Journal of Sedimentary Research, 2007, v. 77, 424–432 Research Article DOI: 10.2110/jsr.2007.035
A COMPARISON OF NANOMETER-SCALE GROWTH AND DISSOLUTION FEATURES ON NATURAL AND SYNTHETIC DOLOMITE CRYSTALS: IMPLICATIONS FOR THE ORIGIN OF DOLOMITE STEPHEN E. KACZMAREK AND DUNCAN F. SIBLEY Michigan State University, Department of Geological Sciences, East Lansing, Michigan 48824, U.S.A. e-mail:
[email protected]
ABSTRACT: This paper describes experiments in which mechanisms of crystal growth are inferred from the surface nanotopography of natural and synthetic dolomite using atomic force and scanning electron microscopy. Synthetic dolomite was formed by replacement of calcite at 218˚ C in Mg–Ca–Cl solutions. The first products to form during dolomite synthesis experiments, as detected with X-ray diffraction (XRD), are invariably poorly ordered (nonideal) dolomite. As the reaction proceeds, the degree of cation order increases, yet the products remain relatively disordered. Upon examination, the surfaces of nonideal dolomite are covered with mounds 10–200 nm wide and 1–20 nm high. Following depletion of the calcite reactant, dolomite becomes relatively well ordered and stoichiometric (ideal). The surfaces of ideal dolomite are covered with broad, flat layers separated by steps that measure tenths to tens of nanometers high. Comparison of natural and synthetic dolomite surfaces after etching in 0.5% H2SO4 for 10–20 seconds indicates that high-temperature synthetic and low-temperature natural nonideal dolomite surfaces are covered with mounds identical to the growth features found on unetched nonideal synthetic dolomite. In contrast, synthetic and low-temperature ideal dolomite surfaces, when etched, are characterized by flat layers with deep, euhedral etch pits. These results suggest that despite the wide range in formation conditions, natural and synthetic dolomites form by the same growth mechanisms. Furthermore, the two different surface nanotopographies described here are consistent with a model in which nonideal dolomite forms by precipitation of amorphous mounds that quickly crystallize to nonideal dolomite while ideal dolomite later replaces nonideal dolomite and grows by spiral growth.
INTRODUCTION
Dolomite, CaMg(CO3)2, is a common mineral in ancient rocks and the thermodynamically stable carbonate phase in modern seawater, yet it is rare in modern marine environments. Why this is so has remained the subject of scientific inquiry of over 200 years. There is very little agreement concerning the details of dolomite formation except that most natural dolomites form at Earth-surface temperatures and pressures (Krauskopf and Bird 1995). Despite such consensus it has been extremely difficult to synthesize dolomite abiotically at temperatures below 100uC, even over many years (Land 1980, 1998). High-temperature laboratory experiments have provided some insight into the nature of dolomite formation (Katz and Matthews 1977; Baker and Kastner 1981; Sibley et al. 1987; Sibley 1990; Malone et al. 1996; Kelleher and Redfern 2002), but there exists no theoretical basis for extrapolating kinetic data over the 25– 250uC temperature range between sedimentary dolomites and temperatures commonly employed in the laboratory. As a result, the applicability of high-temperature experiments to natural low-temperature systems remains debated (Arvidson and Mackenzie 1999). An understanding of dolomite formation has been further complicated by the observation that many ancient dolomites are stoichiometric and well ordered compared to their modern counterparts, which often contain excess calcium, are poorly ordered, and display heterogeneous microstructures when observed with transmission electron microscopy (see Reeder 1992, 2000 for reviews). Similar deviations from ideality have been observed in early dolomite-like phases formed during high-temperature synthesis experiments (Graf and Goldsmith 1956; Gaines 1974; Sibley
Copyright E 2007, SEPM (Society for Sedimentary Geology)
1990; Malone et al. 1996; Politi et al. 2004), the inference being that nonideal dolomite represents a metastable precursor to ideal dolomite (Nordeng and Sibley 1994). Recently, a number of studies have documented the occurrence of amorphous CaMg-carbonate precursors in highly supersaturated solutions with Mg:Ca ratios . 2 that later crystallize to Mg-calcite (Raz et al. 2000; Kelleher and Redfern 2002; Loste et al. 2003; Schmidt et al. 2005; Kwak et al. 2005; DiMasi et al. 2006). Kelleher and Redfern (2002) reported forming a hydrous calcium magnesium carbonate phase in precipitation experiments that was compositionally and structurally identical to recent dolomites from Abu Dhabi based on XRD and infrared spectroscopy. They argued that a hydrous precursor phase could circumvent kinetic barriers associated with dehydration of magnesium ions in solution. Schmidt et al. (2005) showed that two amorphous phases were produced in laboratory growth experiments, a hydrated CaMgcarbonate and an amorphous calcium carbonate. They found that the hydrated CaMg-carbonate had similar characteristics as disordered dolomite from the Coorong region in South Australia. In all cases, amorphous precipitates are characterized by structural disorder and rounded morphologies with a high degree of surface roughness. Micrometer- to nanometer-scale surface topographies on synthetic and mildly etched natural crystals have long been used to provide indirect evidence of crystal growth mechanisms (see Sunagawa 1984 for review). More recently, direct observation of crystal growth in a fluid cells using atomic force microscopy has provided direct evidence that crystals grow by spiral growth at low degrees of supersaturation and polynuclear growth at
