Introduction
The Brazilian economy is highly dependent on the mineral sector. Brazil is the owner of the biggest world reserves of niobium (98.2%), barite (53.3%) and graphite (50.7%), second biggest of tantalum (36.3%) and REE (16.1%), third biggest of nickel (13.7%) and tin (10.0%) and fourth biggest of iron (13,6%), manganese (8.8%) and vanadium (1.3%). The mineral sector employs more than one million people directly and indirectly and the sum of mineral extraction in 2013 was more than 77 billion dollars (4.1% of the Gross Domestic Product GDP) (MME, 2014).
Manganese has great importance in steelmaking industry (to increase the steel strength and resistance to corrosion), to decolourise glass, to make fertilizers, ceramics and especially for the batteries industry (Gonçalves & Serfaty, 1976). As its internal consumption and exportation increases, also increases the necessity of finding new deposits.
Ore is a material from which minerals and metals of intrinsic economic value or interest can be economically extracted at the present time (Peters, 1987; Misra, 1999; Moon et al., 2006; Guilbert & Park, 2007). A prospect is a restricted volume of ground, which is considered to have the possibility of directly hosting an ore body, selected based on some geological idea or an anomalous feature of the environment (Marjoribanks, 2010).
Once a prospect has been defined, exploration work advances through a series of progressive detailing stages where success leads to the next stage and negative results might cause the prospect to be discarded, sold or put on hold until new information, ideas or technology arrive. Exploration techniques will generally go through the main stages: target generation; target drilling; resource evaluation drilling and feasibility studies (Milsom & Eriksen, 2011; Marjoribanks, 2010).
One of the techniques used for the search and investigation of an orebody is exploration geophysics (Reedman, 1979). Its use begins in the reconnaissance stages, where airborne methods are applied in regional scales for the definition of new prospects, and continues into most detailed stages where ground methods are employed directly towards delimitating the orebodies in subsurface (Ford et al., 2007). Geophysical surveys are aimed at measuring rock properties (electrical, magnetic, density, mechanical, etc.), which may reflect or have straight relationships to economic mineralizations. Measurements that are considered anomalous (that is, above area background) are then analyzed to determine its nature, size, position and shape as a prelude for a follow-up detailed exploration stage, usually drilling (Keller & Frishknecht, 1970; Moon et al., 2006; Dentith & Mudge, 2014).
Electrical methods are vastly used in mineral exploration. As they are essentially ground methods, they are mainly applied to local studies like prospect investigation for subsequent drilling (Robinson, 1988). Amongst several geoelectrical methods, the Induced Polarization Method (IP) is important in base metal exploration because it depends on the surface area of the conductive mineral grains rather than their connectivity; therefore, the method is especially sensitive to disseminated mineralization (Keller & Frishknecht, 1970). IP effects can be very strong at the surface of grains of conducting minerals such as graphite and metallic sulfides. Graphite is generally considered a drawback for the method. It is often said to be a ''indicator of pyrite'' due to their similar (both high) polarization responses (Moon et al., 2006).
The IP method has been historically used for the search of disseminated sulfide and gold ores (and graphite in a lesser extent), in the most variable geological settings (Pelton & Smith, 1976; Moreira et al., 2012; Langore & Gjovreku, 1989; Allis, 1990; Izawa et al., 1990; Irvine & Smith, 1990; Locke et al., 1999; Dentith & Barrett, 2003). However, few references can be found about the use of IP method in the search and investigation of manganese primary and secondary ore deposits (Moreira et al., 2014; Ramazi & Mostafaie, 2013).
This paper presents the results of the IP method applied on the study of a supergene manganese occurrence in the southeast of Brazil, with the aim of analyzing the orebody architecture in subsurface based on the contrast between the ore minerals (Mn oxides, hydroxide and graphite) and host rocks.
Study area
The site of study is located in the northeast of São Paulo State, Brazil, around 6 km to the east of Itapira city (Figure 1).
