Structural Diagenesis in Clay Smearing Bands Developed on Plio-Pleistocene Sediments Affected by the Baza Fault (S Spain)

: This study reveals mineral and deformation processes associated with faulting of lacustrine unconsolidated sediments in the Guadix-Baza Basin (Betic Cordillera, S Spain) affected by the Baza Fault. Brittle carbonate and silt sediments develop deformation bands frequently sealed by dolomite crystallization, whereas ductile clay-rich sediments form clay smearing bands where late crystallization of gypsum can be observed. Granular ﬂow and local cataclasis were the main deformation mechanisms in the brittle deformation bands. Flow alignment, grain-boundary sliding, and extrusion were predominant in the clay smearing bands. These water and clay-rich bands reduced shear strength of the faulting process due to their lubricating effect. Beidellitic smectite deﬁnes shear foliation of the smeared bands, but Mg-Fe, a K-rich smectite (Fe + Mg > 1 and K content up to 0.8 a.p.f.u), crystallizes in the micropores surrounding brittle clasts produced by deformation pressure shadows. These data suggest that the interaction of micromechanical events, which increased sediment porosity by the generation of pressure shadows, and the ﬂow and concentration of saline ﬂuids in these pores promoted structural diagenesis processes that favoured the beginning of local illitization.


Introduction
Clay-rich beds play an important role in accommodating fault deformation, mainly by clay smearing processes [1]. Interbedded clay-poor sediments in fault areas develop small fractures and, frequently, deformation bands [2,3]. When materials are poorly lithified and water-rich, sediments from the fault area are mechanically mixed even to the microscale [4]. The clay mineral fraction in fault zones can be produced by mechanical transformation, both brittle (disaggregation, fracturing, cataclasis, grain-boundary sliding), e.g., [5,6], and ductile (grain deformation, smearing, injection), e.g., [7][8][9] and/or mineral/chemical changes during diagenetic/metamorphic reactions (pressure dissolution, diffusive mass transfer, clay precipitation), e.g., [10][11][12][13]. Moreover, water saturation during deformation can especially promote mineral chemical transformations [14][15][16]. Therefore, the origin of small-sized clay minerals in clay bands from fault areas can be associated with detrital, cataclastic and neoformation processes. Laurich et al. [17] suggested that in clay gouges, neoformation can contribute more significantly to producing small-sized clay grains than mechanical deformation processes. Solum et al. [16] indicated that clay authigenesis in fault areas is very frequent in many geological contexts, suggesting that fluids related with fault activity were responsible for transformation processes.
These reactions, commonly referred to as structural diagenesis, contrast with the mineral formation controlled by time and temperature during burial diagenesis. Nanometresized clay grains produced by deformation increase reactive clay surfaces which are responsible for processes of adsorption-desorption involved in many clay transformations. Moreover, tectonically controlled fluid flow in fault areas can generate high salinity pore areas that favour changes in the charge of the octahedral layer (i.e., by the substitution of

