worldly possessions packed into the back of my 1969 International Harvester pickup truck. After unloading my things at the graduate student dorm, Helen Hadley Hall, a boxy 1950’s Eastern-block-inspired construction, I parked my truck in the student garage1. My project was to study the rate of pyrite formation in
marine sediments, and eager to get started, I walked briskly across the parking lot to Kline Geology labs. Here I met my advisor, Bob Berner, who, after showing me around, introduced me to the fast-talking Rob Raiswell (on sabbatical at the time), who, with exaggerated arm movements, enthusiastically related the mysteries of concretion formation, and was convinced that we would find them in Long Island Sound sediments, if only we looked hard enough. I didn’t understand everything that Rob said, but it sounded important, and the next day I sailed with Bob Berner’s soon-to-finish student Joe Westrich onto Long Island to collect sediment cores at three different sites. Back in the lab, and for the next week, I carefully sieved these sediments and poked through the debris anticipating a big find. No concretions emerged, but I got my first look at pyrite framboids, and observed all kinds of different pyrite associations with plant material, diatoms and foraminifera. My interest in pyrite formation was aroused, especially the details of how sedimentary Fe species and sulphide interact. This led to one of my main PhD thesis topics and is explored in more detail below.
I never did return to explore rates of pyrite formation, but I did continue, together with Rob in many cases, to explore the geochemistry of Fe as well as the cycling of Fe through Earth history. As it turns out, Fe interacts with many other elements, and in ways that act to control the chemistry of the environment. In this section, I will look at how Fe interacts with the cycles of many other elements of biogeochemical interest. Many of these interactions are promoted by microbes, the real workhorses in driving biogeochemical cycles. These inter- actions help to inform us on the history of Fe cycling through time, which is the subject of Section 9.
1 . As I recall, the garage fee was about 1/3 of my take-home Yale stipend of 450 dollars a month . It was either beer or garage, and the obvious choice forced me park my truck back in Ohio for the next 2 years . When my salary rose to the point where I could afford the truck, I chased the mice out, started her up, and enjoyed a very memorable ride back to
8.2
Oxygen
As described in previous sections, the geochemistries of Fe and oxygen are inti- mately linked. In oxygenated environments, Fe(III) is the thermodynamically stable oxidation state. Aqueous Fe(III) (and some Fe(II) as well; see Section 2.2) in seawater is present as organic complexes, but, at circum-neutral pH, most Fe(III) is found hydrolysed into a variety of (oxyhydr)oxide phases. In the absence of oxygen, Fe(II) is much more stable and dissolved Fe(II) will often accumulate (although not in the presence of nitrate as we will explore below) until saturation with some mineral phase is reached.
At the interface between oxic and anoxic environments, reaction fronts are frequently established between Fe(II), in its aqueous and solid forms, and oxygen. We can envisage various types of reaction fronts. Perhaps the most obvious is the type established when strong gradients of oxygen meet opposing gradients in aqueous Fe(II) such as found, for example, in aquatic sediments, anoxic lakes, ground water springs, and hydrothermal fluids. Other reaction fronts are established between oxygen and solid Fe(II) phases in, for example, weathering environments where pyrite-rich shale is exposed to the atmosphere. We will consider each of these different types of reaction interfaces starting with the oxidation of aqueous Fe(II) with oxygen.
Reaction Between Fe(II) and Oxygen. Mix an Fe(II) solution into oxygen-
ated seawater, and Fe (oxyhydr)oxides are very rapidly formed. Indeed, if the water remains well aerated, within 10 minutes, or less, the Fe(II) will be completely removed. These kinetics are well described by the following rate law (Millero
et al., 1987; Singer and Stumm, 1970; Stumm and Lee, 1961): –d[Fe2+]/dt = k[Fe2+][O
2] [OH–]2 (8.1)
For a range of temperatures and ionic strengths, the rate constant k = logko – 3.29I0.5 + 1.52I, where logko = 21.56-1545/T, I is ionic strength and T is
temperature in degrees Kelvin (Millero et al., 1987). From this rate law, we can see that the oxidation rate accelerates with higher concentrations of Fe2+, higher
concentrations of O2 and at higher pH.
