Surface chemistry allows for abiotic precipitation of dolomite at low temperature

Significance

Abundant in the geologic record, but scarce in modern environments below 50 °C, the mineral dolomite is used to interpret ancient fluid chemistry, paleotemperature, and is a major hydrocarbon reservoir rock. Because laboratory synthesis of abiotic dolomite had been unsuccessful, chemical mechanisms for precipitation are poorly constrained, and limit interpretations of its occurrence. Here we report the abiotic synthesis of dolomite at 25 °C, and demonstrate that carboxylated surfaces on organic matter catalyze precipitation through complexation between carboxyl groups and Mg²⁺, removing water to make Mg²⁺ available for dolomite precipitation. This mechanism is consistent with dolomite formation in depositional environments rich in organic matter. Our experimental protocol provides opportunities for calibrating conditions of low-temperature dolomite formation throughout the geologic record.

Abstract

Although the mineral dolomite is abundant in ancient low-temperature sedimentary systems, it is scarce in modern systems below 50 °C. Chemical mechanism(s) enhancing its formation remain an enigma because abiotic dolomite has been challenging to synthesize at low temperature in laboratory settings. Microbial enhancement of dolomite precipitation at low temperature has been reported; however, it is still unclear exactly how microorganisms influence reaction kinetics. Here we document the abiotic synthesis of low-temperature dolomite in laboratory experiments and constrain possible mechanisms for dolomite formation. Ancient and modern seawater solution compositions, with identical pH and pCO₂, were used to precipitate an ordered, stoichiometric dolomite phase at 30 °C in as few as 20 d. Mg-rich phases nucleate exclusively on carboxylated polystyrene spheres along with calcite, whereas aragonite forms in solution via homogeneous nucleation. We infer that Mg ions are complexed and dewatered by surface-bound carboxyl groups, thus decreasing the energy required for carbonation. These results indicate that natural surfaces, including organic matter and microbial biomass, possessing a high density of carboxyl groups may be a mechanism by which ordered dolomite nuclei form. Although environments rich in organic matter may be of interest, our data suggest that sharp biogeochemical interfaces that promote microbial death, as well as those with high salinity may, in part, control carboxyl-group density on organic carbon surfaces, consistent with origin of dolomites from microbial biofilms, as well as hypersaline and mixing zone environments.

Although synthesis of dolomite in laboratory settings at high temperature (80–250 °C) has yielded valuable information regarding dolomite formation (1, 2), the validity of extrapolating kinetic data at 250 °C down to 25 °C is questionable. Synthesis of low-temperature dolomite is hindered by slow reaction kinetics (2). Kinetic inhibition is attributed to lack of solution supersaturation (3), sulfate inhibition (1), cation desolvation (4), and lack of nucleation sites (5). Laboratory precipitation at low temperature has only been successful in producing disordered dolomite: from solutions with high salinity (6); through intermittent (7) or complete dehydration (8); by using organic or inorganic compounds that effectively dewater Mg²⁺ ions (9–11); or in the presence of microorganisms, their exudates, or surfaces (12, 13).

Microbial dolomite has been produced in the presence of several different metabolic pathways including sulfate reduction, methanogenesis, methanotrophy, sulfide oxidation, and aerobic respiration (12–16), which may drive precipitation through the supersaturation of solutions with respect to dolomite. Recent work, however, has focused on the role of microbial cells and exopolymeric substances (EPS) as surfaces for dolomite nucleation (17). Whereas these studies clearly demonstrate that these surfaces are involved in dolomite formation, specific mechanisms of nucleation have not been identified, but likely include serving as surfaces for heterogeneous nucleation, epitaxial growth, or a source of surface functional groups that catalytically influence growth rates (18–20).

Undoubtedly, a lack of abiotic ordered dolomite synthesis at low temperature indicates that we have not yet discovered the key to dolomite nucleation in low-temperature sedimentary systems (3), hindering our ability to effectively model its geochemical behavior at low temperature and to apply these relationships to predict its occurrence in the rock record. Here we report an example of abiotic laboratory precipitation of dolomite at 30 °C by nucleation on synthetic, carboxylated surfaces, used as models for natural organic surfaces including microbial EPS.

