Geological field observations were conducted in North Konawe and South Bungku, southeastern Sulawesi, in four stratigraphic sections from south to north: Wiwi-1, Wiwi-2, Bule-1, and Bule-2 ( Figure 1 ). Based on a previous survey by the Indonesia Geological Agency [21] , lithologies traversed in all sections belonged to the Tokala Formation. Because no age data were available and biostratigraphic analysis was not the main focus of this study, the assignment of the formations was later validated based on correlations between lithological characteristics and regional lithostratigraphy [19] - [21] . A total of 72 rock samples were collected from shale and limestone outcrops along road cuts. These samples were considered fresh because the outcrops were excavated to a thickness of at least 15 cm prior to sampling. Two samples showing oil stains was also acquired from limestone in the Bule-1 and Bule-2 sections ( Figure 1 ). To ensure preservation, all samples were stored in well-sealed plastic bags, and the oil-stained rock was wrapped with aluminum foil to avoid contamination [29] .

A total of 27 samples (19 shale, 8 limestone) selected from all sections were used for petrographic analysis to obtain detailed lithological descriptions. Prior to preparation, shale was hardened with epoxy solution (resin and catalyst) and left for 12 h to avoid internal breakup, because the lithology was too friable for primary cutting. After primary cutting, shale and limestone chunks were polished sequentially with mesh #350, #800, #1000, #2000, and #3000, and attached to glass slides. Secondary cutting was performed to thin attached samples to a thickness of ca. 1 - 2 mm, and a Prepa-Lap cutter (Maruto Instrument, Fukuoka, Japan) equivalent to #600 mesh powder, was utilized to grind thick blocks down to a thickness of ca. 30 μm. Finally, the samples were polished sequentially with powder from mesh #1000, #2000, #3000, and #6000. These petrographic slides were observed under a microscope with transmitted and reflected light (H600L; Nikon, Tokyo, Japan).

A total of 49 representative samples (43 shale, 6 limestone) from all sections were subjected to XRF analysis (ZXS Primus II XRF; Rigaku, Tokyo, Japan) to determine the content of major oxides (SiO₂, TiO₂, Al₂O₃, Fe₂O₃, MnO, MgO, CaO, K₂O, and P₂O₅) and some trace elements (Ba, Cr, Cu, Nb, Ni, Pb, Rb, S, Sr, V, Y, Zn, and Zr). Prior to analysis, samples were pulverized to obtain 5 - 6 g of homogeneous powder, which was carefully placed within the assigned pellet rings. The sample powders were compressed using a flat disk to obtain appropriate pellet samples, as described in detail elsewhere [30] .

Source rock characteristics were evaluated by pyrolysis to determine the bulk organic geochemical compositions of 72 samples (40 shales, 10 mudstones, and 22 limestones) using the Rock-Eval 6 apparatus (Vinci Technologies, Nanterre, France). These samples were ground into #300 - #400 mesh powder samples (clay size), of which 0.6 - 0.7 g was placed in crucibles for later analysis, as described previously (Lafargue et al., 1998; Behar et al., 2001). Parameters acquired in Rock-Eval analysis included total organic carbon (TOC); the amounts of free hydrocarbons detected in the sample (S1), and generated via the thermal cracking of nonvolatile organic matter (S2); the amount of carbon dioxide (CO₂) generated through kerogen pyrolysis (S3); and Tmax.

From the south section, the stratigraphic interval of the Tokala Formation consisted of intercalated limestone and shale, with limestone beds predominating in the Wiwi-1 section ( Figure 3 ). The limestone and shale were well stratified with a sharp contact ( Figure 4(A) , Figure 4(B) ), where the Limestone was light gray, matrix-supported, moderately sorted, and massive, with each layer ranging in thickness from 32 to 50 cm ( Figure 4(D) ). Under microscopic observation, the limestone was composed of skeletal grains, mostly foraminifera followed by mollusk grains, in a micritic matrix; therefore, the lithology was assigned as foraminifera wackestone (FW) facies.

