REE Characteristics of Lower Cretaceous Limestone Succession in Gümü¸shane, NE Turkey: Implications for Ocean Paleoredox Conditions and Diagenetic Alteration

: Trace and rare earth elements (REEs) are considered to be reliable indicators of chemical processes for the evolution of carbonate systems. One of the best examples of ancient carbonate successions (Berdiga Formation) is widely exposed in NE Turkey. The Lower Cretaceous limestone succession of Berdiga Formation may provide a case study that reveals the e ﬀ ect of ocean paleoredox conditions on diagenetic alteration. Measurement of major, trace and REEs was carried out on the Lower Cretaceous limestones of the Berdiga Formation, to reveal proxies for paleoredox conditions and early diagenetic controls on their geochemistry. Studied micritic limestone limestones have the potential to contribute important information as to regional paleoceanographic conditions of Tethys during an important period in Earth history.

Although the diagenetic history seems complex, the mechanisms and origins of dolomitization have been well described in recent studies [31]. However, there has not been an attempt to investigate the micritic limestone strata that are least diagenetically altered and should, therefore, provide proxies for paleoredox conditions and early diagenetic controls on their geochemistry. Consequently, this study focuses on the limestone strata outcropping in the Mescitli area in the southern part of Eastern Pontide (NE Turkey). Here, we present a petrographic analysis coupled with REE+Y characterisation of the limestone succession. The main purpose of the current study is (1) to determine ancient seawater composition using marine carbonates of NE Turkey, from inner platform to slope paleoenvironments; (2) to define the redox conditions and discuss the paleoceanographic proxies and (3) to discuss the role of potential diagenetic influence on REE signals of ancient carbonates.
Reworked skeletal grainstone/packstone microfacies (MF-2) conformably overly the benthic foraminiferal packstone microfacies (MF-1) and comprise approximately 80 m thick, light grey to grey coloured, thick to medium bedded (25-75 cm) limestones ( Figure 3b). The predominant carbonate components are diverse, fragmented skeletal assemblage (echinoids, pelecypoda, brachiopoda (punctate shells), thin-shelled ostracoda, bryozoa) and peloids. Most of the biotic components are broken and reworked, except for small benthic foraminifera [64]. Rare plagioclase and basaltic extraclasts are observed in the base of the microfacies. The absence of a diverse and abundant benthic foraminiferal community and the fragmented assemblage of heterotrophic skeletal fragments implies that the deposition of MF-2 took place in a deeper water environment than MF-1 [31,38].  Sponge spicule wackestone/mudstone microfacies (MF-3) constitute most of the upper part of the studied succession (Berdiga Formation). The contact between microfacies MF-2 and MF-3 is a gradual facies change. MF-3 microfacies comprises approximately 180 m thick, grey to dark grey mudstone and wackestone that is mostly characterized by sponge spicules and glauconite. The other carbonate components are allochthonous benthic foraminifera, intraclasts and rarely peloids. Besides, rare small benthic foraminifera assemblage including Lenticulinidae, Nodosariidae, Gaudryina sp., Verneuilina sp., Bolivinopsis sp., Spirillina sp. are also observed in the lower part of the lithofacies. The abundance and size of allochthonous skeletal fragments gradually decrease throughout MF-3 whilst the proportion of micrite increases upward. Planktonic foraminifera, including Microhedbergella, are Sponge spicule wackestone/mudstone microfacies (MF-3) constitute most of the upper part of the studied succession (Berdiga Formation). The contact between microfacies MF-2 and MF-3 is a gradual facies change. MF-3 microfacies comprises approximately 180 m thick, grey to dark grey mudstone and wackestone that is mostly characterized by sponge spicules and glauconite. The other carbonate components are allochthonous benthic foraminifera, intraclasts and rarely peloids. Besides, rare small benthic foraminifera assemblage including Lenticulinidae, Nodosariidae, Gaudryina sp., Verneuilina sp., Bolivinopsis sp., Spirillina sp. are also observed in the lower part of the lithofacies. The abundance and Minerals 2020, 10, 683 6 of 25 size of allochthonous skeletal fragments gradually decrease throughout MF-3 whilst the proportion of micrite increases upward. Planktonic foraminifera, including Microhedbergella, are also observed in the most upper part of the MF-3. The mud-rich texture and decrease in abundance/size of allochthonous skeletal fragments indicate that the depositional environment was deeper than the MF-2. Overall, the assemblage and the presence of the planktonic foraminifera may suggest that this was the deepest part of the succession, corresponding to a deep shelf or slope. This is consistent with coeval carbonate succession, which is reported in the Başoba Yayla area (Trabzon, NE Turkey) [34,37]. It has also been noted that the age of the most upper part of the formation is the Albian owing to the presence of Microhedbergella rischi Moullade, 1974 [65]. Therefore, it can be inferred that the age of MF-2 and MF-3 is likely Albian [31,38].