The purpose of this study was to examine the influence of model organic surfaces on dolomite precipitation with the goal of ascertaining how characteristics of these surfaces contribute to nucleation and precipitation. We used batch laboratory experiments with two marine-type solutions with identical starting pCO₂ and salinity. One solution represents conditions typical of modern seawater (MSW) favoring aragonite and high-Mg calcite precipitation (21), whereas the other broadly simulates Silurian seawater (SSW) in which calcite is the favored carbonate phase (22). These experimental conditions give context to dolomite formation under a range of conditions present during different periods of the geologic past.

Results and Discussion

Aqueous Geochemistry and Mineralogy.

pH decreased from 8.2 to 7.8 and 7.7, respectively (0.82 and 20.3), in MSW and from 8.2 to 7.29 and 7.15, respectively (for polystyrene spheres with diameters 0.82 and 20.3), in SSW after 20 d. Control vessels for MSW dropped to 7.94 in MSW and 7.33 in SSW (Fig. 1); this drop in pH coincided with precipitation of carbonate minerals and therefore changes in saturation state, Mg:Ca and dissolved CO₃²⁻ (carbonate speciated using PhreeqC; 23). Mg:Ca ratio dropped from 5.1 to 3.25 and 2.99, respectively (for polystyrene spheres with diameters 0.82 and 20.3 μm), in MSW and from 1.4 to 1.15 and 1.03, respectively (for polystyrene spheres with diameters 0.82 and 20.3 μm), in SSW after 20 d. After 20 d control vessels for MSW had a Mg:Ca ratio of 5.34 and 1.44 for SSW. Dolomite, calcite, and aragonite were initially supersaturated in both solutions (Table 1).

Fig. 1.

Aqueous geochemistry of experimental vessels. pH (Upper Left), Mg:Ca (Upper Right), CO₃^2- (Lower Left; modeled from titrated alkalinity and measured pH using PhreeqC; 23) and dolomite saturation state (Ω_dolo; Lower Right) in experimental vessels as a function of time, with either MSW or SSW and two different spheres. Values for control experiments are as follows for MSW at 20 d (gray squares): pH = 7.94; Mg:Ca = 5.34; CO₃²⁻ = 0.0173; Ω_dolo = 98; SSW (black squares): pH = 7.33; Mg:Ca = 1.44; CO₃²⁻ = 0.0035; Ω_dolo = 9.3.

Table 1.

Solution compositions and modeled saturation state of carbonate phases

Solution composition	MSW (40)	SSW* (38)
pH	8.2	8.2
Na+†	469	445
Ca2+	10.3	35
Mg2+	52.8	48
Mg:Ca	5.1:1	1.4:1
HCO3−	2.4	2.4
SO42−	28	11
R-COO−(0.82μm)‡	5.1 × 1012	5.1 × 1012
R-COO−(20.3μm)	2.8 × 1012	2.8 × 1012
Ωdol¶	316	234
Ωcal	6.0	13.5
Ωarag	4.4	14.5

^*SSW compositions were taken from ref. 38 except for pH and HCO₃⁻, which were estimated from modern values. For both solutions pCO₂ = 10^−3.4 atm and I = 0.63.

^†All aqueous concentrations are given in mmol L⁻¹.

^‡Carboxyl-group densities are given as groups L⁻¹ and achieved by adding either 1 × 10⁻² mg L⁻¹ 0.82-μm spheres or 1 × 10⁻³ mg L⁻¹ 20.3-μm spheres to the experimental solutions.

^¶Saturation states calculated using PhreeqC at 30 °C using the thermodynamic database phreeqe.dat (K_sp were 10^−8.2, 10^−8.4, and 10^−17.1 for aragonite, calcite, and dolomite, respectively; 23), which uses the extended Debye–Hückel equation suitable for seawater ionic strengths, such as those analyzed in this study.

