Thin sections were made in the Department of Earth Sciences (UCL). First, raw samples (hand specimen) were trimmed using a diamond saw, ground and polished before cleaning with water, drying and mounting with epoxy onto standard 27 × 46 mm glass slides. Then, samples around 30 μm thickness were polished with progressively finer diamond-based compounds, down to the last step with 0.25 mm alumina, to produce polished thin sections suitable for optical microscopy. Finally, thin sections were scanned with a flatbed document scanner to create petrographic maps that could be annotated.

Petrographic descriptions of the sixteen selected thin sections were carried out in the Geological Spectroscopy Laboratory at the Department of Earth Sciences (UCL). The polarising microscope consists of an Olympus BX51 petrographic microscope with UIS2/UIS (Universal Infinity System) optical system. Imaging modes used were transmitted light, crossed polar, and reflected light. Objectives used for petrographic analysis were ×5, 10X, 20X, 50X, and 100X and no immersion oil was used. Petrographic descriptions in polished thin sections allowed the mapping of minerals and sedimentological targets onto the thin section maps to be subsequently analysed by micro-Raman spectroscopy.

Raman spectra and images were obtained in the Geological Spectroscopy Laboratory at UCL with a high-resolution WiTec alpha300 confocal Raman imaging system using a 532−nm wavelength laser with output laser power maintained at ∼9 mW. This system used a grating with 600 lines per mm to provide a spectral resolution of around 4 cm⁻¹ over a bandwidth of 4,000 cm⁻¹. Spectral accuracy was calibrated against an in-house gem-grade diamond standard and was better than 1 cm⁻¹ over years of reproducibility tests. The acquisition time for each spectrum in Raman images was set between 0.5 and 0.8 s, and an optic fibre of 50 μm diameter was selected to provide confocality and acceptable signal intensity.

Data acquisition was made through the WITec Project Four Plus software. It allowed the creation of hyperspectral images for OM and specific minerals using peak intensity mapping for characteristic peaks in a scan in which each point of the scanned area is associated with a Raman spectrum. Raman hyperspectral scans were obtained at variable magnifications, mainly with ×50 and ×100 objectives, yielding variable spatial resolutions of up to 360 nm/pixel. All targets were analysed 1–5 μm below the thin section surface to minimise the detection of possible surface contaminants. In addition, polished thin sections were cleaned with an acetone rub using nitrile gloves and kimwipes to remove excess solvent, which helps to avoid potential surface contamination. Cosmic ray reduction was applied on all Raman spectra, and their backgrounds were fitted to a polynomial function of order 7 and subtracted. Raman hyperspectral images were generated with the WITec Project Four Plus software by mapping the primary peak intensities diagnostic for specific mineral phases, including 465 cm⁻¹ for quartz in blue, 1,090 cm⁻¹ for dolomite in green, 965 cm⁻¹ for apatite in turquoise, 515 cm⁻¹ for feldspar, 990 cm⁻¹ for barite, and 148 cm⁻¹ for anatase. This approach produces micro-Raman images in which colours and spectra are correlated with optical properties to identify OM and associated minerals, which is ideal for understanding their petrographic distribution.

For OM, represented in red colour in hyperspectral images, the Raman spectrum is composed of two prominent peaks: the disordered (D) band (∼1,350 cm⁻¹) and the graphite (G) band (∼1,594 cm⁻¹), that represent represents sp³-hybridised C atoms bonded to variable heteroatoms, and sp²-hybridised C atoms bonded to three other atoms, respectively. When disordered, the OM Raman spectrum presents additional bands named D2 (∼1,620 cm⁻¹), D3 (∼1,510 cm⁻¹) and D4 (∼1,245 cm⁻¹) bands (Beyssac et al., 2002). The bands mentioned above are calculated after deconvoluting the spectra range from 1,090 to 1700 cm⁻¹ using Lorentzian functions. Additional key Raman spectral parameters that helped characterise the OM are the full width at half maximum (FWHM) of D and G bands and the I-1350/1,600 ratio. Figure 3 shows the peak-fitting and deconvolution of the Raman spectrum for Member IV.

