The Fe-Mn crust sample DGCR1314U01 from KC-7 in the Magellan Seamount cluster ( Figure 1 ) was recovered using a dredge line from 16°54.869′N and 151°50.030′E to 16°55.152′N and 151°50.220′E (length: approximately 0.62 km; depth: 1536–1523 mbsl) during the CR1302 expedition (Korea Institute of Ocean Science & Technology, February 14–March 13, 2014; R/V Onnuri). The approximately 104 mm thick Fe-Mn oxide layers were divided into seven distinct layers by optical observation, which included a highly altered basaltic substrate. Co chronometry (Manheim and Lane-Bostwick, 1988; Puteanus and Halbach, 1988) using modified empirical equation (Equation 1 in supplementary) (Moon, 2016) dated the age of the lowest Fe-Mn oxide layer at 40.78 Ma, with an average growth rate of approximately 2.60 mm/Ma observed for the middle Eocene ( Supplementary Tables S1 , S2 ). Dating was performed at one or two points in the middle of each distinct layer to minimize the influence of layer boundary inclination. Fe-Mn crust samples were halved, and powder samples obtained each layer using a microdrill with a tungsten carbide carving bit (Dremel 3000, Ø=0.8 mm), which was then consolidated using Akasel Aka-Resin liquid epoxy and polished using sandpaper. Thin sections were then prepared for geochemical and mineralogical analysis of the basaltic substrate and the lowest part of the Fe-Mn oxide layer.

A Nikon ECLIPSE LV100N POL polarization microscope at the Korea Institute of Ocean Science & Technology (KIOST), Busan, Republic of Korea was used to examine thin sections of the basaltic substrate. Open and crossed Nicol prism conditions were used for optical observation and mineral identification.

Powder samples were obtained using a microdrill and immersed in distilled water in a sonicator to minimize flocculation (Yang et al., 2019). The oriented samples were then placed on glass slides. To determine the mineralogy of the Fe-Mn crust and basaltic substrate, X-ray Diffraction (XRD) was performed using a PAN analytical X’pert Pro system at KIOST, under the conditions of 45 kV and 30 mA and a CuKα radiation source. The receiving slit was set to >0.2° to optimize the resolution of the XRD profile and the peak intensity (Hurst et al., 1997; Jung et al., 2022). XRD profiles were obtained over a 2θ range of 4–60° at scan speed of 1.2°/min and a scan step size of 0.02°, and mineral composition was determined using the Crystallographica Search-Match software (version 2.0.3.1).

Elemental Al, Si, K, and Fe compositions were measured using the Bruker M4 Tornado μ-XRF at the Korea Institute of Industrial Technology, Wonju, Republic of Korea. Analysis was performed using Rh anode tube at 50 kV and collection time of 1 s. Element distribution maps were constructed using geochemical analysis of the basaltic substrate in the Fe-Mn crust, which was performed in 0.025 mm steps with a count time of 600 ms per point. Built-in XRF software (Bruker XRF, Germany) was used for the quantitative analysis of each element.

Microscopic and geochemical analyses were performed using thin sections coated with Pt. Sections were examined using the ZEISS Gemini 500 field emission scanning electron microscope (FE-SEM) at an acceleration voltage of 10 kV and working distance of 5 mm and by energy dispersive X-ray spectrometry (EDS) at an acceleration voltage of 15 kV and a working distance of 8.5 mm. Analysis was performed at the Pusan National University, Busan, Republic of Korea. Backscattered electron (BSE) images were obtained for chemical distinction of the alteration rims and the groundmass. EDS was used to examine the chemical composition of the distinct alteration rims.

