Within the Posidonienschiefer, Hg content fluctuates between 53–154 ppb in core A (Fig. 2). In the equivalent interval of cores B and C, the range of Hg content is 28–397 ppb (Fig. 3) and 32–433 ppb (Fig. 4) respectively. The mean Hg content increases from 85 ppb (core A) to 203 ppb (B) and 225 ppb (C) (Table 1, Fig. 5). The Hg/TOC ratio in core A is in the range of 5–20 ppb/%, 10–56 ppb/% in core B, and 12–83 ppb/% in core C. The T-OAE-associated negative CIE is not clearly identifiable in cores B and C.

In core A, Hg/TOC ratios are very stable below 10 ppb/% throughout the Posidonienschiefer (Fig. 2), while in core B and core C, the Hg/TOC values fluctuate more (Fig. 2, 3, and 4). The Hg/TOC ratios in core A diverge from those for cores B and C (Fig. 6). Above 50 m relative depth core A Hg/TOC is persistently lower (~10 ppb/%) compared to ~20 to 50 ppb/% for cores B and C. The Hg/TS ratio in core A is relatively stable throughout the Posidonienschiefer. Similar to the Hg/TOC signal, the Hg/TS signal in cores B and C diverges above 50 m relative depth from the core A Hg/TS signal (Fig. 6). The Hg/TS in cores B and C shows an increasing trend in the lower to the middle part of the section (max ~80-90 ppb/S%). The signal stabilises toward the top of the Posidonienschiefer in core B but decreases to 10 ppb/S% in core C (Fig. 6).

Correlations between Hg, TOC, and TS in the main high TOC sections are presented in Table 2. All individual correlations between Hg, TOC and TS explain less than 50% (R² <0.5) of the observed variance. With the exception of core B, Hg shows weak but significant (p <0.01) correlations with TOC. A similar relationship is found for TS, and Hg correlates significantly to TS in all cores. For core A and core B, Hg has stronger correlations with TS than with TOC. By contrast, Hg shows a very weak correlation with TS in core C and a weak correlation with TOC.

5. Discussion

5.1. Organic-matter characteristic of the Posidonienschiefer

In all three cores, Hg content in the Posidonienschiefer are higher than in the underlying Lias Zeta and overlying Lias Delta (Fig. 2, 3, and 4). The increased Hg content appear coincident with the much higher TOC in the Posidonienschiefer, a marked change from the marls and carbonate-dominated lithologies stratigraphically above and below, differences that reflect very pronounced changes in depositional redox conditions (Dickson et al., 2022). Within the Posidonienschiefer, the highest average TOC content is found in core C, followed by core B and core A (Table 1). This pattern is most likely a function of higher initial OM content in cores B and C, which, in turn, could be the result of their more distal position of the cores relative to the palaeo-shoreline (Dickson et al., 2022) and subdued siliciclastic input relative to supply of organic matter.

With HI values around 650–750 and oxygen index (OI) values under 20, the Posidonienschiefer can be interpreted as Type I/II kerogen showing a primary algal and amorphous organic-matter-dominated maceral assemblage (Littke et al., 1991). Algal-OM has shown high rates of Hg scavenging in modern lake and marine environments (Wallace, 1982; Zaferani et al., 2018; Outridge et al., 2019). Combined with the elevated TOC values (~5–15%), this interval can be identified as an excellent oil/gas-prone source-rock. Such sediments would undergo TOC breakdown when reaching burial temperature (50–150 °C), which, in turn, might affect Hg content and consequently Hg/TOC after the thermal maturation.

The stable TOC, HI and OI values within the Posidonienschiefer itself suggest only minor changes in the OM type, making the Posidonienschiefer ideal for studying the effects of thermal maturation on Hg contents and normalised Hg. In contrast to core A, HI and OI values throughout cores B and C are extremely low throughout, consistent with the highly mature to post-mature stages of catagenesis experienced by the latter two cores (Tissot & Welte, 1984).

