole% 0a003 according to Equation 2. Because Mg-enrichment is very
rare in dolomite (Reeder, 2000) these dolomites are considered as
stoichiometric. The X-ray determined dam) peak position was reproducible to
within 0.006° 26. This variation leads to an approximate compositional error of
10.33 mole % 0a003. Because XRD measures all parameters in the bulk
sample, heterogeneities in dolomite stoichiometry and degree of cation order are
averaged. Furthermore, due to the resolution limits of powder X-ray diffraction, it
is difficult to detect mineral phases that constitute less than approximately 5
weight% of the bulk sample. Therefore analyses reported as 100% (products or
reactants) may contain impurities that are in abundances below the ~5% X-ray

detection limit.

32

To discriminate between dolomite and calcite structures, dolomite
superlattice reflections were used. Although this method is largely qualitative,
the degree of cation order in the dolomite lattice is commonly reported relative to
the ratio of the {015}:{110} reflection peaks (Fuchtbauer and Goldschmidt, 1965;
Goldsmith and Graf, 1958). The {015} is a principle order reflection that is less
pronounced for Ca-rich dolomites compared to ideal dolomites. The {110} is not
a dolomite ordering reflection so it serves as a reference for changes in order
intensity (Goldsmith and Graf, 1958). As a result, higher {015}:{110} ratios
indicate a higher degree of cation order in dolomite.

Thirty natural dolomite samples analyzed in this study range from
Ordovician to recent in age (T able 1). Compositions and degree of cation order
were determined according to the same procedures for high-temperature
synthetic dolomites. The natural dolomite samples represent a wide range of

geographic locations, ages, and inferred formation conditions.

33

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fomlatlonl 2- Mole°lo
Sample ID Age Group theta Ca003 SINS Orderlnl Pitsllslands
MH-2 Perm. Hueco 31 .00 50.00 S 1.13 pits
MH-6 Perm. Hueco 31.034 50.00 S 1.08 pits
MH-10 Perm. Hueco 31.022 50.00 S 1.08 pits
MH-11 Perm. Hueco 31.026 50.00 S 1.07 pits
MH-20 Perm. Hueco 31 .034 50.00 S 1.1 1 pits
MH-52 Perm. Hueco 31.038 50.00 S 1 .22 pits
SAL 7-5 0rd. Saluda 30.842 53.73 NS 0.82 islands
SAL 7-11 0rd. Saluda 30.87 52.75 NS 0.67 islands
7-9-99-3 0rd. Saluda 30.858 53.23 NS 0.54 islands
Seroe Domi,
SB 42 Plio. Bonaire 30.91 51.50 NS 0.60 islands
Seroe Domi,
SC 100 Plio. Bonaire 30.872 52.75 NS 0.57 islands
Seroe Domi,
SB 35 Plio. Bonaire 30.878 52.67 NS 0.73 islands
RR 17 0rd. Prarie DuChen 31 50.00 S 0.96 pits
RR 19 0rd. Prarie DuChen 31.006 50.00 S 1.00 pits
PdC 1 0rd. Prarie DuChen 31.046 50.00 S 1.33 pits
L2 91 0rd. Prarie DuChen 31.034 50.00 S 0.97 pits
Seroe Domi,
80-8-TS Plio. Aruba 30.85 53.37 NS 0.67 islands
Seroe Domi,
80-8-17-1 Plio. Aruba 30.85 53.37 NS 0.48 islands
Seroe Domi,

Aruba Dol Plio. Aruba 30.846 53.43 NS 0.46 islands
8—9-84-5 Mio. Spain 31 50.00 S 0.44 islands
8-9-84-3 Mio. Spain 31 .002 50.00 S 0.62 islands

13-4 0rd. Galena 31.06 50.00 S 1.14 pits
16-8 0rd. Galena 31 .034 50.00 S 1.04 pits
S-1 Silur. Lockport 31.034 50.00 S 1.07 pits
8-23-86-7 Silur. Lockport 31 .042 50.00 S 1 .04 pits
Burlington
D037 Miss. Keokuk 30.92 51 .33 NS 0.68 islands
Niagara Silur. Niagara 31.046 48.80 S 1 .00 pits
GWP287
123,33 Holo. Abu Dhabi 31.022 50.00 S 0.67 islands
FC 20 Miss. Kentucky 30.878 52.67 NS 1.17 islands
7-22-02-3 Plio. Bonaire 30.886 52.33 NS 0.56 islands

 

 

 

 

 

 

 

 

Table 1. Natural dolomite sample list and data table.

34

 

2.c. Atomic Force Microscopy

Atomic force microscopy was used to observe nanometer-scale features
on the surfaces of natural and synthetic dolomite crystals. Flat-sided dolomite
crystals were chosen with a binocular microscope for AFM analyses. Individual
crystals as well as clusters of dolomite crystals were mounted on magnetic AF M
disks using water-insoluble EM adhesive tabs. Synthetic dolomite crystal
surfaces are assumed to correspond to the dominant {104} growth face because
only the dominant rhombohedral form was observed. Natural dolomites were
cleaved and ultrasonically cleaned in distilled water prior to analysis. All cleaved
crystal surfaces are assumed to correspond to the dominant {104} cleavage
plane.

The atomic force microscope is capable of resolving atomic-scale surface
structures in air, under vacuum, and in liquids (0hnesorge and Binnig, 1993).
The AF M images surface features by measuring the changes between a rastered
sample and a cantilever spring with a fine tip. The cantilever, which remains in
constant contact with the sample, moves in the Z-direction as the surface
topography changes. Cantilever movements are imaged by a laser-beam that is
deflected onto a photodiode. In areas where the sample surface is relatively high
the cantilever is deflected more, thus an increase in topography is recorded by
the laser onto the photodiode. The entire system is linked in a feedback loop that
keeps the cantilever in constant contact with the rastered sample.

AFM analyses were performed on a Digital Instruments Nanoscope III

using constant force mode. Assuming the manufacture’s reported spring

35

constant of 0.06, the average applied force for the crystal surfaces was 1-5nN.
According to empirically determined force calibration procedures an actual
average applied force of 5-20nN was more reasonable. Although the force
between the sample and the cantilever was minimized as best as possible,
cantilever-surface interaction in this force range can potentially alter crystal
surface features (Kessels et al., 2000). Artifacts caused by cantilever-surface
interaction were not observed in any of the analyses. Moreover, no change in
dolomite surface nanotopography was observed for AFM scans of a single
surface that lasted over 20 minutes. Typical scan durations were between 30
seconds and 2 minutes depending on the scanning frequency. Image artifacts
may also arise during imaging due to irregularly shaped cantilevers, as well as
irregular cantilever movement across the crystal surface. Artifacts were detected
by sample rotation and/or changing cantilevers. All images reported in this study
were reproduced and replicated on more than one crystal per sample using
multiple cantilevers.

AFM images were acquired under atmospheric conditions using silicon
nitride cantilevers (NP-20) with an average tip radius (reported as radius of
curvature, R00) of 20 nm. According to the manufacturer (Veeco Probes), 1 in
10 NP-20 tips have an average R00 of 5nm. Many times during the image
collection process, numerous tips were discarded because of poor performance
before images were collected with a tip that provided better resolution. Eggleston
(1994) showed that the height and spacing of AFM resolvable steps could be

calculated by using the following formula:

36

2 2
R = (H + D-) (Equation 3)

2D

where, R is the radius of the cantilever tip, H is the height of a step, and D is the
distance between steps. This tip radius thus allows for near atomic-scale
resolution. Sometimes cantilever tips may contain irregularities or imperfections
that project outward from the tip that make the effective tip area in contact with
the surface even smaller. This can further enhance the resolving power of the
AFM.

The figures presented in this paper represent the highest-quality images of
the type of surface feature, which it is intended to illustrate. Although all results
were confirmed by observing numerous samples, changing AFM tips, as well as
observing multiple areas on any one crystal, the features reported may not
always represent the entire crystal. As a result, size ranges for the described
surface features are reported. Determination of average sizes for the various
surface features was not attempted here because it is impossible to demonstrate
that the images are representative of the entire sample.

Some natural and synthetic dolomite crystals were etched prior to AFM
analyses. Individual dolomite crystals were chemically etched in 0.5% H2S04
solutions (Johnson, 1962) between 10-120 seconds at room temperature.
Samples were then rinsed in distilled water and dried in a vacuum desiccator.
Caution must be used during chemical etching experiments because dissolution
features may reflect the chemistry of the etchant (Keith and Gilman, 1960;

Johnson, 1962) as well as the mechanism of dissolution (Berner, 1981; Pande

37

and Vadrabade, 1990; Morse and Arvidson, 2002). Etched surface topography
has been shown to be especially dependent on the rate of dissolution (Berner
and Morse, 1974; Berner, 1978). Slow interface controlled dissolution that
occurs near equilibrium, produces crystallographically controlled surface
features, such as etch pits and etch hillocks, (Lasaga and Blum, 1986), whereas
rapid dissolution may completely annihilate evidence of microstructure and leave
the crystal with smooth faces and rounded comers (Berner and Morse, 1974;
Berner, 1978; Sangwal, 1987). Folk et al. (1985) suggested that etching could
provide a way to reverse the crystal growth process. The purpose of using
chemical etching in this study is to accentuate crystalline microstructures that
formed during crystal growth. Therefore, the methods used here are used to
minimize artifacts caused by the etching process. The consistency of our results
suggests that the features observed here are in fact real features rather than

artifacts of etching procedures.

2.d. Scanning Electron Microscopy

Following AFM analysis, sample disks were adhered to conductive
scanning electron microscopy (SEM) stubs using double-sided carbon tape and
made electrically conductive with a 10-30nm gold coating. Some of the samples
also contained calcite and aragonite so each crystal was analyzed using SEM
Energy Dispersive Spectroscopy (EDS) following AFM analysis. Because most
samples only contain one dolomite phase (as identified by XRD), the presence of

magnesium peaks above background levels in EDS spectra were used to confirm

38

the presence of dolomite. SEM-EDS analyses were performed at 25kV on the
JEOL JSM-6400V scanning electron microscope. Other synthetic and natural
dolomite samples were prepared with a 2nm osmium coating and analyzed using
the JEOL 6300 field emission SEM (FESEM). Operating conditions were 8mm
working distance with an acceleration voltage of 10kV. All microscopic and EDS
analyses were preformed at the Michigan State University Center for Advanced

Microscopy (MSU-0AM).

2.a. Crystal Size Distributions

Crystal size distributions (CSD) were determined for dolomites
synthesized according to the procedures described above. The only difference
being that CSD dolomite experiments used different size solid reactants. Some
CSD experiments used reagent grade calcium carbonate with a solution
M92‘20a2“ ratio = 0.66 and others used Iceland Spar Calcite (140-160micron)
with a solution M92“:Ca2+ = 1.0. Following sample preparation protocol, dolomite
products were mounted on conductive SEM stubs with double-sided conductive
carbon tape. CSD were then determined by measuring the long axis of flat
dolomite rhombohedra along a linear traverse using scanning electron
microscope images. Between 140 and 502 crystals were measured for each
sample. Measuring protocol required dolomite rhombs to lay relatively flat on
SEM mounts and have at least three well-formed comers to be eligible for CSD

analysis.

39

3. RESULTS

3.a. Synthetic Dolomite Reaction
High-temperature dolomite synthesis experiments are characterized by a
relatively long induction period in which no dolomite is detected for some time
after the onset of experimental conditions. Once initial products begin to form,
however, the carbonate precursor is rapidly replaced by poorly ordered nonideal
dolomite. During the dolomite growth period, rapid replacement by products
occurs at an increasing rate. As reactants become depleted, the reaction rate
slows as the reaction approaches 100% dolomite. Near the point in which initial
reactants are completely consumed (i.e., reactant depletion), ideal dolomite
starts to form at the expense of nonideal dolomite. The overall dolomitization
reaction creates a sigmoidal-shaped reaction curve (refer to Figure 1),
characterized by an initial growth period that is relatively short compared to the
induction period. After reactant depletion, all products are generally ordered
stoichiometric dolomite.

Results from the dolomite synthesis experiments with calcite as the
reactant are recorded in Table 2a. Aragonite synthesis experiment data are

recorded in Table 2b.

40

Table 2a. Experimental results for high-temperature calcite experiments.

Solution Mg:Ca Sample ID Tlme(hrs.) %Prod.

0.43
0.43
0.43
0.43
0.43

0.5
0.5
0.5

0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66
0.66

0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79
0.79

6-1 7-04-D
6—1 7-04-E
6-1 7-04-F
6-1 7-04-G
6-1 7-04-H

4-9-04-A
4-9-04-B
4-9-04-0

4-8-05-A
4-8-05-B
4—8-05-C
4-8-05-D
4-8-05-E
4-8-05-F
4—8—05-G
4-8-05-H
8-23-04-A
8-23-04-8
8-2 3-04-D
8-2 3-04-E
8-23-04-F
8-2 3-04-G
8-23-04-H
1 0-3—04-A
1 0-3-04-B
1 0-3-04-C
1 0—3-04-D

4-1 5-05-B
4-1 5-05-C
4-1 5-05-D
4-1 5-05-E
4-1 5-05-F
4-1 5-05-G
4-1 5-05-H
1 0-20-04-A
1 0-20-04-B
1 0-20-04-C
1 0-20-04-D
1 0-20-04-E
1 0-20-04-F
1 0-20-04-G
1 0-20-04-H
1 020-044

594
762
930
1 122
1 362

219.5
259.75
315

60
61
62
63
63
64
65
66
96
1 00
145
168
1 94
21 7
241
288
309
334
356

46
48
49
56
58
59
61
64
65
65
65

0
3

100
100
100
100
100
100

8:833:on

50
62
47
51
30

90
81
43
1 00

41

M°l°% CaCO3

61.07
61.00
61.07
61.07

60.33
60.67
60.33

59.93
60.13
60.13

59.53
60.07
60.00
60.13
60.13
60.00
59.67
59.53
50.00
59.53
50.00
50.00
50.00
50.00

58.27
59.28
58.90
58.40
58.50
58.78
58.50
58.50
58.67
59.00
58.50
59.00
59.17
58.50
50.00

2-theta Ordering Ratio
30.600 0.00
30.606 0.00
30.600
30.600 0.00
30.618 0.00
30.612 0.00
30.618 0.00
0.00
30.640 0.00
30.630 0.00
30.630 0.00
0.00
30.650 0.10
30.634 0.08
30.638 0.00
30.630 0.00
30.630 0.00
30.638 0.09
30.646 0.27
30.650 0.36
30.970 0.70
30.650 0.45
31.000 0.65
31.000 0.77
31.000 0.73
31.000 0.88
30.690 0.00
30.658 0.00
30.670 0.00
30.686 0.21
30.682 0.13
30.674 0.10
30.682 0.00
30.682 0.00
30.678 0.00
30.660 0.00
30.682 0.17
30.666 0.20
30.662 0.14
30.682 0.00
31.000 0.80

Solution Mg:Ca Sample ID Tlme(hrs.) %Prod.

1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0

1.14
1.14
1.14
1.14
1.14
1.14
1.14
1.14
1.14
1.14
1.14
1.14

1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27
1.27

10-18-04-0
10-18-04-D
10-18-04-E
10-18-04-F
10-18-04-G
10-18-04-H
9-14—04-A
9-14-04-B
9-14-04-c
9-14-04-0
9-14-04-E
9-14-04-F
9-14-04G
9—14-04-H

4-12—05-A
4-12-05-B
4-1 2-05-C
4-12-05-D
4-12-05-E
4-12-05-F
4-1 2-05-G
4-12-05-H
1 0-1 3—04-E
1 0-1 3-04-F
1 0-1 3-04-G
1 0-1 3-04-H

4-6-05-A
4-6-05-8
4-6-O5oC
4-6-05-D
4-6-05-E
4-6-05-F
4-6—05-G
4-6-05-H
1 0-5-04-E
1 0-5-04-F
1 0-5-04-G
1 0-5-04-H
9-28-04-A
9-28-04-8
9-28-04-C
9-28-04-D
9-28-04-E
9-28-04-F
9-28-04-G

33
38
39
39
40
40
160
168
1 84
1 92
21 1
306
336
361

30.5
31.5
32.5
33
33.5
34
34.5
49
35
37
38
41.5

29
29.5
30
30.25
31 .25
32
48
53
30
31
32
33
38.5
40
40
45
48
62
1 17

0
15
71
70
95

8

100
100
100
100
100
100
100
100

68
98
32
1 00
95

67
85
1 7
97

99
1 00
1 00

42

Mole'l. 011003

56.17
55.83
55.83
56.17
55.33
50.00
50.00
50.00
50.00
50.00
50.00
50.00
50.00

53.23

53.97
54.73
54.13
53.63
53.67
53.73
54.17

52.33
52.33
52.75
51.83
52.67
51.13
53.77
52.00
52.00
52.33
51.67
52.00
52.67
52.33
52.67
51.33
50.00
50.00

2-theta Ordering Ratio

30.760
30.770
30.770
30.760
30.790
30.998
31.000
31.000
31.000
31.000
31.000
31.000
31.000

30.858

30.830
30.806
30.826
30.838
30.840
30.842
30.822

30.886
30.886
30.870
30.900
30.878
30.926
30.846
30.894
30.894
30.888
30.906
30.894
30.876
30.886
30.878
30.922
30.986
31 .000

0.00
0.10
0.11
0.17
0.00
0.48
0.50
0.51
0.72
0.77
0.76
0.76
0.82

0.00
0.00
0.00
0.00
0.00
0.00
0.25
0.36
0.00
0.10
0.11
0.00

0.00
0.00
0.00
0.22
0.20
0.00
0.33
0.30
0.00
0.09
0.14
0.00
0.30
0.18
0.25
0.23
0.39
0.54
0.69

Solution Mg:Ca Sample ID Time (hrs.) “I. Prod. Mole% 0a003 2-theta Ordering Ratio

1 .5 8-17-04-0 28.25 0

1 .5 9-4-04-A 28 14 50.00 31 .000 0.00
1.5 9-4-04-B 28 1 1 50.00 31 .000 0.00
1.5 9-4-04-0 30 14 50.00 31 .000 0.00
1 .5 9-4-04-D 32 9 50.00 30.986 0.00
1 .5 9-4-04-E 33 16 50.00 30.958 0.00
1 .5 9-4-04-F 34 65 50.00 30.962 0.08
1 .5 9-4-04-G 35 100 50.00 31 .000 0.40
1.5 9-4-04-H 36 93 51.33 30.918 0.31
1 .5 8-17-04-D 39 55 50.00 31 .000 0.00
1.5 8-13-04-A 77.5 100 50.00 31 .000 0.42
1.5 8-13-04-B 91 100 50.00 31.000 0.39
1.5 8-13-04-0 91 100 50.00 31 .000 0.37
1.5 8-13-04-D 115 100 50.00 31.000 0.55
1.5 8-13-04-E 149 100 50.00 31 .000 0.59
1.5 8-13-04-F 236 100 50.00 31 .000 0.65
1.5 8-13-04-G 236 100 50.00 31 .000 0.67

Table 2b. Experimental results for high-temperature aragonite
experiments.
Mole
Mg:Ca Sample ID Time (hrs) % Prod. %Ca003 2-theta Ordering Ratio

1 .0 4-5—05-A 1 .5 0 0.00
1 .0 4-5-05-B 2.25 0 0.00
1 .0 4-5-05-0 3.25 20 59.83 30.642 0.00
1 .0 4-5-05-D 3.75 50 61 .00 30.606 0.00
1 .0 4-5-05-E 4.5 80 58.33 30.686 0.00
1.0 4-5-05-F 5.5 90 57.17 30.726 0.00
1.0 4-5-05-G 6.5 90 56.83 30.738 0.10
1.0 4-5-05-H 8 100 57.00 30.734 0.10
1.0 3-31—05-A 14 100 52.50 30.882 0.16
1.0 3-31-05-B 16 100 51 .33 30.922 0.22
1.0 3-31-05-0 17.3 100 50.00 30.974 0.25
1.0 3-31-05-D 18.5 100 50.00 30.950 0.20
1.0 3-31-05-E 20 100 50.00 30.950 0.22
1 .0 3-30-05-A 22 100 50.00 30.930 0.35
1 .0 3-30-05-8 24 100 50.00 30.930 0.25
1 .0 3-30-05-0 25 100 50.00 30.978 0.30
1 .0 3-30-05-D 27 100 50.00 30.974 0.32
1 .0 3-30-05-E 30 100 50.23 30.954 0.20
1 .0 3-30-05-F 32 100 50.00 31 .000 0.26
1 .0 3-30—05-G 34 100 50.00 30.998 0.35
1 .0 3-30-05-H 46 100 50.00 30.974 0.39

Table 2a. Experimental results for high-temperature calcite experiments. 2b.
Experimental results for high-temperature aragonite experiments.

43

3. b. Calcite Reactant Experiments

3.b.1. Reaction Rate: Figures 4-11 are plots of time after onset of
experimental conditions versus the percent product (dolomite) for each solution
Mg“’*':0a2+ ratios studied here (0.43, 0.50. 0.66, 0.79, 1.0, 1.14, 1.27, 1.50,
respectively). The plots represent reaction curves for the individual calcite
reactant experiments. Each point represents the analyzed contents from an
individual reaction vessel. There is considerable scatter in the data that cannot
be explained by error in laboratory techniques. Because some reaction vessels
within a single experiment form dolomite faster than others, it appears as though
percent product decreases with time. In addition, some plots are incomplete
because certain reactions did not progress until complete depletion of reactants.
This generally results from extremely long reaction times required by solutions
with relatively low M92*:Ca2+ ratios (e.g. Mg"”’:0a2+ < 0.66). As discussed in the
introduction, many factors are responsible for variations in the rate of
dolomitization in high-temperature experiments. Even in these seemingly well-
controlled experiments there are still factors that can affect the system that have
not been accounted for. Sibley (1990) suggested that discrepancies in high-
temperature dolomitization experiments could be attributed to an inadequately
controlled variable that led to variation in nucleation and/or growth rates.
Because some of the more complete and well-constrained data sets presented
here and in other high-temperature dolomitization studies are very consistent,
generalizations have made regarding the reaction curves for some of the

experiments with more scatter.

44

 

1 00
90
80
70
60
50
40
30
20
1 0

Percent Product

 

Calcite Reactant Experiment, Initial Solution Mg”:0a"’ = 0.43

 

 

 

 

 

 

 

 

«D

 

 

O

A __ -——-———_.————-4
If V l I '_1—

0 200 400 600 800 1000 1 200 1400 .

Time (in hours)

 

Figure 4. Percent product vs. time for calcite reactant experiment with initial
solution Mgz*:0a2* = 0.429.

45

 

i Calcite Reactant Experiment, Initial Solution Mg”:0a2+ = 0.5 .

 

1 00
90
80 .2
70 .2
60 — —-

50
40
30
20
1 0

0 ‘1——_—_‘—"T'_— r 1— I

 

 

 

 

 

 

 

Percent Product

 

 

 

O O

 

 

 

T r 1

l 0 50 100 150 200 250 300 350
L Time (In hours)

 

Figure 5. Percent product vs. time for calcite reactant experiment with initial
solution M92*:0a2* = 0.50.

46

 

Calcite Reactant Experiment, Initial Solution Mgz’:0az’ = 0.66

I 120 4-----—---__~___~____e_. a- a- L -23---- -

 

 

1 00 - # —-+—+—-Fr-—+—o— 4

 

 

 

 

 

 

Percent Product
0)
O
I
I
I
I
I
I
I
I
I

 

 

 

L T
0 100 200 300
Time (in hours)

 

 

Figure 6. Percent product vs. time for calcite reactant experiment with initial
solution M92*:Ca2+ = 0.66.

47

 

Calcite Reactant Experiment, Initial Solution Mlgz“:0a2+ =

 

 

 

 

 

 

 

 

 

 

I
t 0.786
I 120
' 100 3

"3 80 i

E .

.é. 60 o e

O
i. 40 3.
0. O
20 .

I O
I 0 40% ,
I 0 50 100 150 l
I

Time (In hours)

Figure 7. Percent product vs. time for calcite reactant experiment with initial
solution M92“:Ca2+ = 0.786.

48

 

I ——--—-——— _.

Calcite Reactant Experiment, Initial Solution Mg”:0a” = 1.0

 

100 es es s *——-%——O—-+
90 —— --
80 ————————-— ———-— —-—-—— —-— m»
70 0 -2, L- — --—h—
50 -L____ ___ as. -_ LL __.-- -___ L L , 32,2212 _..”
50 **~~~~—W- 1* r1 _1 r“ rr—-—m*
40 m—wr 'ffl —— ~___.__,
30 —— ~~ ———-————————h-+#—A——-—-—~ ~-
20 — --——

I 10 '3 --— -— - -—- ———- —-— -*--——— ——

 

 

 

 

 

 

 

 

 

Percent Product

 

 

 

 

 

 

 

 

0 1 00 200 300
Time (In hours)

Figure 8. Percent product vs. time for calcite reactant experiment with initial
solution M92*:0a2” = 1.0.

49

 

120

Calcite Reactant Experiment, Initial Solution Mgz‘:0a2“ = 1.14

 

100

 

80

J.

 

60

 

40

 

Percent Product

20

 

 

 

 

I

40‘ . .

40 60
Time (in hours)

80

 

100

Figure 9. Percent product vs. time for calcite reactant experiment with initial

solution M92*:0a2* = 1.14.

50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

120

Calcite Reactant Experiment, Initial Solution Mgz*:0a2‘ = 1.27
100 . .k e e
O
90 Q ~
“ 80 _
0
g 60 —
9:, 50 .
c
3 40 __
E 30 3 _
20 ——-~4 -
O
10 . ~—
0 i e r T i I *1
0 20 40 60 80 100
Time (in hours)

Figure 10. Percent product vs. time for calcite reactant experiment with initial

solution M92*:Ca2+ = 1.27.

51

 

 

 

 

 

 

 

 

 

 

Calcite Reactant Experiment, Initial Solution Mg”:0a2+ = 1.5
100 —~—o———e—c——c-«——.s_._-E__L_.+_.
O
a 80 ’1— “m’fffifififiM—f—fi —- __
0
3 e
2 60 --—~——— mmum
a 0
5 40 WWW-*Lfi 3min" ..
2
a I
20 —— -— m - - WW -
I ’3
0 I s r l . ,- l
0 50 100 150 200 250

Time (in hours)

 

Figure 11. Percent product vs. time for calcite reactant experiment with initial
solution M92*:Ca2+ = 1.5.

52

One measure of reaction rate for an individual experiment is the length of
the induction period. As illustrated in the reaction curves, the dolomitization rate
is highest for experiments with relatively high solution M92*:Ca2" ratios and lowest
in experiments with relatively low solution M92*:Ca2" ratios. Figure 12a plots
length of the induction period (on a semi-log-scale) as a function of solution
M92“:Ca2+ ratio. Induction period results for calcite reactant experiments are
recorded in Table 3. Induction period data is also plotted on a linear scale in
Figure 12b. The shortest induction period in the calcite experiments was
approximately 28 hours, which was recorded for the experiment with an initial
solution M92*:Ca2+ = 1.5. Conversely, the longest induction period of 762 hours
was recorded in the experiment with an initial solution M92“:Ca2+ = 0.43. The
relationship between the induction period and Mgz’:0a2” ratio in solution is not
linear. When plotted on linear axes, the data are characterized by a curve that
becomes nearly asymptotic to the axes of the plot at low and high solution
M92":Ca2+ ratios (Figure 12b). For solutions with M92*:Ca2* 21.0, the length of
the induction period is shortened only slightly with increasing Mgz":0a2" ratios in
solution. Conversely, for solutions with M92*:Ca2+ < 1.0 the induction period is
lengthened significantly by decreasing the solution M92“:Ca2+ ratio. For example,
changing the solution M92*:Ca2” ratio from 1.0 to 1.50 shortened the induction
period from 35 hours to 28 hours, whereas changing the solution Mg"’*:0a2+ ratio
from 0.43 to 0.50 reduced the induction period from approximately 762 hours to
220 hours.

53

 

Induction Period vs. Initial Solution Mgz’:0a2"

 

 

 

 

 

 

 

 

 

 

 

I
I
I
1000 I
E .
3
o .
S .
.5 I
2 I
-: 100 — I
w .
a
c .
.2
s I
3 I
13
5 I
10 . 4 . . . . I
0.3 0.5 0.7 0.9 1.1 1.3 1.5 '

Initial Solution Mg”:0a2*

 

-_J

 

Figure 12a. Induction period (log-scale) vs. initial solution Mgz“:0a2+ for calcite
reactant experiments.

54

Induction Period vs. Initial Solution llr‘lgz":0a2+

 

 

”—7

 

910 ~
810 ~
710 1
610 ~
510 1
410 ~
310 d
210 ~
110 -
10 t l
0.4 0.9 1.4

i Initial Solution M92+:Ca2* Ratio

hours)

 

 

Length of Induction Period (in

 

 

 

Figure 12b. Induction period (linear-scale) vs. initial solution M92“:Ca2+ for calcite
reactant experiments.

55

Induction

0.43 762

0.50 219.5
0.66 61

0.79 48
1.0 38
1.14 33
1.27 29.5
1.50 28

 

Table 3. Induction period times for calcite reactant experiments.

56

3.b.2. Dolomite Stoichiometry: Figure 13 indicates a strong relationship
between dolomite stoichiometry for experiments ended prior to reactant depletion
and the initial solution M92“:Ca2+ ratio (R2 = 0.97). Experiments with a solution
M92*:Ca2" of 1.5, yield stoichiometric dolomite (50 mol% CaCO3), whereas
progressively lower solution M92*:Ca2“ ratios yield more calcium-rich
compositions, respectively. For example, the experiment with an initial solution
M92":Ca2+ ratio of 0.43 produced the most Ca-rich dolomite with an average
stoichiometry of 61 mole% 0a003.

