Download PdF-Article - Institut de Physique du Globe de Paris

The Island Arc (2000) 9, 428–438
Thematic Article
Diamond growth during ultrahigh-pressure metamorphism of the
Kokchetav Massif, northern Kazakhstan
K. DE CORTE,1,2 A. KORSAKOV,3 W. R. TAYLOR,4 P. CARTIGNY,5 M. ADER5 AND P. DE PAEPE1
1
Department of Geology and Soil Science, University of Ghent, Belgium ([email protected]),
2
Department of Geology and Mineralogy, Royal Museum for Central Africa, Tervuren, Belgium,
3
Institute of Mineralogy and Petrography, Novosibirsk, Russia, 4Research School of Earth Sciences,
Australian National University, Canberra, Australia, 5Laboratoire de Géochimie des Isotopes Stables,
Université de Paris 7 et Institut de Physique du Globe, France
Abstract Characteristic features of in situ diamonds can be used to retrace diamond formation during ultrahigh pressure (UHP) metamorphism of the Kokchetav Massif,
Kazakhstan. These features include the nitrogen aggregation state in diamond, dissolution
features observed on diamond surfaces, and the carbon and nitrogen isotopic composition
of the diamonds. The minerals in which the diamonds are included provide additional information and support the view that at least some of the Kokchetav microdiamonds are the
products of prograde or peak UHP metamorphism. The coexistence of diamond and
graphite is evaluated within this framework.
Key words: diamond, graphite, ultrahigh-pressure metamorphism.
INTRODUCTION
Approximately 10 years ago, microdiamonds
(< 1 mm) were recognized in metamorphic rocks
of crustal affinity in the Kokchetav Massif of
Northern Kazakhstan (Sobolev & Shatsky 1990).
However, their formation requires much higher
pressure–temperature (P–T) conditions than those
expected within the Earth’s crust. Three fundamentally different models have been proposed to
explain their occurrence.
The premetamorphic relict model postulates
that the diamonds are of kimberlitic origin
(Marakushev et al. 1995). Diamonds of mantle
origin were brought by kimberlites or lamproites
to the Earth’s surface. They were incorporated
into crustal sediments as placer deposits that were
later metamorphosed. It is thought that diamonds
would have survived the metamorphism because
they were enclosed in refractory minerals that
grew early in the metamorphic history.
The metastability model claims diamond genAccepted for publication 27 March 2000.
© 2000 Blackwell Science Asia Pty Ltd.
esis by condensation from a reduced mantle fluid
at temperatures of 600–1050 °C and pressures that
are below their equilibrium pressure in local shear
zones. Diamonds are formed at conditions of
crustal metamorphism and there is no requirement for ultrahigh-pressures (UHP), which are
characteristic of the upper mantle (e.g. Nadejdina
& Posukhova 1990; Dobrzhinetskaya et al. 1994;
F. A. Letnikov, unpubl. data, 1995; Simakov 1995;
Lavrova et al. 1997).
The
ultrahigh-pressure
metamorphism
(UHPM) model states that diamonds grew
steadily within their stable equilibrium as a result
of a major metamorphic event (Sobolev & Shatsky
1990). Crustal rocks subducted to depths of over
150 km were metamorphosed and returned to
the Earth’s surface. Microdiamonds crystallized
at pressures of > 4 GPa and temperatures ≥ 900–
1000 °C.
Studies of the external morphology of diamonds
found in situ within metamorphic rocks of the
Kokchetav Massif (hereafter described as in situ
diamonds) strongly suggest that the diamonds
crystallized in the diamond stability field (Shatsky
et al. 1989; De Corte et al. 1999). The presence
Diamond growth during UHP metamorphism
of diamond populations linked to the host rock;
their characteristic cuboid morphology; the complete absence of diamond abrasion features; the
preservation of Ib diamond; and the isotopic composition of the diamonds, which reflect a crustal
signature and the tectonic setting all support
microdiamond genesis during UHPM. However,
the diamonds may have crystallized during the
latter part of the prograde path, at peak metamorphic conditions, and/or during the early retrograde path of UHPM.
Metamorphic conditions that led to the formation of the diamondiferous UHPM rocks in the
Kokchetav Massif peaked at temperatures of
900–1000°C and pressures of more than 4 GPa
(Shatsky et al. 1995; Zhang et al. 1997; Fig. 1).
These P–T estimates are based on relict high-P–T
minerals that are included in garnet and zircon
because whole-rock mineral assemblages have
been affected to varying degrees by retrograde
metamorphism. Recently, new peak pressure estimates of at least 5 GPa have been proposed (V. S.
