Geochemical Variations of Kerolite, Stevensite, and Saponite from the Pre-Salt Sag Interval of the Santos Basin: An Approach Using Electron Probe Microanalysis

Abstract

This study investigates the mineralogy and chemical characteristics of pre-salt clay minerals, classifies them, and defines assemblages in reactive microsites. Using Electron Probe Micro-Analysis (EPMA), the chemical formulas of Mg-rich clays were determined. Stevensite exhibited low interlayer charge and aluminum content, while kerolite was characterized by a minimal charge. K/S (kerolite/stevensite) mixed layer showed intermediate compositions and charges between these endmembers. Saponite was distinguished by higher levels of Al, K, and Fe, along with a higher interlayer charge. The proposed assemblages are as follows: saponite in mudstone facies (without spherulites/shrubs), with a hybrid matrix; pure kerolite in spherulstone and shrubstone facies, marked by the absence of significant reactions and high preservation of matrix and textures; stevensite in facies with extensive matrix replacement by dolomitization/silicification; and K/S and kerolite in similar facies with intermediate matrix replacement levels and the coexistence of two intimately related clay mineral compositions. This study enables reliable differentiation of these species based on point mineral chemistry and mapping, combined with a microsite approach and conventional techniques. Additionally, it discusses the formation of pre-salt clays, influenced by significant kinetic and chemical interactions during their genesis and burial to depths of approximately 5 km.

1. Introduction

The chemical characterization of magnesium-rich clay minerals poses significant challenges due to the compositional similarities among species and the difficulty in establishing clear boundaries for distinguishing kerolite, stevensite, and saponite. Numerous studies have reported challenges in defining chemical criteria, often requiring complementary techniques for accurate classification of this mineral group [1,2,3,4,5,6,7,8,9,10,11]. These challenges encompass the precise determination of chemical formulas, the application of suitable analytical techniques, and the complexity arising from the coexistence and transformation of these mineral phases.

In order to relate the mineral phases to the geological context of their formation, this work proposes a study approach based on the so-called reactive microsites, which refer to microscale areas where processes and products associated with metastable aggregates are evident. These aggregates are typically rich in fine particles, such as clay minerals, characterized by high reactivity. This approach enables a detailed examination of reaction pathways using microscopy and microanalysis techniques, considering the high susceptibility of clay minerals to reaction due to their large surface area and the intrinsic reactivity resulting from structural disorder.

This methodology is crucial as the studied materials do not reflect equilibrium or homogenization conditions, as seen in metamorphic rocks. Instead, they are governed by kinetic effects over thermodynamic ones. Reaction products can vary significantly within a single sample, depending on the mineral assemblages and the interactions promoting the formation or transformation of other minerals.

Thus, this manuscript aims to distinguish magnesium-rich clay minerals found in rocks collected from the giant oil and gas reservoirs of the Brazilian pre-salt areas through mineral chemistry analyses and chemical mapping via EPMA.

Geologial Setting

The Brazilian pre-salt deposits are predominantly found in the Campos and Santos basins along Brazil’s southeastern margin. They are located along the coastal margin between the Brazilian states of Santa Catarina and Espírito Santo (Figure 1A) and constitute the most productive oil reservoirs of South America. The pre-salt formation dates to the Early Cretaceous and is closely linked to the breakup phases of Western Gondwana. During this period, lacustrine carbonate rocks were deposited under conditions of high alkalinity, marking the transition from a continental to a marine environment [12,13].

The stratigraphy of these basins is illustrated in Figure 1B, based on the model proposed by [14] for the Santos Basin. Both basins share a similar geological evolution, comprising three major megasequences: rift, post-rift, and drift. The rift sequence, associated with the continental phase, is subdivided into the Lower and Upper Rift. The Lower Rift is characterized by basaltic rocks of the Camboriú Formation (134–122 Ma [15]), conglomerates, alluvial sandstones, and lacustrine deposits of the Piçarras Formation. The Upper Rift, correlated with the Jiquiá Stage, includes coquinas and black shales from the Itapema Formation, which is considered an important oil and gas source rock.

The post-rift stage, represented by the Barra Velha Formation, spans the end of the syn-rift and the beginning of the sag phase, marked by the Intra-Alagoas Unconformity, which is similar to the Macabu Formation in the Campos Basin [16]. The Barra Velha Formation, dated to the Aptian, consists of shales, intraclast and hybrid conglomerates, sandstones, dolostones, and limestones (the most abundant rock) [17,18]. Wright and Barnett [19] describe the main constituents of sag phase rocks as a heterogeneous mixture of in situ precipitated and reworked deposits, with a mineralogy rich in carbonates and clays. The origin of different textures in the carbonatic rocks has been debated, with some proposals suggesting microbial origin [20,21,22,23,24], while others indicate chemical precipitation in lacustrine or restricted marine environments [25,26,27,28,29,30,31]. Following these formations, the Ariri Formation consists of evaporites resulting from intense evaporation in lacustrine environments, which gives rise to the “pre-salt” term used for these reservoirs located beneath salt layers. Finally, the drift phase overlays evaporitic rocks with marine carbonatic and siliciclastic sequences [32,33].