We also know from practical experience that in nature microbes are typi- cally associated with environments supporting aqueous Fe(II) oxidation. For example, iron (oxyhyr)oxide-encrusted sheaths of the Fe(II)-oxidiser Leptothrix
spp. or the helical (oxyhyr)oxide-encrusted stalks of Gallionella spp. are typically found populating areas where Fe(II) seeps into the oxygenated surface environ- ment (Emerson et al., 2010). The oxidation of Fe(II) with oxygen by microbes is clearly a thermodynamically favorable process; however, microbes in nature must compete with rapid inorganic oxidation. This is perhaps their biggest challenge, and experiments suggest that at circum-neutral pH, such as typically found in natural environments, they are only marginally successful at doing so. For example, the Fe(II)-oxidising groundwater organism Sideroxydans lithotrophicus did
not accelerate Fe(II) oxida- tion over inorganic rates in experiments at full oxygen saturation of 275 µM and a pH = 6.2 (Druschel et al., 2008). However, a micro- bial effect was observed at reduced oxygen concentra- tions of between 9 and 50
µM. The size of the effect seemed to scale inversely with oxygen concentra- tion (see Fig. 8.1) (Druschel
et al., 2008), where, with the exception of one appar- ently aberrant experiment (Fig. 8.1), microbes acceler- ated Fe(II) oxidation rates from 17 to 75%. Similar
extents of microbial contribution to Fe(II) oxidation were found in experiments with iron-oxidisers isolated from the rhizosphere of wetland soils (Neubauer et al., 2002), where a mixed population of microbial mat Fe oxidisers accelerated rates of Fe oxidation by from 40% to 300% (Rentz et al., 2007).
It makes sense that microbes contribute more to Fe(II) oxidation at lower oxygen concentrations. This is because the rates of the inorganic oxidation process decrease as oxygen concentrations fall (equation 8.1), whereas the enzymes conducting Fe(II) oxidation by microbes are apparently saturated with oxygen even at very low concentrations. To my knowledge, this has not been explored explicitly for Fe oxidisers, but it is true for other aerobic microbes where half- saturation values with respect to oxygen utilisation are in the micromolar to sub-micromolar range (e.g. (Longmuir, 1954; Stolper et al., 2010)). In any event, Fe oxidisers seem to preferentially populate the low-oxygen regions of gradient systems where they are active, presumably providing them with a strong kinetic advantage over inorganic Fe(II) oxidation (Druschel et al., 2008; Emerson et al., 2010; Emerson and Revsbech, 1994a,b). By contrast, at pH <5, the rates of inorganic oxidation are sluggish, and Fe-oxidisers can accelerate rates of Fe(II) oxidation by orders of magnitude over inorganic rates (Ferris, 2005; Singer and Stumm, 1970).
The products of microbial Fe oxidation depend very much on pH. At circum-neutral pH, ferrihydrite predominates initially (Supplementary Infor- mation SI-3 and Cornell and Schwertmann, 2003). At low pH, oxidation prod- ucts include ferrihydrite, goethite, and, if pyrite oxidation provides abundant sulphate, jarosite ([(H, Na, K)Fe3(OH)6(SO4)2] and schwertmannite predominate
(see Supplementary Information SI-6 and Ferris, 2005; Kupka et al., 2007). With time, schwertmannite converts to goethite/haematite (Cornell and Schwertmann, 2003; Davidson et al., 2008) or jarosite (Wang et al., 2006).
Figure 8.1 Ratio of the oxidation kinetics of Fe(II) in a strictly abiotic reaction and in the presence of Fe(II) oxidising-microbe Sid-
eroxydans lithotrophicus at a pH of 6.2.
Microbes accelerate the oxidation kinet- ics over abiotic rates at reduced oxygen concentrations (data from Druschel et al., 2008).
Pyrite Oxidation with Oxygen. Oxygen also interfaces with Fe through the weathering and oxidation of solid phases, of which iron sulphides, and in particular pyrite, are the most abundant and the best studied. This process was briefly discussed in Section 4.4, and it will be explored in more detail here. In considering the pathways of pyrite oxidation, pH is an important consideration. At low pH, the most efficient pathway of pyrite oxidation is only indirectly linked to oxygen. It seems, rather, that in natural settings, the main oxidation pathway for pyrite is by reaction with aqueous Fe(III) (e.g. Nordstrom and Southam, 1997): FeS2 + 14Fe3+ + 8H2O → 15Fe2+ + 2SO42– + 16H+ (8.2)
In principle, reaction 8.2 is an anoxic oxidation pathway. However, the Fe3+
must be made available, and this happens through oxidation with O2 as follows:
14Fe2+ + 3.5O2 + 14H+→ 14Fe3+ + 7H2O (8.3)
The overall reaction then becomes:
FeS2 + 3.5O2 + H2O → Fe2+ + 2SO42– + 2H+ (8.4)
Each of these reactions has possible abiotic and biotic pathways, and after a careful review of the available literature, Nordstrom and Southam (1997) came to the following conclusions:
(a) Reaction 8.4 may represent the sum of two separate pathways (reac- tions 8.2 and 8.3) or a direct oxidation pathway with oxygen. Available experiments suggest, however, that the direct pathway is at least one order of magnitude slower at acid pHs.