5.2. Mercury enrichment, carrier-phase depletion, or mobilisation in mature sediment?

We find that the average Hg content in the Posidonienschiefer is much higher in core B and core C compared to core A (Table 1, Fig. 5, 6). For the Posidonienschiefer, cores B and C have, on average, ~2.3–2.6-fold higher Hg content than core A. Assuming OM and sulfides were the dominant Hg carriers at the time of deposition, the elevated TOC influenced by the relative position of each core within the basin (see discussion §5.1) may have caused higher original Hg content in immature cores B and C. As can be seen in the divergent Hg and TOC- and S-normalised Hg, the most common carrier phases for Hg show a different signal compared to the element itself (Fig 6).

It is not possible to know exactly how much of the original organic-carbon content of the sediments has been lost during maturation. However, in an extreme scenario, up to ca. 60% of the original organic-carbon content of the sediments may have been lost during maturation (Lewan et al., 1979; Tissot & Welte, 1984; Raiswell & Berner, 1987). In turn, loss of organic carbon may have resulted in a significant (relative) enrichment in trace elements, if these were retained in the host rock. Hydrocarbon expulsion from the more mature cores thus potentially would have increased the relative trace-element concentrations through loss of mass. In addition to TOC loss, the removal water from the dehydration of clays during heating also likely contributed to the loss of rock mass (Peters & Cassa, 1994). A significant TOC loss will manifestly cause a more elevated value of normalised Hg/TOC in high-maturity OM-rich rocks (Fig. 5 and 6) if Hg is retained. This potentially results in a bias in both measured Hg and normalised Hg relative to the original signal stored in the rock record.

To assess whether mass loss or hydrocarbon expulsion processes have influenced the composition of our core material we compare our Hg data to previously published trace-element data. If the process of metal concentration due to loss of OM and water played a dominant role, the magnitude of change in Hg should resemble the relative increase in other (immobile) trace elements such as Mo, Zn, and Cd. Trace-metal increases observed between core A and core C did not exceed +45%, except for U (+185%; Table 1) (Dickson et al., 2020, 2022). The increase of +6 , +22, and +45% for Cd, Zn, and Mo concentrations, respectively from core A to C is primarily driven by TOC loss (up to ca. ~21%) during catagenesis, as demonstrated by Dickson et al. (2020) through pyrolysis experiments on the same core material.

The magnitude of change in Hg content compared to previously determined TOC loss (Dickson et al. 2020) shows that less than half the increase in Hg can be explained by mass loss during catagenesis. This implies that, in addition to the constant sum effects discussed above, other processes likely have played role. Such processes could include, for example, internal redistribution of Hg in the sediment or capture of extraneous Hg within the Posidonienschiefer. Hg mobility during maturation might, for example, be attributed to volatility of weakly bound Hg⁰ and Hg²⁺ at relatively low temperatures (60 to 225 °C) (Rumayor et al., 2013), which overlaps with catagenesis temperature experienced by this sequence in the Lower Saxony Basin (50 to 330 °C) (Bruns et al., 2013). In addition to the rise in average Hg content, its value is rather stable in core A, whereas an up-section increase is observed in cores B and C (Fig. 6). While this might be suggestive of Hg migration and (partial) recapture, this is difficult to envisage in lithologies such as the Posidonienschiefer that have extremely low matrix porosity and permeability, especially after burial to depths in excess of 8 km (Hooker et al., 2020).

An alternative explanation might be that the TOC, TS and Hg of core B and C were slightly higher in the upper interval at the time of deposition. Although the basin is generally characterised by euxinic conditions, the presumed greater palaeo-water depths of core B and C and thereby relative thicknesses of the euxinic water column may have resulted in slightly higher TOC, TS and Hg at the time of deposition.

From the correlation between Hg, TOC, and TS in Table 2., Hg correlates significantly with TOC and TS in all cores, without there being correlation between TOC and TS. The relationship indicates the tendency of Hg likely being hosted in both the organic C fraction and sulfur compounds which add complexity to the mechanism of redistribution/re-capture of Hg during thermal maturation.