Foraminifera were diverse, with elongated, oval, and circular shapes. Thin shale layers (5 - 12 cm) intercalated with wackestone were matrix-supported, intensively bioturbated, composed of carbonate grains within argillaceous matrix and occasional carbonaceous layers ( Figure 4(C) , Figure 4(E) , Figure 4(I) ). Carbonate grains were predominantly composed of skeletal fragments derived from foraminifera and mollusks ranging in size from 0.15 to 1.1 mm. Compared to the limestone, skeletal grains in the shale layers tended to be homogeneous. Notably, almost all original structures of the skeletal fragments had already been destroyed by extensive diagenesis, specifically neomorphism, as large calcite crystals grew within shells ( Figures 4(F)-(H) ). Significant concentrations of

amorphous organic matter (AOM) and structured organic matter (SOM) were observed in the size range of 0.025 - 0.060 mm ( Figure 4(I) ). This organic matter was dispersed throughout the matrix, but occasionally occurred as laminas ( Figure 4(E) , Figure 4(F) ). Another significant feature observed during macroscopic examination was the occurrence of bioturbation within the shale layer ( Figure 4(C) ). Later, shale in the Wiwi-1 section was assigned as massive bioturbated shale (MBS) facies.

Another shale/limestone alternation of the Tokala Formation was outcropped in the Wiwi-2 section, but in contrast to that described in the previous section, this alternation was dominated by shale approximately 42 m thick ( Figure 5(A) , Figure 5(C) ). The limestone was poorly stratified and contact between these lithologies was unclear as the layer tended to be discontinuous. Interestingly, deformation was observed, where the more brittle limestone blocks showed signs of fracture and the underlying shale layers tended to be bent due to its ductility ( Figure 5(D) ). Shales were dark gray and could be considered massive based on field observations, but tended to have weak (irregularly sinuous) lamination ( Figure 5(F) , Figure 5(G) ).

The bioturbation intensity within this interval was relatively poorly developed. Organic matter such as AOM tended to be more abundant, with both dispersed and laminar distributions ( Figures 5(H)-(L) ). Framboidal pyrite was common and poorly sorted within the shale facies, predominantly composed of large framboids (10.7 - 17.5 μm), followed by moderately sized framboids (6 - 9 μm; Figure 6 ). We assigned this shale as weakly laminated shale (WLS) facies. Although the shale layers emitted an oily odor, no oil stains were visible during macroscopic or microscopic observation. The limestone layers transitioned into finer grains supported by matrix, in which no skeletal fragments were observed. Therefore, limestone in this section was assigned as lime mudstone (LM) facies. Further north, another Tokala Formation interval was observed as 23-m-thick outcrops with good stratification of limestone and shale layers of equal thickness in the Bule-1 and Bule-2 sections (Tokala Formation; Figure 3 ). The shale layers showed similar characteristics to all of the abovementioned sections; however, skeletal fragments were suspended between argillaceous matrix and organic-rich layers ( Figure 6(G) , Figure 6(I) , Figure 6(J) ).

The bioturbation intensity within this interval was relatively poorly developed. Organic matter such as AOM tended to be more abundant, with both dispersed

and laminar distributions ( Figures 5(H)-(L) ). Framboidal pyrite was common and poorly sorted within the shale facies, predominantly composed of large framboids (10.7 - 17.5 μm), followed by moderately sized framboids (6 - 9 μm; Figure 5 ). We assigned this shale as weakly laminated shale (WLS) facies ( Figure 5 ). Although the shale layers emitted an oily odor, no oil stains were visible during macroscopic or microscopic observation. The limestone layers transitioned into finer grains supported by matrix, in which no skeletal fragments were observed. Therefore, limestone in this section was assigned as lime mudstone (LM) facies.

Further north, another Tokala Formation interval was observed as 23-m-thick outcrops with good stratification of limestone and shale layers of equal thickness in the Bule-1 and Bule-2 sections (Tokala Formation; Figure 3 ). The shale layers showed similar characteristics to all of the abovementioned sections; however, skeletal fragments were suspended between argillaceous matrix and organic-rich layers ( Figure 6(G) , Figure 6(I) , Figure 6(J) ). The bioturbation intensity within

this interval was also poorly developed, while organic matter such as AOM tended to be more abundant in laminar distributions ( Figure 6(G) , Figure 6(I) , Figure 6(J) ). Framboidal pyrite was common and poorly sorted within the shale facies, predominantly composed of large framboids (10.7 - 17.5 μm).