A Carl Zeiss EVO25 scanning electron microscope with an Oxford Instrument Xmax 80 electron dispersive spectrometer (EDS) detector was used at the Institute of Archaeology (UCL). Eight thin sections were selected for analysis and carbon-coated using a Quorum Technologies K975X carbon coater before analysis. Because of the necessary carbon coating, all EDS analyses were performed after micro-Raman examinations. The working distance was 8.5 mm, with an electron beam of 20 keV. Spectra were collected using Oxford Instruments Aztec 5.0 software, optimised against a pure cobalt standard before data collection, with a deadtime of approximately 40% on the cobalt. X-ray maps were collected with a pixel resolution of 1,024^*768 pixels and a dwell time of 200 microseconds per pixel. Detection of distinctive x-ray emissions from the sample was done by energy dispersive spectroscopy (EDS) to detect elements and provide independent, quantitative confirmation of mineral compositions. ZAF correction was applied to minimise the impact of the atomic number, absorption, and fluorescence excitation effects.

Diagenetic spheroids are found in the shelf lagoon and lower slope environment and are primarily formed of carbonate and phosphatic concretions (>1 cm) and rosettes (<200 μm) (Figures 1D, J, L, 4–7). Concretions present variable sizes and morphologies, between 10–20 cm and with an ellipsoidal shape (Figure 1D). Metre size concretions are also found in the area (Figure 1L). However, micrometre size spheroids are mostly dolomitic, with average diameters of 95 ± 35 μm (n=12) (Figure 4A), with spherical (Figure 5C) to ellipsoidal (Figures 5A, B, D, E, H) morphologies. Occasionally, dolomitic spheroids can reach diameters between 250–500 μm (Figures 5A, B, D, E). The core of rosettes observed in Member IV is circularly concentric with an inner core of OM surrounded by dolomite rims, formed by individual dolomite crystals between 5–20 μm (Figures 5B, C) that sometimes form aggregates (Figures 5D, E).

Although restricted to phosphatic concretions (Figure 6A), coccoidal morphologies resembling microfossils are also observed in Member IV. They present variable sizes between 5–20 μm with average diameters of 10.0 ± 3.0 μm (n=277) (Figures 4C, D) and are hosted in pale and dark bands and appear randomly and densely distributed. Raman and SEM results reveal that apatite is the primary component of the pale bands’ matrix (Figures 6D–F, H). The coccoidal shapes preserve their original spheroidal shape with no identifiable internal structures. In general, these are isolated and not interconnected. The vast majority of these microfossil-like objects have been replaced by quartz (Figures 6C, D), whereas other coccoidal shapes show a partial quartz replacement with an inner core of OM surrounded by a quarzitic rim (red arrows in Figure 6G).

Additional diagenetic spheroids in Member IV occur in the upper slope environment in the form of microscopic sub-spheroidal morphologies (Figure 7). They have an average diameter of 59.0 ± 8.0 μm (n=6) and are surrounded by a quartz rim up to 15 μm wide (Figure 7B). The inner core is composed of highly fluorescent OM (HF-OM) and phosphate minerals (Figures 7B, C). In some of these spheres, barite crystals up to 25 μm long are found to replace the phosphatic phase (Figures 7B, C, E). In some cases, the spheroids have sizes up to 50 μm. They present characteristic concentric layers with distinctive mineral zonation alternating pale and dark grey from the core towards the outer rim (Figure 7G). SEM analyses show that the pale-grey layers correspond to the strengite phase based on its Fe and P content (Figures 7G, H, spectrum 1) and that the dark-grey layers correspond to the variscite phase based on its Al and P content (Figures 7G, H, spectrum 2).