Shatsky pers. comm., 1999) on the basis of magnesite and magnesian–calcite inclusions in zircon
from Kumdy–Kol marble (sample K91-16). These
inclusions are inferred to be high-P breakdown
products of dolomite, which is unstable in comparison with aragonite and magnesite at pressures
exceeding 5 GPa (Martinez et al. 1996). Okamoto
et al. (2000) suggest peak pressures of 6 GPa for
eclogites and adjacent garnet peridotites. They
further suggested that SiO2 rods and rutile lamellae in clinopyroxene would require even higher
pressures (7 GPa and higher).
In seeking to unravel the formation history of
diamond and, hence, the complexities of the
P–T–time path, the genesis of the Kokchetav
microdiamonds will be discussed with reference to
the nitrogen aggregation state, its morphological
features and the isotopic composition of the in situ
microdiamonds. The characteristics of some host
minerals are used to further constrain their histories. This information is used to evaluate the
observed coexistence of graphite and diamond in
some samples.
NITROGEN AGGREGATION STATE
Based on Sm–Nd and U–Pb ages, the Kokchetav
microdiamonds spent no longer than 25 Ma at
depths greater than mid-crustal levels and
temperatures >600°C (Claoué-Long et al. 1991;
Shatsky et al. 1995; Hermann et al. 1999). The
429
nitrogen aggregation state of diamond is the only
method that gives precise information about the
prevailing T during nitrogen aggregation and the
time (t) diamonds spent at high temperature.
Diamonds can be classified into two types on the
basis of their infrared (IR) spectra and the form
in which substitutional nitrogen impurities (i.e. N
substituting for C atoms) occur in them. Type II
diamonds have essentially no IR-active substitutional nitrogen, while type I diamonds contain
> 10 ppm of substitutional nitrogen incorporated at
the time of their formation. Type I diamonds are
further characterized by the presence of a dominant form of a nitrogen defect center. The main
subgroups are type Ib, IaA and IaB. In type Ib diamonds, nitrogen substitutes in the form of single
atoms (the C defect center). In type Ia diamonds,
nitrogen-related defect centers are present in
aggregated forms. Type IaA diamond contains
nitrogen atoms arranged in pairs (the A defect
center), whereas in type IaB diamond four nitrogen atoms occur in a tetrahedral arrangement
around a vacancy (the B defect center). The
various nitrogen defects are related to each other
by diffusion, and an aggregation sequence of C to
A to B has been established (Chrenko et al. 1977;
Evans & Qi 1982). Chrenko et al. (1977) showed
experimentally that type Ib diamond could be
transformed into type IaA. The aggregation reaction was shown to follow second order kinetics.
The degree of nitrogen aggregation (or the
nitrogen aggregation state) depends mainly on the
initial nitrogen content, the prevailing T during
aggregation and the time spent at this temperature (Chrenko et al. 1977; Evans & Qi 1982; Taylor
et al. 1996). The activation energy of the Ib to IaA
transformation is relatively low and is dependent
on crystallographic orientation. The activation
energy has been determined experimentally to be
4.4 eV for octahedral growth sectors and 6.0 eV
for cube (cuboidal) sectors (Taylor et al. 1996).
Cathodoluminescence studies of in situ diamonds
of the Kokchetav Massif revealed the presence of
dominating cubic growth sectoring; hence, the
activation energy of 6.0 eV is used.
Singly substituted nitrogen atoms (type Ib
diamond) are preserved in all studied in situ diamonds of the Kokchetav Massif (De Corte et al.
1998a, 1999), which indicates a short residence
time at high pressure and/or a low prevailing temperature (Chrenko et al. 1977; Evans & Qi 1982;
Taylor et al. 1996). However, the temperature to
which the diamonds were exposed must have been
more than 750 °C as below this temperature nitro-
430
K. De Corte et al.
Fig. 1 Peak ultrahigh-pressure metamorphic conditions as defined by (i) Shatsky et
al. (1995) and Zhang et al. (1997); (ii)
Shatsky et al. (unpubl. data); and (ii)
Okamoto et al. (2000). (iv) Considering Si
rods in clinopyroxene and rutile lamellae,
more than 7 GPa is required (Okamoto et al.
2000). The figure is adapted from Zhang
et al. (1997).
gen aggregation does not proceed for reasonable
geologic times (i.e. 750°C is the blocking temperature of nitrogen aggregation; Taylor et al. 1996).
Temperature–time calculations based on nitrogen
content and the aggregation state of nitrogen in
diamonds obtained from garnet-clinopyroxene
rocks all yield average aggregation temperatures
of 900–950°C for a geologic acceptable residence
time of 5 Ma (i.e. the time spent by the diamond in
the high-P environment at T > 750°C; De Corte
et al. 1998a). It should be noted that the nitrogen
aggregation temperature is a time-averaged value
Diamond growth during UHP metamorphism
431
that depends on the postformation thermal history
of the diamond because all time spent at temperatures > 750°C will contribute, to some degree,
toward the final aggregation state of the diamond.