Prior to reactant depletion, overall dolomite stoichiometry does not appear
to be related to percent product, which is shown in Figure 14 for all calcite
reactant experiments (R2=0.023). Dolomite stoichiometry also remains relatively
constant prior to reactant depletion in each of the individual solution M92“:0a2+
ratios. Figures 15 and 16 plot dolomite stoichiometry versus percent product for
experiments with initial solution Mgz*:0a2* ratio = 0.79 (R2=0.0013) and
M92“:Ca2+ ratio = 1.27 (R2=0.0005), respectively. The experiment with a solution
M92*:Ca2+ ratio = 1.14 also has a very low correlation between percent product
and dolomite stoichiometry (R2=0.0172). Two other experiments had significantly
higher correlations, namely the initial solution M92*:Ca2+ = 0.66 experiment (R2:
0.496) and the initial solution M92*:0a2” ratio = 1.0 experiment (R2=0.212).
However, dolomite products from these experiments have very minor
compositional ranges. The M92*:Ca2+ = 0.66 solution experiment, for example,
changed by only ~0.6 mole% CaCO3 (60.13 to 59.53) and the Mgz*:Ca2* = 1.0

solution experiment changed by only ~0.8 mole% Ca003 (56.15 to 55.35). This

57

 

Dolomite Stolchlolmetry vs. Initial Solution llllg"‘:Ca2+
63.00
61 .00 %\‘\

59.00 \

57.00

55.00 \ e
O

53.00 X

51.00 \;

47.00 t 4 r .
0.3 0.5 0.7 0.9 1.1 1.3 1.5

Initial Solution Mg”:0a2+

 

R2 = 0.9688

 

 

 

 

 

 

 

 

Dolomite Stoichiometry (mole %
CaCOa)

 

 

 

w

 

L

Figure 13. Dolomite stoichiometry vs. initial solution M92”:Ca2+ for all dolomites

formed prior to calcite reactant depletion, calcite reactant experiments,
R2=0.9688, N=69.

58

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dolomite Stoichiometry vs. Percent Product, Calcite
Reactant Experiments
33 65.00
% 63.00
g 61.00 a—Ofi. . O :0 ... e. . . .
E 59.00-fi—9—c r ”1r. 0 e ‘o’
E '3 57.00
0 0 ° 0 °
E g 55.00 . 9 g y .
5053.004 . T . . °’. 3"
:75 ° 0 9
51.00 Q—SL—
§ 0 e
E 49.00 -——+w e t
a: 47.00
45.00 , , . . ¥
0 20 40 60 80 100
Percent Product (Reaction Progress)

 

 

 

 

Figure 14. Dolomite stoichiometry vs. percent product for all dolomites formed
prior to calcite reactant depletion, calcite reactant experiments, R2=0.023, N=69.

59

 

 

 

 

 

 

 

 

 

 

 

 

Dolomite Stoichiometry vs. Percent Product, Calcite ,
Reactant Experiment, Initial Solution Mgz":0a2*=0.79 j
I
59.40
g e R2 = 0.0013 I
a 59.20 , I
E ’3 I
o
E 8 59.00 s s ,
.2 g 0 I
o
:75 32 58.80 1 I
8 3 O I
-- g 58.60 I
E v e e e e e
'5 58.40 c
D I. ‘
58.20 . . . . . I
20 40 60 80 100
Percent Product
I

 

Figure 15. Dolomite stoichiometry vs. percent product for dolomites formed prior
to calcite reactant depletion, calcite reactant experiment, initial solution M92*:Ca2”
= 0.79, R2=0.0013.

60

 

Dolomite Stoichiometry vs. Percent Product, Calcite Reactant I
Experiment, Initial Solution Mg”:0a"’ =1.27

 

 

 

 

 

 

 

 

 

 

 

. 54.00 I
3 R2 = 0.0005 o
0
g 53.50
g A 53.00 .
” 0 00
E 8 52.50
.2 is e e e o
O O
«7: 52.00 -—e s .-
3 e
-- 0
_§ 51.50
8 ‘I
51.00 . . . .
0 20 40 60 80 100
I Percent Product

I

Figure 16. Dolomite stoichiometry vs. percent product for dolomites formed prior
to calcite reactant depletion, calcite reactant experiment, initial solution M92*:Ca2”
= 1.27, R2=0.0005.

61

is in contrast to the dolomite products from the Mgz*:0a2+ = 1.27 calcite reactant
experiment, which randomly varied in stoichiometry by approximately 2.5 mol%
Ca003 (R2=0.005). Such a small range in dolomite stoichiometry (< 1 mole%
0a003) is also near the XRD instrumental error of :l:0.33 mole% 0a003.
Moreover, by removing the highest percent product data point (96% dolomite),
the correlation coefficient drops considerably to from 0.496 to 0.33. Experimental
runs with initial solution ratios below 0.66 were not plotted because these
experiments did not proceed to reactant depletion and there was insufficient
data. Although dolomite stoichiometry remains relatively stable prior calcite
reactant depletion, dolomite stoichiometry must increase rapidly because shortly
after reactants are depleted all products are stoichiometric dolomite regardless of
1) the initial solution M92*:Ca2+ ratio, and 2) the composition of the dolomite first
formed in the bombs.

The data also show that no correlation exists between the stoichiometry of
dolomite formed prior to calcite reactant depletion and time after the onset of
experimental conditions. Figures 17 and 18 plot dolomite composition as a
function of time for the solution M92*:Ca2* = 0.79 (R2=0.003) and 1.27
(R2=0.0097) experiments, respectively. These plots exhibit the lowest correlation
between dolomite stoichiometry and time. Although relatively minor, other calcite
reactant experiments, particularly M92":Ca2+ = 0.66 (R2=0.275), M92“:Ca2+ = 1.0
(R2=0.218), and Mt_iiz":0a2+ = 1.14 (R2=0.079) may indicate a potential correlation
between dolomite stoichiometry and reaction progress. Again, small ranges in

dolomite product compositions in the experiments with higher correlation

62

 

Dolomite Stoichiometry vs. Time, Calcite Reactant
Experiment, Solution Mg”:0a2*=0.79

59.40

 

R2 = 0.0003 .
59.20 ,

 

59.00 w—

 

 

58.80 a.

 

58.60

 

58.40 4

 

Dolomite Stoichiometry
(mole % 0a003)

 

 

58.20

 

 

 

 

 

30 40 50 60 70 ;
Time (in hours) I

I

Figure 17. Dolomite stoichiometry vs. time for dolomite formed prior to calcite
rezactant depletion, calcite reactant experiment, initial solution M92*:Ca2+ = 0.79,
R =0.0003.

63

 

Th. i Dolomite Stoichiometry vs. Time, Calcite Reactant Experiment, ,__

 

 

 

 

 

 

 

 

 

 

 

 

Initial Solution Mg”:0a” = 1.27
.\° 54.00
g R2 = 0.0097 0
E 53.50
g A 53.00
" ° 0 e e
E 8 52.50 4
3 8 ‘13 4—
% 52.00 4e e
03 O
,9 51 50 °
5 ' .
8 51.00 - . . . ° .
0 10 20 30 40 50 60

I Time (in hours)

Figure 18. Dolomite stoichiometry vs. time for dolomites formed prior to
calcite reactant depletion, calcite reactant experiment, initial solution M92”:Ca2+ =
1.27, R2=0.0097.

coefficients make it difficult to adequately assess the importance of the trend.
3.b.3. Cation Order: Dolomite ordering reflections first appear in XRD
patterns when initial products compose more than approximately 30% of the bulk
sample. Initial ordering reflections are attenuated, broadened, and shifted
relative to ideal superlattice peaks for dolomite. Samples with less than 30%
products exhibit no ordering peaks. Sibley (1990) argued that absence of
ordering peaks may not indicate lack of order because the small abundance of
products below 30% may result in low relative peak intensities that are not
detected. The first dolomite ordering reflection peak to appear in X-ray diffraction
patterns is at 33.5 degrees 2-theta. This is consistent with the observations of
Sibley (1990) and Kessels et al. (2000) who also reported the same observation.
The 33.5° ordering peak becomes increasingly visible as percent dolomite
increases from 30% to 100%. Additional ordering reflections are detected as
percent dolomite increases. For samples with greater than approximately 50%
products, the ratio of the d{105} and d{110} peaks is used as a measure of the
degree of cation order (Goldsmith and Graf, 1958; Fuchtbauer and Goldschmidt,
1965). The d{105} peak is absent in samples with less than about 50% products.
As a result, samples with less than ~50% product have no {105}:{110} ordering
ratios reported (refer to table 1). The {105}:{110} ordering ratio and thus the
degree of cation order increases with time. Dolomites formed prior to reactant
depletion have ordering ratios that range from 0.0 to 0.45. These dolomites are
considered relatively poorly ordered and are therefore referred to herein as

nonideal synthetic dolomites. In some experiments, the degree of cation order

65

increases relatively fast as calcite is replaced by nonideal dolomite, but appears
to slow after reactant depletion. This is illustrated in Figures 19 and 20, which
show the relationship between cation ordering and time for the calcite reactant
experiments with solution M92*:0a2” ratios of 1.0 and 1.5, respectively. The
slope in each of these curves shallows considerably after consumption of the
reactant, when dolomite products are stoichiometric. Other experiments do not
exhibit a change in sl0pe after reactant depletion. This can be seen in Figure 21
for the calcite reactant experiment with a solution M92*:0a2* ratio of 0.66. A
linear relationship (R2=0.926) is shown between the ordering ratio and time.
Following reactant depletion, all contents of the reaction vessels, regardless of
the initial solution M92“:Ca2+ ratio, are stoichiometric dolomite with ordering ratios
that range from 0.50 to 0.88. Because these dolomites are stoichiometric and
relatively well-ordered they are referred to herein as ideal synthetic dolomites.
Figure 22 shows little covariance between the degree of cation order
(ordering ratio) and the stoichiometry of nonideal dolomite formed prior to
reactant depletion (R2= 0.0342). Furthermore, cation order appears to be
independent of the initial solution Mgz*:Ca2* ratio. For example, sample 8-13-04-
G from the solution ratio M92*:Ca2+ = 1.5 experiment had an ordering ratio of 0.67
after 236 hours, whereas sample 10-20-04-l from solution ratio Mg"’*:0a2+ = 0.79

had an ordering ratio of 0.8 after only 79 hours.

66

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Cation Order vs.Time, Calcite Reactant Experiment, Initial
Solution Mgz+:0a2+ = 1 .0
0.9
O
0.8 . O .
0.7 °
6
E 0.6
or 0.5 .9 L
C
5 0.4
5 0.3
0.2 .
0.1 0
0 e I I r
0 100 200 300 400
Time

 

Figure 19. Cation order vs. time for calcite reactant experiment, initial solution
M92*:Ca2+ = 1.0. Reactant depletion occurs after ~160 hours.

67

 

 

 

 

 

 

 

 

 

 

 

 

 

Cation Order vs. Time, Calcite Reactant Experiment, Initial
Solution Mgz":0a2* = 1.5
0.8
0.7 :
o 0.6 0
'5 e
,9 0.5
g 0.4 4 3j
3 0.3 1
5 0.2
0.1 3
0 H I I . I
0 50 100 150 200 250
Time (in hours) "

 

 

Figure 20. Cation order vs. time for calcite reactant experiment, initial solution
M92*:Ca2* = 1.5. Reactant depletion occurs between 35 and 77 hours.

68

 

1.00
0.90
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00

Ratio

ring

Orde

0

 

 

____

= 0.9263

100

200
Time

Cation Order vs.Time, Calcite Reactant Experiment, Initial
Solution Mg":0a2’ = 0.66

300

 

 

400

Figure 21. Cation order vs. time for calcite reactant experiment, initial solution
M92*:Ca2” = 0.66. Reactant depletion occurs at ~217 hours.

69

l__ _. ___,_..__ _

 

Cation Order vs. Dolomite Stoichiometry, Calcite Reactant

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dolomite Stoichoimetry (mole% 0a003)

Experiments
| 0.45 2:
0.40 o R 0.0342
c
p 0.35 . °
E 0.30 ’4 0 .
a, 0.25 s. *
§020 -. . ’s
a 0.15 6 ° 5
0 ~ 0
I 0.10 . M—H‘
I 0.05 ~
I 000 WW
I 45.00 50.00 55.00 60.00 65.00
I
I

 

Figure 22. Cation order vs. dolomite stoichiometry for all synthetic dolomites
formed prior to calcite reactant depletion. Calcite reactant experiments, N=69,

R2=0.0342.

70

I
I
I

3.b.4. Solution Chemistry: Solution chemistry changes considerably as
dolomite replaces calcite in the high-temperature bombs. The M925032+ ratio in
solution is systematically lowered because magnesium is sequestered by
dolomite and calcium is liberated into solution from dissolving calcite. Table 4
illustrates these effects as a function of dolomite composition and reaction
progress for 50% product and 100% product. When 50% of the calcite has been
replaced, solution M92“:Ca2+ ratios are 10-13 percent lower than initial values
when the experiment started. And depending on the initial solution ratio,
M92“:Ca2+ ratios are further reduced by approximately 20-24 percent after

reactant depletion.

71

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Inltlal Solution Mole% CaCO3 In Final solution
Mg":0a’* °/. Products dolomlte Mg":0a’* °/. Change
0.429 50.0 61 .00 0.377 -11.94
0.500 50.0 60.25 0.443 -11.47
0.667 50.0 60.00 0.596 -10.64
0.786 50.0 58.00 0.701 -10.82
1.000 50.0 56.00 0.889 -11.08
1.143 50.0 54.00 1.011 -11.56
1.273 50.0 52.00 1.119 -12.11
1.500 50.0 50.00 1.308 -12.82
Initial Solution Mole% CaCO3 in Final solution
Mf’xa” % Products dolomite Mg":0a” % Change
0.429 100.0 61 .00 0.330 -23.05
0.500 100.0 60.25 0.390 -22.09
0.667 100.0 60.00 0.531 -20.41
0.786 100.0 58.00 0.623 -20.66
1.000 100.0 56.00 0.790 -21.00
1.143 100.0 54.00 0.894 -21.78
1.273 100.0 52.00 0.984 -22.68
1.500 100.0 50.00 1.143 -23.81

 

Table 4. Changes in solution M92*:Ca2+ chemistry at 50% products and 100%
products. Column three is the average dolomite stoichiometry observed in the
experiments for the initial solution M92*:Ca2" in column one.

72

 

3. c. Aragonite Reactant Experiments

3.c.1. Reaction Rate: The induction period is shorter for aragonite reactant
experiments than calcite reactant experiments. The induction period for
aragonite reactant experiments (solution Mgz“:Ca2+ = 1.0) is 3.25 hours (Figure
23), whereas products are first detected 38 hours into calcite reactant
experiments with the same solution M92“:Ca2+ ratio (refer to Figure 8). After all
aragonite is consumed at approximately 8 hours, all initial calcite, which
comprises half of the starting solid reactants, still remains. ln calcite reactant
experiments, it takes longer than 40 hours to dolomitize half the calcite reactants.
Polarizing light microscope (PLM), SEM, and XRD analyses indicate that
dolomite selectively replaces aragonite when calcite and aragonite are present
(Figure 24a). Calcite grains are generally not affected by dissolution as
evidenced by sharp corners and edges. Conversely, aragonite grains are
rounded and have pitted surfaces. Randomly oriented dolomite rhombs (Figure
24b) often form preferentially on the comers, edges, and steps of the aragonite
crystals. X-ray diffraction data further indicate that aragonite is preferentially
consumed during the aragonite reactant experiments. Samples from bombs with
progressively longer reaction times yield weakened aragonite diffraction peak
intensities, without a change in calcite peak intensities. Although it is
theoretically possible to determine an apparent change in the growth rate
between the calcite and calcite/aragonite experiments, the slopes of the reaction
curves are too steep and the data is not constrained well enough to compare the

reaction curves with any level of confidence.

73

 

T

1 00
90
80
70
60
50
40
30
20
1 0

0

Percent Product

 

I

0
0

0

Aragonite Reactant Experiment, Initial Solution Mgz*:0a2" = 1.0

 

n——————O——————H+O+—O-—O+—-O——O—O—O———————-—O—

 

1)

 

 

1D

 

 

 

A
w

0 10

20 30
Time (in hours)

40

Figure 23. Percent product vs. time for aragonite reactant experiment with initial
solution M92":Ca2” = 1.0.

74

50 pm

 

Figure 24a. SEM photomicrograph showing selective dolomitization of aragonite
when both calcite and aragonite are present.

75

- , 9' 3’20 pm

 

Figure 24b. SEM photomicrograph showing randomly oriented dolomite rhombs
on an aragonite substrate.

76

3.e.2. Dolomite Stoichiometry: Dolomite from aragonite reactant
experiments is more Ca-rich than dolomite from calcite reactant experiments with
the same solution Mgz"‘:0a2+ ratio. Calcite reactant experiments with solution
M92“:Ca2+ ratio = 1.0 generated dolomites with an average mole% Ca003 of 56,
whereas the aragonite reactant experiments produced dolomite with 57-61
mole% 0a003. Unlike the calcite reactant experiments, dolomite products
formed prior to aragonite depletion became more stoichiometric with time (Figure
25; R2=0.939). In addition, dolomite products from aragonite experiments remain
slightly Ca-rich after aragonite reactant depletion. For example, aragonite was
completely consumed in about 8 hours, but it took over 17 hours before
stoichiometric dolomite was detected in the aragonite reactant experiments.
Therefore it took nearly twice the reactant depletion time before dolomite
stoichiometric was detected. Because of short dolomitization times for aragonite
and long dolomitization times for calcite, all aragonite reactant experiments were
terminated prior to reaching calcite depletion. As a result no data was obtained
regarding the time to calcite depletion in these experiments.

3.c.3. Cation Order: Dolomite ordering ratios increase with time for
aragonite reactant experiments (R2 = 0.83; Figure 26). Cation order in dolomite
products from aragonite reactant runs, was however, significantly lower than in

calcite reactant experiments. Furthermore, ordering ratios were not observed

77

 

 

Dolomite Stoichiometry vs. Time, Aragonite Reactant
Experiment, Mgz’:Ca2’ = 1.0
65 00 R2 = 0.841
5 .. 63.00
S 5’ 61.00
f, a 59.00
“,6, ‘3 57.00
I; 3 55.00
E g 53.00
_9 .- 51.00
3 49.00
47.00
45.00 .
0 5 10 15 20 25 30 35 ;
Time (In hours)

 

 

Figure 25. Dolomite stoichiometry vs. time for dolomite formed up to 30 hours
after the onset of experimental conditions, aragonite reactant experiment, initial
solution Mg“”':0a2+ = 1.0, R2=0.841.

78

 

0.50

Cation Order vs. Time, Aragonite Reactant Experiment, Initial
Solution Mg"‘:0a2+ = 1.0

 

R2 = 0.8325

 

0.45
0.40

 

0.35

 

I»
e

 

 

 

 

 

 

 

 

 

 

I T

20 30
Time (in hours)

40

50

 

Figure 26. Cation order vs. time for aragonite reactant experiments, initial

solution Mg°*:Ca°*= 1.0, R2=0.8325.

79

until almost all aragonite was consumed (~90% products), whereas ordering
ratios are first detected at approximately 50% products in the calcite reactant
experiments. Dolomites formed after reactant depletion in aragonite reactant
experiments with initial solution M92":Ca2* ratios of 1.0, have ordering ratios
between 0.1 and 0.39. For calcite reactant experiments with solution Mg"’*:0a2+
ratios of 1.0, dolomites formed after reactant depletion have significantly higher

ordering ratios between 0.48 and 0.82.

3. d. Synthetic Dolomite Crystal Size Distributions
The crystal size distributions (CSD) for five different synthetic dolomite
samples are reported in Figures 27-31. Each histogram is characterized by a

coarse-skewed crystal size distribution. The number of dolomite grains that were

counted for each sample (N), as well as mean grain size of the sample (—X-) is

reported on the individual histograms.

3. 6. Synthetic Dolomite Surface Nanotopography, AFM

3.e.1. Calcite Reactant Experiments: Ex situ AFM observations of
synthetic dolomite growth surfaces reveal two principal features, described
herein as islands and layers. Islands are rounded positive relief features that
measure approximately 10-200nm in diameter and 1-20nm high (Figures 32 and
33). Islands occur pervasively on the growth surfaces of all nonideal dolomites
formed prior to reactant depletion. Individual islands are generally round in

shape without preferred orientation. Island topography is irregular in that the

80

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample 7-22-02-7, X=75.5, N=502
60
50
4‘“ III'
a 30 —
g 2,- IIII
IIIII
10 -
0 - IIIIIIIIIIII
Q N‘" 13’ la?" 45° 4° <8" Q%\\W\"l9\l>9,@°‘\6b
Size Bins (microns)

 

Figure 27. Coarse skewed CSD, sample 7-22-02-7. Reactant: 140-160 micron
size Iceland spar calcite.

81

 

Sample 1-19-87-1 , X=8.06 N=144

 

N
01

 

AN
0

001001
LII

 

_3

 

Fequency

 

 

 

 

l I 1.1 I l-TITT Il‘

oeeoggkewxmh$gc

I

I

I Size Bins (Microns)

 

Figure 28. Coarse skewed CSD, sample 1-19-87-1. Reactant: reagent grade
calcium carbonate.

82

 

i Sample 1-24-87-1 , X=5.26, N=135

 

 

N
O

 

A

 

 

Frequency

 

_\
U‘IOU'I

 

 

o rt, Ix 6 99.9 \1\u\6\%.19

Size Bins (microns)

 

 

Figure 29. Coarse skewed CSD, sample 1-24-87-1. Reactant: reagent grade
calcium carbonate.

83

 

 

 

 

 

Sample 1-17-87-2, x=5.75, N=183

 

 

 

N
O

 

 

 

 

Frequencey

 

 

 

OU'IOU'I
IllI

 

I Size Bins (microns)

Figure 30. Coarse skewed CSD, sample 1-17-87-2. Reactant: reagent grade
calcium carbonate.

84

 

 

F Sample 1-21-87-1, X=9.09, N=154

 

 

 

 

 

 

 

 

 

25
>‘20
O
515
5:110-
: 5 III!"

I ,. llllllIl .
o ‘b 1'0 ek‘tlfagwamergrboéb

Size Bins (microns)

 

 

Figure 31. Coarse skewed CSD, sample 1-21-87-1. Reactant: reagent grade
calcium carbonate.

85

300.0 nu

*‘7150.0 nu

     

0.0 nu
1'50
1.00
NanoScope Contact RPM
Scan size 1.644 nu

Sctpoint
0'50 Scan rat

0 U
c 10.17 Hz
Number of sauples 512

10201505.001

Figure 32. AFM image of a nonideal synthetic dolomite growth surface showing
pervasive island nanotopography. Calcite reactant experiment, solution
Mgz*:Ca2+ = 1.0.

86

 

Cursor Marker Spectrum Zoom Center Line Offset Clear

     

Section Al‘ialysis

800

Spec trun

I .1513-

Surface distance
Horiz distance
Uert distance
Rngle

08301437.001
Cursor: fixed Zoom: 2:1 Cen line: Off Offset: Off

Figure 33. AFM image of a nonideal synthetic dolomite showing cross-section of

island-covered growth surface. Calcite reactant experiment, solution M92":Ca2+ =
0.66

87

islands on dolomite surfaces do not lie horizontally across the same plane.
Instead, islands compose crystal surfaces that are uneven and hummocky
(Figure 34). Generally speaking, no pits or flat layers are observed between
individual islands except for an occasional irregularly shaped topographic
depression. Adjacent islands touch and are ubiquitous across the dolomite
crystal surface. Individual islands do not coalesce with other islands during
growth to form flat layers. Islands are observed on dolomites from samples in
which only a small amount of product has formed as well as samples that are
nearly 100% dolomite. Islands appear equally on dolomites from solutions with
high M92*:Ca2* and low M92“:Ca2+ ratios. For example, islands were observed
on a 30% dolomite with a composition of 61 mole% Ca003 from a 0.429
M92":Ca2" solution ratio as well as a 93% stoichiometric dolomite from a 1.5
M92“:Ca2+ solution ratio experiment. Islands are observed on all nonideal
dolomites prior to reactant depletion despite highly variable solution M92*:Ca2“
ratios, product compositions, and reaction rates. Furthermore, no measurable
geometric differences were observed for islands with different compositions,
solution chemistries, or reaction progress.

Islands were also observed on a few samples that had progressed passed
reactant depletion. These samples were removed from the oven at
approximately the time when all calcite was consumed. Therefore they may
contain a small amount of calcite that is below the XRD detection limit.

After reactant depletion, products are stoichiometric and relatively well-

ordered (ideal) dolomite. Ideal dolomite crystal surfaces display flat, laterally

88

Contact flFll
1.024 um
-1.000 0
6.104 Hz
umber of samples 512

   

 

09040833 . 001

Figure 34. AFM image of a nonideal synthetic dolomite growth surface showing
variable nanotopography caused by islands. Calcite reactant experiment, solution
M92*:Ca2" = 1.5.

89

extensive layers with steps (Figure 35) that measure a few tenths (~0.221nm) to
tens of nanometers high. With an approximate error of :i: 0.06nm (Kessels et al.,
2000) in the step height measurement, 0.221nm layers are similar to the d-
spacing for d{104} dolomite plane = ~0.3nm (1-monolayer). Due to the relatively
large error, it cannot be ruled out that the measured layers are 1/2 of a dolomite
monolayer (i.e. either a calcite-like layer or a magnesite-like layer). Layer
surfaces are extremely flat and do not have islands on top of them. Some layers
are separated by straight-sided edges and steps (Figure 35), whereas others are
defined by more wavy edges and steps (Figure 36). Despite variability in layer
geometry, edges and steps generally run parallel to one another. Although
classic multi-layer, helical growth spirals are not observed on dolomite surfaces,
individual layers are observed to emerge from flat crystal surfaces as predicted
by the spiral growth model. Figure 37a shows a flat surface with an emerging
step (towards the top of the image), which measures approximately 0.2 nm near
the offset. Figure 37b shows that the step completely disappears on the opposite
side of the emerging surface layer (towards the bottom of the image).

3.e.2. Aragonite Reactant Experiments: Dolomite formed in aragonite
reactant experiments have surfaces covered with islands (Figure 38). Islands
are present on dolomite surfaces throughout the reaction. Dolomite growth
islands in the aragonite reactant experiments are the same in size and shape as

those described for dolomite from calcite reactant experiments.

90

 

 

Center Line Offset Clear

Spectrun 200a

 

Section Armlysi's

Spectrun

Surface distance
Horiz distance
‘ Uert distance
angle
1 oudhuumhuuu
DC Hin

101?l348.001

'Cursor: fixed Zoom:
Figure 35. AFM image of an ideal synthetic dolomite growth surface showing
straight-sided steps. Calcite reactant experiment, solution Mgz":Ca2+ = 1.0.

2:1 Cen line: Off Offset: Off

91

Section Atmlysi'

3

200 300

400
nm

Spectrum

‘i.‘v~‘

Surface distance 34.918 nn
Horiz distance 34.841 nm
Uert distance 0.221 nm
‘ Rngle 0.384 deg
10210756.001
Cursor: fixed

 

      

Zoom: 2:1 Cen line: Off

Offset: Off

Figure 36. AFM image of an ideal synthetic dolomite growth surface showing
multiple layers with curved steps. Calcite reactant experiment. solution Mgz*:Ca2"
= 1.5.

92

 

Spectrum

Surface distance
Horiz distance
Uert distance

.7 _ ‘ Rngle

10210850.001
Cursor: fixed Zoom: 2:1 Cen line: Off Offset: Off

 

Figure 37a. AFM image and cross-section of an ideal synthetic dolomite growth
surface showing an emerging growth layer. Calcite reactant experiment, solution
Mgz*:Ca2* = 0.66.

93

 

Section Ai'mlysis

:00

Spectrum

Surface distance
Horiz distance
Uert distance
angle

10210850.001
Cursor: fixed Zoom: 2:1 Cen line: Off Offset: Off

 

Figure 37b. AFM image and cross-section of an ideal synthetic dolomite
showing the emerging growth layer disappears into the surface.

94

NanoScoPe Contact RFH
Scan size

Setpoint

Scan rate

Number of samples

X 200.000 nm/div
2 50.000 nm/div

 

Figure 38. AFM“ image of a nonideal synthetic dolomite showing pervasive island
nanotopography. Aragonite reactant experiment, solution Mg”: Ca2 = 1. 0.

95

3.f. Synthetic Dolomite Surface Nanotopography, SEM

3.f.1. Calcite Reactant Experiments: Dolomite rhombs form
preferentially at edges and corners in crystallographic continuity with the
dissolving calcite substrate (Figure 39). Figure 40 demonstrates the presence of
euhedral etch pits on the calcite substrate that form as a result of dissolution. As
the dolomitization reaction progresses, calcite becomes totally encrusted in
dolomite rhombs (Figure 41). After this occurs, individual dolomite rhombs
continue to nucleate on dolomite surfaces (Figure 42).

Nonideal synthetic dolomites are characterized by crystal surfaces with
what appears to be relatively flat areas separated by jagged and irregular ledges.
Figures 43a, and 43b are FESEM images, which show nonideal surface features
are generally irregular with rounded topographic highs and edges. The rounded
topographic features are on the order of tens to hundreds of nanometers in
diameter, consistent with the island topography observed with AFM in Figures
32-34.

The surfaces of ideal synthetic dolomite, on the other hand, generally
have flatter surfaces that are separated by straight-sided steps and ledges. The
areas in between layers are more flat. Edges and steps are more angular than
observed in nonideal dolomites. Figures 44a and 44b are FESEM images of
ideal dolomite crystal surfaces taken at the same magnifications used in figures
43a and 43b, respectively). Straight-sided steps appear to form from the
intergrowth of individual dolomite crystals and/or growth layers. The flat surfaces

with straight-sided steps are comparable to AFM observed steps in figures 35-37.

96

 

Figure 39. SEM photomicrograph showing crystallographically oriented dolomite
rhombs on the edges and corners of the calcite substrate.

97

 

Figure 40. SEM photomicrograph showing dolomite growth rhombs on a pitted
calcite surface that has been dissolved.

98

 

Figure 41. SEM photomicrograph showing a dolomite encrusted calcite grain.

99

 

Figure 42. SEM photomicrograph of a small dolomite rhomb on a dolomite
surface.

*Calibration bar = 10 micrometers

100

 

Figure 43a. SEM photomicrograph of a nonideal synthetic dolomite surface
showing irregular growth nanotopography.
*Calibration bar = 2 micrometers

101

 

Figure 43b. SEM photomicrograph of the same nonideal synthetic dolomite
growth surface at higher magnification showing rounded topographic highs and
wavy ledges.

