Iron isotope fractionation in a buoyant hydrothermal plume, 5°S Mid-Atlantic Ridge

Available online at www.sciencedirect.com

Geochimica et Cosmochimica Acta 73 (2009) 5619–5634 www.elsevier.com/locate/gca

Iron isotope fractionation in a buoyant hydrothermal plume, 5°S Mid-Atlantic Ridge Sarah A. Bennett a,*, Olivier Rouxel b, Katja Schmidt c, Dieter Garbe-Scho¨nberg d, Peter J. Statham a, Christopher R. German a,b a

National Oceanography Centre, Southampton, University of Southampton, SO14 3ZH UK b Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA c Jacobs University Bremen, Earth and Space Sciences, 28759 Bremen, Germany d University of Kiel, Institute of Geosciences, D-24098 Kiel, Germany

Received 1 June 2009; accepted in revised form 30 June 2009; available online 7 July 2009

Abstract Fe isotopes are a potential tool for tracing the biogeochemical redox cycle of Fe in the ocean. Specifically, it is hypothesized that Fe isotopes could enable estimation of the contributions from multiple Fe sources to the dissolved Fe budget, an issue that has received much attention in recent years. The first priority however, is to understand any Fe isotope fractionation processes that may occur as Fe enters the ocean, resulting in modification of original source compositions. In this study, we have investigated the Fe inputs from a basalt-hosted, deep-sea hydrothermal system and the fractionation processes that occur as the hot, chemically reduced and acidic vent fluids mix with cold, oxygen-rich seawater. The samples collected were both end-member vent fluids taken from hydrothermal chimneys, and rising buoyant plume samples collected directly above the same vents at 5°S, Mid-Atlantic Ridge. Our analyzes of these samples reveal that, for the particulate Fe species within the buoyant plume, 25% of the Fe is precipitated as Fe-sulfides. The isotope fractionation caused by the formation of these Fe-sulfides is dFe(II)–FeS = +0.60 ± 0.12&. The source isotope composition for the buoyant plume samples collected above the Red Lion vents is calculated to be 0.29 ± 0.05&. This is identical to the value measured in end-member vent fluids collected from the underlying ‘‘Tannenbaum” chimney. The resulting isotope compositions of the Fe-sulfide and Fe-oxyhydroxide species in this buoyant plume are 0.89 ± 0.11& and 0.19 ± 0.09&, respectively. From mass balance calculations, we have been able to calculate the isotope composition of the dissolved Fe fraction, and hypothesize that the isotope composition of any stabilised dissolved Fe species exported to the surrounding ocean may be heavier than the original vent fluid. Such species would be expected to travel some distance from areas of hydrothermal venting and, hence, contribute to not only the dissolved Fe budget of the deepocean but also it’s dissolved Fe isotope signature. Ó 2009 Elsevier Ltd. All rights reserved.

1. INTRODUCTION Iron (Fe) is an essential micronutrient that helps regulate primary production in the upper ocean making it important to understand the geochemical cycle of Fe within the marine environment (Martin, 1990; Martin et al., 1994;

*

Corresponding author. E-mail address: [email protected] (S.A. Bennett).

0016-7037/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.gca.2009.06.027

Lefevre and Watson, 1999; Archer and Johnson, 2000; Boyd et al., 2000; Turner et al., 2001; Christian et al., 2002; Moore et al., 2002). In the deep-ocean, dissolved Fe (dFe, fraction passing through a 0.4 lm filter (Cullen et al., 2006)) concentrations range between 0.4 and 1 nM varying with different water masses (Boyle, 1997; Johnson et al., 1997; Bergquist and Boyle, 2006a). This dissolved Fe fraction is composed of both colloidal and soluble Fe species (Wu et al., 2001; Bergquist et al., 2007). The major supply of dissolved Fe to the surface ocean is

5620

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

understood to come from atmospheric deposition (Jickells and Spokes, 2001; Jickells et al., 2005) with smaller contributions from rivers, groundwater and sediments around coastal regions (Wells et al., 1995; Johnson et al., 1999; Elrod et al., 2004; Windom et al., 2006). However, debate continues over the sources of dissolved Fe to the deepocean. Potential inputs include remineralisation of sinking particles as well as benthic inputs such as seafloor sediments and hydrothermal fluids (Berelson et al., 2003; Boyle et al., 2005; Lam et al., 2006; Nishioka et al., 2007; Bennett et al., 2008), but the ability to quantify the impact of each of these on the deep-ocean dissolved Fe budget is extremely difficult (de Baar and de Jong, 2001). Fe isotopes are a potentially powerful tool for tracing the Fe biogeochemical redox cycle and for constraining the contribution of various Fe sources to the ocean. In nature, stable Fe isotopic variations are controlled by both biotic and abiotic redox processes along with a range of isotope (kinetic and/or equilibrium) fractionations arising from non-redox processes (Welch et al., 2003; Croal et al., 2004; Johnson et al., 2004; Anbar et al., 2005; Balci et al., 2006; Dauphas and Rouxel, 2006). Only very recently has the first data been reported for the isotope composition of dissolved Fe in the open ocean. Rouxel (2009) reported Fe isotope values for open ocean seawater from the Southern Equatorial Atlantic (<1000 m) where Fe concentrations were 0.4 nM, yielding d56Fe values clustered at +0.24 ± 0.16& (d56Fe relative to IRMM-14, 2r, n = 4). These values compare well with recent Fe isotope results from the Southern Atlantic reported by Lacan et al., 2008 (d56Fe = 0.14 to +0.23&). Previously, ferromanganese (Fe-Mn) crusts were used to investigate the long-term variability of the Fe isotope composition of seawater (Zhu et al., 2000; Levasseur et al., 2004; Chu et al., 2006; Yang et al., 2006). The origins of this Fe isotope variability in Fe-Mn crusts continues to be debated but may reflect the relative flux of Fe from aerosol particles (at around 0&) and mid-ocean ridge hydrothermal flux (at around 0.5&) (Beard et al., 2003a; Chu et al., 2006). However, d56Fe values as low as 0.7& are commonly observed in Fe-Mn crusts, which would appear to require an unrealistic contribution from hydrothermally-derived Fe (Levasseur et al., 2004). In addition, the potential for Fe isotope fractionation within hydrothermal plumes remains largely unconstrained even though this would be expected to result in modification of the hydrothermal Fe isotope signature. High-temperature submarine vent fluids have end-member Fe concentrations that typically fall in the range 1 to 10 mM (Von Damm, 1995) and Fe isotope compositions, where measured, ranging from 0.69& to 0.12& (Sharma et al., 2001; Beard et al., 2003a; Severmann et al., 2004; Rouxel et al., 2008). On entering the oxygen-rich open ocean, the Fe from hydrothermal vents typically precipitates to form both polymetallic sulfide and oxyhydroxide phases (Rudnicki and Elderfield, 1993), both of which may cause isotopic fractionation of Fe. Severmann et al. (2004) observed a narrow range of isotope values at the Rainbow (ultramafic hosted) hydrothermal site in both particulate samples, filtered in-situ from the non-buoyant plume (0.18 ± 0.05&) and in underlying sediments

(0.19 ± 0.05&), both of which corresponded to the isotope composition of the end-member fluid (0.23 ± 0.04&). At that location, therefore, it was concluded that the isotopic composition of the hydrothermal end-member fluids was preserved in the non-buoyant plume and recorded in underlying hydrothermal sediments. However, because the Rainbow vent field is hosted in ultramafic rocks, the vent fluids show an unusually high enrichment in both vent fluid [Fe] values and associated Fe/H2S ratios (Douville et al., 2002). Consequently, less than 4% of Fe can be precipitated as polymetallic sulfides at Rainbow and Fe isotope fractionation in the overlying plume is dominated by near-quantitative Fe-oxyhydroxide formation (Severmann et al., 2004). This results in the transfer of all the dissolved Fe released by venting into a single phase that is transported to the top of the buoyant plume, with no overall Fe isotope fractionation. Distinct variations were only observed in two samples that collected material from the buoyant portion of the Rainbow hydrothermal plume (Edmonds and German, 2004), where heavy d56Fe values of up to +1.20& were observed and inferred to be due to partial Fe(II) oxidation (Severmann et al., 2004). In basalt-hosted systems, where up to 50% of the Fe may be precipitated as Fe-sulfides (Rudnicki and Elderfield, 1993), preferential settling of these heavy mineral phases might result in Fe being lost from the plume before reaching non-buoyant plume height. Therefore preservation of the original end-member Fe isotope composition in the buoyant and non-buoyant plume products, as observed at Rainbow, may not be typical. Here, we report on Fe isotope variations in end-member fluids from basalt-hosted vent sites at 5°S, Mid-Atlantic Ridge (MAR) and investigate Fe isotope fractionation in the overlying buoyant plume. Samples were collected during two cruises (RRS Charles Darwin CD169 and FS Meteor cruise M68/1) in 2005 and 2006 and provide an excellent opportunity to study Fe isotope systematics in a basalt-hosted system. 2. GEOLOGICAL SETTING The 5°S hydrothermal fields were first identified from co-registered deep-tow and hydrothermal plume investigations to lie within an area of recent volcanic eruptions at the center of a 2nd order segment of the Mid-Atlantic Ridge (German et al., 2008). Subsequent investigations by Remotely Operated Vehicle (ROV) (Koschinsky et al., 2006; Haase et al., 2007) revealed three distinct high-temperature vent fields, with maximum recorded fluid temperatures of 407 °C, 348 °C and 400 °C, respectively: Turtle Pits, Red Lion and Comfortless Cove. All of these sites are basalt-hosted with slightly altered to fresh lavas consistent with very recent (<5 years) volcanic activity in the area (Haase et al., 2007) (Fig. 1). The Red Lion vent field is made up of four high-temperature chimneys (193–348 °C, Haase et al., 2007) – Mephisto, Tannenbaum, Sugarhead and Shrimp Farm, each separated by 10–15 m. Hydrothermal fluids from the Mephisto and Tannenbaum vents were sampled for Fe isotopes. These fluids were cooler than the fluids emitted from the other two vent sites (Turtle Pits,

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR

5621

Fig. 1. Location of vent sites in the 2 km section of ridge at 5°S MAR. The sites are superimposed onto a bathymetry map of the area.

