Mineralogical and geochemical characterization of hydrocarbon microseepage-induced sediments in part of Assam-Arakan Fold Belt, Cachar area, NE India

Long term hydrocarbon microseepages create a reducing environment on the surface rocks and sediments, which induces an array of mineralogical alterations. Among these alterations, the reduction of ferric iron minerals to ferrous iron minerals and precipitation of clay and carbonates are signicant. Several studies have been carried out to characterize these hydrocarbon induced rocks/sediments. Almost all these studies have been carried out for arid to semi-arid climatic regions. The present study attempts to characterize the geochemical properties of the hydrocarbon induced sediments in part of Assam-Arakan Fold Belt (AAFB), NE India characterized by heavy rainfall. Reectance spectroscopy, X-ray diffraction (XRD), X-ray uorescence (XRF) and inductively coupled plasma emission - mass spectrometry (ICP-MS) studies have been carried out on the sediments. The reectance spectroscopy reveals that microseepage-induced sediments have higher clay content and lesser ferric iron mineral content. Geochemical indices also suggest that the hydrocarbon-affected sediments are relatively more altered than the unaffected ones. Studies of trace element patterns indicate that the hydrocarbon-induced sediments are enriched in average Be, V, Cu, Zn, Ga, Zr, and Mo and are depleted in Li, Cr, Co, Ni, Rb, Sr, Sc, and Y. The normalized rare earth element (REE) distribution patterns are the same for both the microseepage affected and unaffected sediments though the microseepage-induced sediments are slightly depleted in the REEs. The present study, thus, points out that the hydrocarbon microseepage-induced alterations are also evident in the high precipitation terrains though the alteration levels are less pronounced than that of the arid to semi-arid climatic regions due to abundant surface and groundwater which mobilize the minerals/elements from the microseepage system and tries to homogenize the compositions. of the continuum removed reectance spectra. The spectra of the standard minerals for the identication of the minerals were taken from the USGS spectral library.


Introduction
Hydrocarbons accumulated in the subsurface reservoirs are needed to be capped by some impervious rocks to prevent their escaping. However, no cap rock is perfectly impervious in nature, and trace amounts of hydrocarbons move vertically or near vertically to the Earth's surface through the micro-pores or fractures in the cap rocks in slowly in the form of hydrocarbon microseepages. The microseepages are invisible but pervasive in comparison to the hydrocarbon macroseepages, which are visible but small (point) in the areal extent (Link 1952;Price 1986;Tedesco 1995;Schumacher 1999;Etiope 2015). The occurrence of hydrocarbon microseepages helps us in detecting subsurface hydrocarbon reservoirs directly beneath them, though the depth and commerciality of the reservoirs cannot be ascertained.
Despite the very small amount of hydrocarbons moving as microseepages, their long term migrations have signi cant effects on the surface rocks and sediment covers. It has been observed that the migrating hydrocarbons are oxidized by the bacteria, and the E h -p H condition of the sediment cover is changed to a reducing environment (Schumacher 1996;Saunders et al. 1999; Warren 2012, Asadzadeh et al. 2017). The reducing condition causes a series of geochemical changes in surface rocks/sediment covers. Among the various geochemical changes, the following three are more prominent: a) Bleaching of red beds: Red beds or rocks having a higher amount of ferric iron minerals become bleached due to the transformation of ferric iron (Fe + 3 ) minerals to ferrous iron (Fe + 2 ) minerals (Donovan 1974;Segal et al. 1984, 1986, Schumacher 1996. The developed ferrous iron minerals may precipitate in situ or mobilize to other places by the groundwater. Thus, the original parent rock becomes de cient in ferric iron minerals and may or may not become enriched in ferrous iron minerals. The common ferric iron minerals involved in the hydrocarbon microseepage alterations are hematite and goethite, and the common ferrous minerals developed are magnetite, pyrite, and siderite. b) Precipitation of clay minerals: The dominance of bacterial action in the hydrocarbon microseepage environment enhances the decomposition of feldspar and mica to form clay minerals (Saunders et al. 1993a;Schumacher 1996;Warren 2012). Though the most important clay mineral is the kaolinite, other clay minerals such as illite, montmorillonite may also form. c) Formation of carbonate minerals: The formation of CO 2 during the decomposition of hydrocarbons reacts with the CaO present in the parent rocks to form carbonate minerals (Donovan 1974;Saunders et al. 1999;Schumacher 1996). Calcite and dolomite are the most common carbonate minerals in the hydrocarbon microseepage environment. However, the formation of these minerals depends upon the availability of calcium in the parent rocks.
Lots of studies have been carried out to characterize the diagenetic changes associated with hydrocarbon microseepages. Almost all these studies have been carried out on the in-situ rocks of the petroleum basins occurring in the arid to semi-arid climatic conditions. However, many of the petroliferous sedimentary basins of the world are located in the heavily precipitated areas and are covered by sediments. The geochemical characterization of the hydrocarbon microseepage-induced sediments of such basins has not been studied in great detail. The present study is a deliberate attempt to characterize the mineralogical and geochemical properties of the sediments affected by hydrocarbon microseepages in part of Assam-Arakan Fold Belt (AAFB) basin, NE India characterized by heavy precipitation.