1527-1404/07/077-424/$03.00
JSR
GROWTH AND DISSOLUTION FEATURES ON NATURAL AND SYNTHETIC DOLOMITES
425
TABLE 1.— Sample list and corresponding analytical data for natural dolomites. Sample ID MH-2 MH-6 MH-10 MH-11 MH-20 MH-52 SAL 7-5 SAL 7-11 7-9-99-3 SB 42 SC 10D SB 35 RR 17 RR 19 PDC 1 L2 91 80-8-TS 80-8-17-1 Aruba Dol 8-9-84-5 8-9-84-3 13-4 16-8 S-1 8-23-86-7 DC37 Niagara GWP 280-290 GWP287 FC 2C 7-22-02-3
Age
Formation
d(104) Position
Mole% CaCO3
Ordering Ratio
Etched Topography
Permian Permian Permian Permian Permian Permian Ordovician Ordovician Ordovician Pliocene Pliocene Pliocene Ordovician Ordovician Ordovician Ordovician Pliocene Pliocene Pliocene Miocene Miocene Ordovician Ordovician Silurian Silurian Mississippian Silurian Holocene Holocene Mississippian Pliocene
Hueco Hueco Hueco Hueco Hueco Hueco Saluda Saluda Saluda Seroe Domi Seroe Domi Seroe Domi Prarie DuChen Prarie DuChen Prarie DuChen Prarie DuChen Seroe Domi Seroe Domi Seroe Domi, Terminal Carb. Complex, Spain Terminal Carb. Complex, Spain Galena Galena Lockport Lockport Burlington Keokuk Niagara Abu Dhabi Sabkha Abu Dhabi Sabkha Borden Seroe Domi
31.00 31.03 31.02 31.02 31.03 31.03 30.84 30.87 30.85 30.91 30.87 30.87 31.00 31.00 31.04 31.03 30.85 30.85 30.84 31.00 31.00 31.06 31.03 31.03 31.04 30.92 31.04 30.89 31.02 30.87 30.88
50.00 50.00 50.00 50.00 50.00 50.00 53.73 52.75 53.23 51.50 52.75 52.75 50.00 50.00 50.00 50.00 53.37 53.37 53.43 50.00 50.00 50.00 50.00 50.00 50.00 51.33 50.00 52.00 50.00 52.75 52.33
1.13 1.08 1.08 1.07 1.11 1.22 0.82 0.67 0.54 0.60 0.57 0.73 0.96 1.00 1.33 0.97 0.67 0.48 0.46 0.44 0.62 1.14 1.04 1.07 1.04 0.68 1.00 0.56 0.67 1.17 0.56
pits pits pits pits pits pits mounds mounds mounds mounds mounds mounds pits pits pits pits mounds mounds mounds mounds mounds pits pits pits pits mounds pits mounds mounds mounds mounds
higher degrees of supersaturation (Gratz et al. 1993; Dove and Hochella 1993; Pina et al. 1998; Bosbach et al. 1998; Teng et al. 2000; Pina et al. 2004). Two-dimensional growth models assume that growth occurs by generation and subsequent spread of surface nuclei (Nielsen 1964). In contrast, spiral growth circumvents the energy barrier associated with nucleation by taking advantage of screw dislocations, which provide a continuous and self-perpetuating source of attachment sites (Burton et al. 1951; Pina et al. 2004). Kessels et al. (2000) used AFM to analyze the surfaces of synthetic dolomite crystals from high-temperature growth experiments. They showed that early in the reaction Ca-rich dolomites were covered with rounded mounds, but late in the reaction, after dolomites became stoichiometric, surfaces were covered with flat layers. The work described here compares growth and dissolution nanotopography on natural and synthetic dolomite crystals with the aim of addressing two hypotheses. (1) High-temperature synthetic dolomites and low-temperature natural dolomites form by the same mechanism. If this hypothesis is unsubstantiated, the applicability of high-temperature experiments to natural low-temperature dolomite is tenuous at best. (2) Precipitation of amorphous metastable CaMg-carbonate is the first step in formation of dolomite. If their occurrence is common in nature, metastable, amorphous precursors could provide an energetically easier and more rapid mechanism by which to form dolomite. Therefore amorphous precursors could be one missing link in a solution to the ‘‘dolomite problem,’’ which has eluded geologists for more than two centuries. METHODS
Preparation of Synthetic Dolomite A series of eight experiments was conducted to dolomitize Iceland spar calcite in Mg–Ca–Cl solutions at 218 6 2uC. Each experiment was
defined by the initial Mg2+:Ca2+ ratio in solution (0.43, 0.50, 0.66, 0.79, 1.0, 1.14, 1.27, 1.50), which was attained by mixing different proportions of 1 M MgCl2 and CaCl2 solutions. Teflon-lined stainless steel reaction vessels were loaded with 0.1 g calcite (ground and sieved to 40–60 mm size fraction and cleaned ultrasonically in distilled water) and 15 ml solution. Reaction vessels were then sealed and heated for a predetermined time. At the end of each run individual reaction vessels were removed from the oven and allowed to air-cool for approximately 30 minutes. The contents of each vessel were then vacuum filtered, rinsed with distilled water, airdried, and stored in a vacuum desiccator until analysis. Preparation of Natural Dolomite Thirty-one natural dolomite rocks ranging in age from Ordovician to Holocene were analyzed (Table 1). Whole-rock samples were broken into pieces with a rock hammer, cleaned ultrasonically in distilled water, dried, and stored in a vacuum desiccator. X-Ray Diffraction (XRD) Powder X-ray diffraction was used to characterize the phases present and to determine composition and degree of cation order for reactionvessel contents and natural dolomites. XRD was carried out using CuKa radiation. Diffraction data were collected between 25 and 45u 2h using a step size of 0.004u with a 1 second count time. A small amount of CaF2 was added to powdered samples as an internal standard to calibrate the position of the d(104) dolomite peak. The X-ray-determined d(104) peak position was reproducible to within 0.006u 2h, or 6 0.33 mole% CaCO3. The percent product, which provides a measure of reaction progress in synthesis experiments, was determined by the peak-height ratio of the d(104) dolomite peak to the d(104) dolomite plus the d(104) calcite peak
426
JSR
S.E. KACZMAREK AND D.F. SIBLEY
FIG. 1.— Plot of induction period vs. initial solution Mg:Ca ratio.