Clay minerals are commonly found as the main constituent of the interstitial matrix of pre-salt carbonatic rocks [21,26,34]. A series of nomenclatures have been used for the classification of pre-salt rocks [21], which represent a challenge for homogenization and application. Due to their high susceptibility to fluid-induced modifications, these minerals are prone to dissolution, transformation, and partial or total replacement processes, which contribute to the formation of secondary porosity in the reservoirs [35]. Clay minerals generally form from lacustrine waters in undisturbed environments, creating laminations. Calcite precipitation (spherulites and shrubs) displaces and deforms these original laminations. In other cases, microcrystalline calcite, dolomite, and silica precipitate, which are later products of diagenesis [26,34,36]. Although the genesis and evolution of clay minerals in pre-salt lacustrine rocks are not extensively discussed, the term stevensite or ’stevensitic clays’ is often used in a generic manner without rigorous mineralogical and chemical characterization. Several studies indicate significant occurrences of kerolite, saponite, and interstratified kerolite/stevensite [9,36,37,38,39], along with other sporadic occurrences of clay minerals such as sepiolite, illite, and smectite, especially within the smectite/chlorite stability fields [11,26,32,40]. Although magmatic and/or hydrothermal processes have been described in the sag interval [17,25,27,28], this study focuses on in situ facies (and one reworked facies) free from these processes. While this work focuses on clay-rich, non-reservoir levels, the features and relationships between carbonates and clay minerals are similar, except for the high degree of dissolution in the reservoir rocks [34].

Recent studies on clay minerals in this context highlight the need to differentiate trioctahedral smectites from kerolite, alongside the analytical and conceptual challenges related to this distinction (Da Silva et al., [9]; Netto et al., [37]). This study aims to apply appropriate methodologies to provide accurate geochemical characterizations of magnesian clay species in the pre-salt context. This manuscript will enable subsequent studies to correlate Mg-clays with diagenetic or hydrothermal processes that impact porosity and reservoir quality, thereby improving models.

Figure 1. (A) Location map of the pre-salt polygon distributed across the Santos and Campos Basins. The area of interest is situated in ultra-deep waters, with depths exceeding 5 km. Information modified from [41,42,43]; (B) stratigraphy of the Santos Basin from the Cretaceous to Recent, highlighting the Aptian-aged Barra Velha Formation in red [14,44].

Figure 1. (A) Location map of the pre-salt polygon distributed across the Santos and Campos Basins. The area of interest is situated in ultra-deep waters, with depths exceeding 5 km. Information modified from [41,42,43]; (B) stratigraphy of the Santos Basin from the Cretaceous to Recent, highlighting the Aptian-aged Barra Velha Formation in red [14,44].

2. Materials and Methods

A total of 24 samples from the Barra Velha Formation, Santos Basin, were analyzed. These samples, retrieved from a Petrobras core collection, are rich in clay minerals and represent distinct facies, characterized by low porosity and permeability, rendering them unsuitable as reservoirs. To preserve the integrity of clay minerals, samples were impregnated and polished using diamond paste, resulting in 25 thin sections (30 µm) and 8 thick sections (100 µm) stained with Alizarin Red S dye for calcite identification. This study extends the previous work presented in Da Silva et al. [9] by expanding the sampling to include all core samples, instead of the selective populations from the earlier study. A key advancement is the application of an operational facies classification, which enables a more detailed classification of lithologies and allows for the use of electron probe microanalysis (EPMA) for the chemical quantification and visualization of the processes acting on the proposed mineral assemblages.

Petrographic analyses were conducted at the IGEO-UFRGS laboratories, focusing on the main constituents, textures, and paragenetic relationships in reactive microsites. Facies classification specific to the pre-salt area, as proposed by [34], was employed.

X-ray diffraction (XRD) analysis for mineralogical characterization was performed at the LDRX/IGEO-UFRGS using Siemens diffractometers (models D-5000 θ-θ and θ-2θ) (Bruker AXS GmbH, Karlsruhe, Germany). Analytical conditions included CuKα radiation (1.5406 Å), 40 kV, 30 mA, and scanning ranges of 2.3–72°2θ for whole rock and 2.3–28°2θ for the <2 µm clay fraction, with step sizes of 0.02° and the time per step varying from 1 to 6 s. Total rock analyses were carried out with finely ground powder placed on a non-oriented sample holder. The mineralogical compositions of the samples were semi-quantified using the Reference Intensity Ratio (RIR) method, in which the diffraction peak intensities (I) are scaled and divided by the intensities of diffraction peaks of a standard (I/Ic), with corundum (Al₂O₃) as the reference. The clay fraction was separated by sedimentation (Stokes’ law) and analyzed in preparations that were air-dried, ethylene glycol-saturated, and heated to 550 °C. Methods adapted to distinguish smectite species followed [9] and were based on [45]. No cation saturation was conducted on the clay mineral samples. The crystallite size data are outside the scope of our research and only serve as a qualitative and comparative observation between species.

Secondary electron (SE) images were acquired using a Zeiss EVO MA10 scanning electron microscope (CMM-UFRGS) (ZEISS Group, Oberkochen, Germany). Fragments measuring 1–2 cm, mounted on carbon tape and coated with gold–carbon, were analyzed at 7–10 kV, 10 pA current, and magnifications up to 100,000x.

Quantitative EPMA analyses were performed at the Electron Microprobe Laboratory (CPGq-UFRGS) on carbon-coated thin sections. Analytical conditions included 15 kV, 15 nA, beam size of 1–5 µm, and 60 s X-ray peak and background acquisition times. Calibration standards from the standard block Minerals from SPI Supplies are as follows: diopside (Mg), sanidine (Si, Al, K), albite (Na), wollastonite (Ca), and hematite (Fe). Matrix corrections for quantitative analyses were based on the PAP method [46]) using PeakSight version 5.1 software for both corrections and electron image processing. Elemental composition maps, including Si, Al, Mg, K, and Fe, were generated with WDS spectrometers, using acquisition times of 10–100 ms/pixel and resolutions up to 1 µm, with color scales representing relative concentrations. According to estimates from counting statistics [47], the mean standard deviations for kerolite analysis (in wt%) are Si (0.39), Al (0.03), Mg (0.42), Fe (0.05), Na (0.06), K (0.03), and Ca (0.03); for stevensite, Si (0.4), Al (0.04), Mg (0.33), Fe (0.05), Na (0.16), K (0.04), and Ca (0.02); and for saponite, Si (0.37), Al (0.1), Mg (0.34), Fe (0.11), Na (0.15), K (0.08), and Ca (0.03). The detection limit ranged between 0.02 and 0.05 wt% for all major elements in all samples.