Geological Context and Materials
We studied deformed materials from the Baza Fault, a 37-km-long structure, striking N-S to NW-SE and dipping around 65 • to the east ( Figure 1) [27,28], which is situated in the central Betic Cordillera. This is an active structure which has been controlling the tectonic and sedimentary development of the Guadix-Baza Basin since the Miocene due to an extensional process that created normal faults [29]. The width of the fault area oscillates from a narrow zone less than 300 m in the northern segment, due to the presence of only one main surface, to a wider southern part (around 7000 m), where it appears up to 13 strands. A 2000 m fault throw was accumulated [29]. Slip rate estimations for the fault oscillate from 0.49 to 0.12 mm/year [27,28,30].
The studied materials belong to the east segment of the Guadix-Baza Basin (Figure 1a), the most important basin of the Betic Cordillera (Spain) during the Neogene. The basin is filled with marine sediments (upper Miocene) and continental materials (fluvial and lacustrine, Pliocene/Pleistocene) [31]. Gibert et al. [32] distinguished three facies zones in the basin (marginal, intermediate and inner). We characterized a set of samples from the "Benamaurel Unit", which comprises sediments of the intermediate and inner facies areas of the basin, including lacustrine carbonates, gypsum-rich sediments, marls, and dark clays with native sulphur.
We studied deformed sediments from a trench (Carrizal trench) near the Barranco del Agua, located at the central part of the Baza Fault (Figures 1b and 2). These materials are affected by two main strands (NNW-SSE orientation) of the Baza Fault, which concentrate most of the movement. Three main types of sediments can be observed in the trench ( Figure 3): white lacustrine carbonates where dolomite predominates over calcite, dark grey lutite/clays, and silts. The ages of these materials oscillate between 3 and 1 × 10 6 years [28].
These materials are situated between two surfaces of the fault zone where deformation is concentrated, producing intense fracturation and disruption of the carbonate, silts and clay levels. Faults are more common in carbonate and silt levels but they are not propagated into the clay beds. Medina-Cascales et al. [26] described the fact that that clay beds are partially continuous and frequently form thin, smeared and injected levels. We collected 11 clay bodies and 12 carbonate and silt levels. Samples were collected from an excavated trench at the Carrizal fault strand (Figure 2), mostly oriented perpendicular to the fault strike, resulting in a total excavation volume of ∼15 m × 15 m × 4 m (See Medina-Cascales et al. [26] for details). Oriented sediment cores were obtained with a stainless tube, ensuring no disruption to the microstructure.      These materials are situated between two surfaces of the fault zone where deformation is concentrated, producing intense fracturation and disruption of the carbonate, silts and clay levels. Faults are more common in carbonate and silt levels but they are not propagated into the clay beds. Medina-Cascales et al. [26] described the fact that that clay beds are partially continuous and frequently form thin, smeared and injected levels. We collected 11 clay bodies and 12 carbonate and silt levels. Samples were collected from an excavated trench at the Carrizal fault strand (Figure 2), mostly oriented perpendicular to the fault strike, resulting in a total excavation volume of ∼15 m × 15 m × 4 m (See Medina-Cascales et al. [26] for details). Oriented sediment cores were obtained with a stainless tube, ensuring no disruption to the microstructure.

Methods
Milled powders and oriented preparations (total sample and <2 μm fractions) were used for the X-ray-diffraction (XRD) study of the sediments. Materials were treated with ultrapure water for salt elimination. Centrifugation was used for obtaining < 2 μm and <0.2 μm fractions. We applied ethylene glycol (EGC) for the identification of expandable clays. An XDR study was carried out with a PANalytical X'Pert Pro diffractometer (CuKα emission, 40 mA, 45 kV) including an X'Celerator solid-state linear detector. A step increment of 0.008° 2θ and a total counting time of 10 s/step was used. Dry samples were scanned from 3° to 62° 2θ on, and glycolated preparations were studied from 2° to 30° 2θ.

Methods
Milled powders and oriented preparations (total sample and <2 µm fractions) were used for the X-ray-diffraction (XRD) study of the sediments. Materials were treated with ultrapure water for salt elimination. Centrifugation was used for obtaining <2 µm and <0.2 µm fractions. We applied ethylene glycol (EGC) for the identification of expandable clays. An XDR study was carried out with a PANalytical X'Pert Pro diffractometer (CuKα emission, 40 mA, 45 kV) including an X'Celerator solid-state linear detector. A step increment of 0.008 • 2θ and a total counting time of 10 s/step was used. Dry samples were scanned from 3 • to 62 • 2θ on, and glycolated preparations were studied from 2 • to 30 • 2θ. HighScore 5.1 software, PANalytical (Eindhoven, The Netherlands) with decomposition routines was used. A mixed Gaussian and Lorentzian Voigt function was applied for peak fitting.
Polished sections were used for textural and compositional characterization with scanning electron microscopy (SEM, Merlin Carl Zeiss electron microscope) operating with a back-scattered electrons (BSE) detector under atomic number contrast mode. Sediment portions were studied by secondary electrons (SE). BSE pictures were acquired using an AsB detector. Conventional and In-Lens detectors were used for obtaining SE images at 15 kV.
Nanometer scale characterization was carried out with a transmission electron microscopy (TEM). Coated Au and Cu nets were employed for preparing samples from a dispersion of finely ground minerals. We checked the monomineralic character of grains with electron diffraction patterns. Moreover, we extracted lamellae from selected smeared samples impregnated with a Dual Beam Helios 650 Focused Ion Beam (FIB). We obtained 4 µm × 3.5 µm lamellae,~60-70 nm thick.
The TEM data was obtained using a HAADF FEI TITAN G2 instrument, working at 300 kV. Analytical electron microscopy (AEM) was used for obtaining quantitative analyses of clays with an EDAX detector coupled to the HAADF FEI TITAN G2 electron microscope. The counting time was established as 100 s except for Na and K (15 s), to avoid problems of alkali-loss [33]. Cliff and Lorimer [34] procedures were employed for obtaining k factors from the following standards: CaS, MnS, albite, spessartine, biotite, titanite, muscovite and olivine.

