Introduction

Understanding the type and rate of reactions between minerals and fluids is important for many applications, ranging from subsurface carbon storage to scaling of membranes in water treatment. Mineral dissolution or growth rates are obtained from laboratory experimental research, whereby well-defined, fairly simple chemical solutions are generally used. However, applying and upscaling those results for predictions in natural settings is challenging due to the chemical complexity, heterogeneity and hydrodynamic conditions of natural systems. Notably, the presence, concentration and type of background ions in solution can impact mineral dissolution and growth rates, as well as mineral polymorph type (Ruiz-Agudo and Putnis 2012). For example, several studies have shown that the presence of Mg2+ or Fe2+ in solution inhibits calcite growth (Astilleros et al. 2010; Di Lorenzo et al. 2017). This decrease in calcite growth rate can be explained based on a solid solution system, the incorporation of impurities in the crystal, its related change in the crystal surface structure, and a decrease in thermodynamic driving force for growth (Astilleros et al. 2010). The growth of a crystal is a complex process of several steps, with each step having an intrinsic activation energy (Di Lorenzo et al. 2017). The first step for the incorporation of a unit in the crystal structure involves the adsorption of cations onto the crystal surface, implying a coordinate bond formation with the crystal surface and alteration of the hydration shell. Then, the adsorbed cation continues to loose water molecules until its coordination is fully replaced by crystal units and the bonds obtain equilibrium values for the specific crystal structure. The adsorption of water molecules on dolomite differs with impurities in the crystals, with Fe weakening the adsorption capacity of water onto dolomite and Al enhancing it (Ao et al. 2022). Moreover, nucleation experiments of calcite in a solution containing also Fe2+ showed that the amount of Fe2+ has a key impact on the induction time and the maximum free calcium (Di Lorenzo et al. 2017).

Dolomitisation is one of the most important diagenetic processes in carbonate rocks, with significant implications for subsurface reservoirs and geological carbon dioxide storage. Notably, dolomite can store more carbon per unit volume, and it is more resistant to acidic corrosion, than calcite. Many earth scientists have been intrigued by the lack of abundant dolomite (CaMg(CO3)2) besides CaCO3 in recent marine sedimentary systems, in contrast to ancient systems on Earth, despite the thermodynamic favourable conditions for dolomite formation (Warren 2000). Slow reaction kinetics impede dolomite formation (Sibley et al. 1987), and one of the most accepted causes is the strong hydration enthalpy of magnesium ions (Brady et al. 1996). In addition, low activity of dissolved carbonate relative to bicarbonate anions has been proposed as a factor that limits dolomite formation (Petrash et al. 2017). Still, given protodolomite, also called very high magnesium calcite, can readily form, additional kinetic factors linked to cation ordering in dolomite must also play a role (Meister et al. 2023). These kinetic factors could be linked to solid–liquid interfacial tension, surface charge and entropic effects, which may all be influenced by the solution composition and thus the specific ions in the solution.

Different background ions, and more specifically, their charge balance, through alkalinity, affect the fluid pH and carbonate speciation, and fluid pH impacts chemical reactions. Classical crystal growth theory links growth rate to degree of supersaturation. Still, the growth rate may also be affected via the Ph, since pH results in a change of the solution speciation, and thus the degree of supersaturation (Ruiz-Agudo and Putnis 2012), and a higher pH decreases the dolomite surface charge (Pokrovsky et al. 1999). In addition, the growth rate may be affected via the ionic strength of the solution, since background inorganic electrolytes can influence solubility due to strong long-range electric fields caused by the background ions reducing supersaturation and by the effect of cation hydration–dehydration dynamics governed by the strength of ion–water and ion–ion interactions, and thus entropic effects (Ruiz-Agudo and Putnis 2012), and a higher ionic strength amplifies the effect of pH on dolomite surface charge (Pokrovsky et al. 1999). Moreover, studies on surface energy and wettability of dolomite have indicated that interfacial tension can be reduced by oxalic acid-based and citric acid-based deep eutectic solvents, with the former outperforming the latter (Sanati et al. 2022). In addition, sulfate ions in water can alter the wettability of dolomite and enhance water adsorption (Safavi et al. 2020). An alkaline fluid pH has been conventionally suggested as being favourable for the formation of carbonate minerals, including dolomite (Arvidson and Mackenzie 1999; Warren 2000), since the carbonate ion is most abundant at alkaline conditions, whereas the bicarbonate ion and carbonic acid are most abundant at neutral and acidic conditions, respectively. Accordingly, naturally occurring carbonate has been explained by formation at alkaline pH conditions, for example, dolomite in sabkhas, playa lakes and hypersaline lagoons (Petrash et al. 2017), or environments with increased alkalinity driven by photosynthesis of cyanobacteria (Martinez et al. 2016). Dolomite is also found in environments with sulfate-reducing bacteria which consume organic acids (Braissant et al. 2007). However, sulfate reduction does not necessarily increase carbonate supersaturation, since it is coupled to H2S oxidation with O2 which produces acidity, as explained for calcium carbonate saturation in cyanobacterial mats (Aloisi 2008; Meister 2014). Some ephemeral lake environments (not linked to seawater) are highly alkaline and show ordered dolomite formation, as documented for e.g., the Deep Springs Lake and Lake Van, which is explained by fluctuations in conditions linked with microbially influenced processes (McCormack et al. 2018) and rate-limited growth model (Meister et al. 2011) related to favouring the stable phase (dolomite) by leaching the metastable phases (Deelman 1999). A similar interpretation was presented for microbial dolomite formation in Lake Neusiedl (Fussmann et al. 2020). Dolomite ripening under oscillating conditions, following Ostwald’s step rule, was originally derived from experimental synthesis results (Deelman 1999; Liebermann 1967). Still, similar experiments with pH cycling led to the formation of magnesite in other studies (dos Anjos et al. 2011; Vandeginste 2021). In addition, the kinetic barrier that is specific to the formation of dolomite, thus the cation ordening, could be of entropic nature, and be related to the mineral surface, in particular the interfacial energy of a growing particle, which can be affected by oscillating environmental conditions (Meister et al. 2023). Furthermore, it has been suggested that alternating dolomite and chert beds may originate from pH changes related to CO2 production by biogenic respiration in lake water and decomposition of organic matter, resulting in pH decrease and dissolution of calcite and precipitation of silica (Kuma et al. 2019). Moreover, the complex nanoscale morphology of biomorphs, consisting of carbonate nanorods interwoven by amorphous silica, is controlled by local pH and supersaturation in an autocatalytic process (Garcia-Ruiz et al. 2009) in solution bulk pH of around 10 to 11 (Eiblmeier et al. 2013).