In general, SiO₂, Al₂O₃, and CaO were the predominant major oxides in our samples, depending on the lithology. Therefore, we describe the composition of major oxides, and provide our analysis of other oxides in Supplementary Data 1A. The shale lithofacies traversed in this research area generally showed varied elemental concentrations. MBCAS facies had 24.06% - 24.11% SiO₂ content, 10.33% - 10.55% Al₂O₃ content, and 28.66% - 28.72% CaO content; WLACS facies exhibited 35.60% - 36.60% SiO₂ content, 14.15% - 14.75% Al₂O₃ content, and 17.61% - 34% CaO content; and SLCAS facies exhibited 19.50% - 23.66% SiO₂ content, 4.34% - 7.27% Al₂O₃ content, and 36.11% - 51% CaO content. Furthermore, FW and LM facies were characterized by high CaO content (47.71% - 53.05%), low SiO₂ content (3.45% - 8.74%), and extremely low Al₂O₃ content (0.24% - 0.81%). Strong correlations were observed among each oxide, particularly the three dominant oxides. Al₂O₃ content was positively correlated with SiO₂ and K₂O content and negatively correlated with CaO content ( Figure 7 ). CaO content was negatively correlated with both SiO₂ and MgO content ( Figure 7 ). These relationships may be related to mineralogical content and could potentially be applied for petrochemical discrimination.

The bulk organic geochemistry of all lithofacies in this research tended to be rich in TOC and had high potential yield (PY). Specifically, the MBCAS and WLACS facies showed high TOC values, ranging from 1.20% to 1.28% and 3.9% to 4.0%, respectively, with corresponding PY values ranging from 2.10 to 2.24 and from 6.14 to 20.91, respectively (Supplementary Data 1C). Hydrogen index (HI) and S2 values were also high in the MBCAS (176 - 182 mg HC/g and 2.03 - 2.06 mg/g, respectively) and WLACS (447 - 500 mg HC/g and 17.42 - 20.24 mg/g, respectively) facies. In contrast, oxygen index (OI) and S3 values tended to be slightly lower in the MBCAS facies, ranging as 54 - 62 and 0.28 - 0.35 mg CO₂/g, respectively, and extremely low in the WLACS facies, ranging as 7 - 11 and 0.30 - 0.36 mg CO₂/g, respectively.

Similarly, the SLCAS facies had high TOC content (2.07% - 2.56%) and PY (7.40 - 13.24). HI values ranged from 384 to 472 mg HC/g in SLCAS facies, and S2 values were high (7.3 - 13.13 mg HC/g). S3 and OI showed low values, at 0.60 - 0.75 mg CO₂/g and 24 - 36 mg HC/g, respectively. Organic matter in the source rock of all lithofacies exhibited early thermal maturity, with Tmax values of 436˚C - 437˚C for MBCAS, 434˚C - 436˚C for WLACS, and 434˚C - 437˚C for SLCAS. Our geochemical characterization of the limestone facies FW and LM

showed low to moderate TOC content. Specifically, TOC content ranged as 0.56 - 0.79 wt.% for FW and 0.29 - 0.33 wt.% for LM, corresponding to low to moderate PY values (0.10 - 3.64 and 0.77 - 0.80, respectively). These facies had significant HI values of 90 - 449 and 229 - 230 mg HC/g, respectively. S2 values were low, ranging as 0.09 - 3.55 for FW facies and 0.75 - 0.76 for LM facies. OI and S3 values ranged as 41 - 54 and 0.21 - 0.32 mg CO₂/g, respectively, in the FW facies and 81 - 88 and 0.28 - 0.29 mg CO₂/g, respectively, in the LM facies (as shown in detail in Supplementary Data 1B), indicating significant concentrations in these facies.