All solutions were initially supersaturated with respect to calcite, aragonite, and dolomite (Table 1). In all experimental vessels changes in pH and CO₃²⁻ values did not vary from the control within the error of measurement whereas values of Mg:Ca and degree of supersaturation for calcite, aragonite, and dolomite decreased (Fig. 1) more than control values after the 20-d period. Control vessels for SSW produced aragonite (CaCO₃), and calcite (CaCO₃); control vessels for MSW produced only aragonite. All experimental vessels containing SSW produced aragonite and calcite, with a smaller peak that is consistent with dolomite (Fig. 2). Precipitates treated with weak acetic acid to remove Ca carbonates (e.g., 24) revealed low yields [based on X-ray diffraction (XRD) intensity] of Mg calcite (19–49 mol % Mg; calculated from XRD with 3; % MgCO₃ = [(−363.96*d-spacing) + 1,104.5]) and disordered (64–75 mol % Mg; calculated from XRD with 3) and ordered dolomite [CaMg(CO₃)₂] (Fig. 3). Experimental vessels containing MSW produced aragonite and calcite. Acetic acid treatments revealed extremely low yields of Mg calcite (5–42 mol % Mg) and disordered dolomite (50–51mol % Mg and, 61–75 mol % Mg), likely due to Mg binding to the sphere surface, thus decreasing the available Mg in solution (25). Because of the extremely low abundance of Mg-bearing precipitates, dolomite mineralogy was also investigated using high-resolution transmission electron microscopy (HRTEM).

Fig. 2.

XRD of mineral precipitates after 20 d (intensity vs. 2 θ). (A) SSW for 20.3-μm polystyrene spheres (black line) and 0.82-μm polystyrene spheres (gray line). Aragonite, calcite, and ordered and disordered peaks of dolomite are present. (B) MSW for 20.3-μm polystyrene spheres (black line) and 0.82-μm polystyrene spheres (gray line). Aragonite is present for both spheres, whereas calcite is present in polystyrene 20.3-μm spheres only. Yield (as inferred by peak intensity) for MSW is much lower than SSW for all carbonates, and dolomite peaks are unclear for MSW due to low yield.

Fig. 3.

XRD of mineral precipitates after 20 d (intensity vs. 2 θ) after leaching of precipitates with weak acetic acid (e.g., 24). SSW for 20.3-μm polystyrene spheres (black line) and 0.82-μm polystyrene spheres (gray line). Ordering peaks for dolomite (labeled using [hkl] notation) at (A) 11(-1) and (B) 333. Note that superstructure reflections are more pronounced for treatments with 20.3-μm polystyrene spheres. Acetic acid leaching of precipitates from MSW treatments was below the detection limit of the instrument.

Dolomite Identification.

Scanning electron microscopy (SEM) and HRTEM analysis demonstrate that microspheres are thinly coated with minerals (Fig. 4) and serve as points of nucleation for nanocrystalline mineral aggregates (Fig. 5; see discussion below), whereas larger aragonite crystals (identified based on elemental analyses and crystal structure) precipitate in the pore space (Fig. 4). HRTEM-based energy-dispersive X-ray (EDX) spectroscopy reveals a predominance of Mg-rich phases at sphere surfaces characteristic of both dolomite (Mg:Ca = 1) and a phase with Mg:Ca ratios greater than 1. Surface-associated carbonate precipitates form as nanoparticles with fine-scale layering, and HRTEM with selected area electron diffraction (SAED) of precipitates shows lattice fringes of 2.89 Å [211] (Fig. 6). Our electron microscopy data and observations therein suggest strongly that Mg-rich phases form preferentially in association with carboxylated surfaces. Absence of Mg-rich phases in control experiments is strong evidence supporting surface mediated nucleation of these phases.

Fig. 4.

SEM photomicrograph of 0.82-μm spheres in SSW after 20 d with carbonates nucleating from sphere surface. (A) Large crystals of aragonite homogeneously nucleated in solution (identified using elemental analysis combined with crystal structure). Note spheres attached onto crystal surface. Scale bar, 1 μm. (B) Thin precipitates form on sphere surfaces. Scale bar, 100 nm.