5.3. Potential mechanisms of perceived Hg enrichment: fractures, and fluid (hydrocarbon) migration or different starting conditions?

The maturation of organic matter and heating of sediments ultimately can lead to fluid overpressures and fracturing of low-porosity/permeability rocks. For example, according to Meissner (1978), the catagenesis of solid kerogen into liquid hydrocarbons, gas, residue and other by-products is accompanied by a volume expansion of up to 25%. It has been suggested that this volume-change reaction would increase local fluid pressures resulting in hydraulic fracturing and formations of veins (e.g., Mandl & Harkness, 1987; Lash & Engelder, 2005). Shale layers in cores B and C contain abundant layer-parallel and oblique veins, mostly filled by calcite cement, with traces of pyrite crystals (Hooker et al., 2020). These oblique and bedding-parallel fractures containing so-called ‘beef’ calcite may have played an important role in the primary migration of petroleum fluids within and from the Posidonienschiefer (Leythaeuser et al., 1988), but, as Hooker et al. (2020) indicate, the potential for vertical migration through the fractures that are filled with calcite is low, as well as the postulated porosity and permeability within the fracture fill. Whether large-scale fractures or networks of fractures existed outside the cored interval and facilitated fluid migration cannot be established with the present material.

Alternatively, core B and C may have had slightly higher Hg during deposition, due to their higher TOC content. However, if we assume that Hg/TOC of the immature core (~11 ppb/%) is representative of the original Hg/TOC of cores B and C and that Hg in these cores did not change, the TOC loss in those cores would need to be ~45% to reach the observed Hg content. A loss of 45% far exceeds the observed TOC loss during pyrolysis experiments (21%, Dickson et al. 2020) and is inconsistent with enrichments other trace element data. This implies that, if starting conditions were different, Hg/TOC for cores B and C should have been substantially higher than for core A. Such higher Hg/TOC and Hg/TS seems unlikely given their distal position in the basin and greater potential for strongly reducing conditions, which typically lead to reduced, not increased, Hg/TOC and Hg/TS (Them et al., 2019; Grasby et al., 2019; Shen et al., 2022; Frieling et al., in review).

5.4. Implications for the use of the Hg proxy in mature sediments

For the compiled data of Grasby et al. (2019) that includes both background and events with active LIPs, average shale values of 62.4 ppb and 71.9 ppb/%TOC are given. The “peak” of Hg contents (397 ppb and 433 ppb) in mature cores B and C (Fig. 3 and 4) are in the range of “Hg spikes” in the geological record previously interpreted to result from LIP activity (e.g., ~20 ppb Hg, associated with the Early Cretaceous Greater Ontong Java LIP (Charbonnier & Föllmi, 2017) to 2517 ppb Hg at the Frasnian–Famennian (Devonian) transition (Racki et al., 2018)). However, the TOC-normalised Hg in the two mature cores (56 ppb/% and 48 ppb/ %) are not in the range of recorded Hg/TOC generally considered as volcanic peaks (e.g., ~175 ppb/ % at the Greater Ontong Java LIP and 7102 ppb/ % at the Frasnian-Famennian transition). The Hg peaks in cores B and C (397-433 ppb) likely resulted from Hg redistribution and rock mass loss during thermal maturation, yet the maximum Hg content is similar to the peak Hg anomaly in Siberian Traps (396 ppb) (Wang et al., 2018) and Deccan Traps (415 ppb) (Sial et al., 2013).

The divergence of the Hg/TOC and Hg/TS above 50 m relative depth (Fig. 6) in cores B and C from a stable signal of Hg/TOC and Hg/TS in core A is difficult to explain without a degree of Hg redistribution and/or slightly higher Hg for the upper Posidonienschiefer in core B and C (see section 5.3). Regardless of the mechanism, the increase in Hg/TOC from TOC alone implies knowledge of thermal maturation history of the material becomes critical for the interpretation of Hg records when analysing sediments that have been exposed to significant temperatures. For example, if the elevated Hg/TOC in core B and C were to be interpreted without the context of core A and associated thermal maturity data, it might be interpreted as increased Hg loading or variability at the time of deposition. It is obvious from our findings that data from successions with significant calcite veins from hydrocarbon migration or other high-temperature fluid migration, similar to those observed in the Posidonienschiefer should be treated with extreme caution. Lastly, we cannot exclude that veins formed during, for example, (non-thermal) fluid escape could have served as similar pathways for Hg redistribution.

