Oxidative conditions can lead to exceptional preservation through phosphatization

Soft-tissue phosphatization, i.e., calcium phosphate mineralization occurring prior to the degradational collapse of cellular tissues, provides the most spectacular fossils of animals, still exhibiting subcellular details (Martill, 1990; Briggs, 2003). Most of the time, soft tissues are pseudomorphically replaced by francolite (a carbonate-rich fluorapatite), as is the case of the iconic ca. 110 Ma fossil fishes from the Santana Formation (Brazil) (Martill, 1988). Laboratory experiments have suggested that phosphatization results from a decay-induced fall in pH under anaerobic conditions, the decrease of pH being responsible for a switch from carbonate to phosphate precipitation (Allison, 1988; Briggs and Kear, 1993; Briggs and Wilby, 1996; Sagemann et al., 1999). However, authigenic phosphates such as francolite may precipitate under a range of redox conditions, including oxic to suboxic conditions (Föllmi, 1996, and references therein). Phosphatization also requires a sufficient concentration of phosphorus (Martill, 1988; Briggs and Kear, 1993; Briggs, 2003). Closed systems, such as those built by biofilms growing around carcasses, promote apatite precipitation by trapping phosphorus (Williams and Reimers, 1983; Martill, 1988; Briggs and Kear, 1993; Wilby et al., 1996), which is released by organic-matter degradation under oxic conditions (e.g., Meunier-Christmann et al., 1989; Mort et al., 2007). Here, we investigate phosphatization from a fossil perspective. We report an in-depth and in situ assessment of the redox conditions having prevailed during the phosphatization of “exceptionally preserved” Cretaceous fossils of fishes and crustaceans from the Jebel oum Tkout (OT1) Lagerstätte of Morocco.

A combination of advanced characterization tools, associating mass spectrometry, X-ray diffraction (XRD), scanning electron microscopy, and infrared (IR) spectroscopy with synchrotron-based methods, was used to achieve a micro-geochemical characterization of a series of fossils from the OT1 Lagerstätte. Major-to-trace elemental composition was determined using synchrotron-based micro–X-ray fluorescence (μXRF) mapping at a spatial resolution of 60–100 μm and at detection limits of a few tens of parts per million (Gueriau et al., 2014, 2018). Samples were scanned in air using a 17.2 keV beam. Under these conditions, the μXRF signal originates from the first hundreds of microns, and absorption by the air of low-energy photons from light elements such as phosphorus narrows the detection to elements from chlorine to uranium (see Gueriau et al., 2018). The spectral decomposition of μXRF data allowed estimating the concentrations of all detected elements, including of rare earth elements (REEs) that were then normalized to the post-Archean Australian shale (PAAS) reference (McLennan, 1989) to reconstruct REE patterns, ratios, and anomalies in the fossils at “local”, submillimeter scales (Gueriau et al., 2015). Finally, the redox states of Fe and Ce were determined using synchrotron-based micro–X-ray absorption spectroscopy (μXAS) at a 3–10 μm spatial resolution. For Fe, continuous Cauchy wavelet transformation of the spectra was performed to establish a more robust determination of the speciation (Muñoz et al., 2003). Additional information on methods is provided in the Supplemental Material¹.

The Upper Cretaceous (Cenomanian, ca. 95 Ma) OT1 Lagerstätte from southeastern Morocco (Fig. 1A) yielded a rich and well-preserved soft-bodied fauna including mollusks, insects, crustaceans, elasmobranchs, and actinopterygian fishes embedded in a pale beige laminated mudstone carved by mudcracks (Fig. 1B). Absence of marine organisms and the presence of mudcracks, unionids, and larvae of insects restricted to freshwater environments strongly suggest a low-energy seasonally dried freshwater deposition environment (Dutheil, 1999; Garassino et al., 2006). A paleoenvironmental model of the whole Kem Kem area is available in Ibrahim et al. (2020).

