Oxidation of water to hydrogen peroxide at the rock–water interface due to stress-activated electric currents in rocks

https://doi.org/10.1016/j.epsl.2009.03.044Get rights and content

Abstract

Common igneous and high-grade metamorphic rocks contain dormant defects, which release electronic charge carriers when stressed. Rocks thereby behave like a battery. The charge carriers of interest are defect electrons h, e.g. electronic states associated with O in a matrix of O2−. Known as “positive holes” or pholes for short, the h travel along stress gradients over distances on the order of meters in the laboratory and kilometers in the field. At rock–water interfaces the h turn into •O radicals, e.g. highly reactive oxygen species, which oxidize H2O to H2O2. For every two h charge carriers one H2O2 molecule is formed. In the laboratory the battery circuit is closed by running a Cu wire from the stressed to the unstressed rock. In the field closure of the circuit may be provided through the electrolytical conductivity of water. The discovery of h charge carriers, their stress-activation, and their effect on Earth's surface environment may help better understand the oxidation of the early Earth and the evolution of early life.

Introduction

Though most oxygen anions in oxide/silicate minerals occur in their common 2− valence state, some can exist in the more oxidized valence 1−, as O. In this paper we present evidence that oxygen anions in the valence state 1− in minerals, which make up the bulk of igneous and high-grade metamorphic rocks, give rise to some unexpected phenomena. The O are normally dormant, forming inconspicuous and electrically inactive point defects. However, they can be “awakened” when rocks are subjected to mechanical stress. Once activated, the electronic state associated with O can produce an electric current, which was previously unknown.

First we discuss how O are introduced into the mineral matrix, why they exist in a dormant form, and how they are become activated when mechanical stresses are applied. Second, we demonstrate that stress-activated electronic charge carriers can flow out of the stressed rock volume and propagate through unstressed rock. When they cross a rock–water interface, they oxidize H2O to H2O2.

It has been shown some time ago that hydroxyl pairs in MgO, introduced into the MgO matrix during crystallization from an H2O-laden melt or recrystallization in an H2O-laden environment, undergo a previously unknown redox reaction: they split off H2 and form peroxy anions, O22− (Martens et al., 1976, Freund and Wengeler, 1982):OHHO ⬄ H2 + ⊕O22−where ⊕ stands for an Mg2+ vacancy site. In the peroxy anion, O22−, the two O are tightly bonded together, forming a self-trapped, localized, electrically inactive point defect.

Evidence suggests that the same type of redox reaction also occurs in fused silica, where O3SiOH pairs are converted into H2 plus peroxy links O3Sisingle bondOOsingle bondSiO3 (Freund and Masuda, 1991, Ricci et al., 2001). Likewise nominally anhydrous silicate minerals (Freund, 1987, Freund, 2003) are known to always dissolve small amounts of H2O when crystallizing in H2O-laden magmas or H2O-laden high temperature environments (Wilkins and Sabine, 1973, Rossman, 1996, Koga et al., 2003). The solute H2O occurs in the form of O3(X,Y)OH, where X, Y = Si4+, Al3+ etc. Hydroxyl pairs are probably the most abundant, but they seem to undergo the same redox conversion, splitting off H2 and forming peroxy links (Freund, 2003):O3Xsingle bondOH HOsingle bondYO3 ⬄ O3X/OO\YO3 + H2

Because crustal rocks consist mostly of nominally anhydrous minerals, which invariably contain small amounts of dissolved H2O, peroxy links may be very ubiquitous.