Study Area
The study area chosen is a part of a known petroliferous basin-Assam-Arakan Fold Belt (AAFB) located in the northeastern part of India. The study area lies to the southern part of the basin and extends in and around Cachar district of Assam, India. The study area is bounded by the latitude 24 0 31'N/24 0 57'N and longitude 92 0 37'E/93 0 07'E covering an area of about 2400 km 2 (Fig. 1a). A false colour composite (FCC) image depicting the distribution of bare lands and vegetational areas of the study area formed by the bands 5-4-3 of the Landsat 8 image in the RGB space is also represented in Fig. 1b. The climate of the area is tropical humid with the dominance of rainy seasons and short dry winter season. The average annual rainfall is very high (about 3100 mm/year). The mean temperature during winter ranges between 12-20 0 C and during the rest of the year 16-30 0 C. A major part of the study area is covered by dense tropical evergreen forests. Physiographically, the study region is bounded by hilly terrains to the north, east, and south. The study area is characterized by plain lands formed by the alluviums of river Barak and its tributaries, with occasional small hillocks rising above the plains.

Geological Background
The Assam-Arakan Fold Belt (AAFB), a typical fold and thrust belt is located at the front of the Indo-Burma Ranges. The basin is bounded to the north by E-W trending Dauki Fault, to the east by N-S trending Kaladan Fault and towards the west by the undeformed foreland of Bengal Basin. The basin was thought to be developed as a result of the eastward subduction of the Indian Plate under the Burmese Plate (Angelier and Baruah 2009). Being located in the collisional zone, the basin has been subjected to intense tectonic disturbances and is characterized by a number of structural elements like folds, faults, and thrusts. In the study area, long narrow north-south trending anticlines are seen to be present with their intervening broad synclines (Ganguly 1983(Ganguly , 1993. The anticlines are dissected by several cross faults limiting the structures (Khar et al. 1984). The surface geological map of the study area is presented in Fig. 2.
The basin comprises thick clastic sediment deposited under marine to continental environments and ranges in age from Paleocene to Recent. The generalized stratigraphic succession of the basin is shown in Table 1. The sedimentary sequence appears to thicken towards the south and east (Chakravorty et al. 2011). Several exploratory wells for hydrocarbons have been drilled in the basin, but the complete stratigraphic succession has not been penetrated. Jenam formation of Barail group is the oldest stratigraphic unit encountered in the basin (Chatterjee et al. 2006). Surma and Tipam Group of sediments belonging to the Mio-Pliocene age are extensively exposed to the anticlinal structures in the basin (Ganguly 1993). A large part covering the plain areas of the study area is covered with alluviums.

Petroleum system
The presence of numerous oil and gas seepages throughout the basin authenticates the existence of an

Fieldwork and sampling
Due to heavy rainfall in most of the year, the eldwork was conducted during the short dry season in the study area to collect near-surface samples. Sediment samples were collected from shallow depths by manual digging. Due care was taken to avoid contamination and root zones of plants. A total of 57 sediment samples were collected in the eld. The samples were then packed in air-tight polythene bags to avoid any degradation. The presence of hydrocarbon microseepages in the samples was established by the occurrence of the anomalous amount of lighter hydrocarbons (ethane to pentane) by extracting the adsorbed gases in the samples. Out of the total samples, only selective numbers of representative specimens were used for spectroscopic and geochemical studies.

