Hydrogen Chemistry of Basalt Aquifers

See allHide authors and affiliations

Science  18 Dec 1998:
Vol. 282, Issue 5397, pp. 2194
DOI: 10.1126/science.282.5397.2194e

In their report “Evidence against hydrogen-based microbial ecosystems in basalt aquifers” (14 Aug., p. 976), Robert T. Anderson et al. found that they could produce only very small quantities of hydrogen gas (H2) from chemical reaction of water and basalt rock (temperature unspecified), and then only under acidic conditions (pH = 6). In nature, however, reaction of water and mafic rocks commonly yields alkaline ground water, and can produce significant quantities of H2. The Semail ophiolite in Oman presents an extreme case (1); there, rainwater has reacted with rock rich in olivine and serpentine at 20° to 50°C to yield highly alkaline ground water (pH to 12.1) and gas seeps of up to 99% (molar) H2.

Thermochemically, the production of hydrogen gas from water can be driven by oxidation of ferrous to ferric iron (2). We have replicated this result in equilibrium computer modeling of basalt alteration (3); for instance, reaction of basalt glass and liquid water in a closed system (25°C; mass ratio of 1 to 5) yields a mineral assemblage dominated by serpentine and smectite clay, water of pH = 11.9 with 186 milligrams of dissolved H2 per kilogram, and a H2 fugacity of 135 bars.

References and Notes


Treiman and Wallendahl suggest, on the basis of a geochemical model, that when subsurface basalt and water are combined in a closed system, a significant amount of hydrogen gas should be produced. However, as we reported, when basalt and ground water are actually mixed together, hydrogen is not produced to any appreciable extent. Furthermore, their model predicts a pH of 11.9, whereas the pH in the real ground water (8.0) is much lower. We believe that the reason for these differences between reality and the predictions of the model are that the model contains faulty parameters. For example, Treiman and Wallendahl consider a basalt glass in their model, whereas the basalts in the aquifer material were crystalline. When it is necessary to chose between theoretical results from a model and direct experimental results, it seems prudent to base conclusions on the real data.

In contrast to the assertion of Treiman and Wallendahl, the study by Neal and Stanger did not demonstrate that the rocks in the Oman system produced hydrogen by reacting with water. This was only inferred from isotopic measurements and several assumptions about isotope equilibrium and exchange. Direct experimental measurements of rock-water interaction with materials from this site are required in order to prove the source of hydrogen in this system. It is also important to note that the study by Neal and Stanger has little relevance to our study because the rocks in the Oman system have a significantly different mineralogy from that of the basalt in the aquifer in question in our study. We specifically stated that our results do not rule out the possibility that hydrogen can be abiotically produced under some natural conditions. However, as detailed in our report, our studies and the results of others strongly suggest that hydrogen produced by basalt-ground-water interactions is not the primary energy source for the microbial community in the Columbia River Basalt aquifer.

Navigate This Article