Fig. 5.

Mg:Ca ratios of carbonate precipitates. HRTEM photomicrograph of 0.82-μm spheres in SSW after 20 d with carbonates nucleating from sphere surface (Left). Arrows indicate locations at which Mg:Ca = 1 and demonstrate a predominance of Mg:Ca = 1 for Mg-bearing phases along the x–y transect (Right). Mg:Ca = 0 indicates Ca carbonate.

Fig. 6.

HRTEM photomicrograph of precipitated carbonates on the sphere surface in SSW after 20 d. Arrows indicated measured fringes and resulting d-spacings consistent with dolomite for the indicated region. White line indicates the approximate boundary between the sphere surface and the mineral precipitates. (Inset) SAED of surface precipitates indicating ordered dolomite. Scale bars, 10 nm.

Dehydration of Mg²⁺ via Surface-Bound Carboxyl Groups.

Because of the spatial association of dolomite exclusively with microsphere surfaces, we propose that natural low-temperature dolomite nucleates preferentially in the presence of carboxylated surfaces, such as certain types of organic matter with high carboxyl sites densities (>0.1 sites Å⁻²; Table 2). Dehydration and carbonation of Mg ions are rate-limiting steps in dolomite nucleation (4). Although increased Mg:Ca ratios in calcium carbonates have been observed in precipitates formed in the presence of dissolved carboxylated molecules (26), ordered, crystalline phases such as the ones in this study have not been reported to form abiotically at low temperature. Surface carboxyl functional groups, however, have been demonstrated to adsorb Mg²⁺ enabling the formation of a [Mg(H₂O)₅(R-COO)]⁺ complex, which results in the ejection of a water molecule (27):

This reaction lowers the energy needed for carbonation and the subsequent attachment of Ca²⁺ when the Mg(H₂O)₅²⁺ complex is bound to an RCOO⁻ group compared with dewatering and carbonation of Mg(H₂O)₆²⁺ (ΔG is 13.6 kcal mol⁻¹ lower when Mg is dehydrated and bound to RCOO⁻; 28). We propose that this sequence creates a thin dolomite template for further precipitation under continued supersaturated conditions (e.g., 9).

Table 2.

Carboxyl character of typical bacteria, organic matter, and polystyrene spheres

Organic material	Site concentration, mole g−1	Carboxyl-site density (group Ǻ-2)
Microspheres (0.82 μm)	8.0 × 10−4	7.0 × 10−1
Microspheres (20.3 μm)	3.8 × 10−4	1.0 × 101
Fulvic acid (NOM)	3.37 × 10−3 (41)	8.1 × 100
Average bacteria	7.7 × 10−4 (33)	6.0 × 10−2
Humic acid (purified peat)	2.3 × 10−3 (42)	5.1 × 10−2
EPS (Desulfovibrio sp.)	1.64 × 10−3 to 2.39 × 10−3 (43)	2.3 × 10−2 to 3.4 × 10−2

Carboxyl-site concentrations, in mole g⁻¹, for organic materials and carboxyl-site densities. Site densities >1 likely represent carboxyl-associated groups that do not terminate at the surface.

In the MSW and SSW experiments, the differing sulfate content did not significantly impact precipitation of Mg-bearing carbonates. This is consistent with other experimental findings (14) that demonstrate that sulfate inhibition (1) is not solely responsible for slow dolomite precipitation at low temperature. Greater effect from carboxyl groups than sulfate is not surprising considering the association between Mg:R-COO⁻ is more favorable than Mg:SO₄ based on aqueous complex data [log K_MgSO4(aq) = 2.4; log K_C10H16MgN2O8 = 2.8; 29].