The area is located in the Alto Rio Grande Fold Belt, Central Sector of Mantiqueira Province, Brazil (Hasui & Oliveira, 1984). The Alto Rio Grande Fold Belt is a Middle Proterozoic tectonic province, thrusted upon fragments of a microcontinent formed in successive orogenies during the Archean and Early Proterozoic. The geological and geochronological history of these terranes show subsequent metamorphic episodes of crust-forming and reworking of continental material since 3.4 Ga until 600 Ma with the Late Proterozoic Brasiliano Orogeny (Tassinari & Campos Neto, 1988; Campos Neto, 1991; Lazarini, 2000).
The surrounding geology is represented by lithotypes of the Amparo Complex (infracrustal basement rocks) and metasediments of the Itapira Group (Mesoproterozoic supracrustal rocks).
The Amparo Complex is composed of highgrade metamorphic rocks including migmatites and ortogneiss of tonalitic to granodioritic composition. The group is marked by a complex structural pattern resulting from superimposed tectono-metamorphic events that has shaped the area since Archean times (Fiori et al., 1978; Artur et al., 1979; Fiori et al., 1980).
The Itapira Group, which hosts the manganese occurrences in the area, is an allochthonous metavolcano-sedimetary sequence, trusted over the older rocks of the Amparo Complex. The group is composed mainly by quartzites, schists, quartz schists, silimanite-garnet paragneisses and subordinated gondites (Wernick, 1976; Arthur, 1980; Veríssimo, 1991; Angeli et al., 2011). The rocks of the Itapira Group show a complex structural pattern, characterized by high angle dextral sense shear zones that are consistent with a transpressive deformation in a ductile to ductile-brittle regime. The main structures are tight folds, refolded or not and sheared folds with steeply-dipping angles in their axial portions.
Artur (1980) recognized at least three phases of regional metamorphism and deformation in the area, resulting from recurring collision events from the Palaeoproterozoic to the Neoproterozoic. The last event, the Brasiliano/Pan-African, was responsible for the reactivation of older crustal structures, folding and shearing, leading to the formation of a sequence of synforms and antiforms with fold axial plane strike NE-SW, perpendicular to the axial plane strikes of the previous deformational event, which is NW-SE.
The individualization of each unit in the field is a difficult task, due to the complex structural pattern caused by the superimposed metamorphism and deformation that affected both groups. However, it is know that the units are organized as a set of sinform-antiform structures, where the sinformsare represented by lithologies of the Itapira Group and the antiforms are represented by the Amparo Complex (Wernick, 1976).
A few lateritic manganese occurrences are found in the area, of which some of them have been studied in terms of their mineralogical and chemical composition, genesis and ore processing (Zanardo, 2003; Veríssimo, 1991; Angeli, 2011). The occurrences are arranged in a NNE-trend, coincident with the area main structural trend and separated from each other by a few hundreds of meters (Figure 2).
The ore is residual and consists of secondary manganese oxides and hydroxides formed by the weathering of silicate and silicate-carbonate protores. Lithiophorite, cryptomelane and pyrolusite make up the highest grade ores, derived from the dissolution and redeposition of Mn from the protores.
Two different types of protores are found in the deposits: silicate protore (essentially quartz and spessartite in equal proportions) and silicate-carbonate protore (rhodochrosite, rhodonite, pyroxenes and amphyboles in addition to quartz and spessartite). Graphite is found in both ore and protore, in amounts that can reach up to 10% (Veríssimo, 1991; Angeli et al., 2011). The mineralogy of the protore and the presence of graphite suggest a terrigeonous source for the metal and a meta-sedimentary origin for the manganese orebodies (Veríssimo, 1991).
The formation of primary sedimentary and residual manganese ore deposits is determined by the interaction of several processes, which may include its extraction from the source rocks, its fluvial transportation and its precipitation when in favorable pH and Eh conditions. In addition, chemical weathering can lead to the development of high-grade residual (secondary) deposits (Stanton, 1972; Roy, 1992; Maynard, 2003; Misra, 1999; Guilbert & Park, 2007; Polgári & Gutzmer, 2012).
Gondites are metamorphosed sedimentary manganese-bearing arenaceous and argillaceous sediments with spessartine and quartz, besides rhodonite and other manganese silicates (Roy & Mitra, 1964). Eventual lateritic alteration can occur to the manganiferous protores under humid tropical to subtropical conditions, causing the remobilization of Mn and its precipitation as secondary manganese minerals (Sethumadhav et al., 2010).