Remarkably, many fossils exhibit soft tissues (muscles, cuticles, and gills; Figs. 1C–1E; Dutheil, 1999; Garassino et al., 2006; Gueriau et al., 2015) pseudomorphically replaced by nanometric apatite crystallites (<30 nm), identified as francolite by infrared spectroscopy (Fig. 1F). XRD analyses on oriented preparations show that the mineralogy of the sedimentary matrix is dominated by illite and quartz, with some illite-smectite mixed layers and kaolinite. These minerals are mainly detrital, except for the accordion-like automorphic crystals of kaolinite. Low trace-metal concentrations (Fig. S1C in the Supplemental Material) with very low degrees of enrichment (PAAS-normalized enrichment factors for V, Ni, Cu, Zn ≤ 1.1, Cr ∼1.9) and chemical index of alteration of ∼78 (Nesbitt and Young, 1982; low range for illite) indicate very moderate weathering. Gypsum filling mudcracks (Fig. 1B), halite crystals locally found on the sample surface (Figs. S2L, S2 M, and S2P), and Fe-rich clay minerals displaying plate-like or sheet-like habits forming roses on the sample surface (Figs. S2G–S2J, S2N, and S2O) are clearly secondary (and rather recent), indicating that sampling and storage conditions may have slightly altered the samples.

Long known for promoting apatite precipitation by trapping phosphorus (Williams and Reimers, 1983; Martill, 1988; Briggs and Kear, 1993; Wilby et al., 1996), fossil biofilms can be observed at the OT1 Lagerstätte. They consist of large (tens of square centimeters) reddish cracked films lying a few millimeters above each of the fossiliferous layers (Fig. 2A). The morphological similarities between these films (Figs. 2B–2D) and modern microbial mats (Fig. 2E) strengthen their identification as well-preserved fossil colonies of microorganisms. A few hundred milligrams of fossil biofilm materials were extracted and milled into powder for XRD analyses, revealing that these large biofilms are composed of Fe₂O₃ [Fe(III) oxide] minerals (Fig. 2F).

Moreover, the fossil skeletons, cuticles, and soft tissues (white to yellowish in color) are largely covered by similar reddish (to sometimes bluish) thin (a few tens of micrometers) patches (Figs. 1B, 1C, and 3A; Fig. S2A) that are interpreted as remains of biofilms. Reddish patches are also found embedded within the phosphatized soft tissues (Figs. 3C–3F; Fig. S3C), suggesting that their precipitation was concomitant with phosphatization. μXRF mapping reveals that these thin biofilm patches are iron rich and contain trace metals (Figs. 3A, 3B, and 3G; Figs. S1 and S3). μXAS investigations at the Fe K-edge reveal that the chemical speciation of iron in these patches is identical to that of iron in the large biofilms and confirms the presence of Fe(III) (Fig. 3H). The spectrum of this Fe-rich mineral is rather inconsistent with hematite (α-Fe₂O₃) or illite {(K,H₃O)(Al,Mg,Fe)₂(Si,Al)₄O₁₀[(OH)₂,(H₂O)]} and less consistent with goethite [α-FeO(OH)] or lepidocrocite [γ-FeO(OH)], while it is consistent with ferrihydrite (Fe₂O₃ · 0.5H₂O) (Figs. 3H and 3I; Fig. S4).

The observed Fe(III) hydroxides surround skeleton tissues and are found around and within the phosphatized soft tissues (Figs. 3C–3F; Fig. S3C), suggesting that they precipitated concomitantly with phosphatization. Although Fe-rich phases found in association with fossils elsewhere have been interpreted as resulting from the oxidation of pyrite, as attested by pseudomorph minerals or molds left in the surrounding matrix (e.g., Gabbott et al., 2004; Osés et al., 2016), this is not the case for the Fe(III) hydroxides observed in our samples. Instead of a cubic or framboidal habit, typical of pyrite weathering products, the observed Fe(III) hydroxides show a honeycomb-like morphology (Figs. 3E and 3F; Fig. S2), rather typical of microbial mats (e.g., Frankel and Bazylinski, 2003; Davies et al., 2016). In particular, one can observe spherical imprints, possibly created by coccoid-shaped bacteria (e.g., Iniesto et al., 2016) or gas bubbles having been produced within the microbial mats (e.g., Davies et al., 2016). The fossil biofilms discussed here have thus not been exposed to significant weathering or recrystallization. In other words, the observed Fe(III) hydroxides are likely primary, i.e., they precipitated during the life of the microbial mats, thereby indicating slightly oxic conditions during phosphatization.