Carbonate Levels
XRD data indicate that dolomite is the most abundant component of the carbonate sediments ( Figure 4a). Phyllosilicates, quartz and feldspars have low contents. The clay assemblage of the total sample and the <2 µm fraction is mainly made of muscovite, paragonite, and kaolinite. A low quantity of smectite (the peak at 14.40 Å) is also detected in the clay assemblage ( Figure 4b). A <0.2 µm size fraction was characterized to determine the specific minerals forming the 14.40 Å peak ( Figure 4c). Air-dried (AD) and EGC patterns completely overlap around the 9 • 2θ (002) area, suggesting the absence of an illite-smectite mixed layers phase. SEM images show that smectite appears as dispersed flakes in the carbonate matrix ( Figure 5). Moreover, AEM analyses reveal that the nature of the smectite of the carbonate sediments is Al-rich dioctahedral. Fe, Mg and K contents are very low in these smectites (Table 1). croscopy (TEM). Coated Au and Cu nets were employed for preparing sa dispersion of finely ground minerals. We checked the monomineralic chara with electron diffraction patterns. Moreover, we extracted lamellae from sele samples impregnated with a Dual Beam Helios 650 Focused Ion Beam (FIB) ~4 μm × 3.5 μm lamellae, ~60-70 nm thick.
The TEM data was obtained using a HAADF FEI TITAN G2 instrumen 300 kV. Analytical electron microscopy (AEM) was used for obtaining qua yses of clays with an EDAX detector coupled to the HAADF FEI TITAN G croscope. The counting time was established as 100 s except for Na and K ( problems of alkali-loss [33]. Cliff and Lorimer [34] procedures were employ ing k factors from the following standards: CaS, MnS, albite, spessartine, b muscovite and olivine.

Carbonate Levels
XRD data indicate that dolomite is the most abundant component of sediments ( Figure 4a). Phyllosilicates, quartz and feldspars have low cont assemblage of the total sample and the <2 μm fraction is mainly made paragonite, and kaolinite. A low quantity of smectite (the peak at 14.40 Å) is in the clay assemblage ( Figure 4b). A <0.2 μm size fraction was characterized the specific minerals forming the 14.40 Å peak (Figure 4c). Air-dried (AD) terns completely overlap around the 9° 2θ (002) area, suggesting the absen smectite mixed layers phase. SEM images show that smectite appears as dis in the carbonate matrix ( Figure 5). Moreover, AEM analyses reveal that the smectite of the carbonate sediments is Al-rich dioctahedral. Fe, Mg and K very low in these smectites (Table 1).

Silt Levels
Silt levels are characterized by the presence of quartz, feldspars, and muscovite, paragonite, and chlorite as phyllosilicates. These minerals form a massive matrix hosting irregular deformation bands, which contain clasts with irregular morphologies of quartz, feldspars and detrital phyllosilicate. Two types of deformation bands are identified: (a) Disaggregation bands. The sediments with coarser grains (50-200 µm) show thick bands (up to 300 µm) developing granular flow processes (grain-boundary sliding or grain rolling) ( Figure 6). Mixing with sediments of smaller size is also observed at distinct bands (Figure 6a). An increase of grain angularity and a decrease of grain size occur due to local grain cracking (Figure 6b,c). Dolomite appears as cement in these bands.
(b) Phyllosilicate bands. Sediments with high contents of coarse muscovite, paragonite and chlorite grains (around 30 µm) develop shear-induced rotation producing phyllosilicate alignment, which forms a special kind of disaggregation band with local fabrics, where the disposition of platy minerals favours frictional grain-boundary sliding (Figure 6d). Dolomite can be observed as cement also in these bands.
Minerals 2022, 12, 1255 7 of size occur due to local grain cracking (Figure 6b,c). Dolomite appears as cement these bands. (b) Phyllosilicate bands. Sediments with high contents of coarse muscovite, paragoni and chlorite grains (around 30 μm) develop shear-induced rotation producing phy losilicate alignment, which forms a special kind of disaggregation band with loc fabrics, where the disposition of platy minerals favours frictional grain-bounda sliding (Figure 6d). Dolomite can be observed as cement also in these bands.