Comfortless Cove) and shore based chemical analyzes have confirmed that the Red Lion fluids were not phase separated (Haase et al., 2007). The Comfortless Cove field includes one active hightemperature vent (Sisters Peak, Haase et al., 2007) which, when sampled, was 13 m high and consisted of two spires, one of which was actively venting acidic, phase separated fluids, at 400 °C (Cl conc. = 224 mM). Finally, the Turtle Pits field is associated with a series of small depressions indicative of lava drain-back – a similar setting to hydrothermal vent sites associated with recent eruptions on the fast spreading East Pacific Rise (EPR) (e.g. Fornari et al., 2004; Soule et al., 2007). A 9 m tall chimney, Southern Tower, had grown by 4 m between initial sampling in March 2005 and May 2006 while, on two adjacent mounds of sulfide debris, numerous small active black smoker vents are observed: the One Boat and Two Boats sites. Fluids emitted from these mounds were ‘‘flashing” indicative of active phase separation at the seafloor (Cl conc. = 288 mM), and yielded the highest vent fluid temperatures yet recorded from the deep-ocean (Koschinsky et al., 2008). Phase separation in high-temperature hydrothermal fluids is the transformation of a homogeneous solution into two or more phases with varying chlorinity, dependent on the pressure and temperature of the system (German and Von Damm, 2004). The Two Boats (407 °C) fluids were also sampled for Fe isotopes, as was a buoyant section of hydrothermal plume, intercepted approximately 30–60 m north of the Red Lion vent field.

3. SAMPLING METHODS 3.1. Vent fluid sampling End-member vent fluids were collected during FS Meteor cruise (M68/1) using a pumped flow-through system (Kiel Pumping system, KIPS) made entirely of inert materials (perfluoralkoxy (PFA), polytetrafluorethylene (PTFE), and high purity titanium) and operated by the ROV Quest (Garbe-Scho¨nberg et al., 2006). A titanium nozzle was inserted directly into the vent orifice and the sample was pumped through PFA tubing into PFA sampling flasks. Parallel to the nozzle was an on-line temperature probe monitoring the in-situ temperature. All tubes and bottles were purged with hydrothermal fluid for at least 2 min before collection of samples for geochemical analysis. In addition to the KIPS system, a Niskin bottle mounted on the front of the ROV was used to sample the plume directly above the vent outlet. Immediately after recovery of the ROV, all KIPS and Niskin samples were sub-sampled in the ship’s laboratory under a class 100 laminar-flow clean hood (Slee, Germany). After sub-sampling for dissolved gas analyzes, an aliquot of the original fluid was transferred into an acid cleaned polyethylene (PE) bottle (50 ml) and acidified with quartz distilled concentrated HNO3 (Q-HNO3) until all the precipitate had dissolved. For some fluid samples, the precipitates had to be dissolved in the home laboratory by pressure digestion with HCl, HNO3 and HF at 160 °C, followed by re-homogenization of the fluid sample.

5622

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

3.2. Buoyant plume sampling A suite of buoyant plume samples were collected during the RRS Charles Darwin cruise (CD169) using 10-L externally sprung Teflon lined Niskin bottles mounted in a CTD rosette consisting of a SeaBird SBE 9/11+ system coupled with a 24-position rosette mounted in a stainless steel frame. Water samples were collected based on realtime feedback from the CTD and in particular, the in-situ transmissometer, interfaced into the SeaBird CTD. On recovery of the CTD rosette, water samples were collected directly from the Niskin bottle into acid cleaned low-density polyethylene (LDPE) 1 L bottles (with 3-fold rinsing) for the analysis of total dissolvable Mn (TdMn) back in a land based laboratory. These samples were acidified to pH 1.6 using Q-HNO3. The Niskin bottles were then removed from the CTD rosette and carried into a clean laboratory on board the ship, re-homogenizing the fluid in the bottles. Here, a portion of the remaining water was filtered under pressure (N2 gas) through a 0.4 lm membrane filter (Whatman polycarbonate, 47 mm) that had been acid cleaned and a portion of the filtrate was collected into acid cleaned low-density polyethylene (LDPE, 2  125 ml) bottles for the analysis of dissolved Fe (dFe). These samples were later acidified in a land based laboratory to pH 1.6 using Q-HNO3. The filters were retained frozen in small sample polystyrene containers and the volume of seawater filtered through each membrane was recorded to enable quantitative analysis of the particulate elements and Fe isotopes. All further analysis was carried out in a land based laboratory. 3.3. Background station Background seawater samples were collected away from any hydrothermal activity using the same CTD rosette as above. Samples were collected at the equivalent plume height to the hydrothermal samples and a constant background transmissometer signal confirmed the absence of any hydrothermal activity. On recovery of the CTD rosette to the ship, samples for phosphate analysis were taken directly from the Niskin bottles into polystyrene tubes and frozen until analysis back in a land based laboratory. 4. ANALYTICAL METHODS 4.1. Element analysis in the vent fluids Element concentrations of Fe, Mn, Zn, Cu and Mg in the high-temperature fluid samples from Red Lion, Turtle Pits and Sisters Peak were measured in non-filtered acidified aliquots with ICP-OES (Ciros Vision) in the laboratory of the University of Kiel. The analytical run precision, defined as percent relative standard deviation (%RSD) of duplicate measurements of an in-house hydrothermal fluid reference standard, was within 1%RSD for Fe, Mn, Cu and Zn. The analytical method precision, expressing the long-term reproducibility of the same sample, was calculated from repeated measurements of the in-house fluid reference standard and was

within 1% RSD for Fe and Mn, within 5% RSD for Zn and within 6% RSD for Cu. Analytical accuracy was ensured by the measurement of spiked artificial hydrothermal fluid. Hydrothermal fluid samples commonly represent variable mixtures of fluid and seawater caused by mixing. Based on the assumption that Mg is quantitatively removed in high-temperature hydrothermal fluids, the end-member fluid composition was calculated based on a regression versus Mg. 4.2. Phosphate analysis in background seawater samples Phosphate measurements in defrosted background seawater samples were determined using a Skalar Sanplus autoanalyser at the National Oceanography Centre, Southampton (NOCS). The method was validated with a commercial nutrient standard (Ocean Scientific International, Petersfield, Hants, UK). 4.3. Quantification of dissolved Fe and Mn in the buoyant plume samples Fe and Mn measurements in the buoyant plume fluid samples were carried out at NOCS using Graphite Furnace Atomic Absorption Spectroscopy (GFAAS) after a preconcentration step involving chelation of the metals in the seawater with ammonium 1-pyrrolidine dithiocarbamate (APDC, Aldrich) and diethyldithiocarbamic acid, as the diethyl ammonium salt (DDDC, Aldrich) followed by solvent extraction with HPLC grade chloroform (CHCl3, Fisher) and back extraction into quartz distilled concentrated HNO3 (Fisher) (Bruland et al., 1979; Statham, 1985). GFAAS measurements were carried out using a Perkin Elmer AAnalyst 800 Atomic Absorption Spectrometer fitted with an auto sampler and calibration was carried out with a diluted 1000 lg L1 Fe and Mn standard (Sigma Aldrich Germany GmbH). The method was validated with an in-house low metal seawater standard (a bulk quantity of ‘background’ seawater) that had previously been certified against a certified reference material (NASS). The analysis of four aliquots of sub-boiled distilled water, treated throughout in the same way as the samples, produced an average blank value of 1.2 nM for Fe and 0.1 nM for Mn. The detection limit was 1.2 nM for Fe and 0.2 nM for Mn (n = 3), calculated as three times the standard deviation of the blank measurement. 4.4. Quantification of particulate Fe (pFe), Mn (pMn), P, V, Cu and Zn in the buoyant plume particulate samples The 47 mm filters were heated to reflux in 10 ml of concentrated HNO3 (Fisher, OPTIMA grade) in closed 30 ml acid cleaned Teflon vials for 72 h at Woods Hole Oceanographic Institution (WHOI). This procedure allows for the digestion of the sulfide and oxide phases, leaving behind the majority of the filter material and therefore minimizing organic interference (German et al., 1991). Two 30 ml Teflon vials containing no filter and one containing a background filter (filtered during the background seawater cast

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR

5623

The analytical procedure required for isotope analysis follows the previously published method (Rouxel et al., 2003, 2004, 2008; Escoube et al., 2009), and is not repeated here. Analyzes of 56Fe/54Fe and 57Fe/54Fe were carried out on a Thermo-Finnigan Neptune MC-ICP-MS at WHOI (medium resolution mode). Two blocks of 25 integrations were used, 4 s each and typical analyzes were done with 1 mL of 200 ppb solution (200 ng). The samples were introduced into the plasma using a double spray quartz spray chamber system (cyclonic and double pass) and a microconcentric PFA nebulizer operating at a flow rate of about 60 ll min1. 54Fe, 56Fe, 57Fe, 60Ni and 62Ni isotopes were counted on Faraday cups. Baseline corrections were made before acquisition of each data block by completely deflecting the ion beam. Although separated, isobaric Cr interferences were always checked before each sample analysis, using peak jumping mode of the Neptune on 52Cr mass. Purified particulate samples with low Fe concentrations were also analysed using a desolvation nebulizer (Cetac Apex) and X-cones (Thermo-Finnigan) (and 1011 O) to improve the sensitivity of the Neptune (Schoenberg and von Blanckenburg, 2005). Instrumental mass bias was corrected for using Ni isotopes as an internal standard and involves simultaneous measurement of a Ni standard solution (Malinovsky et al., 2003; Poitrasson and Freydier, 2005). Also a standard bracketing approach, which normalizes the Fe isotope ratio to the average measured composition of a standard (IRMM-14) was carried out before and after each sample. The two methods, combined, permit the verification of any instrumental artefacts generated by residual matrix elements. The internal precision of the data have 95% confidence levels based on isotopic deviation of the bracketing standards analysed during the session. The IRMM standard processed through the entire chemical procedure was observed to be indistinguishable from pure solutions and