*Calibration bar = 500 nanometers

102

 

Figure 44a. SEM photomicrograph of an ideal synthetic dolomite showing
relatively flat growth surfaces.
Calibration bar = 2 micrometers

103

 

Figure 44b. SEM photomicrograph of the same ideal synthetic dolomite at

higher magnification showing flat growth surfaces separated by straight—sided
steps.

*Calibration bar = 500 nanometers

104

The nanometer-scale surface features described above were observed on
samples coated with 2nm of osmium. Samples coated with 10-30nm of gold did
not reveal such differences between ideal and nonideal dolomite. The gold
coating, as shown in Figure 45, was too thick and covered up most nanometer-
scale surface features. I

3.f.2. Aragonite Reactant Experiments: Dolomite forms preferentially on
aragonite when both aragonite and calcite are present (refer to Figure 27a).
Dolomite grains have random orientation and nucleate on edges, steps and
corners (refer to Figure 27b). Calcite remains in the reaction vessels after

aragonite is consumed, but showed no signs of dolomitization.

3.9. Etched Synthetic Dolomite Surface Nanotopography, AFM

The surfaces of synthetic dolomite crystals chemically etched in dilute acid
are characterized by two distinct surface nanotopographies - dissolution islands
and euhedral etch pits on relatively flat layers. Dissolution islands are observed
on nonideal synthetic dolomites that formed prior to reactant depletion whereas
euhedral etch pits are observed on the surfaces of ideal synthetic dolomites that
formed after reactant depletion. Dissolution islands observed on the surfaces of
nonideal synthetic dolomite are indistinguishable from the growth islands
previously described. Dolomite growth islands shown in Figures 32-34 are
comparable to the dissolution islands in Figure 46. Dissolution islands occur
pervasively across the crystal surface and generally measure 10-200nm in

diameter and 1-20nm high. The sizes and shapes of individual islands are

105

 

Figure 45. SEM photomicrograph showing gold coating on a synthetic dolomite
grain.

106

20.0 nm

10.0 nm

 

1
i:
°.
8
0.0 nm
:3 1.00
3
0.75
ianoScope Contact AFN
.can size 1.238 um
oetpoint -1.000 U
o .can rate 104 Hz
lumber of samples 12

 

 

08301452 . 001

Figure 46. AFM image of an etched synthetic nonideal dolomite showing a
pervasive island dissolution surface. Calcite reactant experiment, solution
Mgz*:Ca"‘+ = 0.66.

107

variable and islands are not observed to have preferred orientation. Adjacent
dissolution islands are in contact with one another, but remain distinct features.
Occasionally, relatively large depressions and voids occur on dissolution island.
Synthetic ideal dolomite crystals etched in dilute acid have surfaces
characterized by relatively flat layers with deep euhedral etch pits (Figure 47).
Euhedral etch pits are commonly rhombic in shape and adjacent pits have
consistent shape and orientation. Therefore etch pits are inferred to be
crystallographically controlled. Etch pits produced on synthetic ideal dolomites
are between 30-200 nm in diameter and 1-30 nm deep. ln cross section, etch
pits are asymmetric and have pointed bottoms (Figures 48 and 49). Etch pit
densities are highly variable between samples. More interesting, however, is
that the distribution of etch pits is very heterogeneous in a single sample. Some
areas on the crystal surface are completely free of etch pits. This makes any
attempt at determining accurate etch pit density extremely difficult. Etch pits
become deeper and wider with increased exposure to etching solutions.
Occasionally, longer etching times resulted in islands being observed on some of
the etched ideal dolomites. These islands are also very similar to those
previously described for growth and dissolution surfaces of nonideal synthetic
dolomite. Uncovering islands during etching was more common on ideal
dolomites with moderately low ordering ratios from experiments that had recently
proceeded to reactant depletion than on ideal dolomites with higher ordering that

had long since proceeded to reactant depletion.

108

NanoScope Contact AFM
Scan size 2.
Setpoint

Scan rate

Number of samples

01181037.001

 

Figure 47. AFM image of an etched ideal synthetic dolomite showing euhedral
etch pits. Calcite reactant experiment, solution Mg2 :Ca2 =0 66.

109

 

Cursor Marker Spectrum 200m Center Line Offset Clear

 

 

Section Analysis

L 90.539
RHS 7.25?
lo DC
Rallcl 0.988
Rmax 5.937
R2 2.281
R2 Cnt 8

750

Spectrum

Surface distance 95.064
Horiz distance 90.539
Uert distance 20.382
angle 12.675

09301228.001
Cursor: fixed Zoom: 2:1 Cen line: Off Offset: Off

Figure 48. AFM image of an etched ideal synthetic dolomite showing a high

density of euhedral etch pits. Calcite reactant experiment, solution Mgz":Ca2+
1.5.

110

Pm
nm

TIM

DH
rlH

Cursor

     
        

"arker Spectrum Zoom Center Line Offset Clear

Section Analysis

400

Spectrum

{i .‘rifi

Surface distance
Horiz distance
Uert distance
angle

10220904.001
Cursor: fixed

 

Zoom:

   

2:1 Cen line: Off Offset: Off

Figure 49. AFM image of an etched ideal synthetic dolomite showing euhedral

etch pits and relatively flat inter—pit surfaces. Calcite reactant experiment, solution
Mgz":Ca2+ = 1.27.

111

3. h. Natural Dolomite Surface Nanotopograph y, AFM

Chemically etched natural dolomite crystal surfaces are also characterized
by the presence of islands and flat layers with euhedral etch pits. Islands occur
on the surfaces of all chemically etched natural nonstoichiometric dolomites as
well as stoichiometric dolomites with relatively low (<0.67) degrees of cation
order (i.e. nonideal). Natural dolomite dissolution islands are pervasive across
the crystal surface (Figure 50). Individual islands are rounded topographical
highs that measure 20 to 300 nm in diameter (average ~60-80nm) and of 1 to 20
nm high (Figure 51). Island-covered natural dolomite surfaces have areas of
lower topography, but euhedral etch pits and flat layers with steps are absent
(Figure 52). Dissolution islands on natural dolomites are very similar to those
observed on synthetic dolomite crystal growth and dissolution surfaces.

The chemically etched surfaces of ideal natural dolomite exhibit euhedral
etch pits with flat inter-pit layers separated by steps (Figure 53). Steps are
generally straight-sided and when layers stacked on one another, steps are
roughly parallel to one another. Euhedral etch pits are commonly rhombic or
near rhombic in shape with sharp edges (Figures 54 and 55). Near-by etch pits
have the same shape and orientation, which suggests that they are
cyrstallographically controlled. Individual pits are 20 to 500 nm wide and 5 to 90
nm deep with pointed bottoms (Figure 56). Etch pit densities on natural dolomite
are highly variable. Some areas have no observable etch pits, only flat faces. In
addition, the distribution of etch pits is very heterogeneous in most samples. As

stated previously, this makes reporting etch pit densities rather ambiguous.

112

NanoScope Contact RFH
Scan size 7
Setpoint

Scan rate

Number of samples

X 200.000 nm/diu
2 250.000 nm/div

 

Figure 50. AFM image of an etched natural nonideal dolomite showing
pervasive island-covered surface.

113

    

            

 

 

 

 

 

Marker Spectrum 200m renter Line Offset ClPai
M Section Anolysrs
8'
L 48.30? nm
_ iims 3.183 nm
° lc DC
Ra(lc) 0.828 nm
Rmax 4.037 nm
R: 2.74? nm
B_ R: Cnt 6
I | l l
r': zrio 400 soo 800
nm Surface distance 53.382 nm
" Spectrum Horiz distance(L) 48.307 nm
Uert distance 10.328 nm
angle 12.068 deg

Surface distance
Horiz distance
Uert distance
angle

Surface distance
Horiz distance
Uert distance
angle

_ Spectral period
nc Hln Spectral freq
Spectral BUS amp

Offset: Oflrr . it

 

05011059.001
fixed

 

 

   
 

  
   

     
 
 

   

     

 

  

     

Cursor: Zoom: 2:1 Cen line: Off
Figure 51. AFM image and cross-section of an etched natural nonideal dolomite
showing dissolution island nanotopography.

114

NanoScope Contact RFH
Scan size

Setpoint

Scan rate

Number of samples

{Hifiwmflfi

0 1.00

04170813.001

 

Figure 52. AFM image of an etched natural nonideal dolomite showing an
island-covered surface with voids and depressions.

115

05081428.001

NanoScope Contact RFH
Scan size 3
Setpoint

Scan rate

Number of samples

 

Figure 53. AFM image of an etched natural ideal dolomite showing euhedral

etch pits with flat inter-pit layers.

116

NanoScope Contact RF"
Scan size 3.
Setpoint

Scan rate

Number of samples

12201204.001

 

Figure 54. AFM image of an etched natural ideal dolomite showing
crystallographically oriented rhombic etch pits.

117

04170856.001

NanoScope Contact RFH
Scan size

Setpoint

Scan rate

Number of samples

 

Figure 55. AFM image of an etched natural ideal dolomite showing euhedral

etch pits with rhombic shape.

118

 

n. Section Analysis

Spectrum

 

DC
05081428.001

 

L 71.244 nm
RNS 8.258 nm
lc

Ra(lc) 1.368 nm
Rmax 5.777 nm
3: 5.777 nx
Rz Cnt 2

 

 

 

 

 

Surface distance 79.028 nm
Horiz distancetL) 71.244 nm
Uert distance 30.469 nm
angle 23.155 deg
Surface distance
Horiz distance
Uert distance
angle

Surface distance
Horiz distance
Uert distance
angle

Spectral period
Spectral freq
Spectral HHS amp

 

 

Figure 56. AFM image and cross-section of an etched natural ideal dolomite

showing euhedral etch pits with pointed bottoms.

119

Upon continued dissolution, the surfaces of ideal dolomite are covered with

wider, deeper etch pits.

3.i. Natural Dolomite Surface Nanotopography, SEM

Unetched natural dolomite surfaces observed with SEM are characterized
by a variety of features including cleavage fractures, variously shaped voids,
relatively flat surfaces, steps, and debris. This is demonstrated in the SEM
image of an ideal dolomite crystal shown in Figure 57. Ultrasonic cleaning and
brief periods (30-60 seconds) of chemical etching in 0.5% sulfuric acid removes
most surface debris, but doesn’t otherwise affect the overall surface topography.
Ideal and nonideal dolomites with short etching times exhibit very similar surface
features even when viewed at high magnification. Ideal dolomites have relatively
flat surfaces with steps and voids. Nonideal dolomites also have relatively flat
surfaces with steps and voids (Figure 58). SEM observed dolomite surfaces are
highly irregular, yet the commonly observed surfaces features occur equally on
all samples. Longer periods of chemical etching (2-15 minutes) did not
accentuate ideal and nonideal dolomite surface nanotopography. In general,
dolomite surfaces become much more irregular and pitted over longer etching
times. Most dissolution voids observed with SEM are irregular and anhedral in
shape, although occasionally some do have straight edges. However, none of
the dissolution voids are euhedral like the AF M observed etch pits. For longer
etching times (e.g., 10 minutes) some dolomites maintain flat surfaces with a low

density of voids, whereas some nonideal dolomite surfaces become highly

120

 

Figure 57. SEM photomicrograph of an unetched natural dolomite surface
showing voids (V), steps (S), and debris (D).

121

 

Figure 58. SEM photomicrograph of an etched natural nonideal dolomite
surface. Arrows mark voids (V), steps (S), and debris (D).

122

irregular with many voids (Figure 59). Some dolomite crystals tend to dissolve
much more rapidly at the center relative to the outer edges. The SEM
observations described here are highly variable and as a result, provide no

definitive way of distinguishing between ideal and nonideal dolomite surfaces.

123

 

Figure 59. SEM photomicrograph of a natural ideal dolomite showing a highly
irregular and dissolved surface after 10 minutes of etching in dilute acid.

124

4. DISCUSSION

4.a. Islands and Layers Are Growth Features:

The islands and layers observed on the surfaces of synthetic dolomite crystals
are interpreted to form during crystal growth, despite the experimental design
permitting the possibility that they form by dissolution. Because dolomite, like
other carbonate minerals, has retrograde solubility (Nordeng and Sibley, 1994;
Langmuir, 1997), dissolution may occur as high-temperature experimental
conditions are terminated as bombs are removed from the oven. Prior to
reactant depletion, however, solution chemistry is likely buffered by the faster
dissolving calcite and/or aragonite reactants (Busenberg and Plummer, 1982).
Using the AF M fluid-cell, Kessels et al. (2000) observed that dolomite islands
remained stable, whereas adjacent calcite surfaces became covered with
dissolution pits as distilled water flowed over the surfaces. This suggests that
dolomite growth surfaces are more stable than calcite surfaces in experimental
solutions undersaturated with respect to the two minerals. The empirical etching
results also indicate that upon dissolution, nonideal dolomites remain covered by
islands. This suggests that islands are pervasive throughout the dolomite
crystals and are not only formed as a surface feature as reaction vessels cool.
Furthermore, the development of euhedral etch pits on etched surfaces of
synthetic ideal dolomites demonstrates that islands do not form by dissolution of

growth layers.

125

It is also possible to initially interpret islands as an amorphous precipitate
that forms upon cooling of the reaction vessels. Not only are the synthesis
experiments carried out within the thermodynamic stability field of dolomite, but
amorphous phases would likely have a higher solubility than the underlying
crystalline dolomite. As a result, any amorphous phases on the surface would
easy dissolve during chemical etching experiments and therefore reveal the
underlying dolomite surface growth topography. Again, the chemical etching
experiments demonstrate that island nanotopography persists after continued
dissolution, further suggesting that islands are pervasive throughout the dolomite
crystals and are not superficial precipitates that form when experimental
conditions are ended.

Due to the limited resolution of the AFM, growth islands may actually be
growth spirals with steps too small to be resolved. However, if mounds are
centered on defects (e.g., spiral dislocations), chemical etching should result in
etch pit formation at the center of the mound where the defect would be located,
because these are areas of enhanced dissolution. Etching of island-covered
surfaces uncovers additional islands, whereas etch pits do not form on any of the
etched nonideal dolomites.

Layers on the surfaces of ideal synthetic dolomite are observed only in
calcite reactant experiments in which all reactants have been consumed.
Following reactant depletion, dissolution of ideal synthetic dolomites is not
buffered by the presence of calcite, therefore ideal dolomites are even more

susceptible to dissolution than nonideal dolomites once experimental conditions

126

are terminated. Two factors argue against the idea that layers form as a result of
dissolution however: 1) The ideal dolomite formed after reactant depletion is
considerably more stable than nonideal dolomite because of its stoichiometric
composition and relatively high degree of cation order, and 2) Flat layers without
etch pits are inconsistent with models of interface-controlled dissolution.
Busenberg and Plummer (1982) showed that dolomite dissolution rates were
independent of advection rates under the experimental conditions pH 0-10, 1.5-
65°C, Pcoz 0-1atm. Furthermore, the chemical etching results clearly
demonstrate that euhedral etch pits form upon dissolution of ideal dolomite. Both
observations suggest that over a wide range of conditions, dolomite dissolution is
an interface-controlled phenomenon. It can also be reasoned that because 1)
chemical etching leads to the development of etch pits on ideal dolomite
surfaces, and 2) no euhedral etch pits are observed on layer surfaces during the

growth experiments, any dissolution during the growth experiments is

insignificant.

4. b. Polynuclear Growth

The growth islands observed on sedimentary minerals like calcite,
gypsum, barite, dolomite, and other crystalline materials, are commonly used as
evidence for polynuclear crystal growth (Hillner et al., 1992; Dove and Hochella,
1993; Bosbach and Rammensee, 1994; Bosbach et al., 1994, 1998; Pina et al.,
1998; Heraty, 1998; Pina et al., 2000; Kessels et al., 2000; Jiang et al., 2001a,b;

Pina et al., 2004). During polynuclear growth, simultaneous nucleation events

127

out-pace lateral attachment to already existing surface features (Nielsen, 1964;
Ohara and Reid, 1973). Because nucleation is the most energetically demanding
step in the crystal growth process, polynuclear growth is associated with crystal
growth in solutions where the degree of supersaturation is relatively high
(Nielsen, 1964; Ohara and Reid, 1973; Kirkpatrick, 1981; Sunagawa, 1984;
Lasaga, 1990). In supersaturated solutions, small critical nuclei (r*) are
continuously deposited on the crystal surface. Some nuclei attain critical size
(2r*) and are therefore stable, whereas other nuclei <r* dissolve back into
solution. The size of the critical nucleus (r*) is dependent on the activation
energy of nucleation, which is a function of the supersaturation of the solution
and the surface free energy of the nucleus (Nielsen, 1964; Walton, 1969; Ohara
and Reid, 1973). Once a nucleus of critical size is deposited on a crystal
surface, it may have material attach to its periphery or it may be covered by
subsequent nucleation events. As a result, surfaces created during polynuclear
growth may be characterized by the presence of growth islands with highly
variable sizes.

To establish the similarities between dolomite growth islands observed
here and those previously observed on the surfaces of other minerals that have
been interpreted as polynuclear growth, it is necessary to outline and review
some of the observations that are pertinent to such a comparison, like the size,
shape, behavior, and conditions under which growth islands generally occur.

Dove and Hochella (1993) reported that rhombohedral growth islands,

which measured 6-9 nm high formed during calcite growth experiments at

128

relatively high degrees of supersaturation (O<2). Upon continued growth, islands
coalesced into even larger islands that measured approximately 200nm in
diameter and up to 31nm high. Based on theoretical growth models, Dove and
Hochella (1993) interpreted calcite islands as evidence of a polynuclear growth
mechanism. They also observed that growth islands were rounded and less
euhedral in growth experiments in which a growth inhibitor was used. They
suggested that the added growth inhibitor disrupted growth processes, thus
resulting in a compromised crystal form. Using in situ AFM techniques, Pina et al.
(1998) observed the formation of euhedral growth islands on barite surfaces
under relatively high degrees of supersaturation ((1:12). Islands had straight
edges that were oriented in crystallographic continuity with barite crystal faces.
Growth islands measured one monolayer high and up to hundreds of nanometers
wide. Pina et al. (1998) interpreted islands to be surface nuclei that formed
during a polynuclear growth mechanism. Heraty (1998) observed the formation
of large calcite growth islands during crystal growth experiments. She observed
the formation of growth islands in solutions that were highly supersaturated with
respect to calcite (O=1.6 to 10.5). Islands were greater than 50nm high and up
to hundreds of nanometers in diameter. Heraty (1998) also demonstrated that
the addition of sulfate and phosphate changed growth island morphology from
euhedral to round and anhedral (Figure 60). Bosbach et al. (1998) reported the
appearance of barite islands nearly 1pm in diameter during growth experiments
at relatively high degrees of supersaturation. Growth islands were interpreted as

evidence of crystal growth by surface nucleation (i.e., polynuclear growth). In

129

addition, Bosbach et al. (1998) observed that nucleation rates were reduced and
island morphology changed dramatically after the addition of a growth inhibitor to
barite growth experiments. Comparison of the two island morphologies revealed
that inhibitor-free environments exhibited sector-zoned (euhedral) islands,
whereas islands grown in solutions with a growth inhibitor had rounded islands
without euhedral edges. Porsche et al. (1998) observed the formation rounded
“quantum islands" measuring approximately 50-100 nm in diameter and 4-18 rim
high that occurred pervasively across the crystal surfaces of synthetic InP and
GalnP during growth experiments. Vennegues et al. (1998) studied GaN growth
on sapphire crystals. TEM cross-sections showed that growth islands with a
pyramidal shape were 50nm to 600nm in diameter and 30nm to 250nm high.
Kessels et al. (2000) observed rounded growth islands on the surface of high-
temperature synthetic dolomite growth surfaces Individual islands measured
<20nm high and <200nm wide. Astilleros et al. (2002) observed rounded growth
islands forming on calcite surfaces that were tens to hundreds of nanometers in
diameter and formed pervasively across crystal surfaces in the presence of Mn, a
proposed growth inhibitor. Astilleros et al. (2002) reported a transition from
layers to islands on freshly cleaved calcite crystal after changing the
concentration of Mn solution. Nuclei were rounded relief features tens to hundred
of nanometers in diameter and 1 monolayer in height. Upon continued growth,
islands became broader and the surface density of islands increased. Islands

eventually coalesced to form layers.

130

 

Figure 60. AFM image from Heraty (1998) showing amorphous growth islands
on a calcite surface in phosphate-doped experiments.

131

If magnesium hydration acts as a growth inhibitor to dolomite growth,
rounded, anhedral growth islands are consistent with the findings from many
other crystal growth studies. Garrels et al. (1960), Hartman (1982), and
Lippmann (1973) all argued that the high hydration energy of Mg2+ is a kinetic
inhibitor of dolomite formation. In order for dolomite to form, Ca2+ and Mg2+
cations must attach to the crystal surface followed by dehydration and
carbonation of each. Calcite has been shown to grow rapidly at relatively low
temperatures (Dove and Hochella, 1993), whereas the findings of Sayles and
Fyfe (1973) showed that magnesite was very difficult to grow in the lab at low
temperatures. Kazmierczak et al. (1982) reported the activation energy for
precipitating calcite to be 9.4 :i:0.9kcal mol". Based on the assumption that
calcite and magnesite grow by identical mechanisms, Arvidson and Mackenzie
(2000) reported the activation energy of magnesite to be 22.2 kcal mol'1 based
on experimental data from Lippmann (1973). This is a difference of 12.8 kcal
mol'1 higher than the value for calcite. Arvidson and Mackenzie (2000) argued
that if dolomite formation is viewed as a process of creating alternating layers of
calcite and magnesite, then the rate of dolomitization is controlled by the
formation of the magnesite layer. Because excess calcium increases the free
energy of dolomite (Navrotsky and Capobianco, 1987; Chai et al., 1995) calcium
will be incorporated if at the given conditions, dehydration of calcium requires
less energy than does the dehydration of magnesium (Arvidson and Mackenzie,
2000). Because the activation energy of dolomite is nearly twice that of

magnesite based on data from Arvidson and Mackenzie (1997, 2000), cation

132

ordering, as well as dehydration and carbonation of the Mgz" cation could both
be rate-limiting steps in dolomite formation.

A common observation from the calcite, gypsum, and barite growth
experiments is that after continued growth, islands often coalesce to form layers
(Hillner et al., 1992; Dove and Hochella, 1993; Bosbach and Rammensee, 1994;
Bosbach et al., 1994, 1998; Pina et al., 1998; Astilleros et al., 2002). Although
layers are observed on dolomite growth surfaces, they appear only after reactant
depletion when islands cease to form. These layers are interpreted to form by
spiral growth and not from coalescing islands. Individual dolomite islands are
much greater than monolayers high and are highly variable in size. If two
adjacent islands have slightly different chemical composition or cation order, it
would be energetically difficult for the islands to completely grow together
because the interface between the islands would have an associated free
energy. This would be analogous to soap bubbles that are situated adjacent to
one another. The bubbles may collide, but maintain their individuality clue to the
associated surface tension. Islands may also not form layers, like other minerals,
because bombs are removed from the oven in while rapid surface nucleation is
still occurring. The dolomite reaction curves show that once dolomite begins to
form, it replaces reactants very rapidly. Because of this, islands are constantly
being formed on the surface, which may prevent the formation of flat layers.
Before multiple islands coalesced to form a flat layer, continued nucleation is
likely to occur and deposit more islands on the surface, therefore covering up a

flat layer. This is interesting because, upon cooling, the saturation state of

133

dolomite is increased and conditions tend toward equilibrium. Thus one would
expect that the potential for layers to form at the end of an experiment is greater
as the chemical conditions change. This is not the case, however; layers are
never observed on the surfaces of nonideal dolomites, which form prior to
reactant depletion. Heraty (1998) observed island nanotopography in calcite
growth experiments persisted during extended periods of growth in the presence
of a phosphate growth inhibitor. After two hours of growth at a high degree of
supersaturation, the sample was covered with rounded calcite growth islands.
Layers from coalescing islands were never observed, but a highly variable
surface topography due to different patches of growth islands (figure 69) with
variable lateral dimensions and vertical relief was reported as being common.
The observation that synthetic dolomite islands remain the dominant
surface feature during chemical etching also supports the interpretation that
islands remain pervasive and distinct throughout the crystal. Following this idea,
one can also infer that because polynuclear growth occurs throughout the
formation of nonideal dolomite and a defect microstructure results from the
mismatch between individual islands. This is consistent with the observations of
Penn and Banfield (1998), who reported on the formation of dislocations at the
interfaces between differently oriented Ti02 (titantia) growth nanoparticles.
Nanoparticles measured 5 to 6nm in diameter according to x-ray diffraction peak
broadening analysis and HRTEM. Penn and Banfield (1998) showed that during
crystal growth, individual titania nanocrystals attached to the crystal growth

surface with slight differences in orientation. Dislocations were subsequently

134

observed to form at these interfaces when adjacent particles were not
anatomically aligned. Banfield et al. (2000) suggested that defects between
nanoparticles could fundamentally alter the way crystals grow by creating growth-
related spiral dislocations for example. The observed nanoparticles may be
comparable to the growth islands observed on dolomite surfaces. If there were
no mismatch between islands then there would be no energetic drive for
dissolution to occur in between individual islands.

Because dolomite islands are consistent in size, shape, and behavior to
other previously observed growth islands, they are interpreted in a manner

consistent with these studies, as evidence for a polynuclear growth mechanism.

4.c. Spiral Growth

The layers observed on the surfaces of ideal synthetic dolomite crystals
are consistent with a number of theoretical growth models. Crystal growth theory
predicts that spiral growth dominates at conditions where the free energy drive is
relatively low because energy barriers to surface nucleation are circumvented by
lateral attachment of growth units to the arm of a screw dislocation. As pointed
out by Sunagawa (1977, 1981), spiral growth is the most common growth
mechanism found in nature. The most commonly cited feature used as evidence
for spiral growth is a multi-layer helical spiral, which occurs on a crystal growth
surface (Hillner et al., 1992; Gratz et al., 1993; Dove and Hochella, 1993;
Bosbach and Rammensee, 1994; Bosbach et al., 1994, 1998; Pina et al., 1998,

2004; Jiang et al., 2001a). Although layers occur on ideal synthetic dolomite

135

surfaces after all calcite reactants have been depleted, no multi-Iayer helical
spirals were observed. However, individual layers were observed to merge into
crystal surfaces (refer to figures 37a and 37b), thus producing a surface feature
geometry that is consistent with an emerging screw dislocation (see figure 2).
The emerging layers on dolomite growth surfaces are consistent with those
observed at spiral dislocation outcrops on other crystalline solids (De Yoreo et
al., 1997; Maiwa et al., 1998; Pina et al., 1998; Jiang et al., 2001). Pina et al.
(1998) observed that emerging layers on the surfaces of barite later became the
sites of developing multi-layer growth spirals after further growth in conditions at
relatively low supersaturation. In the case of synthetic dolomite, multi-layer
growth spirals may not be as common as in other minerals, for numerous
reasons. Individual layers produced at growth spirals may converge before a
large helical feature is produced. Additionally, the Burgers vector of the
dislocation may be very small, which could cause spirals to be more laterally
extensive and less vertically prominent. Paquetle and Reeder (1990, 1995) both
showed that at low to moderate degrees of supersaturation, single growth spirals
on calcite commonly spread laterally for hundreds of micrometers, but may rise
less than 5 microns above the crystal surface. Assuming calcite and dolomite
are similar in this respect, it would be impossible to image laterally extensive
dolomite growth spirals in their entirety because the scan sizes used in AFM
would not be large enough.

Layers are also consistent with the mononuclear growth model. Recall,

however, that mononuclear growth is predicted to be operational at conditions in

136

which the degree of free energy drive is relatively low, yet the mechanism
requires the energetically difficult step of surface nucleation for each new layer
that forms on a crystal surface. According to Nielsen (1964), surface nucleation is
extremely difficult at conditions near equilibrium and as a result, mononuclear
growth is a relatively inefficient growth mechanism. Furthermore, the observation
that multiple dolomite growth layers with roughly parallel steps that are stacked
on top of one another is inconsistent with the mononuclear growth model. The
model is developed on the premise that only one nucleation event occurs per
layer (hence the term mononuclear) because it is far easier to continue attaching
growth units to energetically favorable step and kink sites on an already existing
layer than to nucleate new layers.

Kessels et al. (2000) suggested that the stacked layers they observed on
synthetic dolomite may form in a manner consistent with the birth and spread
model. Birth and spread is somewhat of a hybrid between polynuclear growth
and mononuclear growth in that it is characterized by multiple nucleation events,
but is ultimately out-paced by the lateral growth component. In this case,
attachment of new growth units along the periphery of different surface nuclei
creates layers that grow laterally until they intersect and fuse with other layers.
Therefore, the birth and spread model could explain some of the geometries
observed on synthetic dolomite surfaces, such as stacked layers. In the birth
and spread model, individual surface nuclei are responsible for providing
attachment sites for new material. Because of this, one would expect to see

layers forming at the different nuclei. However, separate, laterally extensive

137

growth layers within the same plane were not observed. Layers on dolomite are
instead stacked on top of one another with steps that are roughly parallel in a
manner similar to the periphery of a helical spiral dislocation (Dove and Hochella,
1993; Gratz et al., 1993; De Yoreo et al., 1997; Maiwa et al., 1998; Bosbach et
al., 1998; Pina et al., 1998; Jiang et al., 2001). Even more important to the
interpretation that layers form by spiral growth is the observation that layers have
steps that merge into crystal surfaces and that deep euhedral etch pits develop
on etched layer surfaces.

The interpretation that layers are evidence of a spiral growth mechanism
is further supported by the fact that growth layers are observed in conjunction
with V-shaped euhedral etch pits that form as a result of chemical etching.
Euhedral etch pits are hundreds of monolayers deep and are therefore indicative
of dissolution at an extended line defect, such as a screw dislocation (see
reviews by Heimann, 1982; Brantley et al., 1986; Lasaga and Blum, 1986;
Sangwal, 1987). Extended line defects are the site of preferential dissolution
because defects have an associated excess free energy (Lasaga, 1990). As a
result, dissolution can occur more readily along the defect, which results in a V-
shaped etch pit. Point defects, such as precipitates and inclusions may also be
sites of etch pits because they too provide an area of elevated reactivity. Point
defects, however, only form shallow etch pits because once the pit forms and the
point defect is eliminated, there is no drive for further dissolution in a direction
normal to the crystal surface. Joshi et al. (1970) and Joshi and Paul (1973)

observed the formation of two types of etch pits in synthetic quartz; deep

138

pyramidal pits and shallow flat-bottomed pits. Upon prolonged etching they
noted that the pyramidal pits grew deeper, but that the shallow pits disappeared.
They argued that shallow, flat bottom pits corresponded to point defects whereas
the deep pointed pits formed at extended line defects. Furthermore, Lin and
Sheri (1993) showed with TEM that dislocations in willemite corresponded to etch
pits that formed on the surface of etched crystals.