Re ectance spectroscopy
Re ectance spectroscopy in the ultra-violet -visible-near infrared and short-wave infrared (UV-VNIR-SWIR) in the interval 250-2500 nm was carried out to identify and quantify constituent minerals in the samples. The eld-collected samples were air-dried, pulverized to micron level in an agate mortar and pestle, and then sieved appropriately to separate any organic matter. The samples were analyzed in a spectrophotometer of Agilent Cary 5000 having a Tungsten-halogen and deuterium light source. The diffuse re ectance values were measured at 1 nm interval of wavelength. The minerals were identi ed by their characteristics absorption features. The quantitative abundance of the minerals was attempted with the help of analysis of the continuum removed re ectance spectra. The spectra of the standard minerals for the identi cation of the minerals were taken from the USGS spectral library.

X-ray diffraction (XRD)
As the collected samples were ne-grained sediments, optical microscopy was not very much helpful in the identi cation of the constituent mineral phases. XRD is considered as the most reliable and powerful tool for the identi cation of minerals in ne-grained rocks and sediments. The XRD was carried out to validate and correlate the re ectance spectroscopic results. The samples were rst air-dried, ground to powder level with a mortar and pestle, homogenized and then sieved to lter out any undesirable particles. The samples were mounted in random orientation to identify all constituent minerals and their crystal planes. The specimens were analyzed in an XPERT-PRO X-ray diffractometer system using a Cu Kα X-rays (λ = 0.15406 nm) over the 2θ range from 4.5 to 101 degrees in the steps of 0.0170 (2θ) degree in continuous scanning mode at room temperature. The generated XRD data were processed and analyzed using the X'Pert High Score Plus software. The identi cation of minerals was carried out manually by matching with the published ICDD cards. Considering the complexities involved in the quantitative analysis of the XRD data, only qualitative analysis was carried out.
X-ray uorescence (XRF) study XRF spectrometry is the most extensively used precise analytical procedure in geochemistry for the determination of major and trace element concentrations (Rollinson 1993). The XRF analysis was carried out to quantify the major element oxide concentrations in the samples. The abundance of eleven major oxides, including SiO 2 , Al 2 O 3 , Fe 2 O 3 (Total), MgO, CaO, MnO, TiO 2 , Na 2 O, K 2 O, P 2 O 5 and LOI (loss on ignition) were determined. To prepare the samples for XRF analysis, the samples were rst air-dried, sieved (< 2 mm), and then ground to ne particle size (powder level). The powered samples were mixed with the binding agents (cellulose) and then palletized with a hydraulic press in stainless steel set by applying the optimum amount of pressure. Due care was taken during the sample preparation process to avoid sample to sample cross-contamination. The pallets were physically checked thoroughly for any cracks or voids before inserting them into the XRF spectrometer.
Inductively coupled plasma emission -mass spectrometry (ICP-MS) study ICP-MS is a relatively new tool for the detection of trace element concentration in rocks/sediments. It is becoming increasingly accepted for its precise detection limit and for a wide range of trace elements within a short time (Rollinson 1993). This study was carried out to determine the concentrations of trace elements, including REEs. The concentrations of Li, Be, V, Cr, Co, Ni, Cu, Zn, Ga, Se, Rb, Sr, Pb, Bi, Cd, Cs, Zr, Nb, Mo, Ge, Ta, Sc, Y, and the REEs were determined. For the ICP-MS analysis, nely powdered samples were used.

