Introduction
The deep biosphere, found at hydrothermal vent areas on the ocean floors, has been recognized to consist mainly of chemoautotrophic microorganisms (Staudigel et al. Reference Staudigel, Tebo, Yayanos, Furnes, Kelley, Plank, Muehlenbachs, Wilcock, DeLong, Kelley, Baross and Cary2004; Edwards et al. Reference Edwards, Bach and McCollom2005). Such organisms use chemical energy stored in minerals or in compounds dissolved in the hydrothermal fluids as an energy source. The most accessible elements and compounds for such organisms in hydrothermal settings are H2, H2S, S, CO, CH4, Mn2+ and Fe2+, and electron acceptors could be CO2, SO42− or O2. At many hydrothermal vent areas iron oxidizing microorganisms appear to play an important role (Boyd & Scott Reference Boyd and Scott2001; Emerson & Moyer Reference Emerson and Moyer2002; Edwards et al. Reference Edwards, Rogers, Wirsen and McCollom2003, Reference Edwards, Bach, McCollom and Rogers2004). Iron oxidizing microorganisms mediate the oxidation of dissolved Fe(II) to ferric oxide/hydroxides Fe(III). Such microorganisms occupy habitats at the interface between two environments with divergent redox chemistry. In contact with oxidized seawater, Fe(II) is out of chemical equilibrium, and redox reactions exploitable by microorganisms can occur. The reactions are relatively slow, and the microorganisms can catalyse the reactions and gain metabolic energy in the process.
Recent observations of encrusted and fossilized microorganisms in drilled samples from sub-seafloor environments indicate that iron is probably the dominating energy source for microbes in such environments as well (Thorseth et al. Reference Thorseth, Torsvik, Furnes and Muehlenbachs1995, Reference Thorseth, Torsvik, Torsvik, Daae and Pedersen2001, Reference Thorseth, Pedersen and Christie2003; Ivarsson et al. Reference Ivarsson, Lindblom, Broman and Holm2008). High contents of iron, iron encrusted cells and association with minerals and materials that contain reduced iron suggest that iron is a crucial element in sub-seafloor environments. Unweathered basalt minerals such as olivine and pyroxene contain high amounts of reduced iron that microbes can scavenge. Volcanic glass is another material that microorganisms have been observed attached to and boring into, and which contain high amounts of reduced iron. So-called tunnel structures or galleries that are a result of microbial activity have during the last ten years been recognized as a widely spread phenomenon in the world's ocean floors, both in modern and in ancient ones (Giovannoni Reference Giovannoni, Fisk, Mullins and Furnes1996; Furnes et al. Reference Furnes, Thorseth, Tumyr, Torsvik and Fisk1996; Fisk et al. Reference Fisk, Giovannoni and Thorseth1998; Furnes & Staudigel Reference Furnes and Staudigel1999; Furnes et al. Reference Furnes, Muehlenbachs, Tumyr, Torsvik and Xenophontos2001; Fisk et al. Reference Fisk, Storrie-Lombardi, Douglas, Popa, McDonald and Di Meo-Savoie2003; Furnes et al. Reference Furnes, Banerjee, Muehlenbachs, Staudigel and de Wit2004; Staudigel et al. Reference Staudigel, Tebo, Yayanos, Furnes, Kelley, Plank, Muehlenbachs, Wilcock, DeLong, Kelley, Baross and Cary2004).