Etch pit distributions on synthetic and natural dolomites were observed to
be highly variable. Furthermore, most samples had very heterogeneous
distributions of etch pits. This is consistent with the findings of Reeder (1982)
who reported that dolomite dislocation densities observed with TEM appeared
variable and heterogeneous. At this point, it is important to note that deep, V-
shaped etch pits may equally form at surface outcrops of edge dislocations as
well as a spiral dislocations. Crystal growth naturally introduces any number of
lattice defects, but screw dislocations are the only known defect that can facilitate
crystal growth. Both edge and screw dislocations may form during growth or
deformation, but screw and edge dislocations have very different roles during
crystal growth. Spiral (screw) dislocations generally take an active role during
growth, whereas edge dislocations, although potentially formed during growth, do
not actively participate in the crystal growth process. Etch pits are commonly
oriented normal to the growth surface and in many instances, they are slightly
tilted and asymmetric. Klapper (1980) showed that spiral dislocations are
commonly linear to sublinear and are most often oriented within ~15 degrees of

the growth normal in a particular growth sector. As Reeder (1992) points out,

139

dislocations of this type are common in dolomite. Reeder (1982, 2000) also
reported that almost all dislocations are aligned nearly normal to the growth
surface in dolomite.

Edge dislocations may also form from processes not associated with
growth. In natural crystals, deformation is probably the most frequent cause of
this (Reeder, 1982). The synthetic dolomites studied here were never exposed
to conditions in which significant deformation would occur. A number of workers
using TEM have reported that ideal dolomites have relatively homogeneous
microstructures with only dislocations and growth defects (Reeder, 1981; Blake
et al., 1982). Reeder (1992) observed that other than in cases where geologic
histories preclude high temperatures and stresses, many of the microstructures
in sedimentary carbonates have resulted from crystal growth. Barber (1977)
noted distinct differences in the dislocation microstructure of dolomite compared
to other sedimentary carbonate minerals. Most importantly, dolomite has been
observed with TEM to have a significantly lower defect density than limestone or
calcite. These differences were attributed to the strong crystallographic control of
dislocation orientation (noted above) and to the fact that dolomite has fewer
deformation mechanisms than calcite. Rosen et al. (1989) reported a higher
density of dislocations in the ideal dolomites compared to the nonideal dolomites
with a heterogeneous microstructure that he studied. His results suggest that
extended line defects may be more common in ideal dolomites whereas a more

heterogeneous microstructure is consistent with a pervasive defect structure.

140

Radke and Mathis (1980) proposed a model to explain the distinctive
macroscopically visible curved faces on saddle dolomite in which Ca substitution
for Mg was accompanied by the introduction of edge dislocations. Subsequent
TEM work by Reeder (1982) of saddle dolomite showed that many saddle
dolomites were characterized by a pervasive array of stacking faults. In addition,
parallel growth dislocations were associated with the stacking faults. Reeder
(1982) interpreted these to represent active screw dislocations within the different
growth sectors.

Although theoretically possible, based on the current observations there is
no unequivocal method to distinguish between an etch pit that formed at a screw
dislocation, which was utilized during crystal growth and a pit that formed at an
edge dislocation. However, the presence of euhedral V-shaped euhedral etch-
pits in conjunction with layers that merge into the crystal observed on synthetic
dolomite growth surfaces supports the interpretation that ideal dolomites grow
primarily by a spiral growth mechanism. In as much as there is no practical way
of distinguishing between euhedral etch pits formed by dissolution at screw
dislocations that facilitated growth and other types of dislocations, the possibility
is very likely that some of the euhedral etch pits are related to edge and perhaps
inactive screw dislocations. Based on the above arguments, deformation seems
to be an unlikely scenario to explain the introduction of dislocations in dolomite.
Furthermore the orientation of most line defects in dolomite, as viewed with TEM,
seem to be consistent with that of screw dislocations. These observations in

conjunction with the layers observed on synthetic dolomite growth surfaces leads

141

to the conclusion that a significant proportion of euhedral etch pits on etched
dolomite surfaces are related to spiral dislocations that facilitated crystal growth.

Otherwise it is extremely difficult to explain how these crystals grew.

4. d. Transition From Polynuclear To Spiral Growth

Crystal growth theory predicts that polynuclear growth and spiral growth
dominate under different thermodynamic conditions. Polynuclear growth
operates under conditions where the free energy drive is relatively high (Nielsen,
1962; Ohara and Reid, 1973; Sunagawa, 1984). Because polynuclear growth is
limited by the formation of surface nuclei, the free energy drive in solution must
be great enough to overcome energetic barriers to surface nucleation in order to
grow by polynuclear growth. Spiral growth, on the other hand, is predicted for
conditions in which the free energy drive is relatively low. Because spiral growth
occurs at energetically favorable step and kink sites created on the crystal
surface by screw dislocations, the energetically difficult step of surface nucleation
is circumvented. In addition to theoretical arguments as a framework for
discussing competing growth mechanisms, relative free energies of the crystal
surfaces provide a foundation to explain why islands form at relatively high free
energy drive and layers form at relatively low free energy drive. Because islands
are characterized by a rougher crystal surface compared to layers, islands are
indicative of a higher surface free energy (Kessels et al., 2000). Regardless of
whether islands nucleate on crystalline defects, like dislocations or defect-free

surfaces, surface features with greater surface area require a higher free energy

142

drive to be stable. Therefore, one might expect the growth mechanism transition
during the calcite reactant experiments (from islands to layers at reactant
depletion) depicted in figure 61 to correspond to a change from relatively high
free energy drive to a relatively low free energy drive in the system.

4.d.1. Energy: According to the experimental design (refer to equation 1),
the reactants (calcite and/or aragonite) provided the only source of carbonate for
dolomite formation. From pH measurements, Sibley (1990) inferred that calcite
dissolution in experimental bombs comes to a steady state within approximately
1 hour at room temperature. At elevated temperatures calcite dissolution is
faster, therefore experimental conditions should achieve equilibrium with respect
to calcite relatively fast. As a result, hydrothermal bombs are likely to be
saturated with respect to calcite (or aragonite) before dolomite begins to form.
Because calcite solubility is assumed, the solution M92“:Ca2+ ratio dictates the
free energy drive in the reaction vessels. According to LeChatlier’s Principle, a
higher solution Mgz*:Ca2” ratio corresponds to a higher degree of free energy
drive for the calcite to dolomite replacement reaction (see equation 3) because
increasing the reactants should drive the reaction forward, thus producing more
product. To this end, higher solution Mgz*:Ca2" ratios have a relatively high free
energy drive and lower solution Mg”':Ca2+ ratios result in a relatively low free
energy drive for the calcite to dolomite replacement reaction. Based on crystal
growth theory, one would predict polynuclear growth to occur at relatively high
solution Mgz":Ca2+ ratios and spiral growth at relatively low solution Mgz*:Ca2"

ratios. Over the entire range of solution Mgz”:Ca2" conditions studied here,

143

however, polynuclear growth dominates until reactant depletion when all
products are ordered, stoichiometric dolomite. Even the experiment with the
lowest initial Mgz*:Ca2* ratio (i.e., 0.43) forms dolomite by polynuclear growth
prior to reactant depletion. Furthermore, the highest solution Mgz“:Ca2+ ratio (i.e.,
1.5) invariably makes the transition from polynuclear to spiral growth after the
calcite is completely consumed. It should be noted at this point that experiments
with low solution Mgt’*:Ca2+ ratios (e.g., < 0.66) did not proceed to complete
reactant depletion over the duration of the experiment and as a result, never
approached the point where the transition from islands to layers is expected.
Perhaps one of the most interesting results from these experiments is that in
spite of theoretical predictions, crystal growth mechanisms in dolomite are not a
function of the Mgz*:Ca2+ ratio in solution.

A measure of free energy drive for dolomite formation can also be inferred
from the reaction rates for the various experiments. Slow reaction rates should
indicate conditions nearer to equilibrium, whereas rapid dolomitization should
indicate conditions with a relatively high free energy drive. The data show that
relatively low solution Mgz*:Ca2* ratios corresponded to slow reaction rates and
relatively high Mgz”:Ca2“ ratios correspond to fast reaction rates. Therefore,
reaction rates are consistent with the solution Mg’*:Ca2+ ratios as an indication of
the free energy drive in the hydrothermal bombs. In the case of low Mg“"‘:Ca2+
ratios, it was difficult to produce dolomite even over very long periods of time. In
the experiment with solution Mgz*:Ca2" = 0.43, for example, one experiment took

approximately 1362 hours to form 30% dolomite, compared to only ~28 hours for

144

the experiment with a solution Mgz“:Ca2+ ratio = 1.5. Below a solution ratio of
Mgz":Ca2‘ = 0.66, the time required to form dolomite increased exponentially with
a decreasing solution Mgz*:Ca2“ ratio (see Figures 12a and 12b). Extrapolating
the curve of reaction rate vs. solution Mgz“:Ca2+ ratio, shows that an initial
solution Mgz‘2Ca2“ = 0.43 requires a very long time to produce dolomite.
Presumably, these long reaction times indicate that an initial Mgz":Ca2+ value of
0.43 is seemingly close to the calcite-dolomite equilibrium or at least close
enough to it where it is impossible to replace calcite with dolomite on reasonable
laboratory time—scales (e.g., months). Furthermore, as the reaction progresses,
solution Mgz“:Ca2+ ratios decrease as dolomite forms at the expense of calcite.
Dolomite sequesters Mgz“ from solution and liberates Ca2+ into solution as the
reaction proceeds. Take for example the experiment with an initial solution
Mgz":Ca2* of 0.43 (sample 6-17-04-G). At the time of reaction termination (1122
hours), 66% products had formed. As outlined in Table 3, the Mgz*:Ca2” ratio in
solution would have dropped to below 0.38. The actual calculation yields a
solution Mgz*:Ca2“ ratio of 0.36 when 66% of the calcite is replaced by dolomite
with 61 mole% CaCOa. Therefore a Mgz*:Ca2+ ratio of 0.36 is near, but still
above the dolomite-calcite stability boundary under the given temperature
conditions. Morrow et al. (1994) determined log (aCa2+ I aMgz“) values for the
calcite-dolomite stability boundary between 220°C and 240°C for a series of
synthetic and natural carbonates to be between 0.4 and 0.9. Converting these
values to Mgz*:Ca2+ ratios yield a calcite-dolomite boundary near 0.4 to 0.125,

which is in good agreement with our results. This is also consistent with the

145

experimental data of Baker and Kastner (1981) and excellent agreement with the
results of Rosenberg and Holland (1964). These results suggest that our
experiment with final solution Mgz":Ca2" ratios of 0.36 is near the calcite-dolomite
stability boundary and therefore solutions are very near equilibrium with respect
to dolomite. Therefore, after reactant depletion, when solution M92":Ca2+ ratios
are at their lowest levels (and solid Mg:Ca ratios are at their highest), conditions
should be ideal to promote the transition from polynuclear to spiral growth.
However, none of the solutions, particularly the experiments with relatively high
Mgz":Ca2+ ratios, had solution Mgz*:Ca2+ ratios that approached the lowest
starting Mgz*:Ca2” ratio of 0.43. This point is important because the experiment
with an initial solution M9250? ratio = 0.43 (final Mgz*:Ca2* ratio = 0.362)
produced islands.

According to the findings presented here, growth mechanisms in dolomite
are inconsistent with crystal growth theory in that a low free energy drive with
respect to dolomite (i.e. low solution Mgz*:Ca2”) does not necessarily promote
growth by a spiral mechanism before reactant depletion. Likewise, experiments
with high free energy drive (i.e. high solution Mgz*:Ca2*) still proceed to spiral
growth after reactant depletion.

Dolomite crystal growth may be analogous to the observations of Dove
and Hochella (1993) for calcite nucleation and growth in the presence of growth
inhibitors. They showed that at high degrees of supersaturation, surface
nucleation continued unimpeded in the presence of phosphate, but that normal

growth was obstructed as evidenced by the formation of amorphous nuclei. This

146

observation was interpreted to reflect surface roughening caused by the sorption
of phosphate, which blocks lateral growth sites on calcium carbonate nuclei. If
hydrated magnesium acts as a growth inhibitor to dolomite formation, polynuclear
growth may dominate over a wide range of conditions because spiral growth is
prevented because lateral attachment sites are poisoned.

Although the dolomitization reaction is typically expressed as a
replacement reaction in which calcite is replaced by dolomite (refer to equation
1), experimental findings suggest there are, in effect, two reactions that occur in
the overall transformation between calcite and ideal dolomite. The first reaction
proceeds as calcite is replaced by nonideal dolomite (Equation 4) and the second

reaction occurs as nonideal dolomite is replaced by ideal dolomite (Equation 5).

2CaC03 +Mg2+ =>nonideal CrzMg(C03)2 + Ca2+ (Equation 4)

nonideal CaMg(C03 )2 :> ideal CaMg(C03 )2 (Equation 5)

Nonideal dolomite always forms by polynuclear growth and ideal dolomite
forms by spiral growth. Therefore, each step in the overall reaction corresponds
to a different growth mechanism. The calcite to nonideal dolomite reaction
occurs by polynuclear growth, whereas the nonideal dolomite to ideal dolomite
reaction proceeds by spiral growth. Because the mechanism transitions from
polynuclear to spiral growth after reactant depletion, it might be predicted based

on crystal growth theory that the free energy drive is lowered after reactant

147

depletion. More specifically, that the free energy drive associated with the calcite
to nonideal dolomite reaction (and polynuclear growth) is higher than the free
energy drive associated with the nonideal dolomite to ideal dolomite reaction
(and spiral growth).

The easiest way to examine how the free energy drive for dolomitization
changes at reactant depletion is to compare the actual free energies of reaction
for the calcite to nonideal dolomite and nonideal dolomite to ideal dolomite
reactions. By doing this it may be possible to better understand the factors
responsible for changing growth mechanisms after reactant depletion. An

important equation to consider when comparing the free energies of reaction is

Equation 6:

AG“, =AGfm +RTan (Equation 6)

In this equation, AG"... is the free energy of reaction, AG°Dm is the standard free
energy of reaction (AGoproducts - AG°maaant3), R is the gas constant, T is
temperature, and Q is the Caz“:Mg2+ activity ratio in solution (Mgz*:Ca2*) divided
by the equilibrium constant (K). In order to solve for the free energy of reaction,
one must know the standard free energies for nonideal and ideal dolomite.
Following an approach by Navrotsky and Loucks (1977), Helgeson et al.
(1978) used Bragg-William theory (Bragg and Williams, 1934) to determine the
thermodynamic consequences of order-disorder in dolomite. According to

Helgeson et al. (1978), the standard free energies of formation for a completely

148

ordered and disordered dolomite are AG°fomnon = -517.98kcal/mol and
AGoformation = -515.873kcallmol, respectively. Navrotsky and Capiobanco (1987)
determined the enthalpy of formation for an ideal dolomite before and after
annealing at 1250°C. Using this value they calculated that the enthalpy of
complete cation disorder was approximately 12kJ/mol. Subsequent work by Chai
and Navrotsky (1996) and Navrotsky et al. (1999) suggest that this value is even
greater at about 20-30kJ/mol. The work of Chai et al. (1995) showed that the
enthalpy of formation of natural dolomites with variable Ca content, increased
from —10kJ/mol for a nearly stoichiometric sample to +10kJ/mol for the most
calcian samples (6 mole% excess calcium). Rock et al. (2001) showed that was
shown that standard free energy of formation values for dolomite compiled from
various studies ranged from as high as -507.13kcal/mol to as low as -
519.25kcal/mol (Wagman et al., 1982; Garrels et al., 1960; Naumov et al., 1974;
Robie et al., 1978). Because these values are highly variable, it is still unclear as
to what impact Ca-Mg composition and cation order have on the relative
energetics of dolomite (Reeder, 2000).

This issue is similar to the problem encountered in assessing magnesium
calcite stability because of issues related to whether compositions reflect true
thermodynamic equilibrium or kinetically controlled processes (Mackenzie et al.,
1983; Arvidson and Mackenzie, 2000). If the free energy drive is responsible for
changing growth mechanisms at reactant depletion, it can be reasoned that
difference between the standard free energies of formation for nonideal

dolomites with islands and ideal dolomites with layers must be considerably

149

smaller than those contributed by Helgeson et al. (1978). In the same respect, it
is important to understand that much of this work indicates that cation order-
disorder has significant effects on the thermodynamic properties of dolomite.
According to Land (1998), the uncertainties about quantitative effects of
stoichiometry and ordering on dolomite solubility remain high. Therefore there is
no accurate way of determining the thermodynamic effects on growth mechanism
transformation at reactant depletion.

4.d.2. Carbonate: Because the thermodynamic framework for discussing
growth mechanisms during the transition between calcite, nonideal dolomite, and
ideal dolomite is poorly constrained, we turn to kinetic considerations for an
alternative approach. Although dolomitization is written as a replacement
reaction, it ultimately proceeds via solution because the reactants must pass
through the state of dissolved ionized species (Lippmann, 1973). As a result, the
dolomitization reaction can be also written as a dissolution-reprecipitation

reaction:

2CaC0, +Mg2+ <=> 2Ca2+ +2C032‘ +Mg2+ ©CaMg(C03)2 (Equation 7)

Although the exact concentrations of calcium, magnesium, and carbonate play
no role in determining the free energy drive for dolomitization of CaCOa (because
thermodynamic stability is a function of the Ca2*:Mgz" in solution assuming the
solution has equilibrated with calcite), Equation 7 suggests that dolomite kinetics

are related to the availability of the reactants, [Caz‘] [M921 [C032]. According to

150

kinetic growth theory, a larger supply of reactants should help to produce
dolomite.

Experimental solutions initially have high concentrations Ca” and Mg”,
but lack carbonate until calcite begins to dissolve. Although these conditions are
not strictly true in nature, they are comparable since seawater’s carbonate
concentration is lower than the concentrations of Ca” and Mg” by approximately
two orders of magnitude. Because carbonate concentration in the reaction
vessels is dictated by reactant solubility, changes in reactant solubility will affect
carbonate activity and, in turn, the ion activity product. At reactant depletion, the
solution aCa”:aMg” is effectively constant so the only significant change to
occur in solution chemistry is when initial reactants are exhausted and the
availability of carbonate is reduced. Therefore, the activity of the carbonate ion in
bomb solutions may be the principle variable affecting reaction kinetics and
growth mechanisms in the dolomitization reaction.

Although the concentrations of Ca”, Mg”, and 0032' cannot be directly
measured in the reaction vessels during an experiment, the concentration of
0032‘ must have decreased after reaction completion when the calcite was
completely consumed and all products were 100% stoichiometric dolomite. At
room temperature calcite is more soluble than dolomite and as a result the
concentration of carbonate ions will be greater in solutions that are able to
equilibrate with calcite. Furthermore, carbonate mineral solubility decreases as
temperature increases (exothermic heat of dissolution). This effect is further

driven by a decrease in CO; solubility with increasing temperature. According to

151

Langmuir (1997), the relative solubility differences between calcite and dolomite
are exacerbated at higher temperatures. Based on thermodynamic data from
Nordstrom et al. (1990), Robie et al. (1978), and Tanger and Helgeson (1988),
Langmuir (1997) used Equation 8 to calculate the solubility of dolomite as a

function of temperature:

—|og K.p (dolomite) = 175.90 + 6835.4/T (K) + 68.727 log T(K) (Equation 8)

According to Langmuir (1997), the solubility of dolomite (based on K,,, = 10'”-09
at 25°C) decreases nearly 14-fold between 0-90°C, whereas the experimentally
determined solubility of calcite reported by Busenberg and Plummer (1982)
decreases only six-fold. Assuming this is applicable to the range of conditions
studied here, the relative solubility differences between calcite and dolomite are
further enhanced by the high temperatures used in the experiments.
Thermodynamic data complied in Carpenter (1980) also show that for a given
solution Ca”/Mg” ratio, increasing the temperature will shift a solution farther
into the dolomite stability field and farther from calcite-dolomite equilibrium. This
further confirms that at higher temperatures, dolomite solubility decreases
relative to the solubility of calcite. The K... = 10'”09 at 25°C dolomite used in
Langmuir (1997) is very similar to the K... for the dolomite reported by Robie et al.
(1978) that was inferred to be representative of a relatively ideal, but not perfectly
ideal dolomite (like the one reported in Helgeson et al., 1978). Because nonideal

dolomite acts as the carbonate source (i.e., the reactant) for the formation of

152

ideal dolomite after reactant depletion, differences in the relative solubilities of
calcite and nonideal dolomite at high temperatures may be sufficient to explain
the change in growth mechanisms after reactant depletion. Assuming that it
approaches equilibrium, carbonate activity must be lowered after calcite is
consumed because nonideal dolomite solubility is lower than calcite solubility.
As stated previously, the Mg”:Ca” ratio of the solution is relatively constant
because the nonideal dolomite is already stoichiometric. As calcite is consumed
and relatively well-ordered stoichiometric dolomite begins to form at the expense
of relatively poorly ordered stoichiometric dolomite, carbonate activity [C0321 in
solution decreases because the solubility of the new reactant (nonideal dolomite)
is lower than the solubility of the initial reactant (calcite). Therefore, less
carbonate is available for dolomite formation once calcite is consumed and the
growth mechanism changes. Again, it is important to note that changes in
solution chemistry at reaction completion do not affect the thermodynamic
stability with respect to dolomite because the energetic drive in the bombs
doesn’t depend on the carbonate activity. Therefore the changing surface
nanotopography from islands to layers at reaction completion may correspond to
a kinetic effect related to [0032'].

The hypothesis that carbonate activity in the hydrothermal bombs is the
major factor in determining reaction kinetics and growth mechanisms in dolomite
is also consistent with the results from our aragonite experiments. The solubility
products of calcite and aragonite at 25°C are K,p = 10‘8'52 and K.p = 10'3“”,

respectively (Helgeson et al., 1978). According to the data compiled by

153

Carpenter (1980) the differences in calcite and aragonite solubility are
exaggerated by more than two-fold at 90°C. Because of the high concentration
of Ca” and Mg” ions in the initial dolomitizing solutions, the compositional
effects of aragonite’s greater solubility on the initial solution Mg”:Ca” ratio will
be relatively minor (approximately 1.2% at 25°C). Conversely, the solubility of
aragonite does lead to a more significant change in the concentration of
carbonate in solution. Because there is no initial carbonate, carbonate
concentration is more dramatically affected by the solubility of the precursor
phase. Based on data reported in Langmuir (1997), the carbonate concentration
is approximately 1.9% greater at 25°C, 2.0% greater at 50°C, and 4.4% greater
at 90°C in solutions with a mixture of aragonite and calcite, than in experiments
with only calcite. The relative differences in solubility of calcite and aragonite are
likely to increase with temperature in a manner consistent with the changes
between 25°C, 50°C, and 90°C. Assuming a linear relationship between
temperature and the relative solubilities of aragonite and calite (a conservative
estimate based on the data above), it is apparent that carbonate concentration in
solution is approximately 10% greater at 218°C in aragonite reactant experiments
vs. calcite reactant experiments.

Preferential replacement of aragonite when both aragonite and calcite are
present can be explained by the fact that the AG"... for the aragonite to dolomite
reaction is greater than the AG”... for the calcite to dolomite reaction. Therefore,
according to thermodynamic arguments aragonite should dolomitize faster than

calcite. However, because thermodynamic factors are unable to fully explain

154

other aspects of dolomitization kinetics, the role of carbonate availablility will be
examined. Although the reactant surface area was not determined, the
aragonite reactants are finer grained than the calcite reactants. Furthermore, the
solubility of aragonite is greater than that of calcite so the dissolution of aragonite
must have occurred much more rapidly than the dissolution of calcite and as a
result the availability of carbonate for dolomite formation is significantly higher in
the aragonite experiments. Selective replacement of aragonite in the aragonite
reactant experiments (calcite-aragonite mixture) further stresses the importance
of carbonate supply in the dolomitization reaction. It was shown earlier that
aragonite’s higher solubility would lead to more Ca” in solution and in turn would
slightly reduce the solution Mg”:Ca”. Based on findings from the calcite
reactant experiments, a lower solution Mg”:Ca” ratio should slow the rate or
dolomitization. However, the rate of dolomitization was nearly three times faster
in the aragonite experiments than any of the calcite experiments. Even the
highest solution Mg”:Ca” ratio in the calcite reactant experiments (Mg”:Ca” =
1.5) took nearly three times longer than the aragonite reactant experiment with a
solution Mg”:Ca” ratio of 1.0. This suggests that the carbonate supply to
solution, which is dictated by the solubility of the precursor, has a dramatic affect
on dolomite growth.

Dolomite and calcite belong to the same crystal system, whereas
aragonite and dolomite are different. Therefore one would expect dolomite to
have a better lattice match with calcite and therefore grow easier on calcite, but

the dolomite always replaces aragonite before calcite in these experiments. The

155

concentration of carbonate ions is presumed to be higher near aragonite
surfaces during the reaction. This effect is enhanced once calcite solubility is
reached in the bombs because further calcite dissolution is prevented and so
local zones around aragonite have high carbonate concentrations relative to the
calcite surfaces. The aragonite reactants were finer than the calcite reactants
and this is likely partly responsible for the increased reaction rate in aragonite
reactant experiments. In fact, Sibley and Bartlett (1987) showed that increasing
the reactant surface area increased dolomitization rates. They also showed,
however, that in experiments comparing the rates of aragonite and calcite
reactants of the same size, aragonite was dolomitized significantly faster than
calcite. Unfortunately, separating the effects of surface area and mineralogy are
difficult given the current data.

In addition to the observation that dolomite selectively replaces aragonite
when both aragonite and calcite are present, another observation in support of
the carbonate supply argument is that dolomite crystals on aragonite are
randomly oriented. This is in contrast to dolomite on calcite, which is always in
crystallographic continuity with the calcite substrate. In terms of surface free
energy, dolomite nucleation should be much easier on calcite than on aragonite
because calcite and dolomite have crystal systems that are much more alike.

The inference that carbonate activity is an important factor in dolomite
growth is also consistent with the observation that dolomite forms preferentially
on edges and corners of aragonite and calcite in high-temperature synthesis

experiments. According to a surface energy standpoint, edges and corners are

156

poor nucleation sites. However, calcite should dissolve faster in areas, like
edges and corners because of the high surface area to volume ratio. Therefore
one may expect to find the highest concentration of carbonate near edges and
corners during calcite dissolution. Edges and corners may also be the best
conversion sites if dolomite nuclei are formed by exchanging Mg” ions for Ca”
ions in the existing calcite crystal. Although this hypothesis was not tested, it is
consistent with the observation that dolomite forms in crystallographic continuity
with the calcite substrate.

Lippmann (1967) showed that adding carbonate to MgClz solutions
allowed norsethite, an ordered, double carbonate mineral [BaMg(COa)2] , to
replace witherite [BaC03] at 20°C and 1 atrn. Previously, Chang (1964) showed
norsethite could be easily formed by replacing witherite in MgCIz solutions at
500°C and 15kbar pressure, but witherite could be kept in magnesium chloride
solutions with various concentrations for months and years at low-temperature
without any evidence of replacement by norsethite (Lippmann, 1973). Forming
norsethite by replacement of witherite was argued by Lippmann (1973) to be
analogous to dolomitization of calcite. However, even with the addition of
carbonate to dolomitization solutions, Lippmann (1973) was unable to form
detectable amounts of dolomite in a year. He argued that the problem with
dolomitization lies in the similar coordination numbers of Ca and Mg. Because of
their larger size, calcium cations are dehydrated easier than magnesium cations
and are therefore more readily incorporated into magnesium sites (the enthalpy

of hydration for Mg is approximately 20% greater than that of Ca). This argument

157

exemplifies Graf and Goldsmith’s (1956), “simplexity principle”, which attempted
to explain the occurrence of metastable magnesium-calcites and disordered
dolomites in the place of ideal dolomite. Malone et al. (1996) formed Ca-Mg
carbonates (41.7 mole% MgCOa) with no observable ordering reflections at
relatively low-temperature (60°C) in CaClz - MgSO. solutions with the addition of
carbonate (N3C03) on the order of a few days. This suggests that carbonate
activity is an important factor in getting magnesium into the carbonate lattice.

A further test of this hypothesis would be to examine dolomites formed on
aragonite after aragonite reactant depletion. If the dolomite crystal growth
mechanism does not change following aragonite depletion in the aragonite
reactant experiments, like it does at reactant depletion in calcite reactant
experiments the hypothesis would be further supported. After calcite is
completely consumed in the calcite reactant experiments, the solubility of poorly
ordered, nonideal dolomite dictates the supply of carbonate to solution because
all calcite is depleted. However, calcite remains in the reaction vessels as the
principle source of carbonate after reactant depletion in the aragonite reactant
experiments. As a result, carbonate concentration should remain relatively high
compared to reactant depletion in the calcite reactant experiments. Therefore,
complete dolomitization of aragonite in the aragonite reactant experiments may

not be followed by a change in growth mechanism.

158

4. e. Synthetic Dolomite Stoichiometry

The experimental results shown in Figure 13 indicate that the
stoichiometry of nonideal dolomite products is highly dependent on the initial
Mg”:Ca” ratio in solution (R2=O.97). Previous high-temperature dolomitization
studies are consistent with this in that they have generally shown solutions with
lower Mg”:Ca” produce Ca-rich dolomite and solutions with higher Mg”:Ca”
yield more stoichiometric dolomite (Gaines, 1974; Sibley, 1990; Sibley et al.,
1994). Although the relationship between initial solution composition and
dolomite stoichiometry is strong (R2=0.97), it is not exactly direct. Despite
solutions becoming increasingly Ca-rich as dolomite replaces calcite (see table
3), dolomite composition appears to be independent of percent dolomite by
remaining relatively constant until reactant depletion (Figures 14-16).
Furthermore, after reactant depletion, Mg”:Ca” ratios in solution reach their
minimum values, yet after reactant depletion synthetic dolomite invariably
becomes stoichiometric. Sibley (1990) also noted that in high-temperature
dolomitization experiments, the composition of the products remained relatively
constant, while the solution Mg”:Ca” decreased from 0.66 to 0.53 after half of
the calcite was replaced by a 65 mole% CaCOa product. Moreover, he noted
that the observed transition to stoichiometric dolomite at reactant depletion
occurred when the solution chemistry was at its lowest value and nearly
constant. These observations suggest that the stoichiometry of nonideal
dolomite only reflects the solution Mg”:Ca” ratio at the beginning of growth and

not the solution chemistry as it changes.