(b) Microbes often sit directly on the pyrite surface during oxidation. Whether they employ reaction 8.4 or the sum of reactions 8.2 and 8.3 is unclear. However, the overall reaction rate is faster if microbes are separated from the surface and oxidation proceeds through reaction 8.2. This relationship could imply that the microbes hinder access of Fe3+ to
the pyrite surface, but it is more likely that microbes conduct reaction 8.3 faster than they can oxidise pyrite through direct contact.
(c) In the presence of microbes, reaction 8.2, an inorganic reaction, seems to be the rate-limiting step. However, since reaction 8.3 is so dramati- cally enhanced through microbial oxidation, the overall process of pyrite oxidation under acidic conditions is considerably enhanced through microbial intervention.
At high pH, ferric iron is nearly insoluble (forming Fe(oxyhydr)oxide phases; see Section 2.2 and Cornell and Schwertmann, 2003), and the pathway of pyrite oxidation represented by equations 8.2 and 8.3 is of less importance. Indeed, the role of microbes in promoting pyrite oxidation at circum-neutral pH in nature is not clear.
Pyrite, however, does oxidise with oxygen in the absence of microbes, and the inorganic kinetics of pyrite oxidation have been well explored through a range of pH (e.g. (McKibben and Barnes, 1986; Nicholson et al., 1988; Smith and
Shumate, 1970; Williamson and Rimstidt, 1994). After compiling the available data, and in light of their own experiments, Williamson and Rimstidt (1994) derived a rate law that applies to the kinetics of oxidation over a broad range of pH (2-10) and dissolved oxygen concentrations (approximately 1 µM to 19 mM; where the high levels are well beyond air-saturation values and represent experi- ments conducted at high pressures of up to 25 bars; Smith and Shumate, 1970). The rate of pyrite oxidation is given in equation 8.5. We saw this rate law previ-We saw this rate law previ- ously in Section 4.4 (equation 4.1) in relationship to understanding the survival time of fine-grained pyrite, apparently derived from continental shelf sediments, during transport in oxygenated marine waters.
Rate of pyrite oxidation (mol m–2 s–1) = 10–8.19 (O
2) 0.5/ {H+}0.11 (8.5)
The data from Williamson and Rimstidt (1994) have been recalculated to a common pH of 6 [RatepH6= Ratemeasured*(MH+measured)0.11/(10–6)0.11] and compared
to oxygen concentration in Figure 8.2. The rate predicted from equation 8.5 at this same pH is also shown.
Figure 8.2 Rates of pyrite oxidation as a function of oxygen with rates normalised to pH = 6. Also shown is the predicted oxidation rates from equation 8.5 (data from compilation in Williamson and Rimstidt, 1994).
A few key points are obvious from this comparison. First, as discussed by Williamson and Rimstidt, (1994), the rate law provides a reasonably good prediction of inorganic pyrite oxidation rate over a range of several orders of magnitude in oxygen concentration. Second, most of the available oxidation rate data is from experiments with oxygen concentrations elevated over those found
rates under nM (or lower) oxygen levels (Anbar et al., 2007; Canfield et al., 2000), as was believed the case in surface waters before the general oxidation of the Earth’s atmosphere around 2.3 to 2.4 billion years ago (Farquhar et al., 2000; Pavlov and Kasting, 2002; Stolper et al., 2010). Such oxygen levels are orders of magnitude lower than the lowest used in experiments to date. So, the question becomes, does the rate law in equation 8.5 still apply at such low levels of O2? It
will be a high priority to explore these kinetics at much lower oxygen levels as relevant for assessing pyrite oxidation on the early Earth and sulphate fluxes to the early Earth ocean.