Alkaline, clay-rich soils have also been described as favourable environments for dolomite formation. Clay minerals are hydrous aluminium phyllosilicates, and thus aluminium is a key chemical element in clay environments. Moreover, previous research suggested that clay minerals may provide a source for magnesium to form dolomite (McHargue and Price 1982). However, more recent studies indicated that diagenetic Mg release from clay minerals cannot account for large scale dolomitisation, and that a seawater origin for magnesium is more likely (Li et al. 2016). The effect of aluminium on dolomitisation was not previously investigated, but there are several studies that show a favourable effect of clay on dolomite formation. Montmorillonite (a type of smectite) can be replaced by dolomite in soils of alkaline lakes (Casado et al. 2014). Dolomite is also often associated with sepiolite and palygorskite in alkaline lakes (Wanas and Sallam 2016; Wang et al. 2017). The presence of smectite clay also favours formation of dolomite in deeper soil levels in semi-arid regions, because water retained by the clay during evapotranspiration changes the local environment and transport properties, which enhance Mg incorporation in the calcite-to-dolomite replacement process (Diaz-Hernandez et al. 2013). Clay minerals have also been used as template for dolomite precipitation (Wanas and Sallam 2016), and negatively charged illite and smectite acted as catalysts and nucleation sites for dolomite formation (Liu et al. 2019). Similarly, a study on Eocene dolomite in evaporative lacustrine deposits suggests that clay surfaces present nucleation sites for the formation of dolomite at low temperature (Wen et al. 2020). Furthermore, a study on dolocrete in NW Australia indicated a close link between microbial extracellular polymeric substances (EPS) and mineralization of clay minerals and dolomite, with dolomite formation being promoted by the negatively charged surfaces of clay and EPS that help bonding and dehydration of Mg2+ and by the slow diffusion of ions through clay or EPS matrix (Mather et al. 2023). Despite the potential role of clay minerals in primary dolomite formation, there are relatively few studies on dolomite–clay mineral systems (Cai et al. 2021).

There are many examples of dolomite formation in alkaline environments, as shown above, and few studies that show dolomite formation in slightly acidic environments, even though several studies invoke pH cycling as a mechanism for dolomite formation, as mentioned above. Sulfide oxidation and resultant acid production was proposed as a mechanism for dolomite formation in the hypersaline coastal lagoon Brejo do Espinho (Moreira et al. 2004). In this process, the pore waters become undersaturated with respect to aragonite and high-Mg calcite, and supersaturated with respect to dolomite, and as they precipitate competitively (Arvidson and Mackenzie 1999), dolomite formation could be thermodynamically favoured (Moreira et al. 2004). Dolomitisation of the microbialites in the Ordovician Majiagou Formation in the Southern Ordos Basin is explained through formation by high-temperature acid fluids associated with the maturation of organic matter (Yang et al. 2022). Experiments with extreme halophilic bacteria suggest that dolomitisation of calcite is related to the presence of acidic amino acids and polysaccharides in the extracellular polymeric substances of the bacteria (Wang et al. 2022). However, in those experiments, dolomite formation is promoted with increasing carbonate alkalinity and pH, as carbonate anions can bind with Ca2+ and Mg2+ ions that are attracted to the negatively charged carboxyl groups of acidic amino acids and polysaccharides (Deng et al. 2010; Wang et al. 2022). Finally, our previous research has shown that zinc ions in saline fluids slightly decrease the fluid alkalinity to a pH of 6.7 and accelerate dolomitisation (Vandeginste et al. 2019). This leads to the following research question: what is the impact of different background ions on the rate of dolomitisation? Our goal with this work is to gain more clarity on the underlying mechanisms and factors, such as fluid acidity, ion hydration enthalpy, and chemical species distribution. Hence, this study is a step towards addressing the challenge of complexity and heterogeneity of natural systems.