Re ectance spectroscopy
Three groups of minerals predominantly present in the hydrocarbon microseepage environments were attempted to identify by re ectance spectroscopy. These were: ferric iron oxides, clay minerals, and carbonate minerals. The most common ferric iron oxides in sediments are hematite and goethite. Hematite shows three distinct absorption bands at 520, 650, and 880 nm (Clark 1999 The spectra of the sediments from both the hydrocarbon microseepage affected as well as unaffected areas show characteristics absorptions near 480 nm, 920 nm, 1400 nm, 1900 nm, 2200 nm, 2340 nm, and 2440 nm (Fig. 3). The absorptions near 480 and 920 nm are characteristics of ferric iron minerals. It is observed that the re ectance spectra of ne-grained goethite (ID: MPCMA2-B of USGS Spectral Library) perfectly matches with the spectra of the samples in the ferric iron absorption wavelengths (Fig. 3). The shallow and asymmetric shape of the absorption feature in the interval 800-1000 nm indicates that the goethite is ne-grained non-crystalline in nature and mostly occur as grain coatings (Sheldon and Tabor 2009).
The characteristics absorption features at 1400 nm, 1900 nm, 2200 nm wavelength indicate the presence of clay minerals. The presence of kaolinite is ruled out by the absence of its characteristics spectral doublets near 2160 and 2210 nm (Fang et al. 2018) in the samples. Also, the absorption features characteristics of montmorillonite, vermiculite, and chlorite do not match with any of the spectra of the samples implying their absence in the samples. The characteristics prominent absorption at 1400 nm, 1900 nm, 2205 nm, and the two weaker absorptions near 2340 nm and 2440 nm indicate that the clay minerals are predominantly composed of illite. The re ectance spectrum of illite with sample ID: IL 101 of USGS Spectral Library perfectly matches with the clay absorption features of the samples (Fig. 3). In the spectra of the sediments, absorption features are observed near 2350 nm, indicating the presence of carbonate minerals, but the absorption feature also coincides with the 2340 nm absorption of illite. Also, the absorption at 2350 nm is not very deep, indicating that the carbonates in the sediments are of very low abundance.
The abundance of minerals was attempted by analyzing the continuum removed (CR) spectral curves of the samples. Continuum removal technique is one of the most e cient methods of determining the abundance of minerals and is considered as a feasible substitute for chemical statistical methods in mineralogy studies (Gomez et al. 2008;Viscarra Rossel et al. 2009). The technique generates a hull of boundary points using the local maxima (Gomez et al. 2008). It removes the background noise and highlights the particular absorption feature (Clark and Roush 1984). The strength and depth of the characteristic absorption feature of a mineral on the whole rock re ectance curve show the abundance of the mineral: the more profound the absorption feature, the more mineral is present in the rock sample (Clark 1999). The band depth (BD) at the characteristic absorption wavelength of a mineral was calculated by subtracting the continuum removed re ectance value from 1 (Viscarra Rossel et al. 2009) and was used as a measure of the mineral abundance. The characteristics absorption wavelength for goethite was taken as 941 nm, and for illite at 2205 nm in the continuum removal curves (Fig. 4). As seen in Fig. 4, the continuum removed spectra of ne-grained goethite standard (MPCMA2-B) of the USGS spectral library, also exhibit absorption at the 2205 nm band wavelength. Thus, it is evident that the presence of goethite in uences the band depth at 2205 nm for illite. Assuming a linear mixing model in the whole rock, the actual band depth of illite at 2205 nm was calculated by subtracting the band depth contribution of goethite from the gross band depth at 2205 nm. The CR band depth calculation shows that the average CR band depth of goethite at 941 nm is lower (0.0227) in the hydrocarbon microseepage bearing sediments than that of the hydrocarbon non-anomalous sediments (0.0280). Again, it is also observed that the average CR band depth of illite at 2205 nm is higher (0.03128) for samples from microseepage bearing areas than that of the hydrocarbon unaffected sediments (0.0265). Thus, it is evident that the average content of ferric iron mineral is lower, and the average clay content is higher in the hydrocarbon microseepage bearing sediments in comparison to the hydrocarbon unaffected sediments.

XRD Studies
The X-ray diffractograms of the sediment samples from both the hydrocarbon affected and unaffected areas show prominent peaks for quartz, illite, muscovite, and feldspar. Illite is identi ed from the prominent peaks near 10 0 A, 4.48 0 A, and 3.33 0 A d-spacings (Fig. 5). The peaks at 4.48 0 A and 3.33 0 A of illite are, however, observed to be coincided with peaks of muscovite and quartz. It is, therefore, the XRD studies of the samples strongly support the spectroscopic observations. However, It is to be noted that though the sediments contain goethite as major ferric iron oxide as revealed by the spectroscopic studies, the prominent peak of goethite at 4.18A is absent in the X-Ray diffractograms. This is due to the fact that the goethites occur as a ne-grained coating around the constituent grains, which are opaque to XRD (Swayze et al. 2000). The X-ray diffractograms of the samples also lacked prominent peaks characteristics of carbonate minerals, indicating very low abundances of these minerals supporting the spectroscopic observations.

