Origin and metamorphism of graphite from Formiga, Minas Gerais (Brazil)

During the Paleoproterozoic Era, the Brazilian cratons experienced orogenic events that modified the archean basement and sedimentary successions. In the southern São Francisco Craton, it can be recognized evidence of an orogenic event that happened between Rhyacian and Orosirian periods. It is related to the closure of an oceanic basin at this time, which led to the collision between the Archean Divinópolis and Campo Belo metamorphic complexes. Graphite schist occurs close to the cities of Formiga and Itapecerica (Minas Gerais), located between these complexes. To contribute to the understanding of the origin and metamorphism of the graphite from Formiga, petrographic studies, X-ray diffraction (XRD) and Raman spectroscopy analyses have been done. XRD and Raman methods revealed that the temperatures record-ed by graphite are around 460°C. However, Raman data showed that the crystallite sizes correspond to higher metamorphic grade conditions (amphibolite to granulite facies). Temperatures of 460°C are probably associated with hydrothermal processes along faults in post-collisional stage. The presence of todorokite, a mineral typical of deep-sea Mn nodules formed by microorganisms, in association with graphite from Formiga, suggests a biogenic origin for the graphite occurrence.


INTRODUCTION
Graphite and diamond are the polymorph occurrence of native carbon on the nature (Harlow 1998). Graphite has a growing economic value due to its modern technological use as graphene source (Simandl et al. 2015). Formed in several geological settings, graphite is most commonly found in metamorphic rocks, especially in orogenic belts. Graphite can be formed through maturation and metamorphism of biogenic carbonaceous material (CM); as precipitation from C-O-H fluids; mantle-derived; and through reduction of carbonates (Simandl et al. 2015).
There is a vast combination of variables like temperature, pressure, kinetics, composition of country rocks and presence or absence of fluids that can influence the formation of this mineral (Wintsch et al. 1981, Luque et al. 1998, Galvez et al. 2013. The graphite formed by biogenic CM undergoes a progressive and irreversible process called graphitization (Buseck and Beyssac 2014). It occurs in temperature and pressure of burial metamorphism, and decreases the hydrogen-to-carbon (H/C) and oxygen-to-carbon (O/C) ratios, transforming disordered and non-crystalline CM in crystalline graphite (Kwiecińska and Petersen 2004) (Fig. 1). The process of graphitization does not depend on the metamorphic pressure, although it is influenced by the oxygen fugacity and metamorphic temperature (Tagiri and Oba 1986). There are two forms to quantify the metamorphic temperature associated to the graphitization of biogenic CM: • by Raman spectroscopy geothermometry, from low (~330°C) to high-grade metamorphism (~600°C) (Beyssac et al. 2002); • by X-ray diffraction (XRD) geothermometry, for metamorphism above 600°C (Wada et al. 1994).
An efficient method to analyse the origin of CM is to evaluate its isotopic composition, inasmuch as Buseck and Beyssac (2014) indicate that this kind of deposit preserves the δ 13 C between -35 and -20‰ biologic signature. A variety of minerals can be formed in this condition, as is the case of todorokite ((Mn, Mg, Ca, Ba, Na, K) 2 Mn 5 O 12 .3H 2 O), which is formed by the accumulation of Mn-oxides (Lowenstam 1981), and can be used as evidence of biogenic activity on seafloor (Burns et al. 1983).
Graphite schist occurs at southern São Francisco Craton, in the Itapecerica Supracrustal Sequence, associated with the collision between the Archean Divinópolis and Campo Belo metamorphic complexes. In Itapecerica (Minas Gerais (MG), Brazil), Miranda et al. (2019) characterized two different forms of graphite. They presented δ 13 C between -21.23 and -27.89‰, indicating biogenic source, and high-grade metamorphism associated with a syn-collisional stage, with average temperature around 729°C (metamorphic graphite), and hydrothermalism associated with post-collisional stage, around 611°C (recrystallized graphite).
In order to complement the study of Miranda et al. (2019), this paper aims to characterize the origin, crystallinity and metamorphism temperature of graphite from an area near the city of Formiga (MG), a city located about 40 km west of Itapecerica (MG) (Fig. 2), using XRD and Raman spectroscopy. The results are compared to previous studies to corroborate to the understanding of the tectonic model proposed for the Rhyacian-Orosirian orogeny in the southern São Francisco Craton.
The basement comprises the Campo Belo, Divinópolis, Bonfim and Belo Horizonte metamorphic complexes (   (Machado Filho et al. 1983, Teixeira et al. 1996. The supracrustal sequence is formed by the Minas Supergroup, which is characterized by clastic-chemical metasedimentary rocks from the Paleoproterozoic (minimum deposition around 2.0 Ga), including the banded iron formations of the Quadrilátero Ferrífero , Moreira et al. 2016, and also the Bambuí Group, formed by pelitic-carbonate sedimentary rocks with Neoproterozoic age (Alkmim and Martins-Neto, 2001).
The Mineiro Belt (Noce et al. 1998, Ávila et al. 2014, Teixeira et al. 2015 was formed by accretionary orogeny and occurred between Rhyacian and Orosirian, which resulted in extensive reworking of regions placed at the margins of the southern SFC (Noce et al. 2007). Chaves et al. (2015) suggested the existence of a Paleoproterozoic event in the Itapecerica region based on chemical ages found in monazites in sillimanite-cordierite-garnet-biotite gneiss (graphite-rich khondalitic rocks), which was confirmed by Carvalho et al. (2017), that found isotopic ages from zircons of 2.05-2.03 Ga (U-Pb) in the interior of the basement.
Occurrences of graphite with δ 13 C between -21.23 and -27.89‰, indicating biogenic origin, (Miranda et al. 2019) are reported in this area, where Itapecerica Supracrustal Sequence has been described , Chaves et al. 2015, Teixeira et al. 2017. The studied area is located near the city of Formiga, where the graphite schist occurs between the Divinópolis and Campo Belo complexes (Fig. 2).