159

Aragonite reactant experiments with initial solution Mg”:Ca” = 1.0 yield
more Ca-rich dolomite than dolomite formed from calcite reactant experiments
with the same initial solution Mg”:Ca”. Aragonite is more soluble than calcite
and as a result, the concentration of Ca” will be higher in solutions that
equilibrate with aragonite than those that equilibrate with calcite. However,
because of the high initial concentrations of Ca” and Mg” in experimental
solutions, the effects of aragonite’s greater solubility have only a relatively minor
effect on the solution Mg”:Ca” ratio (approximately 1.2% at 25°C). Because the
effects of aragonite and calcite solubility on solution Mg”:Ca” are difficult to
determine at 218°C, they may account for the observed stoichiometric
differences between dolomite formed in the calcite reactant experiment and
those formed in the aragonite reactant experiment.

Although dolomite bulk composition remains relatively constant with
reaction progress, the composition of individual islands is likely to be variable. If
the composition and degree of cation order of individual islands were identical,
one might expect islands to coalesce into layers upon continued growth.
According to our observations, islands do not coalesce into layers, but instead
remain distinct surface features as polynuclear growth proceeds. Furthermore,
our chemical etching experiments suggest that island interfaces remain distinct
components of the crystalline microstructure even long after formation. TEM
observations also indicate that the microstructure of natural nonideal Ca-rich

dolomites is heterogeneous due to distinct compositional and structural domains

160

(Reeder and Wenk, 1979; Reeder, 1981, 1992, 2000; Schubel et al., 2000;

Frisia, 1994). This will be discussed in more depth in subsequent sections.

4.f. Synthetic Dolomite Cation Order

In the calcite reactant experiments, the degree of cation order (ordering
ratios) in synthetic dolomites formed prior to reactant depletion with islands have
values between 0.0 and 0.45, whereas synthetic dolomites formed after reactant
depletion with layers have ordering ratios of 0.5 to 0.88. This suggests that a
higher degree of cation order is achieved after the reactant is consumed during
spiral growth. However it could be argued that cation order is related to dolomite
stoichiometry because all well orderd dolomites are stoichiometric. Figure 22
indicates that prior to reactant depletion, the degree of cation order is seemingly
independent of dolomite stoichiometry (R2 = 0.034). Cation order generally
increases with reaction progress, but dolomite stoichiometry remains relatively
constant. Although a certain abundance of magnesium is necessary for cation
order to occur, the results here indicate that cation order is independent of
dolomite stoichiometry. This is further supported by the observation that many of
the stoichiometric dolomites produced in high Mg”:Ca” solutions have ordering
ratios that are lower than Ca-rich dolomites from lower Mg”:Ca” solutions.

Changes in the degree of cation order with time (Figures 19 and 20) also
indicate changing growth mechanisms after reactant depletion. The results show
that degree of cation order increases relatively rapidly during the growth of

nonideal dolomite before all calcite is consumed. The rate at which cation order

161

increases slows considerably after the calcite is consumed and products are
100% stoichiometric dolomite (indicated by a shallower slope). This is precisely
the time when the change from island to layer nanotopography on dolomite
surfaces occurs. According to crystal growth theory, spiral growth is a slower
growth mechanism than polynuclear growth (Sunagawa, 1984; refer to Figure 3).
Therefore after reactant depletion, a slowing rate of increase for cation order with
time is consistent with dolomite changing from polynuclear growth to spiral
growth. Note that the experiment with solution Mg”:Ca” = 0.66 does not follow
the trend of decreasing slope (R2= 0.93; Figure 21). Because the growth rate is
already relatively slow compared to experiments with higher solution Mg”:Ca”
ratios, the change in rate of ordering after reactant depletion in the Mg”:Ca” =
0.66 experiment may not be as dramatic.

Because ideal dolomite is thermodynamically stable relative to nonideal
dolomite, some factor must cause slow nucleation or crystal growth of ideal
dolomite in order for nonideal dolomite to form first. Because it is possible to
produce stoichiometric, but poorly ordered dolomite in our high Mg”:Ca”
experiments, it seems reasonable that cation ordering may be a factor to
consider in explaining why nonideal dolomite forms before ideal dolomite.
Although Ca-enrichment has been shown to destabilize dolomite (Navrotsky and
Capobianco, 1987; Chai et al., 1995), calcium is incorporated more easily into
the lattice than magnesium. In the laboratory, Glover and Sippel (in Bathurst,
1975, page 542) were able to produce magnesium calcites at room temperature

and atmospheric pressure. Although compositions had cation proportions equal

162

to that of some dolomites, the magnesium calcites were completely disordered.
Morrow et al. (1994) argued that the precipitation of partially ordered, metastable
dolomite of stoichiometric composition is favored over that of ideal dolomite
because of nonspecified kinetic inhibitions. This is in contrast to other studies
that envoke thermodynamic arguments.

Kazmierczak et al. (1982) reported the activation energy for precipitating
calcite to be 9.4 :l:0.9kcal mol“. Assuming that calcite and magnesite grow by
identical mechanisms, Arvidson and Mackenzie (2000) reported the activation
energy of magnesite to be 22.2 kcal mol‘1 based on data from Lippmann (1973).
Arvidson and Mackenzie (2000) plotted activation energy as a function of
carbonate composition by drawing a line between the calcite and magnesite
values. Arvidson and Mackenzie (1999) calculated the activation energy for Ca-
rich protodolomite (poorly ordered) to be 31.9 kcal mol“. This value agreed
relatively well with the predicted value for a carbonate of the same composition.
Gaines (1980) synthesized relatively well-ordered dolomites and calculated the
enthalpy of formation to be ~5.5 kcal mol'1 above the value predicted by the
extrapolation by Arvidson and MacKenzie (2000). Because the composition was
stoichiometric, the additional energy is assumed to represent the effect of cation
ordering in the dolomite structure. The additional activation energy calculated by
Gaines (1980) is also in good agreement with the value of 5.5 kcalmol'1 that
Arvidson and MacKenzie (2000) calculated from the data of Malone et al. (1996)
for the rate of recrystallization of disordered Ca-Mg carbonate to ordered—phases

between 149°C and 197°C.

163

For comparable reaction progress, the degree of cation order in dolomite
produced during aragonite reactant experiments is considerably lower than in
dolomite produced during calcite reactant experiments. This is likely related to
the fact that the rate of dolomitization was significantly higher in the aragonite
reactant experiments. Ordering reflects how crystalline material (e.g., ions or
molecules)is deposited into energetically favorable attachment sites. The
processes of nucleation and crystal growth typically occur when material
attaches to a surface followed by surface diffusion to energetically favorable
(ideal) attachment sites (Gratz et al., 1991). When nucleation and growth rates
are extremely high, attachment of new material may out compete surface
diffusion and poor ordering may result. Aragonite reactant experiments
demonstrate a linear relationship between cation order and time (R2 = 0.83).
More specifically, no change in the rate of increasing cation order with reaction
progress is observed after the aragonite is completely consumed. However, if
dolomite formed in the aragonite experiments never transitions from polynuclear
to spiral growth after aragonite is consumed, a change in the rate of increasing

order would not be expected at reactant depletion.

4. 9. Synthetic Dolomite Recrystallization

The only source of carbonate for ideal dolomite after reactant depletion is
poorly ordered stoichiometric (nonideal) dolomite. Immediately after reactant
depletion dolomite still has a relatively low degree of order, and is therefore

metastable and an energetic drive towards equilibrium still exists. Because

164

layers appear only after calcite reactant depletion when dolomite products are
already stoichiometric, they indicate a transition between poorly ordered
(nonideal) dolomite and well-ordered (ideal) dolomite (refer to Figure 61).
According to Bathurst (1975), recrystallization can be defined in terms of a
diagenetic fabric change without accompanying change in carbonate mineralogy.
Growth layers on synthetic dolomite are interpreted, in this context, to be
indicative of recrystallization. After reactant depletion, cation order and surface
nanotopography change without a concomitant change in dolomite composition.
Therefore recrystallization is understood to include changes in ordering and
surface nanotopography.

Machel (1997) suggested the use of the term “significant recrystallization”
to describe a modification in the original texture, ordering, chemical composition,
or magnetic properties that is larger than the original range during dolomitization.
The use of this term is warranted because all synthetic dolomites displaying
evidence of recrystallization (i.e., layers) have higher degree of cation order than
any synthetic dolomites with islands. As the degree of cation order increases in
dolomite, excess strain energy in the crystal must decrease as the internal
microstructure approaches an ideal form. As layers become more pervasive with
increasing reaction times after 100% replacement, the degree of cation order
continues to increase as well. Therefore significant recrystallization has occurred

after the appearance of layers on dolomite surfaces.

165

Reaction Progress

Polynuclear Growth: Spiral Growth:
A Nonideal Dolomite 0 Ideal Dolomite

Islands

tr

   

  

Nonideal Dolomite
Dissolves

   
 

Substrate Substrate

Dissolves

/

Reactant Depletion

 

 

 

Nonideal Dolomite

 

 

 

 

 

Ideal Dolomite

L
7

Increasing Cation Order

Figure 61. Growth mechanism summary diagram for dolomitization.

166

The observations that synthetic dolomite nantopography does not change
as a function of reaction progress, except after reactant depletion and that
nanotopography does not change with solution M92503” ratio are also
consistent with recrystallization. These observations indicate that polynuclear
growth dominates over a wide range of solution conditions prior to reactant
depletion. Furthermore, despite highly variable starting solutions and nonideal
dolomite stoichiometries, all dolomite surfaces are covered with layers after
reactant depletion. In as much as layers occur over a wide range of conditions,
they provide a measure of recrystallization that is independent of solution

chemistry and stoichiometry of the nonideal precursor dolomite.

4h. Natural Dolomite Growth Mechanisms

Most natural stoichiometric dolomites included in this study are relatively
well-ordered (0.96-1.33), whereas most nonstoichiometric dolomites are
relatively poorly ordered (0458-0824). In most cases, surface nanotopography
can generally be differentiated according to stoichiometry alone. For instance, all
nonstoichiometric dolomites have islands, but so do a few stoichiometric
dolomites. These include the dolomites from Spain (samples 8-9-84-3 and 8-9-
84-5) and Abu Dhabi (GWP287 120-130), which are stoichiometric and have
island-covered surfaces. The above dolomites have relatively low ordering
ratios (0.44-0.67), however, and are therefore considered nonideal.
Furthermore, a nonstoichiometric Mississippian dolomite from Kentucky has an

ordering ratio of 1.167, which is higher than some of the stoichiometric samples.

167

This sample has islands. However, the nonstoichiometric composition (i.e.,
excess calcium) indicates that the sample is nonideal. According to the results
presented here, dolomite surface nanotopography is related to ideality, a
combination of stoichiometry and cation order.

4.h.1. Nonideal Dolomite and Polynuclear Growth: Observations of
dissolution islands are relatively rare in comparison to those of etch pits. There
is, however, another dissolution feature that is somewhat analogous to the
dissolution islands reported here. During dissolution experiments, Lin and Shen
(1995) observed the formation of etch hillocks (smooth, topographically high
surface features) prior to the development of etch pits on willemite crystal
surfaces. They reasoned that the hillocks resulted from dissolution-resistant
impurities contained within the willemite lattice. Other etch hillocks have been
interpreted to form due to local surface contaminations (Batterman, 1957), etch
resistant impurities (Stadler, 1963; Lin and Shen, 1995), reaction products,
precipitates (Tuck, 1975; Rozgonyi et al., 1976), and gas bubbles attached to the
surface (Weyher and Van Enckevort, 1983). Regardless of the exact mechanism
involved with their occurrence, etch hillocks are generally understood to form on
the crystal surface where dissolution is suppressed (Tuck, 1975). Therefore
relative dissolution rates at various surface locations dictate the development of
hillocks (Tuck and Baker, 1973). Islands have also formed on calcite as a result
of dissolution in the laboratory (Kirkland et al., 1999) and islands on natural
kaolinites have been interpreted to form by dissolution (Zbik and Smart, 1998).

Dove and Hochella (1993) reported that calcite surfaces were characterized by

168

the presence of flat, rounded islands with flat tops in some of their experiments.
They interpreted these features as being remnant growth islands that were
exhumed during undersaturated conditions.

The formation of dislocations at the interface between nanoparticles as
shown by Penn and Banfield (1998), suggests that it is not unreasonable to
expect a defect-laden boundary between adjacent growth islands. Preferential
dissolution at the interface between islands indicates that there is some driving
force such as chemical or structural discontinuity at the interface between
adjacent islands. The relatively high-energy interface between individual growth
islands can therefore be utilized during etching to uncover an island
nanotopography in ancient dolomites.

Using FESEM, Jones (2004) observed that chemically etched Ca-rich,
Miocene-age dolomite cements exhibited “blocky topography”. Individual
“blocks” were described as square or elongate topographic highs with rounded
edges and corners. Blocks generally measure between 250 and 500
nanometers wide and less than 1 micron long. The edges between individual
blocks were relatively straight, but curved edges were not uncommon. The
features were observed to occur pervasively across the surfaces of Ca-rich
dolomite cements. Jones (2004) also noted that the blocks were oriented
perpendicular to the dominant dolomite growth face. Furthermore, the corners of
dolomite crystals that represented multiple growth faces were characterized by

multiple block orientations. The size, shape, and orientation of blocks on Ca-rich

169

dolomite observed by Jones (2004) are consistent with the islands observed with
AFM. No euhedral etch pits were reported.

In addition to looking at hillocks as areas of suppressed dissolution, as
Tuck (1975) suggested, it is helpful to consider the interface between hillocks
(islands) as areas of enhanced dissolution. The presence of distinct growth
islands on synthetic dolomite surfaces in and of itself suggests a physical
mismatch between adjacent islands that serves as the site of elevated free
energy. The ability to etch and uncover these growth features is further evidence
of lattice mismatch between adjacent growth islands. The finding that islands are
easily exposed on natural nonideal dolomite with chemical etching, even after
hundreds of millions of years, suggests that islands maintain a pervasive defect
network that was formed during polynuclear growth. According to Reeder
(1992), solid-state diffusion is recognized as a relatively slow process; therefore
defect microstructures formed during growth are likely to remain unmodified long
after growth. If two adjacent islands have even slightly different orientation,
structure, stoichiometry, or cation order, and come into contact with one another
during crystal growth, their crystal lattices will be different, thus producing surface
tension at the island-island interface. Surface tension between islands would
result in elevated reactivity at the interface and would therefore serve as the site
of preferential dissolution during chemical etching. Because the interface
between islands follows the shape of an island, dissolution is able to uncover the

polynuclear growth microstructure.

170

According to models of dissolution, the reactivity of island interfaces must
be higher than the reactivity of the islands themselves, othenrvise islands would
dissolve and the interfaces would be resistant. However, it possible to argue that
euhedral etch pits are not observed because the reactivity of the island interfaces
is higher than the reactivity of other naturally occurring defects. Although this
may represent a limitation to uncovering other defects, the argument is not
important to the interpretation that islands form by a polynuclear growth
mechanism. The presence of dislocations in dolomites with island-covered
surfaces does not preclude the interpretation that growth occurred by polynuclear
growth. Because growth islands occur pervasively on dolomite surfaces, the free
energy drive must have been high enough to support nucleation dominate
growth.

Because empirical observations of islands as dissolution features are
limited compared to etch pits, observations from growth and dissolution
experiments for synthetic dolomites become an extremely useful analog for
interpreting natural etching features. Dissolution islands on natural nonideal
dolomites (Figures 5052) are very similar to synthetic dolomite growth (Figures
32-34) and dissolution (Figure 46) islands. As a result, etch features on natural
dolomites are interpreted in a manner consistent with the AFM observations of
synthetic dolomite crystal surfaces. Because islands appear to be ubiquitous in
nonideal dolomites they are not interpreted as post reaction surface precipitates,

reaction products formed during dissolution, or etch resistant impurities. Islands

171

on natural and synthetic nonideal dolomites are interpreted as growth features
that have been exposed during chemical etching.

4.h.2. Ideal Dolomite and Spiral Growth: The euhedral etch pits on
natural dolomite (Figures 53-56) have similar size, shape, and orientation as the
euhedral etch pits on synthetic dolomites (Figures 4749). Because of this,
natural dolomites are interpreted according to the observations from high-
temperature synthetic dolomites as well as theoretical predictions for interface-
controlled dissolution. Spiral growth in synthetic dolomite is inferred from the
presence of steps and layers on ideal dolomite surfaces as well as the
occurrence of euhedral etch pits following dissolution. Therefore the presence of
euhedral etch pits on natural ideal dolomites are interpreted as evidence for
spiral growth.

Deep euhedral etch pits form equally at spiral and edge dislocations
because both promote preferential dissolution and it is not possible to distinguish
the two by means of chemical etching. Although significant deformation of
synthetic dolomites in the bombs likely did not occur, some of the natural
dolomites studied here may have been subjected to conditions favorable for
deformation. Consider, however, that euhedral etch pits are observed only in
ideal dolomites and were never observed in nonideal dolomite. Deformation
probably did occur in some of the natural dolomites studied here, but it is rather
unlikely that deformation preferentially affects ideal dolomites.

Herman and White (1985) analyzed surface topography on polished

natural dolomite crystals after dissolution in distilled water. After 200 hours of

172

dissolution, the etched single crystal had ledges and steps with consistently
oriented pits that ranged from microns to tens of microns in diameter. In the rock
specimen, dissolution occurred preferentially along grain boundaries after 200
hours, but Herman and White (1985) also observed that surfaces were covered
by etch pits with steps that descended into the crystal. Luttge et al. (2003) used
vertical scanning interferometry to analyze natural dolomite surfaces following
experimental dissolution runs. After 8 hours of dissolution in mild hydrochloric
acid (0.001mol/L; pH = 3), dolomite surfaces exhibited variously-shaped etch pits
that measured up to 20pm in width and ~1.7pm in depth. Groups of pits have
consistent orientation. Luttge et al. (2003) argued that dissolution occurred at
inter-pit areas on dolomite surfaces. According to Luttge et al. (2003) vertical
scanning interferometry has a vertical resolution of 2nm and a lateral resolution
of 0.5 to 12pm (depending on the objective). The etch features on natural and
synthetic dolomite observed with AFM (both etch pits and islands) are typically
tens to hundreds of nanometers in diameter and would therefore not resolvable
using this instrument (VSl). As a result, the inter-pit dissolution observed by
Luttge et al. (2003) may actually occurred at pits or islands that were not

resolvable.

4. i. Surface Topography And Internal Microstructure
TEM observations of ancient dolomite microstructures indicate that ideal
dolomites are characterized by a homogeneous microstructure whereas nonideal

(Ca-rich) dolomites exhibit heterogeneous and modulated microstructure (Barber,

173

1977; Reeder and Wenk, 1979; Reeder, 1981, 1983, 1992, 2000; Wenk et al.,
1983; Van Tendello et al., 1985; Rosen et al., 1989; Frisia, 1994; Schubel et al.,
2000). A number of these TEM studies (e.g., Reeder, 1981, 1992, 2000) have
discovered that nonideal dolomites are composed of chemical and/or structural
domains, which cause the appearance of modulations. The modulations have
wavelengths that measure nanometers to tens of nanometers (up to 400
Angstroms). Modulations have been attributed to local compositional
heterogeneities and/or fluctuations in cation ordering (Reeder and Wenk, 1979;
Reeder, 1981,1992, 2000; Reeder and Prosky, 1986; Miser et al., 1987; Fouke
and Reeder, 1992). High-resolution TEM (HRTEM) images have confirmed the
presence of individual domains in nonstoichiometric dolomite (Schubel et al.,
2000). Furthermore, that these domains have cation distributions different from
ideal dolomite and from one another. This suggests that domains are
compositionally distinct and are characterized by different degrees of cation order
(Reeder, 2000). Reeder (2000) argued that modulations found in nonideal
dolomites were largely a result of strain associated with the structure of individual
domains. He suggested that variations in the size, shape and, perhaps local
distortions of adjacent zones could explain the occurrence of distinct domains in
Ca-rich, nonideal dolomite.

Schubel et al. (2000) reported the occurrence of a heterogeneous,
modulated microstructure in ancient Ca-rich dolomites despite a wide range of
inferred dolomitizing conditions and reaction mechanisms. Dolomitizing

environments reportedly ranged from low-temperature (<50°C) vadose

174

conditions to relatively high-temperature (<177°C) hydrothermal conditions.
These observations suggest that similar dolomite microstructures occur over a
wide range of temperature and formation conditions. Fouke and Reeder (1992)
interpreted such differences in microstructures to indicate that ideal and nonideal
dolomites formed by different growth processes. The ability to detect growth
islands in ancient nonideal dolomites with the use of chemical etching suggests
that islands remain as separate entities, which is consistent with the idea that
dolomite has a pervasive crystalline defect microstructure. If two growth islands
with different composition, ordering, or orientation (as suggested by the TEM
results) intersect one another, lattice mismatch is likely to occur and therefore
should result in a defective interface. In order for individual growth islands to
remain distinct, there must be a surface tension associated with the interface
between growth islands. Without an interfacial free energy growth islands would
just grow together and form a continuous growth layer.

Although the size of islands and modulations are not identical, there is
reason to believe that surface nanotopography and crystalline microstructure are
directly related. First, the range of island sizes overlaps the range of modulation
wavelengths. Secondly, modulations have been interpreted as growth features
because they occur in an orientation that is sub-parallel to the growth direction.
Even in areas with multiple growth sectors, modulations are observed to be
consistent with the growth direction (Reeder, 2000). This is consistent with the
AFM observations because regardless of what cleavage face is analyzed, islands

are always observed in nonideal samples.

175

Ideal dolomite microstructures are homogeneous when viewed with TEM
(Barber, 1977; Reeder and Wenk, 1979; Reeder, 1981, 1983, 1992, 2000; Wenk
et al., 1983; Van Tendello et al., 1985; Rosen et al., 1989; Frisia, 1994; Schubel
et al., 2000). This suggests that growth occurred in a way that promotes a
homogeneous microstructure. Rosen et al. (1989) reported a higher density of
dislocations in the ideal dolomites with homogeneous microstructure compared
to the nonideal dolomites with a heterogeneous microstructure that he studied.
His results suggest that extended line defects may be more common in ideal
dolomites whereas a more heterogeneous microstructure is consistent with a

pervasive defect structure.

4.]. Similarities Between Synthetic and Natural Dolomites

Because discrepancies exist between the conditions in which dolomite is
synthesized in the laboratory and those generally recognized as being
associated with natural dolomitization, one might scrutinize comparisons
between them. Despite the lack of direct evidence indicating high-temperature
experimental results are applicable to low-temperature natural dolomites, a
number of observations, in addition to the similarities between surface
nanotopography observed here, imply that empirical findings from high-
temperature experiments are consistent with natural, low-temperature dolomites.

4.j.1. Induction Period: Although there is no theoretical basis for
predicting an induction period, high-temperature synthetic dolomitization

experiments are characterized by a long induction period and relatively short

176

growth period. The data presented here show that the length of the induction
period is highly dependent on the solution chemistry and mineralogy of the
reactant. Other studies have shown that temperature, the reactant surface area,
and the amount of sulfate in solution can also have dramatic effects on the length
of the induction period (Land, 1967; Gaines, 1974; Baker and Kastner, 1981;
Sibley and Bartlett, 1987). In nature, seawater is supersaturated with respect to
dolomite, yet many sediments consisting of aragonite and calcite exist in marine
waters for millions of years without being dolomitized. This demonstrates that
natural environments can have a long induction period for dolomitization.
Furthermore, the observation that many ancient marine limestone deposits
contain no dolomite despite evidence suggesting these rocks were submerged in
seawater for extensive periods of time is also consistent with the idea of an
induction period for natural dolomite. One the other hand, the existence of
Holocene dolomite deposits suggests that dolomite can and does form relatively
rapidly (~103 years) in natural environments (Baltzer et al., 1994; Purser et al.,
1994; Warthman et al., 2000). Therefore the existence of modern dolomites
suggests an induction period for natural dolomitization that is on the order of
~1000 years or longer. Arvidson and Mackenzie (1997) argued against an
induction period in natural dolomites, suggesting that the induction period itself is
an artifact of high-temperature synthesis experiments. Because the majority of
carbonate minerals that reside in marine waters (which are supersaturated with
respect to dolomite) have no detectable amounts of dolomite, we interpret this to

be consistent with the experimental results that indicate an induction period for

177

dolomitization. Moreover, many natural carbonate rocks show evidence of finer
grained components being preferentially dolomitized. This will be discussed
further in a subsequent section, but suffice it to say at this point that this is
consistent with experimental findings, which show reactant surface area plays a
major role in the induction period for dolomitization.

4.j.2. Reaction Rate: The observation that dolomitization in high-
temperature synthesis experiments proceeds relatively rapidly once the induction
period is over is consistent with the findings of Lumsden and Chimahusky (1980)
and Sperber et al. (1984). These studies reported that most dolomitized-
limestones in the rock record have dolomite/limestone ratios that are between
90% and 100% dolomite, which indicates that once dolomitization begins, most
limestone are completely replaced. Given and Wilkenson (1987) plotted the
distribution of dolomite content in Phanerozoic carbonates based on data from
Baum et al. (1985), Chilinger (1956), Langbein et al. (1984), Lumsden and
Chimahusky (1980), Marschner (1968), and Sperber et al., (1984) to show that a
majority of carbonate samples have either a very low percentage of dolomite
(~20%) or a very high percentage of dolomite (~97%). This is consistent with the
observation that most synthetic dolomitization experiments start out relatively
slowly compared to the rapid replacement phase.

The contacts between limestone and dolomite are generally abrupt (Sibley
et al., 1994). Furthermore, Choquette and Steinen (1980) reported that these
transitions are characterized by a decrease in the number of dolomite

rhombohedra, but no change in crystal size across the contact. These

178

observations are consistent with an induction period as well as a rapid
replacement period observed in high-temperature synthetic experiments. Sibley
et al. (1994) argued that if limestone-dolomite contacts correspond to transition
zones where limestone resides in potentially dolomitizing solutions of different
duration, then completely dolomitized rock may indicate only a slightly longer
duration in the fluid than the undolomitized limestone. Based on these
observations, it seems evident that the same process, or processes inherent to
the dolomitization reaction result in either minor or complete dolomitization of
carbonate sediments in both high-temperature and low-temperature reactions.
4.j.3. Selective Replacement: Sibley and Bartlett (1987) reported that the
reactant surface area greatly influenced the length of the induction period in
hydrothermal bomb experiments. They showed that the rate of dolomitization
decreased for finely ground calcite, medium calcite, and coarse calcite,
respectively. They also showed that the induction period was longer for finely
ground aragonite, medium aragonite, and coarse aragonite, respectively. These
observations are consistent with nucleation theory, which predicts that finer
reactants have more active nucleation sites per unit mass and should therefore
have a higher nucleation rate. Sibley and Bartlett (1987) and this study show
that high-temperature synthetic dolomites selectively replace aragonite when the
solid reactant is a mixture of calcite and aragonite. Sibley and Gregg (1986) also
reported that in high-temperature synthesis experiments with mixtures of low
magnesium calcite and aragonite, the aragonite was completely dolomitized

before any calcite showed signs of replacement. Katz and Matthews ( 1977) and

179

Gaines (1980) showed that aragonite is dolomitized faster than LMC in high-
temperature synthesis experiments. The data presented here also show that the
induction period of aragonite is nearly four times shorter than calcite experiments
with the same solution Mgz“:Ca2+ ratio. This rapid rate of dolomitization is
attributed to both the greater surface area and the mineralogy of the aragonite
reactant. Although each individually could possibly account for the increased
dolomitization rate, both have been experimentally shown to influence reaction
rates. The observations from Sibley and Bartlett (1987) are interpreted to be
analogous to selective dolomitization of fine-grained components in natural
limestone reported by Murray and Lucia (1967). Furthermore, McKenzie (1981)
and Patterson and Kinsman (1982) reported that dolomite selectively replaces
aragonite in modern sabkhas.

4.j.4. Crystal Size Distributions: Crystal size distributions (CSD) contain
information about nucleation and crystal growth (Marsh, 1988). Therefore it is
reasonable to use CSD for comparisons between synthetic and natural
dolomites. Although the inability to unambiguously interpret the separate effects
of nucleation and growth presents a major limitation in interpreting CSD, certain
constraints can be established, thus making CSD more useful in understanding
similarities between natural and synthetic dolomites. Sibley et al. (1993)
determined CSD for natural dolomites Recent to Cambrian in age by counting
crystals in thin-section and scanning electron photomicrographs. They analyzed
dolomites from Belize (Recent), the Saluda Formation (Ordovician), the Trenton

Limestone (Ordovician), and the Bonneterre Dolomite (Cambrian) and reported

180

that many of the natural dolomite samples were unimodally distributed and
characterized by a coarse-skewed lognorrnal distribution. Gregg et al. (1992)
looked at Holocene dolomites from Belize and reported that the CSD were nearly
lognorrnal. Kessels (2001) showed zone widths for 1400 dolomite crystals from
the Saluda Formation were coarse-skewed.

The synthetic dolomites studied here display coarse-skewed CSD.
Furthermore, Malone et al. (1996) experimentally investigated recrystallization of
Ca-rich dolomite at 50°C to 200°C for up to one year in solutions with ionic
strength near that of seawater. CSD frequency plots produced from their
cumulative percent crystal size data reveal coarse-skewed CSD for the synthetic
dolomite products. Therefore the CSD for synthetic dolomites representing a
wide range of formation conditions are consistent with those reported for a
number of naturally occurring sedimentary dolomites.

