218
microbial respiration, the detailed mechanisms of siderite
formation needs to be evaluated in future work.
12.4.2 Iron Isotope Fractionations in Lac Pavin
The geochemical model developed in the present contribu-
tion provides an estimate of Fe isotope fractionations for Fe
oxidation-precipitation and reduction-dissolution, and vivi-
anite precipitation in Lac Pavin. Contrasting with some geo-
chemical parameters adjusted in our modeling, the tests of
sensitivity demonstrate that Fe isotope fractionations are rea-
sonably well constrained, and can be compared to data from
the literature.
In Lac Pavin, Fe oxidation-precipitation at the redox tran-
sition zone produces an isotope fractionation of ~1 ‰
(0.9 ± 0.2 ‰) (Table 12.3), leaving the residual aqueous
Fe(II) depleted in heavy isotopes (Fig. 12.7c). This result is
compatible with experimental data available for Fe(II)aq oxi-
dation to Fe(III)aq, showing that Fe(III)aq is systematically
enriched in the heavy isotopes (e.g. Johnson et al. 2002 ;
Welch et al. 2003 ; Anbar et al. 2005 ; Balci et al. 2006 ). The
enrichment in heavy Fe isotopes during oxidation is gener-
ally counterbalanced by a kinetic isotope fractionation dur-
ing precipitation, enriching Fe(III) precipitate in the light
isotopes (Skulan et al. 2002 ). The net isotope fractionation
thus depends on the rate of precipitation but produces in
most cases a precipitate enriched in the heavy isotopes of Fe.
For instance, in experiments of abiotic precipitation of ferri-
hydrite from Fe(II)aq, the precipitate was enriched in heavy
isotopes by ~0.9 ‰ (Bullen et al. 2001 ), which is similar to
the value deduced from our geochemical modeling.
Experimental culture of Fe(II)-oxidizing photoautotrophic
bacteria under anaerobic condition showed that Fe oxyhy-
droxide precipitate was enriched in the heavier isotope rela-
tive to Fe(II)aq by ~1.5 ± 0.2 ‰ (Croal et al. 2004 ). The close
similarity between Fe isotope fractionation in abiotic and
biotic experiments prevents any assessment of life involve-
ment in the Fe oxidation process at Lac Pavin. SEM images
of particulate matter collected on filters at the redox transi-
tion zone shows encrusted microbial shape, suggesting that
Fe oxidation may be bacterially-mediated.
Iron reduction also plays a key role in the Fe biogeochem-
ical cycle at Lac Pavin. The geochemical model predicts a Fe
isotope fractionation of ~1 ‰, with an enrichment in light
isotopes in the produced Fe(II)aq (Table 12.3). This value is
similar in magnitude but reverse in direction to the Fe iso-
tope fractionation determined for oxidation. From various
experimental studies, it was shown that Fe(III) substrate
reduction and dissolution preferentially release light Fe iso-
topes into solution, in good agreement with the results of the
present study (e.g. Crosby et al. 2005 ). Abiotic mineral dis-
solution produces either no or limited Fe isotope fraction-
ation, depending on the mechanism (proton-promoted,
ligand-controlled or reductive dissolution). In contrast, bac-
terial dissimilatory iron reduction (DIR) can be associated
with large Fe isotope fractionation, up to 3 ‰, between
Fe(II)aq and the ferric substrate (Beard and Johnson 1999 ;
Crosby et al. 2005 ; Johnson et al. 2008 ). In Lac Pavin, Fe(III)
reduction is probably bacterially mediated, as demonstrated
by culture experiments on water samples from the monimo-
limnion (Lehours et al. 2009 ). In particular, facultative
Fe(III) reducers such as fermentative and sulphate-reducing
bacteria were enriched in culture experiments while obliga-
tory Fe(III) reducers, such as Geobacter, were not detected.
Although not discriminating, the Fe isotope fractionation
determined from our geochemical modeling is fully compat-
ible with a bacterial Fe reduction in Lac Pavin.
The last Fe isotope fractionation value determined from
modeling corresponds to vivianite precipitation process. To
our knowledge, no previous study determined such fraction-
ation, either experimentally or from theoretical calculation.
The present results thus provide the first estimate of this frac-
tionation. We find that vivianite should be enriched in the
light isotope of Fe by ~0.5 ‰ relative to dissolved Fe(II)
(Table 12.3). This may reflect a kinetic isotope process asso-
ciated with vivianite precipitation, but needs to be explored
further in future studies. An interesting observation is that
the spherical shape of vivianite particles formed in the moni-
molimnion (n) and their close association with organic mat-
ter suggest that microorganisms might be involved in the
precipitation process (Cosmidis et al. 2014 ).12.5 Conclusion and Perspective
Over the last decades, geochemical data and modeling on
dissolved and particulate matters coupled with mineralogical
analyses have provided a global picture of the biogeochemi-
cal machinery in Lac Pavin. Like for other anoxic and ferru-
ginous aquatic systems and seasonally stratified lake, Lac
Pavin is characterized by a “ferrous wheel” or “iron wheel”,
which corresponds to strong Fe cycling through successive
reductive dissolution, diffusion and oxidation-precipitation
reactions. In Lac Pavin, Fe cycle is intimately associated
with P cycle through precipitation of Fe(II)-Fe(III)-
phosphates, P-rich HFO and Fe-oxyhydroxides with high
amount of adsorbed phosphate, at the redox transition zone,
and release of SRP during Fe(III) reductive dissolution under
anoxic conditions. This leads to extremely high concentra-
tions of dissolved Fe(II) and SRP at the lake bottom, induc-
ing ferrous phosphate (i.e. vivianite) precipitation and
sequestration to the sediment.
In addition, phosphorus cycle is also partly controlled by
microorganisms in Lac Pavin, which can store high amounts
of intracellular polyphosphates and/or catalyze the hydroly-V. Busigny et al.