Potential Function of Microbial Mats in Regard to Water Chemistry and Carbonate Precipitation in the Alkaline Waterbody Lake Van (Turkey)

Abstract

In this article, we examine water chemistry and carbonate precipitation in the alkaline waterbody Lake Van in Turkey, analyzing the possible role of microbial communities in stromatolite formation. Lake Van represents a unique environment characterized by high salinity and pH and extensive microbial communities, as revealed by SEM observation. Microbial activity, including that of cyanobacteria, can influence carbonate precipitation processes, leading to the formation of authigenic carbonates through physicochemical or metabolic mechanisms such as photosynthesis or sulfate reduction. In these environments, which are often dominated by cyanobacteria, carbonate precipitation can be influenced by biologically induced processes. This study presents new data on the hydrochemistry of lake water, focusing on the behavior of rare-earth elements (REEs) in this water and the carbon and oxygen isotopic compositions of carbonate microbial mats. The oxygen isotope data suggest that inorganic carbonate precipitation is the dominant process, but a biological influence on inorganic precipitation cannot be ruled out. For a deeper understanding of the role of biological processes in Lake Van, further studies on microbialites are needed.

1. Introduction

Alkaline lakes are unique environments where the charge balance between carbonate ions and bicarbonate is achieved by cations, such as Na and K as well as Ca and Mg (e.g., [1,2]). These lakes exhibit high pH values of up to 10 [3,4,5]. Alkaline lakes are also known to contain vast microbial communities, leading to the widespread development of microbial mats through a mix of processes, which can be biologically induced, controlled, or influenced [6]. Initially, microbial activity governs mineral nucleation, growth, and morphology, simultaneously creating the metabolic conditions necessary for carbonate precipitation.

In biologically influenced processes, organic matter passively mineralizes as external conditions trigger the precipitation of mineral phases that conform to the shapes of the organic components (as reviewed in [6] and its references). Dupraz et al. [6] coined the term “organomineralization” to describe how an organic matrix of extracellular polymeric substances (EPSs), microbial metabolism, and environmental factors impacting calcium carbonate saturation can collectively promote carbonate nucleation.

Broadly, authigenic carbonates can form due to physicochemical or metabolic processes such as photosynthesis or sulfate reduction. Conversely, fermentation, sulfide oxidation, and aerobic heterotrophy are known to promote carbonate dissolution [7]. Hydrothermal activity can occasionally encourage the precipitation of carbonate in alkaline lakes [8,9].

As the world’s largest soda lake, Lake Van exemplifies active hydrothermal processes within a lacustrine setting. It features high salinity, an alkalinity reaching pH 10, and elevated phosphate levels [10], thereby supporting the planet’s most extensive microbialite formations [11].

These formations emerge near the shore where calcium-rich, neutral groundwater combines with the lake’s calcium-poor alkaline water, leading to significant carbonate oversaturation and precipitation and the creation of a milky solution. Microbialites are remarkable carbonate pinnacles (Figure 1c–e), reaching up to 40 m in height and featuring hollow interiors. Their outer surfaces are covered by coccoid cyanobacteria mats, which undergo in situ permineralization by aragonite and calcite [12]. Inside the pinnacles, spring water flushes the internal walls, which are coated with carbonate globules displaying alternating light and dark concentric layers of aragonite and Mg/Mn/Fe-enriched calcite [12]. These nanobacteria-like carbonate globules also appear on the exterior of the pinnacles, embedded within an organic-rich aragonite and Mg–silica matrix [12]. Lòpez-Garcìa et al. [12] performed phylogenetic analyses to identify bacterial diversity within Lake Van carbonates, both on the surface and internally. Their findings revealed a substantial presence of cyanobacteria, along with alkaliphilic and/or heterotrophic bacteria capable of degrading complex organic matter, indicating their crucial role in microbialite formation.

The aim of this study is to provide additional geochemical evidence of the connection between the biotic and abiotic processes that lead to the precipitation of carbonate phases in the lake. This evidence will include new information on the hydrochemistry of lake water, the fabric and O and C isotopic compositions of microbial mats, and the behavior of REEs in water.

2. Geochemical Background

The rare-earth elements all possess the same external electronic configuration ([Xe]4f_n5d₁6s₂; n variable from 0 to 14). The gradual filling of the 4f orbital causes a decrease in their ionic radii (lanthanide contraction), which changes the reactivity of the elements in the series.

In seawater, the aqueous state of REEs determines the stability of their complexes and of autogenesis processes involving the dissolved species; here, REEs are primarily complexed by [CO₃]²⁻ complexes [13,14,15], and their stability increases along the series as a consequence of lanthanide contraction ([16] and references therein).

