The geochemistry of oil in Cornish granites

Oil residues in Variscan granites in Cornwall, SW England, preserve biomarker data which indicate an origin from marine source rocks. The biomarkers also indicate a thermal maturity that excludes an origin from the Devono-Carboniferous rocks intruded by the granites, but is similar to that of Jurassic-sourced oil to the east in the Wessex Basin. A suite of five different samples from the South Crofty tin mine are variably biodegraded, implying alteration after emplacement of oil in the granite. These characteristics are compatible with models for updip flow of fluids from offshore Mesozoic sediments into older granite topographical highs.

dilated by drilling (Harrison & Beer 1977). There is no evidence for post-oil mineralization.
The present study presents organic biomarker data from seven samples of viscous/solid oil residue from mines in Cornwall. The objectives for these data are to: • measure thermal maturity; • determine any constraints on the environment of deposition for the source rock; • determine levels of biodegradation, and assess if this reflects modern and/or ancient alteration; • compare with data from samples in other fractured basement.

Methods
Five of the seven samples are from South Crofty Mine in the Carnmenellis Granite ( Fig. 1; Table 1) but collected at different times, allowing the measurement of a set of samples with variable degrees of alteration. The samples, coded South Crofty A, B, Y, H and NHM, were provided by A. Bromley, A. Bath, B. Young, R. Harrison and the collections of the Natural History Museum (specimen 14433, recorded as Eastwell Crofty, assumed East Wheal Crofty which is now subsumed into South Crofty, archived in 1840-41), respectively. Samples were also measured from North Treskerby Mine, from the collection of the Royal Cornwall Museum, Truro, and from the Wheal Speed Mine, from the collection of the British Geological Survey. The occurrences of oil residues at South Crofty, North Treskerby and Wheal Speed were recorded by Harrison & Beer (1977) and LeBoutillier (1996), by Tweedy (1842), and by Harrison & Taylor (1966), respectively. The sample from North Treskerby is exceptional in containing crystals of sulphide minerals, and the Wheal Speed sample is intermixed with uranium minerals. Gas chromatography of Cornish samples had been attempted previously (Harrison & Beer 1977;Bath et al. 1984) but no evidence of the results has been presented. Samples of 'killas' slate from Treskillard and Looe were extracted to assess if they could yield meaningful data. Data from the Cornish samples were compared with new analysis of a previously studied Jurassicsourced oil seep at Mupe Bay, Dorset, 200 km to the east in the Wessex Basin (Parfitt & Farrimond 1998), and from organic-rich samples at several other stratigraphic levels. The samples were from black Devonian limestone in the Quantock Hills near Bridgwater, Triassic black shale from Blue Anchor and Oligocene lignite at Bovey Tracey (Webby 1965;Edwards 1976;Warrington & Whittaker 1984) (Fig. 1).
Rock samples were prepared by rinsing with distilled water two times, and again with dichloromethane (DCM). The dry rocks were crushed and extracted using a Soxhlet apparatus for 48 h. Solid bitumen and tar samples were ultrasonicated with DCM and methanol (MeOH). All glassware was thoroughly cleaned with a 93:7 mixture of DCM-MeOH. Crushed samples were weighed, recorded and transferred into pre-extracted thimbles. The extracts were then dried down using a rotary evaporator, and separated into aliphatic, aromatic and polar fractions via a silica column chromatography using hexane, hexane-DCM in the ratio 3:1 and DCM-MeOH, respectively. Prior to gas chromatography mass spectrometry (GC-MS) analysis, an internal standard (5β-Cholane, Agilent Technologies) was added to the saturate fraction before injection into the GC-MS machine and subsequent biomarker identification. This was done using an Agilent 6890N gas chromatograph fitted with a J&W DB-5 phase 50 m mass selective detector (MSD) and a quadruple mass spectrometer operating in selected ion monitoring (SIM) mode (dwell time of 0.1 s per ion and ionization energy of 70 eV). Samples were injected manually using a split/splitless injector operating in splitless mode (purge 40 ml min −1 for 2 min). The temperature programme for the GC oven was 80-295°C, holding at 80°C for 2 min, rising to 10°C min −1 for 8 min and then 3°C min −1 , and finally holding the maximum temperature for 10 min. Quantitative biomarker data were obtained for isoprenoids, hopanes, steranes and diasteranes by measuring responses of these compounds on m/z 85, 125, 191, 217 and 218 mass chromatograms and comparing them to the response of the internal standard. Biodegradation was assessed using the ratio of phytane to the C 18 n-alkane (Ph/nC 18 ), which increases with degradation as n-alkanes are altered more rapidly (Peters & Moldowan 1993). Biodegradation also causes the breakdown of steranes at a faster rate than diasteranes, including in the shallow subsurface in aerobic conditions; thus, the diasterane/sterane ratio (D/S) can be elevated due to shallow biodegradation (Seifert & Moldowan 1979) but is also controlled by thermal maturity and source rock lithology (Burton et al. 2018). Thermal maturity was estimated from the 20S/(S + R) ratio for C 29 steranes, based on the increasing proportion of the S isomer with maturation (Peters & Moldowan 1993), and was also determined from the relative proportions of hopane peaks Ts (C 27 18α (H)-22, 30-trisnorneohopane) and Tm (C 27 17α(H)-22,30-trisnorhopane) peaks, and the C 32 22S/(S + R) hopane ratio.