In this work, we show that dolomitisation is faster in saline fluids that are slightly acidic due to the presence of ions with strong hydration enthalpies, in particular Lewis acids (electron pair acceptors). Based on the results of our experiments we identify and quantify the correlation between hydration enthalpy (amount of energy or heat released upon hydration) of the metal ions in solution and dolomitisation induction time. In addition, we present geochemical modelling that simulates the experimental conditions based on growth model and classical nucleation theory. The new insights suggest a different view for interpretations on the mechanisms, conditions and local environments that increase the rate of dolomitisation. We propose that dolomitisation can become faster in a saline, slightly acidic environment by the presence of background ions.

Materials and methods

Experimental design and materials

We hypothesize that background metal ions in solution can affect the rate of dolomitisation beyond the general solution ionic strength effect (influencing dolomite dissolution kinetics (Xu et al. 2013)), and that the hydration enthalpy of the specific metal species correlates with the dolomitisation rate. To test this hypothesis, four types of batch reactor experiments were carried out in this study. Each experimental series followed the same overall procedure, but used different fluid compositions (Table 1). An aliquot of 200 mg calcite (CaCO3) was added to 15 mL solution at room temperature in a Teflon-lined steel batch reactor (with volume of 25 mL), and the reactor was subsequently placed in a preheated oven at 200 °C. A series of reactors with the same fluid and powder was placed in the oven, as each sample measurement (reaction time step) is based on an individual reactor, that was removed from the oven at a given reaction time. This means that for one experimental series of a total of 24 h reaction time and 1 h interval, a total of 24 batch reactors was used. For triplicate measurements, this means 72 batch reactors. The pH was recorded at room temperature both in the starting fluid and in the fluid for each reaction time step upon cooling. The pH could be calculated by conversion from room temperature to 200 °C, and the predicted pH at 200 °C would be lower than that at room temperature. However, also the neutral pH value would be lower at 200 °C than at room temperature. Therefore, the pH is reported as it was measured at room temperature. A temperature of 200 °C and high salinity (2.00 mol L−1 NaCl or slightly less to obtain fluids of ionic strength of 3.5 in all experiments) were selected to enable comparison with a previous study in which the effect of dissolved zinc was tested (Vandeginste et al. 2019). Moreover, elevated temperature and fluid salinity are common in burial diagenetic settings, where dolomitisation can also take place. The amount of CaCO3 and concentration of MgCl2 and CaCl2 in the fluids was kept the same for all experiments, and thus the dolomite saturation state was the same for the start of each experiment, to enable comparison between the different experiments and evaluation of the effect of background ions. All fluids contain 0.30 MgCl2, 0.20 M CaCl2, with Mg/Ca molar ratio of 1.5, shown to result in stoichiometric dolomite (Kaczmarek and Sibley 2011), and added NaCl up to 2 M, which accelerates the replacement reaction (Baker and Kastner 1981).

Table 1 Overview of fluid starting compositions of all experiments. Concentrations are presented in mol L−1
Full size table

The first type of experiments (in triplicate with 1 h reaction time intervals) use the following solution composition: 0.30 mol L−1 MgCl2, 0.20 mol L−1 CaCl2, 1.55 mol L−1 NaCl and 0.03 mol L−1 AlCl3. Addition of AlCl3 in solution resulted in a decrease of the solution pH, leading to a fluid pH of 3.3. In the second type of experiments, the dissolved aluminium concentration was varied in seven duplicate experiments that were sampled after a time period of 8 h. Each of the solutions in those experiments contain 0.30 mol L−1 MgCl2 and 0.20 mol L−1 CaCl2, but the AlCl3 concentration was varied, and NaCl concentration as well to maintain the solution ionic strength of 3.5, resulting in the following concentrations: (1) 2.00 mol L−1 NaCl, (2) 1.85 mol L−1 NaCl and 0.01 mol L−1 AlCl3, (3) 1.70 mol L−1 NaCl and 0.02 mol L−1 AlCl3, (4) 1.32 mol L−1 NaCl and 0.045 mol L−1 AlCl3, (5) 1.10 mol L−1 NaCl and 0.06 mol L−1 AlCl3, (6) 0.80 mol L−1 NaCl and 0.08 mol L−1 AlCl3, (7) 0.50 mol L−1 NaCl and 0.10 mol L−1 AlCl3. The third type of experiments (in triplicate with 1 h reaction time intervals) tests the effect of solution pH. In this case, no AlCl3 was added to the solution, but the solution pH was decreased to 3.3 (as measured in the first type of experiments) by addition of trace element grade HCl. The experimental series uses a solution composition of 0.30 mol L−1 MgCl2, 0.20 mol L−1 CaCl2, 2.00 mol L−1 NaCl, and minor amount of HCl to reach pH of 3.3. Finally, the fourth type of experiments has a solution composition of 0.30 mol L−1 MgCl2, 0.20 mol L−1 CaCl2, 1.97 mol L−1 NaCl and 0.03 mol L−1 KCl. The starting pH of this solution is 9.3.

