Geochemical energy sources for microbial primary production in the ...

Available online at www.sciencedirect.com
Marine Chemistry 108 (2008) 18–31
www.elsevier.com/locate/marchem
Geochemical energy sources for microbial primary production
in the environment of hydrothermal vent shrimps
Caroline Schmidt a,b , Renaud Vuillemin c , Christian Le Gall b ,
Françoise Gaill a , Nadine Le Bris b, ⁎
a Université Pierre et Marie Curie Paris 6, UMR CNRS 7138 SAE, 7 quai Saint Bernard, 75252 Paris Cedex05, France
b Département EEP, IFREMER, BP 70, 29280 Plouzané, France
c Département TSI, IFREMER, BP 70, 29280 Plouzané, France
Received 1 December 2006; received in revised form 5 July 2007; accepted 17 September 2007
Available online 29 September 2007
Abstract
At deep-sea hydrothermal vents, dense invertebrate communities prevail along chemoclines where the relaxation of redoxdisequilibria
sustains chemolithoautotrophic microbial CO 2 -fixation. At the Mid-Atlantic Ridge, swarms of thousands of Rimicaris
exoculata shrimps thus assemble along the turbulent mixing interface between the hydrothermal fluid and oxygenated seawater. It
was suggested that this environment provides ideal conditions for growth to the abundant chemosynthetic microbial epiflora that
colonizes the shrimps' branchial cavity. Sulfide has long been considered as the prime electron donor used by the epibionts but, the
oxidation of iron has recently been hypothesized as an alternative pathway for the iron-rich Rainbow site. In order to examine the
potential energy sources for microbial primary production within the swarms at Rainbow, the chemical conditions along the mixing
gradient have been modeled from field data. This model provides a basis for the quantitative comparison of energy-budgets
available for chemolithoautotrophic primary production based on different oxidative pathways (e.g.: oxidation of sulfide-iron IImethane
and hydrogen by oxygen). A comparison was proposed for TAG, another hydrothermal vent field at the mid-Atlantic
Ridge which is characterized by the presence of similar swarms. Although the narrow temperature range of the shrimp environment
is similar at both sites, their chemically contrasted environments suggest different metabolic pathways would benefit from the
highest energy budgets. While sulfide oxidation is confirmed to be the energetically most favorable pathway at TAG, an original
biogeochemical context is suggested for Rainbow. Here, the highest energy could be derived from iron oxidation. At this site, the
oxidation of hydrogen possibly constitutes another dominant energy source, but this hypothesis still needs to be constrained by
kinetic studies. Methane and sulfide appears as minor energy sources in the environment of shrimps. A wider and original diversity
of the metabolic pathways involved in the microbial epibiosis can be expected at Rainbow in comparison to TAG.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Chemolithoautotrophy; Iron oxidation; Methane; Sulfide; Rimicaris; Rainbow
1. Introduction
⁎ Corresponding author. Tel.: +33 298 22 40 85; fax: +33 298 22 47 57.
E-mail address: nlebris@ifremer.fr (N. Le Bris).
At mid-ocean ridges, hydrothermal venting on the
seafloor creates sharp chemical gradients along centimeter
to decimeter-scales, with opposite trends in
reducing (e.g. sulfide, methane, ferrous iron) and
0304-4203/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.marchem.2007.09.009

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
19
oxidizing species (oxygen, nitrate) (Johnson et al., 1986;
Le Bris et al., 2000; 2006). The dynamic mixing of
hydrothermal fluid and seawater, and the relatively slow
kinetics of redox reactions enable the chemical
compounds to coexist in metastable conditions in the
interfacial zone. The chemical energy released by the
relaxation of redox-disequilibria fuels microbial CO 2 -
fixation which in turn sustains invertebrate communities
that prevail along these chemoclines (Jannasch and
Mottle, 1985). Thermodynamic properties hence place
critical constraints on productivity and distribution of
biological communities. The maximum amount of
geochemical energy that can be used to convert CO 2
into biomass from a redox couple depends on both, the
availability of the corresponding electron donor and
acceptor, and the energy yield of the chemical reaction.
Several studies have focused on the assessment of these
bioenergetic aspects in various hydrothermal environments,
ranging from black smoker plumes to the deepsubseafloor
habitats on ridge flanks (McCollom and
Shock, 1997; McCollom, 2000; Shock and Holland,
2004; Bach and Edwards, 2003).
To date, energy budget calculations have not been
considered in the environment of vent fauna, despite
they constitute the highest biomass in vent ecosystems.
The oxidation of sulfide and methane by oxygen has
been described as main energy-acquisition pathways
driving chemolithoautotrophic metabolisms in deepsea
hydrothermal vent environments. Physiological
and molecular biology studies have shown that thriving
populations of tubeworms and bivalves are sustained
by symbiotic interaction with sulfide-and methaneoxidizing
autotrophic bacteria (Childress and Fisher,
1992; Cavanaugh et al., 2006). Recently, the existence
of hydrogen-based symbioses has been suggested
(Zielinskietal.,2005). Bacterial cultivation was more
successful for free living vent microbes than for symbionts
(e.g. Takaietal.,2005). These studies confirmed
the importance of sulfide as a main electron donor for
autotrophic carbon fixation. Methane and, more recently,
hydrogen have also been identified as potential
substrates for microbial primary producers (Takaietal.,
2004). Although, microbial oxidation of ferrous iron is
known as a major biogeochemical pathway (Cornell
and Schwertmann, 2003), it has long been considered
to be energetically unfavorable for chemosynthetic
growth. Recent studies suggest that they could play a
significantroleinsomedeep-seahydrothermalenvironments.
Strains of chemoautotrophic iron-oxidizing
bacteria were isolated and cultured from dense filamentous
mats of iron oxides at Loihi seamount
(Emerson and Moyer, 2002) aswellasfromaltered
iron sulfide minerals on ridges flanks (Edwards et al.,
2003). These strains were shown to grow in microaerophilic
conditions at ambient temperature for which
they are able to compete with abiotic iron oxidation.
The vent shrimp Rimicaris exoculata, that forms
swarms of up to thousands individuals per square meter
at the mid-Atlantic Ridge hydrothermal vent structures,
constitutes another example of highly productive
chemosynthetically sustained communities. Part of the
abundant microbes that colonize the shrimps' branchial
cavity were shown to fix carbon chemolithoautotrophically
(Wirsen et al., 1993). Gebruk et al. (2000) suggested
that the localization of the shrimps in the mixing
zone could provide ideal conditions for the growth of
these microbes. Sulfide has long been considered as the
sole electron donor used by the shrimp epibionts,
according to the observation of elemental sulfur
associated with individuals from the TAG site (Gebruk
et al., 2000). A first phylogenetic analysis of Rimicaris
samples from Snake Pit emphasized a single ɛ-
Proteobacteria phylotype for this hypothesized sulfidebased
epibiose (Polz and Cavanaugh, 1995). Recently,
Zbinden et al. (2004) have shown that bacteria in the
branchial cavity were closely associated with ferrihydrite
potentially originating from biogenic oxidation of
ferrous iron (Gloter et al., 2004; Anderson et al., in
press). They proposed that iron oxidation may represent
Fig. 1. Temperature measurement and chemical analysis in the shrimp
swarm in close vicinity to individuals (1: ROV temperature probe; 2:
Alchimist inlets for dissolved sulfide and ferrous iron detection circuit;
3: pH electrode with protection).

