Direct Rubidium-Strontium Dating of Hydrocarbon Charge Using Small Authigenic Illitic Clay Aliquots from the Silurian Bituminous Sandstone in the Tarim Basin, NW China

Illitic clay is ubiquitous in clastic hydrocarbon reservoirs, and the host for several radiometric isotopes such as the potassium-argon (K-Ar) and rubidium-strontium (Rb-Sr) systems. This study applied the isotope-dilution thermal ionization mass spectrometry technique for small samples (3–4 mg) to conduct illite Rb-Sr isotope dating of five illitic clay samples from the Silurian bituminous sandstone (SBS) intersected by five drillholes in the Tarim Basin, NW China. The 87Rb/86Sr ratio of clays is fractionated mainly by the addition of Rb during the illitization of mixed-layer illite/smectite (I/S), which is the dominant clay species in the Tarim Basin samples. The subsample-scale Rb/Sr isotope values suggest that each subsample may contain I/S particles of slightly variable degrees of illitization. Three of the analyzed samples (H6, KQ1 and TZ67) generated Rb-Sr isochron ages of 141 ± 61 Ma, 332 ± 32 Ma and 235 ± 8 Ma (errors quoted at 2σ), respectively. These results are similar to the corresponding K-Ar ages (125 Ma, 389 Ma and 234 Ma). The isotopic ages are consistent with the timing of hydrocarbon charge which varies in different drillholes as constrained by basin modelling, indicating that a closed-system behavior is attained by the hydrocarbon charge that inhibits the illitization of I/S. The Rb-Sr isotope analyses of the other two samples (YM35-1 and Q1) that did not yield isochron ages suggest the conditions for producing isochrons were not satisfied, which may be caused by disturbance of the isotope system by a post-charge hydrothermal event. The outcomes of this study show the robust potential of Rb-Sr clay subsample geochronology for cross-checking isotopic ages yielded by other systems (e.g. K-Ar system) and constraining the timing of hydrocarbon charge.

Lower Silurian sandstone is a significant hydrocarbon reservoir in the Tarim Basin and accommodates abundant solid bitumen, and thus the reservoir is also named the Silurian bituminous sandstone (SBS) 5,36 . Hydrocarbons in the SBS have marine molecular and stable isotopic compositions and show genetic affinity to www.nature.com/scientificreports www.nature.com/scientificreports/ the underlying Cambrian-Ordovician source rocks 36 . The SBS is mainly distributed in the area surrounding the Manjiaer and Awati depressions 36 . Illite K-Ar geochronology shows that the timing of hydrocarbon charge in the SBS varies locally, owing to variation in the timing of hydrocarbon generation 5,13 .

Sample Information and Analytical Method
The samples used in this study were 0.3-0.15 μm fractions of five clay samples utilized in the study of ref. 13 . The five illitic clay samples were separated from sandstone drillcores YM35-1, H6, KQ1, Q1 and TZ67 (Fig. 1B), using the method described in ref. 13 . Grain size fractions (<0.15, 0.15-0.3, 0.3-0.5 and 0.5-1.0 μm) were separated in distilled water using a progressive high-speed ultra-centrifuge 13 . The 0.3-0.15 μm fraction was selected for the Rb-Sr isotopic study based on the following considerations: (1) detrital minerals are generally coarser in size, whereas they may be mixed in the 0.5-0.3 and 1-0.5 μm size fractions 5,9 ; (2) the finest fraction (<0.15 μm) may  www.nature.com/scientificreports www.nature.com/scientificreports/ also contain inherited 87 Sr atoms of detrital origin 37,38 . Thus, the 0.3-0.15 μm size fraction is the most appropriate for this study. Scanning electron microscopic (SEM) and X-ray powder diffraction (XRD) investigations by ref. 13 also confirmed the purity and authigenic origin of this size fraction. Ordered mixed layer illite/smectite (I/S) is the dominant species in all these samples with no detrital K-feldspar or illite identified 13 . Sample information, mineralogical information and K-Ar dating results and are listed in Table 1 and Figs 3 and 4.
