Mineralogical and Geochemical Characterization of the Benavila (Portugal) Bentonites

Abstract

This work aims to perform a detailed mineralogical, crystal-chemical, and geochemical characterization of bentonites from the Benavila outcrop, the largest known deposit of bentonites in continental Portugal. Bulk samples and different size fractions were characterized through X-Ray Diffraction (XRD). Structural formulae of the smectites were fitted from point analyses acquired by analytical electron microscopy (AEM) with transmission electron microscopy (TEM). Smectites are the major component with variable amounts of calcite and minor amounts of quartz, feldspar, illite, and chlorite. Occasionally, amphiboles and dolomite have also been identified. The high content of carbonates in different parts of the sampling area is related to the circulation of carbonate-rich fluids. The smectites present high-layer charge, are intermediate terms of the montmorillonite–beidellite series, and also show an intermediate cisvacant–transvacant configuration. Major and trace elements concentrations were determined by ICP-MS. The geochemical analysis of the samples indicates an enrichment in SiO₂ and Al₂O₃ and a depletion of the more clayey materials in REE, HFSE, and Y, among others. The calculation of the PIA and CIA alteration indices, along with other parameters observed, shows the possible alteration pathways of the Benavila deposit. Research to evaluate the ability of these bentonites to be used as engineering barrier systems (EBS) and sealing materials for radioactive waste repositories is ongoing.

1. Introduction

Bentonites are clayey industrial rocks of great economic interest whose main component is smectites. The most important physical–chemical properties of these minerals are their high specific surface area and cation exchange capacity, which are deeply related [1,2,3,4], along with their variable layer charge, allowing them to react with inorganic and organic reagents. Hence, they have swelling and rheological properties and high plasticity, which play a very important role in the industrial field [5,6]. Their applications are diverse, contingent on the aforementioned properties. While their most prominent use is as ad-absorbents, their significance in other domains cannot be overstated. For instance, they are used in civil engineering, as binding agents, as additives in animal nutrition, as catalysts or support for catalysts [5,6,7], which also have to be highlighted. Apart from the typical fields in which these materials have been traditionally used, in recent years there has been a spike in the use of bentonites in the fields of nanocomposites [8,9,10], pharmaceutics [11,12], medicine [13,14,15], among many others. The study of bentonites as engineering barrier systems (EBS) and sealing materials in radioactive waste repositories has to be highlighted, since this research field has been one of the most prominent ones in the last four decades regarding these materials [16,17,18,19,20,21,22].

The largest known deposit of bentonites in continental Portugal is located in Benavila (in the Alentejo region) (Figure 1). This deposit was formed due to the weathering of granitoids from the Benavila igneous massif [23,24], mainly composed of quartz, plagioclase, and amphibole, along with other accessory minerals [24]. The composition of these plutonic rocks will affect the resulting weathering products, which are the bentonites.

The Benavila deposit reserves were estimated at around 5 M.t. of bentonite, with the highest purity levels reaching 2.0–2.5 Mt [28,29]. The mineralogy and applications of these bentonites have been studied by different authors, determining their mineralogy as being primarily constituted by dioctahedral smectites [29,30,31,32,33,34]. However, there are slight differences regarding their classification within the group of dioctahedral smectites, because some authors classify them as Tatatila-type montmorillonites [29] and others just as montmorillonites [31].

This work aims to attain a detailed mineralogical, crystal-chemical, and geochemical characterization of a representative outcrop of bentonites from the Benavila deposit. This work will serve as a basis for further research regarding the applications of these bentonites, mainly towards their use as lining barriers of disposed radioactive waste.

2. Geological Setting

The area under study, corresponding to the Benavila igneous massif, is located close to the Avis area, more precisely in Benavila, Portalegre District (Portugal). It is approximately 7 km long (E-W) and 4 km wide, occupying an area of about 40 km² (Figure 1). This area exemplifies the typical morphology of Alto Alentejo: a high plateau with scattered reliefs. The petrographic characteristics of the Benavila rocks lead them to be considered similar to those of the neighboring Ervedal, Fronteira, and Stª Eulália massifs. However, the rocks of this region, like those of Stª Eulália, are more diverse than those of the first two massifs mentioned, as the presence of gabbros is referred [35].

The Benavila igneous massif was formed during the Lower Silurian [26,27], and it is characterized by the diversity of petrologic types: granites, granodiorites, monzonites, quartz diorites, and, exceptionally, olivine gabbros. Granitoids, quartz diorites, and gabbros have a common characteristic: the presence of green-colored hornblende amphiboles. The massif is cut by numerous veins, running NW-SE, which are mostly constituted of granite. In addition to the aforementioned veins, there are also many enclaves of felsic nature, namely monzonites, always more basic than the surrounding granitoids. The entire massif is also affected by carbonated alterations of Paleogene–Quaternary age, due to the high content of carbonates in the groundwater [24]. Chemically, the global composition of these rocks located at the Benavila igneous massif is mainly constituted by SiO₂, Al₂O₃, Fe₂O_3, and CaO, representing at least 75% of the total oxide content [24].

The subject of study is bentonites of Miocene age, which have been previously described in the literature as “clayey and marly sandstones with a greenish hue, of continental facies, probably Miocene” [26]. These bentonites originated from the weathering of the preexisting granodiorites located within the massif, all of which present the same type of alteration that partially covers the massif, which seems to be favored by fractures in the NNE–SSW direction [26,27] and is dated in the Paleogene–Quaternary. The Geological and Mining Institute of Portugal refers to these clayey materials of residual origin as “very pure bentonites with relevant technological characteristics and difficult exposure and exploitation conditions” [36].

3. Materials and Methods

3.1. Samples

A bentonite outcrop (Universal Transverse Mercator coordinates: 29S 598,193.02 m E 4,330,882.19 m N) with a thickness of approximately 7 m and a NE–SW direction in the proximity of the locality of Benavila (Portalegre, Portugal) (Figure 1) was sampled. The outcrop is characterized by the greenish colors of the bentonites, as well as by the presence of sub-vertical carbonate-rich veins of whitish color (Figure 2a,d). The presence of these veins is the main difference that can be observed in the horizontal axis. In addition, it is possible to observe a higher degree of compaction of the bentonites towards the top of the outcrop, while the bottom shows more disaggregated materials.