20 C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
Fig. 2. Temperature measurement locations in the shrimp swarms at Rainbow (a) and TAG (b).
an alternative energy-pathway for chemosynthetic microbial
communities at Rainbow.
Our study is focused on the shrimp environment at
the Rainbow vent field, that has been investigated
during two submersible cruises in the past 6 years. The
hydrothermal end-member fluid composition is strongly
departing from other previously studied MAR sites
(Charlou et al., 2002). It is highly enriched in hydrogen,
methane, ferrous iron and relatively depleted in sulfide.
These particular features result from the location of the
Rainbow vent field on ultramafic rocks; in contrast to
the more extensively studied basalt-hosted sites on midocean
ridges. Main differences should be expected for
TAG that is hosted on basaltic rocks, where the shrimp
R. exoculata was first discovered and described by Rona
et al. (1986). In contrast to Rainbow, the black smoker
complexes at TAG expel fluids that are enriched in
sulfide and significantly lower in iron (Edmonds et al.,
1996). Although the data set characterizing the shrimp
environment is more limited for this site, a first
comparison with the obtained results for Rainbow is
presented.
Our aim was to address the two following questions:
(1) what are the energetically most favored autotrophic
CO 2 -fixation pathways in the shrimp environment at
Rainbow?, (2) should substantial difference be expected
between the Rainbow vent site and the basalt-hosted
Rimicaris habitat on the MAR? The thermal and
chemical variability of the shrimp environment has
been described from temperature measurements, in situ
chemical analysis and discrete fluid sampling. The
evolution of physico-chemical parameters as a function
of temperature in the hydrothermal fluid-seawater
mixing zone was modeled from these data. In comparison
to tubeworm communities fixed on seafloor
rocks that were previously studied (Le Bris et al., 2006),
the local source fuelling the shrimp habitat was hidden
behind mineral spires and therefore inaccessible for
sampling with the submersible manipulator. Although
the composition and temperature of the local source
remains unknown, the relative depletion of electron
donors and oxygen in the mix with respect to the endmember
and seawater contributions could be constrained
from the data set obtained in the shrimp
environment. On this basis, the energy-yield of various
oxidative pathways in the swarm environment at
Rainbow and TAG could be quantified and consequently
compared.
2. Materials and methods
2.1. Temperature measurements
Temperature data were obtained during the submersible
cruises ATOS (2001) at Rainbow and EXOMAR (2005) at
Rainbow and TAG. Temperature measurements were operated
from the ROV VICTOR6000 and controlled simultaneously
on board of the R/V Atalante. By means of the submersible
manipulator, the probes were positioned in close proximity to
the shrimps, avoiding to touch the wall of neighboring
chimneys. The probe remained at each sampling point for a
maximum of 6 min. All operations were registered and
followed by video control, using a deported camera (Bowtech

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
21
Table 1
Chemolithoautotrophic pathways in oxic deep-sea hydrothermal environments and the corresponding changes in standard Gibbs free energy (ΔG°)
Metabolic process Chemical reaction ΔG° 298K
kJ mol −1
Methanotrophy
(Jannasch, 1995)
Sulfide oxidation
(Jannasch, 1995)
Iron (II) oxidation
(Emerson et al., 2002;
Edwards et al., 2003)
Hydrogen oxidation (Jannasch, 1995; Takai
et al., 2005)
CH 4 +2O 2 →HCO 3 − +H + +H 2 O
HS − +2O 2 →SO 2− 4 +H +
HS − +1/2 O 2 +H + →S°+H 2 O
Fe 2+ +1/4 O 2 +3/2 H 2 O→FeOOH+2H +
H 2 +1/2 O 2 →H 2 O
All thermodynamic constants were calculated from the PHREEQC database, except for iron oxidation 2-line ferrihydrite data reported in Majzlan
et al. (2004) were used instead of the PHREEQC iron hydroxide constants.
−803
−790
−182
−26
−263
Ltd) attached to the probe as described in Le Bris et al. (2005)
(Fig. 1).
In 2001, temperature measurements were performed in two
shrimp swarms located in depressions on the side of two
smoker groups at Rainbow (position markers Iris 7 and Iris
13). A complementary chemical characterization was performed
for the swarm at Iris 13. In 2005, a swarm located at
exactly the same spot was investigated. To characterize the
spatial variation of temperature and dissolved chemical species
within a shrimp assemblage, measurements were done in the
center of the swarm and in bordering areas to unpopulated
chimney walls (Fig. 2a).
At TAG the chimneys were covered by huge shrimp
swarms. Temperature measurements were done in distinct
locations distributed within one of these swarm (Fig. 2b).
Several discharges of black smoke were observed on the flank
of the chimney complex.
2.2. In situ chemical analysis
In 2001 and 2005, the submersible flow analyzer Alchimist
(Le Bris et al., 2000) was installed on the Victor and used to
determine in situ dissolved ferrous iron and sulfide concentrations
by flow injection analysis (FIA). The FIA method
for iron analysis is described in Sarradin et al. (2005). The
fluid was drawn from the inlet positioned with the manipulator
to the colorimetric detection unit installed on the ROV.
The calibration of the device was performed in situ, before
and after each sampling run. The analyses were done with a
frequency of 60 s over a total sampling period of about 6 min
for each position. The primary aim of these measurements
was to correlate temperature and chemical parameters within
the steep gradients encountered in these media. As described
previously (Le Bris et al. 2001), the inlet was tightly attached
to the temperature probe of the Victor. At TAG, combined
Fe II and temperature measurements were obtained in different
location of the swarm, as done in other vent fauna
habitats (Le Bris et al., 2006). At Rainbow, one set of Fe II and
temperature measurements could be obtained in 2001 but the
inlet was accidentally detached from the temperature probe
preventing the direct combination of the data sets. Successive
measurements of temperature and iron at several points within
the swarm enabled to establish a coarse correlation between
these parameters. These measurements were repeated
during the EXOMAR cruise at Rainbow. However, due to
technical problems, measurements were obtained at only one
point within the swarm.
At both sites, Rainbow and TAG, the determination of the
S − II /temperature correlation could not be achieved due to
failure in the sulfide analyzer unit or the temperature probe.
Simultaneous analyses of Fe II and S − II could solely be
obtained in the surrounding of a swarm at Rainbow during the
ATOS cruise.
2.3. Fluid sampling and analysis
Discrete fluid samples were obtained in 2001 and 2005
using the VICTOR6000 pumping fluid multisampler consisting
of 19 evacuated 200 ml titanium bottles, which were
rinsed with 0.1N HCl and deionised water before operation.
In 2005, the ROV temperature probe was coupled to the
sampler inlet, enabling to measure temperature simultaneously
to the fluid sampling. All samples were obtained within the
swarms. In 2001, a few samples were additionally obtained
with gas-tight titanium bottles. For these samples a combined
autonomous probe enabled to set the average temperature at
the sampling point. Since the ROV sampler inlet was not
equipped with a temperature sensor in 2001, temperature
data are therefore not available for these samples. In 2001, a
punctual source of fluid in the immediate environment of
the shrimp assemblage was sampled as well with titanium
syringes.
The pH was measured with a conventional pH electrode
immediately after the sampling bottles were recovered on
board. For oxygen measurements, sub-samples were carefully
transferred in glass rod flasks avoiding the formation of gas
bubbles. In order to fix oxygen, Winkler reagents were immediately
added before subsequent titration (ATOS 2001).