METHODS AND RESULTS
Three samples were collected at the Formiga area (20º26'40.4"S, 40º30'04.3"W). Samples LR44A and LR44C were chipped from an abandoned trench (Fig. 3A) and sample LR44B was chipped about 5 m away. Both collection spots are formed by fine grained graphite schist composed of quartz, graphite and mica, with presence of vertical foliation. In the trench it is possible to notice the presence of quartz veins oblique to the foliation that contain manganese minerals and occur filling fractures. The host rock presents some recrystallization in the contact with the veins. Samples LR44A and LR44B (Fig. 3B) are graphite schist with vertical foliation, while sample LR44C ( Fig. 3C) was taken from an associated quartz vein ( Fig. 3D), with manganese minerals, oriented N40E/75NW. The visual proportion between graphite and manganese mineral in this sample is 4:1. Each sample was between 10 and 20 cm long and weighed approximately 0.5 kg. The studied rocks are fine grained and have granolepidoblastic texture (Fig. 4). They are composed of quartz (50%), graphite (35%) and phengite (15%). Quartz is the dominant mineral in the composition and occurs in anhedral fractured grains. The second main component is graphite, presented in aggregates (LR44A) and in layers parallel with foliation (LR44B). Phengite marks the foliation in both samples, but occurs in larger crystals in sample LR44B than in sample LR44A. The foliation is more evident in sample LR44B than it is in sample LR44A.