The structural formula of clay minerals was calculated based on the number of atoms per unit cell, following Linus Pauling’s rules [48], as revised by [49]. The principle of electroneutrality was applied to ensure charge balance between cations and anions in the mineral structure. For the FTIR study, eight samples, previously separated in the clay fraction (<2 µm) following the same protocols described above, were selected. The samples were analyzed using infrared spectroscopy with a spectral resolution of 2 cm⁻¹ in transmission mode, covering the 400–4000 cm⁻¹ range. Sample preparation involved KBr pellet formation (3 mg of sample mixed with 300 mg of KBr), and analyses were conducted using a Bruker IFS66v FTIR spectrometer (Bruker AXS GmbH, Karlsruhe, Germany) at the laboratories of the Autonomous University of Madrid, Spain.

3. Results

The results are organized according to the petrography and mineralogical characterization by X-ray diffraction of all core samples, followed by the element point analysis and chemical mapping of the mineral assemblages in the representative samples and their respective microsites. By convention, the samples are identified by the last four digits of their depth in the drill core.

3.1. Mineralogy

The mineralogy of the rocks was characterized through semiquantitative X-ray diffraction analyses, and the results are presented in this section in a way that allows for the comparison of the calcitic, dolomitic, clay minerals, and other silicates, as shown in

Table 1. Semi-quantitative data of bulk rock by X-ray diffractometry by calculating the I/Io—RIR method of the core samples.

Sample Depth (m)	Calcite	Dolomite	Clay Minerals 1	Quartz	Detritics 2
5115.15	42%	27%	14%	17%
5116.35	58%	20%	9%	13%
5119.55	16%	* 58%	3%	22%
5124.40	77%	8%	12%	3%
5125.30	76%	5%	10%	10%
5127.60	37%	5%	13%	45%
5128.40	52%	36%	9%	2%
5129.55	75%	7%	15%	3%
5129.85	83%	5%	12%
5130.85	91%	3%	6%	1%
5132.15	82%	5%	13%
5133.70	59%	30%	11%
5136.90	60%	5%	35%
5137.50	48%	3%	18%	3%	28%
5138.15	85%	5%	10%
5138.50	69%	2%	11%	18%
5140.25	81%	10%	5%	2%
5141.68	81%	13%	4%	2%
5141.75	76%	11%	13%
5141.80	78%	7%	12%	3%
5144.15	52%		14%	1%	33%

* The % shown here represents the sum of dolomite and ankerite + magnesite. ¹ kerolite + smectite. ² feldspar + mica.

. Analyses were conducted across the entire core interval.

Calcite predominates throughout the entire interval, although this semiquantitative method does not allow for the differentiation of the various calcite morphologies, which may be present as spherulites, shrub-like structures, microcrystalline forms, or late cementation. These details are summarized in the following section in the facies classification presented in Figure 2.

The same limitation applies to the clay minerals, further compounded by the additional difficulty in quantifying these minerals due to their highly disordered structure. This characteristic leads to broad and sometimes shifted peaks, in addition to a limited number of reference patterns available for comparison. As a result, clay minerals will be more thoroughly explored in specific fractions (less than 2 µm) and appropriate preparations for the analysis of these minerals. Notably, in samples 37.50 and 44.15, a centimeter-thick layer of chert was avoided in this preparation.

3.2. Clay Minerals and Assemblages

The analyzed rock matrices predominantly consist of clay minerals, which occur either as interstitial material in calcite carbonates (spherulites or shrubs) or as laminations in rocks with microcrystalline calcite. All samples exhibit clay minerals, regardless of facies (Figure 2). Two types of matrices were identified: pure, composed exclusively of authigenic clays, and hybrid, which includes detrital minerals. Out of the 24 analyzed samples, 22 have pure matrices and 2 contain hybrid matrices. The primary detrital grains include quartz, feldspars, micas (biotite, phlogopite, and muscovite), rare heavy minerals such as Fe oxides, organic matter, and bioclasts. In the hybrid samples (37.50 and 44.15), saponites were identified with higher counts and the typical behavior of retaining their expansion capacity with ethylene glycol after heating to 550 °C. These samples constitute the saponite and detrital minerals assemblage, associated with the mudstone/chert facies. Mudstone represents a fine, laminated rock, possibly indicative of a fine detrital input (silt) within a lacustrine environment. These facies undergo extensive replacement by microcrystalline calcite, forming laminations interspersed with preserved fine constituents. Chert lenses crosscut the mudstone facies, containing micro- and macrocrystalline silica and chalcedony, with remnants of calcite nearly unrecognizable.

Clay minerals in the pure matrix form aggregates with beige-brown coloration and slight pleochroism toward yellow. Two types of aggregates were observed: lamellar aggregates with high optical continuity and third-order birefringence resembling talc, and chaotic aggregates with massive internal textures or disorganized particle orientations and lower birefringence. Alongside lamellar aggregates, clayey peloids with similar composition but spherical morphology and massive internal structures were also present.

In samples containing only kerolite, lamellar aggregates predominate. This assemblage was termed kerolite-rich, with samples 32.15 and 36.90 selected for chemical analysis. It occurs predominantly as calcite spherulites or shrubs (spherulstone and shurbstone facies) with the matrix partially replaced by rhombohedral dolomites. In samples containing smectite, a loss of expansion capacity is observed after heating to 550 °C. Some of these samples, rich in smectite, exhibit more resolved peaks with higher counts and are classified as stevensite. These form the stevensite-rich assemblage, with intense dolomitization and silicification, as in samples 15.15 (Muddy Spherulstone facies) and 19.55 (dolomitized mudstone/dolostone); the high dolomite content in this sample, as shown in Table 1, is particularly notable.