The Rb-Sr chemistry and mass spectrometry analyses were completed at the State Key Laboratory of Lithospheric Evolution, Institute of Geology and Geophysics, Chinese Academy of Sciences (IGGCAS), Beijing. Five portions of each sample were randomly picked and each subsample was weighed so that 3-4 mg was obtained using an AG104 Mettler Toledo analytical balance. They were then dissolved with 87 Rb-84 Sr isotopic tracers in 0.1 mL HF (22N) and 0.03 mL HNO 3 (14N) in screw-top PFA Savillex vials. The Rb and Sr fractions were separated and purified via a mini-column containing ~30 μL of Sr Spec resin 21 . Isotope ratios for Rb and Sr were determined using a multi-collector Triton plus TIMS instrument 21,22 . 87 Sr/ 86 Sr ratios were normalized to 88 Sr/ 86 Sr = 8.375209 using the exponential law. Duplicate analyses of Sr standard NBS987 during this study yielded a mean 87 Sr/ 86 Sr value of 0.710244 ± 0.000012 (2σ, n = 4) in good agreement with the reported value of 0.710251 ± 0.000016 21 . Analytical uncertainties for 87 Rb/ 86 Sr ratios were less than 1%. The blank during the analytical session was lower than 3 pg for Rb and 6 pg for Sr. The Rb-Sr isochron ages for samples were calculated using the ISOPLOT3.0 software 39 , applying a decay constant (λ 87 Rb) of 1.396 × 10 -11 yr -1 40 . Acids used during the Rb-Sr chemistry were all analytical reagent (AR) grade and were purified utilizing a SavillexTM DST-1000 sub-boiling distillation system. Ultrapure water with resistivity of 18.2 MΩ cm -1 obtained from a Milli-Q Element system was used throughout this work. Errors of 87 Rb/ 86 Sr ratios are 1% (2σ).

Results
The results of Rb-Sr isotope dating of the five illite samples are presented in Table 2, Figs 5 and 6. The details for each sample are as follows: YM35-1: Subsample A-3 has 87 Sr/ 86 Sr higher than the other subsamples (Table 2), and is not included in the calculations. The regression of the remaining four subsamples yields an isochron age of 111 ± 36 Ma (2σ, Fig. 6A). There is no obvious relation between 87 Sr/ 86 Sr and 1/Sr (Fig. 7A).
Q1: Regression of all Rb-Sr data yields an age of 484 ± 51 Ma and a Sr i of 0.6972 ± 0.0082 (2σ, Fig. 6D). A linear trend with a steeper slope is defined by the Rb-Sr isotope data of three subsamples (D-1, D-4 and D-5, Fig. 6D). The linear relation between 1/Sr and 87 Sr/ 86 Sr for the three subsamples suggests that this three-point "isochron" is a mixing line (Fig. 7D).
TZ67: Regression of Rb-Sr data yields a precise isochron age of 235 ± 8 Ma (2σ) with the Sr i of 0.71011 ± 0.00048 (Fig. 6E). There is no relation between 87 Sr/ 86 Sr and 1/Sr, indicating that the isochron is not a mixing line (Fig. 7E).
The mixed layer illite/smectite (I/S) is the dominant clay species in all the analyzed samples, totaling more than 50% of the entire clay composition ( Table 1). The good negative correlation of 87 Rb/ 86 Sr with IR of I/S (Fig. 8A) implies that the 87 Rb/ 86 Sr ratio of the samples is controlled by the illitization of I/S. Smectite illitization is a ubiquitous process in the depositional environment 41,42 . I/S is composed of smectite and illite layers, and the www.nature.com/scientificreports www.nature.com/scientificreports/ smectite can be transformed to illite with sufficient K supply 41,42 , as the negative relation between IR values and K contents shows in Fig. 8B. Because Rb has a geochemical behavior similar to K (Fig. 8C), Rb is also introduced to I/S during the illitization process (Fig. 8D). The broad positive correlation between the Sr contents and IR values of I/S indicates that the variation in Sr contents by illitization may be insignificant (Fig. 8E). Therefore, the Rb/Sr fractionation of clay samples is mainly controlled by the addition of Rb during the illitization of I/S. Subsample-scale Rb/Sr fractionation is also observed and the variation in 87 Rb/ 86 Sr ratio is generally below 1.5 (Figs 6A-E). Such a small degree of Rb/Sr fractionation may be a response to the microscale chemical variation in the precipitation environment, and each analyzed subsample may contain I/S particles of slightly variable degrees of illitization.