(Royce et al. 1971). Product composition was calculated for synthetic and natural samples according to the relative corrected position of the d(104) dolomite peak (Lumsden and Chimahusky 1980). The degree of cation order is reported as the ratio of the d015:d110 superlattice reflection peaks following the methods of Goldsmith and Graf (1958). Atomic Force Microscopy (AFM) and Scanning Electron Microscopy (SEM) Samples used in AFM and SEM analyses were selected for relatively flat crystal faces. Because only the dominant rhombohedral form was observed, synthetic dolomite crystal surfaces are assumed to correspond to the d104 growth face. Natural dolomite crystal surfaces are assumed to correspond to the d104 cleavage plane. Natural and synthetic dolomite crystals used in the etching experiments were submerged in 0.5% H2SO4 solutions for 10–20 seconds at room temperature. Etched samples were then flushed with distilled water and dried in a vacuum desiccator. AFM analysis was conducted on a Digital Instruments Nanoscope III using constant force mode. Images were acquired under atmospheric conditions using silicon nitride cantilevers (NP-20) with an average tip radius of 20 nm. According to the manufacturer (Veeco Probes), 1 in 10 NP-20 tips have a radius of 5 nm. This is important because a smaller tip enhances the ability to resolve smaller surface features (Eggleston 1994). To limit artifacts, the force between each sample and the cantilever was minimized. According to our force calibration procedures the applied force was 5–20 nN. Although artifacts can form as a result of tip-surface interaction in this force range, no change in surface nanotopography was observed during extended scans (e.g., 20 minutes). Artifacts can also arise during imaging because of irregularly shaped cantilevers, as well as irregular cantilever movement across a crystal surface. In an effort to detect and minimize these types of artifacts, samples were periodically rotated and cantilevers replaced. All images reported here were reproduced on more than one crystal from each sample using multiple cantilever tips. Following AFM analyses, samples were prepared with a 2 nm electrically conductive osmium coating for observation with a JEOL 6300 field emission SEM at 10 kV. The mineralogy of individual crystals was verified with energy dispersive spectroscopy (EDS).
FIG. 2.—SEM micrograph of dolomite (d) forming on the edges and corners of a dissolving calcite crystal (c). RESULTS
High-temperature synthesis experiments are characterized by a relatively long induction period in which no products are detected by XRD after the onset of experimental conditions. The length of the induction period, as shown in Figure 1, is a function of the initial Mg:Ca ratio in solution. Products are detected sooner in solutions with high Mg:Ca ratios and later in solutions with low Mg:Ca ratios. In all experiments, the first products detected by XRD are invariably poorly ordered crystalline dolomite. SEM observations indicate that the first dolomite rhombohedra form preferentially on edges and corners, and in crystallographic continuity with the dissolving calcite substrate (Fig. 2). Synthetic dolomite compositions range from stoichiometric to Ca-rich (, 61 mol% CaCO3) with magnesium contents dictated by the initial Mg:Ca ratio in solution (R2 5 0.97). Figure 3 illustrates a marked increase in mole percent MgCO3 with increasing Mg:Ca ratio in solution for runs ended prior to calcite depletion. The overall dolomitization reaction rate is proportional to the initial Mg:Ca in solution, and in all cases, proceeds very rapidly once products are detected. Figure 4 plots percent product vs. time for the Mg:Ca 5 1.27 experiment. Provided that not all calcite reactants have been depleted, the slope of the reaction curve is used as a measure of the reaction rate. After depletion of calcite reactants all dolomites become stoichiometric. Although major reflections indicate that products are crystalline, dolomite from runs ended before 50% replacement lack superstructure reflection peaks, thus indicating poor ordering. Cation order in dolomite from runs with greater than 50% product relatively rapidly increases with time. In some experiments cation order increased linearly with time, but in others it increased less rapidly after calcite reactants were depleted (Fig. 5). Despite differences in solution chemistry and product composition, synthetic dolomites always become well ordered and stoichiometric after calcite reactants are exhausted. Comparison of data from these experiments indicates that synthetic dolomite produced in experiments stopped prior to calcite depletion have ordering ratios between 0.0 and 0.45, whereas dolomites from experiments stopped after calcite depletion have ordering ratios between 0.50 and 0.88. Because of differences in cation order, the terms nonideal and ideal are used herein to describe poorly ordered and relatively well ordered synthetic dolomite, respectively.
JSR
GROWTH AND DISSOLUTION FEATURES ON NATURAL AND SYNTHETIC DOLOMITES
427
FIG. 5.— Plot of cation order vs. time for the Mg:Ca 5 1.5 experiment. FIG. 3.— Plot of dolomite stoichiometry vs. solution Mg:Ca for experiments ended prior to calcite reactant depletion (R2 5 0.97, N 5 69).