Other samples exhibit broader, diffused XRD reflections with lower counts in the smectite region, indiscriminately associated with kerolite. These samples, classified as K/S (kerolite/stevensite) mixed layer, occur in the same facies rich in spherulites or shrubs, with intermediate and partial dolomitization and silicification (less intense than the stevensite-rich assemblage). The selected samples include 16.35 (muddy calcarenite facies, the only reworked facies in the cored interval) and 30.85 (intercalation of muddy shrubstone within the muddy spherulstone facies). This assemblage was termed K/S mixed layer.

3.3. In Situ Morphochemistry

Based on the identification of clay mineral assemblages, we propose the in situ chemical characterization of the microsites, classified as follows: (I) kerolite pure; (II) stevensite-rich; (III) K/S mixed layer; (IV) saponite and detrital minerals.

3.3.1. Mineral Chemistry and Chemical Mapping

Analyses were conducted on two representative samples for each mineral assemblage. For the kerolite-rich assemblage, samples devoid of smectites or interstratified phases were selected. This assemblage primarily consisted of kerolite and carbonates, predominantly calcite spherulites or shrubs, with occasional rhombohedral dolomites. No significant reactions were observed in the matrix. As shown in Figure 3, kerolite occurs as laminated aggregates with notable chemical homogeneity, supported by Si, Mg, and Ca maps. The map indicates that neither calcite nor dolomite exerted a significant influence on the clay matrix. Rhombohedral dolomite was merely precipitated punctually onto the matrix without any intense substitution within the matrix or calcite. Kerolite analyses revealed average compositions of 58–61% SiO₂, 27–29% MgO, 0.2–0.4% Al₂O₃, and 0.6% F−, with totals ranging from 87 to 93%. The structural formula (Equation (1)) yielded average values of 8.05 for Si, 0.05 for total Al, 0.01 for Fe²⁺, 5.63 for Mg, and a charge balance of 0.29. Analytical data for all minerals are provided in

Table 2. Quantitative point analyses of clay minerals conducted by EPMA-WDS on thin sections; the data presented are in oxides and ions on a 22-oxygen basis.

Mineral	Kerolite		K/S Mixed Layer		Stevensite		Saponite
Sample	32.15/36.90		16.35/30.85		15.15/19.55		37.50/44.15
n	118		185		44		19
Oxides	µ	σ	µ	σ	µ	σ	µ	σ
SiO2	59.454	1.555	57.183	2.388	57.612	2.042	52.071	3.059
Al2O3	0.296	0.114	1.070	0.504	0.556	0.081	5.896	0.874
FeO	0.150	0.031	0.190	0.088	0.098	0.023	2.717	1.586
MgO	28.091	0.820	27.154	1.355	26.380	1.065	19.751	2.796
TiO2	0.034	0.016	0.015	0.059	0.024	0.013	0.242	0.332
MnO	0.004	0.015	0.000	0.014	0.000	0.014	0.006	0.015
CaO	0.242	0.147	0.878	0.356	0.196	0.064	0.380	0.289
Na2O	0.563	0.119	0.319	0.135	1.488	0.241	2.029	0.355
K2O	0.206	0.057	0.380	0.192	0.755	0.146	1.879	1.140
SUM	89.873	2.221	88.306	3.065	90.595	3.062	87.869	3.611
F-	0.670	0.170	0.990	0.269	3.047	0.531	0.691	0.239
Cations (22 O)
Si	8.058	0.037	7.939	0.089	8.047	0.036	7.635	0.272
Al	0.047	0.019	0.175	0.088	0.092	0.014	1.008	0.157
Al IV	0.000	0.011	0.084	0.081	0.019	0.011	0.365	0.270
Al VI	0.046	0.020	0.113	0.052	0.090	0.019	0.550	0.239
Fe2+	0.017	0.004	0.022	0.011	0.012	0.003	0.328	0.195
Mg	5.653	0.076	5.610	0.118	5.459	0.098	4.357	0.616
Ti	0.003	0.002	0.001	0.006	0.003	0.001	0.028	0.036
Mn	0.000	0.002	0.000	0.002	0.000	0.002	0.001	0.002
Ca	0.035	0.022	0.134	0.055	0.031	0.010	0.060	0.045
Na	0.147	0.032	0.087	0.039	0.395	0.065	0.571	0.107
K	0.036	0.010	0.067	0.036	0.135	0.029	0.348	0.189
Charge	0.265	0.053	0.430	0.155	0.624	0.104	1.095	0.102

IV—tetrahedral, VI—octahedral.

.

$${\mathrm{M}}_{(0.08-0.51)}\left({\mathrm{M}\mathrm{g}}_{5.46-5.93} {{\mathrm{F}\mathrm{e}}^{2+}}_{0.007-0.031} {{\mathrm{A}\mathrm{l}}^{\mathrm{V}\mathrm{I}}}_{0-0.109}\right)\left({{\mathrm{A}\mathrm{l}}^{\mathrm{I}\mathrm{V}}}_{0-0.015} {\mathrm{S}\mathrm{i}}_{7.854-8.09}\right){\mathrm{O}}_{20}{\left({\mathrm{F}}_{0.7}\mathrm{O}\mathrm{H}\right)}_4$$

(1)

For the stevensite-rich assemblage, samples were selected with a significantly higher proportion of smectites (stevensite) compared to kerolite. These samples illustrate processes of intense dolomitization and silicification, resulting in significant matrix substitution and the formation of lamellar lenses that frequently replace clay minerals.