As shown in Figure 2a, nine samples from the most representative and accessible parts of the deposit were collected in a grid, with the A profile located towards the NE and the C profile towards the SW. In order to obtain non-contaminated samples, the surface of the outcrop was cleaned at the places where these samples were collected. The samples from the top horizontal profile (BEN3) were interpreted in the field as the samples with the higher content of smectites. They are characterized by their high compaction, and, in some cases, they even preserve their original textures (especially towards the C profile) (Figure 2b), although they have been transformed almost entirely to clay minerals. According to the expected evolution of the weathering profile, which gave place to the formation of the deposit, the lower levels (BEN1 and BEN2) present a lower degree of compaction and higher disaggregation, not preserving the original textures (Figure 2c). Specially along the B line, the presence of veins of carbonates (Figure 2d) is usual.

3.2. Methodology

Different granulometric fractions of all the samples were obtained. Silt fractions (θ < 63 μm and θ < 38 µm) were separated from the bulk rock samples by wet sieving, while the clay size fraction (θ < 2 µm) was separated by decantation following Stokes’ law. The granulometric separation was performed not only to identify the variation in the relative proportion of the minerals within them, but also with a view to their possible future use in radionuclide sorption experiments to assess their applicability as lining barriers at radioactive waste disposal sites.

The mineralogical characterization of the samples was carried out by X-ray diffraction (XRD) using a Bruker D8 Advance and Cu kα radiation. All the granulometric fractions were powdered prior to their characterization, and, in addition, the clay fraction (θ < 2 µm) was also prepared as oriented air-dried aggregates (OA), solvated with ethylene glycol (OA + EG), and heated at 550 °C for 2 h (OA + TT). Powdered samples were scanned from 2° to 65° 2θ, and air-dried aggregates from 2° to 45° 2θ, using a speed of 0.05° 2θ/s. The clay size fraction was also measured from 59° to 63° 2θ at a speed of 0.03° 2θ/s to obtain a detailed diffractogram of the (060) reflection of clay minerals, to know the trioctahedral or dioctahedral character of the phyllosilicates present within the samples. Semi-quantification of the mineral phases observed at the diffractograms was performed through the reflecting powers method (RPM) [37], which is a standard in the study of clays, focusing on the ratio of peak areas to semi-quantify the minerals within the samples.

Structural formulae of the smectites were fitted from point analyses acquired by analytical electron microscopy (AEM) with transmission electron microscopy (TEM) at the National Electron Microscopy Center (Madrid, Spain) using a JEOL JEM-1400 microscope (JEOL Ltd., Tokyo, Japan), with an acceleration voltage of 100 kV. The microscope incorporates an OXFORD ISIS EDX spectrometer (Oxford Instruments, Oxfordshire, UK, 136 eV resolution at 5.39 keV), equipped with its own software for quantitative analysis. The TEM samples were prepared by depositing a drop of diluted suspension on a microscopic grid with collodion (cellulose acetate butyrate) and coated with Au. Homoionization was performed on the samples to obtain more accurate structural formulae that allow a more reliable classification of the smectites [38].

Thermogravimetric analysis (TGA) routine was performed using an SDT-Q600 from T.A. Instruments (T.A. Instruments, New Castle, DE, USA). Homoionized fractions < 2 µm were analyzed in an air atmosphere, with a heating rate of 10 °C/min, from room temperature up to 985 °C.

Geochemical analyses of major and trace elements were performed at Activation Laboratories Ltd. (ACTLABS, Ancaster, ON, Canada). For these analyses, 2 g of powdered sample was digested with aqua regia, along with appropriate international reference materials for the metals of interest, and diluted to 250 mL volumetrically. Major elements, along with Sc, Be, V, Sr, Zr, and Ba, were analyzed by lithium metaborate/tetraborate fusion inductively coupled plasma (ICP) and trace elements were analyzed by ICP-mass spectrometry (ICP-MS).

The chemical index of alteration (CIA) and plagioclase index of alteration (PIA) were calculated to elucidate the alteration rock degree [39,40,41,42,43,44]. These indices incorporate major bulk element oxide chemistry from the samples, giving information on the leaching of elements on each of the samples originating from homogeneous parent materials and indicating their relative degree of alteration.

Software package IBM SPSS version 29.0 was used for the statistical analyses.

4. Results

4.1. Mineralogical Characterization

4.1.1. XRD

Powdered samples from all the granulometric fractions were characterized by XRD, as well as the air-dried aggregates of <2 microns fractions, which allowed identifying the mineral assemblages within them (Figure 3). Powdered samples allow the identification of phyllosilicates, carbonates (mainly calcite, along with minor amounts of dolomite), and scarce quantities of tectosilicates (quartz, K-feldspar, and plagioclase) (

Table 1. Semi-quantification of the mineral content of the samples from X-ray diffraction (XRD). (-): present at trace levels. Amphibole (Amp), calcite (Cal), chlorite (Chl), dolomite (Dol), phyllosilicates (Phyllo), plagioclase (Pl), potassium feldspar (Kfs), and quartz (Qz).

Sample	Amp	Cal	Dol	Kfs	Pl	Qz	Chl	Ilt	Sme	C/S
BEN1A		12	2	4	-	6	-	-	74	-
BEN1A < 63 µm		5	-	-	-	2	-	-	90	-
BEN1A < 38 µm		2	-	-	-	-	-	-	96	-
BEN1A < 2 µm		2					-	-	97	-
BEN1B		25	2	-	-	-	-	-	70	-
BEN1B < 63 µm		11	-	-	-	-	-	1	86	-
BEN1B <3 8 µm		9		-	-	-	-	1	89	-
BEN1B < 2 µm		4					-	-	95	-
BEN1C	-	16		-	-	-	-	-	82	-
BEN1C < 63 µm	-	4		-	-	-	-	-	94	-
BEN1C < 38 µm	-	5		-	-	-	-	-	92	-
BEN1C < 2 µm		2					-	-	97	-
BEN2A		17	-	-	-	-	-	-	81
BEN2A < 63 µm		12		4	-	-	-	-	83
BEN2A < 38 µm		5		-	-	-	-	-	93
BEN2A < 2 µm		-					-	-	99
BEN2B		27		-	-	-	-	-	71	-
BEN2B < 63 µm		9		3	-	-	-	-	86	-
BEN2B < 38 µm		15		4	-	-	-	-	80	-
BEN2B < 2 µm		2					-	-	97	-
BEN2C	9	26		2	7	-	-	-	55
BEN2C < 63 µm	4	13					-	-	82
BEN2C < 38 µm	3	5		-	-	-	-	-	90
BEN2C < 2 µm		2					-	-	96
BEN3A			-	-	-	2		-	95	1
BEN3A < 63 µm				-		-		-	97	1
BEN3A < 38 µm				-	-	-		-	96	1
BEN3A < 2 µm								-	98	1
BEN3B		34					-	-	64	1
BEN3B < 63 µm		27				-	-	-	71	1
BEN3B < 38 µm		28					-	-	70	1
BEN3B < 2 µm		1					-	-	96	2
BEN3C	4	38		4	-	2	-	-	51
BEN3C < 63 µm	-	16		-	-	-	-	-	82
BEN3C < 38 µm	2	11		-	-	-	-	-	86
BEN3C < 2 µm		2					-	-	97

