REVIEW ARTICLE
CURRENT SCIENCE, VOL. 108, NO. 2, 25 JANUARY 2015 246
*e-mail: ramananda@ceas.iisc.ernet.in
Silicon isotopes: from cosmos to benthos
Ramananda Chakrabarti*
Centre for Earth Sciences, Indian Institute of Science, Bengaluru 560 012, India
Silicon is the second most abundant element on the
Earth and one of the more abundant elements in our
Solar System. Variations in the relative abundance of
the stable isotopes of Si (Si isotope fractionation) in
different natural reservoirs, both terrestrial (surface
and deep Earth) as well as extra-terrestrial (e.g. mete-
orites, lunar samples), are a powerful tracer of pre-
sent and past processes involving abiotic as well as
biotic systems. The versatility of the Si isotope tracer
is reflected in its wide-ranging applications from un-
derstanding the origin of early Solar System objects,
planetary differentiation, Moon formation, mantle
melting and magma differentiation on the Earth, an-
cient sea-water composition, to modern-day weather-
ing, clay formation and biological fractionation on
land as well as in the oceans. The application of Si iso-
topes as tracers of natural processes started over six
decades ago and its usage has seen a sudden increase
over the last decade due to improvements in mass
spectrometry, particularly the advent of multi-
collector inductively coupled plasma mass spectro-
meters, which has made Si isotope measurements safe
and relatively easy while simultaneously improving
the accuracy and precision of measurements.
Keywords: Mass spectrometry, meteorites, sea water,
silicon isotopes, weathering.
ISOTOPE geochemistry is a discipline that utilizes varia-
tions in the relative abundance of isotopes of elements as
tracers of natural processes. Although there are almost 60
naturally occurring elements that are not radioactive or
have a very long half life (~ one billion years or more and
hence essentially stable) and have more than one isotope,
traditional stable isotope geochemistry studies have typi-
cally focused on the elements H, O, C, N and S. Tradi-
tional stable isotope geochemistry has been successfully
applied to diverse problems, including the origin and evo-
lution of our Solar System, crust and mantle evolution,
origin of life, palaeoclimate reconstruction, and modern
climate studies amongst many others. Breakthroughs in
Earth and planetary science, especially in geochemistry,
follow the development of instrumentation which pro-
vides newer and more precise analytical data on natural
materials. The evolution of the field of ‘non-traditional’,
stable isotope geochemistry in the last ten years is a
direct result of the advent of multi-collector inductively
coupled plasma mass spectrometry (MC-ICPMS; c.f. refs
1 and 2).
Silicon (Si) is one of the most abundant elements on
the Earth as well as the Solar System and has three stable
isotopes with masses 28, 29 and 30. The earliest studies
involving Si isotope ratio measurements date back to the
early 1950s (e.g. Reynolds and Verhoogen
3
). These
measurements were done using gas-source isotope ratio
mass spectrometers (IRMS), which involved the use of
very hazardous gases such as F
2
or BrF
5
. The complexity
of these measurements resulted in limited use of Si iso-
topes as a ‘tracer’ of terrestrial and planetary processes.
The first MC-ICPMS Si isotope data were reported in
2002 by De La Rocha
4
; this new measurement tool elimi-
nated the use of hazardous reagents and has propelled the
use of Si isotopes in geochemical and cosmochemical re-
search. Over 40 papers on Si isotope ratio measurements
in a variety of natural objects using MC-ICPMS have
been published since 2002. Being a widely abundant ele-
ment in nature, Si isotope fractionation has been docu-
mented for a wide variety of natural processes as diverse
as planetary differentiation to secretion of siliceous shells
by microorganisms. Here, I present an overview of the
application of Si isotopes in tracing processes from early
Solar System evolution (cosmos) to biological fractiona-
tion in the marine environment (benthos).