Minerals 2020, 10, x FOR PEER REVIEW 6 of 25 also observed in the most upper part of the MF-3. The mud-rich texture and decrease in abundance/size of allochthonous skeletal fragments indicate that the depositional environment was deeper than the MF-2. Overall, the assemblage and the presence of the planktonic foraminifera may suggest that this was the deepest part of the succession, corresponding to a deep shelf or slope. This is consistent with coeval carbonate succession, which is reported in the Başoba Yayla area (Trabzon, NE Turkey) [34,37]. It has also been noted that the age of the most upper part of the formation is the Albian owing to the presence of Microhedbergella rischi Moullade, 1974 [65]. Therefore, it can be inferred that the age of MF-2 and MF-3 is likely Albian [31,38].

Material and Methods
Upper Jurassic-Lower Cretaceous carbonates are well exposed along the Mescitli section (Figures 1 and 3a-c). The study area is situated in Mescitli-İkisu (40.528656, 39.380240), nearly 10 km north-west of Gümüşhane (NE Turkey). The 230 m of Upper Jurassic to Lower Cretaceous neritic micritic limestone of the Berdiga Formation was examined and sampled, and 100 rock samples were collected with a 20 to 50 cm spacing. A total of 80 samples of the Mescitli section were petrographically analyzed. Microfacies types were described based on fundamental principles of limestone classification and microfacies concept [66,67].
Representative samples from the different microfacies (Figures 2, 3a-c and 4a-i) were selected for geochemical analysis (Table 1). Selected samples are all micritic limestones (MF-1 to MF-3) not exhibiting petrographic evidence of significant post-depositional alteration such as recrystallization. Representative thin sections and a mirror-image slab of each thin section were polished to evaluate (i) weathering (ii) presence of clay minerals or basalt extraclasts (iii) micro-fissures filled with calcites or clastic components and (iv) fracturing, all of which were avoided during geochemical sampling. The micritic orthochemical carbonate particles corresponding to the matrix micrite of each polished slab were sampled by micro-drilling using a hand drill.

Material and Methods
Upper Jurassic-Lower Cretaceous carbonates are well exposed along the Mescitli section (Figures 1  and 3a-c). The study area is situated in Mescitli-İkisu (40.528656, 39.380240), nearly 10 km north-west of Gümüşhane (NE Turkey). The 230 m of Upper Jurassic to Lower Cretaceous neritic micritic limestone of the Berdiga Formation was examined and sampled, and 100 rock samples were collected with a 20 to 50 cm spacing. A total of 80 samples of the Mescitli section were petrographically analyzed. Microfacies types were described based on fundamental principles of limestone classification and microfacies concept [66,67].
Representative samples from the different microfacies (Figures 2, 3a-c and 4a-i) were selected for geochemical analysis (Table 1). Selected samples are all micritic limestones (MF-1 to MF-3) not exhibiting petrographic evidence of significant post-depositional alteration such as recrystallization. Representative thin sections and a mirror-image slab of each thin section were polished to evaluate (i) weathering (ii) presence of clay minerals or basalt extraclasts (iii) micro-fissures filled with calcites or clastic components and (iv) fracturing, all of which were avoided during geochemical sampling. The micritic orthochemical carbonate particles corresponding to the matrix micrite of each polished slab were sampled by micro-drilling using a hand drill.
Major, minor and trace elements, including REE of selected matrix micrite of each sample, were carried out by ACME Analytical Laboratories, Ltd. (Vancouver, BC, Canada). The major elements, trace elements, and REEs in the carbonates were determined by inductively coupled plasma-mass spectrometry (ICP-MS). Analyses used~0.2 g of powdered sample digested in 10 mL 8 N HNO 3 , of which 1 mL was diluted with 8.8 mL deionised water, and 0.1 mL HNO 3 . To monitor the precision and accuracy, 1 mL of an internal standard (including Bi, Sc, and In) was added to the solution. For more details of these methods, please see the website of http://acmelab.com. Detection limit (dtl) of CaCO 3 , MgCO 3 , SiO 2 , Al 2 O 3 , Na 2 O, K 2 O, TiO 2 , P 2 O 5 , MnO is 0.01 wt % with exception of Fe 2 O 3 (dtl = 0.04 wt %) and Cr 2 O 3 (dtl = 0.002 wt %). Detection limits for Ba, Ni and Sc, are 1, 20 and 1 µg/g, respectively. Detection limits for Hf, Zr, Y, La, Ce are 0.1 µg/g. Detection limits for Th and Nd are 0.2 and 0.3 µg/g, respectively. Detection limits for Tm, Tb, Lu are 0.01 µg/g; for Pr, Eu and Hu are 0.02 µg/g; for Er is 0.03 µg/g; for Sm, Gd, Dy and Yb are 0.05 µg/g.