During the last decade an increased amount of research regarding the role of iron in biogeochemical cycles has been carried out, generally with a focus on iron isotopes (e.g. Beard & Johnson Reference Beard, Johnson, Johnson, Beard and Albarède2004; Johnson et al. Reference Johnson, Beard, Roden, Newman, Nealson, Johnson, Beard and Albarède2004; Johnson & Beard Reference Johnson and Beard2006). The possibility of biogeochemical fractionation and the use of iron isotopes as ‘biosignatures’ have been tantalizing both for research connected to environments on our planet as well as a potential way of distinguishing between biological and abiological features in extraterrestrial samples (Anbar Reference Anbar2004; Anand et al. Reference Anand, Russell, Blackhurst and Grady2006). Although there is no doubt that metabolic processing of Fe produces isotopic fractionation (Beard et al. Reference Beard, Johnson, Cox, Sun, Nealson and Aguilar1999, Reference Beard, Johnson, Skulan, Nealson, Cox and Sun2003; Croal et al. Reference Croal, Johnson, Beard and Newman2004; Johnson et al. Reference Johnson, Beard, Roden, Newman, Nealson, Johnson, Beard and Albarède2004), similar Fe isotope fractionations may be produced by abiologic processes (Anbar et al. Reference Anbar, Roe, Barling and Nealson2000; Bullen et al. Reference Bullen, White, Childs, Vivit and Schultz2001; Johnson et al. Reference Johnson, Skulan, Beard, Sun, Nealson and Braterman2002; Roe et al. Reference Roe, Anbar and Barling2003; Welch et al. Reference Welch, Beard, Johnson and Braterman2003; Icopini et al. Reference Icopini, Anbar, Ruebush, Tien and Brantley2004). Thus, it has been questioned whether or not a biosignature based on iron isotopic fractionation can be preserved in the geologic record (Balci et al. Reference Balci, Bullen, Witte-Lien, Shanks, Motelica and Mandernack2006).
In this study we use laser-ablation inductively coupled plasma mass spectroscopy (ICP-MS) to study variations in iron isotopes on a micro-scale. We have studied carbonaceous filamentous structures previously described by Ivarsson (Reference Ivarsson2006) and Ivarsson et al. (Reference Ivarsson, Lindblom, Broman and Holm2008) and which have been interpreted as fossilized microorganisms. We have also studied non-carbonaceous filamentous structures that we interpret to have formed abiotically, comparing their isotopical composition to the isotopical composition of the microfossils.
The samples used in this study were collected during the Ocean Drilling Program (ODP) Leg 197 at the Emperor Seamounts in the Pacific Ocean. During this leg, three seamounts belonging to the Emperor Seamounts were drilled: Detroit, Nintoku and Koko Seamounts. The collected samples consist of drilled ocean floor basalts from all three seamounts. These basalts are fractured and veined to varying degrees depending on where in the volcanic pile they were collected. The veins are filled with secondary, hydrothermally formed minerals such as calcite, aragonite and gypsum (Tarduno et al. Reference Tarduno, Duncan and Scholl2002). The filamentous structures studied in this work are found in the veins where they are attached to the vein walls and entombed in the vein filling minerals.
Methods
The samples were prepared as doubly polished thin sections (~150–200 μm in thickness) due to their advantages in studying microfossils in volcanic rocks (Ivarsson Reference Ivarsson2006).
Filamentous structures and their association with iron oxide phases were first studied with light optical microscopy, environmental scanning electron microscopy (ESEM) and energy dispersive spectrometry (EDS). The ESEM analyses were performed using a Philips XL 30 ESEM-FEG which is a field emission microscope. EDS analyses were performed using a Philips EDAX (energy dispersive analysis of x-rays) instrument. The samples were subjected to a pressure of 0.5 torr and the accelerating voltage was 20 kV. The EDS analyses were performed by standardless quantification. The point detection of the EDS instrument is in the size range of 1 micrometer (μm). Analyses of microfossils with the EDS is thus very precise and the measurements are not mixed with the surrounding minerals.
As a next step they were analysed with laser-ablation ICP-MS. The laser that was used was a New Wave UP 213 that was connected to a Thermo Elemental X Series ICP-MS. The carrier gas in the laser was argon. Different sizes of the laser beam were used (5–60 μm) depending on the size of the filamentous structures. In general, a thicker beam size gives better results. The filaments do not vary much in thickness (~2–10 μm) but do so in length (~10–200 μm). Variations in thickness are usually due to different thicknesses of the deposited iron oxides onto the filaments. The filaments themselves are usually only a couple of micrometres wide. Short filaments are usually more difficult to analyse with a thick beam size compared to long filaments, and thus long filaments were mostly selected for the laser-ablation analyses. Since laser ablation is a destructive method long filaments were also selected because parts of them are still left after the analyses and can be used as a reference material.
Each analysis or run results in a profile through the analysed material. This profile shows a curve for each analysed element or isotope that reflects the occurrence of the element versus time. Depending on which material is analysed, the curves vary with respect to each other. In most profiles certain elements or isotopes are selected over other elements to make the interpretation of the profiles easier. An element like Ca, for example, is usually excluded since it dominates in the calcite and would only make the profiles more difficult to decipher.