). The rest of the minerals identified are always present, just at trace levels. These results agree with previous references [29,30]. Amphiboles were also identified, but only in samples belonging to the C vertical profile (Figure 2). The diffractograms corresponding to the finer granulometries (<63, <38, and <2 µm) show that the relative intensities of phyllosilicates increase compared with the other minerals identified, indicating a preferential concentration of phyllosilicates in the fine-grained fraction, as expected. The oriented aggregates (OA) and their treatments grant the capacity of studying clay minerals in more detail, being able to identify smectite as the main component along with traces of illite, chlorite, and possible kaolinite. Small amounts of a mixed-layer of chlorite and smectite (C/S) have also been identified due to the presence of a reflection at 12.4 Å that does not collapse in the AD + TT [45]. Also, in the OAs, there are small inflections at 12.3 Å (Figure 3b), which indicate a possible mixture of smectites with different predominant interlayer cations, ratified by their swelling in the OA + EG treatment. The dioctahedral character of the smectites was identified due to the d-spacings of the (060) reflections, whose values range between 1.499 and 1.507 Å, and in all the samples studied (Figure 3c).

The X-Ray diffractograms not only allow identifying minerals, but also semi-quantifying them (Table 1). In all the granulometric fractions, the main components are phyllosilicates (smectite plus quantities below 2% of illite, chlorite/smectite mixed-layers, and chlorite), followed by carbonates (calcite plus traces of dolomite) and scarce quantities of quartz and feldspars. In the bulk rock samples, phyllosilicates range from 97% (BEN3A) to 52% (BEN3C), with smectite being the main component (95% and 51%, respectively). Calcite is the major impurity and ranges from 0% (BEN3A) to 38% (BEN3C) with a mean content of 22%. It has to be pointed out that sample BEN3A is the richest sample in smectite because it is the only one that does not contain calcite, and a negative correlation between these two minerals can be observed. The <63 µm fraction shows percentages of phyllosilicates ranging from 99% (BEN3A) and 73% (BEN3B), with smectite being the major mineral, and of calcite ranging from 0% (BEN3A) and 27% (BEN3B). This tendency previously described is also seen in the <38 µm fraction, being the maximum content of phyllosilicates (mainly smectite) 98% (BEN3A) and the minimum, 72% (BEN3B), coinciding with the maximum and minimum contents in calcite (28% and 0% at samples BEN3B and BEN3A, respectively). Finally, in the <2 µm fraction, most of the minerals identified and quantified are essentially smectite with a minimum content of 95% and scarce impurities of chlorite, illite, and mixed-layer chlorite/smectite that can be identified but are difficult to quantify.

4.1.2. TEM-AEM

The top samples (BEN3A, BEN3B, and BEN3C) were used for the crystal-chemical study through TEM-AEM analyses. The TEM representative photographs (Figure 4) show platy particles with subhedral morphologies, which present some straight faces together with wavy and pointy ends (as usual in smectites) on the same particles. This resembles typical morphologies of micaceous materials (euhedral particles with straight faces), which, most likely, are the precursor phases whose alteration gave place to the smectites. The wavy and pointy ends indicate the transformation process and/or growth of smectites from these micaceous particles [46,47,48,49]. The photographs evidence an increase in the wavy and pointy-ended faces towards the A profile (Figure 4).

The point chemical data obtained by AEM analyses (

Table 2. Chemical analysis of major oxides (wt%) and structural formulae of the smectites obtained from point analyses by HRTEM–AEM. Max: maximum value; Min: minimum value; Mean: mean value; SD: standard deviation; ƩTC: sum of tetrahedral cations; ƩOC: sum of octahedral cations; TCh: tetrahedral charge; OCh: octahedral charge; LC: layer charge; n: number of analyses.

Samples	Chemical analysis of major oxides									Mean structural formulae
		SiO2	Al2O3	Fe2O3	MgO	CaO	NaO	K2O	TiO2	Tetrahedral			Octaheral					Interlayer
										Si	Al	ΣTC	Al	Mg	Fe	Ti	ΣOC	Ca	Na	K	TCh	OCh	LC
BEN3A	Max	66.33	32.64	11.02	8.21	9.15	0.10	6.87	0.26
	Min	54.81	18.33	0.00	2.57	0.25	0.00	0.00	0.00
	Mean	63.98	19.90	7.89	6.84	0.92	0.00	0.40	0.06	7.67	0.33	8.00	2.48	1.22	0.71	0.01	4.42	0.12		0.06	−0.33	0.05	0.28
n = 26	SD	1.97	2.58	2.28	0.90	1.60	0.02	1.26	0.09
BEN3B	Max	62.46	19.56	33.00	11.94	2.75	0.00	1.43	1.14
	Min	48.07	11.58	10.40	4.60	0.46	0.00	0.00	0.00
	Mean	56.49	15.20	18.43	8.03	1.20	0.00	0.54	0.11	7.12	0.88	8.00	1.39	1.51	1.75	0.01	4.66	0.16		0.09	−0.88	0.48	0.40
n = 29	SD	3.85	1.72	5.04	2.06	0.60	0.00	0.40	0.29
BEN3C	Max	63.18	16.95	29.57	11.95	2.93	1.83	0.70	0.74
	Min	49.74	11.85	11.77	4.20	1.41	0.00	0.00	0.00
	Mean	56.80	13.35	20.07	7.08	1.89	0.12	0.38	0.12	7.22	0.78	8.00	1.23	1.34	1.92	0.01	4.50	0.26	0.03	0.06	−0.78	0.17	0..61
n = 18	SD	3.37	1.24	4.10	1.79	0.41	0.47	0.24	0.23
BEN3ACa	Max	66.65	29.57	13.24	6.85	16.03	0.00	9.76	0.57
	Min	54.10	15.68	2.04	1.13	2.22	0.00	0.04	0.00
	Mean	61.53	19.19	7.64	4.71	5.49	0.00	1.40	0.06	7.54	0.46	8.00	2.31	0.86	0.70	0.01	3.88	0.72		0.22	−0.46	−1.20	1.66
n = 20	SD	3.42	3.74	3.17	1.41	2.68	0.00	2.88	0.20
BEN3BCa	Max	63.19	17.80	19.73	13.66	6.48	0.00	2.77	1.38
	Min	49.28	12.84	11.67	4.03	1.77	0.00	0.00	0.00
	Mean	58.80	15.13	14.75	6.54	3.67	0.00	0.70	0.42	7.36	0.64	8.00	1.59	1.22	1.39	0.04	4.24	0.49		0.11	−0.64	−0.46	1.10
n = 26	SD	3.36	1.05	1.98	2.17	1.05	0.00	0.81	0.33
BEN3CCa	Max	62.60	16.75	16.39	13.30	8.25	0.00	7.10	1.03
	Min	53.74	12.86	10.97	6.04	2.18	0.00	0.31	0.02
	Mean	57.10	14.73	13.27	8.99	3.74	0.00	1.67	0.50	7.20	0.80	8.00	1.38	1.69	1.26	0.05	4.38	0.48		0.34	−0.80	−0.50	1.30
n = 17	SD	2.49	1.11	1.77	2.30	1.39	0.00	1.94	0.25