The measured REEs data of all micritic limestone samples were normalized to those of post-Archean Australian Shale (PAAS) which are previously reported (e.g., [68]). The equations of (i) Eu/Eu* ratio = Eu N /(Sm N + Gd N )0.5, (ii) Pr anomaly = Pr/Pr* = Pr N /(0.5Ce N + 0.5Nd N ), (iii) La anomaly = La/La* = La N /(3Pr N − 2Nd N ), and (iv) Ce anomaly = Ce/Ce* = 3Ce N /(2La N + Nd N ) are used to express Eu, Pr, La and Ce anomalies in the studied limestones [69,70]. Major, minor and trace elements, including REE of selected matrix micrite of each sample, were carried out by ACME Analytical Laboratories, Ltd. (Vancouver, BC, Canada). The major elements, trace elements, and REEs in the carbonates were determined by inductively coupled plasma-mass spectrometry (ICP-MS). Analyses used ~0.2 g of powdered sample digested in 10 mL 8 N HNO3, of which 1 mL was diluted with 8.8 mL deionised water, and 0.1 mL HNO3. To monitor the precision and accuracy, 1 mL of an internal standard (including Bi, Sc, and In) was added to the solution. For more details of these methods, please see the website of http://acmelab.com. Detection limit (dtl) of CaCO3, MgCO3, SiO2, Al2O3, Na2O, K2O, TiO2, P2O5, MnO is 0.01 wt % with exception of Fe2O3 (dtl = 0.04 wt %) and Cr2O3 (dtl = 0.002 wt %). Detection limits for Ba, Ni and Sc, are 1, 20 and 1 µg/g, respectively. Detection limits for Hf, Zr, Y, La, Ce are 0.1 µg/g. Detection limits for Th and Nd are 0.2 and 0.3 µg/g, respectively. Detection limits for Tm, Tb, Lu are 0.01 µg/g; for Pr, Eu and Hu are 0.02 µg/g; for Er is 0.03 µg/g; for Sm, Gd, Dy and Yb are 0.05 µg/g.
Reworked skeletal grainstone/packstone microfacies (MF-2) comprise a high abundance skeletal component (30-50%), peloids (5-10%) and intraclasts (5-10%) (Figure 4d-f). The bioclastic components are represented by upward decreasing allochthonous skeletal fragments including, echinoids, molluscs, very rare small benthic foraminifera and undifferentiated shell fragments. The most frequent and common bioclasts through the MF-2 are echinoderms and bivalves. Other bioclasts, in order of abundance, are thin-shelled ostracods, sponge and sponge-spicules (Figure 4d-f). Most of the biotic components vary in size from 200 to 600 µm long and are broken and reworked, except for small benthic foraminifera. Benthic foraminiferas are rare and infrequent. Relatively high biotic diversity of bioclasts is observed compared to the MF-1 (Figure 4-f). Sponge spicule wackestone/mudstone microfacies (MF-3) is differentiated from MF-2 mainly by its mud-dominated texture, the predominance of the sponge spicules and lower abundance of reworked bioclasts (Figure 4g-i). MF-3 mostly consists of wackestone to mudstone beds, as well as sponge spicules, rare reworked skeletal components, pseudo-peloids, and rare intraclasts occur. The intraclasts are composed of rounded micritic grains. Reworked skeletal components become thinner and sparser towards the top of MF-3 (Figure 4g), and the most upper part of the microfacies consists of rare reworked skeletal components (Figure 4h), but with a high abundance of planktonic forams, including Calcisphaerulidae (Figure 4i). MF-3 also contains locally common authigenic components (glauconitic grains). Deep-water bioclasts such as calcispheres and sponge spicule are observed at the top of MF-3.

Major and Trace Elements
The chemical analyses of the micritic limestone samples are represented in Table 1  The samples mostly plot in area IIIb of the Pr/Pr* vs. Ce/Ce* ratios [68], which confirm that the Ce anomalies are not an artefact of La-interference (Figure 5a). Eu/Eu* ratios exhibit no correlation with Ba/Sm (Figure 5b). Ce/Ce* ratios show fluctuation (Table 1) with a steady increase throughout the MF-1 part of the studied section. The uppermost part of MF-1 exhibits slightly lower Ce/Ce* anomalies and is followed by a sharp decrease in the Ce/Ce* anomalies at the base of MF-2. MF-2 and MF-3 are represented by relatively low Ce/Ce* anomalies. There is no significant difference in Ce anomaly between MF-2 and MF3.   [68], which confirm that the Ce anomalies are not an artefact of La-interference (Figure 5a). Eu/Eu* ratios exhibit no correlation with Ba/Sm (Figure 5b). Ce/Ce* ratios show fluctuation (Table 1) with a steady increase throughout the MF-1 part of the studied section. The uppermost part of MF-1 exhibits slightly lower Ce/Ce* anomalies and is followed by a sharp decrease in the Ce/Ce* anomalies at the base of MF-2. MF-2 and MF-3 are represented by relatively low Ce/Ce* anomalies. There is no significant difference in Ce anomaly between MF-2 and MF3.

Siliciclastic Impurities
To reveal the possible potential influence of terrigenous input on paleoredox indicators within the studied limestone, it is necessary to identify possible contamination by siliciclastics. This is important because contamination by terrestrial clays can control the primary REEs + Y contents of ancient marine carbonates [4,6,71,72]. A small quantity of terrestrial particulate matter (i.e., shale), which has high REE contents with distinctly non-seawater-like patterns [73], may dramatically modify original REE patterns and cause a decrease in both the extent of LREE depletion and Y/Ho ratios [3,4,74]. Several lines of evidence suggest that the studied limestone displays a diagnostic seawater signature: (1) negative Ce anomalies (0.38-0.81; average 0.57), (2) depleted LREE and (3) superchondritic Y/Ho (average 46.26) (Figure 6a) (e.g., [4]). Although the limestone exhibits a slight enrichment of LREE compared to modern seawater of South Pacific deep water [11] their PAAS normalized patterns exhibit a seawater-like signature compared to the North American Shale Composite (Figure 6a,b) [75]. Furthermore, their general PAAS normalized patterns (i.e., Ce* with an average of 0.57; Nd/YbN with an average of 0.72; Pr/YbN with an average of 0.73 values) are compatible with well-preserved brachiopods [8,75] (Figure 6b,c).