Measurement techniques
Since 2002, most published Si isotope ratio measure-
ments are using a MC-ICPMS. However, the use of
IRMS for Si isotope ratio measurements continues with
recent modifications involving conversion of silica to
Cs
2
SiF
6
followed by acid reaction to produce SiF
4
gas
5
,
laser heating of precipitated silica
6
, etc. Prior to MC-
ICPMS measurements, samples have to be dissolved and
purified. The traditional method of using hydrofluoric
acid (HF) for dissolving silicate samples can potentially
fractionate Si isotopes due to the volatile loss of SiF
4
,
which forms during this dissolution process. It has been
argued that limited use of HF does not fractionate Si iso-
topes
4,7–9
. However, mass bias instability during mass
spectrometry can occur
10
. Hence, samples are fused using
an alkali flux followed by dissolution in dilute inorganic
acids and separation of Si from the sample matrix using
ion-exchange chromatography (e.g. refs 10–13). Ion-
exchange chromatography eliminates interferences from
doubly charged species like
56
Fe
2+
,
58
Fe
2+
and
58
Ni
2+
, and
REVIEW ARTICLE
CURRENT SCIENCE, VOL. 108, NO. 2, 25 JANUARY 2015 247
60
Ni
2+
on masses 28, 29 and 30 respectively, during mass
spectrometry as well as reduces the matrix effect which
can affect isotopic measurements (e.g. ref. 14). However,
molecular species which are generated in the plasma and
cause isobaric interferences (e.g.
14
N
2
+
and
12
C
16
O
+
for
mass 28,
14
N
2
H
+
for mass 29 and
14
N
16
O
+
for mass 30)
have to be resolved. This is typically done by analyses in
medium or high resolution (cf. ref. 15) as well as the use
of H
2
and He in small amounts in a collision cell, if this
option is available in the MC-ICPMS (cf. ref. 11). Analy-
ses in high resolution, however, reduces sensitivity.
Hence Si concentration in samples and standards,
matched to within 10%, have to be between 1 and 5 ppm
for acceptable counting statistics.
Since Si has only three stable isotopes, it is not possi-
ble to use the more definitive double-spiking technique
16
for accurate determination of Si isotope fractionation in
nature. To correct for instrumental mass fractionation, a
sample-standard bracketing technique has to be used.
This technique assumes that the mass bias of the sample
and the bracketing standard are similar. Early IRMS Si
isotope ratio measurements used Cal Tech Rose Quartz as
the bracketing standard
17–24
. The first Si isotopic study to
use the NBS28 standard was done by Molini-Velsko et
al.
25
. For MC-ICPMS measurements, the NBS28 standard
is also chemically processed in the same way as the
sample, prior to analysis. Delta values are calculated as
xx
Si(‰) = [(
xx
Si/
28
Si)
sample
/(
xx
Si/
28
Si)
NBS28
)
–
1] 1000
(permil, ‰), where xx is either mass 29 or 30. After suc-
cessful chemistry and mass spectrometry, all samples and
standards analysed should plot on a mass-dependent frac-
tionation line in a three-isotope plot with a slope close to
one-half. An alternate method to correct for instrument
fractionation during Si isotope ratio measurements is to
simultaneously measure Mg-isotopes in Mg-doped
samples and standards
7,26
assuming that the instrument
fractionation of Mg and Si isotopes is similar. Several
inter-laboratory standards like Diatomite, Big Batch,
IRMM018, NBS28, Harvard-AA and Harvard-HPS
11,27–29
have been prepared and calibrated. These standards are
also chemically processed and routinely measured along
with unknown samples to test for accuracy of measure-
ments. With improved instrumentation, sub-0.1‰ data
for
30
Si can now be obtained routinely using MC-
ICPMS.
In situ measurements of Si isotopes using MC-ICPMS
involve ablation of the sample by lasers prior to introduc-
tion in the plasma for ionization; high-precision data have
been reported using LA-MC-ICPMS
30–32
. In situ Si iso-
tope data can also be obtained using MC-SIMS (secon-
dary ion mass spectrometry). While this technique allows
sampling in high spatial resolution, it is limited to a 2
uncertainty of 0.3‰ for
30
Si (ref. 33), which although a
significant improvement compared to earlier ion-probe
studies (e.g. refs 34–37), is still three times less precise
than MC-ICPMS.