Micro Raman spectroscopy was conducted using a Ramboss star micro-Raman system (DXG) at the Gwangju Institute of Science and Technology, Gwangju, Republic of Korea. A 785 nm laser source (>90 mW) was used to ascertain the clay mineral composition of the alteration rims in the basaltic substrate. Raman spectra were collected with a spectral resolution of ~0.24 cm^–1 using a 1200 gr/mm grating, with both the lower wavenumber (27–1230 cm^–1) and the hydroxyl-stretching regions (3446–3711 cm^–1) examined. The laser power on the sample was operated through a 0.40 NA, 20x objective at about 28.5 mW. Each Raman spectrum was obtained as an average of 10 spectra recorded over 10 s. Spectra were cross-referenced against spectral mineral patterns in the laboratory collection and databases available through CrystalSleuth provided by RRUFF for comparison. Spectrometer calibration was performed using the Raman band for a silicon wafer at 520 cm^–1 (Bernardini et al., 2024). The measured spectrum was baseline corrected and fitted with Voigt function to obtain the phonon wavenumbers and intensities.

Samples for TEM were obtained from the alteration rim of the thin section using an FEI Scios DualBeam focused ion beam. Samples were placed on a holey-carbon TEM Cu grid (TED PELLA Inc.) and a TALOS F200X field-emission transmission electron microscope at Pusan National University was used to identify the microstructure and chemical composition of the clay minerals. TEM was performed at 200 kV to examine the clay minerals on the lattice fringes and for EDS analysis. Bright-field TEM images and fast Fourier transform (FFT) images were evaluated by Gatan digital micrograph software (version 3.01.598.0, Gatan Inc.) to identify the d-spacing of the lattice fringes and analyze the microstructures of the clay minerals. The chemical formula of the smectites was calculated as atoms per formula unit (apfu) based on O₂₀(OH)₄. To identify the different types of smectites, the octahedral chemical compositions were plotted on an ^VIAl–Fe–Mg ternary diagram that included saponite (Mg-rich), montmorillonite-beidellite (Al-rich), and nontronite (Fe-rich) domains (Weaver, 1989; Marinoni et al., 2008; Jung et al., 2022).

Visual observation showed that the Fe-Mn crusts are divided into seven layers, including a basaltic substrate. The texture of the upper layers 1–3 was porous and infilled with sediments ( Table 1 ), while the lower layers 4–6 exhibited more common characteristics of a relatively blocky texture. Both texture differentiation and distinct colors were observed in the layers ( Table 1 ). L1 (at the crust surface) was a reddish-brown color and 23 mm thick; L2 was dark brown and ~18 mm thick; L3 was yellowish brown and ~20 mm thick; L4 was dark gray and ~13 mm thick; L5 was a gray layer with white chunks and ~15 mm thick; and L6 was light gray in color with white chunks and ~15 mm thick ( Table 1 ). The upper Fe-Mn layers were dated to approximately 19.18–2.84 Ma and the lower layers to approximately 40.78–26.45 Ma ( Table 1 ).

The highly altered basaltic substrate (Ba) included dark green matrix material, in which micro-XRF analysis confirmed relatively high Si, Al, and K contents ( Figure 2 ), and a light-yellow matrix ( Figure 3 ). Greenish-brown alteration rims were observed in the dark green matrix ( Figure 3A ) and traces of planktonic foraminifera (Globorotalia menardii) in the light-yellow matrix ( Figure 3B ).

Diffractograms of bulk powder samples from each layer (L1–Ba) were examined to obtain the mineral assemblage of the air-dried samples ( Figure 4 ). Manganese oxide minerals, identified as vernadite (δ-MnO₂), were observed in all manganese layers, appearing as a broad peak at 2θ of 36.6°, while distinct carbonate fluorapatite (CFA) was identified in the lower part of the Fe-Mn crust, L5–Ba ( Figure 4 ).

Several differences between the substrate matrix and alteration rims were observed in the BSE images and elemental maps ( Figures 5A–H ). The matrix exhibited high concentrations of P (average 14.93 wt. %) and Ca (average 31.77 wt.%) while showing relatively low Si (average 0.24 wt.%) and Al (0.03 wt.%) ( Figures 5B, C ). The alteration rims showed higher Si concentrations (average of 21.13 wt. %) and Al (average 8.66 wt.%) compared to the matrix ( Figures 5D, E ), along with relatively higher K (average 3.33 wt.%), Mg (average 2.28 wt.%), and Fe (average 9.40 wt.%) contents ( Figures 5D–H ). These chemical compositions suggest that the alteration rims are likely composed of clay minerals.