22 C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
Table 2
Temperatures measured at different locations in a shrimp swarm at
Rainbow (upper) and TAG (lower) as displayed in Fig. 2 (SD: standard
deviation, N: number of data)
Sampling point T min [°C] T mean [°C] T max [°C] SD [°C] N
1 5.3 6.2 6.7 0.4 12
2 5.3 6.2 7.1 0.4 27
3 6.4 7.7 9 0.6 29
4 9.7 11.9 14.5 1.3 33
5 6.8 8.5 11.3 1 26
6 7.3 10.9 15.8 2.2 29
7 7.2 12.3 18.2 2.2 28
8 9.3 10.6 14.6 1.2 30
9 7.4 9.3 12.7 1.2 32
10 11.4 16 17.3 1.3 28
1 5.3 7.2 11.9 1.2 97
2 2.8 5.3 7.9 1.2 94
3 3.3 5.9 9.1 1.3 97
4 4 7.1 13.3 1.7 86
5 5.1 6.4 8.2 0.7 97
6 4 7.5 12.6 1.6 85
7 3.6 6 8.5 1.1 96
8 4.1 9 17.4 2.5 90
9 4 10.2 16.6 3.2 96
10 3.6 4.6 5.4 0.4 97
Alternatively, oxygen was directly measured using an
amperometric microelectrode (EXOMAR 2005). Fluid subsamples
were preserved in gas-tight flasks under the addition
of Hg 2 Cl 2 for CO 2 and CH 4 on-shore GC-analysis (ATOS
samples). Applied analytical methods are described in Sarradin
et al. (1998). Total iron was preserved in acidified sub-samples
and analyzed on-shore using the ferrozine method after
reduction of ferric iron with ascorbic acid.
2.4. Geochemical modeling
Geochemical calculations were performed applying the
computation code PHREEQC (version 2.8) (Parkhust and
Appelo, 1999) in order to simulate the chemical composition
of the mixing zone as function of temperature in the shrimp
environment. The program is based on chemical equilibria in
aqueous solutions interacting with minerals. The PHREEQC
database accounts for temperature effects on thermodynamic
constants but not for the influence of high pressure (about
23 MPa at 2300 m depth and 35 MPa at 3500 m depth). As
hypothesized in Le Bris et al. (2003) the effect of pressure
should be of minor importance for the investigated reactions
in the temperature range (2–30 °C). The thermodynamic
constants for the considered chemical species were defined
from the PHREEQC database, except for aqueous ferrous
iron-sulfide complexes. For these complexes, the database
has been upgraded using data from Rickard and Morse
(2005). According to these authors, only Fe(HS) + displays a
certain congruency among authors with a determined log K
close to 5.2. Additionally, soluble iron-sulfide molecular
clusters (FeSaq) are expected to dominate labile iron-sulfide
forms in hydrothermal fluids (Luther et al., 2001). According
to Theberge and Luther (1997), these complexes will form at
saturation of the medium with respect to precipitated ironmonosulfide
FeS. The Ks used in our calculation for this
precipitate corresponds to the value defined by Benning et al.
(2000). Full equilibration between the dissolved species and
the solid phase was not allowed, unless specified.
The progressive mixing of the end-member fluid with
seawater was modeled assuming that the chemistry of the
hydrothermal fluid at Rainbow corresponds to the composition
recorded in 1997 as presented in Charlou et al. (2002)
and Douville et al. (2002). The end-member composition
used for TAG refers to Edmonds et al. (1996). Theseendmembers
are termed ‘reference end-members’ in the following
sections. This definition does not infer a stable composition
of the end-members over years, but rather provides
a comparison basis from which the composition of the local
source fluid could be estimated. Fitting the model outputs to
empirical data enabled to quantify the depletion of Fe II and
S − II in the local source fluids with respect to the conservative
dilution of the reference end-members. The concentration
of methane in the mixing gradient was estimated by
similar adjustment to the field data.
Furthermore, different scenarios were considered to
account for oxygen consumption by electron donors which
could spontaneously occur in the mixing zone or in the
sampling bottles. For this purpose O 2 -pH and O 2 -T data-sets
Table 3
Compilation of thermal data (mean values and ranges) in the immediate
environment of the shrimp in different swarms at the Mid-Atlantic
Ridge (RB: Rainbow, SP: Snake Pit, L: Logatchev)
Site Location T [°C] Reference
Diverse Swarm 10–40 Gebruk et al.
(1993)
SP Swarm 10–15 (5–37) Segonzac et al.
(1993)
L Swarm N20 Gebruk et al.
(1993)
RB Swarm 9–25 Desbruyères et al.
(2000)
RB Swarm 13.2 Desbruyères et al.
(2001)
RB Swarm 11 (4.7–25) Geret et al. (2002)
RB Swarm 1 11.8 (3.9–16.6) this study
(ATOS 2001)
RB Swarm 2 8.7 (4.5–18.3) this study
(ATOS 2001)
RB Swarm 3 11.5 (3.2–18) this study
(EXOMAR 2005)
TAG Swarm 7 (2.8–17.4) this study
(EXOMAR 2005)

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
23
where ΔG T and ΔG T ° (J mol − 1 ) denote the change in Gibbs
free energy and the change in standard Gibbs free energy (all
species activities are set to 1 at standard state), respectively.
Πa i product /Πa i reactive represents the activity quotient of the
products and reagents of the chemical reaction. The temperature
dependence of ΔG° can be determined applying the
Van't Hoff law:
ln K T
K 298K
¼
DH-
R d 1
T
1
T 298K
ð2Þ
with K, the equilibrium constant, defined as ΔG T °=−RT ln K.
R represents the ideal gas constant (8.314 kJ mol − 1 ), ΔH° the
change in standard enthalpy of the reaction and T is temperature
in Kelvin.
The total amounts of chemical energy available from a
redox reaction, therefore potentially disposable for microbially
mediated processes, were determined by the following
relationship:
Fig. 3. Schematic representation of the shrimp environment. The
swarm position in the mixing zone between seawater and a local
source of the hydrothermal fluid on the flank of the chimney.
were compared to their modeled trends at equilibrium with
sulfide, ferrous iron and hydrogen. Variable contribution of
these electron donors in the source fluid were considered. The
200 ml mixing zone samples are composed of multiple mixing
fractions with highly diverse compositions and temperature.
pH is a more reliable tracer to asses the contribution of the
hydrothermal end-member fluid to such samples than temperature
at the sampling inlet which only reflects punctual
conditions. For this reason and considering the lack of temperature
data for most samples, pH was first chosen as the
correlation factor in order to constrain the model. Equilibrium
in samples after recovery on-board was modeled for a temperature
of 25 °C. To model the in situ oxygen variation
required to account for the real temperature of the mix and
for the electron donors that achieve equilibrium in situ with
oxygen.
For TAG, only the correlation of iron II concentration with
temperature could be constrained by means of in situ data. As
hydrogen is unlikely to be present in the shrimp environment
at this site, only sulfide was accounted to impact on the
oxygen-pH relation in samples. The sulfide content in the
mixing zone was estimated on this basis.