Calculations using probable prograde and retrograde T–t paths (Taylor et al. 1996) indicate
the peak metamorphic temperature will be some
25–50°C higher than the nitrogen aggregation
temperature range given earlier. Furthermore, the
problem can be inverted so that the residence time
can be calculated based on input temperature. A
residence time of approximately 10 Ma is required
if average formation plus postformation temperatures were 900°C, while only 0.097 Ma is needed if
average temperatures were 1000°C.
Temperature–time calculations thus indicate
that temperatures much higher than 800°C are
required during the aggregation process, so that
diamond crystallization during most of the retrograde phase of metamorphism (i.e. at T less than
about 900°C) can be excluded.
MORPHOLOGY
The diamond stability field covers a wide P/T
range. Once diamonds are formed, they remain
very stable as long as: (i) they are preserved in
favorable P–T and oxygen fugacity (fO2) conditions; and (ii) they do not come into contact with
resorptive agents. Some information about P–T
regimens can be obtained from diamond dissolution features. Signs of pronounced bulk dissolution
(e.g. tetrahexahedroid forms) have not yet been
observed among the Kokchetav diamonds but
surface features indicative of early stages of
resorption are observed (De Corte et al. 1999).
From the 672 in situ diamonds investigated using
scanning electron microscopy, more than 45%
crystals exhibit at least one of the following dissolution characteristics: (i) surface etch pit and frosting; (ii) a relatively small, irregular-shaped hole;
(iii) a rounded face and edge; (iv) a mottled or
blocky {100} face; (v) trigon, hexagon or tetragon;
and (vi) a rough octahedral face (Robinson 1979;
Sunagawa 1984). The negatively oriented etch features seen on Kokchetav microdiamonds (Fig. 2)
are of special interest as they provide evidence of
temperatures in excess of 950°C (Robinson 1979).
Hence, this temperature must have been reached
during UHPM. Water- and carbon dioxide-bearing
fluids are the etchants most likely to be responsible (Robinson 1979), and these fluids must have
had access to the diamonds after their formation.
Fig. 2 Octahedral crystal from zoisite gneiss sample b-94-83 (a) with
detailed view of its negatively-oriented trigon, bar = 10 mm, (b) which is
formed by high-temperature etching bar = 1 mm.
The process by which fluids reached the diamond
surfaces is not fully understood. Three models can
be proposed: (i) the present host minerals crystallized at the same time or just after the diamonds,
so that the etchants had free access to the diamonds; (ii) the diamonds were trapped together
with a coexisting fluid which, under different P/T
conditions, was out of equilibrium with the
diamond and caused etching; and (iii) as soon as
the diamonds were formed they were captured by
minerals but etchants came in contact with the diamonds via fractures in the host minerals. Although
we do not have evidence, the first or second model
seems more realistic and either implies that at
least some of the diamonds were present before
their host minerals.
For crystals with the same habit and from the
same host rock (200 g), surface features of
diamond are highly varied, as illustrated for
octahedra in Fig. 3. This can be explained by
local changes in P–T–fO2 conditions, by different
diamond growth events, by a distinct time of
432
K. De Corte et al.
Fig. 3 Octahedral crystals from zoisite gneiss sample b-94-369 with
(a) smooth faces and (b) a rough surface morphology. Both bars = 10 mm.
crystal capturing by the host minerals, and/or by
the non-uniform presence of active etchants.
HOST MINERALS
Microdiamonds are found today as inclusions in
garnet and clinopyroxene (and their secondary
minerals), zircon, kyanite, zoisite, biotite and
quartz included in garnet. Because the diamonds
are included in these minerals, the diamonds crystallized before or synchronously with the host
minerals. In this section we will explore some
constraints on the conditions of diamond formation
provided by the host minerals.
GARNET
Experimental work by Poli and Schmidt (1995)
demonstrates that the grossular component in
Fig. 4 Garnet compositions in zoisite gneiss sample (a) b-94-83 and
(b) b-93-6. (a) Circles represent matrix garnet compositions () garnet’s
core compositions; () mirrow garnet’s rim composition; +, , lozenges
indicate the composition of garents included in zircon. Garnets in zircon
found in association with high-pressure minerals are marked as lozenges,
with low-pressure minerals as , and isolated garnets are indicated
by (+).
garnet decreases with decreasing pressure.