The quantitative analysis results of the metal elements in the manganese oxide of layer 6 were plotted on a Fe-Mn-(Co + Ni + Cu) ternary diagram (Bonatti, 1972), with the plot showing the distribution of the manganese oxide elements corresponding to the hydrogenetic domain ( Figure 6 ).

Raman spectroscopy offers significant advantages in identifying specific types of clay minerals that are often undetectable (or indistinguishable) by X-ray diffraction measurements. The Raman spectrum measured for the alteration rim within the 250–600 cm^-1 spectral region shows two distinct broad bands with 250–450 cm^-1 and 450–575 cm^-1 ( Figure 7 ). This spectral region includes the complexation of two clay minerals, identified as nontronite and glauconite, by comparison with the reference Raman spectra X050120 (Nontronite) and R070295 from the RRUFF database. The observed spectral band between 250 and 450 cm^-1 is fitted using four distinct bands. The Raman bands at approximately 356 and 418 cm^-1 are attributed to the vibrations of SiO₄ in nontronite (Frost and Kloprogge, 2000). Other bands at 335 and 388 cm^-1 were assigned to glauconite, with the 388 cm^-1 band being the most prominent for glauconite when using a red laser (Coccato et al., 2016). For 450 – 575 cm^-1 spectral region, the observed band was deconvoluted with three spectral bands. The first band at 502 cm^-1 is assigned to nontronite, while the bands appeared 526 cm^-1 and 544 cm^-1 are attributed to glauconite. The Raman spectra results indicate that nontronite and glauconite are the main clay mineral components of the alteration rim in the basaltic substrate, which consistent with the elemental analysis.

The lattice fringes in the EDS spectrum were observed using a representative TEM bright-field micrograph, with FFT images used to observe the clay minerals ( Figures 8A, B ). Clay minerals including chlorite with a 1.4 nm spacing, smectite with a 1.2 nm spacing, and mica with a 1.0 nm spacing, were observed. Smectite and mica were interlayered with Fe-oxyhydroxide with a 0.25 nm spacing ( Figure 8A ). The chemical composition of smectite was [(K_0.47–1.04)(Fe_0.87–1.62Al_1.91–2.57Mg_0.72–1.55)(Al_0.78–1.56Si_6.44–7.22)O₂₀(OH)₄; Table 2 ], and the Al/Si ratio was between 0.34 and 0.50. The EDS results for the octahedral cation were plotted on an ^VIAl-Fe-Mg ternary diagram to compare the smectite group endmembers, confirming the presence of a Fe-smectite adjacent to the nontronite endmember ( Figure 9A ). The interlayer Fe-smectite showed K enrichment characteristics ( Figure 9B ), indicating that the smectite at the alteration rim of the basaltic substrate was K-enriched Fe-smectite.