All solutions were prepared with milliQ water. CaCO3 was purchased from Acros Organics, NaCl and KCl from VWR Chemicals, HCl from Fisher Chemicals, AlCl3.6H2O from APC Pure, CaCl2 from Merck, and MgCl2 from Alfa Aesar.

Analytical methods

The reaction solids (referring to all solid material present in the reactor at the time of sampling) from each batch reactor were collected by centrifugation and three rinsing steps with milliQ water and centrifugation at high speed to remove any potential NaCl salt from the reaction products. Upon decanting the liquid, the reaction products were then dried in an oven at 50 °C for 3 days. After complete drying, the weight of the reaction products was recorded. The composition of the reaction solids assemblage was analysed by powder X-ray diffraction (XRD), using a PANalytical X’Pert Pro X-ray diffractometer with CuKα radiation at a tube voltage of 40 kV and current of 40 mA. The reaction solids were scanned over a sampling range of 5 to 70 ° 2θ, with a step size of 0.0066° and a scan speed of 0.023 ° 2θ per second. Data were analysed for phase identification by comparison with reference spectra within the PDF4 database from the International Centre for Diffraction Data. Identification and quantification of the different minerals in the reaction products was completed using Profex software with Rietveld analysis. In the results below, “(proto)dolomite” indicates the combined amount of protodolomite (or also called very high magnesium calcite) and dolomite, which have the same main peaks, such as the 104 diffraction peak. The main differences in the diffractograms of dolomite and protodolomite is the lack of the ordering peaks (such as 015 diffraction peak) in protodolomite. The reaction solids were also analysed by Fourier Transform Infrared (FTIR) spectrometry using a Bruker Alpha FTIR spectrometer with OPUS software to record transmittance for the wavenumber range of 400–4000 cm−1. Imaging of the reaction solids has been conducted using a JEOL JSM-7100F Field Emission Gun Scanning Electron Microscope. The Brunauer–Emmett–Teller (BET) surface area of the calcite powder used in the batch reactor experiments was determined by triplicate analysis (on samples of more than 1 g) using a Micromeritics 3Flex instrument with nitrogen gas for surface characterization, and has a value of 8.72 ± 0.04 m2/g.

Modelling

In this study, calcite and dolomite are the only minerals considered and their reaction rates are computed kinetically. Dissolution is modelled following the TST approach (Aagaard and Helgeson 1982):

$$r_{ + } = k_{ + } S\prod\nolimits_{i}^{n} {a_{i}^{\nu + } \exp \left( { - \frac{{{\text{Ea}}_{ + } }}{RT}} \right)} \left( {1 - \frac{{\Delta G_{{\text{R}}} }}{RT}} \right),$$
(1)

and precipitation is modelled using the combination of a simplified second-order BCF (Burton Cabrera and Frank) growth model (Burton et al. 1951) and classical nucleation theory (Pham et al. 2011):

$$r_{ - } = k_{ - } S\prod\nolimits_{i}^{n} {a_{i}^{\nu - } \exp \left( { - \frac{{{\text{Ea}}_{ - } }}{RT}} \right)} (\Omega - 1)^{2} - k_{N} \exp \left( { - \Gamma \left( {\frac{1}{{(T)^{2/3} \ln \Omega }}} \right)^{2} } \right),$$
(2)

where r is reaction rate (mol/s), k is the reaction rate coefficient (mol/m2 s), S is the reactive surface area (m2), a is the activity of species i in the aqueous solution, ν is the reaction order, Ea is the apparent activation energy, R is the gas constant, T is the absolute temperature, ΔGR is the free energy of the reaction, Ω is saturation ratio, kN is the nucleation rate constant, Π is a product operator, and Γ is a constant, where it represents the effects of surface tension, molar volume, and the geometric surface shape factor (Pham et al. 2011). The plus and minus subscripts show the dissolution and precipitation reactions, respectively. In the above equations, the reactive surface area is calculated by.

$$S = \beta Mn,$$
(3)

where n is the number of moles of the mineral, M is molar weight, and β is the specific surface area. PHREEQC v3 (Parkhurst 1995) was used to perform the geochemical modelling. The llnl.dat thermodynamic database included with the PHREEQC v3 (Parkhurst 1995), was used for the aqueous speciation and for minerals. The parameters used for Eqs. (1–3) are summarized in Table 2.