– Section 3.3) were treated in the same way to give reagent blanks and procedural blanks. A 1:100 dilution of the leach solution was then directly analysed with inductively coupled plasma mass spectrometry (ICP-MS) using medium resolution on a Thermo-Finnigan Element2 at WHOI for Fe, Mn, P, V, Cu and Zn, with detection limits of 11.6, 0.2, 3.5, 0.2, 0.5 and 3.5 nM, respectively. The detection limit of this technique was calculated from repeat analysis of a blank solution of 2% HNO3. These detection limits were well below the concentrations measured in the plume samples, as each filter had at least 6 L of sample filtered through it. Indium (5 ppb) was added to each sample as an internal standard to correct for changes of instrument sensitivity. Stock 1000 lg L1 standards (Specpure, Spex) of each element of interest were diluted in preparation for instrument calibration (ranging from 5 to 1000 ppb). A number of geo-standards (BHVO1 and IFG, Govindaraju, 1994) were prepared and analysed along with the samples to confirm analytical accuracy. 4.5. Fe isotope analysis of particulate and hydrothermal fluid Fe The Teflon vials containing the filter leached solutions were placed on a hot plate and the solution was evaporated to dryness. The solid residue was re-dissolved in 10 ml 1% HNO3 in a closed vial with gentle warming. The samples were then centrifuged and separated from the filter remaining after the evaporation step. A known quantity of the solution was returned to the original Teflon vial and evaporated to dryness. Preparation of the vent fluids for isotope analysis involved evaporating to dryness a known quantity of vent fluid. Depending on its concentration, between 3 and 5 ml of hydrothermal fluid was evaporated in an acid cleaned Teflon vial and then treated along with the other pFe samples.

Table 1 End-member temperature, element concentrations and Fe isotopes at the 5°S MAR vent sites. Site

ID

Tempa (°C)

Cla (mM)

Red Lion – Mephisto Red Lion – Tannenbaum Turtle Pits – Two Boats

7-ROV-5*

346

558

546

7-ROV-13

345

551

3-ROV-10

407

271

Comfortless Cove – Sisters Peak

12-ROV-5 12-ROV-8 20-ROV-5 20-ROV-6

400

224

Cu (lM)

Zn (lM)

Mg (mM)

No.b

d56Fe

1r

d57Fe

1r

550

7.5

75

49.6

2

0.49

0.05

0.74

0.01

712

500

9.5

95

45.3

3

0.29

0.07

0.44

0.10

3984a

487a

6.9

3

0.26

0.11

0.38

0.13

2.8 13.2 38.1 7.4

2 4 2 2

0.21 0.29 0.50 0.45

0.00 0.10 0.05 0.01

0.36 0.40 0.70 0.70

0.03 0.12 0.03 0.05

Fe (lM)

3380a

Mn (lM)

704a

Temp is the end-member temperature measured at each of the chimneys. Two chimneys were sampled at the Red Lion vent site and one chimney at the Turtle Pits and Comfortless Cove vent sites. Discrete samples were taken from Two Boats and Sister Peak and demonstrate the reproducibility of the sampling method. * Niskin sample. a Data from Koschinsky et al. (2008). b Number of duplicate samples analysed.

5624

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

gave an external precision of 0.09& and 0.14& (2r, standard deviation) for d56Fe and d57Fe, respectively. 5. RESULTS 5.1. End-member vent fluid composition The calculated element concentrations for the end-member fluids sampled at two chimneys at Red Lion were very similar (Table 1), even though the fluid from Mephisto was sampled with a Niskin. The Fe concentration ranged from 546 to 712 lM and the Mn concentration ranged from 500 to 550 lM. The Cu concentration was similar for both the fluids (7.5 and 9.5 lM, respectively) and Zn concentration ranged from 75 to 95 lM. Both samples had entrained similar amounts of seawater, as determined from Mg concentrations and it is unlikely that the Mephisto sample, collected using a Niskin rather than KIPS, had lost any Fe from the plume due to the sampling proximity to the vent outlet. In comparison, the fluids from Turtle Pits and Sisters Peak were much more concentrated, with Fe concentrations of 3940 and 3380 lM, respectively, and have been shown to have different reaction zones to Red Lion, even though they are in close proximity (Haase et al., 2007). The Mn concentrations were similar to those at Red Lion, 487 and 704 lM, respectively. The Fe isotope composition of the vent fluids sampled at 5°S MAR ranged from 0.21 to 0.50& for d56Fe (Table 1). The two chimneys at the Red Lion vent field yielded a 0.2& difference in their Fe isotope composition. At Mephisto, the Fe isotope composition was 0.49& whereas the fluid at Tannenbaum was heavier with an isotope composition of 0.29&. At Turtle Pits, three different samples from the same chimney averaged 0.26& with only a 0.04 standard deviation and two samples from Comfortless Cove averaged 0.48&. The vent fluid isotope compositions measured in the current study therefore lie within the same range as previously reported for Atlantic and Pacific vent fluids (Sharma et al., 2001; Beard et al., 2003b; Severmann et al., 2004; Rouxel et al., 2008).

5.2. Buoyant plume 5.2.1. Plume characteristics Changes in measured parameters with time during the upcast of the CTD deployment that intercepted particle rich seawater between 30 and 60 m north of the Red Lion vent field is shown in Fig. 2. The presence of both positive temperature anomalies and strong transmissometer signals, less than 50 m off the seafloor, indicate that a buoyant hydrothermal plume has been sampled at around 3000 m depth, 40 m above the seafloor. During the CTD cast the ship was located directly north of the Red Lion vent field, where four high-temperature chimneys were actively emitting hydrothermal fluids. Therefore all four chimneys are potential sources of the warm particle rich seawater sampled in this study while the other vents fields, Turtle Pits and Comfortless Cove, are located too far (1.5 and 0.9 km, respectively) to contribute significantly. As the CTD was pulled out of the plume at the end of the CTD cast, Niskin bottles were fired repeatedly, collecting samples throughout a lens of buoyant hydrothermal water. For all samples collected, temperatures were greater than the background temperature, with varying Mn concentrations (Fig. 2 and Table 2). The most Mn rich lens of plume water occurred over a small ‘depth range’ of approximately 30 m, which reflected passage through a lens of buoyant plume water. Therefore rather than collecting a suite of samples vertically through the rising buoyant plume, the CTD drift resulted in an oblique cross section of the buoyant plume being sampled. The inferred trajectory that the CTD took through the buoyant plume is shown schematically in Fig. 3. The highest temperature and particle anomalies coincide with the core of the plume. 5.2.2. Fe concentrations within the buoyant plume A time profile of these samples with pFe, dFe and total Fe (TFe = pFe + dFe) is shown, together with co-registered transmissometer readings, in Fig. 4. During sample collection, total Fe concentrations varied from 71 to 413 nM. As the CTD rosette crossed into the lens of buoyant plume

Fig. 2. Time profile of the CTD cast that intercepted particle rich water. Solid circles represent total dissolvable Mn (TdMn), dashed line represents temperature, dotted line represents depth and solid line represents light transmission.

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR

5625

Table 2 Element concentrations and Fe isotopes measured in samples collected from the buoyant plume. Depth (m)

TdMn (nM)

pMn (nM)

dFe (nM)

pFe (nM)

TFe (nM)

P (nM)

V (nM)

Cu (nM)

Zn (nM)

Fe: (Fe + Mn + Al)

No.a

d56Fe

1r

d57Fe

1r

3009 3006 3002 2996 2992 2989 2985 2981

132.0 91.0 96.4 160.3 202.6 208.2 284.1 356.6

0.1 0.1 0.2 0.1 0.2 0.1 0.1 0.2

49.2 53.2 45.7 28.4 27.2 39.0 26.8 20.7

77.5 21.6 25.3 94.5 168.8 191.3 210.1 392.3

126.7 74.7 70.9 122.9 195.9 230.4 236.9 413.0

3.9 0.6 0.8 7.0 12.2 15.1 17.3 33.6

0.18 0.03 0.04 0.23 0.45 0.53 0.58 1.11

2.9 1.4 1.8 3.0 3.5 4.1 4.9 7.1

10.6 4.2 7.2 13.3 15.3 18.6 19.5 34.9

0.91 0.82 0.90 0.93 0.97 0.98 0.92 0.99

3 3 2 4 2 4 4 4

0.56 0.70 0.56 0.47 0.44 0.39 0.36 0.31

0.05 0.07 0.03 0.05 0.00 0.07 0.03 0.03

0.79 0.96 0.95 0.65 0.65 0.66 0.61 0.58

0.13 0.07 0.05 0.04 0.00 0.03 0.02 0.05

Total dissolvable Mn (TdMn) was measured in samples taken directly from the CTD bottle, prior to filtration. Dissolved Fe (dFe) concentrations were measured in filtered samples. Particulate Mn (pMn), Fe (pFe), P, V, Cu, Zn and Fe isotopes were all measured in the extracted filters. Total Fe was calculated from the sum of (dFe + pFe). a Number of duplicate samples analysed.

the TFe concentration increased, with maximum Fe concentrations observed at 2981 m, corresponding to 65 m above the Red Lion vents. At this depth, if it is assumed that there was no lateral movement of the plume, any plume rising with a 15° half-angle (Turner, 1962) would have spread to approximately 35 m in diameter. Therefore the CTD must, apparently, have been within 20 m of the buoyant plume axis at this time. The pFe concentrations followed the same pattern as TFe, ranging from 22 to 392 nM. Conversely, the dFe concentrations decreased as the CTD was brought up through the water column with much lower concentrations, ranging from 53 nM to 21 nM at the core of the major plume lens. The low dFe concentrations can be attributed to particle precipitation in the Niskin bottles between the time of trip-

ping the bottles at depth and recovery of the CTD aboard ship. It was not until the samples were on board the ship, that the particulate Fe could be separated from any remaining dissolved Fe by filtration. Therefore the Fe content (dissolved and particulate) measured within the samples does not correspond to the partitioning of Fe within the plume at the time of sampling. During sample recovery, the dissolved Fe will have been continually oxidizing and precipitating as Fe-oxyhydroxides. The Fe oxidation rate has been calculated previously for the 5°S vent sites and a half-life of 27 min for dissolved Fe(II) has been determined (Bennett et al., 2008). Between 4 and 8 h elapsed between sampling the plume and processing the samples in a clean laboratory on the ship corresponding to between 8 and 16 half-lives and, consequently, resulting in almost quantitative oxidation of any Fe(II) that had been present. The remaining dFe fraction that we have measured, therefore, is most probably composed of colloidal Fe(III) species as well as a small percentage of truly dissolved Fe that may have been stabilised by organic complexation (Bennett et al., 2008). 5.3. Particulate composition

Fig. 3. Schematic of the potential upcast path the CTD took through the buoyant plume. The temperature and transmissometer graph is to scale with the height shown on the schematic. The dashed line represents temperature and the solid line represents light transmission.