4.j.5. Mimetic Replacement: When dolomite replaces the precursor
carbonate in a way that preserves the shape and texture of cements or
allochems, replacement is reported as mimetic. Nonmimetic replacement
destroys the original fabric of the precursor phase. Mimetic replacement occurs
when the abundance of dolomite nuclei is sufficient and nuclei have the proper
orientation with respect to the precursor phase (Bullen and Sibley, 1984).
Because mimetic/nonmimetic replacement fabrics reflect information about
crystal growth and nucleation, it is relevant to compare natural and synthetic
dolomite in this context. Bullen and Sibley (1984) dolomitized a variety of fossils

in hydrothermal bombs (at 250°C). They observed that the fossils that displayed

18]

mimetic replacement in the experiments commonly exhibit mimetic replacement
in nature (e.g., Cullis, 1904; Murray, 1964; Sibley, 1982). Furthermore, the
fossils that showed nonmimetic replacement in the high-temperature experiments
were consistent with the types of fossils that do not undergo mimetic replacement
in nature. Bullen and Sibley (1984) also pointed out that coralline algae
underwent rapid dolomitization in the experiments and are commonly selectively
dolomitized in nature, whereas microcrystalline, LMC fossils preferentially resist
dolomitization in high-temperature experiments and in nature. Sibley et al.
(1993) reported that dolomite commonly nucleates epitaxially on calcite, thus
preserving the crystallographic orientation of cements and allochems. Aragonite,
on the other hand, is replaced by dolomite in a way that preserves structures, but
not crystallographic orientation in cements and allochems. Bullen and Sibley
(1984) also reported that LMC and HMC fossils were more apt to be mimetically
replaced than aragonite fossils. This is interpreted to be consistent with the
observation that dolomite rhombohedra are randomly oriented on aragonite, but
invariably oriented in crystallographic continuity with the calcite substrate.

4.j.6. Stoichiometry: Compositional zoning in dolomites is often used as
evidence for establishing a relationship between dolomite composition and the
chemistry of the dolomitizing fluids (Katz, 1971; Richter, 1974; Cander et al.,
1988). Thermodynamic considerations suggest that the Mg:Ca ratio in dolomite
should be positively and proportionally related to the activity ratio in the liquid
(Sass and Katz, 1982). Sass and Katz (1982) argued that the two activity ratios

cannot be proportionally related based on the observation that Mgz"‘:Ca2+ ratios

182

vary widely in natural solutions, but dolomite compositions are relatively well
constrained (0.74 to 1.0). The high-temperature synthetic experiments presented
here indicate a strong correlation between the solution Mgz*:Ca2+ ratio and the
stoichiometry of nonideal dolomite products formed prior to reactant depletion (R2
= 0.97). At relatively low Mgz*:Ca2+ ratios, the rate of dolomite formation is
relatively slow, and the first dolomite products are initially Ca-rich. At relatively
high solution Mgz“:Ca2+ ratios, the rate of dolomitization is relatively rapid and
initial dolomite products are more stoichiometric. These observations are
consistent with the findings from a number of studies that have examined natural
dolomites (Schmidt, 1965; Katz, 1971; Richter, 1974; Patterson and Kinsman,
1974; Sass and Katz, 1982; Sass and Bein, 1988). Patterson and Kinsman
(1974) observed increasing Ca-enrichment with increasing depth that followed a
decrease in the Mgz*:Ca2+ ratio of local pore waters of a sabkha environment.
They interpreted this to represent a geologic situation where increasing isolation
resulted in lowering Mgz“:Ca2+ ratios in solution. Sass and Katz (1982)
suggested that dolomite stoichiometry and solution Mgz“:Ca2+ ratios were related
based on observations from a Cretaceous carbonate sequence in the Judean
Mountains in Israel. Their data show greater compositional variations and more
Ca-enrichment in dolomites found lower in the section. Based on dolomite
NalCa ratios and isotope data, they interpreted the compositional changes with
depth to reflect the degree of isolation from overlying seawater. The shallower,
more stoichiometric dolomites were interpreted to represent earlier, less isolated

crystallization, whereas deeper, Ca-rich dolomites crystallized from older, more

183

modified solutions. Rosen et al. (1989) proposed that the Mg:Ca in dolomites of
the Coorong region reflect the solution Mgz":Ca2+ ratio of the parent fluids. TEM
observations of growth microstructures and geochemical (isotopic) analyses
implied that Mg-rich dolomites underwent rapid precipitation from brines with
highly elevated Mgz*:Ca2" ratios, whereas the Ca-rich dolomites precipitated
more slowly from solutions that were less evaporative. Mg-rich dolomites
exhibited extremely defect-rich miscrostructures and heavy 6180 values.
Despite a decreasing Mgz*:Ca2“ ratio in solution as the dolomite replaces
calcite, synthetic dolomite composition is invariably stoichiometric after reactant
depletion. This is consistent with the findings of Schmidt (1965), Langbein et al.
(1984), Lumsden and Chimahusky (1980), and Sperber et al. (1984), who
showed that dolomite in partially dolomitized limestone is generally more
calcium-rich whereas completely dolomitized rocks are found to be more
stoichiometric. This is also consistent with the observation from our experiments
that over most solution Mgz“:Ca2+ ratios studied here, initial products are
generally Ca-rich, but become stoichiometric after reactant depletion regardless
of the solution Mgz*:Ca2" ratio or the stoichiometry of initial products. It is
noteworthy to mention at this point that although no relationship was observed
between the percent product and dolomite stoichiometry in the calcite reactant
experiments, other high-temperature dolomitization studies have reported
increasing dolomite stoichiometry with reaction progress (e.g., Nordeng and
Sibley, 1994). However, the results presented here and those of Sibley (1990)

suggest that there is no simple relationship between the solution composition and

184

that of the solid product. The stoichiometry of dolomites formed prior to reactant
depletion is clearly a function of the initial solution chemistry, but all experiments
proceeding to reactant depletion were stoichiometric. Therefore, without
knowledge of the solution chemistry, stoichiometric dolomites must be interpreted
with caution.

4.j.7. Cation Order: The experimental results suggest that although
dolomite stoichiometry is related to the initial Mgz“:Ca2+ chemistry of the parent
fluids, the degree of cation order is a function of reaction progress and is
independent of solution chemistry. Although it was possible to form
compositionally stoichiometric dolomite directly from calcite, the dolomite was
nonetheless characterized by a low degree of cation order and therefore
interpreted as nonideal. Ordering increased throughout the reaction, both before
and after reactant depletion, albeit at a much slower rate after the calcite reactant
was consumed. Synthetic dolomites formed prior to reactant depletion are always
less ordered than dolomites formed after reactant depletion. Ordering ratios
achieved by synthetic nonideal dolomites with growth islands ranged from 0.0 to
0.45, whereas synthetic ideal dolomites with layers had ordering ratios that
ranged from 0.5 to 0.88. This indicates that dolomites formed earlier in the
reaction are generally poorly ordered compared to those formed later. Gregg et
al. (1992) showed that the degree of cation order in natural dolomites from Belize
increased with depth without a concomitant increase in stoichiometry. These
observations are consistent with the experimental data presented here, which

show cation ordering and dolomite stoichiometry are independent. Furthermore,

185

better cation order is achieved with a spiral growth mechanism than during
polynuclear growth. This observation is consistent with the observations of
natural dolomite samples reported here, which suggest ideal dolomites
correspond to a spiral growth mechanism whereas nonideality corresponds to
dolomites that grew by a polynuclear grovvlh mechanism.

Many recent sedimentary dolomites have a low degree of cation order and
are Ca-rich. Garcia et al. (2001) reported on the occurrence of primary dolomite

lake deposits. Dolomites were calcium-rich (51-54 mol% CaC03) with ordering

ratios that ranged from 0.27 to 0.48 (7? = 0.36). In a DSDP report, Bottcher et al.

(1998) noted relatively low cation ordering for Pliocene and Miocene dolomites.

The degree of ordering ranged between 0.39 and 0.56 (Y = 0.45). Other modern
dolomite from the Coorong area, Australia (Von der Borch and Jones, 1976;
Rosen et al., 1988), Sugarloaf Key, Florida (Carballo et al., 1987), Ojo de Liebre,
Mexico (Pierre et al., 1984), West Caicos, British West Indies (Perkins et al.,
1994) and Western Australia (DeDekker and Last, 1988) have been reported to
be Ca-rich and poorly-ordered.

Based on laboratory work and observations from modern and relatively
recent environments, Bathurst (1975) proposed that a high Mg""‘:Ca2+ ratio was
required for dolomite ordering. He argued that a high Mg activity relative to Ca in
the solution is required to give the chemical potential necessary to achieve the
high degree of order in the dolomite lattice. According to our experimental
results, ordering is neither a function of dolomite stoichiometry nor a function of

the M925032+ ratio in solution. More specifically, low Mg”:Ca2+ ratios were

186

observed to be equally as efficient in producing ordered dolomites than higher
solution Mgz*:Ca2* ratios. Some solutions with lower Mgz":Caz* ratios (e.g., 0.66)
produced dolomites with a higher degree of cation order in less time than it took
some experiments with higher solution Mgz*:Ca2“ (e.g., 1.5). This is interesting
because lower solution Mgz’*:Ca2+ ratios have slower dolomitization rates and
ordering increases with reaction progress, therefore we would predict ordering to

increase at a slower rate in low Mgz”:Ca2+ solutions.

4. k. Implications for Natural Dolomite

The similarities between natural and synthetic dolomites discussed in the
previous section permit observations from synthetic dolomite experiments to be
used to infer factors that affect the kinetics of dolomitization in natural settings
with a reasonable level of confidence. One of the geologically significant
inferences based on the empirical findings is that solution Mg""‘:Ca2+ ratios do not
dictate crystal growth mechanisms in natural dolomite. This is striking because
according to crystal growth theory, one would predict spiral growth to occur under
relatively low solution Mgz*:Ca2+ ratios and polynuclear growth during relatively
high solution Mgz”:Ca2+ ratios. As demonstrated by the experimental results,
however, crystal growth mechanisms are independent of solution Mgz“:Ca2+
ratios over the entire range included in this study. According to seawater curves
for Mg” and Ca2+ from Lowenstein et al. (2001) and Horita et al. (2002),
Mgz*:Ca2” ratios have changed considerably throughout the Phanerozoic.

Though there is no way of knowing the exact compositions of dolomitizing fluids,

187

one can assume that most natural dolomite deposits formed from relatively
normal seawater or seawater derivatives. If this is correct, it is reasonable to
infer that a wide range of solution Mgz*:Ca2" ratios is represented by the natural
samples studied here. To this end, nonideal dolomites formed by polynuclear
growth occur over a wide range of geologic time periods and formation
conditions. For example, based on the seawater chemistry curves of Lowenstein
et al. (2001) and Horita et al. (2002), nonideal dolomites from the Ordovician, a
period of relatively low Mgz":Ca2+ ratios, have the same island nanotopography
as recent dolomites that formed during geologic time periods with relatively high
Mg"’*:Ca2+ ratios. Furthermore, the natural nonideal dolomites studied here,
which formed by polynulcear growth, range from 54 mol% CaCOa to
stoichiometric. This further suggests that natural nonideal dolomites form by a
polynuclear mechanism independently of dolomite stoichiometry, an inference
consistent with the empirical observations.

Earlier discussions emphasized the idea that the transition from
polynuclear growth to spiral growth is indicative of recrystallization in synthetic
dolomite. Because the surface features observed on natural ideal and nonideal
dolomites are consistent with observations of synthetic surface nanotopography,
it may be possible to interpret natural dolomite in the same context. For
example, if two natural stoichiometric dolomites were both determined to have
ordering ratios of 0.8, and the first was observed to have an etched surface
covered with islands and the second with layers and euhedral etch pits, the latter

would be interpreted as a recrystallized dolomite. This is not to say that the

188

dolomite with islands is not recrystallized, only that layers with pits would be clear
evidence of recrystallization according to the experimental observations.
However, because the transition from polynuclear to spiral growth in the high-
temperature synthesis experiments occurs only after calcite reactant depletion, it
is not easy to distinguish between the effects of recrystallization and inferred
changes in carbonate ion concentration. Because the relationship between Ca”,
Mg”, and C032' in nature is often unknown, one cannot determine whether spiral
growth indicates the process of recrystallization, or dolomite formation under
conditions where the carbonate ion concentration at the growth interface is low.
According to the carbonate concentration argument, it may be possible to directly
form a dolomite with layers, without the presence of a nonideal dolomite
precursor, if the carbonate concentration is low enough.

Mimetic replacement of CaCO3 by dolomite indicates that dolomitization
may proceed by micrometer or smaller scale dissolution-reprecipitation. As a
result, carbonate concentration in the bulk solution should be much less
important than the carbonate concentration at the interface between growing a
dolomite crystal and the dissolving carbonate substrate. Following this
argument, the growth mechanism may be more indicative of the carbonate ion
concentration at the surface of the precursor phase (e.g., calcite or nonideal
dolomite) than the carbonate concentration in solution. Two observations from
this study support the importance of carbonate ion concentration at the growth
interface. First, dolomite nucleation occurs preferentially on the edges and

comers of calcite. One presumes nucleation would not readily occur at corners

189

and edges because a nucleus on a corner or edge would have a relatively small
contact area with the substrate. This small contact area would cause the nucleus
to have a relatively high surface free energy. The same size nucleus in a kink
site would have a much larger contact area with the substrate and, therefore, a
lower surface free energy. As a result, the nucleus in the kink site would be more
stable than the nucleus on an edge or corner. In order for the nucleus on the
edge to become more stable than a nucleus in a kink site, it would have to be
considerably larger. To this end, nuclei on edges and corners may be much
larger than those in surface kinks because edges and corners (of CaCOa) are
likely to dissolve faster and dissolution of the substrate is the source of carbonate
ions for the growing dolomite. Second, selective dolomitization of aragonite
occurs when both calcite and aragonite are present. A bomb solution that
equilibrates with respect to aragonite should allow dolomite to form readily on
calcite. In fact, based on surface energy considerations, it is more favorable for
dolomite to grow on calcite instead of aragonite because calcite and dolomite
have the same crystal system. Furthermore, dolomite should form on calcite in
bombs with mixed calcite-aragonite faster than it forms in bombs with just calcite
because the carbonate ion concentration in the bombs with calcite and aragonite
must be higher than in the bombs with just calcite. However, in bombs with
mixed aragonite and calcite reactant, dolomite does not form on the calcite.
These observations are consistent with the model that surface concentration of

carbonate lens has a greater effect on the kinetics of dolomitization than surface

190

free energy of equivalent size nuclei or bulk solution carbonate ion concentration
arguments.

There is no mechanistic basis in crystal growth theory that explains why
dolomite growth mechanisms change when reactants are depleted in the
synthesis experiments. The free energy drive (based on the Mgz":Ca2* in
solution) is relatively constant at reactant completion and so is the composition of
the dolomite phase. However, after reactant calcite depletion, ideal dolomite
utilizes carbonate from the dissolving nonideal dolomite, whereas prior to
reactant depletion nonideal dolomite uses the dissolving calcite substrate as its
carbonate source. The substrates, calcite and nonideal dolomite, have different
solubility products. Following this idea, it may be possible that differences in
solubility are able to explain the change in growth mechanisms at reactant
depletion. Given the Ksp values for nonideal dolomite and calcite from Helgeson
et al. (1978), one can calculate the carbonate concentration in solution for each
of the substrates. If the concentration of Ca2+ is fixed, for example at 103, then
for dolomite dissolution, the concentration of M92+ must be the same as Ca”
because for every mole of Ca2+ that dissolves a mole of magnesium is also
dissolved. This means that the concentration of Mgz+ is also 103. For a nonideal
dolomite with K.,, = 101652, the concentration of Ca2+ times concentration of Mg2+
is 10‘“, and the concentration of C032' squared must be 10"°°52. That means the
concentration of 0032' in solution would be 10'5"!6 when nonideal dolomite is
dissolved. Now for calcite, if Ksp = 10‘“52 and given that Ca2+ is 103, the

concentration of 0032' must be 105'”. This simple calculation demonstrates that

191

the carbonate concentration differences for calcite and nonideal dolomite are
relatively minor. Therefore, the solubility products of calcite and nonideal
dolomite provide very little insight into the observed growth mechanism change
between nonideal dolomite formed on calcite and ideal dolomite formed on
nonideal dolomite substrates.

From field data, it is well known that dolomite weathers more slowly that
calcite (Cowell and Ford, 1980; Atkinson 1983). This has also been confirmed by
experiment when Chou et al. (1989) showed that calcite dissolved eight to nine
times faster than dolomite. In contrast, Busenberg and Plummer (1982) showed
that it took different dolomite specimens 25-250 times longer than calcite to
dissolve. The kinetic details of calcite dissolution seem to be well understood
(Plummer et al., 1979), but dolomite dissolution is not (Busenberg and Plummer,
1982; Chou et al., 1989), except far from equilibrium. Under a range of P002
values, Busenberg and Plummer (1982) measured dissolution rates for a variety
of natural dolomite samples. From this data, a rate expression in terms of three
fonlvard and one backward reaction was derived. Busenberg and Plummer
(1982, 1989) proposed that the CaCOa part of dolomite dissolves more readily
that the MgCOa component. Consequently, the rate of dolomite dissolution is
dependent on the rate of the M9003 dissolution. Chou et al. (1989) later
demonstrated a difference of four orders of magnitude in the rate constants
between magnesite and calcite.

The rate of dissolution for dolomite and calcite are significantly different.

This indicates that the carbonate flux to the growth surface must be highly

192

dependent on the nature of the substrate. Because the rate of dissolution of
nonideal dolomite is significantly lower than the rate of dissolution of calcite
(Busenberg and Plummer, 1982; Chou et al., 1989), the carbonate ion flux to the
surface of a dissolving nonideal dolomite grain must be lower than at the surface
of a dissolving calcite grain. Therefore, a nonideal dolomite substrate will have a
lower carbonate ion concentration at a dolomite growth/substrate dissolution
interface than a calcite substrate. Without the presence of a nonideal dolomite
precursor for ideal dolomite to grow on, the carbonate flux to the dolomite growth
interface would be dictated by more soluble substrates, such as calcite and
aragonite. It has been shown experimentally that calcite and aragonite reactants
support polynulcear growth of nonideal dolomite under a wide range of
conditions. Only when the substrate and the carbonate source changes to
nonideal dolomite, does the growth mechanism change from polynuclear to
spiral. Therefore, the observation of layers only on ideal dolomite is consistent
the carbonate flux argument in that a nonideal dolomite precursor is necessary
for spiral growth. Therefore, it is reasonable to suggest that layers are only
associated with ideal dolomite because it is only with a nonideal dolomite
precursor that the carbonate flux to the interface between the dissolving
substrate and growing dolomite is low enough to for spiral growth to out compete
polynuclear growth.

If a nonideal dolomite must exist in order to provide a substrate for spiral
growth, then the ideal dolomite with evidence of spiral growth inevitably replaced

an older dolomite substrate. In the experiments, the formation of ideal dolomite

193

took place after the nonideal dolomite was already stoichiometric. Therefore, the
only changes to occur between the nonideal and ideal phases were in the degree
of cation order and the surface nanotopography. Mineral replacement of this sort
is consistent with what is generally accepted as recrystallization. Therefore, an
ideal dolomite that grew by spiral growth should be interpreted as a recrystallized
dolomite.

Dolomite solubility increases with excess calcium content and with
decreasing cation order (Carpenter, 1980). It is, therefore, reasonable to
suggest, at least on thermodynamic grounds, that nonideal dolomite (either
poorly ordered and/or Ca-rich) is susceptible to recrystallization. Although the
persistence of nonstoichiometric, nonideal dolomite in seawater over millions of
years is a common phenomena (e.g., Lumsden, 1988), some studies show that
recrystallization occurs in relatively recent dolomitic sediments (e.g., Gregg et al.,
1992). Land (1992) proposed that most ancient dolomites have been
recrystallized; furthermore that it is only through multiple recrystallization events
that the formation of ideal dolomite is possible. Progressive recrystallization is
consistent with the experimental observation, which show that the degree of
cation order continues to increase throughout dolomitization. Furthermore, the
experimental observations presented here indicate that ideal dolomite only forms
by replacing a metastable, nonideal dolomite precursor. In addition, modern
dolomites are characteristically nonideal (typically Ca-rich and poorly ordered),

whereas ancient dolomites are generally more ideal.

194

Providing that conditions are constant, systems tend toward equilibrium
with the passage of time. This suggests that older nonideal dolomites have a
greater potential for recrystallization because older rocks are increasingly
exposed to conditions that may favor dolomitization, such as influx of pore fluids,
as well as elevated temperatures and pressures (Lumsden and Chimahusky,
1980). F uchtbauer (1972) noted the calcium content of natural dolomites
decreased in response to increasing diagenetic change. Sperber et al. (1984)
observed a paucity of Ca-rich dolomites in samples Jurassic in age and older (2
of 55 samples), whereas Lumsden and Chimahusky (1980) observed that 43 of
290 pre-Cenozoic dolomites were Ca-rich. Therefore, it is not clear whether
there is a relationship between dolomite stoichiometry and age. However, given
enough time, solutions supersaturated with respect to dolomite should yield
dolomite with compositions and cation order that tend toward ideality. This
hypothesis is consistent with our experimental results, which show that dolomite
always achieves a stoichiometric composition and relatively high degree of order
given long enough.

Because there is currently no independent test of recrystallization for
dolomite, previous studies have relied on geochemical tests, a somewhat
tenuous approach, as our experiments have shown, when trying to identify
recrystallization in dolomite. As the synthesis experiments have shown,
examining geochemistry alone is not sufficient in attempting to identify
recrystallization. First, there is no change in nonideal dolomite stoichiometry until

the point of reactant depletion, despite a continually increasing degree of cation

195

order. Secondly, nonideal dolomite stoichiometry only reflects the composition of
the solution during the start of dolomitization, but following reactant depletion all
dolomite is stoichiometric.

Recrystallization has important implications for using dolomite
geochemistry to infer conditions about the environment of formation. In
particular, if a dolomite is recrystallized, its stoichiometry may no longer reflect
the chemistry of the solution from which it initially formed, but instead reflect
conditions during the latest recrystallization event. As the experiments here
have shown, dolomite stoichiometry may also not reflect the solution from which
it formed if it has recrystallized. All synthetic dolomites, regardless of the
solutions from which they formed, were stoichiometric after reactant depletion.
As Land (1980) reported, most ancient dolomite is trace-element depleted
relative to Holocene dolomites. This suggests that either Holocene dolomite is
formed under significantly different formation conditions than ancient dolomite, or
that most ancient dolomites have been recrystallized. If the latter is true,
valuable geochemical information about the original formation conditions has
been lost.

According to the seawater curves of Lowenstein et al. (2001) and Horita et
al. (2002) the Ordovician corresponds to low seawater Mg""‘:Ca2+ ratios. This is
interesting because some Ordovician dolomites included in this study are Ca-rich
with surfaces covered with islands, whereas others are ideal and are
characterized by layers with etch pits. According to the arguments described

above, nonideal Ordovician dolomites presumably represent rocks that have not

196

been fully recrystallized and may even contain information about the original
dolomitizing conditions, whereas ideal Ordovician dolomites with layers likely
represent rocks that have been recrystallized. As the synthesis experiments
have shown, examining geochemistry alone is not sufficient in attempting to
identify recrystallization. First of all, there is no change in nonideal dolomite
stoichiometry until the point of reactant depletion, despite a continually increasing
degree of cation order. Furthermore, nonideal dolomite stoichiometry seems
only to reflect the composition of the initial solution. Following reactant depletion,
however, all dolomite is stoichiometric. As the earlier discussions alluded to, the
surface nanotopography approach may be a powerful tool that offers an
independent assessment of recrystallization for ancient dolomites.

Based on the premise that relatively high solution Mgz*:Ca2"
concentrations should promote dolomite-dominated carbonate rocks as well as
more stoichiometric dolomite, Lumsden and Chimahusky (1980) initially
postulated that the percent dolomite in a carbonate rock should be directly
related to its stoichiometry. This is consistent with the data of Schmidt (1965)
and Langbein et al. (1984), who showed a relationship between percent dolomite
and dolomite stoichiometry in some ancient dolomitic carbonate rocks. Swart
and Melim (2000) reported an R2 of 0.13 between the percent dolomite and
stoichiometry in Tertiary sediments of the Great Bahama Bank. Gregg et al.
(2001) reported an R2 of 0.115 between the percent dolomite and stoichiometry
in sediments from Belize. The relatively low correlation coefficients make it

difficult to accurately discern if, in fact, there is a physical relationship. The data

197

presented in this study from a suite of stratigraphically related samples from Abu
Dhabi (GWP samples) also demonstrate a possible relationship between percent
dolomite and dolomite stoichiometry. The results plotted in Figure 62 show a

correlation between dolomite composition (reported as degrees 2-theta) and the

198

 

l Dolomite I (Dolomite + Calcite + Aragonite)

 

R2 = 0.3706

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

fl

l O 0.1 0.2 0.3 0.4 0.5 0.6 0.7
l Percent Product l

 

Figure 62. 2-theta (i.e. dolomite stoichiometry) vs. percent dolomite for GWP
data from Abu Dhabi, R2=0.3706. Note: 2-theta values are proportional to
dolomite stoichiometry.

199

percent products [dolomitez (dolomite + calcite + aragonite) ratio] in the rock
(R2=0.37).

A correlation between percent dolomite in a carbonate rock and dolomite
stoichiometry can be interpreted in two different ways. First, it may be
interpreted to indicate that dolomite stoichiometry is dependent on reaction
progress. In other words, the dolomite becomes increasingly stoichiometric as
more dolomite replaces the precursor carbonate. However, a second
interpretation, one more consistent with the experimental observations, is more
appropriate. It postulates that variations in percent dolomite and dolomite
stoichiometry are the result of different solution chemistry. Our high-temperature
experiments clearly demonstrate that dolomite stoichiometry is highly dependent
on the initial solution Mgz”:Ca2" ratio and that stoichiometry remains remarkably
steady throughout the reaction prior to reactant depletion. These findings
indicate that no correlation exists between reaction progress (i.e., the percent
dolomite in the carbonate fraction) and dolomite stoichiometry (in contrast to the
findings of Malone et al., 1996). Based on the low correlation coefficients
discussed earlier for percent dolomite and dolomite stoichiometry in natural
rocks, it appears that the experimental results and natural observations coincide
relatively well. Recall that the experiments show that the rate of dolomitization
was also heavily dependant on solution M92":Ca2+ ratios. In conjunction with the
observed correlation between solution Mgz*:Ca2" and nonideal dolomite
stoichiometry, higher initial solution Mgz*:Ca2+ and hence faster growth rates

produce greater abundances of more stoichiometric dolomite. Therefore,

200

correlations between percent products and dolomite stoichiometry, such as the
ones observed by Schmidt (1965) and Langbein et al. (1984) for geologically
related rocks, are interpreted to indicate that dolomitization occurred in a
geological environment with variable solution Mgz":Ca"’+ ratios at the time of
dolomite initiation. The argument here rests on the idea that dolomite does not
become more stoichiometric simply because the percent dolomite in the rock
increases, but that more dolomite (i.e., higher percent product) and more
stoichiometric dolomite form because a higher solution Mgz“:Ca2+ ratios lead to
faster rates of dolomitization. In ancient carbonate rocks where the percent
dolomite does correlate with dolomite stoichiometry (e.g., Schmidt, 1965 and
Langbein et al., 1984), areas with higher percentage of dolomite and lower Ca-
enrichment would be interpreted to indicate initial solutions with higher Mgz“:Ca2+
and faster growth rates, not just to represent an area in which dolomitization has
progressed further with time. Conversely, areas with a lower percentage of
dolomite in the carbonate rock and more Ca-enrichment suggest lower Mgz*:Ca2*
in solution and in turn slower growth rates. Therefore the driving force behind
correlations between stoichiometry and percent dolomite are likely due to
differences in the Mgz*:Ca2“ ratio In solution at the onset of dolomitization.
Furthermore, this may indicate an environment in which the solution chemistry is
geologically controlled. One geologically reasonable, and potentially common
situation is envisioned where Mg-rich fluids migrate downward and dolomitize
carbonate sediments. During the migration the Mgz*:Ca2“ ratio of these solutions

is reduced as Mg2+ is consumed and Ca2+ is liberated during the replacement

201

reaction. Therefore, sediments nearer to the Mg-rich fluids are dolomitized faster
and have more stoichioemtric compostions, whereas sediments lower in the
section and farther away from the source of Mg-rich fluids are dolomitized more
slowly and end up being more Ca-rich.

Lumsden and Chimahusky (1980) noted that dolomites became
increasingly Ca-rich with increasing distance from overlying evaporite beds in the
Salina-Niagaran Suite (Sanders and Horicon wells). Zenger (1996) observed a
similar pattern in the dolomites of the Red River Formation in eastern Montana.
Both studies interpreted this correlation to indicate a change in solution
Mgz":Ca2+ ratio with distance from the overlying dolomitizing solutions. In
particular, they suggested that that dolomite formed by downward seeping fluids
that become increasingly depleted in M9“ as dolomite precipitates. This
suggests that geologic controls on Mgz":Ca2+ ratios in nature may also explain
the correlation between percent dolomite and dolomite stoichiometry observed by
Schmidt (1965) and Langbein et al. (1984). If the Mgz*:Ca2+ in dolomitizing
solutions is geologically controlled, we might predict a correlation between
dolomite stoichiometry and percent dolomite in the rock. Once again, according
to our experiments, carbonate sediments closer to the source of magnesium will
dolomitize faster and have a more stoichiometric composition, whereas deposits
down section will be exposed to solutions with lower Mgz*:Ca2+ ratios and
therefore have slower rates dolomitization and more Ca-rich compositions.