XRF Studies
Selective specimens were analyzed for XRF studies from hydrocarbon affected and unaffected areas. The average and median concentrations of the major element oxides are shown in Table. that the major constituent of the sediments/sediments is illite. This is also supported by the spectroscopic studies, which indicated illite as major constituent clay mineral in the sediment samples.

The silici cation index (SI) is a measure of silica content and is de ned as: SI = [ SiO 2 / (SiO 2 + Al 2 O 3 )] * 100 (after Pirajno, 2009)
The higher values of SI indicate a higher amount of silica in the samples. It is observed that the average SI value is lower (72.76) for the microseepage affected samples in comparison to the microseepage unaffected samples (74.51). Thus, it is indirectly evident that some amount of silica has been depleted in the hydrocarbon microseepage induced sediments.
Alteration index (AI) is a measure of the degree of the alteration as suggested by Ishikawa et al. (1976) and is measured by The alteration index calculation reveals that the average AI for the hydrocarbon induced sediments is higher (83.14) than that of the hydrocarbon unaffected sediments (81.29). Therefore, although the average bulk compositions of the sediments are similar, the different alteration indices indicate a higher degree of alteration in the hydrocarbon microseepage bearing sediments. The MIAI calculation shows that average MIAI values are higher (54.95) in the microseepage affected sediments with respect to the hydrocarbon microseepage unaffected sediments (52.54). This indicates that the hydrocarbon microseepage affected sediments have undergone more alteration than that of the hydrocarbon unaffected ones. The results, therefore, prove that the MIAI is an e cient index in differentiating hydrocarbon microseepage induced alterations in sediments.

Trace element (excluding REEs) studies
Limited numbers of specimens were analyzed by inductively coupled plasma-mass spectrometry (ICP-MS). The result of the trace elements (excluding REEs) is shown in Table 3 Table 4. The table indicates that the average REE values are higher in the hydrocarbon unaffected sediments. The PAAS normalized REE plots for the samples are shown in Fig. 6.
The REE patterns show that all the sediment samples exhibit positive Eu positive anomalies. These results, in general, also support the observations by Asadzadeh et al. (2020).