Si isotopes in bulk meteorites and their
components
In the early studies of bulk meteorites, a
30
Si range of
2‰ was observed in the different meteorite groups de-
rived from both primitive as well as differentiated
planetesimals and no resolvable difference was observed
between the average
30
Si in terrestrial igneous rocks,
meteorites and lunar samples. Bulk meteorites showed
mass-dependent silicon isotope fractionation and isotopic
anomalies were not observed (cf. ref. 25). In contrast,
larger (3–4‰) fractionation in
30
Si has been observed in
many refractory inclusions of chondritic meteorites
38–40
(Figure 1). These large variations in
30
Si have been
explained by isotopic exchange reactions between nebu-
lar-SiO gas and silicate condensates forming from it and
can be modelled by a Rayleigh process
38
. Some calcium-
and aluminum-rich inclusions (CAIs) from the chondritic
meteorite Allende (e.g. C1, EK 1-4-1, CG-14, HAL,
labelled ‘FUN’ – unknown nuclear fractionations) show
anomalous compositions of isotopes of many elements.
30
Si values of these FUN inclusions show a range of
Figure 1. Schematic diagram showing Si isotope (
30
Si, mostly MC-ICPMS data) variability in natural material from bulk meteorites and their
components (CAIs, calcium–aluminum-rich inclusions), bulk lunar rocks, terrestrial igneous rocks, and Precambrian and Phanerozoic cherts. Also
shown as shaded regions are the domains of the
30
Si of bulk silicate earth (BSE), river and groundwater (RW, GW), products of terrestrial
weathering like clays and silcretes, land plants, average present-day sea water (SW), diatoms and sponges, etc. Large fractionations observed in
‘FUN’-CAIs inclusions and pre-solar grains are not plotted.
REVIEW ARTICLE
CURRENT SCIENCE, VOL. 108, NO. 2, 25 JANUARY 2015 248
~25‰, which is an order of magnitude higher than most
other natural objects that have been measured for their Si
isotopic composition (e.g. ref. 41). Largest fractionation
(100s ‰) of Si isotopes are observed in pre-solar SiC
grains, which are thought to have been formed in AGB
stars
42,43
.
The range of
30
Si in bulk meteorites and terrestrial
igneous rocks as measured by recent MC-ICPMS meas-
urements (e.g. refs 11, 44) is an order of magnitude less
compared to the older IRMS data for the same (e.g. ref.
25). Reported
30
Si values for bulk carbonaceous and
ordinary chondrites are indistinguishable (e.g. refs 11, 44,
45), although some researchers have argued for a margin-
ally (by ~0.1‰) lighter
30
Si in bulk ordinary chondrites
compared to carbonaceous chondrites
46
. Enstatite chon-
drites, however, clearly show lighter
30
Si compared to
other chondrite groups
9,11,44,46,47
(Figure 1). The light Si is
hosted in metals in enstatite chondrites (as well as in their
differentiated counterparts, aubrites)
9,47,48
. The metal-free
silicate portion of enstatite chondrites shows similar
30
Si compared to other bulk chondrites
47
, thereby sug-
gesting broad homogeneity in the silicon isotopic compo-
sition of silicates condensing from the solar nebula. Non-
chondritic meteorites do not show much variation in
30
Si, which overlaps with that of bulk carbonaceous and
ordinary chondrites
11,44–46
(Figure 1).
Bulk silicate Earth Si isotopic composition
Si isotope data for terrestrial rocks were first reported in
the 1950s (ref. 3). However, the first comprehensive
study of Si isotopes in terrestrial samples was reported by
Douthitt
18
~30 years later. He reported that igneous rocks
show a range of 1.1‰ in
30
Si, which correlated with
silicon content and argued for a mean
30
Si of –0.4‰
(precision of 0.3‰) for terrestrial mafic and ultramafic
igneous rocks. Another decade and half later, Georg et
al.