Normalizing REE concentrations to a standard shale composition eliminates the Oddo–Harkin rule’s even–odd effect, revealing an increasing abundance of carbonate REE complexes from lanthanum (La) to lutetium (Lu).

The inherent instability of Ce³⁺ and its subsequent oxidation to insoluble CeO₂ result in cerium (Ce) depletion, manifesting as a negative anomaly in seawater [17]. The characteristic REE pattern in shale-normalized seawater typically shows an upward trend, which is interrupted by this negative Ce anomaly (Ce/Ce*) [18].

REE patterns can be altered by external contributions, such as the dissolution of atmospheric solids, the removal of dissolved REEs via authigenic mineral crystallization within the water column, or scavenging by suspended particles [19,20,21]. Observing shifts in REE patterns provides insight into the nature of these neogenic processes.

Typically, enrichments or depletions in REEs are evaluated as positive/negative anomalies along their series according to the normalization expressed by the following equation:

$${\displaystyle \frac{{\left[REE\right]}_i}{{\left[REE\right]}_i^{*}}}={\displaystyle \frac{2{\left[REE\right]}_i}{{\left[REE\right]}_{(i+1)}+{\left[REE\right]}_{(i-1)}}}$$

(1)

where the concentrations in square brackets are normalized [22].

In primary processes like new-magma formation and subsequent igneous mineral crystallization, yttrium displays substantial geochemical similarities to heavy rare-earth elements (HREEs)—particularly those from gadolinium (Gd) to lutetium (Lu), with holmium (Ho) being a notable example [23]—due to their highly comparable charges and ionic radii in solid phases. Bau [24] defined “CHARAC (CHarge and RAdius Controlled) processes” as chemical reactions that do not lead to fractionation among Y, Ho, and HREEs. Conversely, the aforementioned similarities between Y and Ho diminish in non-CHARAC processes, such as aqueous reactions, where the electronic configurations of these elements become critical, given that Y³⁺ has a [Kr]4d0 configuration and Ho³⁺ has a [Xe]4f10 configuration [20,21,25].

3. Methods

3.1. Sampling Sites

Lake Van is one of the largest lakes in Turkey, measuring 3713 km², being situated at an altitude of 1646 m, having an average depth of 171 m, and being 430 km in length. The sampling procedures were conducted from 26 to 28 October 2023. Samples were obtained from eleven distinct sites along the coast (Figure 1;

Table 1. Sampling sites and geographic coordinates of studied lake water.

Sample	Location	Longitude (E)	Latitude (N)
ST-01	Tatvan	42.305124°	38.482524°
ST-02	Adabağ	42.452715°	38.637806°
ST-03	Ahlat	42.532471°	38.759970°
ST-04	Adilcevaz	42.795610°	38.786339°
ST-05	Ercis W	43.279248°	38.959671°
ST-06	Ercis E	43.530767°	38.984128°
ST-07	Edremit E	43.311378°	38.431562°
ST-08	Edremit SW	43.180271°	38.382668°
ST-09	Hasbey	42.941726°	38.329994°
ST-10	Bolalan	42.687062°	38.429276°
ST-11	Tatvan E	42.436072°	38.505397°

). A sample of lake water was taken 20 cm below the surface.

3.2. Analytical Methods

A Traceable digital thermometer (±0.1 °C), an Orion 4 Star pH meter (±0.1 U), and an Oaktan pH tester 30 with a Gel-Filled pH electrode and an Orion 4 Star conductivity meter with an Orion conductivity electrode were used to measure the lake water’s temperature, pH, and electrical conductivity in the field (Supplementary Material S1). By using 0.1 M hydrochloric acid and volumetric titration, alkalinity was ascertained.

The water samples were initially filtered using 0.45 m MilliporeMF filters before being collected in LD-PE (low-density polyethylene) bottles, and the aliquot meant for cation measurement was acidified with HNO₃ (Suprapure grade). To determine the main ions, ionic chromatography was performed using a Dionex DX 120 instrument (Thermo Scientific, Monza, Italy). Its measurement accuracy was 5%.

After filtering the water samples for REE analysis in our lab using a Millipore^TM manifold filter (diameter 47 mm, pore size 0.45 µm), we added 5% HNO₃ ultrapure acid solution to bring the pH down to about 2.

