Extensive decentralized hydrogen export from the Atlantis Massif

A continuous supply of hydrogen formed by the interaction of water with mantle rocks may have driven the formation of organic molecules on early Earth and other planets, laying the prebiotic groundwork for life (Martin and Russell, 2007; Sojo et al., 2016). Hydrogen likely fueled the development and evolution of early life, as several lines of evidence point to a central role of H₂-utilizing metabolic pathways in the last universal common ancestor (Weiss et al., 2017).

Fuel alone is not sufficient for the formation of prebiotic organics and early life. Conducive temperatures, availability of inorganic carbon, mineral surfaces, wet-dry cycles, and metal catalysts also determine reaction extents and shape the earliest biomolecules necessary to harness energy and transfer information (e.g., Yu et al., 2013; Camprubi et al., 2017; Frenkel-Pinter et al., 2020). In this context, the locations and environmental conditions in which H₂ forms are as important as the quantities that are generated.

We report the widespread export of H₂ across the Atlantis Massif oceanic core complex (Mid-Atlantic Ridge, 30°N), which is distinct from the circulation system and channeled flow of the nearby Lost City Hydrothermal Field (LCHF). This type of decentralized export starkly contrasts with magmatic-dominated systems at mid-ocean ridges where hot, buoyant fluids discharging from depths of ∼1–3 km are channeled into high-permeability up-flow zones (Fisher, 2004; McCaig et al., 2007). The Atlantis Massif decentralized fluids are elevated in H₂ and migrate through environments that satisfy multiple requirements for the development of early life.

Samples and data were collected during International Ocean Discovery Program (IODP) Expedition 357 (October 2015) and the Lost City 2018 Return expedition (September 2018) (see Text S1 in the Supplemental Material¹). In both expeditions, water column samples collected by hydrocasts were analyzed aboard the ship for volatile concentrations (H₂, CH₄) (Fig. 1; ¹Figs. S1 and S2). In 2018, the remotely operated vehicle (ROV) Jason carried out visual seafloor surveys and was outfitted with in situ sensors that continually monitored oxidation-reduction potential (ORP) and temperature.

The results from Expedition 357 were previously reported by Früh-Green et al. (2017, 2018) and are summarized here. Near-seabed hydrocasts across the top of the massif captured fluids with H₂ concentrations regularly in the tens of nanomolars and with a high of 44 nM above one of the central boreholes, compared to background seawater concentrations of <1 nM (Fig. 1; Table S2). Fluids flushed from the bore-holes during drilling reached H₂ concentrations up to 322 nM (Table S1). In contrast, CH₄ concentrations were low to undetectable, ranging from below the detection limit of 0.3 nM to a high of 5 nM in hydrocast and borehole fluids (Tables S1 and S2).

In 2018, we also directly observed fluid flow at multiple locations across the top of the massif. Series of orthogonal veins along a sediment-free region at the top of the massif are infilled with carbonate (Fig. 1; Fig. S3). Similar fractures have been attributed to volume expansion during serpentinization (Denny et al., 2016). Active flow was confirmed by a 1 °C increase above background on a temperature probe.

Flow through the carbonate sands that cover much of the rocky surface over the central portion of the massif was also observed ∼450 m and ∼700 m north of LCHF (Fig. 1). Two orifices surrounded by bright white deposits were identified while transiting (Fig. 2). When the ROV passed over the orifices, the ORP sensor dropped by 4.3–8.3 mV in conjunction with a 0.09–0.33 °C temperature increase, indicating active flow of a fluid more reducing than seawater.

Extensive fluid flow was also observed along the eastern wall of the Atlantis Massif (Fig. 1; Figs. S4 and S5). The cliff face is the surface of a steep normal fault that cuts gabbroic and ultramafic rocks within the detachment fault zone (Kelley et al., 2005; Karson et al., 2006).

Extensive carbonate deposits on the cliff face extend >200 m laterally at a depth of 800–850 m (Kelley et al., 2005). Large, actively venting, carbonate structures also rise from the cliff face and fall along the east-west major lineament on which most chimneys occur across the LCHF (Fig. 1B) (Kelley et al., 2005). In 2018, we identified a previously undocumented, nearly continuous series of carbonate deposits at depths of 855–870 m and extending laterally over 30 m (Fig. 1; Figs. S4 and S5). Fluids vented directly from the wall with geochemical characteristics (T = 22 °C; pH_22C = 9.1; Mg = 36.6 mM) similar to LCHF fluids.