2.5. Energy budget calculation
Energy calculation were performed for the mild part of the
mixing zone (temperature below 30 °C). The energy release
per mole of reactants for a given reaction at temperature T is
determined by:
DG T ¼ DG T - þ RT ln
Pa product
i
=Pa reactive
i
ð1Þ
DGðtotÞ T
¼ DG T d½e
Š ð3Þ
ΔG(tot) T in kJ kg −1 represents the maximum energy amount
that can be liberated per mass unit of solution. It depends on
both, the Gibbs free energy ΔG T per mole [kJ mol − 1 ]and[e − ],
the concentration of electron donor that can be oxidized,
considering that this quantity is limited by the availability of
oxygen and the number of electron needed to oxidize one
compound (mol kg − 1 ). The reactions considered in this study
are the oxidation of methane, sulfide, hydrogen and ferrous
iron. Both, the complete oxidation of sulfide (as HS − )to
sulfate or partial oxidation to S° are investigated.
To quantify the energy budget available from ferrous iron
oxidation, the definition of thermodynamic properties of the
oxidized form of iron is of main importance. Two-line
ferrihydrite is the mineral that was shown to be formed
dominantly in association with the shrimp epibionts that are
harbored in the branchial cavity (Gloter et al., 2004). The
thermodynamic properties of this compound have only
recently been assessed in experimental conditions. Majzlan
et al. (2004) defined ΔG 298K ° ranging from 26 to 29 kJ mol − 1
and ΔH° of−48 to−47 kJ mol − 1 for the formation of 2-line
ferrihydrite, depending on the duration of precipitation process
(Table 1).
3. Results
3.1. Temperature ranges for the shrimp habitat
Temperature measurements performed in a shrimp swarm
at TAG and Rainbow in 2005 are summarized in Table 2 and
the corresponding sampling locations are shown in Fig. 2a, b.
The temperature ranged between 2.8–17.4 °C at TAG and
5.3–18.2 °C at Rainbow. At both sites, Rainbow and TAG,
the lowest temperature was recorded at the boundary of the
shrimp swarm and the unpopulated rock, while more elevated

24 C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
Iron distribution appears conservative in first approximation
in the studied environment, with a roughly linear
correlation with temperature. The observed trend still lies
significantly below the reference end-member dilution model.
This depletion reflects, either a reduction of the end-member
iron content, or more likely, iron removal by precipitation from
the secondary source fluid before its emission into seawater.
From the in situ iron II measurements performed in 2005, a
ratio of 51 μMFe II °C − 1 was established. This estimate depicts
an iron depletion in the local source fluid of about 23%, with
regard to the end-member dilution measured in 1997 (i.e.
18.4 mM instead of 24 mM when extrapolated to the end-
Fig. 4. (a) Iron as a function of temperature at Rainbow showing the
reference end-member dilution model (dotted line) and the model
defined from 2005 in situ data (solid line). Open circles: total Fe in
samples (2001), triangles: in situ Fe II measurements (2005), crosses:
in situ Fe II measurements (2001). (b) Sulfide as a function of iron at
Rainbow. Bold dotted line: reference end-member dilution model,
solid line: empirical model, thin dotted line: model allowing
equilibrium with FeS precipitate. Open circles: sample contents,
Close circles: Fe II in situ data (ATOS 2001).
temperatures were measured in the center of the swarms. A
temperature gradient was observed along a vertical axis
within the shrimp assemblages. Similar ranges and patterns
were obtained during the ATOS cruise (2001) for two different
swarms at the Rainbow site (Table 3). These features
and the turbulent currents of shimmering fluid observed
around the bottom of the swarm suggest a hot fluid emission
below the swarm as schematized in Fig. 3. The direct outflow
source was not observable, as it was masked by a group of
mineral spires.
3.2. Comparison of empirical data with the end-member
dilution model
3.2.1. Rainbow
Iron concentrations in the swarm environment have been
determined, both, in situ (Fe II alone) and from fluid sampling
(total labile Fe after reduction of Fe III with ascorbic acid). The
iron–temperature correlation for the data sets obtained in 2001
and 2005 is rather consistent (Fig. 4a). Furthermore, the total
Fe-temperature correlation in fluid samples does not significantly
depart from in situ data (Fig. 4a). This suggests that the
contribution of ferric iron to the overall in situ iron budget in
the shrimp surrounding is minor.
Fig. 5. (a) pH as a function of temperature, (b) CO 2 as a function of pH
for Rainbow samples. Solid line: without Fe II oxidation, line with
crosses: allowing Fe II oxidation in samples. Open squares: in the
shrimp habitat (2001), triangles: local hot fluid source (2001), close
squares: in the shrimp habitat (2005). (c) CH 4 as a function of pH for
Rainbow samples. Thin line: reference end-member dilution model,
solid line: empirically fitted model.

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
25
that 40% of the reference end-member methane content is
preserved in the source fluid (Fig. 5c).
Fig. 6. Iron as a function of temperature at TAG showing the reference
end-member dilution model (solid line) and the in situ measurements
(crosses).
member temperature). The sampling data from 2001 are in
agreement or slightly below the conservative dilution model
for this local source (Fig. 4a).
In situ sulfide measurements recorded parallel to iron II at
Rainbow show a linear correlation between these two parameters
(Fig. 4b). Although sulfide is likely to be associated with
iron in the shrimp environment where the Fe II :S − II ratio
exceeds 30, this linear correlation indicates that labile ironsulfide
compounds are present. The hypothesis of in situ
equilibrium with stable iron sulfide precipitates, i.e. mackinawite,
that do not dissociate in the analyzer manifold is not
supported by these results (Fig. 4b). As previously described in
hydrothermal environments (Luther et al. 2001), iron sulfide
complexes FeS° and FeS-colloids more likely dominate the in
situ speciation of sulfide. The Fe II :S − II ratio furthermore
reflects a sulfide content reaching only 40% of the corresponding
end-member contribution for a given mixing ratio
(i.e. 500 μM instead of 1.2 mM when extrapolated to the endmember
temperature). In contrast to in situ data, sulfide
appears completely depleted in discrete samples (Fig. 4b). The
hypothesis that crystalline FeS phases that would not be
detected by the Cline Method have formed during recovery
cannot be discarded. However, equilibrium with this phase
cannot explain the quantitative removal of sulfide from
samples (Fig. 4b). A more relevant explanation to the
undetectable or very low sulfide concentrations in the samples
is the complete oxidation of S − II during recovery.
Data points of both sampling sets (2001 and 2005) lie close
or slightly below the modeled pH–T correlation curve,
allowing sulfide oxidation at 25 °C (Fig. 5a). The deviation
observed (b0.4 pH) may reflect conductive cooling of the fluid
or, underestimation of the temperature corresponding to each
sample. Another possibility is that Fe II oxidation in samples
resulted in acidification of the medium. Part of the data from
2005 support this assumption. A better fit is indeed obtained
for them assuming that iron is in equilibrium with O 2 (Fig. 5a).
However, the CO 2 -pH correlation would then depart significantly
from the model which assumes that iron II does not
reach equilibrium with O 2 .CO 2 -pH data in 2001 samples
rather agree with the later hypothesis (Fig. 5b). The methane–
pH correlation was also consistent with this model assuming
3.2.2. TAG
The iron II concentration at TAG is consistent with the
conservative dilution of the reference end-member (Fig. 6).
Linear extrapolation of in situ ferrous iron to the reference
end-member temperature of 365 °C, 5160 μM, is fairly similar
to the end-member fluid content (5180 μM) as reported in
Edmonds et al. (1996).