Water-rock interactions that influence the global elemental cycles tend to form secondary minerals as products. In marine environments, these minerals form as a result of interactions between seawater and basalt at seamounts or in the oceanic crust (Jarrard, 2003; Yamashita et al., 2019; Ramos et al., 2020). In this study, secondary minerals were observed in the dark green areas of the basaltic substrate, in which relatively high Al, Si, and K concentrations were expected to lead to clay mineral formation ( Figure 8 ). Furthermore, greenish-brown alteration rims indicative of clay minerals and are characterized by higher Al, Si, and K concentrations compared with those of the surrounding matrix were also identified in the dark green areas of the basaltic substrate ( Figures 2 , 5 ). Alteration rims were formed by the alteration of basalt via seawater-basalt interaction as fluids infiltrate veins or fractures (Alt and Honnorez, 1984). Since clay minerals are used as proxies for reconstructing paleo-ocean environments (Yang et al., 2016, 2018; Jung et al., 2022; Kim et al., 2023), the clay minerals in the alteration rim can be used to reconstruct the oxic-suboxic conditions on seamount flanks prior to the formation of the Fe-Mn crust. However, identifying clay minerals in the samples using XRD ( Figure 4 ) was challenging owing to the low abundance and poor crystallinity of the authigenic minerals. Micro Raman spectroscopy allowed for non-destructive analysis of material with poor crystalline characteristics. The micro Raman spectroscopy and TEM analyses confirmed the presence of smectite, and glauconite in the alteration rims ( Figures 7 , 8 ). Fe-smectite is generally observed in submarine environments in which basaltic glass has undergone low temperature hydrothermal alteration (Melson and Thompson, 1973; Alt et al., 1987; Yang et al., 2018; Yamashita et al., 2019). The Fe-smectite that forms in hydrothermal settings typically exhibits low Mg and K contents; however, the Fe-smectite on seamount KC-7 was relatively enriched in Mg and K characteristics ( Figure 9 ). According to previous research (Yamashita et al., 2019), Si and Fe are derived from basalt dissolution during authigenic mineral formation resulting from seawater-basalt interactions, whereas Mg and K are derived from circulating seawater. Moreover, preferential K fixation in nontronite is the result of its large cell dimensions, which are compatible with larger ionic radius of K compared with those of Na and Ca (Meunier, 2005; Yamashita et al., 2019). Therefore, the Fe-smectite in the basaltic substrate of the Fe-Mn crust likely formed via interactions between oxygenated seawater and basalt before the incrustation of the Fe-Mn crust. Additionally, Fe-oxyhydroxides with a d-spacing of 0.25 nm ( Figure 8 ) co-occurred with the Fe-smectite, indicating oxic conditions (Hein et al., 1979; Alt et al., 1987; Chevrier et al., 2007). Therefore, the Fe-smectite in the basaltic substrate formed in an environment in which oxygen was continually supplied by seawater circulation. A steady oxygen supply is necessary to form authigenic clay minerals in marine environments, whereas maintaining non-vigorous seawater circulation is crucial for preventing the dilution of Mg and K derived from seawater (Yamashita et al., 2019). Si and Fe likely originated from the basalt during seawater-basalt interactions, while K, Mg, and O were likely supplied by the non-vigorous seawater circulation driven by upwelling near the seamount flank. In other words, Fe-smectite authigenesis is expected to involve non-vigorous circulation of oxygenated seawater ( Figure 10A ).

Mica interstratified with Fe-smectite was identified as glauconite ( Figure 7 ). The nature of glauconite is challenging owing to subtle differences (e.g., similar physical, chemical, and mineralogical characteristics) from other green clays (López-Quirós et al., 2020; Singh et al., 2023). Glauconite is predominantly a K- and Fe-rich mica that forms in suboxic conditions that typically occur in shallow marine environments (< 500 m) at around 15°C (Odin, 1988; López-Quirós et al., 2019, 2020). However, previous studies have suggested that glauconite forms in deep-sea environments (> 2000 m) at lower temperatures (3–6°C) (Giresse and Wiewióra, 2001; Cuadros et al., 2011; Baldermann et al., 2013; López-Quirós et al., 2019). Thus, the presence of glauconite suggests that the micro-environment in the basaltic substrate may have transitioned from oxic to suboxic conditions ( Figure 10B ). The formation of glauconite on seamount flanks in suboxic micro-environmental conditions likely occurred after the Fe-Mn crust incrustation, as seawater-basalt interaction weakened owing to the global phosphatization event that formed the CFA (Benites et al., 2021; Hein et al., 2017). The porosity of hydrogenetic Fe-Mn crusts is approximately 60% (Conrad et al., 2017; Hein et al., 2017), and CFA is known to saturate the pores of Fe-Mn crusts (Yang et al., 2019). Therefore, CFA precipitation would have weakened seawater-basalt interaction and reduced the supply of oxygenated seawater. As a result, glauconite formation was likely promoted via the remineralization of organic matter and glauconitization in a suboxic micro-environment (López-Quirós et al., 2020). In other words, the presence of authigenic phyllosilicates formed by seawater-basalt interactions suggests that K fixation and Fe cycling occurred on the flanks of seamount KC-7 ( Figure 10 ). The Fe-smectite likely formed in the basaltic substrate first, followed by glauconite, suggesting that the micro-environmental conditions in the basalt on the seamount flanks may have transitioned from oxic to suboxic conditions following the formation of the Fe-Mn crust.