Approximately 80% of the matrix garnets of
diamond-bearing rocks are characterized by a Carich core compared to their rim (Fig. 4a). Therefore, the garnet must have formed at or near the
peak of UHPM and continued its growth during
retrograde metamorphism. Hermann and Green
(1999) support garnet growth during retrograde
metamorphic conditions based on piston cylinder
Diamond growth during UHP metamorphism
experiments in the range 700–1100°C and 2.0–
3.5 GPa using gneissic to pelitic compositions.
As diamonds are observed both in the core and
in the rim of garnet, diamonds must have formed
before or during garnet formation. At least some
of the diamonds crystallized later during prograde
path and/or peak UHPM.
ZIRCON
Zircon is extremely stable and resistant over a
wide temperature and pressure interval (Chopin &
Sobolev 1995) and, therefore, is considered to be
the best UHP mineral container. Garnet, clinopyroxene, phengite, diamond, coesite, graphite, chlorite, quartz, plagioclase, K-feldspar, kyanite and
amphibole have been identified as inclusions in
zircon. From zoisite-bearing gneiss sample B93-6
the inclusion content of 860 zircon grains was
studied. Except for one case, all zircon grains contained either low P (quartz, graphite) or high P
minerals (coesite, diamond) and not much information could be obtained. Only in one zircon grain
of the sample set were both low P and high P
mineral assemblages identified. The low P minerals are located close to the zircon core, while the
high P minerals lie close to the zircon rim. A
similar arrangement of inclusions in zircon was
also described by Katayama et al. (1998). These
authors concluded that the zircon grew along the
prograde path and at peak metamorphism. The
present study supports the growth of some zircons
during prograde metamorphism, as distinct
mineral assemblages are present in well-defined
zones of the zircons.
In zoisite-bearing gneiss sample B-94-83 from
the Barchi locality, a zircon with a garnet inclusion
each in its core and rim has been observed. The
composition of both garnets is depleted in Ca compared to the core and even the rim of matrix
garnet (Fig. 4b). Therefore, we suggest that the
zircon grow prior to garnet crystallization and
thus during prograde metamorphism.
ZOISITE
Diamond inclusions in zoisite occur in core samples
of zoisite-bearing gneiss from the Barchi locality.
Diamonds occur more commonly in symplectitic
zoisite–quartz intergrowths, but again only in
zoisite gneisses from Barchi. The morphology of
the diamonds from these rocks seems to be related
to the presence of symplectitic zoisite. Diamonds
433
with an octahedral habit are only observed in
zoisite-bearing gneisses that are devoid of symplectitic zoisite, while zoisite-bearing gneisses with
symplectitic zoisite are dominated by diamonds of
cuboidal morphology (De Corte et al. 1999).
Infrared spectroscopy studies have shown that
cuboid microdiamonds from symplectitic zoisite
gneisses contain inclusions of carbonates and
water, as is the case for other in situ diamonds (De
Corte et al. 1999). The presence of these inclusions
is interpreted to be the result of the trapping of
fluids during the rapid crystallization of the diamonds (De Corte et al. 1998a). In a fluid-bearing
system the experimental work of Boettcher (1970)
showed that the assemblage zoisite–quartz in the
system CaO–Al2O3–SiO2–H2O is not stable at temperatures > 800–850°C at pressures around 3 GPa.
However, in the coesite stability field, the assemblage Zo + Coe is stable up to 850–900°C. If the
system is dry or at low H2O activity, much higher
temperatures are possible. The nature of
zoisite–quartz symplectites is not clear. As evidence of melting is absent in the rocks, the symplectites are interpreted to be the products of
decomposition of a UHP phase during retrograde
metamorphism.
MICA
Biotite containing cubo–octahedral microdiamonds has been found as an inclusion in garnet
from a biotite gneiss at Kumdy–Kol. The biotite
occurs in close relationship regarding locality with
quartz. In only one sample were diamonds identified as inclusions in matrix biotite of biotite gneiss.
The origin of the biotite is not yet fully understood.
Diamond has been found as an inclusion in phengite, which itself is included in garnet from zoisite
gneiss (Fig. 5). The phengite inclusion is partly
transformed into chlorite. Phengite is stable at
pressures > 2.4 GPa (Schmidt 1996) but chlorite is
not stable above 600°C (Schmidt 1996). Two theories may explain the coexistence of chlorite and
phengite as inclusions in garnet.
(1) Phengite was present at a UHPM phase and
grew during or after diamond formation. However,
the phengite stayed uncaptured until 600°C so
that it partly reverted to chlorite. Only then was
the phengite trapped by garnet, which implies that
the garnet is a retrograde product.