Table 2 Parameters used for equations in geochemical modeling of this study
Full size table

The initial fluid compositions used in PHREEQC modelling are listed in Table 1, and the initial calcite in the simulations was 200 mg in 15 ml solution. The temperature in the simulation was 200 °C.

Results

Dolomitisation reaction and byproduct formation

The experimental results show that the addition of 0.03 mol L−1 AlCl3 to the NaCl–MgCl2–CaCl2 solution causes a low starting pH of 3.3 ± 0.1 due to the hydrolysis of Lewis acid AlCl3 in water. This leads to dissolution of about 20% of the calcite reactant within the first hour based on reaction solids weight of 157 ± 4 mg. Reaction products are detected after 3 h (Fig. 1) based on XRD (Fig. 2a) and FTIR analyses (Fig. 3a). The reaction does not only result in the formation of (proto)dolomite, but up to 30% of the reaction solids is hydrotalcite (Figs. 2a, 4a) with its highest content measured at 8 h reaction time (Fig. 1). Hydrotalcite is a layered double hydroxide, and occurs as a rare natural mineral on Earth (Tao et al. 2019). Hydrotalcite consists of positively charged brucite-like oxide layers [Mg3Al(OH)8], charge-balancing interlayer carbonate anions, and interlayer water (Ishihara et al. 2013). In addition, a minor amount of pseudoboehmite [AlO(OH)], a synthetic pure pseudocrystalline aluminium hydroxide gel (Calvet et al. 1952), is identified based on the XRD patterns of the reaction solids (Figs. 1, 2a), which are similar to that of well-crystallized boehmite but with broadening and different relative intensities of the peaks (Santos et al. 2009; Yang et al. 2012).

Fig. 1
figure 1

Reaction solids calcite, (proto)dolomite, hydrotalcite and pseudoboehmite mineral cumulative percentages as a function of reaction time in the 0.03 mol L−1 AlCl3 experiment

Full size image
Fig. 2
figure 2

Powder X-ray diffractograms of reaction products in a the 0.03 mol L−1 AlCl3 experiment at 10 h reaction time, b the 0.045 mol L−1 AlCl3 at 8 h reaction time and c the 0.10 mol L−1 AlCl3 at 8 h reaction time

Full size image
Fig. 3
figure 3

FTIR spectra of the reaction products in a the 0.03 mol L−1 AlCl3 experiment at the labelled reaction times in hours, b the experiments with the labelled aluminium concentrations for 8 h reaction time, c the HCl experiment with starting fluid pH of 3.3 at the labelled reaction time in hours, and d the 0.03 mol L−1 KCl experiment at the labelled reaction times in hours. Note the selected 650–950 cm−1 interval in a, c and d in contrast to the full 400–4000 cm−1 in b

Full size image
Fig. 4
figure 4

Scanning electron microscope images of reaction solid from a 0.03 mol L−1 AlCl3 experiment at 24 h reaction time, with rhombohedral dolomite crystals and hexagonal shaped platelets of hydrotalcite, b 0.10 mol L−1 AlCl3 experiment at 8 h reaction time with pseudoboehmite, and c HCl pH 3.3 experiment with rhombohedral dolomite crystals

Full size image

The experiments with differing Al concentrations (up to 0.10 mol L−1 AlCl3) in saline fluids indicate that the Al concentration has a significant impact on the fluid starting pH and on the time-dependent reaction solids composition (Fig. 5). For a reaction time of 8 h, the weight of the reaction solids decreases roughly linearly from 191 to 117 mg with increasing Al concentration of up to 0.08 mol L−1 AlCl3, and then a larger decrease down to 29 mg for the 0.10 mol L−1 AlCl3 experiment (Fig. 5a). The experiments with up to 0.06 mol L−1 AlCl3 have a pH of 5.2–5.5, the 0.08 mol L−1 AlCl3 experiment a pH of 4.6, and the 0.10 mol L−1 AlCl3 experiment a pH of 1.5 (Fig. 5b). The 0.045 mol L−1 AlCl3 experiment (at 8 h reaction time) has the highest relative percentage of (proto)dolomite with only minor contents of calcite and pseudoboehmite, and lack of hydrotalcite (Figs. 2b, 5c). The reaction solid of the 0.10 mol L−1 AlCl3 experiment consists entirely of pseudoboehmite (Fig. 4b) based on XRD (Fig. 2c) and FTIR spectra with disappearance of the asymmetric stretching of CO3 band at 1390–1510 cm−1 (Fig. 3b).

Fig. 5
figure 5

Experimental data at 8 h reaction time for dolomitisation with fluids of differing aluminium concentrations. a Reaction solids weight decreases with increasing AlCl3 concentration. b Final fluid pH is similar for AlCl3 concentration up to 0.06 mol L−1 and decreases significantly with even higher AlCl3 concentration. c Relative content of calcite, (proto)dolomite, hydrotalcite and pseudoboehmite in the reaction solids changes with different AlCl3 concentrations in the fluids

Full size image

In the experiments with HCl and those with KCl, no byproducts were detected based on SEM (Fig. 4c), XRD and FTIR (Fig. 3c, d). The HCl experiments have a starting fluid pH of 3.3, and the pH increases to 5.7 ± 0.3 within 1 h, without much dissolution of calcite, based on the reaction solids weight of 197 mg. A much higher starting fluid pH (9.3) is measured in the experiments with 0.03 KCl, and no calcite dissolution phase precedes the interface-coupled dissolution–precipitation of the dolomitisation process.