) of isolated particles of smectite of the natural samples show that the main oxide (considering the mean values) is SiO₂, ranging from 63.98% (BEN3A) and 56.49% (BEN3B). Regarding the second most abundant oxide, there are two cases: sample BEN3A has more content in Al₂O₃ than Fe₂O₃ (19.90 and 7.89%), while samples BEN3B and BEN3C have higher contents in Fe₂O₃ (18.43 and 20.07%) than in Al₂O₃ (15.20 and 13.35%). Finally, the rest of the oxides analyzed follow the same trend in all the samples studied: MgO > CaO > K₂O > TiO₂ > NaO. It is also noteworthy that they show a broad compositional variability, as indicated by the differences between the maximum and minimum values, as well as the standard deviation, which is related to the continuous compositional variations in all the studied samples.

The structural formulae were calculated from the mean oxide contents of the point chemical analyses (Table 2). The resulting formulae from the natural samples correspond to dioctahedral smectites. Their average number of octahedral cations ranges from 4.42 to 4.66. These formulae exhibit positive octahedral charge ranging from 0.05 to 0.48 and a very low interlayer charge ranging from 0.30 to 0.61, indicating that the results are unreliable. To properly fit the smectite structural formulae, the samples were homoionized with Ca²⁺ [38]. The homoionized analyses show similar distributions as in non-homoionized samples (also considering the mean values): SiO₂ is the main oxide, ranging from 61.53% (BEN3ACa) and 57.10% (BEN3CCa). The second most abundant oxide is always Al₂O₃ (in opposition to the non-homoionized samples), which ranges from 19.19% (BEN3ACa) and 14.73% (BEN3CCa), while Fe₂O₃ is the third, ranging from 14.75% (BEN3B) and 7.64% (BEN3A). Samples BEN3BCa and BEN3CCa show MgO as the fourth oxide in terms of abundance (6.54 and 8.99% respectively) and CaO as the fifth (3.67 and 3.74% respectively), while this trend is inverted in sample BEN3ACa (5.49% of CaO and 4.71% of MgO). The rest of the oxides analyzed follow the same trend in all the samples studied: K₂O > TiO₂ > NaO. The structural formulae also correspond to dioctahedral smectites. The mean octahedral charges vary, being −1.66 (BEN3ACa), −0.46 (BEN3BCa), and −0.50 (BEN3CCa), and an increase in the interlayer charge. It is possible to see that the main difference between non-homoionized and homoionized samples is their content in Fe₂O₃, which suggests that iron was most likely adsorbed on the non-homoionized samples and removed during their homoionization. The K₂O content did not decrease in the homoionic analyses, indicating the presence of interestratified micaceous layers in the smectite particles, which supports their detrital origin.

The structural formula obtained from both the non-homoionized and homoionized samples was used to obtain a ternary plot of the octahedral content (Mg, Fe, Al) of all the analyses, along with the octahedral content of different smectite samples from the literature [38,50,51] (Figure 5). The ternary plot of the octahedral content of non-homoionized samples shows different tendencies for Benavila samples. BEN3A is the most enriched in aluminum, therefore, being the closest to both beidellite and montmorillonite. Samples BEN3B and BEN3C display higher compositional dispersion and show higher contents of iron, which in some cases bring them close to the nontronite octahedral content. However, when the homoionized octahedral content is plotted, it is easily observed that no analyses fall near the notronite octahedral content but rather fall in intermediate fields. There is also an evident trend that can be observed in both ternary plots, which is a progressive transformation towards purest dioctahedral smectites from the BEN3C to the BEN3A sample, due to the progressive loss of magnesium and iron.

The smectites from Benavila had already been classified as Tatatila type montmorillonites [29], meaning they have 15–50% of the charge originating from tetrahedral substitutions and a layer charge > 0.85 [45]. Other studies have simply classified them as montmorillonites, without defining their type [31]. However, recent classifications of dioctahedral smectites [52,53], which consider more variables, such as the layer charge, the octahedral structure, the iron content, and the charge location, allow for giving more descriptive names of the subjects of study. For such purpose, the structural formulae obtained from the homoionized samples were used.

Sample BEN3ACa has a layer charge (LC) of 1.66, a Fe³⁺ content (IC) that represents 18% of the octahedral cations, and a tetrahedral charge (TC) that accounts for 28% of the total charge, and is therefore classified as a high-charged ferrian beidellitic montmorillonite [52]. Sample BEN3BCa has an LC of 1.10, an IC of 33%, a TC of 58% and is classified as a high-charged ferrian montmorillonitic beidellite. Sample BEN3CCa has an LC of 1.30, an IC of 29%, and a TC that is 62% of the total charge, being classified as a high-charged ferrian montmorillonitic beidellite. Therefore, it can be summarized that there are two types of samples: sample BEN3A is a ferrian beidellitic montmorillonite, while BEN3B and BEN3C are ferrian montmorillonitic beidellites. All the samples show high layer charges.

4.1.3. TGA

The thermal behavior of Benavila bentonite samples, essential for their possible use in radioactive storage, was determined through thermogravimetric analysis. These analyses also enable an in-depth characterization of the octahedral sheet of the smectites. For that purpose, the homoionized < 2 µm fraction of the three samples from the top level (BEN3A, BEN3B, and BEN3C) was selected. The TGA analyses (Figure 6a–c) show that all the samples present the same tendencies: the most important weight loss (around 15% of their masses) takes place from room temperature up to ~175 °C, due to the dehydration of the absorbed and interlayer water of the smectites. They keep losing small percentages of weight until the range comprised between 550 and 650 °C, where they lose approximately 5% of their mass due to the dehydroxilation of smectites. DTG analyses (Figure 6a–c) give more detailed information. The dehydration of the interlayer water occurs in two peaks due to the presence of different interlayer cations [54], and the dehydroxilation events are marked by defined peaks, although they can be broader or narrower peaks, depending on the samples.