(2) Garnet enclosed phengite but, via a fracture,
the phengite reacted to become chlorite at lower
P–T conditions. Detailed investigations of the
garnet composition support the second hypothesis
434
K. De Corte et al.
by positive mean d15N values of = +5.9 and +8.5‰,
respectively. Their mean carbon isotopic composition corresponds to -10.5 and -10.2‰, respectively. The nitrogen isotopic composition of the
diamonds is in accordance with the positive values
of metasediments. The carbon isotopic composition of the diamonds is consistent with a mixture
of crustal carbonates and biogenic material, which
are characterized by initial mean d13C values of 0
(e.g. Veizer & Hoefs 1976) and - 25‰ (e.g. Schidlowski 1983), respectively. A contribution of mantle
carbon and nitrogen (with d13C = -4.5‰ and d15N
= -6‰, e.g. Javoy & Pineau 1991; Cartigny 1997)
is thought to be negligible but can not be completely ruled out from the present data set.
Additional support for the mixing of mainly
crustal carbonate and biogenic carbon is provided
by the presence of the diamonds in the area. In
eclogites or any other rock type that is unlikely to
contain a pre-existing carbonaceous component
(biogenic carbon), no single diamond has been
found to date (Shatsky et al. 1995). In rocks containing none or only minor carbonate contents,
diamonds have also not been found (I. Katayama,
pers. comm., 1999). Hence, the following diamond
reaction is favored:
Fig. 5 (a) Phengite inclusion (arrow) in garnet from zoisite gneiss
sample b-95-42 (microphotograph by A. Korsakov) Bar = 500 mm. (b)
Detailed view of phengite partly transformed to chlorite. Note the presence of a halo around the diamond. De Corte et al. (1998b) reported
earlier of observing reddish colors at the border of some diamonds that
were studied by cathodoluminescence and they interpreted it as the result
of annealing after radiation damage. This implies that the diamonds came
in contact with a radioactive medium, most probably fluids enriched in
uranium. Bar = 100 mm.
as the garnet core is enriched in Ca, indicating
garnet crystallization at peak metamorphism.
ISOTOPIC COMPOSITION
Carbonate and water inclusions in in situ cuboids
give evidence that the metamorphic diamonds are
crystallized from a C–O–H fluid (De Corte et al.
1998a, 1999). The main nitrogen and carbon source
of the diamonds is identified as recycled nitrogen
and carbon of metasedimentary origin (Cartigny
et al. 1998). Diamonds from garnet–clinopyroxene
rock (formerly described as ‘garnet clinopyroxenites’ by De Corte et al. 1998a, 1999 based on the
work of Shatsky et al. 1995; sample 2-4) and
marble (also known as ‘garnet–clinopyroxene
dolomitic rock’; sample K92-99) are characterized
CH4 + CO2 = 2C + 2H2O
Decarbonatization of carbonates producing CO2rich fluids is believed to have interacted with
reduced carbon phases bearing a biogenic component to form diamond. The reduced phase is represented here by CH4, although its precise nature
is unknown (bituminous material, CH4 and/or any
reduced hydrocarbon-rich fluids) and requires
further research. In a reduced carbonaceous sediment layer H2O released by a dehydration reaction
could lead to CH4 formation via the reaction:
C + 2H2O Æ CH4 + O2
A low intrinsic fO2 value for the rock will drive
the reaction to the right. In the framework of the
diamond genesis model proposed earlier, the coexistence of diamond and graphite is evaluated.
COEXISTING DIAMOND AND GRAPHITE
In the Kokchetav rocks, as a rule graphite and
diamond coexist (Fig. 6). Graphite is observed both
in intergranular sites and as inclusions in minerals
that also contain diamond. The latter case is of
Diamond growth during UHP metamorphism
435
zation. The transformation of diamond into
graphite involves the breaking of C–C bonds in the
diamond structure. As the number of bonds to be
broken is highest for {111} faces, the activation
energy for graphitization is theoretically the
highest for {111} faces (1060 kJ/mol; Pearson &
Nixon 1996). The activation energy for {110} faces
is 760 kJ/mol (Davies & Evans 1972). Experimental work has demonstrated that the rate of graphitization on {100} surfaces is much slower than that
of {110} surfaces and, therefore, the activation
energy for {100} faces is greater than 760 kJ/mol
(Evans 1979). At present no data about the precise
activation energy for graphitization of {100} faces
are available. The presence of some gases in
natural systems may catalyze graphitization, and
hence lower the activation energy.