The destructiveness of this method means that each run needs to be carefully targeted (beam size, path and length of the run, surrounding phases) and each element or isotope required for the analysis needs to be selected in advance. Only one opportunity is given and the material is relatively limited.
Results
Filamentous structures
The filamentous structures found in the samples were divided into two groups according to chemical composition: carbonaceous filamentous structures and non-carbonaceous filamentous structures (Fig. 1). Carbon-rich filaments from these samples have previously been described by Ivarsson et al. (Reference Ivarsson, Lindblom, Broman and Holm2008). The following is a summary of the descriptions of the carbon-rich filaments and the non-carbonaceous filaments.
Microphotograph of carbonaceous filamentous structures in A and B. Non-carbonaceous filamentous structures in C and D. Scale bars in A and D: 10 μm, in B and C: 5 μm.
Carbonaceous filamentous structures
The carbon-rich filamentous structures are found attached to the vein walls, which consist of volcanic glass or altered basalt, and are embedded in the vein-filling carbonates. Some filament structures are attached with both ends to the vein walls. Usually, the filament structures have a ‘worm-like’, smoothly curved appearance and are ~10–100 μm long and ~2–10 μm thick.
The filaments were divided into five groups according to morphology: sheaths, segmented filaments, twisted filaments, amorphous filaments and branched filaments (Fig. 2). Sheaths are made up of hollow, tube-like filamentous structures that resemble one thinner tube inserted into a thicker one. This type is typically attached to the vein walls at both ends as opposed to the other classes which are attached only at one end. This is also the most common kind. Segmented filaments resemble the previous type but lack the consistent tube shape. Instead, they consist of individual segments making up a long filament. Amorphous filaments are microstructures with no visual inner texture. The only characteristic feature is a thicker, spherical end at the top of the filament like a bean on a stalk. Twisted filaments are the most uncommon type. They were only seen to occur in one single sample. They resemble the two first types in shape but lack the tube-like outline as well as the segments. Instead, the filaments are twisted in a screw-like fashion. Branched filaments consist of a group of filamentous structures that branch off from each other creating a complicated network. This type is abundant where they occur and occupy much space. In some samples they fill a whole vug or parts of veins. All filament types show a similar range of thicknesses (2–10 μm) except the amorphous filaments that never exceed 2 μm in thickness.
Microphotographs obtained with transmission light microscope showing the different types of fossilized filaments according to morphology. A: sheaths (1206A-18R-1, 104), B: segmented filament (1204B-12R-1, 24), C: amorphous filament (1205A-6R-2, 53), D: twisted filament (1203A-30R-1, 63), E: branched filaments (1203A-57R-3, 40). Scale bars: 5 μm.
EDS analyses showed that the carbon-rich filaments have an iron content of ~10–50 wt% Fe. The filaments are also closely associated with globular iron oxides both in their close vicinity and deposited onto the surfaces of the filaments. EDS analyses of the iron oxides found in the close vicinity of the filaments show iron contents up to 80 wt% but no carbon content. The iron oxides observed on the fossilized filaments have iron contents similar to the iron contents of the filaments.
The carbon-rich filamentous structures have a carbon content that ranges between ~10 and 50 wt% C (Table 1). Time-of-flight secondary ion mass spectrometry (ToF-SIMS) analyses showed a concentration of C2H4, phosphate and lipids to the filamentous structures. Staining with the pigment PI (propidium iodide), which is a dye that binds to dead bacteria cells and remnants of DNA, showed that PI binds to parts of the filaments (Ivarsson et al. Reference Ivarsson, Lindblom, Broman and Holm2008).
Chemical compositions of carbonaceous filamentous structures
Non-carbonaceous filamentous structures
Non-carbonaceous filamentous structures resemble the carbonaceous filamentous structures morphologically to some extent. They are attached to the vein walls and embedded in carbonates just like the carbonaceous filaments. They are in the same size range as the carbonaceous filaments except in length. Non-carbonaceous filaments seldom exceed 30 μm in length while carbonaceous filaments can reach up to ~100 μm in length. The non-carbonaceous filaments are bright reddish in color and have an undulating appearance that increases as the filament becomes longer. The short filaments are usually straight and have a much stiffer appearance than carbonaceous filaments. The non-carbonaceous filaments appear to have no internal morphology in comparison to the sometimes quite complex internal morphology of the carbonaceous filaments. The greatest difference is the characteristic edifice with globules (~1–2 μm in diameter) connected to each other with short and thin interfaces in between (Fig. 2C). A filament consisting of several globules almost adopts the appearance of a string of pearls (Fig. 2D). At the end of a filament it is possible to view the filament-like interface that connects two globules as well as an indication of a new globule yet to become whole (Fig. 2C).