Siliciclastic Impurities
To reveal the possible potential influence of terrigenous input on paleoredox indicators within the studied limestone, it is necessary to identify possible contamination by siliciclastics. This is important because contamination by terrestrial clays can control the primary REEs + Y contents of ancient marine carbonates [4,6,71,72]. A small quantity of terrestrial particulate matter (i.e., shale), which has high REE contents with distinctly non-seawater-like patterns [73], may dramatically modify original REE patterns and cause a decrease in both the extent of LREE depletion and Y/Ho ratios [3,4,74]. Several lines of evidence suggest that the studied limestone displays a diagnostic seawater signature: (1) negative Ce anomalies (0.38-0.81; average 0.57), (2) depleted LREE and (3) superchondritic Y/Ho (average 46.26) (Figure 6a) (e.g., [4]). Although the limestone exhibits a slight enrichment of LREE compared to modern seawater of South Pacific deep water [11] their PAAS normalized patterns exhibit a seawater-like signature compared to the North American Shale Composite (Figure 6a,b) [75]. Furthermore, their general PAAS normalized patterns (i.e., Ce* with an average of 0.57; Nd/Yb N with an average of 0.72; Pr/Yb N with an average of 0.73 values) are compatible with well-preserved brachiopods [8,75] (Figure 6b,c). The Y/Ho ratios can be used as a tracer to assess whether carbonate samples reflect primary marine signature or siliciclastic components [2,4,[77][78][79]. Y and Ho are isovalent trace elements which have a similar charge and radius, and they are considered to display extremely coherent behaviour as twin pairs in a geochemical system characterized by charge-and-radius-controlled (CHARAC) [10]. According to the previous work [77,78,80], an aqueous marine system is characterised by non-CHARAC trace element behaviour, and electron structure must be considered as an important additional parameter which may cause trace element fractionation processes such as chemical complexation. The complexation behaviour of a trace element is additionally influenced by its electron configuration and by the character of chemical bonding between a central ion and a ligand [79,80]. For this reason, Ho is scavenged more readily from seawater than Y due to differences in geochemical behaviour in surface complexation [79][80][81][82][83][84]. According to the literature, high Y/Ho ratios (average of 61 ± 12) occur in the upper 200 m of seawater [7,11,12,64,[85][86][87][88][89][90][91]. However, other studies have proposed that typical Y/Ho ratios in marine carbonates that are free of contamination from terrigenous material show a wider range of values  or a chondritic value of 28 [78,80]. The studied limestones display high Y/Ho ratios with an average of 46.26 (29.33-70.00), which is consistent with detritus-free marine sediments (average 45.50) [69]. Further, the lack of correlation The Y/Ho ratios can be used as a tracer to assess whether carbonate samples reflect primary marine signature or siliciclastic components [2,4,[77][78][79]. Y and Ho are isovalent trace elements which have a similar charge and radius, and they are considered to display extremely coherent behaviour as twin pairs in a geochemical system characterized by charge-and-radius-controlled (CHARAC) [10]. According to the previous work [77,78,80], an aqueous marine system is characterised by non-CHARAC trace element behaviour, and electron structure must be considered as an important additional parameter which may cause trace element fractionation processes such as chemical complexation. The complexation behaviour of a trace element is additionally influenced by its electron configuration and by the character of chemical bonding between a central ion and a ligand [79,80]. For this reason, Ho is scavenged more readily from seawater than Y due to differences in geochemical behaviour in surface complexation [79][80][81][82][83][84]. According to the literature, high Y/Ho ratios (average of 61 ± 12) occur in the upper 200 m of seawater [7,11,12,64,[85][86][87][88][89][90][91]. However, other studies have proposed that typical Y/Ho ratios in marine carbonates that are free of contamination from terrigenous material show a wider range of values  or a chondritic value of 28 [78,80]. The studied limestones display high Y/Ho ratios with an average of 46.26 (29.33-70.00), which is consistent with detritus-free marine sediments (average 45.50) [69]. Further, the lack of correlation between Y/Ho ratios, Ce/Ce*, and Al and Zr contents in the studied limestone implies little or negligible terrestrial contamination [74].