Dark Clays
Dark clay-rich levels are characterized by the abundant presence of phyllosilicate quartz and feldspars. XRD data (total sediment and <2 μm fractions) show a clay assem blage made of smectite, muscovite and low quantities of kaolinite, chlorite and paragoni The presence of chlorite, slightly hidden by the smectite (001) reflection in AD diagram is revealed in the EGC treated samples (Figure 7). Broad (001) 12-15 Å peaks appe around 17 Å in the ethylene glycol treated samples. When the AD and EGC treated di grams are superposed, the EGC diagram shows a slightly asymmetric area of higher i tensity around the theoretical position of the (002) smectite peak around 8.5 Å, whi could be produced by the presence of dispersed illitic layers in some smectites [35].
From the textural point of view, clay minerals show evidence of flow, reorientatio and extrusion produced by deformation to form clay smearing bands at micrometric sca (Figure 8a-c), where slip surfaces can also occur.

Dark Clays
Dark clay-rich levels are characterized by the abundant presence of phyllosilicates, quartz and feldspars. XRD data (total sediment and <2 µm fractions) show a clay assemblage made of smectite, muscovite and low quantities of kaolinite, chlorite and paragonite. The presence of chlorite, slightly hidden by the smectite (001) reflection in AD diagrams, is revealed in the EGC treated samples (Figure 7). Broad (001) 12-15 Å peaks appear around 17 Å in the ethylene glycol treated samples. When the AD and EGC treated diagrams are superposed, the EGC diagram shows a slightly asymmetric area of higher intensity around the theoretical position of the (002) smectite peak around 8.5 Å, which could be produced by the presence of dispersed illitic layers in some smectites [35].
From the textural point of view, clay minerals show evidence of flow, reorientation and extrusion produced by deformation to form clay smearing bands at micrometric scale (Figure 8a-c), where slip surfaces can also occur. A late gypsum precipitation is observed sealing some deformation bands (Figure 8d). TEM images of the clay smearing bands reveal the presence of oriented beidellitic smectite with very fine grain size, muscovite, and paragonite defining the shear foliation produced by strain (Figure 9a). Bigger grain clasts of quartz, calcite and feldspar appear inside the clay smearing bands. These irregular large grains are enclosed by the oriented phyllosilicates (Figure 9b). Clay bending of the minerals surrounding the clasts can be observed.  A late gypsum precipitation is observed sealing some deformation bands (Figure 8d). TEM images of the clay smearing bands reveal the presence of oriented beidellitic smectite with very fine grain size, muscovite, and paragonite defining the shear foliation produced by strain (Figure 9a). Bigger grain clasts of quartz, calcite and feldspar appear inside the clay smearing bands. These irregular large grains are enclosed by the oriented phyllosilicates (Figure 9b). Clay bending of the minerals surrounding the clasts can be observed. A late gypsum precipitation is observed sealing some deformation bands (Figure 8d). TEM images of the clay smearing bands reveal the presence of oriented beidellitic smectite with very fine grain size, muscovite, and paragonite defining the shear foliation produced by strain (Figure 9a). Bigger grain clasts of quartz, calcite and feldspar appear inside the clay smearing bands. These irregular large grains are enclosed by the oriented phyllosilicates (Figure 9b). Clay bending of the minerals surrounding the clasts can be observed.  Pressure shadow micropores are developed by the clast fragments. Small nanoparticles of smectite, with random orientation, rich in Fe, Mg and K crystallize in these micropores (Figure 9c). Pressure shadow micropores are developed by the clast fragments. Small nanoparticles of smectite, with random orientation, rich in Fe, Mg and K crystallize in these micropores (Figure 9c). The AEM microanalyses reveal two groups of smectite composition. Smectites defining shear foliation in the clay smearing bands have beidellitic composition with high Al dioctahedral contents and low Mg, Fe, and K contents. However, smectites crystallized in the micro pressure shadows are richer in octahedral Mg and Fe (Mg + Fe > 0.9, a.p.f.u. adjusted to 11 oxygens), which produces an octahedral sum appreciably greater than 2 ( Table 1). Regarding the interlayer composition, the amounts of Na and K are high with a sum of interlayer cations frequently greater than 0.6 a.p.f.u. In some cases, K + Na content can be close to 0.8, near to an illitic composition. Assignation of Mg to the octahedral is The AEM microanalyses reveal two groups of smectite composition. Smectites defining shear foliation in the clay smearing bands have beidellitic composition with high Al dioctahedral contents and low Mg, Fe, and K contents. However, smectites crystallized in the micro pressure shadows are richer in octahedral Mg and Fe (Mg + Fe > 0.9, a.p.f.u. adjusted to 11 oxygens), which produces an octahedral sum appreciably greater than 2 ( Table 1). Regarding the interlayer composition, the amounts of Na and K are high with a sum of interlayer cations frequently greater than 0.6 a.p.f.u. In some cases, K + Na content can be close to 0.8, near to an illitic composition. Assignation of Mg to the octahedral is only considered for presentation purpose, but part of Mg could probably be present in the interlayer [36].