Samples are primarily composed of clays, carbonates, phosphates and OM. In addition, accessory minerals such as barite, pyrite, feldspar and anatase are also present. Carbonate minerals are exclusively found in Zhimaping (proximal section) (Figure 5F–H) and in Xiajiaomeng (distal section) (Figures 5A–E). These carbonate minerals are generally present as euhedral to subhedral crystals and commonly form spheroids and ellipsoids. The carbonate phase is identified as dolomite from the Raman peaks at 157, 283, 714 and 1,090 cm⁻¹ (Figure 5J). Note that the dolomite peaks in Zhimaping and Xiajiaomeng are slightly shifted compared to distinctive dolomite peaks (170, 290, 720 and 1,095 cm⁻¹). However, the EDS spectra shown in Figure 5I from carbonates in Figure 5G, with peaks for Ca, Mg, C, and O, support the occurrence of dolomite as the primary carbonate phase. Phosphate minerals in Member IV are only found in the shelf lagoon (Zhimaping) and the upper slope (Taoying). In the shelf lagoon, phosphates appear in horizontal apatite bands, up to 200 μm thick, that host variable amounts of black and white microfossil-like structures (Figure 6). Characteristic Raman peak at 965 cm⁻¹ (Figure 6E) and EDS spectra with peaks for P, Ca and F (Figure 6H) reveal that the phosphate phase corresponds to fluorapatite. In the upper slope, however, phosphates appear to form microscopic spheroidal morphologies (Figures 7B, C, E, G). Raman spectra show these spheroidal morphologies are associated with highly fluorescent OM (HF-OM in Figure 7B) and surrounded by a rim of similarly sized quartz crystals and elongated barite grains (Figure 7B). However, complementary SEM analyses reveal the presence of strengite [Fe³⁺PO₄ ⋅ 2H₂O) and variscite [Al³⁺PO₄ ⋅ 2H₂O) associated with the highly fluorescent OM fraction (Figures 7G, H). The cryptocrystalline nature of these minerals in the spheroids is consistent with the high fluorescence and lack of peaks other than OM in Raman spectra.

Sulfur-bearing minerals in the studied samples predominantly occur in the form of barite [BaSO₄], jarosite [KFe₃(SO₄)₂(OH)₆] and pyrite [FeS₂]. Barite is identified in proximal and distal sections from its diagnostic Raman peaks at 455 and 989 cm⁻¹ (Figures 7D, E). In the upper slope, barite is present as small and dispersed irregular crystals with sizes <10 μm, associated with quartz, filling voids and partially replacing organic-rich spheroidal morphologies (Figures 7B, C, E). In the basin section, barite appears as isolated euhedral crystals up to 50 μm embedded in an organic-rich matrix with no signs of grain dissolution (Figure 8B). Jarosite is exclusively found in the Fengtan section (distal environment). These crystals are subhedral brownish grains up to 40 μm in size without any preferential orientation (Figure 8B). This mineral is confirmed by Raman spectra, which show characteristic peaks at 221, 433, 625, 1,008 and 1,102 cm⁻¹ (Figure 8E). Pyrite is also commonly found in all the studied samples, except those from Taoying. It is identified from its yellow reflective colour and characteristic Raman peaks at 345 and 381 cm⁻¹ (Figure 5J). Pyrite mostly occurs as rounded, framboidal, subhedral, and euhedral habits with sizes between 10 μm and 2 mm.

Feldspars are also present in the studied samples. They exhibit variable sizes between 5–30 μm in the upper slope (Figure 7F) and 5–40 μm in the deep basin (Figure 8D). In these environments, feldspars are generally subangular or elongated in shape and with no preferred orientation. Characteristic Raman peaks confirm this mineral at 467 and 513 cm⁻¹ (Figure 7D) correspond to either polymorph orthoclase or microcline [KAlSi₃O₈] based on elemental analyses with 11% K, 9% Al, 32% Si and 48% O of targeted minerals (Spectrum 3 Figure 8F).

Anatase (TiO₂) is present in all studied environments and shows characteristic Raman peaks at 146 (very strong), 398, 517 and 638 cm⁻¹ (Figures 6E, 7D, 8E). However, anatase crystals have different size distributions related to specific mineral associations. For example, in samples associated with sulfur-bearing minerals (pyrite and/or jarosite), anatase is randomly distributed within the matrix and mainly occurs in the form of subhedral and elongated crystals <7.5 μm and most often <5 μm (Figure 8B). However, in samples where it is associated with apatite or barite, anatase has rounded and sub-rounded morphologies. It occurs as isolated crystals or forming aggregates between 5 and 15 μm (Figure 7F).