The peaks related to the dehydroxilation of the smectites in the DTG curve (derivative of the thermal analyses) are used to calculate the percentage of cis-vacant (cv) and trans-vacant (tv) configurations of the octahedral sheet [55]. The two configurations can be differentiated due to their dehydroxylation temperatures. The cv varieties dehydroxylate at temperatures between 650 and 700 °C, while tv varieties dehydroxylate at temperatures of about 500–550 °C, being the boundary temperature between the two configurations at around 600 °C [55,56,57]. These percentages were obtained by performing the deconvolution of the DTG curve between 300 and 800 °C using the software package Fityk (v. 1.3.1) [58] (Figure 6d), where it can be seen that a curve is deconvoluted into three curves plus a straight line (the latter is used to subtract the background). Considering the center of the curves obtained from the deconvolution, it is determined whether the dehydroxilation takes place at temperatures below or over 600 °C, and using their respective areas, it is possible to calculate the percentage that cv (peaks > 600 °C) and tv (peaks < 600 °C) varieties represent (Table S1) [57,59]. All the samples are classified as tv/cv, with a progressive decrease in the cv character towards sample BEN3A.

Considering the previous classification from the TEM-AEM analysis of the samples according to recent references [52,53] and focusing on the octahedral configuration, we can conclude that sample BEN3A is a high-charged tv/cv ferrian beidellitic montmorillonite, while BEN3B and BEN3C are high-charged tv/cv ferrian montmorillonitic beidellites.

4.2. Geochemistry

The major (Table S2) and trace elements (Table S3) contents were obtained by means of ICP-MS for bulk samples and different size fractions.

Table 3. Maximum (Max), minimum (Min), and mean values of major elements in weight percentage (wt%) of all the granulometric fractions, along with their standard deviation (SD).

	Bulk Sample				<63 µm				<38 µm				<2 µm
	Max	Min	Mean	SD	Max	Min	Mean	SD	Max	Min	Mean	SD	Max	Min	Mean	SD
SiO2	51.86	29.76	39.27	6.96	48.76	32.96	42.36	4.57	49.33	34.20	43.31	4.22	50.23	45.27	48.14	1.50
Al2O3	13.58	8.26	11.37	2.00	14.39	8.72	12.02	2.24	14.48	8.55	12.14	2.32	16.08	10.38	13.47	1.89
Fe2O3	7.81	6.10	6.96	0.62	10.60	5.27	7.56	1.50	11.52	5.30	8.10	1.81	10.53	4.85	7.63	2.02
MnO	0.13	0.06	0.10	0.02	0.09	0.04	0.06	0.02	0.09	0.04	0.07	0.02	0.07	0.02	0.05	0.02
MgO	6.18	4.30	4.93	0.54	5.42	3.99	4.61	0.43	5.44	3.96	4.67	0.47	5.57	4.66	5.14	0.28
CaO	21.65	1.67	13.80	6.32	17.32	1.32	8.15	4.42	15.84	1.38	7.31	3.94	4.47	1.27	2.64	0.93
Na2O	1.20	0.04	0.34	0.38	0.27	0.05	0.12	0.08	0.24	0.02	0.08	0.07	1.08	0.22	0.55	0.28
K2O	1.49	0.50	0.87	0.31	0.56	0.29	0.43	0.10	0.52	0.23	0.39	0.09	0.59	0.25	0.35	0.10
TiO2	1.02	0.40	0.65	0.15	0.75	0.39	0.61	0.12	0.77	0.35	0.61	0.13	0.34	0.12	0.19	0.07
P2O5	0.21	0.09	0.15	0.04	0.23	0.03	0.13	0.06	0.26	0.03	0.13	0.07	0.02	0.01	0.01	0.00
LOI	26.20	16.19	20.88	3.32	28.34	22.11	24.20	1.79	26.72	20.87	22.94	1.62	23.78	19.00	21.35	1.26
Total	100.00	98.43	99.31	0.60	100.80	99.18	100.24	0.52	100.80	98.59	99.75	0.72	100.90	98.70	99.53	0.81

and

Table 4. Maximum (Max), minimum (Min), and mean values of trace elements (µm) of all the granulometric fractions, along with their standard deviation (SD).