Mechanical Deformation
The predominant mechanism of deformation (brittle or ductile) can be controlled by pore fluid pressure [6] and mineral composition of the shear band, e.g. [9]. Overburdened water-saturated sediments that undergo deformation can experience fluidization processes that produce fluid-like and oriented structures, due to lateral escape [37,38]. Fluidization has been frequently related to seismic activity [28,39]. Moreover, platy morphology and low friction coefficients of phyllosilicate favour ductile deformation and foliation structure development [40]. Therefore, high content in water and clays favours ductile deformation. Sediments from the studied trench show a variety of deformation styles developed in poorly lithified sediments saturated in water during the deformation processes: (a) brittle deformation (fracturation and grain size reduction concentrated in deformation bands) in carbonate and silt sediments; and (b) ductile deformation (folding and smearing) in clay-richer sediments.
SEM images suggest shearing processes of carbonate beds and silicate-rich layers of silts, which developed flow and local cataclasis of grains leading to the formation of deformation bands. Rotevatn et al. [41] and Torabi et al. [42] described similar structures (e.g., millimetric local shear zones rather than individual glide surfaces, localized grain size reduction) during the fracturation process of partially unconsolidated sediments developed by grain cataclasis. Medina-Cascales et al. [26] suggested that the predominance of silts in the stratigraphic column affected by the Baza fault led to more brittle deformation.
The presence of dolomite and gypsum cements in these structures reveals intense fluid flow during deformation and late microsealing by precipitation from fluids during this process. Torabi et al. [3] and Romher et al. [43] indicated that precipitation of carbonates is responsible for permeability reduction in deformed sediments due to microsealing of fluid flow.
Clay smearing bands are mainly characterized by fluid-like features and ductile deformation. Shear strength can be reduced by the inclusion of water and clay-rich sediments along deformation bands, due to the lubricating effect of adsorbed water on the mineral surfaces by electrochemical forces [44,45]. In the studied sediments, the presence of beidellitic smectite in the shear structure of the bands, which is bent around the brittle clasts could produce a reduction of strength due to the microphysical processes (rotation, delamination and breaking of bonds) working during clay deformation [46][47][48].
The contact between clay and silt/carbonate beds is frequently characterized by detachment structures produced by the behaviour of the clay-rich beds. The attractive forces between clay and other mineral surfaces can allow easy mechanical disaggregation that favours detachment during shear [49]. Clay beds are strongly deformed to accommodate deformation by thinning and smearing, whereas the brittle silt layers adjust to deformation by fracturation. Sometimes, microsilty beds are inserted between the clay smears. Ductile layers enriched in clays can accommodate large proportions of strain during faulting [26,[50][51][52].
Deformation styles of clay and silty layers are interconnected. The sediments in the trench were saturated in water during deformation. Deformation of highly water-rich clay beds may produce fluidization processes leading to the release of fluids and material from the layered clay, causing a collapse [53,54]. The breakdown of these layers can produce fracturation and tilting of the adjacent brittle sediments of the studied trench. Fluidized clay can escape by these fractures, favouring smearing and injection structure development [1,55,56].

