Methane generation during experimental serpentinization of olivine.

Oze et al. (1) observed generation of methane (CH4) during experimental serpentinization of olivine, and use the results to make inferences about sources of methane in natural serpentinites. Their data interpretation, however, is directly contradicted by results of prior studies, critically undermining the credibility of their conclusions. Several prior studies conducted using experimental methods essentially identical to those of Oze et al. have demonstrated that natural minerals and other sources contribute significant amounts of background methane during laboratory experiments (e.g., refs. 2–4). In particular, experiments using C-labeled reactants to track carbon sources have shown that background sources generate methane at μmol·kg levels, comparable to the concentrations reported by Oze et al. (1). Furthermore, the labeled experiments show that only a small fraction of the methane produced in the experiments occurs as CH4, demonstrating that the vast majority of the methane comes from carbon already present in reduced form at the onset of the experiments. In one experiment, for example, serpentinization of olivine at 300 °C in the presence of dissolved HCO3 − yielded <2 μmol·kg of CH4, whereas most of the methane generated in the experiment was unlabeled (∼11 μmol·kg CH4). Oze et al. (1) reported no effort to document the source of carbon for the methane observed in their experiments, nor do they report adequate control experiments to assess background methane levels. Indeed, as they did not list any carbon source at all among the reactants, the carbon contained in methane must have been derived from background sources, most likely some form of reduced carbon present in the reactants. There are other reasons to be skeptical of their interpretations. Although the authors infer that methane in their experiments was derived from reduction of inorganic carbon catalyzed by magnetite formed during serpentinization (1), the rate of methane generation in their experiments was comparable to that reported in previous studies (2–4) even though those studies were performed at higher temperatures, contained much higher concentrations of reactants (H2 and CO2), and included substantially larger amounts of magnetite. In addition, in a closely related study by the same authors (5), methane yields were actually lower when an inorganic carbon source in the form of bicarbonate was added to the reactants, opposite to what would be expected for formation of methane by catalyzed carbon reduction. [It is perhaps worth noting that experiment 2 of Oze et al. (1) appears to be identical to experiment 1 reported by Jones et al. (5), although this is never mentioned by Oze et al.] The results of prior studies indicate that most, if not all, of the methane reported by Oze et al. (1) [as well as in their previous study (5)] is derived from background sources already present in the reactants, rather than being derived from reduction of carbon during the experiments. Consequently, conclusions reached by Oze et al. concerning H2/CH4 ratios in natural serpentinites must be viewed with extreme skepticism. Thomas McCollom Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, CO 80309