Similar in situ sulfide data could not be obtained due to
analyzer failure. The S − II :Fe II end-member ratio of 0.6
determined by Edmonds et al. (1996) enables to set an upper
limit of the sulfide concentrations in the mixing zone. As
sulfide is the dominant electron donor in the mixing zone at
TAG, the oxygen content in the samples can be used to
constrain in situ sulfide levels. Assuming conservative dilution
of the end-member fluid with respect to sulfide (i.e. 100% endmember
total sulfide concentration contributes to the source
fluid) and total sulfide oxidation with oxygen in samples, will
significantly under-estimate the oxygen content in samples
(Fig. 7). This suggests that sulfide is depleted in the source
fluid with regards to Fe II , possibly caused by FeS 2 precipitation.
A better fit for the O 2 sample measurements is obtained
for a total sulfide concentration reaching about 20% of the endmember
contribution.
Fig. 7. Oxygen content as function of pH (a) and temperature (b) for
TAG showing the reference end-member dilution model (dotted line),
the model assuming that only 20% of the reference end-member sulfide
contributes to the mix (solid line) and the sample contents (crosses).

26 C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
content is preserved in the source fluid fuelling the mixing
zone.
Consequently, two cases can be considered: (1) H 2 is fully
oxidized in the mixing zone and the available O 2 is defined
by its equilibrium with H 2 (Fig. 9a), (2) H 2 does not oxidize
spontaneously in the mixing zone and its concentration is at a
maximum of 40% of the level expected from reference endmember
dilution (Fig. 9b).
The minimum in situ O 2 boundary in this last case is more
difficult to constrain since no redox couple is assumed to
control oxygen variations in the mixing zone. An arbitrary
upper limit of 100 °C was set considering that kinetically
inhibited reactions at low temperature would be much more
rapid at this temperature and that electron donors would be in
large excess.
A similar assumption was done for oxygen at TAG. The
oxygen trend was fit to the O 2 concentrations in the fluid
samples obtained during the EXOMAR Cruise in 2005.
Assuming 80% depletion of the end-member sulfide contribution
in the source fluid and its total oxidation in the mixing
zone sets the oxic–anoxic interface around 30 °C.
3.4. Chemical energy budgets
Fig. 8. O 2 as a function of pH (a) and temperature (b) in Rainbow
samples (open squares: 2001; filled squares: 2005). The model curves
reflect total oxidation of sulfide in samples and the oxidation of
hydrogen assuming variable contribution in the source fluid with
respect to the reference end-member: 0% (line with crosses), 20% (thin
line), 40% (bold line), 100% (dotted line). Filled squares: 2005 shrimp
habitat samples, open squares: 2001 shrimp habitat samples, triangles:
2001 local fluid source samples.
The energy available to chemolithoautotrophic organisms
per mole of electron donor was assessed by considering the
3.3. Potential controls on the oxygen content in samples
On the basis of these empirically estimated S − II and Fe II
contents in the Rainbow source fluids, different scenarios were
considered assuming that redox equilibrium is achieved with
oxygen for one or several of the most abundant reducing
species in Rainbow fluids, S − II ,Fe II or H 2 . Fig. 8 presents the
comparison of modeled O 2 -pH correlations with field data.
The conversion of sulfide into sulfate during sample recovery
is suggested by its depletion in samples but this process cannot
solely explain the O 2 depletion (Fig. 8a). Fe II /Fe III equilibrium
with O 2 /H 2 O is insufficient as well to account for the observed
oxygen trend. Additional scenarios accounted for the presence
of hydrogen in the source fluid. It is likely that hydrogen
would equilibrate with O 2 over several hours in the samples,
since their reaction rate is expected to be quite rapid. However,
if 100% of the hydrogen contained in the reference endmember
was preserved during mixing with seawater, the
oxygen concentration in the sample should be distinctly lower
than measured (Fig. 9). The O 2 -pH trend agrees much better
with the empirical data as long as only 20 to 40% of the endmember
hydrogen contribution is accounted in the model.
From these computations, it can be estimated that a maximum
hydrogen contribution of 40% of its reference end-member
Fig. 9. Variation of oxygen (line with crosses), ferrous iron (bold lines)
and hydrogen (bold dotted line) in the two scenarios considered for
energy budgets calculation at Rainbow: (a) as function of pH, (b) as
function of temperature.

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
27
change in Gibbs free energy (ΔG) in the mixing gradient
(Table 4). Although the formation of 2-line ferrihydrite at
neutral to alkaline pH results in an increase in the Gibbs free
energy of the reaction, the less energetic reaction on a molar
basis is the oxidation of iron. The conversion of HS − to
elemental sulfur is slightly more energetic. Comparatively, the
ΔG per mole of electron donor are almost one order of
magnitude higher for the oxidation of sulfide to sulfate or
methane oxidation to HCO 3 − . The oxidation of hydrogen
constitutes an intermediate with a ΔG being four times lower
than the oxidation of methane or sulfide to sulfate.
Multiplying these values with the equivalent quantity of
electron donor that can be oxidized per unit mass of solution
enables to assess the total amount of energy that can be
released for a given reaction (ΔG(tot) in kJ kg − 1 ). Consequently,
the variation of ΔG(tot) is important over the mixing
range, showing a progressive increase up to a maximum value
before decreasing due to oxygen limitation (Fig. 10). For the
Rainbow site hydrogen-related assumptions influence the
overall available energy-budget to a large extend. If H 2 is
considered to reach equilibrium with oxygen in the mixing
zone, the oxygen boundary would be slightly below 30 °C
(Fig. 9b). In these conditions, the maximum energy yield is
reached around 15 °C for ferrous iron oxidation. The available
energy budgets from the oxidation of methane or sulfide to
sulfate reach their maximum at around 20–24 °C. Still, they
remain half the energy available from iron. In the second
scenario, hydrogen is preserved in the mixing zone and the
oxygen level remains much higher. Under these conditions, the
largest energy can still be derived from the oxidation of iron II
and its maximum shifted to slightly higher temperature (about
20–22 °C if an oxygen boundary at 100 °C is assumed).
Hydrogen constitutes another substantial source of energy in
this case. The energy available from the oxidation of hydrogen
lies in the same order of magnitude, at least up to 20 °C. The
provided energy increases up to 25 °C where it reaches a
maximum. The energy budget for sulfide and methane
oxidation are significantly lower. They are not limited by the
availability of oxygen below 30 °C.
At TAG these energy maxima are shifted above 30 °C,
when the maximum oxygen limit in the mixing zone is
arbitrary set at 100 °C (Fig. 10c). If this boundary is moved
toward lower temperature, an energy maximum may be
reached below 30 °C. In any case, the conversion of sulfide to
sulfate would be the most energetic process. Iron oxidation
Table 4
Minimum and maximum ΔG values calculated for the mixing zone for
Rainbow and TAG
T
(°C)
FeII/
FeOOH
HS − /
2−
SO 4
CH4/
−
HCO 3
HS − /
S°
H 2 /
H 2 O
TAG 2.2 −69.5 −757.3 −786.5 −78.3 –
32 −69.1 −740.2 −770.0 −93.3 –
Rainbow 3.9 −77.8 −776.8 −796.2 −94.2 −225.9
32 −68.4 −746.1 −772.0 −95.7 −229.7
Fig. 10. Potential energy yield (per kilogram fluid) for the accounted
chemosynthetic processes in the mild part of the hydrothermal fluidmixing
zone as a function of temperature at Rainbow (1rst scenario:
hydrogen equilibrates with O 2 in the mixing zone (a), 2nd scenario:
hydrogen remains available in the mixing zone (b)), and TAG (c).
still constitutes a substantial energy source. Methane is not
enriched in TAG end-member and should therefore not
significantly contribute to the energy budget.