It can be shown that the rate of graphitization
per unit time can be expressed as
Dx/dt = C e-(DE+PDV)/RT
Fig. 6 (a) Graphite crystals (arrow) coexisting with 20 mm octahedral
diamonds from zoisite gneiss sample b-94-331. Bar = 100 mm. (b)
Graphite coating (arrow) on a 105 mm diamond (garnet–clinopyroxene
rock sample 2-4). Bar = 100 mm.
special interest. Three subgroups are recognized:
(i) intergrowth of diamond and graphite; (ii) separate graphite crystals (without adjacent diamond)
and; (iii) graphite forming a coating on diamond.
Diamond and graphite are often intergrown.
Katayama et al. (1998) describe the particular case
of graphite rimmed by diamond. We believe that a
small influx of CO2-rich fluid reacted with bituminous material or CH4-rich fluid, which resulted in
diamond crystallization. The opposite case, that is,
in which a large amount of CO2-rich fluid reacted
with a low volume of bituminous material or CH4rich fluid, is believed to form separate diamond
crystals (without adjacent graphite).
Graphite-coated diamonds are present in all
rock types and no correlation between the coated
diamond and its size has been observed. Graphite
coats are observed on diamonds from any morphology. The graphite coats can either be
explained as late overgrowths forming in the
graphite stability field (hence, during the retrograde path), or as the result of diamond graphiti-
with C a constant, DV the activation volume and
DE the activation energy. The low activation
volume for graphitization (~10 cm3/mol) means
that the influence of the pressure on the reaction
is small and that the graphitization rate is thus
largely dependent on temperature (Pearson et al.
1995). The transformation of diamond to graphite
has been used (e.g. Pearson et al. 1995; Leech &
Ernst 1998) to place constraints on the temperature conditions experienced by rocks bearing such
graphitized diamonds.
The experimentally determined activation
energy for {110} faces is used as an indication of
the maximum graphitization rate of the Kokchetav
diamonds. Calculations show (Fig. 7) that the
conversion of an in situ diamond of average 15 mm
size (Shatsky et al. 1995)—which corresponds to
a mass of approximately 1 ¥ 10-11 g—to graphite
would require a time scale of less than 106 years at
1000 °C and less than 100 years at 1200 °C. At temperatures below 900 °C, graphitization takes more
than 10 million years, which is an unrealistic situation in the Kokchetav case. Hence, the graphite
crystals and coats may, in principle, be the result
of graphitization during prograde path, peak or
the early (T > 900°C) retrograde UHPM path.
Graphite found in close relationship with
diamond or other UHPM indicators, could be: (i)
graphite formed during prograde metamorphism
in the graphite stability field and preserved as
metastable phase in the diamond stability field; (ii)
graphite formed by complete graphitization of dia-
436
K. De Corte et al.
pared with their rims. This is considered to be evidence for garnet growth during the peak (core)
and retrograde (rim) conditions. Because diamonds are observed in both garnet cores and rims,
at least some of the diamonds must have formed
during prograde or peak metamorphism. The
study of diamond included in zircon confirms this
finding.
The nitrogen isotopic compositions of the diamonds reveal that the source of nitrogen was
metasedimentary. Their carbon isotopic signature
is consistent with a carbon source derived from a
mixture of crustal carbonates and reduced carbonaceous material. A reaction between CO2 and
CH4 to form diamond and H2O is proposed as the
dominant mechanism for diamond formation.
ACKNOWLEDGEMENTS
Fig. 7 Time required for half the mass of a diamond octahedron to
become graphite by conversion on {110} at 2 GPa (modified from
Pearson et al. 1995 and references herein).
mond; or (iii) ‘new’ graphite, grown during retrograde metamorphism in its own stability field.
The coexistence of diamond and graphite does
not imply diamond crystallization at or near the
graphite–diamond stability line as was stated in
the past. Our study proposes alternative explanations, which are not related to the P–T conditions
of diamond crystallization.
CONCLUSION
The Ib–IaA nitrogen aggregation state of the
Kokchetav diamonds requires that temperatures
higher than 800°C prevailed after their formation
so that the diamonds must have formed during
prograde metamorphism, peak metamorphism, or
during the high temperature part of the retrograde path. The dissolution features observed on
more than 45% of diamond surfaces suggest that
diamonds were etched, in some cases by temperatures in excess of 950°C. As diamonds in the
UHPM rocks are present as inclusions in silicates,
the characteristics of the host minerals may be
used to further constrain diamond history. Garnet
is of special interest as approximately 80% of the
garnet have a higher Ca content in their core com-
The authors express their sincere thanks to Prof
V. S. Shatsky, Prof J. Klerkx and Ac N. Dobretsov
for their interest and discussions. Dr A.
Zayakovsky, Dr K. Theunissen and Dr O. Navon
are thanked for their constructive comments. This
research was made possible by a grant received
from the Flemish Institute for Support of Scientific and Technological Research in the Industry
(Belgium).