	Bulk Sample				<63 µm				<38 µm				<2 µm
	Max	Min	Mean	SD	Max	Min	Mean	SD	Max	Min	Mean	SD	Max	Min	Mean	SD
Zr	119.00	35.00	72.44	23.22	122.00	49.00	69.67	22.23	127.00	44.00	65.00	24.09	56.00	31.00	40.33	8.16
Nb	6.60	3.40	4.76	0.96	6.10	2.00	4.21	1.26	6.10	2.10	4.24	1.21	1.30	0.20	0.76	0.28
Hf	2.70	1.30	1.89	0.40	3.20	1.70	2.19	0.50	3.40	1.60	2.12	0.55	2.20	0.90	1.49	0.36
Ta	0.74	0.23	0.35	0.15	0.71	0.09	0.30	0.17	1.25	0.52	0.78	0.20	0.15	0.01	0.08	0.04
Tl	0.38	0.16	0.26	0.08	0.21	0.08	0.11	0.04	0.14	0.05	0.09	0.03	0.07	0.05	0.05	0.01
Th	5.22	0.79	2.86	1.83	4.78	0.77	2.54	1.59	5.21	0.87	2.70	1.70	3.27	0.40	1.59	1.17
U	1.22	0.43	0.80	0.30	1.21	0.55	0.71	0.21	1.19	0.50	0.71	0.22	0.24	0.09	0.15	0.04
Sc	33.00	18.00	27.33	4.52	40.00	20.00	30.89	5.26	40.00	21.00	31.22	5.65	42.00	20.00	31.00	5.93
V	265.00	53.00	111.67	61.02	194.00	45.00	89.22	47.79	208.00	46.00	93.89	52.20	171.00	27.00	81.44	48.44
Cr	810.00	20.00	400.00	256.95	850.00	30.00	443.33	282.13	1080.00	40.00	528.89	372.48	890.00	60.00	442.22	302.65
Co	34.00	10.00	21.89	6.35	17.00	11.00	14.33	1.89	22.00	10.00	17.44	3.72	17.00	6.00	11.44	3.59
Ni	120.00	30.00	68.89	34.78	100.00	20.00	51.25	30.18	130.00	20.00	61.25	40.45	120.00	30.00	55.71	32.89
Cu	40.00	20.00	25.00	7.07	30.00	10.00	17.50	6.61	30.00	10.00	20.00	5.00	20.00	10.00	15.00	5.00
Zn	80.00	50.00	64.44	8.31	80.00	40.00	48.89	12.86	90.00	40.00	55.56	14.99	60.00	30.00	43.33	10.54
Rb	68.00	19.00	43.44	17.68	47.00	15.00	24.78	9.59	44.00	13.00	23.67	8.73	16.00	7.00	10.89	2.51
Sr	175.00	81.00	122.56	26.58	98.00	80.00	90.44	5.89	96.00	80.00	88.56	6.45	133.00	93.00	108.78	12.73
Cs	3.70	0.70	2.42	0.95	2.30	0.80	1.51	0.48	2.20	0.60	1.43	0.48	1.50	0.60	1.08	0.27
Ba	604.00	176.00	348.56	145.07	363.00	106.00	225.33	71.50	327.00	98.00	208.67	67.81	556.00	111.00	277.89	153.05
Pb	11.00	5.00	8.25	2.17	27.00	6.00	13.67	9.46	8.00	5.00	6.00	1.22	12.00	5.00	8.50	3.50
La	24.30	6.73	13.97	6.35	21.20	3.11	9.98	5.45	19.20	2.77	9.30	4.89	8.57	0.71	4.27	2.99
Ce	36.80	15.70	24.77	8.06	35.90	10.40	20.39	7.91	33.60	10.40	20.69	7.70	26.20	2.93	12.40	7.09
Pr	6.34	2.47	4.05	1.28	5.27	1.64	2.85	1.11	5.06	1.44	2.77	1.05	1.84	0.18	0.97	0.65
Nd	26.50	10.30	17.42	5.46	23.70	7.65	12.59	4.97	22.60	6.46	11.81	4.72	6.39	0.88	3.61	2.25
Sm	6.34	2.69	4.27	1.29	5.77	2.02	3.09	1.14	5.30	1.73	2.88	1.10	1.19	0.19	0.65	0.37
Eu	1.55	0.67	1.06	0.30	1.35	0.47	0.76	0.26	1.16	0.47	0.71	0.22	0.23	0.05	0.13	0.07
Gd	6.75	2.43	4.06	1.33	5.96	2.00	3.03	1.26	4.84	1.67	2.64	1.04	0.88	0.19	0.47	0.23
Tb	1.09	0.37	0.65	0.21	0.91	0.33	0.49	0.18	0.77	0.30	0.44	0.16	0.12	0.03	0.07	0.03
Dy	6.52	2.20	3.92	1.26	5.34	2.05	2.87	1.09	4.56	1.80	2.62	0.99	0.65	0.15	0.38	0.15
Ho	1.30	0.43	0.76	0.26	1.06	0.39	0.57	0.22	0.90	0.36	0.51	0.19	0.12	0.03	0.07	0.03
Er	3.72	1.30	2.18	0.70	3.12	1.14	1.67	0.63	2.54	1.03	1.47	0.54	0.34	0.10	0.21	0.08
Tm	0.54	0.20	0.33	0.10	0.45	0.17	0.24	0.09	0.36	0.16	0.21	0.07	0.05	0.01	0.03	0.01
Yb	3.29	1.36	2.11	0.55	2.90	1.14	1.59	0.54	2.35	1.07	1.44	0.45	0.33	0.08	0.19	0.07
Lu	0.46	0.21	0.32	0.08	0.44	0.17	0.25	0.08	0.38	0.17	0.23	0.07	0.05	0.01	0.03	0.01
Be	3.00	1.00	1.89	0.57	3.00	1.00	1.67	0.67	2.00	1.00	1.33	0.47	2.00	1.00	1.50	0.50
Ga	15.00	10.00	12.44	1.77	15.00	11.00	12.78	1.47	15.00	12.00	13.11	1.10	14.00	12.00	13.22	0.63
Ge	1.40	0.70	0.94	0.23	0.90	0.50	0.70	0.14	0.90	0.50	0.69	0.14	0.70	0.60	0.65	0.05
Y	39.50	12.30	21.91	7.86	32.90	11.80	16.09	6.94	26.60	10.00	14.46	5.71	3.00	0.80	1.82	0.69
Sn	5.00	1.00	2.33	1.05	6.00	2.00	2.89	1.37	6.00	2.00	3.22	1.40	5.00	2.00	2.78	1.13
Sb	0.30	0.20	0.23	0.04	3.50	0.20	1.54	1.02	0.40	0.20	0.27	0.09	135.00	0.50	38.03	46.11
W	4.00	0.60	1.43	1.03	2.70	0.50	1.26	0.68	8.00	0.90	2.38	2.22	1.90	0.60	1.05	0.42

show the maximum, minimum, and mean values of the chemical analyses of all the granulometric fractions, along with their standard deviation. The trace elements considered were the high field strength elements (HFSE), transition trace elements (TTE), large-ion lithophile elements (LILE) and rare earth elements (REE), as well as other trace elements such as Be, Ga, Ge, As, Y, Mo, Ag, In, Sn, Sb, W and Bi.

The major elements concentrations are in accordance with the mineralogical characterization. Samples with a higher content in phyllosilicates present higher percentages of SiO₂ and Al₂O₃, consistent with their identification as being mainly constituted by dioctahedral smectites, and samples with a higher content in calcite have more CaO (Table S2). This is observed in Figure 7, where the contents of SiO₂, Al₂O_3, and CaO are plotted, and it is possible to observe an enrichment in SiO₂ and Al₂O₃ and a depletion in CaO towards the finer grained fractions (higher content in phyllosilicates, Table 3). The trends observed are corroborated by the high correlation coefficients among SiO₂, Al₂O_3, and CaO shown in Table S4 and by Figure 8, which shows the contents of these elements in the finer grain size fractions normalized with respect to the contents in the bulk rock samples.

Focusing on the HFSE, there are no clear differences according to the mineralogy of the samples, as shown by the correlation matrices performed (Table S5). Nb, Tl, Th, and U are significantly correlated with quartz, and it also appears that samples with higher contents in phyllosilicates also have lower total contents in HFSE, showing a progressive depletion towards finer granulometries (Table 4, Figure 8). This is explained by their relative immobility and their association with heavier and more resistant minerals present in the coarser fractions [60,61,62,63,64].