Results
The granite-hosted samples all yielded extracts adequate for GC-MS. Some chromatograms for the samples show prominent 'humps' (unresolved complex mixtures) that indicate strong alteration ( Fig. 2) but, nonetheless, the data allow determination of quantitative parameters (Table 1). The proportion of diasteranes ( Fig. 3), as indicated by the diasterane/sterane ratio, is relatively low for all five samples from South Crofty but higher for the samples from North Treskerby and Wheal Speed (Figs 4 and 5). The South Crofty samples have a wide range of Ph/nC 18 ratios (Fig. 5). The sterane thermal maturity parameter C 29 20S/(S + R) ratios of 0.35-0.5 indicate onset of oil generation to early oil generation. The hopane thermal parameters of Ts/Tm 0.34-0.65 and C 32 22S/(S + R) ratio 0.37-0.65 (excluding one aberrant value of 0.78) similarly indicate early oil generation.
The 'killas' slate samples yielded a very limited extract that was too degraded to allow meaningful interpretation. However, other organic-rich samples from further east in SW England yielded extracts from which environmental and thermal maturity data could be obtained.
The relative proportions of C 27 , C 28 and C 29 steranes are clustered between C 27 and C 29 predominance (Fig. 6). The Jurassic-sourced oil sample from Mupe Bay plots in the middle of this array. The samples of Devonian, Triassic and Oligocene rocks plotted distinctly from the array.

Oil composition
The chromatograms are typical of the residues of oils derived from marine source rocks. The consistently low content of C 28 steranes makes a non-marine source very unlikely. Some samples have unresolved complex mixtures indicative of alteration. The South Crofty samples exhibit variable degrees of biodegradation, as evidenced from the range of Ph/nC 18 values. This is to be expected where the samples were collected from different sites and fluid ingress has continued up to the present time (Edmunds et al. 1984). The least degraded sample is relatively fresh, and suggests that biodegradation occurred once the oil had entered the granite rather than prior to migration. If biodegradation had occurred in an intermediate reservoir, the Ph/nC 18 ratios should be less variable. The least degraded sample is oil that has exuded from fractures which dilated during mining activity (Harrison & Beer 1977), and so experienced negligible exposure to alteration. Some other samples may have been in vuggy cavities, where they had much greater exposure to alteration. The samples from the other two mines also have very low Ph/nC 18 ratios, but have a higher content of diasteranes. This distinction may be related to the evidence of metalliferous mineralization coeval with the oil emplacement, which suggests the possibility of hydrothermal alteration.
The spatial relationship with uranium mineralization at Wheal Speed could have caused the alteration of the organic matter by irradiation from the uranium (Landais et al. 1990;Court et al. 2006). Irradiation causes cross-linking of organic bonds, which results in greater molecular complexity and lower solubility, so that biomarker compositions can become altered. However, such alteration is not evident in the data from Wheal Speed. Where Mesozoic rocks are preserved further east in SW England, in the Wessex Basin, oils exposed at the current surface exhibit variable biodegradation, and thermal maturity C 29 20S/(S + R) ratios in the same range as the Cornish samples (Parfitt & Farrimond 1998). Their sterane composition is also comparable (Fig. 6) and it is possible that the Cornish oil residues are derived from a similar Jurassic source rock. Well 88/2-1 in the Plymouth Bay Basin (Fig. 1) contains organic-rich Jurassic shales which are oil-prone, and further indicates the possibility of a Jurassic source in the region (Evans et al. 1981;Evans 1990). The oil seep at Mupe Bay has an assumed Liassic source, and the rocks penetrated in the Plymouth Bay Basin are Liassic (Evans et al. 1981), but oil has also been generated at other levels within the Jurassic. The data reported here do not distinguish different levels of the Jurassic. Of the other samples from SW England, the two sample of Devono-Carboniferous 'killas' slates yielded very limited extracts which were inadequate for any interpretation, while the Devonian sample remote from granitic metamorphism and the Triassic and Oligocene samples gave measurable extracts. However, their sterane compositions, which are rich in C 27 (Devonian) or C 29 (Triassic, Oligocene), are distinct from those of the granite-hosted samples, and suggest they are not correlative.