The particulate P, V, Cu and Zn concentrations relative to pFe are shown in Fig. 5a and b. The oxyanions, P and V, are positively correlated to pFe with an r2 of 0.995 and 0.999, respectively, and a negative y intercept. Phosphorus ranged from 0.6 to 34 nM and vanadium ranged from 0.03 to 1.11 nM (Table 2). P/Fe ratios lie between 0.03 and 0.09 and compare well with the Atlantic hydrothermal vent fields, and V/Fe ratios lie between 0.001 and 0.003, comparable to the Pacific hydrothermal vent fields (Feely et al., 1998). The phosphate concentration in the background CTD cast ranged between 1.60 and 1.64 lmol L1, between 2930 and 2952 m. The chalcophiles, Cu and Zn, are also positively correlated with pFe, with r2 values equal to 0.976 and 0.981, respectively, and with positive y intercepts. Copper concentrations ranged from 1.4 nM to 7.1 nM and zinc concentrations ranged from 4.2 to 35 nM (Table 2). Cu/Fe ratios lie between 0.02 and 0.08, similar to the ratios measured at TAG and Rainbow (0.006–0.080 and 0.003–0.015,

5626

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

respectively (German et al., 1991; Edmonds and German, 2004)). By comparison, the Cu/Fe ratio in the end-member fluids at Red Lion is less than 0.01 which suggest that Cu is preferentially precipitated as polymetallic sulfides relative to Fe. The Zn/Fe ratios on the other hand ranged between 0.09 and 0.29, much greater than the ratios observed at TAG and Rainbow (0.001–0.008 and 0.001–0.066). However, the most likely reason for these high ratios is the

abnormally high Zn concentration in the end-member fluids at Red Lion, which have a Zn/Fe ratio of 0.13 (Table 1). 5.3.1. Iron isotope composition of the particulate Fe in the buoyant plume The Fe isotope composition of the particulate Fe varied from 0.31& to 0.70& for d56Fe (Fig. 4 and Table 2) and appeared to become heavier with increasing pFe. The high-

Fig. 4. Time profile of light transmission (solid line) and particulate Fe (pFe, solid circles), dissolved Fe (dFe, open circles), Total Fe (TFe = dFe + pFe, crosses), and Fe isotope composition of the particulate Fe (crosshair and dashed line)

a

b

Fig. 5. (a) Relationships between particulate P, V relative to particulate Fe (pFe) (b) Relationships between particulate Cu, Zn relative to particulate Fe (pFe)

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR

est d56Fe values (0.31&) are observed in the core of the plume and compare well with the average d56Fe value of high-temperature hydrothermal fluids at Red Lion (0.39 ± 0.13&). In contrast, d56Fe values down to 0.70& encountered in the outer edges of the plume are significantly lower than vent fluid values. 6. DISCUSSION 6.1. Variations in the isotope composition of the end-member vent fluids Fe isotope results at 5°S MAR provide further evidence that hydrothermal fluids have negative d56Fe values relative to the bulk silicate Earth, determined at 0.09& (Sharma et al., 2001; Beard et al., 2003a; Severmann et al., 2004; Rouxel et al., 2008). However, the parameters affecting the Fe isotope composition of hydrothermal fluids are still poorly constrained. So far, neither phase separation nor geological setting (i.e. basaltic or ultramafic rocks) have been found to produce significant isotope variability (Beard et al., 2003b). At 5°S, the Fe isotope compositions of the fluids vary by 0.3& (Table 1), but d56Fe data do not correlate with Cl concentrations. Hence, although the fluids have undergone phase separation, this process has not affected the Fe isotope composition of the end-member fluid. Haase et al. (2007) found that fluid and vent field characteristics at 5°S MAR have a large variability between the vent fields, which is unexpected within such a limited area. Accordingly, each of the sites at 5°S may have a different reaction zone which cold seawater has percolated down into, heated up, reacted with surrounding rock, and undergone phase separation before venting to the seafloor. These processes would be expected to affect the isotope composition of the Fe in the end-member fluids. Rouxel et al., 2003 observed the results of low-temperature hydrothermal crust alteration on the isotope composition of basaltic crust at ODP site 801. In particular, highly altered basalts that were depleted by up to 80% from their original Fe concentration displayed an increase in d56Fe values relative to fresh rocks (up to 1.3&), which suggested preferential leaching of light Fe isotopes (between 0.5& and 1.3&) during alteration. Similar isotopic fractionation processes may occur through the formation of secondary minerals (e.g. Mg-Fe amphibole) in the high-temperature reaction zone. Other fractionation processes that could be occurring sub-surface include the precipitation of Fe-sulfides due to conductive cooling of the fluid or mixing with seawater (Rouxel et al., 2004, 2008). However, the importance of such processes is difficult to predict based on the limited number of samples available in this study. 6.2. Buoyant plume 6.2.1. Source of the plume particles The seafloor close to a hydrothermal vent is a dynamic and turbulent environment. Therefore, a first priority has to be to demonstrate that the particulate samples are not affected by entrained resuspended particles from the surrounding sediments. The classification developed by

5627

Bostrom et al. (1969) using particulate Al, Fe and Mn concentrations has previously been employed (German et al., 1991) to distinguish between hydrothermal and lithogenic components. Al, Mn and Fe have high solid phase concentrations within the sediments, whereas in the end-member fluids, Al is present as a trace element and Mn is present as a dissolved reduced species that is only slowly removed, following mixing with the surrounding seawater, from the non-buoyant plume (Trocine and Trefry, 1988). Fe on the other hand, initially present as a dissolved reduced species, precipitates rapidly forming solid phase Fe-sulfides and Feoxyhydroxides as soon as the vent fluids first mix with seawater in the buoyant plume. Therefore, a high value of the ratio Fe/(Fe + Mn + Al) can be used to indicate that samples have little detrital input, with particulate Mn and Al concentrations both low. Here, all the samples collected during the CTD upcast had Fe/(Fe + Mn + Al) values greater than 0.8 (Table 2) indicating a high proportion of hydrothermal material (more than 80%) with little influence from any resuspended sediments. This is consistent with the minimal sediment cover associated with the fresh basalts surrounding the vent sites at 5°S. Therefore these samples can be assumed to be predominantly hydrothermal precipitates, which is consistent with the high-temperature anomalies intercepted in the plume. 6.2.2. Composition of the particulate Fe fraction As the buoyant plume evolves, Fe-sulfide and Fe-oxyhydroxides are precipitated. During oxyhydroxide formation, oxyanions (i.e. HPO42 and VO4H2) are co-precipitated and scavenged (Trefry and Metz, 1989; Feely et al., 1991, 1996, 1998; Rudnicki and Elderfield, 1993) with constant molar ratios of P/Fe and V/Fe dependent on the local dissolved phosphate concentration (Feely et al., 1998). In the buoyant plume at 5°S, both P and V were observed to be linearly correlated to pFe (Fig. 5a), similar to the vent systems at Rainbow and TAG (German et al., 1991; Edmonds and German, 2004). However, both the P to Fe and V to Fe relationships have negative intercepts suggesting the presence of a particulate Fe species that does not associate with the oxyanions (i.e. Fe-sulfides). In contrast to oxyanions, chalcophile elements, such as Cu and Zn sourced from the hydrothermal end-member, are incorporated with Fesulfides in hydrothermal plumes. It is also possible that a fraction of those elements are adsorbed onto Fe- and/or Mn-oxyhydroxide phases at seawater pH, but this process is probably not significant in hydrothermal plumes where the concentration of H2S is sufficient to quantitatively precipitate the chalcophile elements. Under those assumption, the particulate Cu and/or Zn relationship to Fe has been used previously to demonstrate a preferential loss of Fe-sulfides from the plume through gravitational sinking and/or the removal of Fe-sulfides via oxidative dissolution (Trocine and Trefry, 1988; German et al., 1991). In the buoyant plume at 5°S, the relationship between Cu and/or Zn with Fe is linear and shows no negative departure (Fig. 6). In fact, the Cu and/or Zn relationship to Fe has a positive y intercept, indicating a higher proportion of Cu and Zn relative to Fe in the low pFe samples. This is likely to be due

5628

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

to a higher proportion of Fe-sulfides in these samples, however a greater adsorption of Cu and/or Zn onto Fe-oxyhydroxide phases present in these low Fe samples cannot be discounted. To understand these relationships, quantitative determination of the amount of Fe-oxyhydroxides versus Fe-sulfides in these plume samples can be carried out using particulate P/Fe ratios. Feely et al., 1998 have shown that the P/Fe ratios measured in the plume particles at various vent sites across the ocean basins correlate linearly to the phosphate concentration in the background seawater at each particular site. The dissolved phosphate concentration at a background station at 5°S was 1.6 lmol L1, this was slightly greater than that at TAG (1.38 lmol L1, Feely et al., 1998) and by using the linear relationship in Feely et al. (1998), the P/Fe ratios in Fe-oxyhydroxide particles is expected to be at around 0.11 (on a molar basis). In comparison, the V/Fe ratios are inversely correlated to pFe, because even though vanadate is scavenged by the Fe-oxyhydroxides, vanadate does not compete as well as phosphate for the available sites on the Fe-oxyhydroxides. Therefore the V concentrations on the hydrothermal particles are lower when the dissolved phosphate concentrations are high. By using the inverse relationship in Feely et al., 1998 an average V/Fe ratio of 0.004 can be estimated for Fe-oxyhydroxide particles at the measured phosphate concentration at 5°S MAR. The fraction of Fe-oxyhydroxide (hereafter referred to XFeOOH) present in the plume particles can be calculated using Eqs. (1), using P (X P FeOOH ) and (2), using V (X V FeOOH ): X P FeOOH ¼ ðP=pFeÞmeasured =ðP=FeÞexpected