The natural dolomites studied here have compositions that range from stoichiometric to Ca-rich (, 54 mol% CaCO3) and ordering ratios that range from 0.44 to 1.22 (Table 1). In general, XRD reflection peaks for ancient dolomites are stronger than those observed for synthetic dolomites with comparable magnesium contents. Examination of crystal surfaces using SEM and AFM shows that ideal and nonideal dolomites have different surface topographies. SEM observations show that nonideal synthetic dolomites have hummocky surfaces with irregular growth fronts (Fig. 6A), whereas ideal synthetic dolomites are generally flatter, with layers separated by straight-sided
FIG. 4.— Plot of percent product vs. time for the Mg:Ca 5 1.27 experiment.
steps (Fig. 6B). Despite these differences, only AFM provides the resolution necessary to clearly distinguish the nanotopography of nonideal and ideal dolomites. AFM observations of synthetic dolomite growth surfaces also reveal two principal features, described herein as mounds and layers. Mounds are rounded positive-relief features (Fig. 7) 10–200 nm in diameter and 1– 20 nm high. Growth mounds cover the surfaces of all nonideal dolomites prior to calcite depletion despite highly variable solution Mg:Ca ratios and stoichiometries. Following calcite reactant depletion, ideal dolomite crystal surfaces are characterized by flat, laterally extensive layers with steps that measure a few tenths (, 0.2 nm) to tens of nanometers high. Layers have very flat surfaces and do not have mounds on top of them. Some layers have straight-sided steps (Fig. 8A), whereas others are curved steps (Fig. 8B). In either case, steps generally run parallel to one another. In some cases, layers emerge from otherwise flat crystal faces (Fig. 8C), but no classic multilayer, helical growth spirals (hillocks) were observed. Following reactant depletion, all ideal synthetic dolomites observed with AFM had layers regardless of the solution Mg:Ca ratio. Chemically etched synthetic dolomite crystals are characterized by two distinct surface nanotopographies: mounds, and layers with euhedral etch pits. Mounds characterize etched nonideal synthetic dolomites, whereas euhedral pits characterize the surfaces of etched ideal synthetic dolomites. Mounds on etched crystals are indistinguishable from the dolomite growth mounds described above. Euhedral etch pits have pointed bottoms and occur on relatively flat layers. Etch pits measure 30– 200 nm in diameter and 1–30 nm deep (Fig. 9). Chemically etched natural dolomite crystal surfaces have the same features as chemically etched synthetic dolomites. Mounds were observed on all chemically etched natural nonstoichiometric dolomites as well as stoichiometric dolomites with cation ordering ratios below 0.67, whereas layers with euhedral pits were observed on the surfaces of stoichiometric, relatively well ordered ($ 1.0) dolomite (Table 1). Individual mounds are rounded topographical highs (Fig. 10) that measure 20–300 nm in diameter and 1–20 nm high. Mound-covered nonideal dolomites do not have etch pits or flat layers with steps. Holocene dolomites from Abu Dhabi have mounds on both etched and unetched surfaces. Prior to chemical etching the mounds were less pronounced, but they became more pronounced after being exposed to etching solutions for a few seconds. Chemically etched ideal natural dolomites exhibit euhedral etch
428
S.E. KACZMAREK AND D.F. SIBLEY
JSR
FIG. 7.—AFM image of a nonideal synthetic dolomite from the Mg:Ca 5 1.0 experiment showing mound nanotopography.
FIG. 6.— SEM micrographs of dolomite growth surfaces comparing synthetic dolomites from the Mg:Ca 5 1.0 experiment. A) Nonideal dolomite showing hummocky growth features. B) Ideal dolomite showing relatively flat growth surfaces separated by straight-sided steps.
pits with flat inter-pit layers separated by steps that run parallel to one another (Fig. 11). Euhedral etch pits are typically rhombic and measure 20–500 nm wide and 5–90 nm deep with pointed bottoms (Fig. 12). DISCUSSION
Mounds observed on the surfaces of natural and synthetic dolomites indicate that natural and synthetic dolomites form in a similar manner, but the mechanism is unclear. Mounds are not characteristic of any of the three classic crystal growth models—mononuclear, polynuclear, and spiral growth (Nielsen 1964: Ohara and Reid 1973; Sunagawa 1984). None of these models predict rounded topographic highs that do not have steps or crystal faces. Polynuclear growth features, often referred to as islands, have been observed on the surfaces of calcite, gypsum, and barite. Islands tend to be flat topped and straight edged with layers one to several monolayers thick (Hillner et al. 1992; Dove and Hochella 1993; Bosbach and Rammensee 1994; Bosbach et al. 1998; Pina et al. 1998; Teng et al. 2000; Pina et al. 2004). It is possible that dolomite mounds have steps that are below the resolution of the AFM. If steps of any height are separated by only a few nanometers, then the AFM tip may ride over them with little or no deflection. The best indication of the size and spacing of steps