Graf and Goldsmith (1958) first documented a covariance between

evaporite conditions and near-stoichiometric dolomite. Fuchtbauer and

202

Goldsmith (1965) found dolomites in evaporite-related deposits have more
stoichiometric compositions than normal marine deposits. Fuchtbauer (1972)
also noted the stoichiometric dolomite is associated with evaporite basin
conditions as well as the presence of anhydrite. Fuchtbauer (1974) related
calcium content in dolomite to the salinity of dolomitizing fluids. He documented
a change from Ca-rich to nearly stoichiometric dolomite in going from humid to
arid modern environments. Lumsden and Chimahusky (1980) observed that
dolomicrites have a bimodal distribution in which near stoichiometric samples
(50-52% CaC03) consist of mainly evaporitic rocks, whereas Ca-rich samples
(53-55% CaCOa) were mostly from non-evaporitic setting. These observations
are important because evaporitic settings promote higher Mgz“:Ca2+ ratios in
dolomitizing solutions due to precipitation of gypsum and calcite (Alderman,
1965; Land, 1967). The occurrence of accessory anhydrite in dolomites and
limestones with near-stoichiometric dolomite supports the interpretation that high
Mgz*:Ca2* ratios in solution result in near-stoichiometric dolomite (Fuchtbauer,
1974). Higher Mgz‘YCa2+ also correspond to higher dolomitization rates, which is
consistent with the occurrence of dolomite in modern environments with high
evaporation rates (llling et al., 1965; Deffeyes et al, 1965; Shinn et al., 1965).
Although percent dolomite and dolomite stoichiometry seem to covary in
some cases where carbonate rocks are stratigraphically related (e.g., Schmidt,
1965; Langbein et al., 1984; Figure 62-GWP), the correlation is lacking when
unrelated rocks are examined. Lumsden and Chimahusky (1980) observed no

correlation between percent dolomite in carbonate rocks and dolomite

203

stoichiometry (R2 = 0.037) for 290 Phanerozoic shelf carbonates from the US.
Sperber et al. (1984) examined an additional 54 carbonate samples and also
observed no correlation, suggesting that dolomite stoichiometry is independent of
reaction progress for geologically unrelated samples. These finding are
consistent with the experimental results (Figure 14), which indicate that the
dolomite stoichiometry is independent of percent dolomite when data from all the
experiments are compared. Because the data presented in Lumsden and
Chimahusky (1980) and Sperber et al. (1984) include carbonate rocks from many
different and unrelated depositional environments, a relationship between
stoichiometry and percent dolomite is not expected, nor is it observed.

Based on the empirical findings, one would predict that times of higher
Mgz“:Ca2+ in seawater would correspond to times of greater dolomite abundance
because it would form faster. However, the Early Paleozoic is a period of high
dolomite/limestone ratio (Given and Wilkenson, 1987 and references therein;
Sun, 1994 and references therein), but low inferred seawater Mgz“:Ca2+ ratios
(Lowenstein et al, 2001). No correlation between these secular variations of the
dolomite/limestone ratio and seawater M92“:Ca2+ chemistry throughout the
Phanerozoic implies that either (1) Mgz“:Ca2+ ratio of seawater was not a major
factor determining the amount of dolomite in ancient rocks, (2) extensive
dolomitization in the early Paleozoic may have actually lowered seawater
Mgz":Ca2+ ratios, as proposed by Wilkenson and Algeo (1989), or (3)
dolomitization of most carbonate rocks occurs in fluids that are significantly

different from seawater. The second hypothesis does not preclude the

204

importance of Mgz“:Ca2+ ratios in natural dolomitization. However, it does imply
that other factors may have been more important. The third hypothesis is
consistent with the idea that dolomitizing fluids are generally seawater derivatives
(Given and Wilkenson, 1987), but not necessarily normal seawater. As it
stands, unmodified seawater appears to be relatively inefficient at dolomitizing
sediments as modern and many ancient marine carbonates have not been
dolomitized.

The experimental results presented here are consistent with the argument
of Clayton et al. (1968), Lippmann (1973), and Hardie (1987), that carbonate may
be a major factor in affecting dolomitization kinetics. As Hardie (1987) notes,
high Mgz“:Ca2+ ratios are not necessary to increase the solubility product of
dolomite, particularly when carbonate is readily available. But as Burns et al.
(2000) outlined, secular variations in seawater alkalinity are also unable to
explain dolomite abundance through the Phanerozoic. As discussed in
Lippmann (1973), however, a more local source of carbonate in marine
environments may be supplied by anaerobic bacterial sulfate reduction.
Vasconcelos and McKenzie (1997) proposed the microbial dolomite model as a
means to explain dolomitization of recent sediments in the lagoonal waters of
Lagoa Vennelha, Rio De Janeiro, Brazil. The model is appealing in two respects:
as a part of their normal metabolic processes, sulfate reducing bacteria 1)
release carbonate into solution (in the form of 0032' or HCO3'), and 2) remove
sulfate, a proposed kinetic inhibitor to dolomitization (see Baker and Kastner,

1981). Therefore microbial sulfate reduction would in effect, create an

205

environment on the submicrometer-scale that is more favorable for dolomite
formation. The model is supported by the findings of Warthmann et al. (2000),
who produced dolomite at low temperatures (30°C) using sulfate-reducing
bacteria. Furthermore, Burns et al. (2000) showed a correlation between periods
of extensive dolomitization and times during the Phanerozoic when redox
conditions were favorable for anaerobic bacteria. The similarities between the
high-temperature inorganic dolomites and ancient, low-temperature dolomites
presented here are remarkable. Although available data for organogenic
dolomites is extremely limited, they do indicate spherulitic growth and not growth
islands. At this point, it is still difficult to evaluate whether organogenic dolomites
are capable of explaining observations of natural dolomites in the same way that

the high-temperature experiments can.

206

5. CONCLUSIONS

High-temperature dolomitization of calcite or aragonite reactants is
characterized by the formation of an initial poorly ordered, nonideal synthetic
dolomite. Following reactant depletion, initial nonideal dolomite products are
later replaced by relatively well-ordered, ideal dolomite. Dolomitization rates are
highly dependent on the initial solution Mgz*:Ca2* ratio. High solution M92”:Ca2+
ratios dolomitized reactants relatively rapidly (e.g., 30 hours), whereas low
solution Mgz*:Ca2+ ratios require very long reaction times (e.g., >1000 hours).
The stoichiometry of the initial nonideal dolomite is also highly dependent on the
starting Mg""‘:Ca2+ ratio in solution (R2 = 0.97). Relatively high solution Mgz*:Ca2”
ratios yield initial products that are stoichiometric, whereas lower solution
Mg""‘:Ca2+ ratios yield initial products that are calcium-rich. The stoichiometry of
the nonideal dolomite remains relatively steady prior to reactant depletion despite
a solution chemistry that continually evolves with reaction progress. After
reactant depletion, dolomite is stoichiometric regardless of the initial or final
solution Mgz*:Caz* ratio, suggesting that the relationship between solution
Mgz":Ca2+ and dolomite stoichiometry is not direct. Cation order increases with
reaction progress as reactants are dolomitized. After calcite reactant depletion,
ordering continues to increase as ideal dolomite forms at the expense of nonideal
dolomite. In some experiments, ordering increases more slowly after reactant

depletion.

207

Two distinct nanometer-scale growth features, described here as islands
and layers, characterize synthetic dolomite growth surfaces. Prior to calcite
reactant depletion, growth islands are observed on the surfaces of nonideal
dolomite. Islands are analogous to surface nuclei observed on other minerals
and are consistent with a model of polynuclear growth. After calcite reactant
depletion, ideal dolomite nanotopography is characterized by layers with steps
that emerge from the crystal. Layer geometry is consistent with a theoretical
model of spiral growth. According to crystal growth theory, polynuclear growth is
indicative of conditions where the free energy drive is relatively high, whereas
spiral growth is indicative of conditions where the free energy drive is relatively
low. Calcite reactant experiments demonstrate, however, that despite highly
variable solution Mgz*:Ca2* ratios, nonideal synthetic dolomites always grow by a
polynuclear mechanism prior to calcite reactant depletion and ideal dolomites by
a spiral mechanism after calcite reactant depletion. These findings contradict the
hypothesis proposed by Sibley et al. (1987) that induction period of dolomitization
marks a change in growth mechanisms from polynuclear to spiral growth.

A number of observations presented here indicate that dolomitization may
proceed by micrometer or smaller scale dissolution-reprecipitation. After calcite
reactant depletion, the flux of carbonate to the dolomite growth interface must be
lowered because [0032“] is dictated by slower dissolving nonideal dolomite. This
suggests that rate of dissolution of the reactant and thus the flux of carbonate to
dolomite growth surfaces dictate dolomitization kinetics in high-temperature

synthesis experiments. The importance of carbonate flux to the growth interface

208

is further supported by the observations that dolomite preferentially replaces
aragonite when both calcite and aragonite are present, and that dolomite nuclei
form preferentially on the edges and corners of calcite. The present observations
contradict predictions based on crystal growth theory. However, although growth
mechanisms are processes that must obey thermodynamics, they are ultimately
dictated by geochemical kinetics. Therefore, theoretical growth models may not
be able to fully explain complex geochemical phenomenon for complex minerals
like dolomite. Furthermore, because growth layers are only observed after all
initial calcite reactants are consumed, they are indicative of recrystallization in
high-temperature synthetic dolomite.

Chemically etched synthetic dolomites that grew by polynuclear and spiral
growth yield islands and euhedral etch pits, respectively. Dissolution islands are
indistinguishable from the growth islands observed on synthetic dolomite growth
surfaces. Euhedral etch pits occur on relatively flat surfaces and are interpreted
to form by preferential dissolution along linear defects, such as spiral
dislocations. Chemically etched natural dolomites exhibit dissolution islands and
deep euhedral etch pits on flat layers. Despite a large range of ages and
stoichiometries, Ca-rich and relatively poorly ordered stoichiometric (nonideal)
dolomites are invariably characterized by dissolution islands. Relatively well-
ordered and stoichiometric (ideal) dolomites are characterized by deep euhedral
etch pits on bread flat layers. Islands and etch pits observed on natural
dolomites are nearly identical to islands and etch pits found on etched synthetic

dolomite crystal surfaces. This suggests that by using chemical etching, it is

209

possible to uncover ancient nanometer-scale growth features on the surfaces of
natural dolomite crystals.

The findings presented here also demonstrate the importance of
examining mineral textures when attempting to understand the evolution of
ancient carbonate rocks. Growth layers on ideal dolomite appear to be
indicative of a growth mechanism that only occurs during the replacement of a
nonideal dolomite precursor. It is only with slow dissolving nonideal dolomite
substrate that carbonate flux to the ideal dolomite growth interface is low enough
for spiral growth to out compete polynuclear growth. Without the presence of a
nonideal dolomite precursor for ideal dolomite to grow on, the carbonate flux to
the dolomite growth interface would be dictated by more soluble substrates, such
as calcite and aragonite. Therefore the presence of layers appears indicative of
recrystallization in ideal dolomite.

The dolomite problem is rooted in the kinetics of nucleation and growth.
Because of sluggish kinetics at low temperatures, high-temperature synthetic
dolomites are often used to better understand dolomitization in natural settings.
To this end, similarities between natural and synthetic dolomite nanotopography
suggests that synthetic and natural dolomites form by the same crystal growth
mechanisms despite large discrepancies in formation conditions. Therefore,
synthetic dolomite should be used as a proxy for natural low temperature
dolomite with a higher level of confidence than previously acknowledged.
Therefore, a long induction period followed by rapid dolomitization is the best

model to explain the absence of dolomite in modern evironements.

210

REFERENCES CITED

-Alderman, AR, 1965, Dolomitic sediments and their environment in the South-
East Australia, Geochimica et Cosmochimica Acta, Vol. 29, pp. 1355-1365.

-Arvidson, RS. and Mackenzie, F .T., 1997, Tentative kinetic model for dolomite
precipitation rate and application to dolomite distribution, Aquatic Geochemistry,
Vol. 2. pp. 273-298.

-Arvidson, RS. and Mackenzie, F .T., 1999, The dolomite problem: control of
precipitation kinetics by temperature and saturation state American Journal of
Science, Vol. 299, pp. 257-288.

-Arvidson, RS. and Mackenzie, F.T., 2000, Temperature dependence of mineral
precipitation rates along the CaCO3-MgCO3 join, Aquatic Geochemistry, Vol. 6,
pp. 249-256.

-Astilleros, J.M., Pina, C.M., Femandez-Diaz, L., and Putnis, A., 2002, Molecular-
scale surface processes during the growth of calcite in the presence of
manganese, Geochimica et Cosmochimica Acta, Vol. 66, No. 18, pp. 3177-3189.

-Atkinson, TC, 1983, Growth mechanisms of speleothems in Castleguard Cave,
Columbian lcefields, Alberta, Canada, Arctic Alpine Research, Vol. 15, pp. 523—
536.

-Avrami, M., 1939, Kinetics of phase change, I. General theory, Journal of
Chemical Physics, Vol. 7, pp 1103-1112.

-Baker, PA. and Kastner, M., 1981, Constraints on the formation of sedimentary
dolomite, Science, Vol. 213, No. 4504, pp. 214-216.

-Baltzer, F., Kenig, F., Boichard, R., Plaziat, J.C., and Burser, B.H., 1994,
Organic matter distribution, water circulation and dolomitization beneath the Abu
Dhabi Sabkha (UAB), in Dolomites: A Volume in Honor of Dolomieu, Purser,
B.H., Tucker, ME, and Zenger, D.H., eds., Special Publication, International
Association of Sedimentologists, Vol. 21, pp. 409-427.

-Banfield, J.F., Welch, S.A., Zhang, H., Ebert, T.T., and Penn, R.L., 2000,
Aggregation-based crystal growth and microstructure development in natural iron
oxyhydroxide biomineralization products, Science, Vol. 289, pp. 751-754.

-Barber, DJ, 1977, Defect microstructures in deformed and recovered
dolomites, Tectonophysics, Vol. 39, pp. 193-213.

211

-Bathurst, R.G.C., 1975, Carbonate Sediments and Their Diagenesis,
Developments in Sedimentology, Vol. 12, Elsevier Scientific, New York.

-Battermann, B.W., 1957, Hillocks, pits, and etch rate in germanium crystals,
Journal of Applied Physics, Vol. 28, pp. 1236-1241.

-Baum, G.R., Harris, W.B., Drez, PE, 1985, Origin of dolomite in the Eocene
Castle Hayne Limestone, North Carolina, Journal of Sedimentary Research, Vol.
55, No.4, pp. 506-517.

-Becke, F., 1890, Tschemarks Min. Petr. Mitteil, Vol. 11, p. 349.

-Bemer, RA, 1978, Rate control on mineral dissolution under earth surface
conditions, American Journal of Science, Vol. 278, pp. 1235-1252.

-Berner, RA, 1981, Kinetics of weathering and diagenesis, In, Lasaga, AC. and
Kirkpatrick, R.J., eds., Kinetics of Geochemical Processes: MSA, Reviews in
Mineralogy, Vol. 8, pp. 111-134.

-Bemer, RA, and Morse, J.W., 1974, Dissolution kinetics of calcium carbonate
in sea water, IV. Theory of calcite dissolution, American Journal of Science, Vol.
274, pp. 108-134.

-Bemer, RA, and Holdren, GR, 1977, Mechanism of feldspar weathering:
Some observational evidence, Geology, Vol. 5, No. 6, pp. 369-372.

-Blake, D.F., Peacor, DR, and Wilkenson, B.H., 1982, The sequence and
mechanism of low-temperature dolomite formation: Calcian dolomite in a

Pennsylvanian echinoderrn, Journal of Sedimentary Petrology. Vol. 52, pp. 59-
70.

-Blum, A.E., Yund, RA, and Lasaga, AC, 1990, The effect of dislocation
density on the dissolution rate of quartz, Geochimica et Cosmochimica Acta, Vol.
54. pp.283-297.

-Bosbach, D. and Rammensee, W., 1994, In situ investigation of growth and
dissolution on the (010) surface of gypsum by Scanning Force Microscopy,
Geochimica et Cosmochimica Acta, Vol. 58, No. 2, pp. 843-849.

-Bosbach, D., Hall, C, and Putnis, A., 1998, Mineral precipitation and dissolution
in aqueous solution: in-situ microscopic observations on barite with atomic force
microscopy, Chemical Geology, Vol. 151, pp. 143-160.

-Bottcher, M.E., Mart, Y, and B, H.J., 1998, Data Report: Geochemistry of
Pliocene and Miocene carbonates from the Eratosthenes Seamount (Site 965),

Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 160, pp. 447-
451.

212

-Bragg. W.L. and Williams, E.J., 1934, The effect of thermal agitation on atomic
arrangements in alloys, Royal Society of London Proceedings, Vol. 145A, pp.
699-730.

-Brantley, S.L., Crane, S.R., Crerar, D.A., Hellmann, R., and Stallard, R., 1986,
Dissolution at etch pits in quartz, Geochimica et Cosmochimica Acta, Vol. 50, pp.
2349-2361.

-Budd, DA, 1997, Cenozoic dolomites of carbonate islands: their attributes and
origin, Earth Science Reviews, Vol. 42, pp. 1-47.

-Bullen, SB. and Sibley, DE, 1984, Dolomite selectivity and mimic replacement,
Geology, Vol. 12, pp. 655-658.

-Bums, S.J., McKenzie, J.A., and Vasconcelos, C., 2000, Dolomite formation and
biogeochemical cycles in the Phanerozoic, Sedimentology, Vol. 47, pp. 49-61.

-Burton, W.K., Cabrera, N., and Frank, 1951, Philosophical Transaction Royal
Society of London A, Vol. 243, pp. 299.

-Busenberg, E. and Plummer, L.N., 1982, The kinetics of dissolution of dolomite

in COzHZO systems at 1.5 to 65°C and 0 to 1 ATM PCOz, American Journal of
Science, Vol. 282, pp. 45-78.

-Busenberg, E. and Plummer, ML, 1989, Thermodynamics of magnesian calcite
solid-solutions at 25° C and 1 atm total pressure, Geochimica et Cosmochimica
Acta, Vol. 53, Issue 6, pp.1189-1208.

-Cabrera, N. and Levine, MM, 1956, Philosophy Magazine, Vol. 1, pp. 450.

-Cander, H.S., Kaufman, J., Daniels, L.D., and Meyers, W.J., 1988, Regional
dolomitization of shelf carbonates in the Burlington-Keokuk Formation
(Mississippian), Illinois and Missouri: Constraints from cathodoluminescent zonal

stratigraphy, In Sedimentology and Geochemistry of Dolostones, SEPM Special
Publication, No. 43, pp. 129-144.

-Carballo, J.D., Land, LS, and Miser, DE, 1987, Holocene dolomitization of
supratidal sediments by active tidal pumping, Sugarloaf Key, Florida, Journal of
Sedimentary Petrology, Vol. 57, pp. 153-165.

-Carpenter, AB, 1980, The Chemistry of dolomite formation l: The stability of
dolomite, in Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds. Concepts and
Models of Dolomitization: SEPM Special Publication, Vol. 28, pp. 111-121.

-Chai, L., and Navrotsky, A, 1996, Thermochemistry of carbonate-pyroxene
equilibria, Contributions to Mineralogy and Petrology. Vol. 114, pp. 139-147.

213

-Chai, L., Navrotsky, A., and Reeder, R.J., 1995, Energetics of calcium-rich
dolomite, Geochimica et Cosmochimica Acta, Vol. 59, pp. 939-944.

-Chang, L.L.Y., 1964, Synthesis of MBa(COa)2 compounds. American
Mineralogist, Vol. 49, pp. 1142-1143.

-Chilingar, G.V., 1956, Relationship between CalMg ratio and geological age,
American Association of Petroleum Geologists Bulletin, Vol. 40, pp. 2256-2266.

-Choquette, P.W. and Steinen, RP, 1980, Mississippian non-supratidal
dolomite, Ste. Genevieve Limestone, Illinois Basin: Evidence for mixed-water
dolomitization, in Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds.
Concepts and Models of Dolomitization: SEPM Special Publication, Vol. 28, pp.
163-196.

Chou, K., Garrels, RM. and Wollast, R., 1989, Comparative study of the kinetics
and mechanisms of dissolution of carbonate minerals, Chemical Geology, Vol.
78.pp.269-282.

-Clayton, R.N., Jones, RF. and Berner, RA, 1968, Isotope studies of dolomite

formation under sedimentary conditions, Geochimica et Cosmochimica Acta, Vol.
43. pp. 415-432.

-Cowell, D.W. and Ford, DC, 1980, Hydrochemistry of a dolomite karst: the
Bruce Peninsula of Ontario, Cananda. Journal of Earth Science, Vol. 17, pp.
520—526.

-Cu|lis, 0.6., 1904, Then mineralogical changes observed in the cores of the
Funafuti boring, in Bonney, T.G., ed., The atoll of Funafuti, Royal Society of
London, pp. 392-420.

-DeDekker, P. and Last, W.M., 1988, Modern dolomite deposition in continental,
saline lakes, western Victoria, Australia, Geology, Vol. 16, pp. 29-32.

-Deelman, J.C., 1981, Dolomite formation: Why thermodynamics failed, N. Jb.
Miner. Abh., Vol. 141, pp. 30-58.

-Deer, W.A., Howie, RA, and Zussman, J., 1992, An Introduction to the Rock-
Forrning Minerals, Longman.

-Deffeyes, K.S., Lucia, F.J., and Weyl, PK, 1965, Dolomitization of Recent and
Plio—Pleistocene sediments by marine evaporite waters on Bonaire, Netherlands
Antilles, In LC. Pray and RC. Murray, eds., Dolomitization and limestone
diagenesis, SEPM special publication, Vol. 13., pp. 71-88.

214

-De Yoreo, J.J., Land, T.A., Rashkovich, L.N., Onischenko, T.A., Lee, J.D.,
Monovskii, O.V., and Zaitseva, NP, 1997, The effect of dislocation cores on
growth hillock vicinality and normal growth rates of KDP {101} surfaces, Journal
of Crystal Growth, Vol. 182, pp. 442-460.

-Dove, PM. and Hochella Jr, M.F., 1993, Calcite precipitation mechanisms and
inhibition by orthophosphate: In situ observations by Scanning Force Microscopy,
Geochimica et Cosmochimica Acta, Vol. 57, pp. 705-714.

-Eggleston, CM, 1994, High-resolution scanning probe microscopy: tip-surface
interaction, artifacts and applications in mineralogy and geochemistry, In
Scanning Probe Microscopy of Clay Minerals, Eds. Nagy, KL and Blum, A.E.,
Clay Minerals Society, Boulder, CO, pp. 3-90.

-Folk, R.L., Chafetz, HS, and Tiezzi, PA, 1985, Bizarre forms of depositional
and diagenetic calcite in hot-springs travertines, central Italy, in Schneiderrnann,
N., and Harris, P.M., eds., Carbonate Cements, SEPM Special Publication Vol.
36, pp. 349-369.

-Folk, R.L., 1993, SEM imaging of bacteria and nanobacteria in carbonate
sediments and rocks, Journal of Sedimentary Petrology, Vol. 25 no. 6, pp. 990-
999.

-Folk, R.L. and Land, LS, 1975, Mg/Ca ratio and salinity: Two controls over
crystallization of dolomite, American Association of Petroleum Geologists
Bulletin, Vol. 59, pp. 60-68.

-Fouke, B.W. and Reeder, R.J., 1992, Surface structural controls on dolomite
composition: Evidence from sectoral zoning, Geochimica et Cosmochimica Acta,
Vol. 56. PP. 4015-4024.

-Frank, F .C., 1951, Capillary equilibria of dislocated crystals, Acta
Crystallographica, Vol. 4, pp. 497-501.

-Friedman, GM. and Sanders, J.E., 1967, Origin and occurrence of dolostones:
In G.V. Chilingar, H.J. Bissell, and R.W. Fairbridge, Eds., Carbonate Rocks, Part
A, Elsevier, New York, pp. 267-348.

-Frisia, S., 1994, Mechanisms of complete dolomitization in a carbonate shelf:
comparison between the Norian Dolomia Principale (Italy) and the Holocene of
Abu Dhabi Sabkha, Special Publication, Int. Ass. Sediment, Vol. 21, pp. 55-74.

-Fuchtbauer, H., 1972, Influence of salinity on carbonate rocks of the Zechstein
formation, northwest Germany, In: Geology of saline deposits, ed. G. Richter-
Bernburg, Unesco, Paris, pp. 23-31.

215

-Fuchtbauer, H., 1974, Sediments and Sedimentary Rocks, Halsted Press, New
York.

-Fuctbauer, H. and Goldschmidt, H., 1965, Beziehungen zwischen calciumgehalt
und bildungs-bedinggungen der dolomite, Geologische Rundschau, Vol. 55., pp.
29-40.

-Gaines, A.M., 1968, An experimental investigation of the kinetics and
mechanism of the formation of dolomite, PhD Dissertation, The University of
Chicago.

-Gaines, A.M., 1974, Protodolomite synthesis at 100°C and atmospheric
pressure, Science, Vol. 55, pp. 84-101,

-Gaines, A.M., 1980, Dolomitization kinetics: recent experimental studies, SEPM
Special Publication No. 28, pp. 81-86.

-Garcia Del Cura, M.A., Calvo, J.P., Ordonez, S., Jones, BF, and Canaveras,
J.C., 2001, Petrographic and geochemical evidence for the formation of primary,
bacterially induced lacustrian dolomite: La Roda “white earth” (Pliocene, central
Spain), Sedimentology, Vol. 48, pp. 897-915.

-Garrels, RM. and Thompson, ME, 1962, A chemical model for sea water at
25°C and one atmosphere total pressure, American Journal of Science, Vol. 260,
pp. 57-66.

-Garrels, R.M., Thompson, ME, and Siever, R., 1960, Stability of some
carbonates at 25°C and 1 atmosphere total pressure, American Journal of
Science, Vol. 258, pp. 402-418.

-Gilman, J.J., Knudson, C., and Walsh, WP, 1958, Journal of Applied Physics,
Vol. 29. PP. 601.

-Given, R.K. and Wilkenson, B.H., 1987, Dolomite abundance and stratigraphic
age: Constraints on rates and mechanisms of Phanerozoic dolostone formation,
Journal of Sedimentary Petrology, Vol. 57, No. 6, pp. 1068-1078.

-Goldsmith, JR. and Graf, D.L., 1958, Structural and compositional variations in
some natural dolomites, Journal of Geology, Vol. 66, pp. 678-693.

-Graf, D.L. and Goldsmith, JR, 1956, Some hydrothermal synthesis of dolomite
and protodolomite, Journal of Geology, Vol. 64, pp. 173-187.

-Graf, D.L., Blyth, CR, and Stemmler, RS, 1967, One dimensional disorder in
carbonates. Illinois State Geological Survey Circular, Vol. 408.

216

-Grandstaff, DE, 1978, Changes in surface area and morphology and the
mechanism of forsterite dissolution, Geochimica et Cosmochimica Acta, Vol. 42,
pp. 1899-1901.

-Gratz, A., Manne, S., and Hansma, PK, 1991, Atomic force microscopy of
atomic-scale ledges and etch pits formed during dissolution of quartz, Science,
Vol. 251, Issue 4999, pp. 1343-1346.

-Gratz, A.J., Hillner, PE, and Hansma, PK, 1993, Step dynamics and spiral
growth on calcite, Geochimica et Cosmochimica Acta, Vol. 57, pp. 491-495.

-Gregg, J.M., Howard, SA, and Mazzullo, S.J., 1992, Early diagenetic
recrystallization of Holocene (<3000 years old) peritidal dolomites, Ambergris
Cay, Belize, Sedimentology. Vol. 39, pp. 143-160.

- Gregg, J. M., Shelton, K. L., Johnson, A. W., Somerville, I. D., and Wright, W.
R., 2001, Dolomitization of the Waulsortian Limestone (Lower Carboniferous)
Irish Midlands: Sedimentology, Vol. 48, pp. 745-766.

-Hardie, LA, 1987, Dolomitization: a critical view of some current views, Journal
of Sedimentary Petrology, Vol. 57, pp. 166-183.

-Hartman, P., 1982, On the growth of dolomite and kaolinite crystals, N. Jb.
Miner. Mh., pp. 84-92.

-Heimann, RB, 1975, Auflosung von Kristallen - Theorie und Technische

Anwendung, Springer, New York, Russ. Translation, 1979, Rastvorenie Kristallov
(Nedra, Leningrad).

- Heimann, RB, 1982, Principles of chemical etching, in: Crystals - Growth,
Properties and Applications, ed. J. Grabmaier, Springer, Berlin, Vol. 8.

-Helgeson, H.C., Delany, J.M., Nesbitt, H.W., and Bird, OK, 1978 Summary and
critique of the thermodynamic properties of rock-forrning minerals, American
Journal of Science, Vol. 278-A, pp. 229.

-Heraty, L.J., 1988, ln-situ AFM examination of calcite nanomorphology in the
presence of know anionic growth inhibitors 8042' and PO43: Masters Thesis,
Michigan State University.

-Herman, JS. and White, W.B., 1985, Dissolution kinetics of dolomite: Effects of

lithology and fluid flow velocity, Geochimica et Cosmochimica Acta, Vol. 49, pp.
2017-2026.

-Hillner, P.E., Gratz, A.J., Marine, S., and Hansma, PK, 1992, Atomic-scale

imaging of calcite growth and dissolution in real time, Geology, Vol. 20, pp. 359-
362.

217

-Hochella, M.F., 1990, Atomic structure, microtopography, composition, and
reactivity of mineral surfaces, In. M.F. Hochella and AF. White eds., Mineral -
Water interface geochemistry, Reviews in Mineralogy, Vol. 23, pp. 87-128.

-Hochella, M.F. Jr., Eggleston, C.M., Elins, VB, and Thompson, MS, 1990,
Atomic structure and morphology of the albite {101} surface: An atomic-force

microscope and electron diffraction study, American Mineralogist, Vol. 75, pp.
723-730.

-Horita, J, Zimmennann, H, and Holland, H., 2002, Chemical evolution of
seawater during the Phanerozoic: Implications from the record of marine
evaporites, Geochimica et Cosmochimica Acta, Vol. 66, No. 21, pp. 3733-3756.

-Hom, PH, 1952, Screw dislocations, etch figures, and holes, Philosophical
Magazine, Vol. 43, pp. 1210-1213.

-Hsu, K., 1963, Solubility of dolomite and composition of Florida ground waters,
Journal of Hydrology, Vol. 1, pp. 288-310.

-Hull, D. and Bacon, DJ, 1984, Introduction to Dislocations, Pergamon Press,
Oxford, UK.

-Illing, L.V., Wells, A.J., and Taylor, J.C.M., 1965, Penecontemporary dolomite in
the Persian Gulf, In L.C. Pray and RC. Murray, eds., Dolomitization and
limestone diagenesis, SEPM special publication, Vol. 13., pp. 89-111.