ð1Þ

X V FeOOH ¼ ðV=pFeÞmeasured =ðV=FeÞexpected

ð2Þ

The results are shown in Table 3 and both relationships with P and V appear to result in similar Fe-oxyhydroxide concentrations with a maximum difference of 0.11 (11%). The fraction of Fe-sulfide (XFeS) in the particulates, deter-

Fig. 6. The isotope composition measured in the particulate samples relative to the percent pFe (%pFe). An excellent linear regression falls through the samples with more than 60% pFe (solid circles). The samples with less than 60% pFe (open circles) are not included in the regression analysis.

mined as XFeS = 1 XPFeOOH decreases with increasing particulate Fe. This relationship is also consistent with the increase of Cu/Fe ratio at low particulate Fe reported in Fig. 5b, indicating a greater proportion of Fe-sulfides. However, relative to the total Fe (dissolved and particulate), the fraction of Fe-sulfide is relatively constant (Table 3), making up 25% of the total Fe in all of the buoyant plume samples. Given that these samples were collected from the early stages of the buoyant plume before much of the Fe(II) has oxidized, and that the fraction of FeS relative to the total Fe is constant, loss of Fe due to settling or addition of Fe from resuspended sediment is minimal in these samples. 6.2.3. Isotope fractionation in the buoyant plume and the isotope composition of the buoyant plume source Four main isotope fractionation processes can occur in the buoyant plume: (1) kinetic Fe isotopic fractionation during Fe-sulfide precipitation, preferentially precipitating the light isotope (Butler et al., 2005; Polyakov et al., 2007; Rouxel et al., 2008); (2) partial oxidation of Fe(II)aq to Fe(III)aq which enriches the ferric Fe fraction in the heavy isotope (Welch et al., 2003; Anbar, 2004; Balci et al., 2006); (3) precipitation of Fe(III)aq into ferrihydrite, which has the potential to kinetically fractionate the light isotope and can be dependent on the rate of precipitation (Skulan et al., 2002) and (4) adsorption of Fe(II) onto Fe-oxyhydroxides (Icopini et al., 2004; Teutsch et al., 2005; Matthews et al., 2008; Mikutta et al., 2009). Oxidative precipitation of ferrihydrite is a combination of (2) and (3) because of the instability of Fe(III)aq species in seawater, producing an overall enrichment in the heavy isotope (Bullen et al., 2001). Further transformation of ferrihydrite to goethite has also been observed to cause isotope fractionation over a time scale of 70 h, but the age of the samples analysed in this study are much younger (Clayton et al., 2005). Adsorption and desorption of Fe(II) onto Fe-oxyhydroxide or Fe-sulfide phases is another potential fractionation mechanism in the plume, with isotopically heavy Fe(II) adsorbing to mineral surfaces. Isotopically heavy Fe can then preferentially be desorbed resulting in an isotopically heavy dissolved Fe fraction (Icopini et al., 2004; Teutsch et al., 2005; Matthews et al., 2008; Mikutta et al., 2009). However, under oxygenated conditions, it is likely that most of the adsorbed Fe(II) will undergo oxidation on the mineral surface and will not contribute further to the total dissolved Fe pool in the buoyant plume. There may also be dissolution of Fe-sulfide particles to Fe(II) or the formation of stabilised dissolved Fe complexes (Fe-ligands, FeL). Currently the isotopic fractionation caused by organic ligand complexation with Fe is unconstrained and only a couple of studies have identified potential fractionation pathways. Brantley et al. (2004) demonstrated ligand effects during mineral dissolution, producing an enrichment in the light isotopes in the dissolved phase, whereas Dideriksen et al., 2008 demonstrated ligand effects with free siderophore ligands and dissolved Fe. Here the organically bound Fe was enriched in the heavier isotope. This supported data from organically rich river water where the dissolved Fe was found to be isotopically heavy

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR

5629

Table 3 Calculated fractions of Fe-oxyhydroxides (FeOOH) and Fe-sulfides (FeS) in the buoyant plume using P and V concentrations. Calculated from phosphorus a,c

In particulate fraction XPFeOOH

X

0.46 0.25 0.29 0.68 0.66 0.72 0.75 0.78

0.54 0.75 0.71 0.32 0.34 0.28 0.25 0.22

a b c

P

FeS

Calculated from vanadium In particulate fractionb,c

Relative to total Fe

XdFe

XVFeOOH

XVFeS

XVFeOOH

XVFeS

XdFe

0.39 0.71 0.64 0.23 0.14 0.17 0.11 0.05

0.57 0.34 0.39 0.59 0.65 0.68 0.67 0.69

0.43 0.66 0.61 0.41 0.35 0.32 0.33 0.31

0.35 0.10 0.14 0.46 0.56 0.56 0.60 0.66

0.26 0.19 0.22 0.31 0.30 0.27 0.29 0.29

0.39 0.71 0.64 0.23 0.14 0.17 0.11 0.05

Relative to total Fe X

P

FeOOH

0.28 0.07 0.10 0.52 0.57 0.59 0.66 0.74

X

P

FeS

0.33 0.22 0.25 0.25 0.29 0.24 0.22 0.21

X P FeOOH = (P/pFe)measured/(P/Fe)expected. X V FeOOH = (V/pFe)measured/(V/Fe) expected. XFeS = 1 XFeOOH.

and hypothesized to be because of organically bound Fe (Bergquist and Boyle, 2006b; Escoube et al., 2009). At present, the mechanism for ligand stabilisation within hydrothermal plumes is unknown and has the potential to follow either of these pathways. For example, microbial dissolution of minerals in the hydrothermal environment has the potential to produce organically stabilised Fe (Edwards et al., 2004; Homann et al., 2009), which could be entrained into the buoyant plumes. Also organic ligands and elevated dissolved organic carbon has been detected in diffuse flow areas (Lang et al., 2006; Sander et al., 2007), which if entrained into the buoyant plume, could complex with dissolved Fe species (Bennett et al., 2008; Toner et al., 2009). All the processes identified above can modify the original isotopic composition of the vent fluid as the Fe precipitates in the buoyant hydrothermal plume. In this particular case, Fe-sulfide precipitation will have occurred in the first few seconds of venting and, by the time these samples were filtered on deck, complete oxidation of all Fe(II)aq will also have occurred. Additionally, the majority of Fe remaining in the dissolved Fe fraction is expected to be present as colloidal Fe-oxyhydroxides, and potentially as Fe-organic complexes although these will only make up a small fraction of the total Fe (Bennett et al., 2008). Therefore any fractionation caused by Fe-sulfide formation will have occurred prior to sampling and any fractionation caused by partial Fe oxidation or ferrihydrite formation will have occurred in the sampling Niskin bottle by the time these samples were processed on deck (Fe(II) half-life 27 min (see earlier)). The formation of any organic Fe complexation is unknown but could have occurred early on in the buoyant plume, or later, in the Niskin bottle. Considering that the process of Fe oxidation and ferrihydrite formation in the plume and in the Niskin bottle are likely analogous, the time lag between plume sampling and on board sample filtration should not significantly affect our interpretation. By plotting the particulate Fe isotope compositions against %pFe (Fig. 6), we can see that as the particulate Fe fraction increases, so does its isotope composition. When more than 60% of the Fe has precipitated, an excellent linear correlation is observed, with an r2 of 0.92. As there has been no loss or gain of Fe to the plume (Section

6.2.2) and any isotope fractionation processes have already occurred, a simple mixing process between two isotopically different pools, can be used to explain this relationship. Isotopically light particulate Fe must be gaining isotopically heavy Fe from the dissolved fraction. This occurs as the dissolved Fe-oxyhydroxide complexes increase in size (due to colloidal coagulation) and as a result are separated into the particulate fraction. This relationship is very useful as it can be used to determine the source isotope composition of the buoyant plume of 0.29 ± 0.05& (at 100% particulate Fe precipitation, error calculated through regression analysis of the data (95% confidence level) followed by Monte Carlo propagation), the same as the fluid measured at the Tannenbaum Chimney at Red Lion. Without this relationship it would be impossible to know the isotopic composition of the buoyant plume as only two out of the four chimneys at Red Lion have been sampled and analysed for their isotope composition, varying by 0.2&. The two samples with less than 60% Fe precipitation were sampled just outside the major buoyant plume lens (Fig. 4, at 5785 and 5805 s) and have the lowest total Fe concentration. The transmissometer profile (shown in Fig. 4) is different for these two samples compared to those in the major plume and suggests a different source. Whether this is a different high-temperature source or resuspended material from the seafloor is difficult to determine. However, they still show very similar Fe-sulfide compositions to the samples in the major buoyant plume lens (Table 3). 6.2.4. Isotope fractionation factor for Fe-sulfide precipitation In Fig. 7, the percent Fe-sulfide is shown relative to the isotope composition in the particulate Fe fraction, demonstrating an excellent linear correlation (r2 = 0.89). Since the isotope composition becomes lighter as the percent Fe-sulfide increases, the Fe isotope composition of the pure Fesulfide end-member is estimated at d56Fe = 0.89 ± 0.11& while the Fe-oxyhydroxide end-member is 0.19 ± 0.09& (error calculated through regression analysis of the data (95% confidence level) followed by Monte Carlo propagation). The majority of the Fe-sulfides will be present within the particulate phase and therefore the isotope composition of the particulate phase is controlled by the balance of