Synchrotron-based μXRF analyses of the phosphatic fossil tissues reveal that they incorporated REEs, strontium, and thorium, which substitute for calcium in apatite minerals (Fig. 3G). The notable incorporation of REEs occurs after decay, with concentrations increasing by three to four orders of magnitude in 10³–10⁴ yr depending on diagenetic conditions (Herwartz et al., 2013). The obtained REE patterns display a very limited enrichment in intermediate REEs (“bell-shaped” patterns) and a slightly negative Ce anomaly (Fig. S5A; Gueriau et al., 2015). The PAAS-normalized La/Yb versus La/Sm values confirm that the fossils investigated underwent very limited weathering and recrystallization (Fig. 3J; Reynard et al., 1999; Lécuyer et al., 2003). The observed negative Ce anomaly thus reflects early diagenetic redox conditions (German and Elderfield, 1990) and indicates that slightly oxic conditions persisted after decay. Ce L₃-edge spectroscopy confirms that Ce has a mixed valence (Gueriau et al., 2015), with oxidized Ce [i.e., Ce(IV)] contributing to 20 at% of the total Ce (Fig. S5B).

Instead of a fall in pH under anoxic conditions, the fine-scale micro-geochemical characterization conducted here strongly suggests that (slightly) oxic conditions prevailed during phosphatization in the OT1 Lagerstätte. This conclusion is consistent with some of the latest experiment results evidencing that chemical microenvironments generated by microbial mats may turn oxic after several weeks despite initially anoxic conditions (Iniesto et al., 2015). One could argue that the observed Fe(III) hydroxides may have been produced by anaerobic neutrophilic iron oxidizers (e.g., Miot et al., 2009; Hedrich et al., 2011). This would not be inconsistent with the present interpretation, i.e., the establishment of slightly oxic conditions during phosphatization, because neutrophilic iron oxidizers require the presence of nitrates to survive, i.e., slightly oxic conditions (anaerobic and anoxic are not strict synonyms). Altogether, the present results establish that exceptional preservation through phosphatization and oxidative conditions are not antithetic.

We acknowledge SOLEIL synchrotron (Saint-Aubin, France) for provision of beamtime under project 20131308. J.-G. Bréhéret (Université de Tours, Tours, France) performed and interpreted the XRD analyses on the bulk sediment. V. Beltran and J. Kaddissy (IPANEMA, France) performed infrared spectroscopy. We thank S. Blanchandin (SOLEIL) for assistance with X-ray powder diffraction of the fossils; D. Vantelon and N. Trcera at the LUCIA beamline (SOLEIL Synchrotron); and teams at the Institut des Sciences Analytiques (UMR 5280 CNRS, Lyon, France) and Service d’Analyze des Roches et des Minéraux (CRPG, UMR 7358 CNRS-INSU, Vandœuvre-lès-Nancy, France) for carrying out the mass-spectrometry measurements. The 2012 field expedition and measurements were supported by the Muséum national d’Histoire naturelle, France (MNHN) through the “ATM Biodiversité actuelle et fossile” and by the UMR 7207 CR2P. We thank the organizers and field workers of the expedition. This work was developed as part of the IPANEMA-MNHN agreement on collaborative research. We thank Paul Wilson and the anonymous reviewers for constructive suggestions to improve the quality of this manuscript.