A problem arose when our lithofacies discrimination method was applied to the multiple shale layers found in every section. Rock described as massive mudstone in the field could be massive, weakly, or strongly laminated mudstone according to our thin section observations. As it would be impossible to prepare thin sections from all shale layers, we turned to quantitative petrochemical classification. The content of three major constituents is commonly used for mudstone and shale classification: quartz, carbonates, and clay minerals [31] - [33] , with their content typically determined by XRD analysis. Not specific only for shale and mudstone, the determination mineralogical composition obtained from XRD analysis even also commonly used for soil or slag [34] . As quantitative mineralogical data were not available, the content of major elements determined by XRF analysis was used in this study (Supplementary Data 1A). This type of classification approach has been applied in previous studies [35] for example, proposed chemical discrimination for sedimentary rocks, dividing shale into “shale” and “Fe-shale,” and interpreting relative Al₂O₃ content as being related to grain size [36] . In the present study, all samples were initially classified as mudstone (sensu stricto) or shale, when fissility was observed. Therefore, we applied a direct classification scheme based on the content of the three major shale components [31] - [33] , in which quartz, carbonate, and clay minerals are represented by SiO₂, CaO, and Al₂O₃, respectively. This approach is justified by the correlations detected among these compounds. Our samples showed a positive correlation between SiO₂ and Al₂O₃. In some cases, increased SiO₂ content was followed by decreased Al₂O₃ content due to fluctuation between biogenic and detrital inputs [37] . Given the observed positive correlation, these oxides were interpreted as being related solely to either detrital fractions or argillaceous materials (aluminosilicates). The fluctuation between biogenic and detrital inputs was further shown by the negative correlations of CaO with SiO₂ and Al₂O₃, as increasing carbonate proportions were associated with reduced proportions of argillaceous and siliceous minerals ( Figure 8 ).

In regard to that, our classification scheme distinguished three groups,

argillaceous, “mixed,” and calcareous shales ( Figure 8 ), where the term “mixed” pertains to the balance among the three major components (CaO, SiO₂, and Al₂O₃), resulting in classifications such as argillaceous-calcareous (ar-ca) and calcareous-argillaceous (ca-ar), which are terms adapted from [38] .

In regard to that, he petrochemistry further discriminated the observed lithofacies, resulting in the differentiation of samples into distinct clusters. For example, the WLS facies exemplified the ar-ca group, which was typified by WLACS from the Wiwi-2 section, whereas the MBS and SLS facies further represented the ca-ar group, which was typified by MBCAS facies from the Wiwi-1 section and SLCAS facies from the Bule-1 and Bule-2 sections. The SLCAS facies had higher CaO content, indirectly suggesting higher carbonate content compared to other shale facies. The WLACS facies had the highest SiO₂ and Al₂O₃ content, which may be related to high argillaceous input. The MBCAS facies showed similar proportions of SiO₂, Al₂O₃, and CaO ( Figure 8 ). This balance between carbonate and argillaceous content (CaO-SiO₂) may have been related to the deposition energy, with higher carbonate levels correlated with more stable energy [39] [40] and higher argillaceous content related to more sediment supply due to high energy, perhaps associated with the supply of detritus materials from the land ( Figure 8 ).

The source rock maturation process can significantly reduce TOC content and S2 values [41] - [43] . Therefore, these parameters do not represent initial organic matter conditions, and thermal maturity information is required prior to further geochemical characterization. According to our data, all facies had mostly reached the early mature stage (T_max: 435˚C - 437˚C), with some samples still in the immature stage [T_max < 435˚C; 37]. This thermal stage is the best condition for understanding source rock characteristics, as the organic matter has not been altered by maturation and still represents initial syn-depositional conditions.

The geochemical characterization of source rock has been conducted on several Mesozoic shales, including Tokala Formation, within neighboring basin of our research area, Tomori Basin [18] exhibiting 0.32 - 3.46 wt.% of TOC. These data aligned with the all of our shale facies (MBCAS, WLACS, SLCAS) which also show high value of TOC in range from 1.20 - 4.0 wt.%. This similarity indicates that the Tokala Formation from Tomori Basin and from our research area in southern Sulawesi represent potential source rock characteristics in the term of quantity organic matter.