44
reported the first high-precision MC-ICPMS data for
a wide suite of mafic and ultramafic terrestrial igneous
rocks and concluded that
30
Si of the bulk silicate Earth
(BSE) is –0.38 0.06‰. Subsequent studies have argued
for similar
30
Si (BSE) values ranging from –0.28
0.02‰ based on peridotites and ocean-island basalt
analyses
46
, –0.33 0.08‰ based on analyses of San Car-
los olivine
9
; –0.38 0.02 based on analyses of terrestrial
basalts from different tectonic settings and localities,
dunites as well as mantle-derived olivine and pyrox-
enes
11
, to –0.24 0.09 based on mafic and ultramafic
USGS rock standards
26
. A more elaborate study of mafic
and ultramafic terrestrial rocks yielded a
30
Si (BSE) of
–0.29 0.08‰ (ref. 49), which is similar within analyti-
cal uncertainties to other studies (Figure 1).
High-temperature fractionation of Si isotopes is expected
to be minimal. This is consistent with the lack of measur-
able Si isotopic variability in co-existing mantle minerals
(e.g. Chakrabarti and Jacobsen
11
). However,
30
Si of
evolved bulk igneous material (andesite, dacite, rhyolite,
etc.) is heavier compared to
30
Si (BSE) and correlates
with increasing silica content
50
(Figure 1). This observa-
tion is consistent with the original suggestion of
Douthitt
18
, which was based on less precise data. The cal-
culated
30
Si of average continental crust with an average
SiO
2
content of 60 wt%, derived solely from magmatic
differentiation of the mantle is however indistinguishable
from the BSE value
50
.
Core formation on Earth and the BSE–chondrite
30
Si difference (or the lack of it)
It has been long known that the density of the Earth’s
Fe–Ni core is lighter than a pure Fe–Ni alloy
51,52
, indicat-
ing the presence of elements lighter than Fe in the Earth’s
core
53
. Silicon is considered to be one of the probable
light elements in the Earth’s core
54
. If true, this should be
accompanied by silicon isotope fractionation between
metal (core) and silicates (mantle), as observed in
aubrites and enstatite chondrites
9,47
, high pressure–
temperature experiments
8,55
as well as theoretical Si
isotope fractionation calculations
56
. If indeed Si is parti-
tioned into the Earth’s core during differentiation, this
should result in an unique Si isotopic composition of the
Earth’s mantle (BSE) and the extent of fractionation
would depend on the fractionation factor between co-
existing metal and silicate at the core–mantle boundary
conditions, which are approximated by other geochemical
and isotopic proxies (e.g. refs 57–60), and the extent of Si
entering the core.
Early work on Si isotopes did not indicate any differ-
ence in
30
Si of the BSE and chondritic meteorites
25
.
This could, however, be due to larger analytical uncer-
tainties of these measurements. Georg et al.
44
first sug-
gested that
30
Si of the silicate Earth is heavier by 0.2‰
relative to meteorites (
30
Si
BSE–chondrite
= 0.2). A similar
argument was put forward by Fitoussi et al.
46
, although
they argued for a smaller
30
Si
BSE–chondrite
of 0.08 0.04
(1SD). In contrast, Chakrabarti and Jacobsen
11
found
30
Si
BSE–chondrite
of 0.035 0.035, which limited the
amount of Si in the Earth’s core to less than 3.84% and
that of oxidized Fe in the mantle during the first 90% of
planetary accretion to as low as ~1%. Armytage et al.
45
argued for a larger
30
Si
BSE–chondrite
(excluding enstatite
chondrites), which translated to 2.5–16.8 wt% Si in the
Earth’s core. The use of Si isotopes to estimate how much
Si is present in the Earth’s core depends on the Si iso-
topic composition of the starting material that formed the
Earth, often termed the chondritic uniform reservoir
(CHUR). This might not be any particular chondrite
group. Based on other geochemical and isotopic proxies,
the composition of enstatite chondrites is quite similar to
the BSE. However,
30
Si values of enstatite chondrites
REVIEW ARTICLE
CURRENT SCIENCE, VOL. 108, NO. 2, 25 JANUARY 2015 249
do not overlap with those of terrestrial material
9,11,44,46,47
.