Microsites in these samples display high chemical diffusion and silica mobility, as shown in Figure 4A. The Si compositional map highlights pure silica (red tones, representing higher counts) infiltrating as small veins and concentrating near dissolved clay minerals. Stevensite appears in green (intermediate silica content), while carbonates show minimal counts. In the Ca map (Figure 4B), calcite is represented by red tones, corresponding to spherulite cores and small crystals in the matrix. Yellow tones indicate rhombohedral dolomite in the matrix, occasionally at the edges of spherulites. A final episode of dolomitization, represented by green tones (low Ca content), is pervasive within spherulites and the matrix, forming microcrystalline masses at the edges of calcite and euhedral dolomite. This episode appears to postdate silica mobility, as it overlaps with the silica veins.

The analyses confirmed a typical stevensite composition, with average values of 54–59% SiO₂, 25–27% MgO, 0.5% Al₂O₃, and 3.0% F− and with totals ranging from 85 to 93%. The structural formula, based on 22 oxygen atoms (Equation (2)), yielded average values of 8.05 for Si, 0.09 for total Al, 0.02 for Fe²⁺, and 5.46 for Mg and a charge ranging from 0.5 to 0.67. Compared to kerolite, stevensite exhibits a slightly higher charge and fluorine content. Stevensite is present in all assemblages containing minor kerolite.

$${\mathrm{M}}_{(0.4-0.87)}\left({\mathrm{M}\mathrm{g}}_{5.17-5.64} {{\mathrm{F}\mathrm{e}}^{2+}}_{0.004-0.03} {{\mathrm{A}\mathrm{l}}^{\mathrm{V}\mathrm{I}}}_{0.03-0.16}\right)\left({{\mathrm{A}\mathrm{l}}^{\mathrm{I}\mathrm{V}}}_{0-0.05} {\mathrm{S}\mathrm{i}}_{7.94-8.08}\right){\mathrm{O}}_{20}{\left({\mathrm{F}}_{3.0}\mathrm{O}\mathrm{H}\right)}_4$$

(2)

For the K/S mixed layer assemblage, “S” is interpreted as stevensite based on its chemical composition and thermal behavior. Selected samples exhibited reflections in the smectite region, poorly resolved peaks, and kerolite in similar proportions. These samples lack detrital minerals but show intermediate and partial dolomitization and silicification processes. Compositional maps for the K/S mixed layer assemblage revealed two distinct chemical populations within the clay aggregates, suggesting reactions or disequilibrium at a very small scale due to differing proportions of kerolite and stevensite. Figure 5 illustrates these variations in a clay aggregate located between calcite crystals with shrub morphology. The Mg map highlights two magnesium-rich compositions, one slightly depleted in Mg and heterogeneously distributed within the microsite. A similar pattern is observed in the Si map, while the Al, Ca, and K maps inversely confirm these observations.

The analyses of K/S mixed layer samples showed values distributed between the kerolite and stevensite fields. Average compositions were 56–57% SiO₂, 27% MgO, 1.1% Al₂O₃, and 1.0% F−, with totals ranging from 87 to 89%. Based on the structural formula (Equation (3)), the average values were 7.94 for Si, 0.17 for total Al, 0.02 for Fe²⁺, and 5.63 for Mg, with a charge of 0.43 and a standard deviation of 0.16. This standard deviation is at least three times greater than those observed for kerolite and stevensite, indicating more chemically diverse populations within these samples.

$${\mathrm{M}}_{(0.18-0.88)}\left({\mathrm{M}\mathrm{g}}_{5.2-5.88} {{\mathrm{F}\mathrm{e}}^{2+}}_{0-0.04} {{\mathrm{A}\mathrm{l}}^{\mathrm{V}\mathrm{I}}}_{0-0.30}\right)\left({{\mathrm{A}\mathrm{l}}^{\mathrm{I}\mathrm{V}}}_{0-0.17} {\mathrm{S}\mathrm{i}}_{7.46-8.08}\right){\mathrm{O}}_{20}{\left({\mathrm{F}}_{1.0}\mathrm{O}\mathrm{H}\right)}_4$$

(3)

Saponite occurs in distinct microsites compared to kerolite and stevensite. The analyzed samples (37.50 and 44.15) represent mudstone facies containing chert, with sample 37.50 including a small interval of muddy spherulstone facies. The assemblage comprises saponite and a variety of detrital minerals, including quartz, feldspars, micas, and diagenetic calcite. In clay-rich levels, diagenetic products are minimal, with only sporadic rhombohedral dolomite and microcrystalline pyrite. Saponite aggregates exhibit smaller particle sizes compared to stevensite or kerolite, reduced optical continuity, and significant disruption by detrital minerals. Both samples analyzed contain centimeter-scale silica veins, though no silicification reactions are observed in the clay-rich zones.

Compositional maps (Figure 6) highlight geochemical variations, with Al and K enriched in both the clay matrix and detrital minerals (notably potassium feldspars, biotite, and muscovite). The magnesium map identifies two distinct phases: an Mg-rich phase (saponite) and a less Mg-rich phase, enriched in Al, K, and Si (probably detrital clays like illite). Silicon maps confirm the distribution of quartz grains and clay heterogeneity, while calcium maps reveal the dominance of calcite and the absence of dolomite in these microsites. The analyses of saponite revealed 52–53% SiO₂, 20% MgO, 5–6.2% Al₂O₃, and 0.7% F⁻, with totals ranging from 87 to 88%. The structural formula (Equation (4)), calculated based on 22 oxygen atoms, indicates values of 7.6–7.8 for Si, 0.9–1.0 for total Al, 0.2–0.4 for Fe²⁺, 4.2–4.7 for Mg, and a charge between 1.0 and 1.1. Saponite is distinguished by its elevated Al and Fe contents, likely derived from feldspars and micas present in the matrix.