Dolomitisation rate

The 0.03 mol L−1 AlCl3 experiment shows a dramatic decrease in induction time, from 10.5 to 3.5 h, and in timing of full conversion from calcite to (proto)dolomite, from 23 to 16 h, in comparison with control experiments (Fig. 6a). The (proto)dolomite and hydrotalcite content in the reaction solids increases between 4 and 8 h reaction time, after which the (proto)dolomite content keeps increasing, in contrast to the hydrotalcite content which shows a general slight decrease after 8 h reaction time (Fig. 1). The dolomitisation rate increases with aluminium concentration up to 0.045 mol L−1 AlCl3 (Fig. 5c). However, the experiments with concentrations of 0.06 to 0.10 mol L−1 AlCl3 lead to the formation of pseudoboehmite (Figs. 2c, 3, 4b, 5c).

Fig. 6
figure 6

a Calcite-to-dolomite replacement reaction curves presenting (proto)dolomite on total carbonate percentage against reaction time. The plot presents experimental data from this study for the 0.03 mol L−1 AlCl3, 0.03 mol L−1 KCl and HCl added to reach pH 3.3, whereas the control experiment and 0.03 mol L−1 ZnCl2 experimental data are from our previous work (Vandeginste et al. 2019). b Avrami plot for dolomitisation in the different experiments with linear best fit trend lines to derive time exponent and rate constant

Full size image

The HCl (pH 3.3) experiments show an induction time of 6.5 h and full conversion time of 18 h (Fig. 6a) based on XRD and FTIR analysis (Fig. 3c) with the bands at 714 and 728 cm−1 being assigned to ν4 in-plane bending of CO3 in calcite and dolomite, respectively (Wang et al. 2019). The 0.03 mol L−1 KCl experiments indicate an induction time of 13.5 h, which is longer than that of the control experiments (10.5), and a full conversion time of 24 h (Fig. 6a) based on XRD measurements and FTIR spectra (Fig. 3d). The dolomitisation reaction rates are fitted using the Avrami equation, with reaction extent y = 1 − exp [– (k(t – t0))n] whereby t0 is the induction time, k the rate constant and n the time exponent (Fig. 6b). The Avrami equation simulates the transformation from one phase to another at constant temperature and has been used for fitting the rate of dolomitisation in some previous studies (Rodriguez-Blanco et al. 2015; Xia et al. 2009). The derived reaction rate constants from our results increase in the order of HCl < control (NaCl) < AlCl3 < KCl < ZnO < ZnCl2 experiments, indicating the lack of overall correlation between decreasing induction time (AlCl3 < ZnCl2 = ZnO < HCl < control (NaCl) < KCl) and increasing Avrami rate constant (Table 3).

Table 3 Best fit parameters for the Avrami model of the dolomitisation rates in the different experiments with t0 the induction time, k the rate constant and n the time exponent for the Avrami equation with reaction extent y = 1 – exp [– (k(t – t0))n]
Full size table

Discussion

The presence of background compounds in solution influences the starting pH of the solution, which in turn can influence calcite dissolution and dolomite precipitation. Even though the ionic strength (which affects the activity of species in solution) is kept the same in all solutions, the measurements show that the fluid starting pH is still influenced by the different interaction of ion assemblages with water in the experiments. At alkaline conditions, the carbon speciation increases in the order of CO2 < HCO3− < CO32−, resulting in a high carbonate mineral saturation index, favourable for carbonate formation. According to classical crystal growth theory, the growth rate is related to degree of supersaturation. However, the presence of background ions may induce non-classical crystallization pathways. For example, calcite nucleation and growth rate decrease with increasing alkaline pH from 7.5 to 12 in supersaturated solutions of low ionic strength of 0.1, due to strongly hydrated hydroxyl ions that lead to increased frequency of water exchange with hydrated calcium ions, causing increased solid–liquid interfacial tension and reducing the nucleation rate (Ruiz-Agudo et al. 2011a). The alternating Ca2+ and CO32− ions in calcite generate a polar hydrophilic surface, with which the hydroxyl group of water strongly interacts through electrostatic interaction (between O from hydroxyl and Ca from the calcite surface) and by hydrogen bonding (between H from hydroxyl and O from the calcite surface) (Hakim et al. 2017). These studies show that faster carbonate nucleation and growth is not always linked to higher alkalinity due to interfacial tension effects. The carbon speciation in our experiments was reconstructed through PHREEEQC modelling, and shows a decrease in the order CO2 > HCO3− >  > CO32− in the 0.03 mol L−1 AlCl3 and HCl experiments with starting fluid pH of 3.3, and HCO3− > CO2 >  > CO32− in the 0.03 mol L−1 ZnCl2 experiments with starting pH of 6.7 (Fig. 7). The control and 0.03 mol L−1 KCl experiments with starting pH of 8.5 and 9.3, respectively, have a carbon speciation with HCO3− > CO32− > CO2 (Fig. 7). The modelled pH at the end of the simulation is 4.9, 5.8, 6.9, 8.5 and 9.3 for the 0.03 mol L−1 AlCl3, HCl, 0.03 mol L−1 ZnCl2, control and 0.03 mol L−1 KCl experiments, respectively. Hence, these PHREEQC results based on classical nucleation theory and growth model indicate that CO32− ions are not the dominant carbon species. Notably, the carbon species absolute values show that the CO32− activity in the ZnCl2 experiment (with the fastest complete conversion of calcite to dolomite) is less than half that in the other experiments. Still, we need to consider that the CO32− activity could remain low during the dolomitisation process, where released carbonate ions from calcite dissolution could immediately be incorporated in the precipitating dolomite.