The method developed by Raso et al. [26] was used to perform REE tests on the lake water. To facilitate the precipitation of solid Fe(OH)₃, an excess amount of FeCl₃ (1%) solution was added to each sample (1 L), followed by an appropriate amount of NH₄OH (25%) solution to reach a pH of 8. Zr, Hf, and REEs were scavenged on the crystallizing solid’s surface during this process, and they were then filtered out of the residual liquid. For 48 to 72 h, the solution was kept in a closed flask in a stirrer manifold to ensure that the crystallization of Fe(OH) ₃ would be complete. Thus, the recovery process was evaluated by measuring the iron concentration, which was consistently greater than 95%. A membrane filter (Millipore^TM manifold filter, with a diameter of 47 mm and a pore size of 0.45 µm) was used to collect precipitated Fe(OH)₃. Following the dissolution of the solid filtrate in 3 M HCl, the resulting solution was diluted to 1 M with ultrapure water and subjected to an external calibration process for analysis using a Quadrupole-ICP-MS (Agilent 7500 series). Raso et al. provide details on the overall process and a thorough analysis of the corresponding potential analytical mistakes in [26].

Inductively Coupled Plasma Mass Spectrometry (ICP-MS) was used to analyze the treated samples using an Agilent 7500ce device that had a collision cell. Variations in instrumental sensitivity were tracked with 103Rh, 115In, and 185Re. To maximize the number of elements examined and reduce isobaric and molecular ion interferences, the instrument was adjusted every day and used as the manufacturer advised. The interference caused by Ba-oxide and hydroxide on Eu was monitored using a standard solution with a Ba/Eu molar ratio of 100:1. When compared to internal standard solutions, the computed REE concentrations had an accuracy of less than 5%. We measured samples from every bottle for every site.

To determine the equilibrium aqueous speciation of REEs and saturation indices with regard to the primary mineralogical phases of water, Visual MINTEQ software (Version 3.1) was utilized.

δD and δ¹⁸O versus V-SMOW were determined using an online pyrolysis system (TC/EA) with a CF-IRMS (Delta XP, Thermo Bremen, Bremen, Germany) and the CO₂–water equilibration conventional technique with a CF-IRMS (Delta V Plus connected to Gas Bench II); precision was ±0.1‰ for δ¹⁸O and ±1‰ for δD. δ¹³C_TDIC analyses were performed on the chemical and physical stripping of CO₂ by adding 200 μL of 100% H₃PO₄ according to the method described by Capasso et al. [27], with a precision of ±0.15 delta per mille. The isotopic values in carbonate and TDIC were measured using Thermo Delta V Plus mass spectrometer coupled with Gas Bench II.

Oxygen and carbon isotope compositions of the carbonate were determined. About 0.2 mg of each sample was deposited into a vial subjected to helium flow (99.9996 vol%) and then acidified with approximately 200 μL of 100% H₃PO₄ [27]. The results are reported in delta per mille vs. V-PDB with standard deviations of δ¹³C and δ¹⁸O of ±0.2 delta per mille.

During laboratory manipulations, only ultrapure substances were employed. Using an Arium^® small system (Sartorius, Italy), ultrapure water with a resistivity of 18.2 MΩ cm or higher was acquired. We bought hydrochloric acid, ammonia solution, and 65% (w/w) nitric acid from J.T. Baker Chemicals. By gradually diluting the multi-element stock standard solutions produced by DBH, Merck, or CPI International (1000 ± 5 mg L⁻¹) in a 1 M HCl medium, working standard solutions of the elements under study were created every day. Every piece of polyethene, polypropylene, or Teflon labware was used, and every volumetric device’s calibration was confirmed.

Material obtained after filtration onto cellulose nitrate membranes was subjected to scanning electron microscopy (SEM) of suspended particles. A gold-coated aluminum stub was used to mount the dried solids. A LEO 440 SEM with an EDS system OXFORD ISIS Link and Si (Li) PENTAFET detector was used in the SIDERCEM SRL facility (Caltanissetta, Italy) to perform SEM-EDS analyses.

4. Results and Discussion

4.1. General Aspects of Water Chemistry

The chemical compositions of the major elements in the lake water and their physicochemical parameters are given in Supplementary Materials S2. The water temperature of the lake varied between 16.4 and 16.9 °C depending on the date of collection of the samples. While this temperature is very cold in winter, it rises to 25–30 °C in summer. The water samples show electrical conductivity ranging from 17.5 to 26.8 mS cm⁻¹ and a quite constant pH of between 9.67 and 9.72. Figure 2a shows that the water samples fall within the area of alkaline elements, exhibiting a composition similar to that reported by Sasmaz et al. [10], with a major degree of homogeneity, while the dominant anions (Figure 2b) are chloride and bicarbonate/carbonate.