X-ray diffraction
The XRD analyses of graphite were performed in the X-ray Laboratory at CPMTC-IGC-UFMG. The mineral was previously isolated from other minerals of the graphite schist by brushing the graphite-rich portion of the rock, followed by graphite flotation in aqueous environment and drying of the floated grains. PANalytical XPert PRO diffraction instrument with theta-theta geometry was used to record the XRD data, using a Cu Kα X-ray source (40 kV and 45 mA). The step size and scan step settings were 0.02°, 2θ and 0.5 s. The high accuracy of the lattice parameter was guaranteed using Rietveld methods (Young 1993) to fit the diffraction data, with starting parameters close to realistic values and equally applied to both samples LR44A and LR44B. Figure 5 shows the XRD results of sample LR44C, composed of todorokite and manganite instead of graphite. The XRD data from the graphite schist samples are organized in Table 1 and their respective diffractogram are presented in Figure 6.
To estimate the crystal size along stacking direction (L c(002) ), we used the Equation 1 provided by Baiju et al. (2005). L c(002) = kλ/β (002) cosθ (1) In this equation, k means the shape constant (0.9), λ is the X-ray wavelength in angstroms (1.5406), β (002) represents the full width at half maximum of the peak in radian and θ correspond to the angle of diffraction in radians. The graphitization degree (GD) has been calculated using the Equation 2 from Tagiri (1981), were d (002) is the interplanar spacing: (2) Moreover, the Equation 3 from Wada et al. (1994) has been used to calculate the metamorphism temperature: The average temperature found for samples LR44A and LR44B were 442°C and 449°C respectively.
The graphs in Figure 7 show correlations that could be done with the results obtained from the XRD data. Figure 7A presents the relationship between interplanar spacing (d (002) ) and crystallite size (L c(002) ), that, in accordance with Tagiri and Oba (1986), can give information about the GD. Both samples LR44A and LR44B were classified as graphite as opposed to fully ordered graphite. The two samples presented similar XRD temperatures and are randomly arranged in the line that correlated GD with metamorphic temperature in the Figure 7B.  Table 1. X-ray diffraction data from LR44A and LR44B, graphite schist samples, in which d(002) represents interplanar spacing, FWHM is the full width at half maximum of the G-band, Lc(002) is the crystal size along stacking direction, GD is the graphitization degree and T is the metamorphic temperature.

Raman spectroscopy
The Raman spectroscopy analyses were performed in the Technological Center of Nanomaterials (CTnano) at Technological Park of Belo Horizonte -BHTec. The equipment used was a confocal microspore Alpha 300R WITec (Wissenschaftliche Instrumente und Technologie GmbH ® , Ulm, Germany) equipped with Nd-YAG laser with double frequency (2.49 mW, λ = 532.2 nm). The same methodology as applied by Rantitsch et al. (2016) has been used, in which the spectra were obtained with 50× lens objective. Five scans in the 1,000-3,200 cm -1 spectra (first order = 1,000 -2,000 cm -1 ; second order = 2,200 -3,200 cm -1 ) were performed with an acquisition time of 30 s. The analyses were made using the powder extracted from both samples LR44A and LR44B.
In each sample, five fields were randomly selected (Fig. 8A, represented by squares), in which three to five analyses were done per square field (Fig. 8B, represented by crosses). The results of Raman spectroscopy are presented in Table 2, and the respective spectrum are presented in Figure 9 (sample 44A) and Figure 10 (sample 44B).
The temperature has been calculated by using the Interactive Fitting of Raman Spectra (IFORS) method, provided by Lünsdorf and Lünsdorf (2016). The average temperatures found for each sample were 486°C for sample LR44A and 463°C for sample LR44B. The graphs from Figure 11 classifies, according to Rantitsch et al. (2016), the metamorphic facies based on interplanar spacing and width at half maximum of the G-band (G HWHM) or area ratio (R2). The area ratio is calculated as described as Beyssac et al. (2002) using the first-order peaks at ~1,350 cm -1 (D1 band), ~1,580 cm -1 (G band), and ~1,610 cm −-1 (D2 band), by the Equation 4: The results, showed in Figure 11, revealed that both samples present graphite. In Figure 11A both samples are plotted in upper amphibolite facies, while in Figure 11B they are classified as granulite. The reason why this divergence occurs is explored in the discussion section.