5630

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

Fe-oxyhydroxides present in the dissolved and particulate phase. Considering an initial hydrothermal fluid d56Fe value of 0.29 ± 0.05& a fractionation factor for Fe-sulfide precipitation can be calculated for the hydrothermal plume at Red Lion of 1.0006 with a dFe(II)–FeS = +0.60 ± 0.12&. This is consistent with experimental studies by Butler et al. (2005), where a kinetic isotope fractionation process of dFe(II)–FeS = +0.85& ± 0.30& has been observed for zero age Fe-sulfides over a temperature range between 2 and 40 °C. During aging, the degree of fractionation decreased between the precipitates and the fluid to 0.3&. This initial kinetic isotope fractionation observed with zero age precipitates has been interpreted to be due to the ligand exchange between hexa-aqua Fe(II) and aqueous sulfide complexes. In the buoyant plume samples the precipitation of Fe-sulfides will be rapid upon mixing of the hydrothermal fluid with seawater and therefore it is expected that a significant kinetic Fe isotope effect will be observed. These results are also consistent with previous studies by Rouxel et al. (2008) reporting low d56Fe values of suspended particles from Ti end-member sample bottles (down to 1.8&) precipitating upon sampling of high-temperature hydrothermal fluids at EPR 9–10°N. 6.2.5. Prediction of the isotope composition of the dissolved Fe species in the buoyant plume It has been suggested that no Fe has been lost from (or gained to) the plume (see above). Therefore the isotope composition of total Fe in each plume sample (collected in the Niskin bottle, referred to as d56FeTFe) should remain constant and should be similar to the initial composition of the buoyant plume (d56Fe value of 0.29 ± 0.05&). This enables us to establish an Fe isotope budget in the buoyant plume at 5°S MAR (Eq. (3)):

with XFeOOH, XFeS, XdFe being the relative proportion of Fe-oxyhydroxide, Fe-sulfide, and dissolved Fe in each plume sample, respectively, as reported in Table 3. d56FeFeOOH and d56FeFeS are assumed to be, respectively, 0.19 ± 0.09& and d56Fe 0.89 ± 0.11& as determined above. This mass balance equation can be used to quantify the isotope composition of dissolved Fe species in each of the plume samples with their respective errors using a Monte Carlo propagation (Fig. 8). The calculated isotope compositions of the dissolved Fe fractions are predicted to be heavier than the original isotope composition of the buoyant plume. However, as shown in Fig. 8, these values become more uncertain as the dissolved Fe fraction decreases, with increasing error bars. For example, the remaining 5% dissolved Fe has a d56FedFe value of +0.8 ± 2.2&, heavier, but still within error of the original isotope composition of the buoyant plume. This fractionation is potentially a result of the following, previously discussed processes; (1) incomplete Fe(II) oxidation, but as discussed earlier, unlikely due to complete oxidation of the sample (2) Fe-sulfide precipitation, making any nanoparticulate Fe-oxyhydroxides isotopically heavier than the original vent fluid isotope composition, (3) the adsorption and desorption of Fe(II) onto suspended mineral precipitates and/or (4) organic stabilisation of dissolved Fe. However, the large errors associated with our isotope mass balance approach preclude estimating a diagnostic signature for a stabilised dissolved Fe fraction escaping from hydrothermal plumes into the deep-ocean. Direct Fe isotope analysis of the dissolved fraction would be required to further address this issue. 7. CONCLUSION

ð3Þ

It was shown in the first isotopic plume study carried out by Severmann et al. (2004), at the ultramafic hosted Rainbow vent site, that the isotope composition of the

Fig. 7. The relationship between the percent of Fe-sulfide (%FeS), as calculated from phosphorus, in the particulate fraction and the isotope composition of the particulate Fe. The solid circles are more than 60% pFe (see Fig. 6) and the open circles are less than 60% pFe. Regression analysis was carried out only on the samples with more than 60% pFe.

Fig. 8. The predicted isotope composition of dissolved Fe relative to the percent dissolved Fe (% dFe). The dotted line represents the original composition of the buoyant plume. The solid circles are more than 60% pFe (see Fig. 6) and the open circles are less than 60% pFe. Errors were determined using Monte Carlo simulation with additional errors propagated for XFeOOH and XFeS using a 10% error on the P/Fe ratio.

d56 FeT ¼ X FeOOH d56 FeFeOOH þ X FeS d56 FeFeS þ X dFe d56FedFe

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR

hydrothermal end-member fluid was preserved in the non-buoyant plume and recorded in local hydrothermal sediments. In that system, no Fe was lost during plume rise and therefore it is not unexpected that the isotope composition of Fe in the non-buoyant plume was the same as that in the original vent fluid. However, in basalt-hosted systems, between 25% (this study) and 50% (Rudnicki and Elderfield, 1993) of Fe is lost during buoyant plume rise as Fe sulfide precipitates, potentially resulting in fractionation of the original Fe-isotope composition. In this study, we have determined the isotope composition of various Fe species within a buoyant hydrothermal plume from a basalt-hosted vent site. From the isotope composition of the particulate Fe and its relationship with percent Fe-sulfide, an isotope fractionation for Fe-sulfide precipitation has been determined (dFe(II)–FeS = +0.60 ± 0.12&). As the majority of hydrothermal venting, especially at fast spreading ridges, occurs at basalt-hosted systems, Fe is expected to be fractionated by Fe-sulfide precipitation. By non-buoyant plume height, at least some Fe-sulfide precipitates are lost from the plume, due to preferential settling, leading to the removal of isotopically light Fe. From this study, it can be hypothesized that a stabilised dissolved Fe fraction may have an isotopic signature heavier than the original vent fluid and that this could be used to trace hydrothermally sourced dissolved Fe throughout the deep-ocean. Considering the potential of stabilised organic colloids and nanoparticulate Fe to be transported some distance from areas of hydrothermal venting, the Fe isotope signatures of hydrothermally-derived Fe and its record in ferromanganese crusts warrant further investigations. ACKNOWLEDGMENTS We thank the captain, crew and technical team on board the RRS Charles Darwin, CD169 and FS Meteor, M68/1 who contributed to the success of this work and to Alan Evans for creating the location map (Fig. 1). We also thank Karin Kissling for help with the ICP-OES measurements, Doug Connelly for help with the GFAAS measurements and Mark Stinchcombe for phosphate measurements. We thank Lary Ball and Dave Schneider for access to the WHOI Plasma Mass Spectrometry Facility. We thank Mark Rehka¨mper, Alan Matthews, Silke Severmann and an anonymous reviewer for their constructive comments on this paper. Support for O. Rouxel was provided by funding from the WHOI Deep Ocean Exploration Institute, Frank and Lisina Hoch Endowed Fund and NSF-OCE 0648287. S.A. Bennett was funded by a NERC Ph.D. studentship, NER/S/A/2004/12631 with additional travel awards from the Kristin Bruhn Foundation and The Challenger Society for Marine Science.

REFERENCES Anbar A. D. (2004) Iron stable isotopes: beyond biosignatures. Earth Planet. Sci. Lett. 217, 223–236. Anbar A. D., Jarzecki A. A. and Spiro T. G. (2005) Theoretical investigation of iron isotope fractionation between Fe(H2O)63+ and Fe(H2O)62+: implications for iron stable isotope geochemistry. Geochim. Cosmochim. Acta 69, 825–837. Archer D. E. and Johnson K. (2000) A model of the iron cycle in the ocean. Global Biogeochem. Cycles 14, 269–279.

5631

Balci N., Bullen T. D., Witte-Lien K., Shanks W. C., Motelica M. and Mandernack K. W. (2006) Iron isotope fractionation during microbially stimulated Fe(II) oxidation and Fe(III) precipitation. Geochim. Cosmochim. Acta 70, 622–639. Beard B. L., Johnson C. M., Skulan J. L., Nealson K. H., Cox L. and Sun H. (2003a) Application of Fe isotopes to tracing the geochemical and biological cycling of Fe. Chem. Geol. 195, 87–117. Beard B. L., Johnson C. M., Von Damm K. L. and Poulson R. L. (2003b) Iron isotope constraints on Fe cycling and mass balance in oxygenated Earth oceans. Geology 31, 629–632. Bennett S. A., Achterberg E. P., Connelly D. P., Statham P. J., Fones G. R. and German C. R. (2008) The distribution and stabilisation of dissolved Fe in deep-sea hydrothermal plumes. Earth Planet. Sci. Lett. 270, 157–167. Berelson W., McManus J., Coale K., Johnson K., Burdige D., Kilgore T., Colodner D., Chavez F., Kudela R. and Boucher J. (2003) A time series of benthic flux measurements from Monterey Bay, CA. Cont. Shelf Res. 23, 457–481. Bergquist B. A. and Boyle E. A. (2006a) Dissolved iron in the tropical and subtropical Atlantic Ocean. Global Biogeochem. Cycles, 20, GB1015. Bergquist B. A. and Boyle E. A. (2006b) Iron isotopes in the Amazon River system: weathering and transport signatures. Earth Planet. Sci. Lett. 248, 54–68. Bergquist B. A., Wu J. and Boyle E. A. (2007) Variability in oceanic dissolved iron is dominated by the colloidal fraction. Geochim. Cosmochim. Acta 71, 2960–2974. Bostrom K., Peterson M. N. A., Joensuu O. and Fisher D. E. (1969) Aluminium-poor ferromanganoan sediments on active oceanic ridges. J. Geophys. Res. 74, 3261–3270. Boyd P. W., Watson A. J., Law C. S., Abraham E. R., Trull T., Murdoch R., Bakker D. C. E., Bowie A. R., Buesseler K. O., Chang H., Charette M., Croot P., Downing K., Frew R., Gall M., Hadfield M., Hall J., Harvey M., Jameson G., LaRoche J., Liddicoat M., Ling R., Maldonado M. T., McKay R. M., Nodder S., Pickmere S., Pridmore R., Rintoul S., Safi K., Sutton P., Strzepek R., Tanneberger K., Turner S., Waite A. and Zeldis J. (2000) A mesoscale phytoplankton bloom in the polar Southern Ocean stimulated by iron fertilization. Nature 407, 695–702. Boyle E. (1997) What controls dissolved iron concentrations in the world ocean? – a comment. Mar. Chem. 57, 163–167. Boyle E. A., Bergquist B. A., Kayser R. A. and Mahowald N. (2005) Iron, manganese, and lead at Hawaii Ocean Time-series station ALOHA: temporal variability and an intermediate water hydrothermal plume. Geochim. Cosmochim. Acta 69, 5165–5166. Brantley S. L., Liermann L. J., Guynn R. L., Anbar A., Icopini G. A. and Barling J. (2004) Fe isotopic fractionation during mineral dissolution with and without bacteria. Geochim. Cosmochim. Acta 68, 3189–3204. Bruland K. W., Franks R. P., Knauer G. A. and Martin J. H. (1979) Sampling and analytical methods for the determination of copper, cadmium, zinc, and nickel at the nanogram per liter level in sea-water. Anal. Chim. Acta 105, 233–245. Bullen T. D., White A. F., Childs C. W., Vivit D. V. and Schulz M. S. (2001) Demonstration of significant abiotic iron isotope fractionation in nature. Geology 29, 699–702. Butler I. B., Archer C., Vance D., Oldroyd A. and Rickard D. (2005) Fe isotope fractionation on FeS formation in ambient aqueous solution. Earth Planet. Sci. Lett. 236, 430–442. Christian J. R., Verschell M. A., Murtugudde R., Busalacchi A. J. and McClain C. R. (2002) Biogeochemical modelling of the tropical Pacific Ocean. II: Iron biogeochemistry. Deep-Sea Res. II – Topical Stud. Oceanogr. 49, 545–565. Chu N. C., Johnson C. M., Beard B. L., German C. R., Nesbitt R. W., Frank M., Bohn M., Kubik P. W., Usui A. and Graham I.