Recent studies have suggested that a maximum of 15%
enstatite chondrites together with CO and CI chondrites
could approximate the CHUR value
48
.
The Earth–Moon system
The silicate Earth and the Moon show considerable geo-
chemical differences, although they show identical
isotopic compositions of different elements like O, Cr,
W, Ti, etc.
61–64
. In addition, it has been long considered
based on numerical simulations that ~80% of the Moon is
derived from the Mars-sized impactor that collided with
the proto-Earth
65
. To explain the isotopic similarity of the
BSE and the Moon, a large-scale isotopic equilibration of
the proto-lunar disk and the Earth has been suggested
66
.
Pahlevan et al.
67
suggested that the Moon and the BSE
should have a 0.14‰ offset in
30
Si for the Moon to have
a Fe/Fe + Mg ratio twice that of the BSE.
Silicon isotope analyses of lunar samples and meteor-
ites date back to almost 40 years ago (e.g. ref. 19). In
contrast to the difference in
30
Si between BSE and
primitive meteorites, the extent of which is debated,
30
Si
of the average Moon is identical to that of the BSE
11,44,68
.
No systematic differences exist between bulk samples of
different lunar lithologies, e.g. the high-Ti and low-Ti ba-
salts, Highland rocks and lunar glass
68
(Figure 1). The in-
distinguishable Si isotopic composition of the Earth and
the Moon does not require any later large-scale Si iso-
topic equilibration in the vapour cloud
66
after the Moon-
forming impact. Instead, it is consistent with recent simu-
lations suggesting that the proto-lunar disk was derived
primarily from the mantle of a fast-spinning proto-Earth
after the giant impact
69
.
Si isotopes in the modern Earth system
As discussed in the previous sections, Si isotope fractiona-
tion at high temperatures is limited.
30
Si of the average
continental crust is indistinguishable from the BSE, with
estimated
30
Si varying between –0.3‰ and –0.4‰. Sil-
ica sources to the oceans include continental weathering,
hydrothermal fluids and seafloor weathering of basalts.
Of these, continental weathering is the dominant source
of silica to the oceans, delivered primarily as dissolved
silicic acid (H
4
SiO
4
) in river water
70
as well as ground-
water
71
. The flux of suspended silica-rich particles in
river water is high, but long time-durations of dissolution
of this particulate matter result in little contribution of
silicic acid to the oceans from this source
70
. Si isotopes
are fractionated during continental weathering and associ-
ated clay formation; lighter isotopes of Si are preferentially
sequestered into secondary clay minerals, which conse-
quently display lower
30
Si values
72–77
(Figure 1). Based
on first principles calculations, a 1.6‰ fractionation is
estimated between quartz and kaolinite at ~27C at equi-
librium conditions
78
.
30
Si of suspended silica in rivers is
similar to average igneous rocks and shales, whereas
30
Si of the dissolved load is higher (~+0.8; Figure 1);
this is a consequence of isotopic mass balance after the
precipitation of
30
Si-depleted secondary minerals
79,80
.
Groundwaters are also depleted in the light Si isotopes
71
(Figure 1). Modern soils show depletion in
30
Si;
30
Si
fractionation during adsorption of silica on iron-oxide/
hydroxide particles to the extent of ~1.1‰ for ferrihy-
drite and ~1.6‰ for goethite has been proposed as a pos-
sible explanation
76,81
. Lowest
30
Si (–5.7‰) measured in
relatively young terrestrial samples is from silcretes (sili-
ceous cements)
34
(Figure 1). Si isotope fractionations
have been observed in the siliceous phytoliths of certain
land plants and can be explained by silica transport and
biomineralization processes
82–84
. In rice plants, both silica
concentration as well as
30
Si progressively increase
from roots, to stems and leaves and to the husk; rice
grains show little silica concentration but very high
30
Si,
as high as 6.1‰ (ref. 82). In bamboo trees, silica content
increases from stems, through branches to leaves, while
30
Si decreases from roots to stems, but increases from
stems through branches to leaves
84
(Figure 1).