Minerals 2020, 10, x FOR PEER REVIEW 12 of 25 between Y/Ho ratios, Ce/Ce*, and Al and Zr contents in the studied limestone implies little or negligible terrestrial contamination [74]. Finally, several trace elements including Hf, Sc, Th, Zr and Al2O3 are considered to be particularly sensitive to shale contamination, as they occur in detrital alumosilicate minerals in much greater volume in average shale than in marine carbonates [4][5][6][17][18][19]64,71,85,92,93]. The studied micritic limestones predominantly exhibit low Hf (<0.02 to 0.5 µg/g), Sc (<1 to 2 µg/g) Th (<0.2 to 0.9 µg/g) contents, which supports the observation that contamination by detrital minerals is minor to negligible (Table 1). Relatively low Zr (1.90-20.40 µg/g, an average of 5.27 µg/g) and Al2O3 (0.14-3.57 wt %, an average of 0.83 wt %), with a good correlation between Zr and Al2O3, suggest a minor amount of shale contamination (Figure 7a). However, total REE concentration is poorly correlated with Zr and Al2O3, implying that this small volume does not dramatically contaminate the REE concentration in these samples [2] (Figure 7b,c). Plotting Y/Ho, Eu/Eu* and Ce/Ce* versus and Zr contents in Figure 7d-f shows no positive correlation, as would be expected if there was low detrital contamination [3][4][5].
To sum up, the contamination of carbonate samples by terrestrial materials can be defined as minor or negligible based on these multiple and independent analyses including (1) their seawater signature; (2) similar Y/Ho ratios to those of the marine carbonates free of contamination and (3) lower Sc, Th, Zr and Al2O3 contents as discussed above.  To sum up, the contamination of carbonate samples by terrestrial materials can be defined as minor or negligible based on these multiple and independent analyses including (1) their seawater signature; (2) similar Y/Ho ratios to those of the marine carbonates free of contamination and (3) lower Sc, Th, Zr and Al 2 O 3 contents as discussed above.
Many studies have focused on using REEs to provide a better understanding of the influence of thermodynamic and kinetic parameters on the partitioning of individual REEs during calcite precipitation [90,91,95,96]. Mineral dissolution is considered unlikely to destroy the geological record encoded by the distribution pattern of REEs in calcites based on experimental studies, though some have questioned this interpretation because of differences between the experimental system and natural settings (e.g., REE concentration in solution) [96]. Further, a large amount of information on REEs in natural carbonates have been compiled [95,96], and several parameters have been investigated to assess the potential diagenetic influence on the REE signature of carbonates [1][2][3][4]. It has been shown that the progressive REE scavenging during post-depositional modification may produce a positive correlation between the Ce anomaly and REE [70,71,91]. However, our samples do not display a notable positive correlation between Ce and REE content (Figure 8a). Furthermore, the studied carbonates show a negative correlation between Eu/Eu* and REE concentration (Figure 8b), implying that such a diagenetic influence is absent [91]. Finally, a negative correlation between Ce/Ce* and La/Sm N ratios have been used as evidence of diagenetic influence on Ce [97][98][99], and this correlation is not apparent in our carbonates (Figure 8c). Similarly, an absence of correlation between Ce/Ce* and Eu/Eu* implies a negligible or absence of influence of post-depositional alteration on the measured Ce/Ce* [3,7,74,84,87,100]. Our samples display a weak correlation (Figure 8d-f), which implies negligible or no diagenetic impact on Ce anomalies. This conclusion is also supported by petrographic observations of the micritic limestone, which do not show significant recrystallization.
Given these conditions, the relationship between Ce/Ce* and Pr/Pr* can be used to characterize redox conditions using shale-normalized REE contents [69]. In assessing whether an apparent Ce anomaly is true, and not a function of La-enrichment, the data was analysed using the Ce/Ce* versus Pr/Pr* analysis [69] (Figure 5a), whereby area (I) no Ce anomaly; (IIa) positive La anomaly causes apparent negative Ce anomaly; (IIb) negative La anomaly causes apparent positive Ce anomaly; (IIIa) real positive Ce anomaly; (IIIb) real negative Ce anomaly; (IV) positive La anomaly disguises positive Ce anomaly [69,99]. The position of samples from this study in area IIIb indicates that they display real Ce anomalies (Figure 5a). In addition, Eu/Eu* can be contaminated by interference with Ba during ICP-MS analysis [101], but Eu/Eu* ratios exhibit no correlation with Ba/Sm, implying that calculated anomalies are real positive Eu anomalies (Figure 5b). It has been shown that the micritic limestones mainly show a seawater signature, with LREE depletion, positive La anomaly (La*) and a mostly negative Ce anomaly (Ce/Ce* = 0.38 to 0.81, average 0.57). However, they also exhibit a wide-range of Eu anomalies (Eu*/Eu = 1.01-1.65; 1.29). This pattern is missing in the general characteristics of normal marine carbonates [102] and is generally considered as an indicator of reducing, high-temperature fluids [77,83,103,104]. Similar patterns are recorded in many distinct marine systems affected by hydrothermal fluids associated with ferromagnesian rocks [71]. Such hydrothermal fluids are usually reported at mid-ocean ridges and back-arc spreading tectonic setting, where water-rock interactions with mafic rocks provide additional REE to ambient seawater [69,77,80]. Therefore, the Eu/Eu* ratios in this study may imply water-rock interaction with basalts. However, previous studies have shown that basaltic rocks are absent throughout the Lower Cretaceous, and there is no evidence of Early Cretaceous magma generation in the study area (Eastern Figure 8. Plots of the ΣREE (µg/g) contents vs. Ce/Ce* ratios (a); ΣREE (µg/g) contents vs. Eu/Eu* ratios (b); Ce/Ce* vs. La/Sm N ratios (c); Ce/Ce* vs. Dy/Sm N ratios (d); ΣREE (µg/g) vs. P 2 O 5 (wt %) contents (e); Eu/Eu* vs. Ce/Ce* ratios (f); Mg/Ca vs Eu/Eu* (g) and Mg/Ca vs. Ce/Ce* ratios (h).