4. Discussion
4.1. Availability of electron donors and oxygen in the
mixing zone
The comparison of empirical data and modeled
trends provides quantitative information on the availability
of electron donors and oxygen with respect to the
conservative end-member dilution. At Rainbow, contents
of the electron donors, are only about 77% for
ferrous iron and 40% sulfide and methane with regard to
their reference end-member. At TAG, the distribution of

28 C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
iron II in the mixing zone may be conservative, but
sulfide appears depleted by about 80% of its endmember
value. These results suggest that the shrimp
environment is fuelled by a secondary fluid source from
which chemical compounds have been partly removed
by precipitation or microbial consumption as shown in
other hydrothermal environments (Le Bris et al., 2003;
Von Damm and Lilley, 2004).
The oxygen variation with pH at Rainbow substantially
departs from conservative mixing. These data
provided clues to identify the redox couple that could
reach equilibrium with O 2 /H 2 O, either in situ or during
the sample recovery process. As previously documented
(Le Bris et al., 2006), sulfide appears to be oxidized
during sample recovery. This result is consistent with
the fast sulfide oxidation rate in presence of Fe II (Zhang
and Millero, 1994). A similar assumption can be done
for hydrogen, which is a major constituent of Rainbow
fluids. Its presence in the source fluid is suggested by
oxygen content in the samples, which is lower than it
could be explained by sulfide oxidation. A maximum
amount of 40% hydrogen in the source fluid with
respect to the reference end-member (Charlou et al.,
2002) was estimated from these data, assuming that iron
oxidation is kinetically inhibited.
The presence of hydrogen in diffuse fluids has been
reported by Butterfield et al. (2004) and Von Damm and
Lilley (2004). It is not surprising that hydrogen could be
enriched in the secondary fluid source at Rainbow, as it
is highly enriched in the hydrothermal end-member
fluid. However, H 2 may be completely oxidized in the
oxic part of the mixing zone, as long as the reaction is
fast enough to be completed in this dynamic environment.
Alternatively, hydrogen may be maintained to a
substantial level if the renewal rate is high enough to
balance its fast oxidation. Two scenarios were proposed
on this basis. In order to draw conclusions on the
accuracy of either of them it would be required to
perform hydrogen in situ measurements. It is indeed
unlikely that H 2 could remain stable for hours in
oxygenated fluid samples. Therefore, hydrogen determination
from samples would always be afflicted with
imprecision. Comparison of in situ oxygen measurements
with discrete sample contents would provide
further clues to assess the presence of hydrogen in the
medium. Unfortunately these data are lacking to date.
4.2. Potential chemosynthetic energy pathways
In both scenarios, iron oxidation appears as a major
energy source in the mild part of the mixing zone at
Rainbow. This peculiar result is due to the exceptional
enrichment of iron in the local vent fluid. The enrichment
of hydrogen in the source fluid could have as well important
biogeochemical implications. In the first scenario,
assuming that hydrogen is fully oxidized in the
mixing zone, the O 2 availability will be limited to the low
temperature range (b25 °C). In the second scenario
hydrogen remains available in the whole mixing zone.
The energetical yield for hydrogen oxidizers would be in
the same range as for iron oxidizers below 18 °C and
slightly more above. This last scenario provides an upper
estimate of the energy available as the hydrogen and
oxygen contents where maximized in this assumption. In
comparison, sulfide and methane only appear as secondary
energy sources for the shrimp epibionts at Rainbow.
In contrast, at the TAG site, total sulfide oxidation to
sulfate appears as the dominant potential energy source
below 30 °C. In this thermal range, the energy that
would be available for chemosynthetic microbes using
iron II is much lower but still substantial. Methane is
only a minor component of TAG fluids and should not
constitute a significant electron donor for primary
production. The energy that could be supplied from
the oxidation of these reduced compounds in the shrimp
habitat at the two Mid-Atlantic Ridge vent sites thus
depict very different patterns.
4.3. Energetic and physico-chemical characteristics of
the shrimp habitat
Although the shrimp are highly mobile, they aggregate
in swarms of several thousand of individuals within
very sharp thermal limits. Similar temperature ranges
have been determined for two swarms at two locations at
the Rainbow site in 2001, and again in a swarm at one of
these location in 2005 (Table 3). Slightly lower data
were obtained for a swarm at TAG.
This result is generally consistent with the thermal
ranges reported in the literature for the R. exoculata
habitat, but it does not confirm the highest values
presented in some studies (Table 3). Results acquired at
Rainbow and at TAG show a certain homogeneity. This
suggests that the thermal range defined in this study may
be representative for Rimicaris swarms. Even though the
presence of shrimp at higher temperatures cannot be
ruled out, Ravaux et al. (2003) demonstrated that the
shrimp expresses a biochemical response to thermal
stress above 25 °C. We believe that the present study
provides a more accurate definition of the thermal
conditions in the Rimicaris habitat than early studies.
Temperature ranges were carefully assessed from large
data sets and followed by close-up video-control in order
to ensured the precise location of the probe tip. These

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
29
short-term studies however may not fully reflect the
temporal temperature variation pattern within a swarm.
Over these narrow ranges, the potential energy sources
available to the epibionts for the chemolithoautotrophic
CO 2 fixation is unambiguous. The conditions within the
shrimp habitat at Rainbow appear energetically optimal
for chemolithoautotrophic growth relying on iron II, and
potentially on hydrogen oxidation. At TAG, the oxidation
of sulfide provides the main energy budgets. These
bioenergetic considerations support previous studies that
have suggested iron oxidation as an important metabolic
pathway, based on the analysis of iron oxide deposits in
the branchial cavity of shrimp sampled at Rainbow
(Gloter et al., 2004; Zbinden et al. 2004). The fast abiotic
oxidation of iron in alkaline to neutral media at ambient
temperature (∼25 °C) is expected to limit the microbially
mediated conversion to low O 2 conditions. According to
the temperature dependence of the kinetic rate constant
(Millero et al. 1987), this effect should not be limiting in
the tepid Rimicaris habitat.
Although the presence of H 2 in the shrimp environment
at Rainbow remains speculative, it is interesting to
note that the available energy budget provided by
hydrogen oxidation could be in the same range as the
one liberated by the oxidation of iron. Reaction kinetics
is a determinant criteria that should be investigated in
order to further asses the importance of this electron
donor. The commonly described electron donors in
chemosynthetic ecosystems, i.e., sulfide and methane,
do not represent the predominant energy sources at
Rainbow. Still, their oxidation may constitute a possible
chemosynthetic pathway, as some energy can be derived,
although the yield is much lower than for iron and
hydrogen oxidation.