that can be imaged on mounds is the size and spacing of steps that we have imaged in etch pits. We have clearly imaged steps that are less than 0.2 nm in height and 4 nm apart. We are also certain that it is possible to image straight edges that are at least six nanometers long because this is the shortest straight edge we have imaged on an etch pit. If steps are present on mounds, then they must be less than 0.2 nm high and 4 nm apart. Mounds also differ from islands in that islands are flat topped whereas mounds have rounded tops. Growth hillocks centered on spiral dislocations with steps too small to resolve could appear rounded, like the mounds we observe. In this case, pits should form at the apices of dislocations when chemically etched, but etching of mounds only reveals more mounds. These comparisons indicate that mounds are most likely smooth and rounded topographic highs, which are different from crystalline growth features formed by mononuclear, polynuclear, or spiral growth. Because dolomite growth mounds are rounded and lack crystal faces, they are similar in shape to the rounded morphologies of amorphous calcium carbonate (ACC) (Reddy and Nancollas 1976; Raz et al. 2000; Kelleher and Redfern 2002; Loste et al. 2003; Schmidt et al. 2005). Reddy and Nancollas (1976) added Mg2+ to [Ca2+]:[CO322] 5 1 solutions and observed the formation of amorphous precipitates after 3 minutes. After 20 days combinations of aragonite spherulites and irregularly shaped calcite were formed. Raz et al. (2000) documented the formation of magnesium calcites via a partially stabilized amorphous phase that contained up to 21 mol% magnesium. Loste et al. (2003) demonstrated that ACC was the first phase to form during precipitation from highly supersaturated solutions containing magnesium and that ACC later crystallized into Mg-calcite. They also showed that the magnesium content of the ACC increased with the Mg:Ca ratio in solution, which is consistent with our findings (see Fig. 3). Stolarski and Mazur (2005) etched biogenic calcium carbonate believed to have formed via amorphous calcium carbonate precursors and found mounds similar to those observed in this study. DiMasi et al. (2006) found that during biomimetic mineralization of CaCO3, the formation of amorphous calcium carbonate films was partially controlled by magnesium. The mound-shaped nanotopography on the surfaces of nonideal dolomites is consistent with an amorphous CaMg-carbonate precursor, but crystallization to nonideal dolomite must take place soon after deposition because nonideal dolomite was detected by XRD throughout the reaction. This is not unreasonable at the high temperatures employed in our experiments. At room temperature amorphous CaMg-carbonates
JSR
GROWTH AND DISSOLUTION FEATURES ON NATURAL AND SYNTHETIC DOLOMITES
429
FIG. 8.—AFM images of ideal synthetic dolomite growth surfaces with layers. A) Dolomite growth layers from the Mg:Ca 5 1.0 experiment showing a straight-sided step that measures 0.6 nm high. B) Stacked layers with curved steps on a dolomite growth surface from the Mg:Ca 5 1.5 experiment. Step height between arrows is 0.2 nm. C) Dolomite surface from the Mg:Ca 5 0.66 experiment showing emerging growth layers.
have been shown to crystallize to calcite and/or aragonite in minutes to days (Reddy and Nancollas 1976; Tracy et al. 1998, Loste et al. 2003), and according to Schmidt et al. (2005) the time for crystallization of amorphous CaMg-carbonate decreased with increasing temperature. The ability to detect growth mounds in ancient nonideal dolomites with chemical etching indicates that mounds maintain their growth structure and remain as separate physical entities over a long time. If adjacent growth mounds with slightly different composition or cation order meet
during growth, lattice mismatch is likely to result. In this case, the interfacial free energy between mounds could prevent them from coalescing in a way that anneals the boundary. If such interfaces exist, crystals grown by mounds would have a pervasive internal defect structure. This is consistent with the modulated microstructures observed with TEM in ancient Ca-rich dolomites (Barber 1977; Frisia 1994; Reeder 2000; Schubel et al. 2000). Modulations, which measure nanometers to tens of nanometers in wavelength, are attributed to chemical and/or
430
S.E. KACZMAREK AND D.F. SIBLEY
JSR
FIG. 9.— AFM image of an etched ideal synthetic dolomite from the Mg:Ca 5 0.66 experiment showing euhedral etch pits on flat layers.
FIG. 11.—AFM image of an etched natural ideal dolomite surface (Sample S-1) showing euhedral etch pits with flat inter-pit layers.
structural domains caused by fluctuations in stoichiometry and ordering created during growth (Reeder 2000). Without defective growthinterfaces, mounds would coalesce into defect-free layers. As a result, these interfaces would be undetectable with chemical etching because the energetic drive for preferential dissolution between individual growthmounds would not exist. Layers and euhedral etch pits on the surfaces of ideal dolomite indicate either spiral or polynuclear growth. In spiral growth, the most common growth mechanism found in natural crystals (Sunagawa 1977, 1981), material is added to step and kink sites along layers created by spiral dislocations. The most convincing evidence for spiral growth is the presence of multilayer spiral hillocks (Hillner et al. 1992; Gratz et al. 1993; Bosbach and Rammensee 1994; Bosbach et al. 1998; Pina et al.
1998; Sunagawa 2005). Some layers observed on ideal synthetic dolomite merge into the crystal surface (see Fig. 10C), thus producing a geometry that is consistent with a spiral dislocation outcrop (Maiwa et al. 1998; Pina et al. 1998; Teng et al. 2000). However, no multi-layer helical growth spirals were observed. Kessels et al. (2000) suggested that stacked layers on synthetic dolomite form by birth and spread. Birth-and-spread growth is a variety of polynuclear growth characterized by attachment of growth
FIG. 10.— AFM image of an etched natural nonideal dolomite (Sample SB 35) with a mound-covered surface.
FIG. 12.— AFM image of an etched natural ideal dolomite surface (Sample MH-2) showing crystallographically oriented rhombic etch pits.
