2. Experiments
The experimental methods employed follow those used previously to study abiotic methane and other organic syntheses by Fischer-Tropsch-type (FTT) reactions under hydrothermal conditions (McCollom et al., 2010; McCollom and Seewald, 2006). Briefly, a flexible-cell reaction cell composed of a gold bag with titanium fittings is contained inside a stainless-steel pressure housing (Seyfried Jr et al., 1987). An external valve connected to the reaction cell with titanium tubing allowed for collection of volatile gases at the end of the experiments. Prior to the experiments, the reaction cell components were heated for two hours at 450°C in air to remove organic contaminants and to create an inert TiO2 layer on the surface of the titanium fittings. Previous experiments with the same reaction system have demonstrated that the reactor itself is not catalytic for organic synthesis under the experimental conditions (McCollom and Seewald, 2006; McCollom et al., 2010).
The experiments were performed by heating carbon monoxide (CO) and native iron (Fe0) in the presence of water at a range of temperatures. The experiments were initiated by placing 1.5 g of Fe0 (Fe powder, -22 mesh; Alfa Aesar) and ~25 g of DI water in a 55 mL gold reaction cell, so the H2O/Fe molar ratio was ~ 50. The native Fe is included both as catalyst and as a source of molecular hydrogen (H2) since it rapidly reacts with water at elevated temperatures to form magnetite (Fe3O4) and H2 according to the reaction:
3Fe0 + 4H2O → Fe3O4 + 4H2. (2)
To promote reaction of Fe0, the water was acidified to a pH of about 2 with HCl. Individual experiments were conducted over a range of temperatures from 130°C to 300°C. The target pressure for the experiments was 15 MPa, but in practice the pressures varied up to 3 MPa from this value during some experiments.
After sealing the reaction cell with the Ti closure piece, it was placed in the pressure containment vessel, pressurized at room temperature, and the entire apparatus was placed horizontally and heated in a furnace. Once the target temperature was attained, CO was injected through the sampling valve using the following procedure. First, a loop of stainless-steel tubing (12.6 mL) was filled with CO at a pressure of 0.93 MPa (the maximum of the pressure regulator on the CO tank) and attached to the sampling valve. A solution with the same composition as that in the reaction cell was then pumped into the tubing until pressure was equal to that within the heated pressure vessel. The valve was then opened and the CO along with 13 mL of solution was pumped into the reaction cell, with an equal amount of water bled from the pressure containment vessel to maintain constant pressure during injection. In practice, some amount of CO remained in the injection loop after this procedure, so it is not possible to determine exactly how much CO was contained in the reaction cell following injection. However, assuming 90% injection efficiency, the amount of CO injected would have been about 4.3 millimoles. Considering that 38 mL of water (25 mL initial + 13 mL added during injection) corresponds to ~2.1 moles H2O, the starting CO/H2O molar ratios of the experiments are estimated to be approximately 2 ⨯ 10-3.
The tank of CO used in the experiments was the same as that used in McCollom et al. (2010) and has a δ13C of -28‰ V-PDB. The water used in this study is deionized water from Univ. Colorado tap water. The δD of DI water sampled in April 2022 and April 2023 was –119‰ V-SMOW, similar to measurements of Boulder tap water measured in older studies (–116±1‰; (Landwehr et al., 2014)). The molecular hydrogen (H2) generated from reaction 2 subsequently reacts with CO to form methane, according to the net reaction:
CO+3H2=CH4+H2O (3)
Reaction 3 is a Fischer-Tropsch reaction that utilizes CO for oxidized carbon rather than CO2. It is distinct from the Sabatier reaction (reaction 1) that may be similar to reactions in natural hydrothermal systems where CO2 is available in large amounts. Carbon monoxide was used here because it is more active in hydrothermal FTT reactions than CO2 (McCollom et al., 2010), which allowed for obtaining measurable CH4quantities under our experimental timescales. In addition to CO reduction, the so-called “water-gas shift reaction” (McCollom and Seewald, 2006) was prominent under the experimental conditions; reaction between CO and water led to the oxidation of a fraction of the CO to form CO2:
CO + H2O = CO2 + H2 (4)
Through a competition between reaction 3 and 4, CO is rapidly and predominantly converted to CH4 and CO2, with much of the CO2 precipitating as siderite (FeCO3). Reaction 4 is anticipated to be dominant over reaction 3 (McCollom et al., 2010).
The experiments were heated for 18 to 138 hours, with longer incubation times used at lower temperatures to allow for the synthesis reactions to proceed. Following reaction, the reactors were gradually cooled to room temperature while pumping water into the containment vessel to maintain elevated pressure. The pressure was then reduced to ~5 MPa by releasing water from the pressure containment vessel in order to allow volatile gases to exsolve and accumulate at the top of the reaction cell. Within 24 h of experiment termination, the gas phase was vented through the sampling valve and transferred to 300 mL stainless steel cylinders (Swagelok) for isotope analysis.