5. Conclusion
Estimations on the energy budget as a function of the
mixing ratio suggested that chemolithoautotrophic primary
producers may find optimal conditions for growth
in association with highly active shrimps that aggregate
in swarms in the hydrothermal fluid-seawater mixing
interface. This study confirmed that the chemical energy
sources that could be utilized by primary producers
associated with R. exoculata differ substantially between
Rainbow and TAG, even though they colonize
similar habitats. Although the shrimps were observed in
comparable narrow temperature ranges, they may harbor
a highly diversified microbial epiflora at different sites,
depending on the hydrothermal fluid chemistry. Consideration
based on thermodynamic constraints need to
be further supported by microbial studies since the most
energetic processes are not necessarily the one which
dominantly fuel the shrimp epibionts. Rimicaris shrimps
maintain themselves in a quite narrow environmental
range within the mixing zone. The electron donors that
are commonly described to fuel chemosynthetic growth,
like sulfide and methane, do not appear as the main
energy sources at Rainbow. Here, the oxidation of iron
yields the maximum energy in the mixing zone. A firstorder
calculation revealed that hydrogen could act as
well as a major electron donor for CO 2 -fixation in the
shrimp habitat, as long as its residence time is long
enough to allow microbial uptake. Field studies are still
required to provide a direct evidence of this hypothesis
and to estimate to what extend hydrogen will be preserved
in the mixing zone.
Acknowledgments
This work was financially supported by IFREMER,
University Pierre and Marie Curie-Paris 6, and the
European Community (PhD grant to C.S. / MOMARNET
RTN contract 2004-5050026). The authors would like
to particularly acknowledge the chief scientists of the
research cruises, Pierre-Marie Sarradin for ATOS and
Anne Godfroy for EXOMAR, the captains and crews of
the RV Atalante and the Victor 6000 operation group, as
well as instrumentation engineers and technicians for their
essential support at sea.
References
Anderson, L.M., Halary, S., Lechaire, J.-P., Boudier, T., Frebourg, G.,
Marco, S., Zbinden, M., Gaill, F., in press. Tomography of
bacteria-mineral associations within the deep sea hydrothermal
vent shrimp Rimicaris exoculata. C.R. Chimie.
Bach, W., Edwards, K.J., 2003. Iron and sulfide oxidation within the
basaltic ocean crust: Implifications for chemolithoautotrophic
microbial biomass production. Geochim. Cosmochim. Acta 67
(20), 3871–3887.
Benning, L.G., Wilkin, R.T., Barnes, H.L., 2000. Reaction pathways in
the Fe–S system below 100 degrees C. Chem. Geol. 167, 25–51.
Butterfield, D.A., Roe, K.K., Lilley, M.D., Huber, J.A., Baross, J.A.,
Embley, R.W., Massoth, G.J., 2004. Mixing, reaction and
microbial activiry in the sub-seafloor revealed by temporal and
spatial variation in diffuse flow vents at axial volcano. In: Wilock,
W.S.D., DeLong, E.F., Kelley, D.S., Baross, J.A., Cary, S.C.
(Eds.), The subseafloor biosphere at Mid-Ocean ridges. American
Geophysical Union, Washington, DC.
Cavanaugh, C.M., McKiness, Z., Newton, I.L.G., Stewart, F., 2006.
Marine chemosynthetic symbioses. In: Dworkin, M., Falkow, S.,
Rosenberg, E., Schleifer, K.-H., Stackebrandt, E. (Eds.), The
prokaryotes: A handbook on the biology of bacteria. Springer, Verlag.
Charlou, J.L., Donval, J.P., Fouquet, Y., Jean-Baptiste, P., Holm, N.,
2002. Geochemistry of high H 2 and CH 4 vent fluids issuing from
ultramafic rocks at the Rainbow hydrothermal field (36°14′N,
MAR). Chem. Geol. 191, 345–359.

30 C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
Childress, J.J., Fisher, C.R., 1992. The biology of hydrothermal vent
animals: physiology, biochemistry, and autotrophic symbioses.
Oceanogr. Mar. Biol. Ann. Rev. 30, 337–441.
Cornell, R.M., Schwertmann, U., 2003. The Iron oxides – Structure,
properties, reactions occurrences and uses, 2nd edition. Wiley-
VCH, Weinheim.
Desbruyères, D., Almeida, A., Biscoito, M., Comtet, T., Khripounoff,
A., Le Bris, N., Sarradin, P.M., Segonzac, M., 2000. A review of
the distribution of hydrothermal vent communities along the
northern Mid-Atlantic Ridge: dispersal vs. environmental controls
Hydrobiologia 440 (1–3), 201–216.
Desbruyères, D., Biscoito, M., Caprais, J.-C., Colaco, A., Comtet, T.,
Crassous, P., Fouquet, Y., Khripounoff, A., Le Bris, N., Olu, K.,
Riso, R., Sarrin, P.-M., Segonzac, M., Vangriesheim, A., 2001.
Variations in deep-sea hydrothermal vent communities on the Mid-
Atlantic Ridge near the Azores plateau. Deep-Sea Res. I 48,
1325–1346.
Douville, E., Charlou, J.L., Oelkers, E.H., Bienvenu, P., Jove Colon,
C.F., Donval, J.P., Fouquet, Y., Prieur, D., Appriou, P., 2002. The
rainbow vent f luids (36_14VN, 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., German, C.R., Green, D.R.H., Huh, Y., Gamo, T.,
Edmond, J.M., 1996. Continuation of the hydrothermal fluid
chemistry time series at TAG, and the effects of ODP drilling.
Geophys. Res. Lett. 23 (23), 3487–3489.
Edwards, K.J., Rogers, D.R., Wirsen, C.O., McCollom, T.M., 2003.
Isolation and characterization of novel psychrophilic, neutrophilic,
Fe-oxidizing, chemolithoautotrophic α-and γ-Proteobacteria from
the deep sea. Appl. Environ. Microb. 69 (5), 2906–2913.
Emerson, D., Moyer, C.L., 2002. Neutrophilic Fe-oxidizing bacteria
are abundant at the Loihi seamount hydrothermal vents and play a
major role in Fe oxide deposition. Appl. Environ. Microbiol. 68
(6), 3085–3093.
Gebruk, A.V., Pimenov, N.V., Savvichev, A.S., 1993. Feeding
specialization of bresiliid shrimps in the TAG site hydrothermal
community. Mar. Ecol., Prog. Ser. 98, 247–253.
Gebruk, A.V., Southward, E.C., Kennedy, H., Southward, A.J., 2000. Food
sources, behaviour, and distribution of hydrothermal vent shrimp at the
Mid-Atlantic-Ridge. J. Mar. Biol. Assoc. UK 80, 485–499.
Geret, F., Riso, R., Sarradin, P.M., Caprais, J.C., Cosson, P., 2002.
Metal bioaccumulation and storage forms in the shrimp, Rimicaris
exoculata, from the Rainbow hydrothermal field (Mid-Atlantic-
Ridge); preliminary approach to the fluid-organism relationship.
Cah. Biol. Mar. 43, 43–52.
Gloter, A., Zbinden, M., Guyot, F., Gaill, F., Colliex, C., 2004. TEM-
EELS study of natural ferrihydrite from geological-biological
interactions in hydrothermal systems. Earth Planet. Sci. Lett. 222,
947–957.
Jannasch, H.W., Mottle, M.J., 1985. Geomicrobiology of deep-sea
hydrothermal vents. Science 229, 717–725.
Jannasch, H.W., 1995. Life at the seafloor. Nature 374, 676–677.
Johnson, K.S., Beelher, C.L., Sakamoto-Arnold, C.M., Childress, J.J.,
1986. In situ measurements of chemical distributions in a deep-sea
hydrothermal vent field. Science 231, 1139–1141.