Our data indicate that the precipitated mineral suite is a function of solution composition, whereas surface carboxyl-group density dictated the kinetics of precipitation, with 10 sites Å^-2 spheres (20.3 μm) exhibiting slightly faster precipitation kinetics (Fig. 1), and, in general, more crystalline Mg-bearing phases. Although the presence of carboxylated surfaces facilitates precipitation of dolomite nuclei at low temperature, solution chemistry fundamentally controls the reactions. Mg:Ca ratios less than 2, such as those thought to have existed in Cambrian-through-Devonian and Cretaceous seawater (21; and modeled with SSW solution in this study), promote calcite and dolomite precipitation. Mg:Ca ratios greater than 2, such as those thought to have existed in the Permo-Carboniferous, Triassic, and Neogene seawater (21; and modeled with MSW solution in this study) promote aragonite precipitates.

Implications for Dolomite Formation at Low Temperature.

We have identified a viable abiotic mechanism for primary dolomite nucleation that relies on the carboxyl-group density of surfaces as a means to overcome the limitation of Mg dehydration and that such surfaces contribute to stabilization of nanocrystals, limiting dissolution, and promoting crystallinity. Organic matter is a typical constituent of carbonate rocks and a potential source of carboxylated surfaces that can bind and dehydrate dissolved Mg ions (30) and facilitate dolomitization (Table 1).

There is a common association of low-temperature dolomite precipitation with environments rich in organic matter. Such organogenic dolomites are commonly proposed to form in association with sulfate-reducing and methanogenic metabolism having modified the seawater pore fluid chemistry (31), yet such metabolism does not consistently result in dolomite. Others propose that exopolymeric substances (EPS) are necessary nucleation sites for dolomite (17). These studies suggest that EPS serves as a site for heterogeneous nucleation of dolomite, but may also concentrate solution geochemistry favoring supersaturation due to limitation of diffusion. Our data support a surface-nucleated mechanism for EPS in the presence of supersaturated solutions and provide a quantifiable mechanism for the data and observed relationships between dolomite and EPS reported in the literature. Because our model does not require metabolic activity, only solution supersaturation achieved by any means, it extends beyond microbial surfaces as we propose that any type of organic matter with sufficient carboxyl-site density is capable of seeding dolomite precipitation.

Organic matter with highly carboxylated surfaces is common but not ubiquitous in marine sediments. Degraded natural organic matter (NOM) with a high fulvic acid character and certain microbial surfaces are known to have high carboxyl-group densities (Table 2). Thus, understanding the controls on distribution of organic matter with high carboxyl-group densities in marine sedimentary settings may provide the key to identifying sediments that have the potential to be dolomitized at low temperature. For example, sedimentary settings with microbial activity promoting rapid cell turnover or death, such as a change from sulfate reduction to methanogenesis (32), can result in an order of magnitude increase in surface site density (33) on organic remains. Mixing zone environments (34) and changing-salinity environments such as sabkhas (35) may promote dolomitization via carboxyl-group density increases resulting from salinity changes (36). It is also possible that the delivery of weathering products may enhance the formation of high-carboxyl organic matter. For example, the presence of iron oxides and clay minerals may concentrate carboxyl-rich organic carbon due to sorption (37).

Abiotic precipitation of low-temperature dolomite reported here clarifies geology's “dolomite problem,” and is not only consistent with, but identifies, mechanisms operative in previous studies implicating microorganisms in this process. The results show that the changing chemistry of seawater through time has had a predictable effect on the distribution of dolomite and other marine minerals. It demonstrates the need to concentrate organic surface functional groups as seeds for dolomite nucleation, which would ultimately form the nuclei for more extensive dolomitization of carbonate rocks and sediments during later times of fluid migration. It may be that environments that lead to accelerated cell turnover or death, changing or high salinity, or certain settings of weathering product delivery may be the ones where highly carboxylated surfaces are generated. Once the controls on those settings are understood, then the first step in the process of dolomitization will be available to model the ultimate controls on dolomitization. Success in developing an experimental regime for low-temperature synthesis of ordered dolomite opens up a wide realm of opportunities for controlled experiments for calibrating interpretations of conditions of dolomite precipitation for much of the geologic record.

Materials and Methods

Batch Reactors.