Thermal history
The circumstantial evidence of oil occurrences in a region where the country rocks were metamorphosed to dark slates during the Variscan Orogeny led Harrison & Beer (1977) and Parnell (1988) to speculate that the shale precursors of the slates had yielded the oil during intrusion. The intruded Devono-Carboniferous rocks of SW England have subsequently yielded vitrinite reflectance values mostly above 2% (equivalent to the dry gas zone) (Cornford et al. 1987(Cornford et al. , 2011, presenting conceptual difficulties that they could be a source of oil. Measurement of an offshore Devono-Carboniferous well sample similarly gave a reflectance of 4.52% (Evans et al. 1981). Further, the country rocks had already been metamorphosed before granite intrusion (Chen et al. 1993), and intrusion occurred at a depth of 7-9 km (Charoy 1986), making oil generation during this event very unlikely. Selley (1992) suggested that rocks well above the level of granite intrusion might still be in the oil window and could yield oil into mineralizing convection cells but the reflectance data are equally inconsistent with that possibility. The thermal maturity data reported here indicate typical early oil window temperatures, and show no evidence of anomalous heating that would be expected if the organic matter had experienced mild metamorphism. The data from two samples (South Crofty NHM, Wheal Speed) indicate slightly higher maturities than the others, which could reflect some homogeneity in the oil. Higher maturity is also reflected in the slightly increased proportion of C 27 steranes (Sajgó 2000).
Oil generation from a source rock deposited after the intrusion of the granites and associated orogenic deformation is conceivable, given the much lower thermal maturities in the postintrusion section for SW England (Macquaker et al. 1986;Suan et al. 2012). This alternative possibility was difficult to sustain  when data from the offshore Plymouth Bay Basin indicated that in most wells the oil-prospective Jurassic section was missing (Ruffell et al. 1995). However, palaeogeographical and burial history reconstructions suggest that there had formerly been a thick cover of Jurassic sediments now largely removed by erosion (Cope 2004), and oil-prospective Jurassic sediments are preserved in well 88/2-1 in the Plymouth Bay Basin (Evans et al. 1981;Evans 1990). Importantly, there is strong evidence from isotopic and fluid-inclusion data that at least the younger stages of mineralization were derived from basinal brines, expelled from the offshore basins (Gleeson et al. 2000;Psyrillos et al. 2003).
The thermal history of the Carnmenellis Granite has been studied in some detail, and may be relevant to the origin/composition of the oil. Apatite fission-track analysis for the surface of the pluton indicates cooling to below c. 125°C at 155-137 Ma (Chen et al. 1996). The composition of the oil suggests that it is unlikely to have experienced long-term heating at above 125°C, and thus post-dates this late Jurassic age. This is compatible with derivation of the oil from a Mesozoic source rock. The groundwaters currently penetrating the granite are meteoric and up to 50°C (Edmunds et al. 1984). These fluids could readily cause alteration, including biodegradation, up to the present day.

Biodegradation in basement rocks
Other studies of hydrocarbons in basement rocks have similarly shown that they underwent biodegradation within the basement (Parnell et al. 2017a, c). This would have been enhanced by the high susceptibility to fluid flow in such rocks. Basement rocks which contain hydrocarbons do so especially where the basement forms a topographical high updip from adjacent sedimentary basins, and/or the basement is separated from a basin by a basin-bounding fault; both of which can allow fluid ingress from basin to basement. The fluid flow is not limited to hydrocarbons, and hydrocarbon emplacement could be accompanied or followed by aqueous fluids that might alter the hydrocarbons. Where basement rocks form a topographical high, this also makes them more likely to be penetrated by meteoric water. The enhanced susceptibility to updip and meteoric fluid is evident in some granites of SW England where they have been extensively altered to kaolinite (Bristow 1993;Psyrillos et al. 2003). Meteoric water, in particular, has the potential to introduce microbial communities from the surface that could cause the biodegradation of hydrocarbons.