Lateritic manganese cappings are originated from the physicochemical weathering of primary manganese deposits, including gondites. The weathering and transport of elements inside the lateritic profile modify the primary mineral content of the deposit, both vertically and horizontally (Wolf, 1976; Taylor, 2011). The leaching of manganese from higher surface levels causes them to become impoverished in that element and enriched in iron, leading to the formation of lateritic iron ore cappings on the surface and manganese enriched surfaces in the intermediate parts of the lateritic profile. As the laterization processes continues, manganese lateritic surfaces are formed in the intermediate parts of the alteration profiles, as a result of the vertical and oblique element mobilization (Wolf, 1976; Taylor, 2011). The leaching of the undesirable elements and the concentration of the manganese in the lateritic profiles might lead to the formation of high grade, economic deposits (Stanton, 1972; Park & MacDiarmid, 1975).
Lateritic ore cappings are a very important feature in terms of manganese exploration, as they are easily recognized in the field because of the characteristic black color they bring to the ground. However, their greater areal extent when compared to the primary deposits very often causes the overestimation of the concealed primary deposits.
Methods
The occurrence was studied using Induced Polarization Tomography (IPT), carried out within an area of approximately 500m by 600m. A total of 10 lines were positioned perpendicular to the local main structural trend, NNE, parallel and spaced 50 m to each other. The total length was 420 m for each line (two multi-electrode cables with 21 stations each), with an along-line electrode separation of 10m for all lines. Were performed a total of 10280 measures, of which half the readings of electrical resistivity and half of chargeability.
Data were collected through a WennerSchlumberger array, multi-electrode cables (21 take-outs each) and non-polarizable porouspot electrodes (Cu-CuSO4). The WennerSchlumberger array has good signal-to-noise ratio and imaging resolution (Dahlin & Zhou, 2004), and has been successfully used in the prospecting and 2D and 3D modeling of mineral deposits (Moreira et al., 2012; Moreira et al., 2014). The opening and saturation of cavities for placement of the electrodes with solution of CuSO4, resulted in minimum contact resistance and a great signal-to-noise ratio.
The equipment used was an ABEM Terrameter LS (Sweden), which consists in automatic and programmable single transmission/ reception module with the capacity to acquire Spontaneous Potential (SP), DC resistivity (ER) and Induced Polarization (IP) field data. Data are then automatically registered into the equipment internal memory, in the chosen data file format, without any human interference (ABEM, 2012).
The survey acquisition parameters were: injected current = 500mA; injection time = 2s; acquisition time delay = 0.2s and a single acquisition time window of 0.1s, with the adoption of a ceiling of 3% of maximum standard deviation in relation to the average of measured values. These parameters were fixed after preliminary tests in the field, for verification and analysis of disturbances as power lines, telluric noise, EM coupling, among other.
The inverse modeling was done using the Res2dinv software (Geotomo Software, 2003), where 2D model sections of chargeability were generated, with the addition of data from topography. The Res2dinv is a 2D inversion software, which automatically defines a bidimensional model of the subsurface (in terms of distance versus approximate vertical depth) from resistivity and IP data, obtained from geoelectrical surveys (Griffiths & Barker, 1993).
The 2D model sections were then exported from Res2dinv and re-imported into the Oasis Montaj Platform (Geosoft), in order to create a 3D visualization model for the chargeability, without adjustment of topographic data. 3D visualization models generated from geophysical data are of great help in the understanding of complex geological structures and hydrological problems, like the flow of pollutants and modeling of ore deposits (Chambers et al., 2006; Aizebeokhai et al., 2011; Moreira et al., 2012). The visualization model was created interpolating the IP data from each section, using the Minimal Curvature algorithm.
Discussion
The data presented in terms of pseudosections feature areas of high chargeability in the central portion, with a tendency of continuity subvertical and reduction of values with the increase of depth (Figure 3). This pattern is accentuated in sections calculated with highlight of central areas and outlining a flank of high values toward the end of the lines are joined on-site deleted in inversion models.
The maximum depth obtained in pseudosections was 70m. The maximum depth after the process was 85m, defined in the form of automatic during the processing, before the unavailability of direct data for calibration, as testimonials or contacts geological outcropping.