Discussion
Re ectance spectroscopic studies show that hydrocarbon microseepage bearing sediments are de cient in ferric iron mineral content in comparison to the hydrocarbon unaffected sediments. On the other hand, hydrocarbon affected sediments are enriched in clay content in relation to the non-hydrocarbon bearing sediments. Both of these observations suggest that microseepage bearing sediments have undergone a higher degree of bleaching of ferric iron minerals and alteration of other minerals (feldspar and mica) to clay minerals. The bleaching of ferric iron minerals and an increase in clay content may be attributed to the effects of hydrocarbon microseepages. The spectroscopic studies also reveal that both the hydrocarbon affected and unaffected samples have low abundance of carbonate minerals which may be due to the facts that the parent rocks being mostly shales and sandstones do not contain any appreciable amount of Ca or the carbonates formed in the microseepage environment are easily dissolved in the abundant surface and groundwater in the acidic conditions and are mobilized away from the hydrocarbon microseepage systems. The spectroscopic results, therefore, indicate that the alteration effects of hydrocarbon microseepages are also evident in the high rainfall areas.
The qualitative XRD studies strongly support the spectroscopic observations on the presence of illite and low abundance of carbonate minerals. The lack of prominent peaks of goethite in the X-ray diffractograms is interpreted in terms of their occurrences as ne-grained coatings which are opaques to XRD. The lack of prominent peaks of carbonate minerals in the XRD points out the low abundance of carbonates in the sediments supporting the spectroscopic observations.
The major element oxide analysis by XRF reveals that except for SiO 2 , the distributions of other oxides are not signi cantly different between hydrocarbon microseepage affected and unaffected samples. The similar bulk compositions of the sediments indicate that both the microseepage affected and unaffected sediments are derived from the same provenance/precursor rocks. The depletion of SiO 2 in the hydrocarbon-bearing sediments may be indicative of silica leaching from the microseepage system. Again, as the Fe 2 O 3 value represents a total of ferric and ferrous iron oxide concentration, no de nite conclusions can be drawn from the Fe 2 O 3 distribution. The slight abundances of K 2 O and CaO in the hydrocarbon-induced samples are in consistent with the current study by Asadzadeh et al. (2020). Among the different geochemical indices, the average alteration index (AI) is higher (83.14) in the hydrocarbon induced sediments in comparison to the hydrocarbon unaffected ones (81.29). This indicates that the hydrocarbon affected areas have undergone a higher degree of alteration. Lower silici cation index (SI) value in the hydrocarbon induced sediments (72.76) with respect to the non-hydrocarbon ones (74.51) show that these sediments have been subjected to silica leaching. The nearly same values of CIA (~ 85) in the hydrocarbon affected and unaffected sediments indicate that the areas have undergone same degree of weathering. The higher average MIAI value in the hydrocarbon induced samples clearly indicates a higher degree of hydrocarbon induced alteration effects. Thus, AI, SI, and MIAI can effectively differentiate between the hydrocarbon affected and unaffected sediments though their differences are small. The small difference in geochemical indices may be attributed to the presence of abundant surface and groundwater in the study area, which mobilizes the alteration products from the hydrocarbon microseepage systems.
Hydrocarbon affected sediments are generally enriched in trace metals like V, Cr, Zn, Cu, Co, Mo, Pb in response to the reducing environment created by microseepages (Duchscherer, 1984;Schumacher 1996).
The trace element distribution pattern of the samples re ects that the hydrocarbon-bearing sediments are enriched in average Be, V, Cu, Zn, Ga, Zr, and Mo and are depleted in Li, Cr, Co, Ni, Rb, Sr, Sc, and Y. However, the median values of Ni, Rb, and Sc are slightly higher, and the median value of Zn is lower in the microseepage affected sediments. Thus, the results, in general, are in conformity with the trace element distribution of the hydrocarbon microseepage system. The distribution of the trace elements revealed in the present study also supports the studies by Petrovic et al. (2012) and Asadzadeh et al. (2020).
The PAAS normalized REE distribution patterns are more or less similar in both the specimens from hydrocarbon affected and unaffected areas. However, the hydrocarbon affected sediments are more depleted in the REEs in comparison to the hydrocarbon unaffected ones, which may be due to the higher degree of alterations, the hydrocarbon affected sediments have undergone. Both the samples from hydrocarbon affected and unaffected areas show Eu positive anomalies. The REEs are good indicators of the provenance because they are insoluble in water and are chie y transported as particulate matter (Rollinson 1993). The similar patterns of the REEs indicate that the sediments are derived from the same provenance. The study, thus, indicates that hydrocarbon microseepage affected and unaffected sediments cannot be differentiated effectively on the basis of REE patterns. The consistent patterns of the REEs also support the current study by Asadzadeh et al. (2020).

Conclusions
The re ectance spectroscopy on the hydrocarbon microseepage-induced sediments in the study area indicates that the two most important alteration characteristics-bleaching of red beds and the increase in clay content, are clearly evident. The geochemical indices like AI, SI, and MIAI also re ect that the hydrocarbon induced sediments have undergone a little higher degree of alterations. The hydrocarbon affected sediments are also enriched in some trace elements characteristics of hydrocarbon microseepage environments. Though, the PAAS normalized REE patterns are similar for both the sediments, hydrocarbon affected sediments are slightly depleted in the average normalized REE values.
The study, therefore, shows that alteration characteristics of hydrocarbon microseepages are evident in the high precipitation areas also. However, the alteration effects in high rainfall areas are not as pronounced as that of the arid to semi-arid climatic regions due to the fact that the abundant surface and groundwater mobilizes minerals and elements from the hydrocarbon-induced microseepage systems and tends to homogenize their distributions in the sediments.

Declarations
Availability of data and materials All data presented in the text of the article are fully available without restriction from authors upon request.

Competing interests
The authors declare that they have no competing interests.

Funding
The authors received no nancial support for the research.
Authors' contributions SG and PD conceived, designed and carried out the research. DM supervised the research work. SG drafted the manuscript. All authors contributed to data interpretation, discussion, and revision of the manuscript.      Re ectance spectra of hydrocarbon microseepage-affected and unaffected sediments. Continuum-removed re ectance spectra of hydrocarbon microseepage-affected and unaffected sediments.

Figure 5
Typical X-ray diffraction patterns of the hydrocarbon microseepage-affected (upper) and unaffected (lower) sediments.