Globular iron oxides (~1–2 μm in diameter) are found throughout the samples, usually assembled in clusters of several hundred globular iron oxides. These can be found close to both carbonaceous and non-carbonaceous filamentous structures as well as completely non-associated with either one. The globules are seldom mistaken for filamentous structures but the globules that are the building blocks of non-carbonaceous filaments are similar in appearance.
EDS analyses show that the non-carbonaceous filaments almost exclusively contain iron, with an Fe2O3 content ranging between ~80 and ~95 wt% Fe (Table 2). Minor amounts of SiO2 (~10 wt%) occur and in some rare cases MgO (~5 wt%). These elements are probably a result of the mixing of surrounding phases such as calcite and basalt as well as minor incorporations of such elements from the phases mentioned during the formation of the filaments. No carbon is detected in the filaments and the dye PI is not binding to them.
Chemical compositions of non-carbonaceous filamentous structures (NCFS)
Appearance of laser-ablation profiles
Around 50 filaments were analysed with laser ablation. Some filaments that were over 100 μm in length were analysed more than once to observe potential differences or to confirm earlier measurements in the same filaments. In some cases the angle of entrance into the filament was not completely straight but oblique to achieve a longer run within the filament and a more accurate analysis.
All profiles were started and stopped in the vein mineral that the fossilized filaments are embedded in, at a distance from both the filament and the basalt which is a guarantee of no contamination. In the middle of the run the filaments are entered and after some time, depending on how thick the filaments are and the beam size that is used, the filaments are left and vein minerals are entered once again. The profiles have a relatively mutual and characteristic appearance. In the beginning and at the end when the beam is in the vein minerals, the curves are relatively flat with no distinctive features. Peaks or dips occur sporadically due to homogeneities in the calcite composition and minor inclusions of various phases. Calcite and aragonite contain many iron oxides/oxyhydroxides, manganese oxides or parts of altered basalt/volcanic glass/palagonite. Such occurrences of various phases result in local anomalies, but otherwise vein minerals are characterized by high calcium peaks.
As soon as the filaments are entered most elements decrease and iron increases. In some filaments elements such as Mn, Mg and S sometimes follow Fe and increase in comparison to the other elements, but usually only iron increases. This increase of iron in the filaments results in a characteristic bump in the middle of all the filament profiles (Fig. 3).
Laser-ablation profiles in carbonaceous filamentous structures. Microphotographs obtained with a light transmission microscope. Rectangle areas show the profile of the filamentous structure. Scale bars: 10 μm except profile 3:5 μm.
In several cases there is a dip in the iron curves in the middle of the filaments which results in a twin iron peak in the filaments. In between these dips the elements of the vein minerals increase. These dips probably reflect the inner morphology of some of the filaments that are tubular and hollow inside.
Elements
The focus of the analyses was to study differences in iron isotopes but some elements were also analysed contemporaneously to detect correlations or variations between some elements. The elements that were analysed were Fe, Mn, Mg, S, P and Ca. The most obvious correlation is between Mn and Fe, for usually the heavier Fe isotopes. There is also a correlation between Mg and Fe as well as S and Fe. However, these correlations are occasional, and S in particular sometimes shows the opposite behavior. In some samples P and S correspond with each other and have a mutual opposite relationship with Fe in the filaments. In such cases P and S are depleted in the filaments while Fe is enriched. However, when S correlates with Fe in the filaments both S and Fe are enriched.