$${\mathrm{M}}_{(0.92-1.2)}\left({\mathrm{M}\mathrm{g}}_{3.3-5.3} {{\mathrm{F}\mathrm{e}}^{2+}}_{0.18-1.0} {{\mathrm{A}\mathrm{l}}^{\mathrm{V}\mathrm{I}}}_{0.15-0.93}\right)\left({{\mathrm{A}\mathrm{l}}^{\mathrm{I}\mathrm{V}}}_{0-0.9} {\mathrm{S}\mathrm{i}}_{7.08-8.03}\right){\mathrm{O}}_{20}{\left({\mathrm{F}}_{0.7}\mathrm{O}\mathrm{H}\right)}_4$$

(4)

The classification of the samples in the classical ternary diagram of [50] (Figure 7B) and in the binary system of Figure 7A allows the following observations: I) the kerolites exhibit some charge, always lower than that of stevensite, but considerable, with a slight displacement from the theoretical ideal of talc, which has zero interlayer charge; II) stevensite has a low charge, sometimes below the theoretical minimum for smectites, as indicated by the dashed field in Figure 7B; III) the K/S mixed layer samples show intermediate compositions between kerolite and stevensite, with greater overlap with stevensite analyses, corroborating the predominance of this mineral, as observed in X-ray diffraction, where the 15 Å reflections are more common in these samples; and IV) saponite samples tend to form a trend aligned with the other samples, due to their slightly higher charge, and also show greater dispersion in the diagrams due to more variable values of Si, Al, and Mg, maintaining an acceptable coherence within the sample set.

FTIR data (Figure 8) for the regions corresponding to the Si–O (tetrahedral) and Mg₃–OH (octahedral) bonds indicated values between 1019 and 1020 cm⁻¹ and 670 cm⁻¹, respectively, for kerolite (sample 36.90) and stevensite (sample 15.15). These values are consistent with those described for talc (1018–1021 cm⁻¹) and hectorite (~1018 cm⁻¹), according to Madejová et al. [51]. For saponite (sample 44.15), these bands were observed to be between 1005 and 1010 cm⁻¹ for Si–O (with [51] reporting values close to ~1006 cm⁻¹) and between 653 and 657 cm⁻¹ for Mg₃–OH (compared to ~661 cm⁻¹ in saponites with some Fe²⁺ in octahedral coordination, as reported by [51]), indicating a lower occupancy of tetrahedral sites by Si and octahedral sites by Mg in saponite relative to the other samples.

In the region associated with the Mg₃–OH (octahedral) bond, a band at 3681 cm⁻¹ was identified for saponite, slightly higher than the values obtained for the other samples, which showed an average around 3677 cm⁻¹. The presence of this band between 3677 and 3681 cm⁻¹ supports the identification of the clays as magnesium trioctahedral, as per [51,52].

Although the band near 670 cm⁻¹ is characteristic of the bending vibration of the OH group coordinated to octahedral Mg, a slight shift to the region around 653 cm⁻¹ was observed in the saponite sample. Madejová et al. [51] suggest that the presence of Fe²⁺ in octahedral coordination may shift this band to approximately 666–661 cm⁻¹, which is consistent with the chemical analysis data, showing higher iron enrichment in saponite compared to the other two clays analyzed.

3.3.2. Morphologies of Clay Aggregates

The morphologies of aggregates and particles were analyzed using secondary electron images from scanning electron microscopy, which highlighted the significance of processes active in the assemblages proposed in this study and contributed to the textural characterization, which is still limited for Mg-clays. In general terms, kerolite, stevensite, and the K/S mixed layer display very similar morphological features (Figure 9).

Kerolite-rich is distinguished by well-preserved lamination, with high optical continuity and organization (Figure 9A), which explains its high birefringence under the microscope. When viewed at higher magnifications and in cross-sections perpendicular to the lamination, kerolite aggregates show elongated, irregular features with slightly undulated lamellar habits at this observation scale (Figure 9B). In samples dominated by stevensite, clay aggregates are frequently associated with dolomitization (Figure 9C) and silicification. In these cases, the aggregates may be residual and chaotic (Figure 9D) or “twisted” (Figure 9E), lacking the continuous lamination typical of kerolite and thus reinforcing the predominance of chaotic aggregates with low optical continuity and birefringence under the petrographic microscope.

In samples dominated by the K/S mixed layer, well-oriented lamellar aggregates like those of kerolite are observed; however, smaller, disoriented particles associated with larger lamellae are also visible (Figure 9F), supporting the presence of the two observed chemical compositions. Saponite aggregates exhibit unique features, consisting of slightly smaller particles and chaotic aggregates mixed with detrital minerals that hinder the continuity of the clay lamellar particles (Figure 9G,H).

4. Discussion

The clay minerals identified in this study are predominantly pure magnesian phases, including kerolite, stevensite, and their irregular mixed layers. These minerals are closely linked to sedimentary environments characterized by evaporitic deposition, typical of arid to semi-arid climates with high magnesium input and elevated alkalinity [53]. In saline, closed, and often ephemeral lacustrine systems—such as playa-type lakes—the limited and intermittent availability of water promotes the progressive concentration of brines over time. This geochemical setting fosters the precipitation of magnesian clay minerals alongside associated evaporitic phases [54].

Based on the facies overlap proposed in this study, it can be stated that the conventional facies classification [34] does not directly correlate with the clay minerals present in the matrix. However, certain facies were exclusively associated with specific assemblages. For instance, mudstones with a hybrid matrix were found to contain the saponite and detrital assemblage. All facies exhibiting a high degree of matrix replacement through dolomitization (including dolostones) and silicification predominantly feature the stevensite > kerolite assemblage. Facies such as spherulstone or shrubstone, which show a lower degree of matrix replacement, are associated with kerolite-rich or K/S mixed layer assemblages. Finally, the only reworked facies identified in this study are associated with the K/S mixed layer assemblage.