The TTE shows different correlation coefficients between phyllosilicates and carbonates, being negative in the first case and positive in the latter, considering the total TTE content as well as the content in certain elements within the group (Co, Ni, Cu) (Table S6). However, when observing Figure 8, it is possible to observe that the trend is not as clear as expected, even having an enrichment in TTE in sample (BEN3C), which is almost entirely formed by smectite (98%) and small quantities of calcite (2%). This points to an unclear linkage between TTE and mineralogy.

LILE appear to be less concentrated in samples with higher phyllosilicate contents, shown by their depletion in finer granulometries in comparison to the bulk rock fraction (Table 4, Figure 8), except in the case of the finest fraction (<2 µm). However, the correlation matrix (Table S7) only shows a negative significant correlation of phyllosilicates and Sr, because this element shows higher concentrations in calcite-rich samples due to its similarity to Ca [65,66,67]. As with the TTE, it is not possible to identify a clear link between LILE and mineralogy, with the exception of Sr and carbonates.

The REE present lower concentrations in the samples with higher contents in phyllosilicates (Figure 8, Table 4 and Table S8), something that can be attributed to their low solubility and their preferential accumulation in the coarser fractions in trace minerals as zircons [68]. However, there is another mechanism that can potentially affect the REE content: their complexation due to the action of carbonate-rich fluids and possible precipitation within calcite veins [69,70,71]. After chondrite normalization [72] (Figure 9), all samples are enriched both in light REE (LREE) and heavy REE (HREE). The fact that, in the majority of the clay size (<2 µm) fractions, the HREE are much lower than in the other fractions may partially explain the negative correlation between REE and smectite. All the samples studied show negative Eu anomalies. Two samples from the B profile have significant negative Ce anomalies in bulk and silt size fractions. In the clay size fractions, a tendency for positive anomalies was found.

The other trace elements analyzed show that most of them are below detection limits in most samples. Therefore, it is important to focus on the elements above the aforementioned detection limits. Yttrium shows high affinity with the REE and therefore has the same tendency as them (Table S9), with its content being lower within more clay-rich samples (Table S3). On the other hand, Ga is apparently more concentrated in samples with higher contents of phyllosilicates (Figure 8, Table 4 and Table S9), because it can substitute Al³⁺ in the octahedral layer of these minerals [73,74]. In the case of the rest of the other elements analyzed, it is not possible to identify clear links between their content and the mineralogy.

5. Discussion

The mineralogical characterization allows observing a variation in the horizontal direction, suggesting that the alteration process that formed the bentonites followed specific pathways. The presence of amphiboles within the igneous rocks whose alteration gave place to the formation of the clay minerals is cited in the literature [23,24]. Therefore, the presence of those minerals is interpreted as an indicator of a lower degree of weathering of the C profile in comparison with the other vertical profiles. The high content of carbonates in different parts of the sampling grid is related to the circulation of carbonate-rich fluids and the resulting carbonated crusts during the Paleogene–Quaternary [24], events that happened after the weathering that gave place to the bentonite deposit. It can be concluded that the mineral composition of the samples reflects an almost complete weathering of the primary minerals (amphiboles, quartz, feldspars, and micas), as these are present in relatively small proportions. Even if their content was recalculated, obviating the presence of carbonates (deposited after the weathering event), they still represent minor quantities, corroborating their almost complete transformation.

The TEM photographs show an apparent increase in the wavy and pointy-ended faces towards the A profile, therefore evidencing an increase in the development of smectites [46,47,48,49]. The TEM-AEM analyses show an increment of the Al₂O₃ and SiO₂ (more abundant within dioctahedral smectites) content from the C profile towards the A profile, as well as a clear decrease in Fe₂O₃ and TiO₂ (less abundant within dioctahedral smectites and more abundant within illite and amphiboles). Additionally, the presence of Fe₂O_3, TiO_2, and K₂O together with the particle morphology is a clear indicator that the precursor phases of the smectites were micas. The results of the smectite crystal-chemical characterization reveal interesting interrelated facts. The first one is that there is a progressive increase in the percentage of cv configurations of the octahedral sheet towards sample BEN3C (Table S1). The TEM-AEM study of the samples also enables observing a shift from beidellitic montmorillonite in the A profile to montmorillonitic beidellites in the B and C profiles. In agreement with the interpretations from the XRD characterization of the bulk samples (lower degree of alteration of the C vertical profile deduced from the identification of amphiboles in XRD), it is possible to suggest that lower percentages of tv configurations and beidellitic montmorillonites are characteristic of less weathered samples, while higher percentages of tv configurations and montmorillonitic beidellites are characteristic of samples which underwent a higher degree of weathering.

Major elements match the expected distribution, in accordance with the mineralogical characterization of the samples. These data were used to calculate two chemical indices, which are adequate for silicate-rich materials [39,40,42,44] and indicate the degree of alteration of the samples (

Table 5. Plagioclase index of alteration (PIA) and chemical index of alteration (CIA) of all the samples and granulometric fractions, along with the indices obtained through TEM-AEM.

	Whole Rock		<63 µm		<38 µm		<2 µm		TEM-AEM
	PIA	CIA	PIA	CIA	PIA	CIA	PIA	CIA	PIA	CIA
BEN1A	50.68	55.80	69.69	74.79	72.16	76.25	81.81	84.14
BEN1B	33.53	40.18	53.07	60.36	56.95	62.68	72.57	73.37
BEN1C	51.81	55.59	67.61	72.63	69.83	73.87	82.40	84.69
BEN2A	49.78	52.89	54.13	61.38	69.85	73.42	84.99	86.00
BEN2B	30.80	37.44	66.81	71.61	52.23	58.13	75.09	74.16
BEN2C	35.03	34.18	51.74	55.23	57.36	59.62	60.17	62.38
BEN3A	76.09	81.27	84.27	88.20	85.09	88.46	85.02	84.47	91.90	90.41
BEN3B	25.46	33.27	31.93	42.86	33.77	42.47	76.89	78.73	86.57	84.61
BEN3C	27.89	30.79	38.92	45.51	44.75	49.05	72.56	74.11	88.73	76.77

, Figure 10). These indices are the plagioclase index of alteration (PIA), which considers the content in Al₂O₃, CaO, Na₂O and K₂O [41]; and the chemical index of alteration (CIA), which considers the molecular percentages of Al₂O₃, Na₂O, K₂O and the CaO (CaO*) that is solely contained within the silicates [41,43]. The CaO* was also used in the calculation of the PIA index to avoid the influence of carbonates. Both chemical indices show a positive correlation between higher values and a higher alteration degree, with the highest values indicating higher clay mineral contents. A correlation matrix analysis (Table S10) reveals strong positive and significant correlations (0.896 and 0.917) with phyllosilicates and negative and significant coefficients with calcite and amphiboles, validating their calculation.