Nearly all of the samples were in equilibrium or oversaturated with respect to aragonite, calcite, and dolomite according to the saturation indexes of the major mineralogical phases given in Figure 3; two samples were saturated in terms of hydromagnesite.

4.2. Isotopic Compositions of Water and Carbonates

The results of the δD and δ¹⁸O analyses (vs. V-SMOW) of the water samples are reported in Supplementary Materials S2. The values fall within a narrow range, being between −3.0 and −13.8‰ and between 0.8 and −2.1‰ for δD and δ¹⁸O, respectively. The data plotted in the δD-δ¹⁸O diagram in Figure 4, along with the Global Meteoric Water Line (GMWL, [28]) and Local Meteoric Water Line (LMWL, [29,30]), fall on the right side of the GMWL, along a typical evaporation pattern, with a slope of 3, slightly distant from that of evaporitic basins [31], falling in the range of 4–6. Lake Van’s salinity (TDS = 25 g/L) is not high enough to foster isotopic fractionation effects, which generally occur at a salinity >35 g/L, even though local effects could not be entirely excluded.

The isotopic composition of total dissolved inorganic carbon (TDIC) in Lake Van appears to be homogeneous in all the samples, showing values of between 3.0‰ and 3.7‰ (vs. PDB—see Supplementary Materials S3). On the contrary, carbonates show greater variability (Supplementary Materials S3). The δ¹⁸O and δ¹³C values measured in the carbonates range from −6.3‰ to +5.8‰ and from −1.3‰ to +6.9‰ (vs. PDB), respectively.

Considering the chemical boundary conditions, the level of oxygen in the carbonates is at isotopic equilibrium with that of water, behaving in accordance with the isotopic exchange equation:

CaC¹⁶O_3(s) + H₂¹⁸O_(l) = CaC¹⁸O_3(s) + H₂¹⁶O_(l)

(2)

According to the calcite–water isotopic fractionation curves proposed by O’Neil et al. [32], our data show an enrichment factor (10³ lnα) of approximately 29.7‰, which yields an equilibrium temperature of 19 °C (Figure 5), compatible with that measured at the time of sampling, suggesting a dominant carbonate inorganic precipitation process.

Regaclayton [33] and Jimenez-Lopez et al. [34] found a 10³ lnα_cl-HCO3⁻ of 0.9‰ for pure calcite at 25 °C in the inorganic precipitation process. Romanek et al. [34] obtained enrichment factors for calcite and aragonite: in the temperature range of 10–40 °C, they were quite constant, being 1.0 and 2.7‰, respectively. Jimenez-Lopez et al. [35] reported a fractionation factor in synthetic magnesian calcite solution (10³ lnα_{Mg-cl-HCO3⁻}) for different Ca/Mg ratios and pH values ranging from 8.0 to 9.2 that was equal to 0.82 ± 0.09‰ at 25 °C.

In accordance with Jimenez-Lopez et al.’s fractionation factor [35], because the chemical composition of the synthetic solution in their study is similar to that of Lake Van and adding their 10³ lnα_{Mg-cl-HCO3⁻} at the measured δ¹³C_TDIC for Lake Van, we obtained the theoretical δ¹³C of carbonates, as reported in

Table 2. Values of = δ¹³C measured in TDIC (column A) and carbonate samples (column B). The difference (B-A) in the case of simple inorganic processes should be equal to 0.82 [35].

	A	B	B-A
ST-01	3.47	2.83	−0.64
ST-02	3.48	4.53	1.04
ST-03	3.18	2.94	−0.24
ST-05	3.36	6.47	3.11
ST-06	3.60	4.13	0.53
ST-07	3.59	−1.31	−4.90
ST-08	2.99	0.39	−2.60
ST-09	3.65	2.77	−0.88
ST-10	3.43	−0.26	−3.69
ST-11	3.64	6.90	3.27

.

Comparing the measured and calculated δ¹³C values for carbonates, Table 2 indicates that only samples ST-02 and ST-06 are compatible with pure inorganic processes, whereas appreciable differences can be found for the other samples (Figure 6). These differences can be explained by considering the presence of microorganisms, which can alter the isotopic equilibrium with the dissolved bicarbonate, as already evidenced by McCormack et al. ([36] and references therein).