We carried out hydrocasts along the eastern wall to cover a broader geographical area. The casts captured fluids with elevated H₂ and CH₄ (18.7–40.8 nM and 5.4–10.6 nM, respectively) at 880–904 m depth, deeper than either identified contour of carbonate deposits or any known vents (Table S2; Figs. S6 and S7). Elevated H₂ concentrations (11.2–19.4 nM) were also detected extensively over the detachment fault termination boundary northeast of the LCHF (Fig. 1; Table S2).

The hydrocasts and seafloor surveys demonstrate that reducing fluids with elevated H₂ concentrations exit the seafloor over a wide geographical area and depth range. Because H₂ has a residence time of ∼10 h in the water column (Kadko et al., 1990), its flux must be sufficient to continually resupply near-bottom waters. Three potential sources of this H₂ are: (1) entrainment of LCHF fluids with millimolar H₂ concentrations dispersed into a water column plume; (2) subsurface migration of LCHF vent fluids to distal exit points; or (3) localized sources distinct from LCHF circulation pathways.

The H₂/CH₄ ratios of the fluids can distinguish among these possibilities. LCHF fluids have H₂/CH₄ ratios of 0.9–9.2, with the highest values occurring in the hottest fluids (Proskurowski et al., 2006, 2008) (Fig. 3; Figs. S8 and S9). Subsurface microbial processes lower this ratio through the consumption of H₂ and production of minor amounts of CH₄ (Lang et al., 2012; Proskurowski et al., 2006). Similar loss of H₂ and gain of CH₄ due to subseafloor microbial activity occur in a myriad of other hydrothermal systems (Von Damm and Lilley, 2004; Wankel et al., 2011). Export into the water column dilutes the fluids, with no impact on this ratio, and promotes the oxidation of both volatiles by microorganisms. Because H₂ is consumed 10[.dotmath] more rapidly than CH₄ in the oxic water column (Kadko et al., 1990), fluids exported from the LCHF have H₂/CH₄ ratios lower than 9.2 (Larson et al., 2015). Therefore, fluids with H₂/CH₄ ratios higher than LCHF fluids (>9.2) cannot be attributed to export from the plume or subsurface migration.

The fluids seeping from the eastern wall have H₂/CH₄ ratios similar to or lower than those of the LCHF (Fig. 3; Figs. S8 and S9) and are likely an extension of the circulation pathway supplying the field. The fluids collected over the detachment fault boundary have H₂/CH₄ ratios (7.0–9.2) indistinguishable from LCHF fluids. However, it is unlikely that they are sourced from the main circulation pathway of LCHF because they egress 7–11 km to the northeast.

The majority of fluids exiting across the top of the massif have H₂/CH₄ ratios greater than those from LCHF (>9.2), pointing to an independent source (Fig. 3; Figs. S8 and S9; Table S2). Most samples were collected deeper than 900 m, precluding input from the LCHF plume (Fig. 3; Table S2). The exception is the near-seabed fluids captured at the summit of the massif, at depths similar to LCHF chimneys (818 and 848 m; Früh-Green et al., 2017). The LCHF plume could have impacted these fluids, but the high H₂/CH₄ ratios (10–148) and evidence of fluid flow from nearby orifices are more consistent with largely local sources.

The most likely source of H₂ across the top of the Atlantis Massif is ongoing, widespread serpentinization (Fig. 4). Geophysical modeling suggests that in addition to vent fluids channeled along relatively narrow flow paths, slower-moving fluids circulate in “low-flux backwaters of the system” (Titarenko and McCaig, 2016, p. 325). These backwater fluids would not reach the elevated temperatures of those in the LCHF, which are heated to 150–250 °C before conductively cooling on ascent (Proskurowski et al., 2006; Seyfried et al., 2015). Rates of H₂ production due to serpentinization are substantially lower at lower temperatures, but are not entirely halted (Mayhew et al., 2013; McCollom et al., 2016).

In contrast to H₂, temperatures >350 °C appear to be necessary to overcome the kinetic inhibition to CH₄ formation (McCollom and Seewald, 2007). The isotopic signatures of CH₄ in LCHF fluids indicate that it formed abiotically at temperatures higher than those present in the modern system (Wang et al., 2018). Therefore, CH₄ likely formed early in the system when temperatures were hotter, was stored over millennia, and was stripped from the rocks into modern fluids (McDermott et al., 2015; Wang et al., 2018).