Fe isotopes in carbonaceous filamentous structures
The iron isotopes follow a characteristic trend throughout the analyses of the filaments (Fig. 3). The heavier iron isotopes, 57Fe and 58Fe, are exclusively enriched in the interior of the filaments and 56Fe is depleted. 54Fe usually follows the trend of the heavier isotopes and is enriched in the filaments but in some cases it is depleted. 56Fe, on the other hand, is almost always enriched in the outer layers of iron oxides that are precipitated or deposited onto the fossilized filaments. These outer zones of the filaments are also always characterized by the lack of 57Fe and 58Fe. In the cases when 54Fe is depleted from the interior of the filaments it sometimes follows the trend of 56Fe and is enriched in the outer iron oxides, but usually it is totally missing. Thus, the characteristic profile of a fossilized filament is a peak of 56Fe at the entrance of the filament. This peak is followed by a dip of 56Fe and a major peak of 57Fe, 58Fe and usually also 54Fe. This peak that represents the interior of the filaments is in most cases characterized by a small dip in the middle due to the hollow structure of the filaments. The profile ends with a decrease in 57Fe, 58Fe and 54Fe and a peak in 56Fe due to entrance into the outer iron oxides. When the beam exits the filaments, there is a decrease in all iron peaks and the peaks of the host mineral increase.
Fe isotopes in non-carbonaceous filaments and host rock
The laser-ablation profiles of the non-carbonaceous filamentous structures differ considerably to the profiles of carbonaceous filaments (Fig. 4). The enrichment of 56Fe in the outer layers and depletion in the interiors as well as the reverse behaviour of 57Fe, 58Fe and to some extent 54Fe is not seen in non-carbonaceous filaments. Instead, the profiles of non-carbonaceous filaments show one major peak in which 56Fe, above all, is enriched. The other iron isotopes show slight enhancements. 54Fe is the isotope that follows 56Fe the most. 58Fe and in particular 57Fe show only minor changes. The 57Fe profile is sometimes completely unaffected when a non-carbonaceous filament is entered.
Laser-ablation profiles in non-carbonaceous filamentous structures. Rectangle areas show the profile of the filamentous structure. Scale bars in profile 1: 5 μm and in profile 2: 10 μm.
Laser-ablation profiles that start in the calcite and then enter the host rock show a similar trend as the non-carbonaceous filaments with enhanced 56Fe and 54Fe values (Fig. 5). 57Fe and 58Fe show less enhancements and 57Fe is sometimes unaffected. The profiles of the basalts also show much stronger irregularities due to mineral grains or alteration zones at the vein boundaries.
Video microscope image and related laser-ablation profile showing the variation of iron isotopes in basalt. Scale bar: 100 μm.
Discussion
Origin of the carbonaceous filamentous structures
The carbon content, the detection of C2H4, phosphates and lipids with the ToF-SIMS as well as the binding of PI led Ivarsson et al. (Reference Ivarsson, Lindblom, Broman and Holm2008) to suggest a biogenic origin for the carbonaceous filaments. Close morphological resemblance to contemporaneous microorganisms and other fossilized microorganisms further support that interpretation.
The microorganisms were interpreted to have lived in the veins as long as fluids were circulating in them (Ivarsson et al. Reference Ivarsson, Lindblom, Broman and Holm2008). As soon as the carbonates were precipitated and the veins were filled, the fluid circulation ended and the microorganisms died and were trapped within the mineral phases. Considering the high grade of preservation, the mineral formation process was probably a relatively quick event.
Several circumstances concerning the appearence of the filaments point in the direction of microbial involvement in iron oxidation. The high carbon content, the close association with iron oxides and the deposition of iron oxides on the surfaces of the filaments suggest involvement in iron oxidation. Thorseth et al. (Reference Thorseth, Torsvik, Furnes and Muehlenbachs1995, Reference Thorseth, Torsvik, Torsvik, Daae and Pedersen2001, Reference Thorseth, Pedersen and Christie2003) made similar interpretations and suggested that iron-rich encrustations on the cells were a result of microbial iron oxidation to gain metabolic energy. Furthermore, cell-like globules that were interpreted as fossilized cells and found in cavities in zeolites in our samples have acted as nuclei for iron precipitation and formation of goethite and hematite (Ivarsson et al. Reference Ivarsson, Lindblom, Broman and Holm2008).
According to Ghiorse (Reference Ghiorse1984), iron oxidizing bacteria produce unique morphological structures such as sheaths and stalks that characterize this group of microbes. Features like attachment to a solid, a bean-shaped cell at the top of a stalk and a twisted stalk structure are all characteristics of Gallionella spp (Ehrlich Reference Ehrlich1996). Furthermore, the tubular-like sheath textures resemble the bacterial sheaths described by Emerson & Moyer (Reference Emerson and Moyer2002) from samples taken at Loihi Seamount. Emerson & Moyer (Reference Emerson and Moyer2002) interpreted them as encrusted sheaths of iron oxidizing Leptothrix ochracea. In some samples the filaments are covered by globular iron oxides as if they were deposited on the fossilized bacteria. According to Ghiorse (Reference Ghiorse1984) sheathed bacteria are more likely to be iron-depositing than iron-oxidizing, meaning that they bind preoxidized iron at their cell surface. Transitions between being iron oxidizing and iron depositing behaviour are common among several iron bacteria, and it is usual that iron oxidizing bacteria deposit the oxidized iron in their sheaths and stalks (Ehrlich Reference Ehrlich1996).