It has been shown that the micritic limestones mainly show a seawater signature, with LREE depletion, positive La anomaly (La*) and a mostly negative Ce anomaly (Ce/Ce* = 0.38 to 0.81, average 0.57). However, they also exhibit a wide-range of Eu anomalies (Eu*/Eu = 1.01-1.65; 1.29). This pattern is missing in the general characteristics of normal marine carbonates [102] and is generally considered as an indicator of reducing, high-temperature fluids [77,83,103,104]. Similar patterns are recorded in many distinct marine systems affected by hydrothermal fluids associated with ferromagnesian rocks [71]. Such hydrothermal fluids are usually reported at mid-ocean ridges and back-arc spreading tectonic setting, where water-rock interactions with mafic rocks provide additional REE to ambient seawater [69,77,80]. Therefore, the Eu/Eu* ratios in this study may imply water-rock interaction with basalts. However, previous studies have shown that basaltic rocks are absent throughout the Lower Cretaceous, and there is no evidence of Early Cretaceous magma generation in the study area (Eastern Pontide, NE Turkey). Although it is widely accepted that the Early Cretaceous was a stable tectonic regime due to a lack of igneous activity [50,61]. The eastern Pontide experienced multistage basaltic magma generation during the Late Cretaceous [43,49] and early Cenozoic acidic magma generation is also recorded [44][45][46]63]. This Upper Cretaceous to Cenozoic magma generation has led to recrystallisation and geochemical modification of the underlying dolomites at a late-stage of their diagenetic history [32][33][34]. However, the studied Lower Cretaceous limestones do not show petrographical evidence of such recrystallisation (Figure 4a,b). Furthermore, the studied limestone samples exhibit low Mg-calcite chemistries (Ca 98.81-99.52 Mg 0.48-1.19 (CO 3 )) [105], indicating that they are chemically stable [106][107][108][109][110][111][112][113]. Further, there is neither petrographical evidence for recrystallization which associated with Late Cretaceous to Cenozoic magma generation nor geochemical pieces of evidence, for example, the lack of apparent correlation between Mg/Ca vs. Ce/Ce* and Eu/Eu* ratios which imply less influence of late diagenetic processes on their REE chemistry (Figure 8g,h). Therefore, Late Cretaceous to Cenozoic magma generation during this later diagenetic stage is probably not responsible for high the Eu/Eu* ratios of the studied limestone.
Nevertheless, Lower Cretaceous calc-alkaline lamprophyre and high-Nb alkaline basaltic dykes have been discovered in the Eastern Pontide [114], and this may have provided hydrothermal fluids to the depositional environment, which changed the temperature and/or redox conditions prior to diagenetic stabilization (e.g., [81]). To test this, Sm/Yb vs. Eu/Sm and La/Yb vs. La/Sm plots were constructed (Figure 9a,b). The data show an overlap with hydrogenetic-Fe-Mn crusts and relatively higher La/Yb N values than seawater are observed, implying the presence of some early absorption processes [3,115] (Figure 9b). It has been proposed that hydrothermal fluid possess higher Sm/Yb and chondritic Y/Ho values (∼28) than modern seawater [71]. Plotting Y/Ho versus Sm/Yb (Figure 9c), studied samples overlap with the samples of Pongola Iron Fm. which were affected by hydrothermal fluid during their sedimentation processes [71]. They also plot on the line of hydrothermal fluid (1-5%) suggesting a contribution of hydrothermal fluid Figure 9c. Although the studied samples have relatively lower Eu/Sm ratio compared to the high-T hydrothermal fluids. They possess similar Eu/Sm ratio to the hydrogenic Fe-Mn crusts (Figure 9d) and samples plot in an area between seawater and hydrogenetic-Fe-Mn crusts which may confirm the influence of the hydrothermal fluid. To sum up, even though the studied limestone formed in an open marine environment, (i) the positive Eu anomalies, (ii) high Sm/Yb and La/Yb N ratios and (iii) the Y/Ho range between seawater and hydrogenetic Fe-Mn crusts suggest some modification by hydrothermal fluids which are most likely to be associated with Early Cretaceous magma generation (Hauterivian-Albian) [77,83,103,104,116,117].