From a semiconductor viewpoint, an O in a matrix of O2− represents a defect electron or hole, also known as positive hole (Griscom, 1990) or phole for short, symbolized by h. Accordingly, a peroxy link represents a positive hole pair, PHP (Freund et al., 2006), in which two O are self-trapped, immobilized and, hence, electrically inactive. PHPs break apart when rocks are subjected to deviatoric stresses, e.g. to stresses that exceed the elastic limit causing dislocations to be mobilized and new dislocations to be generated (Freund et al., 2006, Takeuchi et al., 2006). As dislocations move through the mineral grains, they intersect a PHP causing the peroxy bond to break:peroxy link + O2− ⬄ loosely bound e' + O = h O3X/OO\YO3 + [XO4]n ⬄ O3X/OO/YO3 + [XO4](n  1)−

Eq. (3) describes a process, by which an electron e' is transferred from a neighboring O2− anion (here represented by [XO4]n) onto the broken peroxy bond, which captures the electron e'. The donor O2− turns into O (here [XO4](n  1)−), e.g. a defect electron or phole h. In other words, Eq. (3) describes the generation of an electron–hole pair.

In semiconductors like silicon electron–hole pairs recombine very rapidly, often within nanoseconds (Sze, 1981, Van Zeghbroeck, 1997). In the case of the e'–h pairs generated by the break-up of peroxy links in silicate matrices, the broken peroxy bond that has captured electron appears to relax in such a way as to delay recombination. Thus, e'–h pairs formed in silicates and, hence, in rocks are long-lived (Freund et al., 2006, Freund and Sornette, 2007).

The h charge carriers in the silicates are associated with O 2sp energy levels at the upper edge of their valence bands (Canney et al., 1999). When mineral grains are in physical contact, the valence bands are electrically connected. Though their energies will shift from grain to grain, all valence bands form an energy continuum along which h charge carriers can propagate. The mode of propagation probably involves a phonon-assisted electron transfer, whereby h hop from O2− to O2−, maximally at the frequency of thermally activated lattice phonons, ~ 1012 Hz (Shluger et al., 1992). The O turn insulating silicate minerals into p-type semiconductors. From a chemistry perspective, an O is a highly oxidizing •O radical.

Hurovitz et al. described experiments, where freshly ground basaltic minerals immersed in aqueous solution produced detectable amounts of H2O2 (Hurowitz et al., 2007). The H2O2 production is thought to occur when water reacts with unspecified defects formed at the mineral surfaces during crushing. Possibly h charge carriers, activated by the high mechanical stresses levels during crushing, contribute to the oxidizing properties of freshly crushed rock surfaces.

In the experiments as described here, where only a subvolume of a large piece of rock is subjected to increased levels of mechanical stress, h charge carriers flow from the stressed into the unstressed rock volume. The stressed rock volume acts as the anode in a battery from where an electric current flows out (Freund et al., 2006). The unstressed rock acts as the electrolyte, through which the h flow to reach the Cu contact at the unstressed end of the rock as depicted in Fig. 1a. This Cu contact acts as the cathode.

The electrons e', co-activated according to Eq. (3), cannot spread from the stressed rock into the unstressed rock. However, they can pass from the stressed rock into a metal contact, in our case the steel pistons, and then proceed through a Cu wire to reach the same Cu contact attached to the far end of the unstressed rock. The Cu wire thereby closes the battery circuit. At the interface between the Cu contact and the unstressed rock the e' “shake hands” with h and recombine. Knowing that h = O, we can write:O + e ==> O2−

In the course of this study we addressed three questions:

  • (i)

    Can the h charge carriers flow not only through solid rock as depicted in Fig. 1a but also through water as depicted in Fig. 1b?

  • (ii)

    What happens at the rock–water interface?

  • (iii)

    How can the rock battery circuit be closed, if no wire is available to connect the stressed rock volume with the unstressed rock volume?

Section snippets

Experimental

We used a gabbro from Shanxi, China, ~ 40 modal% plagioclase, 30% augitic clinopyroxene with alteration rims to amphibole and chlorite, plus 25% opaques, a porosity of about 0.3%, and less than 1% total water mostly due to hydroxyl-bearing minerals such as amphiboles (Parkhomenko, 1967).