Illitzation has the potential to homogenize the initial Sr isotopic composition of clays and the 87 Rb/ 86 Sr and 87 Sr/ 86 Sr values for subsamples generally define a linear relation (Figs 6A-E). Regressions of the data of three samples, H6 (Fig. 6B), KQ1 (Fig. 6C) and TZ67 (Fig. 6E), yield isochron ages similar to the corresponding K-Ar ages and reflect the timing of illitization ( Table 1). The consistency between Rb-Sr and K-Ar ages suggest that Sr isotopic homogeneity was attained during the illitization. Furthermore, the Rb-Sr age for TZ67 (235 ± 8 Ma, Fig. 6E), which contains 100% I/S (Table 1), has a better precision than H6 (141 ± 61 Ma, Fig. 6B) and KQ1 (332 ± 32 Ma, Fig. 6C), and this suggests that Sr isotopic homogenization may be easier attained within I/S particles than among different Sr-bearing phases (Table 1). Geological significance of the Rb-Sr isochron ages. Geological factors such as burial, hydrothermal activity, and hydrocarbon charge can influence illitzation and reset the Rb-Sr isotope system in clays 2,5,9 . Burial-induced temperature increments may increase the degree of illitization of clays, as temperatures increase with depth, facilitating illitization of I/S 2,43,44 . Burial history analyses show that the maximum temperatures for the Silurian strata in H6, KQ1 and TZ67 were ~138 °C, ~180 °C and ~150 °C, respectively (Figs 9A-C) 5,24 , and not relevant to the respective IR of I/S (Table 1). Therefore, illitization of I/S in these samples may be more dependent on the availability of reactants.
Potassium, the dominant interlayer cation in illite, is a significant reactant for illitization and is more depleted in hydrocarbons compared to formation water in the reservoir. The hydrocarbon charge may, therefore, inhibit the illitization process and reset the Rb-Sr isotope chronometer 3,5 . The Rb-Sr ages for H6, KQ1 and TZ67 are consistent with their K-Ar ages and the timing of hydrocarbon charge is constrained by basin modelling (Figs 9A-C. ref. 13 ). Therefore, the Rb-Sr ages likely represent the timing of hydrocarbon charge in respective regions. www.nature.com/scientificreports www.nature.com/scientificreports/ The new Rb-Sr age data therefore support the hypothesis of ref. 13 that the timing of hydrocarbon charge in the Silurian reservoir varies locally. The hydrocarbon charge occurred earlier in the east of the Manjiaer depression (late Caledonian-Hercynian) compared to the southwest (Indosinian) and northwest of the depression (Yanshanian) (Fig. 9D). The Manjiaer depression is a major tectonic unit accommodating mature source rocks in the Tarim Basin and hydrocarbons discovered in the SBS around the depression have been demonstrated to be generated by source rocks within the depression 45 . Modelling implies that source rocks in the east side of the Majiaer Depression reached the maturity window earlier than in the west side of the depression 45 , resulting in the earlier timing of hydrocarbon accumulation in KQ1 than H6 and TZ67 (Fig. 9D). The earlier timing of hydrocarbon charge in TZ67 than in H6 may be due to the shorter distance between drillhole TZ67 and the source kitchen, which was located near the Tazhong uplift during the late Hercynian 5 .
implications for Rb-Sr clay hydrocarbon charge geochronology. This study presents new Rb-Sr isotope data for five SBS illitic clay samples from the Tarim Basin and shows that a sufficient spread in 87 Rb/ 86 Sr occurs at the subsample scale to allow construction of an isochron. Regressions of Rb-Sr data for samples H6, KQ1, and TZ67 yield three ages: 141 ± 61 Ma (2σ, Fig. 6B), 332 ± 32 Ma (2σ, Fig. 6C) and 235 ± 8 Ma (2σ, Fig. 6E), respectively. These Rb-Sr isochron ages are consistent with the timing of hydrocarbon charge as determined by  Table 1) and basin modelling results 13 . Therefore, the dating method used in this study has the potential to broadly constrain the timing of hydrocarbon charge.