The studied limestone is interpreted to have experienced progressive modification through interaction with reducing, high-temperature diagenetic fluids, and this caused the slight enrichment of Eu. However, no notable positive correlation between Eu anomalies versus Ce are present, which may indicate that modification of Eu was not accompanied by the modification of their Ce* values. Consequently, the wide range of negative Ce distributions may point to a change in paleoceanographic conditions from slightly suboxic to oxic paleo-redox state of ancient seawater but still remaining <0.9 [3]. MF-1 located in the lower part of the studied limestone succession is interpreted to have been deposited in an inner platform paleoenvironment during the Hauterivian-Aptian based on the microfacies characteristics and foraminifera assemblages (e.g., the high abundance of the benthic
The studied limestone is interpreted to have experienced progressive modification through interaction with reducing, high-temperature diagenetic fluids, and this caused the slight enrichment of Eu. However, no notable positive correlation between Eu anomalies versus Ce are present, which may indicate that modification of Eu was not accompanied by the modification of their Ce* values. Consequently, the wide range of negative Ce distributions may point to a change in paleoceanographic conditions from slightly suboxic to oxic paleo-redox state of ancient seawater but still remaining <0.9 [3]. MF-1 located in the lower part of the studied limestone succession is interpreted to have been deposited in an inner platform paleoenvironment during the Hauterivian-Aptian based on the microfacies characteristics and foraminifera assemblages (e.g., the high abundance of the benthic characteristics of foraminifera assemblages, lack of deep-water bioclasts and low biotic diversity) [31,67] (Figure 1.) It exhibits a wide range of Ce anomalies (0.38-0.81), with a steadily increasing trend upward in the succession, coincident with interpreted increasing palaeo-water depth ( Figure 10, Table 1). The uppermost part of MF-1 exhibits slightly weaker Ce/Ce* anomalies implying a relatively suboxic state of the ancient seawater. This is followed by a sharp decrease in the Ce/Ce* anomalies at the base of MF-2, suggesting a sudden change in the redox state of contemporaneous seawater. There is then a gradual transition to the overlying MF-3, indicating oxic seawater ( Figure 10, Table 1) which remained relatively stable until the end of MF-3. This is despite the progressive deepening of the basin, which might have been caused a slight decrease in oxygenation as relative sea-level increased [15].
Minerals 2020, 10, x FOR PEER REVIEW 17 of 25 characteristics of foraminifera assemblages, lack of deep-water bioclasts and low biotic diversity) [31,67] (Figure 1.) It exhibits a wide range of Ce anomalies (0.38-0.81), with a steadily increasing trend upward in the succession, coincident with interpreted increasing palaeo-water depth ( Figure 10, Table 1). The uppermost part of MF-1 exhibits slightly weaker Ce/Ce* anomalies implying a relatively suboxic state of the ancient seawater. This is followed by a sharp decrease in the Ce/Ce* anomalies at the base of MF-2, suggesting a sudden change in the redox state of contemporaneous seawater. There is then a gradual transition to the overlying MF-3, indicating oxic seawater ( Figure 10, Table 1) which remained relatively stable until the end of MF-3. This is despite the progressive deepening of the basin, which might have been caused a slight decrease in oxygenation as relative sea-level increased [15]. The differences in microfacies characteristics and foraminifera assemblage between MF-1 and overlying MF-2 may also confirm a change in paleoceanographic conditions because MF-2 is represented by the predominance of allochthonous bioclastic components which are broken and reworked, high biotic diversity and scarcity of small benthic foraminifera implying a deeper shelf environment relative to the MF-1. Furthermore, the presence of the rare plagioclase and basaltic extraclasts in the transition zone from MF-1 to MF-2 ( Figure 4c) is likely associated with the basaltic magma generation, which has been recently reported [114]. This basaltic magma generation is probably caused by synsedimentary tectonic activity. The syn-sedimentary tectonic regime may also cause progressive deepening of the basin. Furthermore, the submarine basaltic magma generation may also influence on the sudden change in paleo-redox conditions of ancient depositional environment. This likely result in the oxygen-deficient conditions, which produce our relatively high Ce/Ce* data (0.71 to 0.81) in the uppermost part of MF-1.
The Ce/Ce* anomaly at the MF1 / MF2 boundary is marked, and coincides with a change in facies from shallow water, benthic foraminiferal packstone to reworked skeletal packstone that is The differences in microfacies characteristics and foraminifera assemblage between MF-1 and overlying MF-2 may also confirm a change in paleoceanographic conditions because MF-2 is represented by the predominance of allochthonous bioclastic components which are broken and reworked, high biotic diversity and scarcity of small benthic foraminifera implying a deeper shelf environment relative to the MF-1. Furthermore, the presence of the rare plagioclase and basaltic extraclasts in the transition zone from MF-1 to MF-2 ( Figure 4c) is likely associated with the basaltic magma generation, which has been recently reported [114]. This basaltic magma generation is probably caused by synsedimentary tectonic activity. The syn-sedimentary tectonic regime may also cause progressive deepening of the basin. Furthermore, the submarine basaltic magma generation may also influence on the sudden change in paleo-redox conditions of ancient depositional environment. This likely result in the oxygen-deficient conditions, which produce our relatively high Ce/Ce* data (0.71 to 0.81) in the uppermost part of MF-1.