Hydrothermal fluids often contain high amounts of Fe2+, at least acidic hydrothermal solutions, which the microorganisms can scavenge and oxidize to Fe3+. Another source of reduced iron is the basalts. Unweathered basalt minerals such as olivine and pyroxene as well as volcanic glass contain Fe2+ which the microorganisms can use as an energy source. Rogers & Bennet (Reference Rogers and Bennett2004) showed that microorganisms preferentially colonize and weather silicates that contain the limiting nutrients Fe and P, while leaving similar non-nutrient silicates uncolonized and unweathered. It is a well-known fact nowadays that iron oxidation microorganisms live around hydrothermal vents on the ocean floors where they scavenge ferrous iron from the hydrothermal fluids, and a similar mode of life cannot be excluded for sub-seafloor microorganisms (Boyd & Scott Reference Boyd and Scott2001; Emerson & Moyer Reference Emerson and Moyer2002; Edwards et al. Reference Edwards, Rogers, Wirsen and McCollom2003, Reference Edwards, Bach, McCollom and Rogers2004, Reference Edwards, Bach and McCollom2005).
Origin of the non-carbonaceous filamentous structures
The lack of carbon, biomarkers and the inability to bind to PI indicates that the non-carbonaceous filamentous structures are not biogenic in origin. Their chemical composition, dominated by mainly iron and oxygen, is concordant with ordinary iron oxides, and their close morphological resemblance with the globular iron oxides found throughout the samples supports an abiotic origin. However, the fact that they occur as filamentous structures and not assembles in clusters like ordinary iron oxides suggests that their formation is different relative to other iron oxides. Their close morphological resemblance to biogenic filamentous structures is also a reason why their origin as well as their internal structure and morphology are of interest. The ability to distinguish between biogenic and abiogenic filamentous structures is a primary target in research areas such as the sub-seafloor biosphere, the search for the earliest life on Earth as well as the search for an extraterrestrial fossil record (Schopf Reference Schopf1993; Farmer & Des Marais Reference Farmer and Des Marais1999; Gibson et al. Reference Gibson, McKay, Thomas-Keprta, Wentworth, Westall, Steele, Romanek, Bell and Toporski2001; Brasier et al. Reference Brasier, Green, Jephcoat, Kleppe, Van Kranendonk, Lindsay, Steele and Grassineau2002; Schopf et al. Reference Schopf, Kudryavtsev, Agresti, Wdowiak and Czaja2002; Brasier et al. Reference Brasier, Green, Lindsay, McLoughlin, Steele and Stoakes2005).
One possible origin of the non-carbonaceous filaments is that globular iron oxides have simply attached to each other and formed these long filament-like structures. However, it is more likely that globular iron oxides would form irregular clusters, as elsewhere throughout the samples, than filaments. Besides, the appearance of the globules connected with some sort of interface, as well as the thin, knob-like structure that they end with, suggest that they ‘grow’ outwards in length. The embryo globule that is seen at the end of a filament in Fig. 2C further supports the suggestion that the globules grow in situ. It appears as if the globules are formed sequentially; when one globule is finished a knob is formed at the end, and from that knob a new globule is formed.
Abiotically formed filamentous micro-structures are known to form in volcanic rocks on the seafloor when lava meets cold seawater (Perfit et al. Reference Perfit, Cann, Fornari, Engels, Smith, Ridley and Edwards2003). Bubbles of vaporized seawater rise through the base of lava flows and result in the formation of both centimetre-sized drip-like lava stalactites and micrometre-sized filamentous structures similar to the morphology of filamentous microorganisms.