The sample set, scale, and objectives of this study do not permit an in-depth analysis of hydrological and climatic cycles, such as those detailed by [38] for the pre-salt paleolake in the Santos Basin. Additionally, it is not feasible to draw definitive conclusions about the paleoenvironment or environmental zoning (as proposed by Millot, [55]) based on data from a single core. Two primary mechanisms for the authigenic formation of clay minerals [53,56] were identified: (i) neoformation during periods of low flooding and high evaporation and (ii) transformation of precursor phases during evaporation after flooding episodes (detrital input). Within this framework, the two saponite-bearing intervals and their parageneses are interpreted here as indicative of atypical flood events rather than as evidence of lacustrine zoning.

The microsites approach demonstrates a clear differentiation among the mineral assemblages of stevensite, kerolite, K/S, and saponite. The saponite and detrital assemblage occurs in specific microsites within the sample set. These microsites are predominantly composed of inherited material, where mapping allows for the identification of quartz, feldspar, and mica grains embedded in a clay matrix. The interstitial clay matrix is complex, with numerous smaller particles enveloped by a homogeneous mass. This complexity poses a significant challenge for pinpointing the chemistry of this assemblage, as the abundance of fine materials complicates conducting analyses without interference, which explains the limited number of analyses with satisfactory chemical closure obtained for this set.

We propose that an effective approach is to define the analytical points using backscattered electron (BSE) imaging or compositional mapping, although the latter requires considerable acquisition time. It is not unequivocally clear whether the saponites are authigenic or detrital. However, the hypothesis of an authigenic origin seems plausible, given the compositional homogeneity observed in the interstitial matrix, particularly around finer grains. Furthermore, at this reaction scale, detrital minerals—especially those closest to the analysis—play a crucial role in clay chemistry, particularly in the occupancy of interlayer sites.

Stevensite predominates in microsites strongly impacted by pervasive and disseminated dolomitization and silicification, resulting in a less well-preserved clay matrix. Chemical mapping revealed the sequential progression of calcite reactions, initially forming spherulites and microcrystals that displace and replace the clay matrix. This was followed by a localized, low-intensity episode of dolomite precipitation (represented by yellow tones in the Ca map, Figure 4) and subsequently by a generalized and highly intense dolomitization phase, characterized by a lower Ca content, that envelops all constituents. Twisted and deformed aggregates observed through scanning electron microscopy, along with the predominantly chaotic features of the matrix identified via optical microscopy, support this interpretation. These features further suggest that this authigenic assemblage underwent significant equilibration with diagenetic products.

In contrast, kerolite-rich assemblage is better preserved, exhibiting regular lamination, optical continuity, and high birefringence, which indicates the greater cohesion of the lamellar particles (as observed through scanning electron microscopy). This suggests a less reactive system that is less affected by diagenetic reactions. The low extent of reaction percolation in these microsites contrasts with the reactive environment observed for stevensite. In this assemblage, the primary product replacing the matrix consists of rhombohedral dolomite. Although these dolomite crystals are significant, they do not appear to have induced substantial reaction mobility, seemingly “floating” within the matrix. The chemical maps clearly illustrate the behavior of these microsites in this assemblage, showing a homogeneous kerolitic matrix with no evidence of pronounced reaction percolation, as indicated by the evaluated elements.

The K/S mixed layer assemblage is represented by microsites where both phases coexist, illustrating a potential mineralogical transition between them. These microsites exhibit two distinct compositions that, without mapping, would be indistinguishable, as the average chemistry could easily be interpreted as low charge smectite data. The differing Al and K concentrations in the K/S mixed layer and Mg and Si in kerolite suggest a progressive chemical evolution. Matrix replacement processes are present in much smaller proportions compared to samples of the stevensite assemblage, yet they still occur. Aggregates predominantly resemble the features observed for kerolite under scanning electron microscopy. However, under optical microscopy, they exhibit chaotic or massive orientation with low birefringence. This characteristic supports the idea that these aggregates evolve from kerolite and are metastable, transitioning toward the stability field of K/S.

The clay minerals identified in this study are interpreted as authigenic products, since syngenesis refers to minerals formed simultaneously or in close association with deposition, either on the surface or at the sediment–water interface [57]. We do not expect these clay minerals to remain completely unaffected by greater or smaller modifications over millions of years and at this depth. The low permeability of these deposits likely restricted further dissolution processes during burial, suggesting that significant diagenesis occurred early, as noted by [58].

Clay matrix deformation caused by calcite growth implies that laminated clay aggregates were already forming during the transition between syngenesis and eodiagenesis, but the effects of compaction do not seem to play a role in the species present (details of this effect are in [59]). No hydrothermal features were observed in the sample set (details of this effect are in [28]).

Netto et al. (2022) and Wright and Tosca (2016) [37,60] propose clay precipitation from gels under extremely calm conditions. This model appears inconsistent with the large pre-salt magnesium clay deposits and frequent silt occurrences. Alternatively, clays may precipitate within the water column and settle as particles, forming laminated aggregates with high optical continuity, as supported by [19,58].

The HK effect [61] has often been invoked to explain the evolution of stevensite into talc (or kerolite) through burial-induced dehydration, as evidenced by Tosca and Wright (2018) [35]. This effect involves a heating, saturation, and cation mobility test, also applied to distinguish montmorillonite from beidellite [62]. Studies show that stevensite reliably loses its expandability at temperatures around 400–550 °C [9,45], significantly exceeding diagenetic conditions, even for the pre-salt area. Reference [58] suggests that stevensite transitions to kerolite through dehydration; however, the results of this study challenge this generalized view. Stevensite is observed in specific assemblages, and its persistence at depths exceeding 5 km contradicts a simple dehydration cycle. The coexistence of stevensite and kerolite at various depths indicates that this process is not merely burial-dependent dehydration. Instead, it is likely governed by the chemical and hydrosedimentary conditions of the lacustrine water. Netto et al. (2022) [38] demonstrated a cyclic succession of stevensite and kerolite in varying proportions and depths.