JSR
GROWTH AND DISSOLUTION FEATURES ON NATURAL AND SYNTHETIC DOLOMITES
units (ions or molecules) along the periphery of individual surface nuclei. These nuclei can advance laterally until they intersect and fuse with other nuclei to form layers. Therefore, the birth-and-spread model could also explain some of the geometries observed on synthetic dolomite surfaces, such as stacked layers. Euhedral etch pits that are hundreds of monolayers deep are indicative of dissolution at extended line defects, such as screw dislocations at the center of growth spirals (Brantley et al. 1986; Sangwal 1987; Lin and Shen 1993). Extended line defects are the site of preferential dissolution because defects have excess strain energy (Lasaga 1990). Point defects, such as precipitates and inclusions, may also be sites of etch pit formation because they too are zones of elevated reactivity. Point defects, however, promote only shallow, flat-bottomed etch pits, because once a pit nucleates and the point defect is removed, the drive for dissolution in a direction normal to the crystal surface is eliminated. Gratz et al. (1991) observed the formation of shallow and deep pits in quartz crystals during chemical etching experiments. Hillner et al. (1992) and Liang et al. (1996) made similar observations in calcite using AFM. These authors argued that shallow, flat-bottom pits corresponded to point defects, whereas deep, pointed pits formed at extended line defects. Based on his observations using in situ AFM techniques, Teng (2004) reported that the value of undersaturation at which deep etch pits began to form on freshly cleaved calcite crystals was in good agreement with the value predicted by dislocation theory for pit formation at line dislocations. Lendvay et al. (1971) found that formation of etch pits in undersaturated conditions corresponded to sites of spiral-growth hillocks in experimentally grown crystals. Therefore, flat growth layers with steps observed in conjunction with deep, euhedral etch pits on both natural and synthetic ideal dolomites are consistent with the spiral-growth model. Mounds always form first during dolomitization of calcite, whereas layers form only after calcite reactants have been depleted. Therefore, over the conditions tested, dolomite surface nanotopography is independent of the Mg:Ca ratio in solution. Because the overall rate of the reaction is proportional to the Mg:Ca ratio in solution, surface nanotopography is also independent of the overall rate of dolomitization. Although we do not fully understand the mechanism for changing surface topography, it appears to be a function of carbonate ion concentration because the carbonate ion concentration probably decreases significantly when the calcite reactant is used up. Based on the experimental design, the carbonate ion concentration for any Mg:Ca ratio is likely to remain relatively constant as long as calcite is present because prior to reactant depletion calcite solubility dictates the concentration of carbonate ions in solution. As dolomite begins to form and sequesters carbonate ions, more calcite dissolves. Following calcite depletion, however, metastable, nonideal dolomite is the only source of carbonate reactants for the growth of ideal dolomite. Although it is not possible to measure the carbonate activity in the reaction vessels, it must be lower after calcite depletion because nonideal dolomite is less soluble than calcite (Carpenter 1980; Chai et al. 1995). Therefore, mounds form early in the reaction when the carbonate ion concentration is relatively high and layers form after calcite depletion when carbonate ion concentration is relatively low. This is consistent with the fact that all low-temperature precipitation of synthetic amorphous calcium carbonate has been accomplished in solutions with high carbonate ion concentrations (Reddy and Nancollas 1976; Tracy et al. 1998, Kelleher and Redfern 2002; Loste et al. 2003). CONCLUSIONS
AFM and SEM observations of natural dolomites reveal two distinct etch features: amorphous mounds, and layers with euhedral pits. Mounds are observed on the surfaces of Ca-rich and/or disordered (nonideal) dolomite, whereas layers with pits are characteristic of stoichiometric and relatively well-ordered (ideal) dolomite. Mounds and layers with etch pits
431
also characterize the surfaces of high-temperature synthetic nonideal and ideal dolomites, respectively. The observations presented here indicate that despite differences in formation conditions, high-temperature synthetic dolomites and ancient dolomites have the same surface nanotopography, thus indicating that they form by similar mechanisms. Therefore, the results of this study are further evidence that hightemperature synthetic dolomitization reactions may be accurate analogs for low-temperature natural dolomitization. In all cases, the formation of well-ordered, stoichiometric synthetic dolomite proceeds through a poorly ordered, metastable phase characterized by rounded growth mounds. The presence of mounds on the surfaces of nonideal dolomites is inconsistent with standard models of crystal growth but is indicative of an amorphous precursor. The two different surface nanotopographies described here are consistent with a model in which nonideal dolomite forms by precipitation of amorphous mounds that quickly crystallize to nonideal dolomite while ideal dolomite replaces nonideal dolomite and grows by spiral growth. This new model for dolomite crystal growth will have to be tested with actively forming natural dolomites before we can be confident that it is accurate. If nonideal dolomites form via an amorphous precursor, interpretations of the origin of many dolomites will change. For example, Schmidt et al. (2005) postulated that amorphous precursors to dolomite might be important in understanding the oxygen isotope composition of dolomite. Also, if nonideal dolomites form via amorphous precursors, then nonideal dolomites probably form in solutions characterized by high degrees of supersaturation with respect to dolomite and high carbonate ion concentrations. ACKNOWLEDGMENTS
The authors would like to thank Kitty Milliken, Leslie Melim, Patricia Dove, John Southard, and an anonymous reviewer for helpful editorial comments during the review process. We would also like to further acknowledge the thoughtful comments of Patricia Dove and thank her for introducing us to the idea of amorphous carbonate precipitation. REFERENCES
ARVIDSON, R.S., AND MACkENZIE, F.T., 1999, The dolomite problem: control of precipitation kinetics by temperature and saturation state: American Journal of Science, v. 299, p. 257–288. BAKER, P.A., AND KASTNER, M., 1981, Constraints on the formation of sedimentary dolomite: Science, v. 213, p. 214–216. BARBER, D.J., 1977, Defect microstructures in deformed and recovered dolomites, Tectonophysics:, v. 39, p. 193–213. 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, v. 58, p. 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, v. 151, p. 143–160. 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, v. 50, p. 2349–2361. BURTON, W.K., CABRERA, N., AND FRANK, F.C., 1951, The growth of crystals and the equilibrium structure of their surfaces: Royal Society (London), Philosophical Transactions, v. 243, p. 299–358. CARPENTER, A.B., 1980, The chemistry of dolomite formation I: The stability of dolomite, in Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds., Concepts and Models of Dolomitization: SEPM, Special Publication 28, p. 111–121. CHAI, L., NAVROTSKY, A., AND REEDER, R.J., 1995, Energetics of calcium-rich dolomite: Geochimica et Cosmochimica Acta, v. 59, p. 939–944. DIMASI, E., KWAK, S., AMOS, F.F., OLSZTA, M.J., LUSH, D., AND GOWER, L.B., 2006, Complementary control by additives of the kinetics of amorphous CaCO3 mineralization at an organic interface: In-situ synchrotron X-ray observations: Physical Review Letters, v. 97045503, p. 1–4. DOVE, P.M., AND HOCHELLA, M.F, JR., 1993, Calcite precipitation mechanisms and inhibition by orthophosphate: In situ observations by Scanning Force Microscopy: Geochimica et Cosmochimica Acta, v. 57, p. 705–714.