4.3. SEM Observations Made Regarding the Carbonates

The hypothesis of microbial involvement in carbonate precipitation is strengthened by our SEM-based observations (Figure 7a–e) of suspended materials, which reveal authigenic carbonates associated with abundant biogenic detritus, including lithic fragments, cyanobacteria, and diatoms. Furthermore, Lake Van hosts microbialites whose surfaces are colonized by coccoid cyanobacteria mats and permineralized by aragonite and calcite and whose internal structures contain carbonate globules (Figure 7f) associated with cyanobacteria (Figure 7c,e) and alkaliphilic/heterotrophic bacteria involved in organic matter degradation, playing a key role in microbialite formation. Therefore, while the evidence points towards significant inorganic precipitation, the isotopic discrepancies and microscopic observations indicate that biologically influenced carbonate precipitation cannot be excluded.

4.4. REE Behavior in Lake Water

The concentrations of REE in the lake water samples (reported in Supplementary Materials S4) range between 208 and 1358 ng L⁻¹ (La to Lu). The distribution of REE speciation abundance can be approximated using the equation for the total REE content; the quantities of dissolved Cl⁻, F⁻, SO₄²⁻, and carbonate species; and the chemical–physical conditions that exist in the lake:

$$\left[{REE}_i{L}_n^{m\pm }\right]=\left[{REE}_i\right]\times \left[L_n^{m\pm }\right]\times {\beta }_{L_n^{m\pm }}^{{REE}_i}$$

(3)

where [REE_i] indicates the concentration of the ith REE element, $\left[L_n^{m\pm }\right]$ denotes the ligand concentration (Cl⁻, CO₃²⁻....), and ${\beta }_{L_n^{m\pm }}^{{REE}_i}$ is the stability constant of the complex. The distribution of REE species in Lake Van water could be calculated by using compilations of the stability constants of REE complexes [37,38,39,40], even though the stability constants utilized for the computations are given at an ionic strength of around 0 at 25 °C.

In the lake water, we discovered that the complex ion abundance shows the following diminishing sequence: [REE(CO₃)₂]⁻ >> [REE(CO₃)]⁺. A prior study by Sasmaz et al. [10] discovered a similar pattern.

Consistent with Zr and Hf’s higher affinity for biological surfaces, their respective contents range from 250 to 1113 and 1.395 to 3.849 ng L⁻¹ [41]. Zr/Hf and Y/Ho had mean values of 234 and 35.14, respectively. Zr is complexed as [Zr(OH)₄] and [Zr(OH)₅]⁻.

The REE distribution patterns of the lake water are shown in Figure 8. The slight differences between the diverse distribution patterns could be related to possible anthropogenic contributions (like in ST-1, close to the city center of Tatvan), mixing with thermal water (ST-2), or mixing with surface water in rivers flowing from volcanic areas (ST-5 and ST-6). The lake water samples show ascending patterns leading to enrichment in HREEs (from Ho to Lu). A similar pattern was found in stromatolites growing in the alkaline lake Specchio di Venere on the volcanic island Pantelleria [42]; it was attributed to the role of microbes and EPS in stromatolite formation. Eu/Eu* ratios ranging from 0.88 to 1.38 and Ce/Ce* ratios ranging from 0.73 to 1.56 were found. Neither ratio range shows evidence of any significant anomalies; a previous study [10] found higher Eu/Eu* values of up to 4.8, but these were found in thermal water.

The observed increase in concentrations along the lanthanide series could be related to the stability constant values of di-carbonates and carbonate REE–complexes [15,43] able to produce [REE(CO₃)₂]⁻ complexes [44,45,46,47]. This interpretation is consistent with that previously reported by Sasmaz et al. [10]. Alternatively, the HREE enrichment could be due to the presence of microbes. According to Takahashi et al. [48], certain bacteria may have binding sites on their cell surfaces with numerous phosphate groups that preferentially bind HREE.

5. Conclusions

According to the results of petrographic, mineralogical, geochemical, and isotopic studies, it is reasonable to hypothesize that inorganic processes are dominant in the formation of the giant microbialites at Lake Van. The high saturation indexes of carbonate species, the δ¹³C values of dissolved C, and the SEM observations suggest inorganic precipitation. However, the SEM images of the suspended materials in the lake water show authigenic carbonate associated with abundant biogenic detritus and lithic fragments. Cyanobacteria and diatoms are abundant in the lake and associated with carbonates. Lake Van harbors microbialites whose external surfaces are covered by coccoid cyanobacteria mats, in situ-permineralized by aragonite and calcite, and internally covered by carbonate globules resembling nanobacteria comprising cyanobacteria and heterotrophic bacteria in addition to alkaliphilic bacteria that can degrade complex organics and thus play an essential role in microbialite genesis. Organic precipitation could not be entirely excluded. Further investigation of the REE distribution on microbialites could give more insights into the role of biological processes in their formation.