Le Bris, N., Sarradin, P.M., Birot, D., Alayse-Danet, A.M., 2000. A
new chemical analyzer for in-situ measurement of nitrate and total
sulfide over hydrothermal vent biological communities. Mar.
Chem. 72 (1), 1–15.
Le Bris, N., Sarradin, P.-M., Pennec, S., 2001. A new deep-sea probe
for in situ pH measurement in the environment of hydrothermal
vent biological communities. Deep-Sea Res. I. 48, 1941–1951.
Le Bris, N., Sarradin, P.M., Caprais, J.C., 2003. Contrasted sulphide
chemistries in the environment of 13°N EPR vent fauna. Deep-Sea
Res. I 50, 737–747.
Le Bris, N., Zbinden, M., Gaill, F., 2005. Processes controlling the
physico-chemical micro-environments associated with Pompeii
worms. Deep-Sea Res. I 52 (6), 1071–1083.
Le Bris, N., Govenar, B., Le Gall, C., Fisher, C.R., 2006. Variability of
physico-chemical conditions in 9°50′N EPR diffuse flow vent
habitats. Mar. Chem. 98 (2–4), 167–182.
Luther, G.W., Rozan, T.F., Taillefert, M., Nuzzio, D.B., Di Meo, C.,
Shank, T.M., Lutz, R.A., Cary, C., 2001. Chemical speciation
drives hydrothermal vent ecology. Nature 410, 813–816.
Majzlan, J., Navrotsky, A., Schwertmann, U., 2004. Thermodynamics
of iron oxides: Part III. Enthalpies of formation and stability of
ferrihydrite (Fe(OH)3), schwertmannite (FeO(OH)3/4(SO4)1/8),
and Fe2O3. Geochim. et Cosmochim. Acta 68 (5), 1049–1059.
McCollom, T.M., Shock, E.L., 1997. Geochemical constraints on
chemolithoautotrophic metabolism by microorganisms in seafloor
hydrothermal systems. Geochim. Cosmochim. Acta 61 (20),
4375–4391.
McCollom, T.M., 2000. Geochemical constraints on primary productivity
in submarine hydrothermal vent plumes. Deep-Sea Res. I 47,
85–101.
Millero, F.J., Sotolongo, S., Izaguirre, M., 1987. The oxidation
kinetics of Fe(II) in seawater. Geochim. Cosmochim. Acta 51,
793–801.
Parkhust, D.L., Appelo, T., 1999. User's guide to PHREEQC – a
computer program for speciation, batch reaction, one dimensional
transport, and inverse geochemical modelling. U.S. Geol. Survey.
water-resource Invest. (http://www.brr.cr.usgs.gov/projets/
GWC_coupled/phreeqc).
Polz, M.F., Cavanaugh, C.M., 1995. Dominance of one bacterial
phylotype at a Mid-Atlantic Ridge hydrothermal vent site. Proc.
Natl. Acad. Sci. USA. 92 (16), 7232–7236.
Ravaux, J., Gaill, F., Le Bris, N., Sarradin, P.M., Jollivet, D., Shillito,
B., 2003. Heat-shock response and temperature resistance in the
deep-sea vent shrimp Rimicaris exoculata. J. Exp. Biol. 206,
20345–20354.
Rickard, D., Morse, J.W., 2005. Acid volatile sulfide (AVS). Mar. Chem.
97 (3–4), 141–197.
Rona, P.A., Klinkhammer, G., Nelsen, T.A., Trefry, J.H., Elderfield,
H., 1986. Black smokers, massive sulfides and vent biota at the
Mid-Atlantic-Ridge. Nature 321, 33–37.
Sarradin, P.M., Caprais, J.C., Briand, P., Gaill, F., Shillito, B.,
Desbruyères, D., 1998. Chemical and thermal description of the
environment of the Genesis hydrothermal vent community (13°N,
EPR). Cah. Biol. Mar. 39, 159–167.
Sarradin, P.M., Le Bris, N., Le Gall, C., Rodier, P., 2005. Fe analysis
by the ferrozine method: adaptation to FIA towards in situ analysis
in hydrothermal environment. Talanta 66 (5), 1131–1138.
Segonzac, M., de Saint Laurent, M., Casanova, B., 1993. L’énigme du
comportement trophique des crevettes Alvinocarididae des sites
hydrothermaux de la dorsale médio-atlantique. Cah. Biol. Mar. 34,
535–571.
Shock, E.L., Holland, M.E., 2004. Geochemical energy sources that
support the subsurface biosphere. In: Wilock, W.S.D., DeLong, E.F.,
Kelley, D.S., Baross, J.A., Cary, S.C. (Eds.), The subseafloor
biosphere at Mid-Ocean ridges. American Geophysical Union,
Washington DC, pp. 153–166.
Takai, K., Gamo, T., Tsunogai, U., Nakayama, N., Hirayama, H.,
Nealson, K.H., Hoirikoshi, K., 2004. Geochmical and microbiological
evidence for a hydrogen-based, hyperthermophilic

C. Schmidt et al. / Marine Chemistry 108 (2008) 18–31
31
subsurface lithoautotrophic microbial ecosystem (HyperSLIME)
beneath an active deep-sea hydrothermal field. Extremophiles 8,
269–282.
Takai, K., Campbell, B.J., Cary, S.C., Suzuki, M., Oida, H., Nunoura,
T., Hirayama, H., Nakagawa, S., Suzuki, Y., Inagaki, F., Horikoshi,
K., 2005. Enzymatic and genetic characterization of carbon and
energy metabolisms by deep-sea hydrothermal chemolithoautotrophic
isolates of epsilonproteobacteria. Appl. Environ. Microbiol.
71, 7310–7320.
Theberge, S.M., Luther, G.W., 1997. Determination of the electrochemical
properties of a soluble aqueous FeS species present in
sulfidic solutions. Aquatic Geochem. 3, 191–211.
Von Damm, K.L., Lilley, M.D., 2004. Diffuse flow hydrothermal
fluids from 9°50′N East Pacific Rise: origin, evolution and
biogeochemical controls. In: Wilock, W.S.D., DeLong, E.F.,
Kelley, D.S., Baross, J.A., Cary, S.C. (Eds.), The subseafloor
biosphere at Mid-Ocean ridges. American Geophysical Union,
Washington DC.
Wirsen, C.O., Jannasch, H.W., Molyneaux, S.J., 1993. Chemosynthetic
microbial activity at Mid-Atlantic-Ridge hydrothermal vent
sites. J. Geophys. Res. 98 (B6), 9693–9703.
Zielinski, F., Pape, T., Wenzhöfer, F., Seifert, R., Dubilier, N., 2005.
Hydrogen may be an energy source for endosymbiotic bacteria of
the vent mussel Bathymodiolus puteoserpentis. EOS Trans. AGU
86 (52) (Fall Meet. Suppl., abstract OS33A-1456).
Zbinden, M., Le Bris, N., Gaill, F., Compère, P., 2004. Distribution of
bacteria and associated minerals in the gill chamber of the vent
shrimp Rimicaris exoculata and related biogeochemical processes.
Mar. Ecol. Prog. Ser. 284, 237–251.
Zhang, J.Z., Millero, F.J., 1994. Investigation of metal sulfide
complexes in sea-water using cathotic stripping square-wave
voltammetry. Anal. Chim. Acta 284 (3), 497–504.