432
S.E. KACZMAREK AND D.F. SIBLEY
EGGLESTON, C.M., 1994, High-resolution scanning probe microscopy: tip-surface interaction, artifacts and applications in mineralogy and geochemistry, in Nagy, K.L., and Blum, A.E., eds., Scanning Probe Microscopy of Clay Minerals: Clay Minerals Society, Workshop, v. 7, p. 3–90. 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: International Association of Sedimentologists, Special Publication 21, p. 55–74. GAINES, A.M., 1974, Protodolomite synthesis at 100uC and atmospheric pressure: Science, v. 55, p. 84–101. GOLDSMITH, J.R., AND GRAF, D.L., 1958, Structural and compositional variations in some natural dolomites: Journal of Geology, v. 66, p. 678–693. GRAF, D.L., AND GOLDSMITH, J.R., 1956, Some hydrothermal synthesis of dolomite and protodolomite: Journal of Geology, v. 64, p. 173–187. GRATZ, A., MANNE, S., AND HANSMA, P.K., 1991, Atomic force microscopy of atomicscale ledges and etch pits formed during dissolution of quartz: Science, v. 251, p. 1343–1346. GRATZ, A.J., HILLNER, P.E., AND HANSMA, P.K., 1993, Step dynamics and spiral growth on calcite: Geochimica et Cosmochimica Acta, v. 57, p. 491–495. HILLNER, P.E., GRATZ, A.J., MANNE, S., AND HANSMA, P.K., 1992, Atomic-scale imaging of calcite growth and dissolution in real time: Geology, v. 20, p. 359–362. KATZ, A., AND MATTHEWS, A., 1977, The dolomitization of CaCO3: an experimental study at 252–292uC: Geochimica et Cosmochimica Acta, v. 41, p. 297–308. KELLEHER, I.J., AND REDFERN, S.A.T., 2002, Hydrous calcium magnesium carbonate, a possible precursor to the formation of sedimentary dolomite: Molecular Simulation, v. 28, p. 557–572. KESSELS, L.A., SIBLEY, D.F., AND NORDENG, S.H., 2000, Nanotopography of synthetic and natural dolomite crystals: Sedimentology, v. 47, p. 173–186. KRAUSKOPF, K.B., AND BIRD, D.K., 1995. Introduction to Geochemistry: New York, McGraw-Hill, 647 p. KWAK, S.Y., DIMASI, E., JIRHAN, Y., AIZENBERG, J., AND KUZMENKO, I., 2005, Orientation and Mg incorporation of calcite grown on functionalized self-assembled monolayers: a syncrotron X-ray study: Crystal Growth and Design, v. 5, p. 2139–2145. LAND, L.S., 1980, The isotopic and trace element geochemistry of dolomite: the state of the art, in Zenger, D.L., Dunham, J.B., and Ethington, R.L., eds., Concepts and Models of Dolomitization: SEPM, Special Publication 28, p. 87–110. LAND, L.S., 1998, Failure to precipitate dolomite at 25uC from dilute solution despite 1000-fold oversaturation after 32 years: Aquatic Geochemistry, v. 4, p. 361–368. LASAGA, A.C., 1990, Atomic treatment of mineral-water surface reactions, in Hochella, M.F, JR, and White, A.F., eds., Mineral–Water Interface Geochemistry: Reviews in Mineralogy, v. 23, p. 17–85. LENDVAY, E., KUHN, A., CHEVY, A., AND CEVA, T., 1971, Dislocation etching of GaSe single crystals: Journal of Materials Science, v. 6, p. 305–308. LIANG, Y., BAER, D., MCCOY, J., AMONETTE, J., AND LAFEMMINA, J., 1996, Dissolution kinetics at the calcite–water interface: Geochimica et Cosmochimica Acta, v. 60, p. 4883–4887. LIN, C.C., AND SHEN, P., 1993, Role of screw axis in dissolution of willemite: Geochimica et Cosmochimica Acta, v. 57, p. 1649–1655. LOSTE, E., WILSON, R.M., SESHADRI, R., AND MELDRUM, F.C., 2003, The role of magnesium in stabilizing amorphous calcium carbonate and controlling calcite morphologies: Journal of Crystal Growth, v. 254, p. 206–218. LUMSDEN, D.N., AND CHIMAHUSKY, J.S., 1980, Relationship between dolomite nonstoichiometry and carbonate facies parameters, in Zenger, D.H., Dunham, J.B., and Ethington, R.L., eds., Concepts and Models of Dolomitization: SEPM, Special Publication 28, p. 111–121. 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, v. 186, p. 214–223.