Organic matter presents the characteristic Raman D and G peaks at ∼1,335 and 1,604 cm⁻¹ (Figures 5J, 6H, 7D, 8E). Three different types of kerogen structures are identified in Member IV, as reported by Cañadas et al. (2022), who concluded that, after evaluating the potential impact of variable degradation and maturation, the different kerogens corresponded to variable dominance of organic carbon sources (photosynthetic vs. secondary). Kerogen “type-A” represents OM sourced from photosynthetic production and presents a wide D1 band, a more intense and narrower G band, and an I-1350/1,600 ratio of 0.6–0.7 (Table 1, Figures 5J, 6H). Kerogen “type-B” corresponds to OM resulting from secondary productivity and is characterised by narrower D1 and G bands and a resolvable shoulder of the D4 band at ∼1,220 cm⁻¹, with an I-1350/1,600 ratio of 0.8 (Table 1, Figure 8E). And lastly, kerogen “type-C” represents mixed OM formed by primary and secondary production fractions, with combined features from type-A and B kerogens based on a more intense G band, like type-A, and narrower D1 and G bands, with an I-1350/1,600 ratio of 1 (Table 1 and Figure 7D). The three kerogen types are characterised by a specific distribution within the basin and a specific mineral association: 1) Kerogen “type-A” occurs in the shelf lagoon and lower slope and is primarily associated with dolomite, phosphates (apatite), anatase and pyrite; 2) Kerogen “type-B” is found in the basin and appears predominantly associated with Ba- and Fe-sulfates, anatase and large euhedral pyrite; and 3) Kerogen “type-C” occurs in the upper slope and is associated with phosphates (strengite-variscite), barite and alkali-bearing phases (feldspar). Table 1 present a detailed summary of the Raman parameters for OM, alongside the associated mineralogy and Table 2 presents characteristics of the diagenetic spheroid according to the mineral association identified in the organic-rich shales of Member IV.

Carbonate diagenetic spheroids are widespread in organic-rich shales and are found in a large range of size and morphology, although usually spherical and ellipsoidal. Their origin is commonly explained as the result of different styles of abiotic physical growth. The two principal styles are replacement growth, where host sediments become incorporated into a concretionary body via extensive cementation of pore spaces, and displacive growth, which occurs when sedimentary beds are forced apart by the development of carbonate minerals (Sellés-Martínez, 1996). In Member IV, two principal types of carbonate diagenetic spheroids are documented. The first group, concretions, includes metre to centimetre size concretions, between 2 cm and 1 m, generally isolated, ellipsoidal and aligned with the sedimentary bedding (Figures 1D, L, J) (Dong et al., 2008; 2013; Zhou et al., 2017). In addition, this group presents characteristic displacive growth where host rock laminations are bent around the concretion contact, which is generally related to early diagenetic concretionary growth before compaction (Sellés-Martínez, 1996). The second group, carbonate aggregates and rosettes, encloses micrometre spherical and ellipsoidal concretions formed via replacement growth and characterised by cores of OM surrounded by outlying dolomite of the concretionary bodies (Figure 5) (Dong et al., 2008; 2013).

The formation of the first group (concretions) has been traditionally explained as the result of carbonate supersaturation in porewaters driven by the generation of bicarbonate and, therefore, high alkalinity from microbial activity around concretion nuclei in the sediments (Sellés-Martínez, 1996). Then, the nuclei were sustained during growth by a pool of inorganic bicarbonate that facilitated rapid precipitation of the concretion matrix during early diagenesis (Raiswell and Fisher, 2000; Gaines and Vorhies, 2016). Thus, the δ¹³C of carbonate concretions generally depends on the source of carbonate ions in pore waters. As a result, δ¹³C values of carbonate concretions are usually characterised by ¹³C depletions relative to seawater, often used as fingerprints of specific microbial metabolic processes that trigger and maintain concretion growth (Raiswell and Fisher, 2000). However, limestone concretions with ¹³C-enrichments in the geological record have also been attributed to methanogenesis (e.g., Birgel et al., 2015).

An alternative mechanism to form these spheroid structures has been attributed to wave action in high-energy shallow marine environments (Akin et al., 2013). However, the estimated sedimentation rates of the studies sections argue against this mechanism. In the areas where spheroidal morphologies are present, sedimentation rates are remarkably low (between 0.02 and 0.34 cm/kyr, calculated as total sediment thickness divided by total time), and shales are relatively homogeneous in granulometry. Thus, it is likely that shales were deposited below or near storm wave base, as already proposed based on other sedimentological evidence such as planar laminations and only rare intraclasts (McFadden et al., 2008), where slow-moving currents prevented coarse-grained sediment from migrating into low-energy depositional environments. These observations are inconsistent with a high-energy environment; therefore, the wave agitation model is not considered a potential mechanism of concretion formation in this study.