-Jiang, X.N., Yuan, D.R., Sun, D.L., Lu, M.K., Zhang, G.H., and Guo, S.Y.,
2001a, Atomic Force Microscopy studies on growth mechanisms and defect
formations on (110) faces of Cadmium Mercury Thiocynate crystals, Crystal
Research Technology, Vol. 36, No. 6, pp. 601-608.

-Jiang X.N., Xu 0., Sun D.L., Yuan D.R., Lu M.K., Zhang G.H., and Fang Q,
2001b, Formation of small straight steps among large rough steps on the {110}
face of cadmium mercury thiocyanate crystal, Journal of Crystal Growth,
November 2001, Vol. 233, No. 1, pp. 318-325(8).

-Johnston, W.G., 1962, Dislocation etch pits in non-metallic crystals, In: Progress
In Ceramic Science Vol. 2, (Ed. J.E Burke).

-Jones, B., Luth, R.W., and MacNeil, A.J., 2001, Powder X-ray diffraction

analysis of homogeneous and heterogeneous sedimentary dolostones, Journal
of Sedimentary Research, Vol. 71, No. 5, pp. 790-799.

218

-Jones, B., 2004, Petrography and significance of zoned dolomite cements from
the Cayman Formation (Miocene) of Cayman Brac, British West Indies,
Journal of Sedimentary Research, Vol. 74, pp. 95-109.

-Joshi, M.S., Kotru, P.N., and lttyachen, MA, 1970, Kristallografiya, Vol. 15, pp.
103.

-Joshi, MS. and Paul, B.K., 1973, Mineralogical Magazine, Vol. 39, pp. 482.

-Katz, A., 1971, Zoned dolomite crystals, Journal of Geology, Vol. 79, pp. 38-51.
-Katz, A. and Matthews, A., 1977, The dolomitization of 0a003: an experimental
study at 252-295°C, Geochimica Cosmochimica Acta, Vol. 41, 297-308.

-Kazmierczak, T.F., Tomson, MB, and Nancollas, G.H., 1982, Crystal growth of
calcium carbonate. A controlled composition kinetic study, Journal of Physical
Chemistry, Vol. 86, pp. 103-107.

-Keith, RE, and Gilman, J.J., 1960, Dislocation etch pits and plastic deformation
in calcite, Acta Metallurgica, Vol. 8, pp. 1-10.

-Kessels, L.A., Sibley, DP, and Nordeng, S.H., 2000, Nanotopography of
synthetic and natural dolomite crystals, Sedimentology, Vol.47, pp.173-186.

-Kessels, LA, 2001, Interpreting crystal growth kinetics of ancient dolomites,
PhD Dissertation, Michigan State University.

-Kirkland, B.L., Lynch, F.L., Rahnis, M.A., Folk, R.L., Molineux, l.J., and McLean,
R.J.C., 1999, Alternative origins for nannobacteria-like objects in calcite,
Geology, Vol. 27, No.4, pp. 347-350.

-Kirkpatrick, R.J., 1981, Kinetics of crystallization of igneous rocks, in Eds. A.C.
Lasaga and R.J. Kirkpatrick, Kinetics of Geochemical Processes, Reviews in
Mineralogy, Vol. 8, Chapter 8, pp. 321-398

-Klapper, H., 1980, Defects in non-metal crystals. In: Characterization of Crystal
Growth Defects by X-Ray Methods, eds., B.K. Tanner, D.K. Bowen, Plenum
Press, new York, pp. 133-160.

- Krauskopf, KB. and Bird, OK, 1995, Introduction to Geochemistry, McGraw-
Hill, New York.

-Land, LS, 1967, Diagenesis of skeletal carbonates, Journal of Sedimentary
Petrology, Vol. 37, pp. 914-930.

219

-Land, LS, 1980, The isotopic and trace element geochemistry of dolomite: the
state of the art, SEPM Special Publication No. 28, Eds., D.L. Zenger, J.B.
Dunham, and R.L. Ethington, pp. 87-110.

-Land, LS, 1985, The origin of massive dolomite, Journal of Geological
Education, Vol. 33, 112-125.

-Land, L. S. and Macpherson, G. L., 1992, Origin of Saline Formation Waters,
Cenozoic Section, Gulf of Mexico Sedimentary Basin: AAPG Bulletin, Vol. 76, pp.
1344—1362.

-Land, LS, 1998, Failure to precipitate dolomite at 25°C from dilute solution
despite 1000-fold oversaturation after 32 years, Aquatic Geochemistry, Vol. 4,
pp. 361-368.

-Langbein, R. von, Landgraf, K.F., and Milbrodt, E., 1984, Calciumuberschusse
im Dolomit ais indicator des sedimentationsmilieus in deonischen
Karbonatgesteinen: Chemie der Erde, Vol. 43, pp. 217-227.

-Langmuir, D., 1997, Aqueous Environmental Geochemistry, Prentice Hall, New
Jersey, pp. 193-230.

-Lasaga, AC, and Blum, A.E., 1986, Surface chemistry, etch pits and mineral-
water reactions, Geochimica et Cosmochimica Acta, Vol. 50, pp. 2363-2379.

-Lasaga, AC, 1990, Atomic treatment of mineral -water surface reactions, In:
Hochella Jr, M.F. and White, A.F., eds. Mineral-Water Interface Geochemistry,
Reviews in Mineralogy, Vol. 23, pp. 17-85.

-Liang, Y., Baer, D.R., McCoy, J.M., Amonette, J.E., and LaFemina, JP, 1996,
Dissolution kinetics at the calcite-water interface, Geochimica et Cosmochimica
Acta, Vol. 60, pp. 4883-4887.

-Lin, CC. and Sheri, P., 1993, Role of screw axis in dissolution of willemite,
Geochimica et Cosmochimica Acta, Vol. 57, pp. 1649-1655.

-Lin, CC. and Sheri, P., 1995, Incubation time of etch pits at dislocation
outcrops, Geochimica et Cosmochimica Acta, Vol. 59, No. 14, pp. 2955-2963.

-Lippmann, F ., 1967, Die syntheses des norsethite bei 20°C und 1 at. Ein Modell
zur Dolmitiisierung, Neues Jahrb. Mineral. Monatash, pp. 23-29

-Lippmann, F., 1973, Sedimentary Carbonate Minerals, Springer-Venag, Berlin.

220

-Lorens, RB, 1981, Sr, Cd, Mn, and Co distribution coefficients in calcite as a
function of calcite precipitation rate, Geochimica et Cosmochimica Acta, Vol. 45,
pp. 553-561. '

-Lowenstein, T.K., Timofeeff, M.N., Brennan, S.T., Hardie, LA, and Demicco,
R.V., 2001, Oscillations in Phanerozoic seawater chemistry: evidence from fluid
inclusions. Science, Vol. 294, pp. 1086-1088.

-Lumsden, D.N., 1979, Discrepancy between thin sections and X-ray estimates
of dolomite in limestone, Journal of Sedimentary Petrology, Vol. 49, pp. 429-436.

-Lumsden, D.N., 1988, Characteristics of deep marine dolomite, Journal of
Sedimentary Petrology, Vol. 58, pp. 1023-1031.

-Lumsden, D.N. and Chimahusky, J.S., 1980, Relationship between dolomite
nonstoichiometry and carbonate facies parameters, In: Concepts and Models of
Dolomitization (Eds. D.H Zenger, J.B. Dunham and R.L. Ethington), Special
Publication Society of Economic Paleontologists and Mineralogists, Vol. 28, pp.
11 1-121.

-Luttge, A., Winkler, U., and Lasaga, AC, 2003, Interferometric study of the
dolomite dissolution: A new conceptual model for mineral dissolution,
Geochimica et Cosmochimica Act, Vol. 67, pp. 1099-1 116.

-Machel, H. , 1997, Recrystallization versus neomorphism, and the concept of
‘significant recrystallization’ in dolomite research, Sedimentary Geology, Vol.
113, pp. 161-168.

-Maclnnes, IN. and Brantley, S.L., 1992, The role of dislocations and surface
morphology in calcite dissolution, Geochimica et Cosmochimica Acta, Vol. 56,
pp. 1113-1126.

-Maclnnes, IN. and Brantley, S.L., 1993, Development of etch pit size
distributions on dissolving minerals, Chemical Geology, Vol. 105, pp. 31-49.

Mackenzie, F.T., Bischoff, W.D., Bishop, F.C., Loijens, M., Schoonmaker, J. and
Wollast, R., 1983, Magnesian calcites: low temperature occurrence, solubility and
solid-solution behavior, in Carbonate: Mineralogy and Chemistry, Reviews in

Mineralogy. ed. Reeder, R.J., Minerals Society of America, Vol. 11, Chapter 4,
pp. 97-144.

-Maiwa, K., Plomp, M., van Enckevort, W.J.P., and Bennema, P., 1998, AFM

observation of barium nitrate (111) and (100) faces: spiral growth and two
dimensional nucleation growth, Journal of Crystal Growth, Vol. 186, pp. 214-223.

221

-Malone, M.J., Baker, PA, and Burns, S.J., 1994, Recrystallization of dolomite:
evidence from the Monterey Formation (Miocene), California, Sedimentology,
Vol. 41. pp. 1223-1239.

-Malone, M.J., Baker, PA, and Burns, S.J., 1996, Recrystallization of dolomite:
An experimental study from 50-200°C, Geochimica et Cosmochimica Acta, Vol.
60, No. 12. Pp. 2189-2207.

Marschner, H., 1968, Ca-Mg distribution in carbonates from the Lower Keuper in
NW Germany, in Muller, G and Friedman, G.M., eds., Recent Developments in
Carbonate Sedimentology in Central Europe, Springer-Verlag, New York, pp.
128-135.

-Marsh, B.D., 1988, Crystal size distribution (CSD) in rocks and the kinetics and

dynamics of crystallization: I. Theory, Contributions to Mineralogy and Petrology,
Vol. 99, pp. 277-291.

-Mazzullo, S.J., 1992, Geochemical and neomorphic alteration of dolomite: A
review, Carbonates and Evaporites, Vol. 7, pp. 21-37.

-McClay, KR, 1977, Pressure solution and cobble creep in rocks and minerals:
a review, Journal of the Geological Society of London, Vol. 134, pp. 57-70.

- McKenzie, J.A. 1981, Holocene dolomitization of calcium carbonate sediments
from the coastal sabkhas of Abu Dhabi, U.A.E.: A stable isotope study: Journal of
Geology. Vol. 89, pp. 185—198.

-Meike, A., 1990a, A micromechanical perspective on the role of dislocations in
selective dissolution, Geochimica et Cosmochimica Acta, Vol. 54. PP. 3347-3352.

-Meike, A, 1990b, Considerations for quantitative determination of the role of
dislocations in selective dissolution, Earth-Science Reviews, Vol. 29, pp. 309-
320.

-Miser, D.E., Swinnea, J.S., and Steinfink, H., 1987, TEM observations and X-ray
crystal structure refinement of a twinned dolomite with a modulated
microstructure, American Mineralogist, Vol. 72, pp. 188-193.

-Morrow, D.W. and Rickets, B.D., 1988, Experimental investigation of sulfate
inhibition of dolomite and its mineral analogues, In: Shukla, V., and Baker, P.A.,
eds., Sedimentology and Geochemistry of Dolostones: SEPM Special
Publication, Vol. 43, pp. 25-38.

-Morrow, D.W., Gorham, BL, and Wong, J.N.Y., 1994, Dolomite-calcite

equilibrium at 220 to 240°C at saturation vapour pressure: Experimental data,
Geochimica et Cosmochimica Acta, Vol. 58, pp. 169-177.

222

-Morse, J.W., and Arvidson, RS, 2002, The dissolution kinetics of major
sedimentary carbonate minerals, Earth Science Reviews, Vol. 58, No. 1-2, pp.
51-84.

-Murphy, W.M., 1989, Dislocation and feldspar dissolution, European Journal of
Mineralogy, Vol. 1, pp. 315-326.

-Murray, RC, 1964, Preservation of primary structures and fabrics in dolomite, in
Imbrie, J. and Newell, N., eds., Approaches to paleoecology: New York, John
Wiley, pp. 388-403.

-Murray, RC. and Lucia, F.J., 1967, Cause and control of dolomite distribution by
rock selectivity, Geological Society of America Bulletin, Vol. 78, pp. 21-35.

-Murray, RC. and Pray, LC, 1965, Dolomitization and limestone diagenesis-An
introduction: In L.C. Pray and RC. Murray, Eds., SEPM Special Publication, Vol.
13, Pp. 1-2.

-Naumov, G.B., Ryzhenko, B.N., and Khodakovsky, LL, 1974, in Barnes, I. and
Speltz, V., eds., Handbook of thermodynamic data, USGS Water Research
Reports 74-001, pp. 239. '

-Navrotsky, A. and Capobianco, C., 1987, Enthalpies of formation of dolomite
and magnesian calcite, American Mineralogist, Vol. 79, pp. 782-787.

-Navrotsky, A., Dooley, D., Reeder, R., and Brady, P., 1999, Calorimetric studies
on the energetics of order-disorder in the system M91.xFexCa(C03)2, American
Mineralogist, Vol. 84, pp. 1622-1626.

-Navrotsky, A. and Loucks, D., 1977, Calculation of subsolidus phase relations in

carbonates and pyroxenes, Physics and Chemistry of Minerals, Vol. 1, pp. 109-
127.

-Nielsen, A.E., 1964, Kinetics of Precipitation, Macmillan, New York.

-Nordeng, SH. and Sibley, OF, 1994, Dolomite stoichiometry and Ostwald’s
Step Rule, Geochimica et Cosmochimica Acta, Vol. 58, pp. 191-196.

-Nordstrom, D.K., Plummer, L.N., Langmuir, D., Busenberg, E., May, H.M.,
Jones, BF, and Parkhurst, D.L., 1990, Revised chemical equilibrium data for
major water-mineral reactions and their limitations, In eds. Melchior, DC, and

Bassett, R.L., Chemical modeling of aqueous systems II, American Chemical
Society, pp. 398-413.

-Ohara, M. and Reid, RC, 1973, Modeling Crystal Growth Rates From Solution.
Prentice Hall, Engelwood Cliffs, NJ, New York.

223

-Ohnesorge, F. and Binnig, G., 1993, True atomic resolution by atomic force
microscopy through repulsive and attractive forces, Science, Vol. 260, pp. 1451-
1456.

-Paquette, J., Hojatollah, V., and Mountjoy, E., 1999, Novel TEM approaches to
imaging of microstructures in carbonates: Clues to growth mechanisms in calcite
and dolomite, American Mineralogist, Vol. 84, pp. 1939-1949.

-Paquette, J. and Reeder, R.J., 1990, New type of compositional zoning in
calcite: Insight into crystal-growth mechanisms, Geology, Vol. 18, pp. 1244-1247.

-Paquette, J. and Reeder, R.J., 1995, Relationship between surface structure,
growth mechanism, and trace element incorporation in calcite, Geochimica et
Cosmochimica Acta, Vol. 59, No.4, pp. 735-749.

-Pande, DR. and Vadrabade, SR, 1990, Etch pits on basal cleavage faces of
apophyllite crystals, Mineralogical Magazine, Vol. 54, pp. 559-565.

-Patel, A.R., Bahl, GP, and Vagh, AS, 1965, Acta Crystallographica, Vol. 19,
pp. 757 and 1025.

-Patterson, R.J. and Kinsman, D.J.J., 1982, Formation of diagenetic dolomite in
coastal sabkha along the Arabian (Persian) Gulf, AAPG Bulletin, Vol. 66, No. 1,
pp. 28-43.

- Penn, R.L. and Banfield, J.F., 1998, Imperfect oriented attachment: dislocation
generate in defect-free nanocrystals, Science, Vol. 281, pp. 969-971.

-Perkins, R.D., Dwyer, G.S., Rosoff, D.B., Fuller, J., Baker, PA, and Lloyd, RM,
1994, Salina sedimentation and diagenesis: West Caicos, British West Indies,

International Association of Sedimentologists, Special Publication, Vol. 21, pp.
37-54.

-Pierre, 0., Ortlieb, L., and Person, A., 1984, Supratidal evaporitic dolomite at
Ojo de Liebre lagoon: mineralogical and isotope arguments for primary
crystallization, Journal of Sedimentary Petrology, Vol. 54, pp. 1049-1061.

-Pina, C.M., Becker, U., Risthaus, P., Bosbach, D., and Putnis, A., 1998,

Molecular-scale mechanisms of crystal growth in barite, Nature, Vol. 395, pp.
483-486.

-Pina C.M., Enders M., and Putnis A., 2000, The composition of solid solutions
crystallizing from aqueous solutions: the influence of supersaturation and growth
mechanisms, Chem. Geol., Vol. 168, pp. 195-210.

224

-Pina, C.M., Putnis, A., and Astilleros, J.M., 2004, The growth mechanisms of
solid solutions crystallizing from aqueous solutions, Chemical Geology, Vol. 204,
pp. 145-161.

Plummer, L.N., Parkhurst, D.L. and Wigley, T.M.L., 1979, Critical review of the
kinetics of calcite dissolution and precipitation. In: Jenne, E. Editor, 1979.
Chemical Modelling — Speciation, Sorption and Kinetics in Aqueous Systems
American Chemical Society, Washington, DC, pp. 537—573.

-Porsche, J., Ruf, A, Geiger, M., and Scholz, P., Size control of self-assembled
InP/GaP quantum islands, Journal of Crystal Growth, Vol. 195, pp. 591-595.

-Purser, B., Tucker, M., and Zenger, D, 1994, Problems, progress and future
research concerning dolomites and dolomitization, in Dolomites: A Volume in
Honor of Dolomieu, Special Publication of the lntemational Association of
Sedimentologists, No. 21 ., pp. 3-20.

-Radke, EM. and Mathis, R.L., 1980, On the formation and occurrence of saddle
dolomite, Journal of Sedimentary Petrology, Vol. 50, pp. 1149-1168.

-Randazzo, AF. and Cook, D.J., 1987, Characterization of dolomite rocks from
coastal mixing zone of the floridian aquifer, USA, Sedimentary Geology, Vol.
-Reeder, R.J., 1981, Electron optical investigation of sedimentary dolomites,
Contributions to Mineralogy and Petrology, Vol. 76, pp. 148-157.

-Reeder, R.J., 1982, Crystal growth defects in sedimentary carbonate minerals,
Estudios Geologica, Vol. 38, pp. 179-183.

-Reeder, R.J., 1983, Crystal chemistry of the rhombohedral carbonates. In
Reeder, R.J. Ed., Carbonates: Mineralogy and Chemistry, Reviews in
Mineralogy, Vol. 11, pp. 1-47.

-Reeder, R.J., 1992, Carbonates: Growth and alteration microstructures. In.
Minerals and reactions at the atomic scale; transmission electron microscopy,
Reviews in MineraIOQY. Vol. 27, pp. 380-424.

-Reeder, R.J., 2000, Constraints on cation order in calcium-rich sedimentary
dolomites, Aquatic Geochemistry, Vol. 6, pp. 213-226.

-Reeder, R.J. and Prosky J.L., 1986, Compositional sector zoning in dolomite,
Journal of Sedimentary Petrology, Vol. 56, pp. 237-247.

-Reeder, R.J. and Sheppard, CE, 1984, Variation of lattice parameters in some
sedimentary dolomites, American Mineralogist, Vol. 69, pp. 520-527.

225

-Reeder, R.J. and Wenk, HR, 1979, Microstructures in low temperature
dolomites. Geophysical Review Letter, Vol. 6, pp. 77-80.

-Richter, OK, 1974, Entstehung und diagenese der Devonischen und
Perrnotriassischen Dolomite in der Eifel: In Contributions to Sedimentology, Vol.
2, Schweizerbartsche Verlagsbuchhandlung, Stuttgart, pp. 101.

-Robie, R.A., Hemingway, 3.8., and Fisher, JR, 1978, Thermodynamic
properties of minerals and related substances at 298.15K and 1 bar pressure and
at higher temperatures, US. Geological Survey Bulletin, Vol. 1452, pp. 456.
-Rock, PA, Manell, G.K., Casey, WA, and Walling, EM, 2001, Gibbs energy
of formation of dolomite from electrochemical cell measurements and theoretical
calculations, American Journal of Science, Vol. 301, pp. 103-111.

-Rosen, M.R., Miser, DE, and Warren, J.K., 1988, Sedimentology. mineralogy,
and isotopic analysis of Pellet Lake, Coorong region, South Australia,
Sedimentology. Vol. 35, pp. 105-122.

-Rosen, M.R., Miser, D.E., Starcher, MA, and Warren, J.K., 1989, Formation of
dolomite in the Coorong region, South Australia, Geochimica et Cosmochimica
Acta, Vol. 53. pp. 661-669.

-Rosenberg, PE. and Holland, H.D., 1964, Calcite-dolomite-magnesite stability
relations in solutions at elevated temperatures, Science, Vol. 145, pp. 700-701.

-Royce, C.R., Wadell Jr., J.S., and Petersen, LE, 1971, X-ray determination of

calcite-dolomite: an evaluation, Journal of Sedimentary Petrology. Vol. 41, pp.
483-488.

-Rozgonyi, G.A., Mahajan, 8., Read, M.H., and Brasen, D., 1976, Sources of

oxidation-induced stacking faults in Czochralski silicon wafers, Applied Physics
Letters, Vol. 29, pp. 531-533.

-Sangwal, K., 1987a, Etching of Crystals: Theory, Experiment, and Application,
North-Holland, Oxford.

-Sangwal, K., 1987b, Etching processes and the investigation of minerals, In:
Geochemistry and Mineral Formation in the Earth Surface, eds. R. Rodriguez-
Clemente and Y. Tardy, Proceedings of the international meeting, Geochemistry
of the earth surface and processes of mineral formation.

-Sass, E. and Bein, A., 1988, Dolomites and salinity: a comparative geochemical
study, in Shukla, V. and Baker, P.A., eds., Sedimentology and Geochemistry of
Dolostones: SEPM Special Publication, Vol. 43, pp. 223-233.

226

-Sass, E. and Katz, A., 1982, The origin of platform dolomites: new evidence,
American Journal of Science, Vol. 282, pp. 1184-1213.

- Sayles, FL and Fyfe, W.S., 1973, The crystallization of magnesite from
aqueous solution, Geochimica et Cosmochimica Acta, Vol. 37, pp. 87-99

-Schmidt, V., 1965, Facies, diagenesis, and related reservoir properties in the
Gigas Beds (Upper Jurassic), northwester Germany: In, Prey, LC. and Murray,
R.C., eds., Dolomitization and Limestone Diagenesis: SEPM, Special
Publication, Vol. 13, pp. 124-169.

-Schubel, K.A., Elbert, DC, and Veblen, DR, 2000, lncommensurate c-domain
superstructure in calcian dolomite from the Latemar buildup, Dolomites, Northern
Italy, American Mineralogist, Vol. 85, pp 858-862.

-Shinn, E., Ginsburg, RN, and Lloyd, RM, 1965, Recent supratidal dolomite
from Andros Island, In L.C. Pray and RC. Murray, eds., Dolomitization and
limestone diagenesis, SEPM special publication, Vol. 13., pp. 112-123.

-Sibley, D.F., 1982, The origin of common dolomite fabrics, Journal of
Sedimentary Petrology. Vol. 50, pp. 1149-1168.

-Sibley, D.F. and Bartlett, TR, 1987, Nucleation as a rate limiting step in
dolomitization, International Symposium on Crystal growth Processes in the
Sedimentary Environment, 2 , Granada, 1986, Madrid, Institute de Geologica,
C.S.I.C, pp. 733-741.

Sibley, D.F., Dedoes, RE, and Bartlett, TR, 1987, Kinetics of dolomitization,
Geology. Vol. 15, 1112-1114.

-Sibley, DP. and Gregg, J.M., 1987, Classification of dolomite rock texture,
Journal of Sedimentary Petrology, Vol. 57, pp. 967-975.

-Sibley, D.F., Gregg, J.M., Brown, R.G., and Laudon, PR, 1993, Dolomite
crystal size distribution, in Rezak, R., and Lavoie, D., eds., Cabonate
Microfabrics: New York, Springer-Verlag, pp. 195-208.

-Sibley, D.F., Nordeng, S.H., and Borkowski, ML, 1994, Dolomitization kinetics
in hydrothermal bombs and natural settings, Journal of Sedimentary Research,
Vol. A64, No. 3. Pp. 630-637.

-Sibley, D.F., 1990, Unstable to stable transformations during dolomitization,
Journal of Geology, Vol. 98, 739-748.

227

-Sperber, C.M., Wilkinson, B.H., and Peacor, DR, 1984, Rock composition,
dolomite stoichiometry, and rock/water reactions in dolomitic carbonate rocks,
The Journal of Geology, Vol. 92, No. 6, pp. 609-622.

-Stadler, H.L., 1963, Etched hillocks in BaTi03, Journal of Applied Physics, Vol.
34. PP. 570-573.

-Stipp, S.L.S., Eggleston, C.M., and Nielsen, 3.8., 1994, Calcite surface structure
observed at microtOpographic and molecular scales with atomic force microscopy
(AFM), Geochimica et Cosmochimica Acta, Vol. 58, No. 14, pp. 3023-3033.

-Sun, 8.0., 1994, A reappraisal of dolomite abundances and occurrence in the
Phanerozoic, Journal of Sedimentary Research, Section A 64, No. 2, pp. 396-
404.

-Sunagawa, I., 1977, Natural crystallization, Journal of Crystal Growth, Vol. 42,
pp. 214-223.

Sunagawa, |., 1981, Characteristics of crystal growth in nature as seen from the
morphology of mineral crystals, Bulletin Mineralogie, Vol. 104, pp. 81-87.

-Sunagawa, I., 1984, Growth of crystals in nature. In: Materials Science of the
Earth’s Interior, Ed. I. Sunagawa, Terra Scientific Publishing, Tokyo, pp. 63-105.

-Swart, PK. and Melim, LA, 2000, The origin of dolomite in Tertiary sediments
from the Great Bahama Bank, Journal of Sedimentary Research, Vol. 70, pp.
738-748.

-Tanger, J.C. IV, and Helgeson, HQ, 1988, Calculation of the thermodynamic
and transport properties of aqueous species at high pressures and temperatures,
Revised equations of state for the standard partial molal properties of ions and
electrolytes, American Journal of Science, Vol. 288, pp. 19-98.

-Tuck, B. and Baker, A.J., 1973, Chemical etching of{111} and {100} surfaces of
InP, Journal of Materials Science, Vol. 8, pp. 1559-1566.

-Tuck, B., 1975, The chemical polishing of semiconductors, Journal of Materials
Science, Vol. 10, pp. 321-339.

Usdowski, E., 1994, Synthesis of dolomite and geochemical implications, In
Dolomites: A Volume in Honor of Dolomieu, Eds. B.H. Purser, M.E. Tucker, and
D.L. Zenger, lntemational Association of Sedimentology, Special Publication,
Vol. 21, pp. 345-360.

228

-Van Tendello, G., Wenk, HR, and Gronsky, R., 1985, Modulated structures in
calcian dolomite: A study by electron microscopy, Physics and Chemistry of
Minerals, Vol. 12, pp. 333-341.

-Vasconcelos, C, McKenzie, J.A., Bemasconi, S., Grujic, D, and Tien, A., 1995,
Microbial mediation as a possible mechanism for dolomite formation at low
temperature, Nature, Vol. 377, pp. 220-222.

-Vasconcelos, C. and McKenzie, J.A., 1997, Microbial mediation of modern
dolomite precipitation and diagenesis under anoxic conditions (Lagoa Verrnelha,
Rio De Janeiro, Brazil), Journal of Sedimentary Research, Vol. 67, No. 3, pp.
378-390.

-Vennegues, P., Beaumont, B., Haffouz, S., Vaille, M., and Gibart, P., 1998,
Influence of in situ sapphire surface preparation and carrier gas on the growth
mode of GaN in MOVPE, Journal of Crystal Growth, Vol. 187, pp. 167-177.

-Vogel, F.L., Pfann, W.G., Corey, H.E., Thomas, E.E., 1953, Physics Review,
Vol. 90. PP. 489.

- Von Der Borch, CC. and Jones, J.B., 1976, Spherular modern dolomite from
the Coorong area, South Australia, Sedimentology, Vol. 23, pp. 587-591.

-Wagman, D.D., Evans, W.H., Parker, V.B., Schumm, R.H., Halow, I., Bailey,
S.M., Churney, KL, and Nuttall, R.L., 1982, The NBS tables of chemical
thermodynamic properties: selected values for inorganic and C1 and C; organic
substances in SI units, Journal of Physical and Chemical Reference Data, Vol.
11 (Supplement No. 2), pp. 392.

-Walton, AG, 1969, Nucleation in Liquids and Solutions, Zettlemoyer, AC
Nucleation, New York.

-Warthman, R., van Lith, Y, Vasconcelos, C., McKenzie, J.A., and Karpoff, A.M.,
2000, Bacterially induced dolomite precipitation in anoxic culture experiments,
Geology, Vol. 28, No. 12, pp. 1091-1094.

-Wenk, H.R., Barber, D.J., Reeder, R.J., 1983, Microstructures in carbonates. In
Carbonates: Mineralogy and Chemistry (ed. R.J. Reeder); Reviews in Mineralogy
Vol. 11, pp. 301-367.

-Wenk, HR, HU, MS, and Frisia, S., 1993, Partially ordered dolomite-
Microstructural characterization of Abu-Dhabi sabkha carbonates, American
Mineralogist, Vol. 78, pp. 769-774

229

-Weyher, J. and Van Enckevort, W.J.P., 1983, Selective etching and
photoetching of {100} in gallium arsenide CrOa-HF aqueous solutions. II., The
Nature of etch hillocks, Journal of Crystal Growth, Vol. 63, pp. 292-298.

-Wilkenson, B.H. and Algeo, T.J., 1989, Sedimentary carbonate record of
calcium magnesium cycling, American Journal of Science, Vol. 289, pp. 1158-
1 194.

-Zbik, M. and Smart, R. St 0., 1998, Nanomorphology of kaolinites: comparative
SEM and AFM studies. Clays and Clay Minerals, Vol. 46, pp. 153-160.

-Zenger, DH, 1996, Dolomitization of the “0’ zone, Red River Formation (Upper

Ordovician) in a deep core, Williston basin, Richland County, eastern Montana,
Contributions to Geology, University of Wyoming, Vol. 31, No. 1, pp. 57-75.

230

   

lIllllllljlllIljlllljljllllljlllll