Silicon isotopic composition of modern hydrothermal
fluids from the East Pacific Rise
85
is similar to that of
terrestrial igneous rocks.
30
Si of hydrothermal siliceous
precipitates (e.g. from Mariana and Galapagos) shows
low values (–0.4‰ to –3.1‰), similar to sinter deposits
from continental hot springs
86
. Clearly, amorphous silica
precipitates are isotopically lighter than the ambient fluid,
although the Si isotope fractionation factor between
amorphous silica precipitates and water remains to be
determined accurately and could vary as a function of the
precipitation mechanism.
The major sink for dissolved silica in the modern
oceans is biological uptake, primarily by diatoms as well
as radiolaria and sponges. This leads to relatively low sil-
ica concentration in the surface waters, which increases
to a relatively constant value at greater depths where sili-
ceous skeletons commonly dissolve
85
. The modern ocean
is silica-undersaturated with the average concentration of
H
4
SiO
4
at ~70 M (2 ppm) (ref. 70). However, silica con-
centrations show strong lateral and vertical heterogeneity
(from <1 to 15 ppm) (refs 70, 87) depending on nutrient
availability in surface waters and skeleton dissolution at
depth
88
. The residence time of Si in the global oceans is
~15,000 years, although relative to biological uptake from
surface waters, it is only ~400 years (ref. 70), implying that
Si delivered to the modern oceans is recycled ~40 times
through the biological cycle before it departs into a sedi-
mentary sink. Silica secreting organisms preferentially take
up lighter Si isotopes with Si isotope fractionation
factors ranging from –1.1‰ for diatoms to –3.5‰ in
marine sponges
89–91
. Although spatial and depth variation
in the sea water
30
Si is well-documented, the mean
REVIEW ARTICLE
CURRENT SCIENCE, VOL. 108, NO. 2, 25 JANUARY 2015 250
30
Si of modern sea water is estimated at +1.1‰ (ref. 85)
(Figure 1), which is higher than the
30
Si of continental
and hydrothermal inputs of silicic acid to the oceans. This
primarily reflects the preferential uptake of light Si iso-
topes by silica-secreting organisms in the oceans.
Si isotopes and evolution of sea-water
composition through time
Silica-precipitating organisms have played a dominant
role in the marine silica cycle throughout the last 550
million years (Ma, Phanerozoic Eon), although the major
users of dissolved silica have changed from sponges and
radiolaria in the early part of the Phanerozoic to diatoms
over the last 200 Ma (refs 92–94). During the Precam-
brian (older than 550 Ma), silica fluxes into the oceans
were likely similar to today, if not higher, given the
higher hydrothermal fluxes, inferred from Sr isotopes, in
a mantle-buffered Archean ocean (e.g. ref. 95). Silica
must continually have left the oceans by means of physio-
chemical processes, primarily as precipitates during the
early stage evaporation of sea water and is preserved
primarily as early diagenetic cherts in shallow-water car-
bonate successions
87,96
. Another significant but more
temporally restricted sink for silica are banded iron for-
mations (BIFs). BIFs generally contain 43–56% SiO
2
by
weight
97
and as most BIFs, particularly those of Archean
age, reflect offshore, commonly basinal deposition, silica
precipitation was probably not forced by evaporation of
sea water. The recently proposed ‘iron shuttle’ model
98
,
involving adsorption of silica on Fe-oxyhydroxides, is a
viable mechanism for transporation of silica to deeper
parts of the basin and be incorporated into BIFs. Some
silica could have been incorporated into clays formed
authigenically within sediments (e.g. ref. 99). In the ab-
sence of any known biological sink, it is estimated that
silica concentration must have been much higher in the
Precambrian sea water, perhaps close to amorphous silica
saturation
87,94
.