Fig. 7
figure 7

Plots of the PHREEQC modelled carbon species distribution against time, simulating the control and ZnCl2 experiments (Vandeginste et al. 2019), and the 0.03 mol L−1 AlCl3, HCl and KCl experiments presented in this study

Full size image

Our laboratory experiments indicate faster dolomitisation in slightly acidic fluids. The fluid pH decreases in solutions with cations that have a strong hydration enthalpy (i.e., energy or heat released upon hydration of the cations). Acidic fluids can cause faster calcite dissolution, which then leads to a faster increase in calcium and carbonate ions in solution; and the pH was less than 6 at 1 h reaction time in all experiments. Considering a first order reaction rate for the precipitation of dolomite, a faster increase of carbonate ion concentration in solution could explain a faster dolomitisation rate, and PHREEQC modelling results support faster dolomitisation in slightly acidic than in alkaline fluids (Fig. 8). This could be interpreted based on the thermodynamic drive that the solution is undersaturated with respect to calcite and oversaturated with respect to dolomite when Ca2+ comes into solution upon calcite dissolution.

Fig. 8
figure 8

Plots of the PHREEQC modelled number of moles calcite and dolomite against time, simulating the control and ZnCl2 experiments (Vandeginste et al. 2019), and 0.03 mol L−1 AlCl3, HCl and KCl experiments described in this study

Full size image

Furthermore, we compared the effect of AlCl3 versus HCl in solution at the same pH of 3.3. The experimental results demonstrate that the dolomitisation induction time does not only depend on the fluid pH but also on the ion assemblage, with Lewis acid AlCl3 being more effective in shortening the dolomitisation induction time than Brønsted acid HCl (Fig. 9). The PHREEQC modelling results do not completely align here with the experimental data (Fig. 8), and also the shorter induction time for the AlCl3 experiment in comparison with the ZnCl2 experiment is not evident in the models. Given that hydrotalcite and some pseudoboehmite form besides dolomite in the 0.03 mol L−1 AlCl3 experiments, the dolomitisation rate may be more complicated in this case. The general trends in the experimental dolomitisation work are consistent with the PHREEQC results, and given these results are based on classical crystallization theory, we think that the slight discrepancies between the laboratory experimental data and the PHREEQC results are probably caused by specific ion effects, inducing non-classical crystallization pathways.

Fig. 9
figure 9

Plots of the relation between dolomitisation induction time (based on experimental results in this study) with the positive hydration enthalpy values of the cation added to the solution (black squares) and the initial fluid pH (based on the experimental data in this study; red diamonds). The hydration enthalpy of Na+ is used for the control experiment, since the amount of NaCl in solution is decreased in the other experiments for addition of other salts to maintain equal solution ionic strength in all experiments. A power law trend line equation is derived for positive hydration enthalpy values against dolomitisation induction time