In this research, we adopted a common source rock classification [44] . The carbonate-dominated facies (CaO > 50%), particularly the LM and FW facies, are considered carbonate-poor, as indicated by their low TOC and S2 levels ( Figure 9(A) ). The shale facies tended to have higher levels of organic matter, increasing in the order MBCAS (good) < SLCAS (good to very good) < WLACS (very good to excellent) ( Figure 9(A) ).

Furthermore, in terms of the quality of organic matter, all of our facies tended to be composed of type II kerogen ( Figure 9(B) ) with HI increasing in the order FW < MBCAS < LM < SLCAS < WLACS. This HI trend may be related to the amount of hydrogen-rich organic material, either based on the original organic matter or excellent preservation of organic matter in the sediments [43] .

All shales in the research area tended to have higher organic matter content than “true carbonate” (CaO > 50%) intervals. This phenomenon diverged slightly from the definition of carbonate source rock as organic-rich sedimentary rock comprised of at least 50% carbonate minerals, but also corresponds to layers containing less than 50% carbonate but with substantial amounts of marl [39] . Although two important geological scenarios for organic matter enrichment in carbonate rocks have been highlighted [45] ., i.e., high primary productivity and elevated organic preservation, we believe that organic matter enrichment and other source rock characteristics are closely related to lithological, mineralogical, and petrochemical aspects of the rock. For example, the defined carbonate source rock [39] appears to be composed of a mixture of carbonate and argillaceous materials, which is a mineralogical matter regardless of depositional and enrichment processes.

The clearest distinction between poor and good source rocks was observed from petrochemical and mineralogical perspectives, in terms of the inverse

relationship of CaO-SiO₂ and Al₂O₃ ratios. Increasing carbonate content was followed by slight decreases in TOC content and PY ( Figure 10 ). Moreover, high siliciclastic material input was negatively correlated with source rock quantity and quality (TOC, PY, and HI). Therefore, the quantity and quality of carbonate source rock are functions of carbonate and siliciclastic content, which are linked to sedimentation processes. Slow carbonate accumulation results in poor organic matter preservation, leading to the degradation of organic matter. Moreover, rapid siliciclastic transportation and sedimentation during mudstone deposition

can dilute the OM concentration with high clay input [43] - [46] .

Not restricted to sedimentation processes, the presence of argillaceous content is also important for the formation of carbonate source rock, as clay minerals can adsorb organic matter [47] - [49] . To explain the possibility of clay adsorption, we utilize Al₂O₃ as a representative clay mineral. Positive correlations were observed between Al₂O₃ and PY, TOC, and HI values ( Figures 10(E)-(G) ), suggesting that clay minerals influenced the organic matter enrichment [49] [50] , in particular within the source rock, as demonstrated in our study. Therefore, the occurrence of clay minerals in the depositional environment should be considered along with organic preservation and productivity [45] .

Another interesting observation is that MBCAS facies trends were always shifted, such that increases in Al₂O₃ were not followed by corresponding increases in TOC, PY, and HI. Such conditions could potentially be influenced by macro-scale factors other than mineralogical and petrochemical factors.

The decreases in TOC content, PY, and HI, even with decreasing CaO and increasing Al₂O₃ content, in MBCAS facies may have been related to bioturbation. Bioturbation causes organic matter destruction, and therefore significantly influences organism activity and oxic conditions [43] . Organism activity can destroy accumulated organic matter (as food), and oxic conditions may affect organic matter degradation, as discussed in previous studies [e.g., [51] and [52] ] where bioturbation was shown to have caused destruction of organic matter. With regard to laminated intervals, our SLCAS facies had high organic matter quantity and quality, which are also typical for carbonate source rock [39] [52] [53] , and the WLACS facies showed excellent organic matter quality. Lamination indicates slow, stable deposition, which is suitable for organic matter preservation [39] [45] . Based on these factors, we summarized the lithological, mineralogical, and petrochemical characteristics associated with our facies variations to understand which facies are preferable as source rock and prospective horizons ( Figure 11 ).