Regarding the K/S mixed layer, there is no consensus on their analogy to classical systems like I/S or C/S, which are formed through burial processes [63,64,65]. Transformations such as smectite-to-illite, typical in diagenetic environments at low temperatures (60–120 °C), depend on the initial composition and local conditions. However, the evolution of K/S in the pre-salt area appears to follow distinct pathways related to the local specific chemical and depositional processes [66].

The results of this study demonstrate a correlation between kerolite and K/S, suggesting an evolution between them, moving toward interstratification with increasing matrix replacement processes. It appears that stevensite may represent a more advanced stage of this process, though this hypothesis, while plausible, does not suggest that kerolite is its precursor. The data do not identify the clay precursor precipitated directly from the lake or the degree of modification during lithification, and nor could they, since any phase (stevensite or kerolite) could precipitate in the lake and be modified later, as per models [53,67,68,69], and no defined models or methodologies exist to determine a precursor in this context. The observable changes, including the nucleation of aggregates, particle growth, compaction, and carbonate precipitation, do not justify classifying the pre-salt clays as entirely diagenetic, although the stevensite in this study, with notable matrix replacement, is the most likely candidate to be in equilibrium with the later paragenesis recorded. The only existing models can characterize the transformation of kerolite and stevensite into sepiolite under specific conditions [53,70,71], but this was not observed in this study.

Saponite samples exhibit the highest XRD peak intensities and narrower widths at half maximum for the first peak, suggesting relatively greater structural organization compared to kerolite, stevensite, and K/S, which display broader and less defined peaks. As [37] discussed, there remains no clear correlation between depth and the degree of ordering (“crystallinity”) in Mg-rich clay minerals. However, the more plausible explanation is that the saponites in this study inherited this structure, making them more organized. This complexity further complicates the use of conventional tools to understand the relationship between diagenesis, burial, and clay minerals.

The theoretical difference between saponite and stevensite is that the former derives its charge from the tetrahedral site, while the latter has octahedral vacancies. In contrast, kerolite is thought to lack an interlayer charge. From the formulas of the pre-salt clays, saponites showed a higher interlayer charge (charge = 1.1) and aluminum content (Al_tot = 1) compared to stevensite (charge = 0.6 Al_tot = 0.09). The values of Al and charge for kerolite are not zero, as theoretically predicted, but are much lower compared to the other phases (charge= 0.2 and Al_tot = 0.04). Meanwhile, the K/S mixed layer recorded intermediate charge values (charge = 0.4) and had much higher Al_tot = 0.17 compared to stevensite. The content of Fe²⁺ is also noteworthy as a diagnostic feature, with values of 0.3 for saponites and between 0.01 and 0.03 for the other species. Magnesium content is highest in kerolite (5.6), followed by stevensite (5.4), and lower in saponite (4.3). These data provide a reliable and robust methodology for determining the composition of these minerals and confirm some of the subtleties and challenges discussed in the global literature on Mg-clays [2,6,7,8,9,72,73,74].

This study emphasizes the importance of chemical maps for identifying reactive interactions in microsites and understanding the stability of clay phases. The results extend the understanding of diagenetic processes in the pre-salt area and may guide future studies focusing on mineralogical evolution in complex systems, with the aid of additional in situ tools.

5. Conclusions

• It is not possible to determine the precursor clay mineral for the phases studied, as all have undergone burial of approximately 5 km deep for over 100 million years. Clay minerals are sensitive and reactive, especially magnesian ones, which are more disordered and exposed to lake fluids under initial conditions. Additionally, the differences between kerolite and stevensite are subtle, and one phase can easily shift to the stability field of the other.
• Saponite is easily distinguishable, with significantly higher values of Al, Fe, and K and lower values of Si and Mg. In the studied rocks, it is consistently associated with detrital minerals, suggesting it may either be a product of the transformation of these minerals or, in part, a detrital clay mineral deposited in the system.
• Stevensite is distinguished from the other phases by its intermediate charge and very low Al content. It is always associated with kerolite in varying proportions. When dominant, it is closely linked to dolomitization and silicification processes, suggesting possible stabilization under these conditions.
• Kerolite is the most abundant clay mineral in the interval and is characterized by a minimal Al content and charge. When occurring in its pure form, it is observed to be free from more intense transformation processes, with only sporadic rhombohedral dolomite, indicating that it was less affected by later diagenetic processes.
• The interstratified K/S mixed layer occurs exclusively with kerolite. In the microsites, it appears as a meta-stable aggregate with two compositions: one intermediate between kerolite and stevensite, and the other equivalent to kerolite. These samples show dolomitization, silicification, and late calcite processes, with varying degrees of matrix substitution, but always with lower intensity compared to the assemblies where stevensite predominates.
• Electron probe microanalysis (EPMA) proved to be an essential tool for in situ chemical analysis, enabling the distinction of clay compositions especially in complex assemblages of strongly interrelated phases. Chemical WDS mapping is a powerful approach to visualizing the elements mobility and distribution and clarifying the processes and products of reactions in microsites scale. Future research should explore additional assemblages and micro-sites from other pre-salt targets (such as the Campos Basin) and analogous areas to validate the findings and proposals made in this study.

Even though the sampled intervals from clay-rich, non-reservoir levels do not encompass the full regional heterogeneity of the Barra Velha Formation, our approach stands out as original and distinct compared to studies focused on reservoir rocks, where clays have undergone intense dissolution and/or replacement. This methodology also anticipates the formation and evolution of clay mineral-induced secondary porosity, which significantly influences reservoir rock quality.

Furthermore, our study recommends that this approach—centered on microsites and reaction pathways, supported by appropriate analytical tools—be extended to regional investigations and facies succession analyses. Such applications would contribute to paleoenvironmental reconstructions and the development of evolutionary models for lacustrine systems.