Despite the expected decline in the PIA and the CIA values from the top to the bottom of the vertical profile due to the field interpretation, it is possible to observe an increase in the values of the indices from top to bottom, indicating more weathered materials at the base of the deposit. However, there is an exception, the A profile, which shows higher values at the top. This can be related to the fracture system running in the NNE–SSW direction that seems to have favored the weathering [26], since it is possible that those fluids came from the NNE and affected the A profile in a slightly different manner. A clear link between the alteration indices and the amount of primary minerals (tectosilicates and amphiboles) is observed. The variation along the horizontal levels shows values that are typically higher towards the A profile, although the B and C profiles have higher variability in their values. The circulation of the carbonated fluids, especially in the B profile, affects the resulting CIA and PIA indices (even after the neglection of the Ca contained within carbonates by the use of the CaO* for their calculation), obscuring their true values and making it challenging to determine the weathering pathways of the deposit (Table 5, Figure 10). When calculating these indices for the finer granulometries, it is possible to observe how the influence of the carbonated fluids is progressively less important (as indicated by their higher values), especially at the <2 µm fraction, which is essentially formed of phyllosilicates. However, these fluids still have some influence on the calculated indices.

To completely avoid the influence of these carbonated fluids on the calculation of the indices and only focus on pure silicates, for which these indices were thought, the results from the TEM-AEM point analyses of the non-homoionized samples were used (Table 5), to avoid the additional input of Ca during the homoionization procedure. The indices calculated this way have slightly higher values than the ones calculated using the ICP-MS analyses of all the granulometric fractions. This is due to the presence of non-silicate minerals (carbonates) within the samples analyzed by ICP-MS, indicating that the corrections performed to completely neglect carbonates to calculate the indices are not 100% efficient. The resulting CIA indices are the most consistent with the previous data (from the mineralogical and crystal-chemical characterizations), showing a higher degree of alteration of the A profile, followed by the B profile, and ending with the C profile. It is especially interesting considering that there is a fracture system which seems to have favored the weathering of the samples running in NNE–SSW direction [26] and that the A profile is located towards the NE and the C profile towards the SW, which can point to a northern origin of the fluids that weathered the samples.

Although clay minerals can adsorb REE or even incorporate them in octahedral positions [75], providing excellent recipients for REE-rich deposits, the richest fractions in phyllosilicates are found to be depleted in such elements. This can be explained by a preferential accumulation of REE in the area within coarser minerals, such as zircons, as well as by their complexation and mobilization by the carbonate-rich fluids and their precipitation within calcite veins [68,69,70,71]. This would explain the significant correlation coefficients that REE have both with quartz and with calcite (Table S8). The Ce and Eu anomalies shown in Figure 9 and detailed in

Table 6. Ce and Eu anomalies of all the samples and granulometric fractions.

	Whole Rock		<63 µm		<38 µm		<2 µm
	Ce/Ce*	Eu/Eu*	Ce/Ce*	Eu/Eu*	Ce/Ce*	Eu/Eu*	Ce/Ce*	Eu/Eu*
BEN1A	1.01	0.81	1.11	0.74	1.19	0.82	1.13	0.79
BEN1B	1.04	0.79	1.30	0.71	1.43	0.81	2.27	0.76
BEN1C	1.24	0.78	1.60	0.81	1.59	0.83	1.87	0.74
BEN2A	0.93	0.78	0.42	0.71	1.14	0.84	1.46	0.76
BEN2B	0.31	0.72	1.07	0.78	0.40	0.71	0.77	0.68
BEN2C	1.12	0.80	1.24	0.87	1.44	0.84	2.55	0.75
BEN3A	1.05	0.81	1.14	0.80	1.14	0.88	1.05	0.75
BEN3B	0.33	0.77	0.36	0.70	0.49	0.70	1.85	0.72
BEN3C	0.93	0.77	1.04	0.81	1.18	0.88	1.80	0.70

were calculated as detailed by previous authors [76]. All the samples present negative Eu anomalies. The origin of this anomaly is due to the dissolution of Eu-rich minerals (plagioclase) present on the parent rock by weathering processes [77,78]. On the other hand, there are samples with positive Ce anomalies and others with negative Ce anomalies, indicating the first ones have more oxidizing environments than the second ones [79,80]. In Table S11, a correlation matrix between the values of these anomalies is displayed. It is possible to observe that the Eu anomaly is not correlated with any mineral, probably because there are no major differences within its values. However, the Ce anomaly shows significant positive correlations with the phyllosilicates and a highly significant negative correlation with calcite. This indicates that the weathering that originated phyllosilicates took place in more oxidizing environments, while the negative anomalies are probably caused by more reducing conditions during the precipitation of calcite due to the circulation of carbonate-rich fluids. This corroborates the important role these fluids play in the chemistry of the preexisting materials and how they can impact the indices calculated as well as their interpretations.

6. Concluding Remarks

The Benavila deposit is mainly formed by dioctahedral smectites, followed by carbonates (Paleogene–Quaternary carbonated crusts due to the circulation of carbonate-rich fluids) and minor amounts of tectosilicates and other phyllosilicates. The almost complete transformation of primary minerals (as tectosilicates, amphiboles, and micas) suggests an intense weathering of the parent rock. In addition, the presence of amphibole within the C vertical profile is interpreted as indicative of a lower degree of alteration due to the primary character of that mineral.

The crystal-chemical characterization reinforces the importance of performing a homoionization of the samples prior to their analysis. The smectites studied are classified as ferrian montmorillonitic beidellites and beidellitic montmorillonites. Smectites with octahedral configurations with lower percentages of tv sites and classified as montmorillonitic beidellites could also indicate a lower degree of weathering within the deposit.

The geochemical analysis of the samples indicates a preferential depletion of the more clayey materials in HFSE, REE, and Y, among other heavy elements. In addition, it also points towards the pathway of weathering along which the bentonite deposit originated. However, the presence of carbonated crusts of Paleogene–Quaternary age due to the circulation of carbonated fluids overprints the deposit in certain parts and hardens the task of obtaining useful weathering indices.

The values of the CIA index are consistent with the aforementioned mineralogical and crystal-chemical criteria regarding the weathering of the parent rocks that originated the Benavila deposit. This study reveals the alteration pathways of the Benavila deposit (NE-SW direction, with the most weathered materials towards the NE) as well as the smectite variability within it. The variations of smectite mainly originate from lower degrees of weathering and the circulation of carbonated fluids that changed the already weathered bentonitic deposit, possibly affecting the properties of the bentonite.