We studied precipitation of carbonate mineral phases in replicate laboratory batch experiments (13) containing solutions representative of modern and Silurian seawater compositions, which span a range in Mg and sulfate concentrations thought to impact precipitated carbonate mineralogy. MSW solution compositions (pH = 8.2; Mg:Ca = 5.1; [SO₄^2-] = 28 mmol L⁻¹) represent conditions that favor aragonite and high-Mg calcite precipitation (32), whereas SSW solutions (pH = 8.2; Mg:Ca = 1.4; [SO₄^2-] = 11 mmol L⁻¹) broadly simulate conditions in which calcite is favored (38). Stock powders of NaCl, Na₂CO₃, Na₂SO₄, MgCl₂, and CaCl₂ were mixed with distilled water and the pH of the solution was adjusted to 8.2. Solutions were then sterilized and dispensed into 60-mL serum bottles. Salinity and pCO₂ (speciated from measured pH and titrated alkalinity using PhreeqC; 23) were identical for both solutions (Table 1; 21). Experimental vessels were seeded with polystryrene spheres (Bangs Laboratories, Inc.), one with 0.82-μm diameter and R-COO⁻ density of 796 ueq g⁻¹, the other with 20.3-μm diameter and R-COO⁻ density of 380 ueq g⁻¹ (with group spacings of 1.4 Å group⁻¹ and 0.1 Å group⁻¹, respectively). Experimental vessels contained a bulk concentration of ∼10¹² R-COO⁻ L⁻¹. The experimental vessels were capped and solution chemistry was allowed to evolve with gentle agitation in the dark at 30 °C for 20 d. Replicate experimental vessels were analyzed at 5, 10, and 20 d. Control vessels contained no polystyrene spheres and were analyzed at 20 d.

Aqueous Geochemistry.

Beginning solution formulations are given in Table 1. Replicate vessels were killed and analyzed at 5, 10, and 20 d. Solutions were analyzed for pH, alkalinity (39), and cations via inductively coupled plasma. Control vessels were analyzed at 20 d. Solids were filtered, rinsed with deionized water, and freeze-dried for 24 h. Saturation state (Ω = ion activity product/equilibrium constant) was calculated and dissolved inorganic carbon was speciated using PhreeqC (23).

Characterization of Solids.

Precipitates were rinsed with ultrapure water, air-dried, and analyzed with XRD using a Bruker SMART APEX single-crystal diffractometer equipped with a Bruker MicroSTAR high-brilliance microfocus Cu rotating anode X-ray generator, a graphite monchromator, MonoCap collimator, and a SMART APEX charge-coupled device area detector in the University of Kansas Small-Molecule X-Ray Crystallography Laboratory. Subsamples were leached with 1% acetic acid for 1 min (24), then rinsed with distilled water, air-dried, and reanalyzed to better isolate Mg-rich phases and dolomite superstructure reflections. Subsamples for SEM were dried, stub-mounted, and gold sputter coated for 1 min and examined using an LEO 1550 Field Emission SEM equipped with an energy-dispersive spectroscopy detector (EDS; EDAX Phoenix detector; Fig. 4). Samples for HRTEM were dried and mounted on lacy amorphous carbon with copper grid supports and analyzed using an FEI Tecnai F20 X-Twin Field Emission Transmission Electron Microscope equipped with EDX detector. High-resolution images of nanocrystals were performed in standard transmission mode at 200kV whereas diffraction reflection imaging of these crystals was performed by placing the machine in STEM mode (scanning transmission electron micrscopy; diffraction), tuning the rhonchigram over an adjacent region of amorphous carbon near the crystal, and then directly placing the beam (<10-nm diameter) over the crystal's location as chosen from the STEM previewed image (Fig. 6). The beam itself was small enough that the subspecimen image plane aperture was not necessary. In this manner, extremely precisely sampled electron diffraction patterns could be captured from the exact location where EDS would be later performed. Dark-field STEM images were collected using a high-angle annular dark-field detector, set at a working distance of 222 mm, so that crystalline phases were visible. Lattice fringes were measured using Image J version 1.46r (44) and d-spacings were determined using Single Crystal v. 2 in Crystal Maker (CrystalMaker Software Limited).