Self-assembled micrometre-sized filaments, so-called biomorphs, have been synthesized in laboratory experiments (Garcia-Ruiz et al. Reference Garcia-Ruiz, Hyde, Carnerup, Christy, Van Krankendonk and Welham2003; Hyde et al. Reference Hyde, Carnerup, Larsson, Christy and Garcia-Ruiz2004). Such biomorphs display morphologies similar to microbial filaments, and they have been used to argue for an abiotic origin of precambrian microfossils from the Warrawoona Chert in Western Australia, which have been claimed to represent the oldest known microbial life on Earth (Garzia-Ruiz et al. Reference Garcia-Ruiz, Hyde, Carnerup, Christy, Van Krankendonk and Welham2003). The biomorphs are grown out of silica-carbonate solutions under ambient pressure and temperatures and at pH ranging from 8.5 to 11. Even though biomorphs containing iron have not yet been produced, a similar formation process could have formed the filamentous iron oxides observed in our samples.
Isotopic variations on a micro-scale
A significant proportion of the Fe isotope literature has focused on the use of Fe isotopes as a biosignature for life. Such a Fe isotope biosignature would require isotopic fractionations that can only be produced by biology. Although there is no question that metabolic processing of Fe produces isotopic fractionation (Beard et al. Reference Beard, Johnson, Cox, Sun, Nealson and Aguilar1999, Reference Beard, Johnson, Skulan, Nealson, Cox and Sun2003; Croal et al. Reference Croal, Johnson, Beard and Newman2004; Johnson et al. Reference Johnson, Roden, Welch and Beard2005), it is also true that similar Fe isotope fractionations may be produced by abiologic processes (Anbar et al. Reference Anbar, Roe, Barling and Nealson2000; Bullen et al. Reference Bullen, White, Childs, Vivit and Schultz2001; Johnson et al. Reference Johnson, Skulan, Beard, Sun, Nealson and Braterman2002; Roe et al. Reference Roe, Anbar and Barling2003; Welch et al. Reference Welch, Beard, Johnson and Braterman2003; Icopini et al. Reference Icopini, Anbar, Ruebush, Tien and Brantley2004). Balci et al. (Reference Balci, Bullen, Witte-Lien, Shanks, Motelica and Mandernack2006) concluded that any characteristic iron isotope fractionation related to microbial oxidation that might constitute a biosignature is likely to be overprinted by rapid isotopic equilibration between Fe(II)aq and Fe(III)aq pools. Thus, it is unlikely that such a biosignature would be expressed in the geologic record.
Iron isotopic ratios are usually expressed as Δ56Fe which is normally calculated as:
![\rmDelta ^{\setnum{56}} {\rm Fe \equals }\left[ {{{^{\setnum{56}} {\rm Fe\sol }{}^{{\rm \setnum{54}}}{\rm Fe}_{{\rm sample}} } \over {^{\setnum{56}} {\rm Fe\sol }{}^{{\rm \setnum{54}}}{\rm Fe}_{{\rm bulk\ earth}} }} \minus 1} \right] \times 10^{\setnum{3}}](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20151007145020566-0314:S1473550408004199_eqn1.gif?pub-status=live){kind=small}
where 56Fe/54Febulk earth is defined by a wide variety of terrestrial and lunar rocks.
Since our results are described by profiles of isotopic variations rather than Δ56Fe values, we can show isotopic trends that are not visual in traditional isotopic measurements. Our profiles show isotopic changes on a micro-scale and not the mean value of a sample. Therefore we are able to show isotopic variations within a fossilized filament and not only give a general isotopic value of a rock that is known to contain fossilized filaments. Our results could perhaps give more detailed isotopic information than a Δ56Fe value and could maybe also indicate something about the formation processes.
Even though the isotope patterns of microfossils in our samples are consistent and differ notably compared to the non-carbonaceous filaments, the results are too preliminary to be used as a biosignature. However, the profile patterns show a trend that is probably a result of some fractionation process.
Significant Fe isotope variations in nature are generally restricted to low-temperature systems and occur at phase transformations such as the fluid–vapor or fluid–mineral. The oxidation of dissolved Fe(II) to solid Fe(III) is one process when fractionation of the isotopes normally occurs. In many cases the phase transformations and thus the fractionation can occur in several steps, for example
Microorganisms process iron when it is energetically favorable to do so, and where they are able to out-compete abiotic redox reactions. Iron is generally biologically processed in three different ways: (1) lithotrophic or phototrophic metabolism where Fe(II) acts as an electron donor for energy generation and/or carbon fixation; (2) dissimilatory Fe(III) reduction, where Fe(III) acts as an electron acceptor for respiration; and (3) assimilatory Fe metabolism which involves uptake and incorporation into biomolecules. For lithotrophic, phototrophic and dissimilatory Fe metabolism, electron transfer occurs between the cell and Fe that is bound to the cell surface or incorporated in the outer membrane.