REFERENCES
BOETTCHER A. L. 1970. The system CaO-Al2O3-SiO2H2O at high pressures and temperatures. Journal of
Petrology 11, 337–79.
CARTIGNY P. 1997. Concentration, composition
isotopique et origine de l’azote dans le manteau
terrestre. PhD thesis, Université Denis Diderot –
Paris 7, Institut de Physique du Globe de Paris p. 259.
CARTIGNY P., DE CORTE K., SHATSKY V. S., SOBOLEV N.
V. & JAVOY M. 1998. Microdiamonds from UHP
metamorphic rocks of the Kokchetav massif and
their bearing on carbon and nitrogen recycling.
Mineralogical Magazine, Goldschmidt Conference
(Toulouse) 62A, 276–7.
CHOPIN C. & SOBOLEV N. V. 1995. Principal mineralogical indicators of UHP in crustal rocks. In Coleman
R. G. & Wang X. eds. Ultrahigh Pressure Metamorphism, pp. 96–131. Cambridge University Press,
Cambridge.
CHRENKO R. M., TUFT R. E. & STRONG H. M. 1977.
Transformation of the state of nitrogen in diamond.
Nature 270, 141–4.
CLAOUÉ-LONG J. C., SOBOLEV N. V., SHATSKY V. S. &
SOBOLEV A. V. 1991. Zircon response to diamond-
Diamond growth during UHP metamorphism
pressure metamorphism in the Kokchetav massif,
USSR. Geology 19, 710–13.
DAVIES G. & EVANS T. 1972. Graphitization of diamond
at zero pressure and high temperature. Proceedings
of the Royal Society of London 328, 413–27.
DE CORTE K., CARTIGNY P., SHATSKY V. S., DE PAEPE
P., SOBOLEV N. V. & JAVOY M. 1999. Characteristics
of microdiamonds from UHPM rocks of the
Kokchetav Massif (Kazakhstan). In Gurney J. J.,
Gurney L. G., Pascoe M. D. & Richardson S. H.
eds. Proceedings of the 7th International Kimberlite Conference, April 13–17, 1998, pp. 174–82.
Cape Town, Red Roof Design cc, Goodwood, South
Africa.
DE CORTE K., CARTIGNY P., SHATSKY V. S., JAVOY M. &
SOBOLEV N. V. 1998a. First evidence of inclusions
in metamorphic microdiamonds from the Kokchetav
Massif, Northern Kazakhstan. Geochimica et Cosmochimica Acta 62, 3765–73.
DE CORTE K., SHATSKY V. S., TAYLOR W., CARTIGNY P.,
KORSAKOV A. & TRAUTMAN R. 1998b. What in situ
microdiamonds tell us about UHPM (extended
abstract). In International Workshop on UHP
Metamorphism and Exhumation, December 3–6,
1998, pp. A-67-69. Stanford University, CA, USA.
DOBRZHINETSKAYA L. F., BRAUN T. V., SHESHKEL G. G.
& PODKUIKO Y. A. 1994. Geology and structure of
diamond bearing rocks of the Kokchetav massif
(Kazakhstan). Tectonophysics 233, 293–313.
EVANS T. 1979. Changes produced by high temperature
treatment of diamond. In Field J. E. ed. The Properties of Diamond, pp. 403–25. Academic Press,
London.
EVANS T. & QI Z. 1982. The kinetics of the aggregation
of nitrogen atoms in diamond. Proceedings of The
Royal Society of London A381, 159–78.
HERMANN J. & GREEN D. 1999. Experimental constraints on continental rocks in ultra-high pressure
metamorphism. In 9th Annual V. M. Goldschmidt
Conference, August 22–27, 1999, p. 123. Harvard
University, Cambridge, Massachusetts.
HERMANN J., RUBATTO D., KORSAKOV A. & SHATSKY V.
S. 1999. Age and exhumation rates of diamondiferous, deeply subducted continental crust in the
Kokchetav massif, Kazakhstan (abstract). In 4th
International Eclogite Field Symposium, August
14–18, 1999, pp. 12–13. Kokchetav, Kazakhstan.
JAVOY M. & PINEAU F. 1991. The volatile record of a
‘popping’ rock from the Mid-Atlantic ridge at 14°N:
Chemical and isotopic composition of gas trapped in
the vesicules. Earth and Planetary Science Letters
107, 598–611.
KATAYAMA I., ZAYACHKOVSKIY A. A. & MARUYAMA S.
1998. A window to prograde metamorphism of the
Kokchetav massif, northern Kazakhstan (extended
abstract). In International Workshop on UHP
Metamorphism and Exhumation, December 3–6,
1998, pp. A-91-95. Stanford University, CA, USA.