The Rb-Sr isotope data of samples YM35-1 and Q1 did not yield acceptable isochron ages (Fig. 6A,D). For the mixed-layer illite/smectite, the K and Ar atoms reside in the interlayer space 6 . The closure of the K-Ar isotope system is mainly influenced by heat-induced Ar-exchange 46 and the closure temperature is estimated to be 260 ± 30 °C 47 . The burial history of the studied area shows that the maximum temperature of Silurian strata was 100-180 °C, which is below this temperature 48 . Therefore, the K-Ar isotope system is unlikely to be disturbed by later events. For the Rb-Sr isotope system, besides being hosted in the interlayer sites, a portion of the Rb-Sr atoms are absorbed by the external surface of I/S particles 49,50 . The Rb-Sr atoms hosted by the interlayer sites should be inert to external influence, whereas those absorbed by the external surface are readily removed by hydrothermal fluids 46 . Therefore, the Rb-Sr isotope system in I/S is more sensitive to hydrothermal fluids 46 . There was widespread hydrothermal flow in the Tarim Basin during the early Permian, associated with the 300-275 Ma Tarim Large Igneous Province (LIP) 51,52 . The hydrothermal events extensively influenced the western and Central Tarim Basin (Fig. 1B) [53][54][55] . Sample KQ1 was collected from the eastern Tarim Basin, where the influence of the Tarim LIP is insignificant (Fig. 1B), whereas samples YM35-1, H6, Q1 and TZ67 were collected from the area  www.nature.com/scientificreports www.nature.com/scientificreports/ affected by the Tarim LIP (Fig. 1B). The hydrocarbon charge in samples H6 and TZ67 occurred subsequent to the Tarim LIP (Fig. 10) 13 , thus, the Rb-Sr isotope system in these samples can record the timing of hydrocarbon charge, which is the latest event in the region. In contrast, the timing of hydrocarbon charge in samples YM35-1 and Q1 is older than the Tarim LIP (Fig. 10). Therefore, the Rb-Sr isotope chronometer that originally recorded the timing of hydrocarbon charge was most likely disturbed by hydrothermal activity (e.g. hydrothermal leaching of Rb/Sr atoms at easily-exchangeable sites of clays) associated with the LIP. Hydrothermal alteration may result in extensive subsample-scale redistribution of Rb-Sr atoms (e.g. sample YM35-1), or Sr isotopic heterogeneity, which further results in an apparent age for Q1 (484 Ma, an early Ordovician age) that is older than the formation age of the host (early Silurian).

conclusions
This study involved subsample scale (3-4 mg) Rb-Sr isotopic analysis of illtic clays utilizing samples from the Silurian bituminous sandstone (SBS) in the Tarim Basin, NW China. The results show that the Rb-Sr dating method has potential for dating hydrocarbon systems.
(2) Mixed-layer illite/smectite (I/S) is the dominant clay species (>50%) in all the samples. The illitization of smectite layers in I/S introduces Rb to clays, thus fractionating the 87 Rb/ 86 Sr ratios. Minor Rb/Sr fractionation is observed at the subsample scale and suggests that each subsample may contain I/S particles of slightly variable degree of illitization. Smectite illitization also has the potential to homogenize the initial Sr isotopic composition of clays as evidenced by the Rb-Sr isochrons yielded in this study. Hydrocarbon charge may cease the illitization process and result in closed-system behavior of the Rb-Sr isotope system in clays. Therefore, the Rb-Sr isochron ages for samples H6, KQ1 and TZ67 are interpreted as recording the timing of hydrocarbon charge: they are consistent with basin modelling results.
(3) Rb-Sr isotope analysis for samples YM35-1 and Q1 did not yield acceptable isochrons for constraining the timing of hydrocarbon charge. This may be caused by post-charge hydrothermal activity associated with the early Permian Tarim LIP. Hydrothermal alteration may result in extensive subsample-scale redistribution of Rb-Sr atoms (e.g. sample YM35-1), or generate Sr isotopic heterogeneity, which results in an apparent age that is older than the formation age of the host (e.g. Q1).