The outer layer of the microfossils enriched in Fe is interpreted as iron oxides deposited or precipitated on the surfaces of the filaments by the microorganisms. The iron oxides should thus represent the rest products of the microbial oxidation of iron to achieve metabolic energy. If that is the case, microbially precipitated iron oxides are enriched in lighter iron isotopes, especially 56Fe, and depleted in heavier iron isotopes. The opposite appears to be the case for the microorganisms themselves which are enriched in heavy iron isotopes and depleted in light iron isotopes. It has been shown previously that ferric hydroxides produced by Fe(II) oxidation under anaerobic conditions have high δ56Fe values relative to the Fe(II)aq sources (Croal et al. Reference Croal, Johnson, Beard and Newman2004). Iron reducing bacteria, on the other hand, are known to produce ferric oxides that contain heavy iron isotopes compared to the iron released to solution that is enriched in 56Fe (Brantley et al. Reference Brantley, Liermann, Guynn, Anbar, Icopini and Barling2004).
It is difficult to establish the underlying processes for the behavior of iron in the microfossils but two major outlines seem possible: (1) isotopically heavy iron is taken up by the microorganisms and assimilated in their cells which creates a pool of 56Fe that will precipitate as ferric oxides; and (2) both heavy and light iron is taken up by the microorganisms but only 56Fe is processed by the microorganisms and precipitated as ferric oxides in a later stage. The 56Fe is either enzymatically or non-enzymatically processed. The latter may involve the ligand destruction of iron chelates (Ehrlich Reference Ehrlich1996).
Implications
Our results show a characteristic pattern of the isotopic variations of microfossils that may be used to distinguish between biogenic structures and abiogenic structures. The results further suggest that microorganisms produce ferric iron oxides enriched in 56Fe and depleted in 57Fe and 58Fe. However, to simply use the enrichment of 56Fe in iron oxides as a sign of biological activity can be misleading, since the non-carbonaceous filaments are enriched in 56Fe as well. It is crucial to recognize the characteristic isotopic pattern, and thus we conclude that a micro-scale study of the isotopic variations is required to distinguish between biologically and abiologically formed 56Fe precipitates. However, it is too early to use this method alone to establish biogenicity. Micro-scale studies of iron isotopes should be used as a complement to other methods in an attempt to establish biogenicity of putative microfossils.
Conclusions
Carbonaceous filamentous structures that are interpreted as fossilized microorganisms and non-carbonaceous filamentous structures that are interpreted to have formed abiotically are observed in samples of volcanic rock drilled at the Emperor Seamounts in the Pacific Ocean. Both groups of filamentous structures are found to be attached to vein walls in fractured basalts and entombed in vein filling carbonates. The similar morphological appearance and the similar occurrence in the rocks show that there is a need to distinguish the two groups from each other, in an attempt to establish biogenicity.
Laser-ablation ICP-MS analysis of the various filamentous structures shows characteristic profile patterns in iron isotopes that differ between the two groups. The centers of the carbonaceous filaments are enriched in 57Fe and 58Fe and depleted in 56Fe. The surficial parts of the filaments display an opposite behavior of the iron isotopes and are thus enriched in 56Fe and depleted in 57Fe and 58Fe. 54Fe usually follows 57Fe and 58Fe, but in some cases it follows 56Fe instead. The outer, surficial parts enriched in 56Fe have been interpreted as iron oxides precipitated on the surfaces of the microorganisms as they mediate oxidation of the iron to achieve metabolic energy. This may suggest that microbially formed iron oxides are enriched in 56Fe. The laser-ablation profiles of the abiotically formed non-carbonaceous filamentous structures do not show the same characteristics as the carbonaceous filaments but only irregular elevations of 56Fe.
It is too early to suggest that the characteristic laser-ablation profiles of the microfossils could be used as a biosignature, but we do conclude that micro-scale studies of iron oxides may resolve some obstacles in attempts to distinguish biologically formed iron oxides from those that are abiologically formed.