The formation of the second group (aggregates and rosettes) has been attributed to possible gas bubbles produced by oxygenic phototrophs in microbial mats and stromatolites (Bosak et al., 2010; Sallstedt et al., 2018), gas bubbles replaced by carbonate formed by microbial decomposition of OM and anaerobic oxidation of methane near the sediment-water interface (Dong et al., 2008) or microbial biomass replaced by carbonate precipitation under slightly alkaline conditions (Varkouhi et al., 2022). In all cases, gases from OM decomposition would have created cavities or bubbles subsequently filled by carbonate precipitated from anaerobic methane oxidation. However, as explained next, gas bubbles from organisms or OM decomposition are not exclusive mechanisms to form concretions.

The conventional characteristics of carbonate concretions are that they preferentially occur in organic-rich mudstones, where the decomposition of OM facilitates the continuous supply of bicarbonate in pore water, and that they are formed at shallow depths during the early stages of diagenesis (Sellés-Martínez, 1996; Raiswell and Fisher, 2000). However, ellipsoidal carbonate concretions have been reported in organic-poor glacial deposits (Vuillemin et al., 2013), organic-poor mudstones (Gaines and Vorhies, 2016) and organic-poor green silty shale (Liu A. Q. et al., 2019), indicating that concretion growth may be sustained even under limited organic content conditions. In addition, the decomposition of OM and anaerobic oxidation of produced gases do not explain relevant attributes of concretions, such as their spherical to ellipsoidal concentricity with sharp boundaries or the factors that terminate concretion growth. To further understand these questions, an alternative model of formation for diagenetic concretions must include biological and abiotic processes, or a combination of both because they are not mutually exclusive and because diagenesis involves both microbial and physico-chemical processes. Indeed, abiotic processes could have played a significant role in forming diagenetic concretions in Member IV.

Phosphate spheroids in organic-rich shales commonly form during very early diagenesis, allowing the preservation of original depositional materials such as fossils and other sedimentological features (Raiswell and Fisher, 2000). This could be the case for the morphologies found in the shelf lagoon (Zhimaping) area that resemble coccoidal cyanobacterial mats (Figures 6C, D, F) or the upper slope (Taoying) with characteristic organic-rich spheroidal and rosette morphologies (Figures 7A–C, E, G). The coccoidal cyanobacterial mats are enclosed in phosphatic concretions (Figure 6A) and present an excellent level of preservation of these objects, indicating that the growth of phosphate concretions occurred during early diagenesis. Lastly, the two different spheroidal morphologies in Taoying, organic-rich spheroidal and rosettes, likely have different origins as the former also resembles cyanobacteria-like microfossils (Figures 7B, C, E) with a phosphatic core whose origin could be attributed to microfossil replacement by phosphate precipitation, and the latter present concentric laminations whose origin could be attributed to COR (Figure 7G) as discussed in the next section.

One may argue that an abiotic origin of the coccoidal morphologies, such as grains partially silicified due to compaction, cannot be ruled out. However, the homogeneous morphology and sizes, with average values of 9.7 ± 3.2 μm, spatial organisation within a phosphatic concretion and chemical composition, and the geochemical context in which they formed support the interpretation that they are likely fossils. Also, while coccoidal microfossils are typically aggregated (e.g., She et al., 2013), they can occur as individual spheres scattered across a bed. For example, similar fossils of colonial coccoidal cyanobacteria preserved as individual spheres with similar sizes were reported in Silurian organic-rich cherts (Kremer, 2006), Mesoproterozoic stromatolites (Golubic and Seong-Joo, 1999) and late Devonian benthic cyanobacterial mats (Kazmierczak et al., 2012), which are compared to modern coccoid cyanobacteria (Chroococcus). These studies suggested that cyanobacteria and algal biomass were the dominant microbial population of large blooms and formed macroaggregates that rapidly sank to the seafloor and were overgrown by benthic coccoidal cyanobacterial mats. In such a dynamic scenario, the decomposition of biomass would release phosphate during shallow burial that would have favoured early diagenetic phosphatisation as calcium, barium, iron cations, and halogens were available in pore water. Interestingly, the occurrence of apatite bands (Figure 6), iron-rich phosphate phases (Figure 7) and large barite grains associated with phosphates (Figure 7F) described in this study suggest that the abiotic decomposition of biomass may have also played a key role in the formation of these mineral phases in Member IV (more in the next section).