JSR
MALONE, M.J., BAKER, P.A., AND BURNS, S.J., 1996, Recrystallization of dolomite: An experimental study from 50–200uC: Geochimica et Cosmochimica Acta, v. 60, p. 2189–2207. NIELSEN, A.E., 1964. Kinetics of Precipitation: New York, Macmillan, 151 p. NORDENG, S.H., AND SIBLEY, D.F., 1994, Dolomite stoichiometry and Ostwald’s step rule: Geochimica et Cosmochimica Acta, v. 58, p. 191–196. OHARA, M., AND REID, R.C., 1973. Modeling Crystal Growth Rates From Solution: Englewood Cliffs, New Jersey, Prentice Hall, 271 p. PINA, C.M., BECKER, U., RISTHAUS, P., BOSBACH, D., AND PUTNIS, A., 1998, Molecularscale mechanisms of crystal growth in barite: Nature, v. 395, p. 483–486. PINA, C.M., PUTNIS, A., AND ASTILLEROS, J.M., 2004, The growth mechanisms of solid solutions crystallizing from aqueous solutions: Chemical Geology, v. 204, p. 145–161. POLITI, Y., ARAD, T., KLEIN, E., WEINER, S., AND ADDADI, L., 2004, Sea urchin spine calcite forms via a transient amorphous calcium carbonate phase: Science, v. 306, p. 1161–1164. RAZ, S., WEINER, S., AND ADDADI, L., 2000, Formation of high-magnesium calcites via an amorphous precursor phase: possible biological implications: Advanced Materials, v. 12, p. 38–42. REDDY, M.M., AND NANCOLLAS, G.H., 1976, The crystallization of calcium carbonate: Journal of Crystal Growth, v. 35, p. 33–38. REEDER, R.J., 1992, Carbonates: Growth and alteration microstructures, in Busek, P.R., ed., Minerals and Reactions at the Atomic Scale; Transmission Electron Microscopy: Reviews in Mineralogy, v. 27, p. 380–424. REEDER, R.J., 2000, Constraints on cation order in calcium-rich sedimentary dolomites: Aquatic Geochemistry, v. 6, p. 213–226. ROYCE, C.R., WADELL, J.S, JR., AND PETERSEN, L.E., 1971, X-ray determination of calcite–dolomite: an evaluation: Journal of Sedimentary Petrology, v. 41, p. 483–488. SANGWAL, K., 1987. Etching of Crystals: Theory, Experiment, and Application: Oxford, U.K., North-Holland, 497 p. SCHMIDT, M., XEFLIDE, S., BOTZ, R., AND MANN, S., 2005, Oxygen isotope fractionation during synthesis of CaMg-carbonate and implication for sedimentary dolomite formation: Geochimica et Cosmochimica Acta, v. 69, p. 4665–4674. SCHUBEL, K.A., ELBERT, D.C., AND VEBLEN, D.R., 2000, Incommensurate c-domain superstructure in calcian dolomite from the Latemar buildup, Dolomites, Northern Italy: American Mineralogist, v. 85, p. 858–862. SIBLEY, D.F., 1990, Unstable to stable transformations during dolomitization: Journal of Geology, v. 98, p. 739–748. SIBLEY, D.F., DEDOES, R.E., AND BARTLETT, T.R., 1987, Kinetics of dolomitization: Geology, v. 15, p. 1112–1114. STOLARSKI, J., AND MAZUR, M., 2005, Nanostructure of biogenic versus abiogenic calcium carbonate crystals: Acta Palaeontologica Polonica, v. 50, p. 847–865. SUNAGAWA, I., 1977, Natural crystallization: Journal of Crystal Growth, v. 42, p. 214–223. SUNAGAWA, I., 1981, Characteristics of crystal growth in nature as seen from the morphology of mineral crystal: Bulletin de Mineralogie, v. 104, p. 81–97. SUNAGAWA, I., 1984, Growth of crystals in nature, in Sunagawa, I., ed., Materials Science of the Earth’s Interior: Tokyo, Terra Scientific Publishing, p. 63–105. SUNAGAWA, I., 2005, Crystals: Growth, Morphology and Perfection: Cambridge, U.K., Cambridge University Press, 308 p. TENG, H.H., 2004, Controls by saturation state on etch pit formation during calcite dissolution: Geochimica et Cosmochimica Acta, v. 68, p. 253–262. TENG, H.H., DOVE, P.M., AND DE YOREO, J., 2000, Kinetics of calcite growth: Surface processes and relationships to macroscopic rate laws: Geochimica et Cosmochimica Acta, v. 64, p. 2255–2266. TRACY, S.L., WILLIAMS, D.A., AND JENNINGS, H.M., 1998, The growth of calcite spherulites from solution II. Kinetics of formation: Journal of Crystal Growth, v. 193, p. 382–388. Received 5 December 2005; accepted 2 November 2006.