A model of extensive lower-temperature fluid flow across the top of the massif is also supported by alteration mineralogies, δ¹⁸O signatures, and rare earth element abundances in the Expedition 357 cores, which indicate that localized fluid pathways across the southern wall are interconnected to form independent 100-m- to 1-km-sized cells (Roumejon et al., 2018). Relict olivine is present in central and eastern boreholes (M0068, M0072, M0076; Früh-Green et al., 2017, 2018), where H₂ is most abundant (Tables S1 and S2), indicating early serpentinization did not go to completion and could be ongoing. The high permeability in this region results from microfractures and the mesh texture of serpentinized peridotite (Roumejon et al., 2018), common features in ultramafic rocks due to the volume expansion and uptake of water during serpentinization (MacDonald and Fyfe, 1985).

Serpentinization-fueled hydrothermal systems have been proposed as potential locations for the origin and evolution of the first biochemical pathways due to the thermodynamic drive supplied by H₂, the alkaline conditions that are favorable for prebiotic chemistry, and the redox and pH gradients that promote biochemical reactions (Martin and Russell, 2007; Sojo et al., 2016).

Proponents of a “small warm pond” for the origin of life emphasize the important roles of wet-dry cycles to promote polymerization (Frenkel-Pinter et al., 2020) and of clay minerals for forming and replicating biopolymers with specific sequences (Yu et al., 2013). Dispersed, low-flow circulation across the Atlantis Massif may create similar micro-environments of fluctuating water availability as water is consumed during serpentinization, followed by stages of higher water activity (Nascimento Vieira et al., 2020; Roumejon et al., 2018). Furthermore, serpentinization leads to the formation of serpentine and smectite clays (e.g., saponite) (MacDonald and Fyfe, 1985), which are similar to the montmorillonite clays most commonly invoked in origin-of-life hypotheses (Yu et al., 2013). Therefore, the widespread, localized flux of low-temperature, H₂-bearing fluids through micro-environments containing clay minerals and fluctuating water availability may have been more conducive to the synthesis of the first biomolecules than the channelized flow of higher-temperature fluids through the central field.

Problematically, ancient serpentinization systems may have been starved for sulfur. Metal-sulfide minerals are thought to have been central to the origin of biochemical pathways as the catalytic templates of the first enzymes (Camprubi et al., 2017; Martin and Russell, 2007; Weiss et al., 2017). Sulfur in modern ultramafic rocks is largely derived from seawater sulfate (Alt et al., 2013), and ancient prebiotic oceans had <1% of the inorganic sulfur content of modern oceans (Crowe et al., 2014). The peridotite-dominated southern wall, where nonchannelized fluid flow is dominant, contains ∼30% gabbro intrusions (Boschi et al., 2006; Früh-Green et al., 2018; Karson et al., 2006; Roumejon et al., 2018), and there is evidence that the sulfur in these mafic rocks is transferred into the peridotites (Liebmann et al., 2018). Therefore, the nonchannelized fluids may have entrained sufficient quantities of sulfur to make metal-sulfide catalysts available for primitive biochemical reactions in the serpentinite subsurface.

We documented the export of reducing, H₂-rich fluids over a large geographical area across the Atlantis Massif, likely due to continuing serpentinization of the high-permeability, peridotite-dominated southern wall. The elevated H₂/CH₄ ratios of these widely distributed fluids indicate that they have a localized source distinct from the central LCHF. The geological features leading to this type of extensive circulation are common in oceanic core complexes and therefore possibly widespread in the world's ocean. On ancient Earth, similar circulation of low-temperature, H₂-rich fluids through micro-environments within the ocean crust could have provided the fuel, sulfur, and fluctuating water availability necessary for the formation of early enzymes.

We thank the captains and crews of the RRS James Cook, R/V Atlantis, the British Geological Survey's Remote Drill Two (RD2), Maurum's Meeresboden-Bohrgerät (MeBO), and ROV Jason, the scientific parties of both expeditions, Daniel Fornari for the use of the MISO camera, and John Baross for inspirational conversations. Funding was provided by National Science Foundation awards OCE-1536702, OCE-1536405, and OCE-1535962, the National Aeronautics and Space Administration (NASA) Astro-biology Institute “Rock Powered Life” team (CAN 7), the U.S. Science Support Program of the International Ocean Discovery Program (IODP), European Consortium for Ocean Research Drilling (ECORD), the Deep Carbon Observatory, and Swiss National Science Foundation project 200021_163187 (to G. Früh-Green). This is Pacific Marine Environmental Laboratory (PMEL) contribution 5151. We thank editor Gerald Dickens and two anonymous reviewers for comments that improved the manuscript.