Chemical Deformation: Illitization Process
The results of this study suggest that the interaction of micromechanical processes that generated pores by pressure shadow, and the flow and concentration of saline fluids in these pores may have been involved in the authigenesis of clays. Thus, the presence of pressure shadow micropores promoted the interaction of beidellitic smectite of sedimentary origin, with the Mg-and Fe-rich saline fluids produced in the lacustrine basin where sediments were deposited.
A detrital origin for chlorite and mica (muscovite and paragonite) in sediments for the lacustrine sequence of the area has been suggested by Jiménez-Millán et al. [57] and Sánchez-Roa et al. [53], who proposed the metamorphic rocks of the Betic Cordillera as the source region. These original detrital minerals were transformed to kaolinite and smectite during wet climate events favouring the deposition of clay-rich levels in the lake.
Gypsum and dolomite crystal precipitation that seals microstructures produced by deformation, indicates that fluids integrated in the sediments by deformation were of high salinity. Gibert et al. [32] indicated that hydrological conditions in the eastern part of the lake in the Guadix-Baza Basin favoured the creation of Mg-rich hypersaline evaporitic brines.
Deocampo [58], Deocampo [59], and Deocampo et al. [60] suggested that porewater hydrochemistry controls the crystallization and nature of neoformed clay minerals, favouring Mg incorporation in phyllosilicates when Mg/Ca ratios rise in the remaining fluid. Diffusive mass transport during deformation favours pressure shadows produced in sediments rich in high salinity waters acting as restricted sites with extreme conditions that favour the processes of mineral alteration by pressure dissolution and authigenesis during clay smearing processes. In these environments, saline fluids can react with pre-existing clays, producing fast mineral transformations that favour K, Mg or Fe uptaking from fluids [61]. Clay mineral reactive surfaces play an important role in the adsorption of elements that can be included in the neoformed clays. Deocampo [59] showed that clays in saline fluids produced octahedral cation modifications that could raise layer charge, promoting the conversion of smectite to illite. The uptake of Mg and Fe and the beidellitic replacement created enough negative charge to enable the inclusion of K to begin an illitization process of low temperature related to microsites formed during clay smearing. Thus, the reaction of the illitization process in the pressure shadows was promoted by high K concentration in the pore waters and the Mg and Fe uptake in the octahedral sheet, which was produced by the coupled substitutions of Al for Si in the tetrahedral sheet and of Mg and Fe for Al. However, at a late stage, dolomite crystallization could have diminished Mg availability in the fluid, decreasing porosity of sediment, which weakened the processes of interaction between smectite and fluid and limited the advance of the illitization process.

1.
Mineral composition of the sediment is an important factor controlling brittle or ductile mechanism of deformation in the unconsolidated sediments of the Guadix-Baza Basin (Betic Cordillera, S Spain).

2.
SEM and TEM images suggest that flow and local cataclasis of grains were the main mechanisms involved in the shearing processes of carbonate beds and silicate-rich layers of silts, producing deformation bands. 3.
The predominance of beidellitic smectite in the shear structure of the clay smearing bands produced by the phyllosilicate-rich sediments, as well as the bending of these smectites around the brittle clasts suggest a ductile behaviour of these sediments.

4.
Precipitation of gypsum and dolomite sealing deformation bands could indicate that fluids integrated in the sediments by deformation were of high salinity, which may be related to the lacustrine waters of the basin.

5.
The interaction of micromechanical processes that generated pores as pressure shadow, and the flow and concentration of saline fluids in these pores were important variables promoting the authigenesis of clays and the structural diagenesis process.

6.
Hypersaline-restricted sites at the pressure shadows formed during clay smearing can react with pre-existing clays, promoting fast mineral transformations that could favour K, Mg or Fe uptaking from fluids. 7.
A future characterization of in situ measurements of rock properties (e.g., permeability and Young's modulus) will help to clarify the influence of mechanical deformation and the formation of authigenic clays on the variation of permeability in the Baza fault zone.