The IP data revealed a low chargeability pattern for the investigated area as a whole. Apart from its central portion, where high chargeability values were detected in the lines 4 and 6, the average chargeability rarely overcomes 8mV/V. The lines 4 and 6, in the central portion of the area, showed significant IP anomalies in their central-eastern parts, characterized by chargeability values higher than 20mV/V, which were correlated to the manganese ore (Figures 4 A and 4B).
The errors of adjustment in both sections (12.9 and 14.1) reflect the standard deviation or variability of the data in relation to the average of the values. The high chargeability areas, presented in sections 4 and 6, have good vertical continuity into the deeper portions, with no end in sight, based on structural data obtained in the field, although it could be considered some smoothing effect during the data processing. The anomalies, which are coincident with the manganese lateritic cover in the surface, extend at least 50 m downplunge towards the west at angles of about 40o (Figure 4A).
When projected to the surface, the central anomalies define an area much smaller than that of the manganese ore capping mapped and recognized during the geophysical survey (Figure 4C). A possible explanation for the discrepancy between the mapped occurrence in the field and the modeled geophysical orebody can be presented in terms of survey configuration. The 10 m electrode spacing used in the data acquisition resulted in a 2D modeled section with a minimum investigation depth of about 3 m, which is possibly deeper than the lower limit of the lateritic cover, with a maximum thickness of 1m verified in the field. Therefore, the high chargeability values are possibly related to the primary ore, instead of being related to the ore capping itself.
Finally, an isosurface was created based on the IP voxel model with the aim of evaluating the morphology of the high chargeability zones (values above 20mV/V) and their correlation with the deep mineralization (Figure 5).
The chargeability isosurface revealed the existence of two independent orebodies in the subsurface, instead of a single one elongated parallel to the regional strike (NNE) as suggested by the boundaries of the lateritic cover. Also, the isosurface showed a good vertical continuity of the orebodies at least to a depth of 80 m below surface, with a dip of about 40o to the WNW direction.
The orebodies are elongated along a WNW-ESE direction, therefore roughly perpendicular to the area main structural trend and the regional alignment of the lateritic occurrences. A possible explanation for this is that the orebodies were most likely originated previously to the last deformational event that culminated with the generation of the synform and antiform structural pattern with fold axis striking NE-SW.
Conclusion
The IP method showed good efficiency in the reconnaissance and morphological characterization of the orebodies in subsurface, mainly due to the presence of disseminated graphite (up to 5%) in the lateritic/manganese ore and protore, besides the Mn oxides and hydroxides, essentially due to the presence of graphite in the ore, because the polarization of oxides is very low.
These results demonstrate the applicability of the geophysical ground methods, specially the IP, as a support tool in the identification and selection of exploration targets for test drilling. The use of relatively inexpensive tools for the identification and selection of best drilling location can be a good strategy for saving time and money in mineral exploration projects.
Lateritic manganese ore cappings are key elements in the reconnaissance phases in regional mineral exploration, as they are easily identified in the field due to their characteristic black color. However, their great areal extent in the field as the result pedogenetic processes often lead to miscalculations on the resources and reserves of the concealed primary deposits.
The 3D visualization model was very important for delimitating both the areal extent and depth continuity of the concealed primary deposits, hidden below the lateritic surface. In addition, the 3D model was fundamental for the individualization and morphological characterization of the two orebodies in subsurface, dislocated from the central region of the lateritic cover and elongated perpendicular to the main regional structural trend.
The presence of graphite in the ore and protore was considered crucial for the success in the applicability of the IP method, due to its high polarizability in relation to the host rocks, which are quartz-feldspatic in composition. The genesis of the graphite is related to the origin of the manganese ore in a marine environment, under strong reducing conditions that allowed for the preservation of the dissolved organic matter in the sediments. Regional deformational and metamorphic events resulted in the conversion of the organic matter into graphite and in the generation of the gondites, which did supergene enrichment processes then concentrate.
The orientation of the orebodies in subsurface, contrary to the main structural trend, brings new possibilities for the revaluation of the several other manganese occurrences in the area through geophysical investigation. The IP method revealed to be a very efficient tool for the characterization and 3D modeling of the manganese orebodies in subsurface, in aid on the target location for direct sampling and chemical analysis.