Deposits of amorphous silica (now microcrystalline
quartz) called cherts (including BIFs) have an ubiquitous
presence in the sedimentary rock record. These unique
rocks document the evolution of the global silica cycle
through geologic time
92,100
. Based on our understanding
of modern-day processes, both chert precipitation by
evaporation of shallow sea water and clay formation are
accompanied by Si isotopic fractionation that leads to
enrichment of heavier Si isotopes in ambient sea water.
Laboratory experiments have shown that adsorption of
silica on Fe-oxyhydroxide particles leads to the progres-
sive enrichment of heavy Si isotopes in the ambient
water
81
, implying that silica deposition in deeper-water
BIFs by the ‘iron shuttle’ mechanism
98
must also have
been accompanied by Si isotope fractionation. Hence,
30
Si of Archean and Proterozoic cherts (Figure 1) can
potentially provide insights into the sources and sinks of
marine silica before the evolution of silica-precipitating
marine organisms like sponges, radiolarians and diatoms
in the Phanerozoic. An additional advantage of cherts is
their high Si content, which makes them less susceptible
to such post-depositional changes and Si isotopes are not
thought to be affected by late diagenesis
101
.
However, the use of cherts to directly infer the compo-
sition of Precambrian sea water has limitations because
most of these Precambrian siliceous rocks formed during
the diagenesis of precursor sediments, particularly car-
bonates. This is evident from the preservation of micro-
fossils in Precambrian cherts
102,103
. Several recent studies
have reported high-precision MC-ICPMS Si isotope
measurements of Archean cherts
31–33,36,37,104–107
, while
high-precision MC-ICPMS data for Proterozoic cherts are
limited
33,108
(Figure 1). Robert and Chaussidon
35
reported
in situ measurements of the Si isotopic composition of an
extensive set of Archean, Proterozoic and some Phanero-
zoic chert samples using an ion-microprobe technique
with a relatively coarse measurement precision (1‰, 2
)
(Figure 1). They interpreted their combined O and Si iso-
tope data from Precambrian cherts to argue for ~70C pa-
leo-sea-water temperatures, a conclusion that has not
received much support
109
. Subsequent studies have high-
lighted the importance of the source of silica, mechanism
of silica precipitation and depositional setting in inter-
preting the
30
Si of preserved cherts
31,106,107
.
30
Si of Precambrian sea water at any given time can be
modeled as a function of
30
Si of continental and hydro-
thermal inputs and outputs dominated by peritidal chert
precipitation and BIF formation, their relative fluxes and
the isotopic fractionation factors related to precipitation
of silica from sea water or porewater and/or adsorption of
silica onto Fe-hydroxide particles during BIF formation.
The long residence time of Si relative to ocean mixing
timescales indicates that Si in Precambrian cherts was
derived from sea water with relatively uniform
30
Si.
Hence,
30
Si variations in cherts over time would indi-
rectly reflect the temporal variation of
30
Si of the sea
water. A detailed analysis of factors than can cause varia-
tion in the sea water
30
Si is presented in Chakrabarti et
al.
108
and is briefly discussed here.
Although there is marked isotopic variation in
30
Si in
cherts from individual Proterozoic basins which can be
modelled using Rayleigh-type isotope fractionation
107,108
,
there is a clear pattern of change in
30
Si from globally
distributed cherts throughout the Proterozoic Eon
35,108
.
30
Si increases from 3.8 till 1.5 Ga followed by a
decreasing trend which continues into the Phanerozoic
(Figure 1). The increasing trend from Archean to
Mid-Proterozoic could reflect a gradual change in the
dominant silica source from hydrothermal (low
30
Si) to
continental (high
30
Si) as well as isotopic fractionation
related to precipitation of cherts and BIFs. The highest
30
Si observed in Mid-Proterozoic cherts could reflect an