Full size image

The effect of specific ions in solution at mineral or other material surfaces are determined by a balance between different forces that are influenced by the characteristics of the ion, the solvent and the surface properties (Moghaddam and Thormann 2019). Therefore, solution composition, mineral surface, surface charge, interfacial tension and entropic effects are all important factors that may influence crystal nucleation and growth rate. Hydration of ions plays thereby a major role, with competition between ion–water (electrostatic) and water–water (hydration) interactions, and also the material surface properties. For example, dolomite growth can be accelerated with carboxylated polystyrene spheres due to the favourable impact of the surface carboxyl functional groups on complexation and dehydration of magnesium ions (Roberts et al. 2013). The results of our study, showing a shorter dolomitisation induction time in fluids with stronger cation hydration enthalpy, are different to the impact of background ions on calcium carbonate nucleation in dilute solutions, where the induction time decreases with increasing ionic radii of background ions (Burgos-Cara et al. 2017), generally correlating with weaker hydration enthalpy. In dilute solutions, an entropic effect related to changes in ordering of water molecules due to arrangement as hydration shells around ions (that build up the mineral) plays a dominant role (Ruiz-Agudo et al. 2011b). In fluids of high ionic strength, there are more interactions between the background ions and water molecules, and thus the competition for hydration water is higher (Ji et al. 2022), and ion hydration becomes the dominant factor controlling kinetics (Ruiz-Agudo et al. 2011b). A study on the effect of background ions on the rate of dolomite dissolution reported a higher dolomite dissolution rate in saline solutions (with ionic strength of 1) with NaCl in comparison with KCl (Ruiz-Agudo et al. 2011b). Since Na+ has a stronger hydration enthalpy and smaller ionic radius than K+, these results imply a faster dolomite dissolution rate in fluids with background ions that have a stronger hydration enthalpy. In solutions of high ionic strength, the relative impact of background ions on the dolomite dissolution rate is determined by water–water interactions and effects associated with the strong hydration of calcium and magnesium ions (Ruiz-Agudo et al. 2011b). The K+ ion weakens the structure of water (due to higher ion mobility), and competition for hydration water between the bulk and the solute ions is lower in a less structured solvent, and therefore, the water exchange frequency will be lower, resulting in slower dolomite dissolution (Ruiz-Agudo et al. 2011b). Our study concerns dolomite formation rather than dolomite dissolution, and still, our results show a similar effect, namely, faster dolomitisation in saline fluids (with even higher ionic strength) with background ions that have a stronger hydration enthalpy. Magnesium ion dehydration and incorporation in the dolomite crystal during dolomitisation is thus more favourable in water composed of background ions with a stronger hydration enthalpy, and thus more structured water. This is interpreted by the fact that removal of magnesium ions from the water may have a small entropic effect on the structured water, and that competition for hydration water is very high in these highly saline fluids, so that that ions with stronger hydration enthalpy in the water would compete for hydration water, and in this way, favour magnesium ion dehydration.

In contrast to induction time, the Avrami derived rate constant for dolomitisation in the different solutions seems to be not only affected by the hydration enthalpy but also by the complexity of the fluid composition and the formation of byproducts in the dolomitisation reaction. Our experiments focused on the effect of positively charged background ions in solution. Negatively charged background ions can also accelerate mineral growth by enhanced ion dehydration. However, this may be counteracted by ion pair formation which lowers mineral supersaturation, as documented, for example, by the inhibiting effect of sulfate on the formation of magnesite (King et al. 2013) and dolomite (Baker and Kastner 1981; Vandeginste et al. 2019).

The rate of geochemical reactions is commonly derived from laboratory experiments, where the experimental conditions are simplified to derive fundamental correlations based on one or a few variable parameters. These laboratory-derived rates then serve as data and input for models that simulate real natural scenarios. However, the combined effect of several parameters together could lead to different results than the simple combination of the effects of the parameters tested separately, for example, by the specific ion effect linked to the presence of background ions. As a consequence, interpretations on natural environments or prediction for geological carbon storage could then be inaccurate. Several studies have highlighted the complexity of natural systems and associated challenges in interpretations. There are discussions around the influence of associated minerals on dissolution processes in carbonate rocks, since additional metal ions, such as Fe3+, in acidic fluids, have shown to change the fluid properties and dissolution behaviour (Ma et al. 2021). A study reports the role of the fluid environment (pH and composition) in determining the surface charge of the interface of minerals and brine, with implications for wettability in reservoirs (Mohammed et al. 2022). Different minerals behave differently in an environment of varying composition and salinity (Mohammed et al. 2022). Fluid inclusion studies have documented thermal and chemical fluid heterogeneity, which is common in large paleohydrothermal systems with recharge of surface-derived fluids (Wilkinson 2010). Understanding and reconstructing the geological history of mineralization is important in ore field studies, for example, Zn–Pb deposits in dolomitized host rocks (Wilkinson 2010). Studies on smart waters for enhanced oil recovery in carbonate reservoirs report that the complexity of the fluids/rock system governs the effects and that the fundamental knowledge is still lacking, and thus many questions and uncertainties remain (Gachuz-Muro et al. 2017). The challenge of complexity of natural systems and resulting inaccurate predictions could be overcome by taking measurements in the field or using the actual formation fluids in laboratory experiments. Still, this offers only a solution case by case, and not a global approach. A more fundamental approach that documents the effect of background ions could help in achieving more realistic predictions for natural scenarios. Previous experiments demonstrate the need to incorporate the specific ion effect in predictive models that simulate mineral formation and dissolution in natural complex systems (Ruiz-Agudo et al. 2011b). Our study contributes in unravelling the effect of some background ions on the rate of dolomitisation.

Conclusion

We have documented that the dolomitisation induction time decreases with stronger hydration enthalpy of background ions in saline solutions of the same ionic strength. This observation is mainly caused by the slightly acidic pH, and by the specific ion effect. These results highlight the importance of fluid complexity or heterogeneity with implications for interpretations of natural environments or predictions for geological carbon storage.