DISCUSSION
To characterize the graphite schist from Formiga, we correlate X-ray diffraction and Raman spectroscopy from samples LR44A and LR44B. The Table 3 compares metamorphic temperatures acquired from both methods that yielded temperatures around 460°C. According to Lünsdorf (2015), Raman analyses are reliable for temperatures between 330 and 600°C, so Raman data presented here are reliable. This temperature disagrees with the high metamorphic degree expected for the area (Chaves et al. 2015). However, it agrees with the results found by Miranda et al. (2019) for hydrothermal recrystallized graphite.   (002)) (Tagiri and Oba 1986). All samples are classified as graphite, but they aren't fully ordered; (B) Graphitization degree versus X-ray diffraction data temperature (Wada et al. 1994). Samples presented similar X-ray diffraction temperature and are randomly arranged in the trendline. Table 2. Raman spectroscopy data with values obtained by the IFORS method (Lünsdorf and Lünsdorf 2016) from graphite schist samples LR44A and LR44B. G HWHM represents the width at half maximum of the G-band in cm-1, R2 is the area ratio and T is the metamorphic temperature in °C.  The XRD data (Tab. 1) was plotted in graphs to analyze the GD accordantly to its crystallinity. Figure 7A correlates interplanar spacing (d (002) A) and crystallite size (L c(002) A), and, as suggested by Tagiri and Oba (1986), shows that both samples are classified as graphite between the fields of disordered graphite and fully ordered graphite. This corresponds to the temperatures which were found in this work, as graphite becomes fully ordered around 600°C (Beyssac et al. 2002, Lünsdorf 2015. The GD and XRD data temperature (T) are plotted using the equation for graphitization in pelitic rocks (Wada et al. 1994), and data from samples LR44A and LR44B behave similarly, leading to the conclusion that they both have been metamorphized under the same conditions. The data acquired by Raman spectroscopy (Tab. 2) gave quite controversial results when plotted in graphs that classify the metamorphic facies as proposed by Rantitsch et al. (2016). Figure 11A correlates the interplanar spacing d (002) and half width at half maximum of the G-band (G HWHM) and classifies the sample LR44B as related to the amphibolite facies while the sample LR44A to the upper amphibolite facies. Figure 11B shows the correlation between d (002) and area ratio (R2 from Beyssac et al. 2002) in which both samples are classified as related to the granulite facies. This result was expected assuming the metamorphism considered for the area (Chaves et al. 2015), but it disagrees with the temperatures found by both methods. Also, the samples are classified    Table 3. Comparison between X-ray diffraction and Raman spectroscopy data from samples LR44A and LR44B of graphite schist. d(002) represents the interplanar spacing, GD is the graphitization degree, T is the metamorphic temperature in °C, G HWHM represents the width at half maximum of the G-band and R2 is the area ratio. as graphite (Kwiecińska and Petersen 2004) in both graphs presented in Figure 11. An explanation for why the temperatures found by the XRD and Raman methods (Tab. 1), for both graphite schist samples, disagrees with the syn-collisional metamorphism (Chaves et al. 2015), and the classifications based on crystallinity (Rantitsch et al. 2016), revealed in Figure 11, could be associated with a post-collisional stage. In this phase, the reactivation of old faults creates open space for fluid percolation (Carvalho et al. 2017, Miranda et al. 2019. It is possible to observe quartz veins apparently related with this event (Fig. 3D). According to Miranda et al. (2019), hydrothermal process can decrease the crystallite size, placing the samples into high-grade metamorphic fields, even when they were metamorphized in much lower temperatures, that is, around 460°C. In their analyses, samples from hydrothermal recrystallized graphites had results similar to those presented here. They are classified as amphibolite facies in the correlation of Figure 8A and granulite facies in the correlation of Figure 11B, exactly as it happens to the samples analysed here.

Raman spectroscopy data
The graphite schist from the Itapecerica region, which is adjacent to the present area of study, has biogenic origin (Miranda et al. 2019). As in Formiga area, todorokite (Mn-oxide mineral typical of deep-sea Mn nodules formed by microorganisms, as suggested by Lowenstam 1981) is also present. Given the geological and mineralogical similarities between the two areas, it is possible to assume that the graphite schist from Formiga also has biogenic origin. This carbonaceous material would have been deposited in an oceanic basin between the Campo Belo and Divinópolis complexes in the pre-collisional stage. During the Rhyacian-Orosirian orogeny, graphite schist would be metamorphized under granulite facies conditions and faults were formed. In post-collisional stage, the faults would have been reactivated turning into pathways to fluids which changed the graphite structure, decreased its crystallite size and changed the temperature signature to around 460°C.

CONCLUSION
The graphite schist from Formiga presented temperature around 460°C by the XRD and Raman analyses. The hydrothermalism associated with a post-collisional event explains the decrease in the crystallite size of the graphite mineral and its low temperature, while temperature typical of granulite facies was expected. The occurrence of todorokite in quartz veins and the presence of graphite with δ 13 C between -21.23 and -27.89‰ in the adjacent area (Miranda et al. 2019) suggest a biogenic origin for the carbonaceous material that resulted in the graphite schist from Formiga.