Comparison of oil residues in UK basement
There are several other examples of oil residues within basement rocks within the UK, including Longmyndian, Moinian and  Lewisian metasediments, and the Weardale and Helmsdale granites (Parnell 1988;Trice 2014;Parnell et al. 2017a, b, c). However many factors control the fate of such residues, including the time gap between basement and oil, temperature, origin of porosity and proximity to the surface, so the residues are diverse.
The basement rocks are all topographical highs, which is why they have been amenable to sampling. This has allowed oils to enter them, updip or laterally, from adjacent sedimentary basins. Fracturing associated with basin development facilitated oil ingress into the Helmsdale Granite from the adjacent Moray Firth Basin, into Longmyndian metasediments from the adjacent Carboniferous and into Moinian metasediments from the adjacent Devonian. In the Helmsdale and Moinian cases, the sedimentary rocks were juxtaposed against basement by basin-bounding faults, while the Longmyndian was penetrated partly beneath the sub-Carboniferous unconformity. The Cornish granite was similarly penetrated below an unconformity surface. The other pluton, the Weardale Granite, was distinct in incorporating oil into an active hydrothermal convection system. Only some of the residues exhibit strong biodegradation, found in Moinian and Longmyndian metasediments. In both of these cases the basement had been uplifted so that oil at a shallow level (<2 km depth, at temperatures allowing microbial life) could enter the basement surface. Recent biodegradation of oil in the Cornish granite was similarly at a shallow level. By contrast, oil entered the Helmsdale Granite and Weardale Granite at deeper levels, and so was not susceptible to biodegradation.

Fluid source
Notably, despite a wealth of fluid-inclusion data for mineralization in the Carnmenellis Granite and other plutons in Cornwall (e.g. Dominy et al. 1994;Gleeson et al. 2000;Psyrillos et al. 2003), there are no published records of oil fluid inclusions. This is in contrast to some other occurrences of oil in fractured basement, where oil inclusions were entrapped in vein minerals (e.g. Feely et al. 2017;Parnell et al. 2017c). The absence of oil inclusions in Cornwall implies that the ingress of oil occurred later than the youngest measured vein minerals (generally quartz in north-south crosscourses), although there is evidence of syn-oil mineralization at Fig. 6. Composition of C 27 , C 28 and C 29 steranes (αββ S + R) in samples from Cornwall, determined from m/z 218 chromatograms. Compositions cluster between C 27 and C 29 predominance. Jurassic-sourced oil from Mupe Bay, Dorset, plots within the cluster. However, the sterane compositions of the Devonian, Triassic and Oligocene samples are distinct from those of the granite-hosted samples, and suggest that they are not correlative. North Treskerby and Wheal Speed in intermixtures with ore minerals, and the oil represents a relatively late reactivation of fluid flow through the mineralized structures.
Models for the migration of mineralizing brines from offshore basins in the region emphasize derivation from the Permo-Triassic offshore section, based on geochemical modelling and the age of late vein mineralization Gleeson et al. 2000;Psyrillos et al. 2003). However, fluid migration probably continued through the Cretaceous (Psyrillos et al. 2003), when any Jurassic section that survived erosion could have contributed oil, and there is evidence of post-Jurassic expulsion of basin fluids elsewhere in SW England (Haggerty et al. 1996). This is consistent with the late context of the oil in the paragenesis. The model of Psyrillos et al. (2003) can be adapted to the story of hydrocarbon generation, migration and degradation deduced in this study (Fig. 7).
Much of the mineralization in the granites occurs close to their upper margins; hence, most of the observations of mineral species, and oil residues, are in this setting. This may reflect the focusing of fluids updip from offshore basins, along the sub-Mesozoic unconformity, which was probably not far above the present erosion surface (Cope 2004). Notably, in a counterpart on the south side of the English Channel, mineralization in the Hercynian granites of the French Massif Central similarly is attributed to updip migration of fluids from the Aquitaine Basin along the Mesozoic granite unconformity (Boiron et al. 2002). The granite in the Massif Central likewise contains oil residues (Powers 1932).

Conclusions
The organic biomarker data for oil samples from Cornish granites can be interpreted as follows: • Thermal maturity data indicate early oil window temperatures but not anomalous heating, so the oil has not experienced Variscan metamorphism. • The samples exhibit variable levels of biodegradation, ranging from low to high. The large range suggests that degradation was relatively recent, reflecting emplacement of oil in different sites in the granite. • The oil composition is typical of a marine source rock, and comparable with data from Jurassic oil in the Wessex Basin, but does not match Devonian, Triassic and Oligocene samples so well. • The implication of a post-granite, possibly Jurassic, source makes it likely that the oil migrated updip into the granites from offshore sedimentary basins, which were the depocentres of post-granite sedimentation.