5632

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

(2006) Evidence for hydrothermal venting in Fe isotope compositions of the deep Pacific Ocean through time. Earth Planet. Sci. Lett. 245, 202–217. Clayton R. E., Hudson-Edwards K. A., Malinovsky D. and Andersson P. (2005) Fe isotope fractionation during the precipitation of ferrihydrite and transformation of ferrihydrite to goethite. Mineral. Mag. 69, 667–676. Croal L. R., Johnson C. M., Beard B. L. and Newman D. K. (2004) Iron isotope fractionation by Fe(II)-oxidizing photoautotrophic bacteria. Geochim. Cosmochim. Acta 68, 1227– 1242. Cullen J. T., Bergquist B. A. and Moffett J. W. (2006) Thermodynamic characterization of the partitioning of iron between soluble and colloidal species in the Atlantic Ocean. Mar. Chem. 98, 295–303. Dauphas N. and Rouxel O. (2006) Mass spectrometry and natural variations of iron isotopes. Mass Spectrom. Rev. 25, 831–832. de Baar H. J. W. and de Jong J. T. M. (2001) Distributions, sources and sinks of iron in seawater. In The Biogeochemistry of Iron in Seawater (eds. D. R. Turner and K. A. Hunter). John Wiley & Sons Ltd., Chichester. Dideriksen K., Baker J. A. and Stipp S. L. S. (2008) Equilibrium Fe isotope fractionation between inorganic aqueous Fe(III) and the siderophore complex, Fe(III)–desferrioxamine B. Earth Planet. Sci. Lett. 269, 280–290. Douville E., Charlou J. L., Oelkers E. H., Bienvenu P., Colon C. F. J., Donval J. P., Fouquet Y., Prieur D. and Appriou P. (2002) The Rainbow vent fluids (36 degrees 140 N, MAR): the influence of ultramafic rocks and phase separation on trace metal content in Mid-Atlantic Ridge hydrothermal fluids. Chem. Geol. 184, 37–48. Edmonds H. N. and German C. R. (2004) Particle geochemistry in the Rainbow hydrothermal plume, Mid-Atlantic Ridge. Geochim. Cosmochim. Acta 68, 759–772. Edwards K. J., Bach W., McCollom T. M. and Rogers D. R. (2004) Neutrophilic iron-oxidizing bacteria in the ocean: their habitats, diversity, and roles in mineral deposition, rock alteration, and biomass production in the deep-sea. Geomicrobiol. J. 21, 393– 404. Elrod V. A., Berelson W. M., Coale K. H. and Johnson K. S. (2004) The flux of iron from continental shelf sediments: a missing source for global budgets. Geophys. Res. Lett. 31, L12307. Escoube R., Rouxel O. J., Sholkovitz E. and Donard O. F. X. (2009) Iron isotope systematics in estuaries: the of North River, Massachusetts (USA). Geochim. Cosmochim. Acta 73, 4045– 4059. Feely R. A., Trefry J. H., Massoth G. J. and Metz S. (1991) A comparison of the scavenging of phosphorus and arsenic from seawater by hydrothermal iron oxyhydroxides in the Atlantic and Pacific Oceans. Deep-Sea Res. A – Oceanogr. Res. Pap. 38, 617–623. Feely R. A., Baker E. T., Marumo K., Urabe T., Ishibashi J., Gendron J., Lebon G. T. and Okamura K. (1996) Hydrothermal plume particles and dissolved phosphate over the superfastspreading southern East Pacific Rise. Geochim. Cosmochim. Acta 60, 2297–2323. Feely R. A., Trefry J. H., Lebon G. T. and German C. R. (1998) The relationship between P/Fe and V/Fe ratios in hydrothermal precipitates and dissolved phosphate in seawater. Geophys. Res. Lett. 25, 2253–2256. Fornari D. J., Tivey M. K., Schouten H., Perfit M. R., Yoerger D. R., Bradley A. M., Edwards M. H., Haymon R., Scheirer D., Von Damm K. L., Shank T. M. and Soule S. A. (2004) A submarine lava flow emplacement at the East Pacific Rise 9 degrees 500 N: implications for uppermost ocean crust stratig-

raphy and hydrothermal fluid circulation. In Mid-Ocean Ridges: Hydrothermal Interactions between the Lithosphere and Ocean (eds. C. R. German, J. Lin and L. M. Parson). AGU Monograph. Am. Geophys. Union, Washington, DC. Garbe-Scho¨nberg D., Koschinsky A., Ratmeyer V., Jahmlich H. and Westernstroer U. (2006) KIPS – a new multiport valvebased all-teflon fluid sampling system for ROVs. Geophys. Res. Abstr. 8, 07032. German C. R. and Von Damm K. L. (2004) Hydrothermal Processes. In The Oceans and Marine Geochemistry. Treatise on Geochemistry (ed. H. Elderfield). Elsevier-Pergamon, Oxford. German C. R., Campbell A. C. and Edmond J. M. (1991) Hydrothermal scavenging at the Mid-Atlantic Ridge: modification of trace element dissolved fluxes. Earth Planet. Sci. Lett. 107, 101–114. German C. R., Bennett S. A., Connelly D. P., Evans A. J., Murton B. J., Parson L. M., Prien R. D., Ramirez-Llodra E. Z., Jakuba M., Shank T. M., Yoerger D. R., Walker S. L., Baker E. T. and Nakamura K. (2008) Hydrothermal activity on the southern Mid-Atlantic Ridge: tectonically and volcanically controlled venting at 4–5 degrees S. Earth Planet. Sci. Lett. 273, 332–344. Govindaraju K. (1994) Compilation of working values and sample description for 383 geostandards. Geostand. Newslett. 18, 1– 158. Haase K. M., Petersen S., Koschinsky A., Seifert R., Devey C. W., Keir R., Lackschewitz K. S., Melchert B., Perner M., Schmale O., Suling J., Dubilier N., Zielinski F., Fretzdorff S., GarbeSchonberg D., Westernstroer U., German C. R., Shank T. M., Yoerger D., Giere O., Kuever J., Marbler H., Mawick J., Mertens C., Stober U., Ostertag-Henning C., Paulick H., Peters M., Strauss H., Sander S., Stecher J., Warmuth M. and Weber S. (2007) Young volcanism and related hydrothermal activity at 5 degrees S on the slow-spreading southern Mid-Atlantic Ridge. Geochem. Geophys. Geosyst. 8, Q11002. Homann V. V., Sandy M., Tincu, J. A., Templeton A. S., Tebo B. M., Butler A. (2009) Loihichelins A#F, a suite of amphophilic siderophores produced by the marine bacterium Halomonas LOB-5. J. Nat. Prod. Icopini G. A., Anbar A. D., Ruebush S. S., Tien M. and Brantley S. L. (2004) Iron isotope fractionation during microbial reduction of iron: the importance of adsorption. Geology 32, 205–208. Jickells T. D. and Spokes L. J. (2001) Atmospheric iron inputs to the oceans. In The Biogeochemistry of Iron in Seawater (eds. D. R. Turner and K. A. Hunter). John Willey & Sons Ltd., Chichester. Jickells T. D., An Z. S., Andersen K. K., Baker A. R., Bergametti G., Brooks N., Cao J. J., Boyd P. W., Duce R. A., Hunter K. A., Kawahata H., Kubilay N., LaRoche J., Liss P. S., Mahowald N., Prospero J. M., Ridgwell A. J., Tegen I. and Torres R. (2005) Global iron connections between desert dust, ocean biogeochemistry, and climate. Science 308, 67–71. Johnson K. S., Gordon R. M. and Coale K. H. (1997) What controls dissolved iron concentrations in the world ocean? Mar. Chem. 57, 137–161. Johnson K. S., Chavez F. P. and Friederich G. E. (1999) Continental-shelf sediment as a primary source of iron for coastal phytoplankton. Nature 398, 697–700. Johnson C. M., Roden E. E., Newman D. K. and Nealson K. H. (2004) Isotopic constraints on the biogeochemical cycling of Fe. Rev. Miner. Geochem. 55, 359–408. Koschinsky A., Devey C., Garbe-Scho¨nberg D., German C. R., Yoerger, D. R. and Shank T. M. (2006) Hydrothermal exploration of the Mid-Atlantic Ridge, 5–10°S, using the AUV ABE and the ROV quest – a brief overview of RV