In South China, coccoidal microfossils have been previously reported in phosphorites of the Doushantuo Fm. (Xiao, 2004; Muscente et al., 2015), located in the same proximal environments as those in the present study. Similarly to the microfossils found in Member IV, the well-preserved cellular structures reported in these studies suggest phosphatisation occurred immediately after deposition. Different P sources likely accounted for the accumulation and preservation of apatite. This is because, during the Ediacaran, continental weathering and P remineralisation were the principal sources of phosphorus in the environment (Laakso et al., 2020; Dodd et al., 2023), which favoured the extensive production of OM (Cañadas et al., 2022) and the proliferation of microbial communities (She et al., 2013), like the coccoidal cyanobacteria shown in Figure 6. The organic decay of these microbial communities during burial caused the release of P remineralised from OM, probably via bacteria sulfate reduction. The redox conditions that prevailed during Member IV deposition in the Ediacaran oceans, with increasing oxic conditions in the shallow water column and anoxic (or euxinic) conditions dominating in the deep basin (Cañadas et al., 2022; Li et al., 2024) promoted iron-(oxyhydr)oxide (FeOOH) particulates to efficiently adsorb dissolved phosphate and transported the Fe-bound phosphorous (FeOOH·PO₄) into marine sediments, ultimately precipitating as diagenetic apatite bands and phosphorite deposits (Cui et al., 2015; 2016).

Phosphatic spheroids and rosettes occur in organic-rich laminated horizons at Zhimaping and Taoying sections (Figures 7A–D, G). However, although these spheroids have been previously reported to occur in black shales (Kidder et al., 2003), they are more commonly associated with other lithologies such as carbonates and chert (Mossman et al., 2005; Papineau et al., 2016). The occurrence of apatite spheroids and rosettes in organic-rich sediments from proximal environments is consistent with a phosphorus-rich environment that stimulated a high level of primary productivity (Cañadas et al., 2022), similarly as described for Member II (She et al., 2013; She et al., 2014) and the stromatolitic phosphorite from the late Paleoproterozoic Aravalli Supergroup (Papineau et al., 2016). Zhimaping and Taoying are located in proximal areas where carbon and nitrogen isotope studies suggest high photoautotrophic production (Cañadas et al., 2022) and, together with the excellent preservation of microfossils, suggest the influence of microorganisms during phosphogenesis.

The spherical and ellipsoidal morphologies of the spheroids shown in Figure 5 and the concentric layered pattern observed in Figure 7G resemble the circularly concentric and radially patterned structures previously described as ‘diagenetic spheroids’ (Papineau et al., 2016; 2017; 2021; Dodd et al., 2018; Papineau, 2020; Gabriel et al., 2021; Goodwin and Papineau, 2022). For all these examples of diagenetic spheroids, diagenetic oxidation of OM through out-of-equilibrium reactions has been proposed as the principal mechanism of pattern formation with circularly concentric layers. In addition, these examples of diagenetic spheroids present similar sizes, generally between 100 and 200 μm, and mineral assemblages that commonly include ¹³C-depleted carbonate, ferric-ferrous, and phosphate phases mixed with OM and sulphide.

Evidence for the diagenetic oxidation of OM in Member IV is seen in the systematically negative δ¹³C_carb values down to −8‰ and coeval δ¹³C_org down to −39‰ (Figure 2B) (Cañadas et al., 2022). Hence, a ¹³C-depleted oxidised source of OM was assimilated by authigenic carbonate and phosphate in Member IV. Furthermore, the core characteristic of the different spheroids described in the organic-rich shales of Member IV is that OM is a ubiquitous primary component. Hence, it is reasonable to propose that OM played a relevant role in forming these diagenetic spheroidal structures.

In this regard, the degradation of OM occurs through different reactions depending on the diagenetic stage. Thus, in the early stages of diagenesis, OM releases free carboxylic acids containing new metabolic intermediates that other organisms readily use. This process occurs through biotic decarboxylation reactions, such as Tricarboxylic Acid Cycle (TCA cycle), sulfate reduction and fermentation. As diagenesis progresses, carboxylic acids decompose abiotically via decarboxylation reactions to form CO₂. Finally, during late diagenesis, abiotic decarboxylation becomes the principal process that produces CO₂, ultimately contributing to carbonate and phosphate supersaturation for precipitation and concretionary growth.