Fig. 2 depicts a block of this gabbro, 60 × 7.5 × 7.5 cm3, loaded at the center through two rectangular steel pistons with a contact area of 5 × 10 cm2 covered by Cu tape with graphite-based conductive adhesive (3M).

Results

Fig. 3 plots the battery current I flowing into and through Pool 1 integrated over time. After 24 h, the total charge C passing through the rock–water interface was 0.5 mC, ≈ 2 × 1015 h charge carriers. C as a function of time (t) is given by:C(t)=t=0tI(t)dt.

When the h current flows into the water, interface reactions have to be considered. Being chemically equivalent to O, the h charge carriers are expected to act as •O radicals at the rock–water interface. One possible reaction for the •O

Discussion

Electric potentials and ground currents are common in nature. Self-potentials can be generated in conductive materials through transfer of charges. Known sources of self-potentials are streaming currents driven by the flow of water through porous rocks, thermoelectric currents associated with temperature gradients, and electrochemical currents associated with a transfer of electrons during redox reactions or with the diffusion of ions (Cox and Singer, 1986, Bernabé, 1998, Revil and Linde, 2006,

Conclusion

By studying the flow of stress-activated battery currents through a rock–water system we simulate oxidation processes that take place in the crust of a tectonically active body like Earth and that may have taken place throughout Earth's history. Electrochemical oxidation driven by stress-activated battery currents at rock–water interfaces may have played an important role in the evolution of the early Earth, specifically in the global oxidation, and in the evolution of early life.

Acknowledgments

This work was supported in part by a grant from the NASA Astrobiology Institute (NAI) Cooperative Agreement NNA04CC05A (to F.T.F and L.J.R.). M. Balk acknowledges a travel grant from the Netherlands Organization for Scientific Research (NWO). G.E. acknowledges a NASA Senior Fellowship through the National Research Council and the Oak Ridge Associated Universities (ORAU).

References (43)

  • RevilA.

    Thermodynamics of transport of ions and water in charged and deformable porous media

    J. Colloid Interface Sci.

    (2007)
  • TakeuchiA. et al.

    Current and surface potential induced by stress-activated positive holes in igneous rocks, Phys

    Chem. Earth

    (2006)
  • BatlloF. et al.

    Dissociation of O22− defects into paramagnetic O− in wide band gap insulators: a magnetic susceptibility study of magnesium oxide.

    J. Appl. Phys.

    (1990)
  • BernabéY.

    Streaming potential in heterogenous networks

    J. Geophys. Res.

    (1998)
  • BrinesL.M. et al.

    Understanding the mechanism of superoxide reductase promoted reduction of superoxide

    Europ. J. Inorg. Chem.

    (2007)
  • CanneyS.A. et al.

    Electronic band structure of magnesium and magnesium oxide: experiment and theory

    J. Phys. Condens. Matter

    (1999)
  • CoxD.P. et al.

    Mineral deposit models

    US Geolog. Survey Bull.

    (1986)
  • CrespyA. et al.

    Detection and localization of hydromechanical disturbances in a sandbox using the self-potential method

    J. Geophys. Res.

    (2008)
  • DismukesG.C. et al.

    The origin of atmospheric oxygen on Earth: the innovation of oxygenic photosynthesis

    Proc Natl. Acad. Sci. USA.

    (2001)
  • EigenbrodeJ.L. et al.

    Late Archaen rise of aerobic microbial ecosystems

    Proc. Natl. Acad. Sci. USA

    (2006)
  • Espinosa-GarciaJ.

    New theoretical value of the enthalpy of formation of the OOH radical

    Mol. Phys.

    (1993)
  • Cited by (0)

    1

    Present address: NIOO-KNAW Centre for Limnology, Rijksstraatweg 6, 3631 AC Nieuwersluis, The Netherlands.

    2

    Present address: Department of Physics and Astronomy, University of California, Riverside, CA 92521, USA.

    3

    Present address: Karlsruhe Forschungszentrum, Weberstr. 5 D-76133 Karlsruhe, Germany.

    View full text