The Ce/Ce* anomaly at the MF1 / MF2 boundary is marked, and coincides with a change in facies from shallow water, benthic foraminiferal packstone to reworked skeletal packstone that is dominated by fragmented, allochthonous, largely heterotrophic fauna. This could reflect a facies transition from shallow water, platform top (MF1) to deeper water slope (MF2) setting, consistent with the overall deepening of the succession. However, the presence of fragmented organisms suggests depositional energy remained moderate, and water depths were not sufficiently deep for oxygenation to be reduced; the overlying MF3 sediments are more likely to be deeper water but remain oxic. The decrease in the Ce/Ce* anomaly coincides with a decrease in ΣREE, Y/Ho and Eu/Eu* anomaly ( Figure 10). Dating of benthic foraminifera indicates that a timing that is broadly consistent with the onset of OAE1a [122], where global temperatures rose, and ocean acidification occurred as a result of rising CO 2 concentrations during emplacement of Large Igneous Provinces. It is, therefore, possible that the MF1/MF2 boundary records this event, both in the increase in hydrothermal activity -which modified Eu/Eu*-and the decrease in seawater oxygenation that decreased Ce/Ce*. In this sense, the change in facies from MF1 to MF2 might not simply reflect a rise in relative sea level, and it might also indicate decreasing carbonate productivity and preservation.
Paleo-redox evolution can be controlled by the function of temperature, fluid chemistry of paleo-oceanic system triggered by complex tectono-sedimentary evolution of the Eastern Pontide. It is therefore recommended that paleoceanographic studies demonstrate more conclusively the occurrence of oxygen-deficient conditions in the Tethys ocean, within the study area. Our results have the potential to contribute new information to this discussion because a paleoceanographic reconstruction of the studied section can be established based on the faunal content, microfacies characteristics, and REE data. MF-1 represents an inner platform setting with mostly oxic conditions during the Hauterivian-Aptian. However, a sudden change in redox-state of ancient seawater, corresponding dysoxic conditions, is recorded in the most upper part of the MF-1. This is followed by an abrupt deepening paleoenvironment with a relative increase in the oxic state of the seawater and the deposition of the deeper water microfacies (MF-2 and MF-3) above a sharp transition. This represents a relative sea-level rise triggered by the evolution of the Tethys in the Eastern Pontide resulting deepening of the basin and is also recorded in age-equivalent carbonates in NE Turkey [37]. Previous studies have indicated that the extensional tectonics and rifting in the Cretaceous in NE Turkey [25,37,50,61] terminated in the middle Cretaceous, such that deeper paleoenvironmental conditions could be consistent with thermal subsidence, even though carbonate sedimentation continued until the end of Albian or Turonian [25,37,123]. Similarly, in this study, MF-1 represents an inner platform setting with mostly oxic conditions during the Hauterivian-Aptian. However, a sudden change in redox-state of ancient seawater, corresponding dysoxic conditions, is recorded in the most upper part of the MF-1 ( Figure 10). This is followed by an abrupt deepening paleoenvironment with a relative increase in the oxic state of the seawater and the deposition of the deeper water microfacies (MF-2 and MF-3) above a sharp transition. Moreover, an abrupt paleoenvironmental change is represented by the presence of MF2 and MF3 which include dark grey limestones with chert nodules, mud-rich textures, allochthonous skeletal fragments and sponge spiculitics and presence of the planktonic organism, all consistent with deeper water sedimentation. These characteristics can be indicative of an increase in the water depth. The gradual transition to MF-3 with a predominance of mud-rich texture, a decrease in abundance of allochthonous skeletal fragments support the transgressively deepening of the paleoenvironment which continued up to late Albian. It has also been recorded the sponge spicule wackestone with Microhedbergella rischi within the most upper part of MF-3 [37,51]. Thus, the presence of the planktonic organism may characterise the deepest part of the paleoenvironment which corresponding to the slope [65]. Further, their REEs data confirm that their oxic conditions remained stable throughout the deeper facies (MF-2 and MF-3), even considering the distal microfacies. The current work, therefore, supports that the shallower part of the paleo-ocean remained relatively less oxic and became suboxic during the Late Aptian-Albian, while deeper facies, which display an overall transgressive motif, were developed in relatively more oxic paleo-oceanic conditions up to the end Albian ( Figure 10).

Conclusions
The study of the Lower Cretaceous limestone succession, corroborated by detailed microfacies analysis and REE geochemistry, allows us to present new data on ocean paleoredox conditions ( Figure 10) and the extent of diagenetic alteration for the studied stratigraphical section in Mescitli area (Eastern Pontides, NE Turkey). Our primary results are as follows: 1.
Analyzed micritic limestone samples mainly exhibit a seawater signature including (1) (4)  The studied sections exhibit negative Ce* anomalies, varying from 0.38 to 0.81, which may confirm mostly oxic to dysoxic paleoceanographic conditions. Further, dyspoxic (Ce* = 0.71-0.81) conditions are also recorded in the most upper part of the MF-1 Microfacies during the late Aptian-early Albian.

4.
The current work suggests that the shallower part of the paleo-ocean remained relatively less oxic and became suboxic during the Late Aptian-Albian, while deeper facies displaying overall transgressive trend were developed in relatively more oxic paleo-oceanic conditions up to end of Albian.