The diagenesis of biosilica remains poorly known, but is increasingly important for reconstructing past oceanic silica levels using silicon isotopes. Here, we present SEM and XRD analyses of sponge spicules from the Late Jurassic, Late Cretaceous, and Eocene, compared with modern samples, to reveal their modifications with time. Modern spicules are composed of opal-A, distinct from sinter-derived opal-A. Many Eocene spicules also preserve opal-A, but show signs of early transformation. Most Eocene and all Cretaceous spicules consist of opal-CT, while quartz occurs in some Cretaceous and all Jurassic samples. Eocene opal-A spicules are macroscopically glassy, whereas opal-CT spicules exhibit a milky and/or opaque appearance, due to the presence of silica microspheres. Cretaceous spicules range from milky to opaque, and Jurassic spicules are typically opaque, containing microspheres, silica nanogranules, and microquartz. The structural and mineralogical evolution is reflected in decreasing Full Width at Half Maximum (FWHM) values of diffractograms with increasing age. Some Cretaceous and Jurassic spicules contain quartz blocks, formed by the fusion of silica nanogranules, while euhedral quartz occurs in both Cretaceous and Jurassic samples. Although diagenetic stages can vary within a single spicule, all retain at least some elements of their original structure and morphology. The observed mineralogical transitions reflect dominant solid-state maturation through dehydration and sintering and/or Ostwald ripening, rather than dissolution, mould formation, and reprecipitation from external fluids. Our findings indicate that sponge spicules preserved through solid-state transformation are preferred targets for silicon isotope studies. However, the assumption that such spicules retain their original isotopic signal must be further verified through integrated structural and isotopic investigations.