Fe isotopes in a buoyant hydrothermal plume, 5°S MAR Meteor Cruise M68/1. EOS, Transactions, American Geophysical Union, vol. 87. Fall Meet, Suppl., OS34A-05 (abstr.). Koschinsky A., Garbe-Scho¨nberg D., Sander S., Schmidt K., Gennerich H. and Strauss H. (2008) Hydrothermal venting at pressure–temperature conditions above the critical point of seawater, 5°S on the Mid-Atlantic Ridge. Geology 36, 615–618. Lacan F., Radic A., Jeandel C., Poitrasson F., Sarthou G., Pradoux C. and Freydier R. (2008). Measurement of the isotopic composition of dissolved iron in the open ocean. Geophys. Res. Lett., 35. doi: 10.1029/2008gl035841. Lam P. J., Bishop J. K. B., Henning C. C., Marcus M. A., Waychunas G. A. and Fung I. Y. (2006). Wintertime phytoplankton bloom in the subarctic Pacific supported by continental margin iron. Global Biogeochem. Cycles, 20. doi: 10.1029/2005gb002557. Lang S. Q., Butterfield D. A., Lilley M. D., Johnson H. P. and Hedges J. I. (2006) Dissolved organic carbon in ridge-axis and ridge-flank hydrothermal systems. Geochim. Cosmochim. Acta 70, 3830–3842. Lefevre N. and Watson A. J. (1999) Modeling the geochemical cycle of iron in the oceans and its impact on atmospheric CO2 concentrations. Global Biogeochem. Cycles 13, 727–736. Levasseur S., Frank M., Hein J. R. and Halliday A. (2004) The global variation in the iron isotope composition of marine hydrogenetic ferromanganese deposits: implications for seawater chemistry? Earth Planet. Sci. Lett. 224, 91–105. Malinovsky D., Stenberg A., Rodushkin I., Andren H., Ingri J., Ohlander B. and Baxter D. C. (2003) Performance of high resolution MC-ICP-MS for Fe isotope ratio measurements in sedimentary geological materials. J. Anal. Atomic Spectrom. 18, 687–695. Martin J. H. (1990) Glacial–interglacial CO2 change: the iron hypothesis. Paleoceanography 5, 1–13. Martin J. H., Coale K. H., Johnson K. S., Fitzwater S. E., Gordon R. M., Tanner S. J., Hunter C. N., Elrod V. A., Nowicki J. L., Coley T. L., Barber R. T., Lindley S., Watson A. J., Vanscoy K., Law C. S., Liddicoat M. I., Ling R., Stanton T., Stockel J., Collins C., Anderson A., Bidigare R., Ondrusek M., Latasa M., Millero F. J., Lee K., Yao W., Zhang J. Z., Friederich G., Sakamoto C., Chavez F., Buck K., Kolber Z., Greene R., Falkowski P., Chisholm S. W., Hoge F., Swift R., Yungel J., Turner S., Nightingale P., Hatton A., Liss P. and Tindale N. W. (1994) Testing the iron hypothesis in ecosystems of the equatorial Pacific Ocean. Nature 371, 123–129. Matthews A., Emmanuel S., Levi L., Gvirtzman H. and Erel Y. (2008) Kinetic fractionation of Fe isotopes during transport through a porous quartz-sand column. Geochim. Cosmochim. Acta 72, 5908–5919. Mikutta C., Wiederhold J. G., Cirpka O. A., Hofstetter T. B., Bourdon B. and Von Gunten U. (2009) Iron isotope fractionation and atom exchange during sorption of ferrous iron to mineral surfaces. Geochim. Cosmochim. Acta 73, 1795– 1812. Moore J. K., Doney S. C., Glover D. M. and Fung I. Y. (2002) Iron cycling and nutrient-limitation patterns in surface waters of the World Ocean. Deep-Sea Res. II – Topical Stud. Oceanogr. 49, 463–507. Nishioka J., Ono T., Saito H., Nakatsuka T., Takeda S., Yoshimura T., Suzuki K., Kuma K., Nakabayashi S., Tsumune D., Mitsudera H., Johnson W. K. and Tsuda, A. (2007) Iron supply to the western subarctic Pacific: importance of iron export from the Sea of Okhotsk. J. Geophys. Res. – Oceans, 112. doi: 10.1029/2006jc004055. Poitrasson F. and Freydier R. (2005) Heavy iron isotope composition of granites determined by high resolution MC-ICP-MS. Chem. Geol. 222, 132–147.

5633

Polyakov V. B., Clayton R. N., Horita J. and Mineev S. D. (2007) Equilibrium iron isotope fractionation factors of minerals: reevaluation from the data of nuclear inelastic resonant X-ray scattering and Mossbauer spectroscopy. Geochim. Cosmochim. Acta 71, 3833–3846. Rouxel O. (2009) Natural variations of Fe isotopic compositions of seawater determined by MC-ICPMS. European Geosciences Union General Assembly 2009, Geophys. Res. Abstr., 11, EGU2009-10517. Rouxel O., Dobbek N., Ludden J. and Fouquet Y. (2003) Iron isotope fractionation during oceanic crust alteration. Chem. Geol. 202, 155–182. Rouxel O., Fouquet Y. and Ludden J. N. (2004) Subsurface processes at the lucky strike hydrothermal field, Mid-Atlantic Ridge: evidence from sulfur, selenium, and iron isotopes. Geochim. Cosmochim. Acta 68, 2295–2311. Rouxel O., Shanks W. C., Bach W. and Edwards K. J. (2008) Integrated Fe and S isotope study of seafloor hydrothermal vents at East Pacific Rise 9–10 degrees N. Chem. Geol. 252, 214– 227. Rudnicki M. D. and Elderfield H. (1993) A chemical-model of the buoyant and neutrally buoyant plume above the TAG vent field, 26 degrees-N, Mid-Atlantic Ridge. Geochim. Cosmochim. Acta 57, 2939–2957. Sander S. G., Koschinsky A., Massoth G., Stott M. and Hunter K. A. (2007) Organic complexation of copper in deep-sea hydrothermal vent systems. Environ. Chem. 4, 81–89. Schoenberg R. and von Blanckenburg F. (2005) An assessment of the accuracy of stable Fe isotope ratio measurements on samples with organic and inorganic matrices by high-resolution multicollector ICP-MS. Int. J. Mass Spectrom. 242, 257– 272. Severmann S., Johnson C. M., Beard B. L., German C. R., Edmonds H. N., Chiba H. and Green D. R. H. (2004) The effect of plume processes on the Fe isotope composition of hydrothermally derived Fe in the deep ocean as inferred from the Rainbow vent site, Mid-Atlantic Ridge, 36 140 N. Earth Planet. Sci. Lett. 225, 63–76. Sharma M., Polizzotto M. and Anbar A. D. (2001) Iron isotopes in hot springs along the Juan de Fuca Ridge. Earth Planet. Sci. Lett. 194, 39–51. Skulan J. L., Beard B. L. and Johnson C. M. (2002) Kinetic and equilibrium Fe isotope fractionation between aqueous Fe(III) and hematite. Geochim. Cosmochim. Acta 66, 2995– 3015. Soule S. A., Fornari D. J., Perfit M. R. and Rubin K. H. (2007) New insights into mid-ocean ridge volcanic processes from the 2005–2006 eruption of the East Pacific Rise, 9 degrees 460 N-9 degrees 560 N. Geology 35, 1079–1082. Statham P. J. (1985) The determination of dissolved manganese and cadmium in sea-water at low nmol/l concentrations by chelation and extraction followed by electrothermal atomic-absorption spectrometry. Anal. Chim. Acta 169, 149– 159. Teutsch N., von Gunten U., Porcelli D., Cirpka O. A. and Halliday A. N. (2005) Adsorption as a cause for iron isotope fractionation in reduced groundwater. Geochim. Cosmochim. Acta 69, 4175–4185. Toner B. M., Fakra S. C., Manganini S. J., Santelli C. M., Marcus M. A., Moffett J. W., Rouxel O., German C. R. and Edwards K. J. (2009) Deep-sea hydrothermal iron(II) in stable association with organic carbon in plume particles. Nat. Geosci. 2, 197– 201. Trefry J. H. and Metz S. (1989) Role of hydrothermal precipitates in the geochemical cycling of vanadium. Nature 342, 531–533.

5634

S.A. Bennett et al. / Geochimica et Cosmochimica Acta 73 (2009) 5619–5634

Trocine R. P. and Trefry J. H. (1988) Distribution and chemistry of suspended particles from an active hydrothermal vent site on the Mid-Atlantic Ridge at 26-degrees-N. Earth Planet. Sci. Lett. 88, 1–15. Turner J. S. (1962) The ‘starting plume’ in neutral surroundings. J. Fluid Mech. 13, 356–368. Turner D. R., Hunter K. A. and De Baar H. J. W. (2001) Iron Limitation in the oceans. In The Biogeochemistry of Iron in Seawater (eds. D. R. Turner and K. A. Hunter). John Wiley & Sons Ltd.. Von Damm K. L. (1995) Controls on the chemistry and temporal variability of seafloor hydrothermal fluids. In Seafloor Hydrothermal Systems: Physical, Chemical, Biological and Geological Interactions (eds. S. E. Humphris, R. A. Zierenberg, L. S. Mullineaux and R. E. Thomson). AGU Monograph Am. Geophys. Union, Washington, DC. Welch S. A., Beard B. L., Johnson C. M. and Braterman P. S. (2003) Kinetic and equilibrium Fe isotope fractionation between aqueous Fe(II) and Fe(III). Geochim. Cosmochim. Acta 67, 4231–4250.

Wells M. L., Price N. M. and Bruland K. W. (1995) Iron chemistry in seawater and its relationship to phytoplankton: a workshop report. Mar. Chem. 48, 157–182. Windom H. L., Moore W. S., Niencheski L. F. H. and Jahrike R. A. (2006) Submarine groundwater discharge: a large, previously unrecognized source of dissolved iron to the South Atlantic Ocean. Mar. Chem. 102, 252–266. Wu J. F., Boyle E., Sunda W. and Wen L. S. (2001) Soluble and colloidal iron in the oligotrophic North Atlantic and North Pacific. Science 293, 847–849. Yang Y., Rouxel O., Shi X. and Liu J. (2006) Fe isotope composition of Fe-Mn crusts in Pacific and its significance for paleoceanography. European Geosciences Union General Assembly 2008. Geophys. Res. Abstr. Zhu X. K., O’Nions R. K., Guo Y. L. and Reynolds B. C. (2000) Secular variation of iron isotopes in North Atlantic deep water. Science 287, 2000–2002.

Associate editor: Mark Rehkamper