Thermally mature sediments have been previously used to assess the potential role of volcanic Hg enrichment in the T-OAE (e.g., East Tributary and core 1-35-62-5W6 (Canada) with Tmax > 450 °C shows peak Hg/TOC of 58 ppb/% and 28 ppb/% respectively (Them et al., 2019), which is in the range of Hg/TOC found in cores B and C). However, if in fact the mercury content increased simultaneously with the loss of TOC during thermal maturation, as suggested by our results here, it could imply that the Hg and Hg/TOC in these more mature and overmature Canadian cores might overestimate Hg loading compared to immature sections elsewhere. When comparing the Canadian T-OAE core material to other T-OAE sections (e.g., Bornholm, Denmark section with Hg/TOC 2590 ppb/% (Percival et al., 2015) and El Peñon, Chile with Hg/TOC 234 ppb/% (Fantasia et al., 2018)), there is no tell-tale sign they are systemically biased. This is likely because the increase in (normalised) Hg introduced by thermal maturation is small relative to the local and (supra)regional variability in Hg deposition (Lepland et al., 2010; Leipe et al., 2013; Percival et al., 2018; Them et al., 2019) and would likely go unnoticed unless the effects of burial-related heating can be isolated, as we have done here. For periods when all records are thermally mature as would be progressively more likely with sediment deposited further back in time, systematic increases in Hg and normalised Hg with thermal maturity could lead to overestimated volcanic (Hg) emissions.

6. Conclusions

We studied three cores spanning the same formation with different levels of thermal maturity from the Lower Toarcian Posidonienschiefer in the Lower Saxony Basin, Germany. We show the increases of Hg content in mature core B and overmature core C were as high as +132% to +157% relative to the immature core A. Because thermal maturation also reduced the TOC content of the host rock at high levels of maturity, Hg/TOC ratios from mature sediments are considerably inflated. Consequently, the use of Hg and Hg/TOC ratios in mature organic-rich shale could lead to overestimates of Hg inventories in the geological past. Further, the magnitude of Hg enrichment is much larger than observed for other trace metals (e.g., Mo, Cd, Zn), which implies that the loss of mass caused by expulsion of hydrocarbons as a result of thermal maturation might not be the only process involved in causing elevated Hg content.

Mature cores B and C showed high Hg content variability profile in contrasts with the relatively stable Hg content profile in core A. We argue that the pronounced difference in the stratigraphic signals might be explained by the mobility and volatility of Hg in the temperature range of thermal maturation in the LSB, the precise mechanisms of mobilisation and re-sequestration remain elusive. Specifically, the observed fractures in cores B and C may have allowed Hg to migrate and re-adsorb at key levels in the stratigraphy, although there is still large uncertainty how this would effectively permeate the bulk of the lithology due to the extremely low matrix porosity and permeability of the Posidonienschiefer.

In general, the magnitude of elevated Hg content and TOC-normalised Hg in the mature cores (B and C) found here is not as high as Hg peak values interpreted as mercury perturbations caused by LIP volcanism. Nonetheless, our results show that the thermal history of sediments must be considered when using Hg as a proxy for volcanism. Specifically, the >2-fold enrichment in Hg content associated with maturation and the potential influence of fluid migration leading to levels of relative Hg enrichment warrant a cautious approach when using (over)mature and especially fractured sedimentary successions as an archive for far-field subaerial volcanic activity.

Hg mobilization processes during thermal maturation, hydrocarbon expulsion and migration history should be considered in evaluating Hg in oil and gas reservoirs. It is essential not only to think about Hg content in the original immature source rock but also how sediment thermal maturity level might enrich or, oppositely, reduce Hg content in the final product