437
LAVROVA L. D., KARPENKO S. F., LYALIKOV A. V. et al.
1997. Diamond formation in the age succession of
geological events in the Kokchetav Massif: Evidence
from isotopic geochronology. Geochemistry International 35, 589–95.
LEECH M. L. & ERNST W. G. 1998. Graphite pseudomorphs after diamond? A carbon isotope and spectroscopic study of graphite cuboids from the
Maksyutov Complex, south Ural Mountains, Russia.
Geochimica et Cosmochimica Acta 62, 2143–54.
MARAKUSHEV A. A., PERTSEV N. N., ZOTOV I. A.,
PANEYAKCH N. A. & CHERENKOVA A. F. 1995. [Some
petrological aspects of diamond genesis.] Geology of
Ore Deposits 37, 105–21 (in Russian).
MARTINEZ I., ZHANG J. & REEDER R. 1996. In situ
X-ray diffraction of aragonite and dolomite at high
pressure and temperature: Evidence for dolomite
breakdown to aragonite and magnesite. American
Mineralogist 81, 611–24.
NADEJDINA E. D. & POSUKHOVA T. V. 1990. The morphology of diamond crystals from metamorphic
rocks. Mineralogicheskiy Zhurnal 12, 3–15.
OKAMOTO K., LIOU J. G., OGASAWARA Y. & MARUYAMA
S. 2000. Petrology of the diamond-grade eclogite in
the Kokchetav Massif, northern Kazakhstan. The
Island Arc 9, 382– 402.
PEARSON D. G., DAVIES G. R. & NIXON P. H. 1995. Orogenic ultramafic rocks of UHP (diamond facies)
origin. In Coleman R. G. & Wang X. eds. Ultrahigh
Pressure Metamorphism, pp. 456–510. Cambridge
University Press, Cambridge.
PEARSON D. G. & NIXON P. H. 1996. Diamonds in young
orogenic belts: Graphitized diamond from Beni
Bousera, N. Morocco, a comparison with kimberlitederived diamond occurrences and implications for
diamond genesis and exploration. African Geoscience Review 3, 296–316.
POLI S. & SCHMIDT M. W. 1995. H2O transport and
release in subduction zones: Experimental constraints on basaltic and andesitic system. Journal of
Geophysical Research 100, 22299–314.
ROBINSON D. N. 1979. Surface textures and other features of diamonds, PhD thesis, University of Cape
Town, Cape Town, South Africa.
SCHIDLOWSKI M. 1983. A 3,800-million-year isotopic
record of life from carbon in sedimentary rocks.
Nature 333, 313–18.
SCHMIDT M. W. 1996. Experimental constraints on recycling of potassium from subducted oceanic crust.
Science 272, 1927–29.
SHATSKY V. S., RYLOV G. M., YEFIMOVA E. S., DE
CORTE K. & SOBOLEV N. V. 1998. Comparative study
of microdiamonds from metamorphic rocks, kimberlite and alluvial placers. Russian Geology and
Geophysics 39, 942–55.
SHATSKY V. S., SOBOLEV N. V. & VAVILOV M. A. 1995.
Diamond-bearing metamorphic rocks of the
Kokchetav Massif (northern Kazakhstan). In
438
K. De Corte et al.
Coleman R. G. & Wang X. eds. Ultrahigh Pressure
Metamorphism, pp. 427–55. Cambridge University
Press, Cambridge.
SHATSKY V. S., SOBOLEV N. V. & YEFIMOVA E. S. 1989.
Morphological features of accessory microdiamonds
from metamorphic rocks of the earth’s crust (extended abstract). In 28th International Geological Congress, Workshop on diamonds, pp. 94–5,
Bochum, Germany.
SIMAKOV S. K. 1995. Diamond formation in metamorphic crustal rocks. Transactions (Doklady) of the
Russian Academy of Sciences 343, 182–6.
SOBOLEV N. V. & SHATSKY V. S. 1990. Diamond inclusions in garnets from metamorphic rocks: A new
environment for diamond formation. Nature 343,
742–6.
SUNAGAWA I. 1984. Morphology of natural and synthetic
diamond crystals. In Sunagawa I. ed. Materials
Science of the Earth’s Interior, pp. 303–30. Terra
Scientific Publishing Company, Tokyo.
TAYLOR W. R., CANIL D. & MILLEDGE H. J. 1996. Kinetics of Ib to IaA nitrogen aggregation in diamond.
Geochimica et Cosmochimica Acta 60, 4725–33.
VEIZER J. & HOEFS J. 1976. The nature of 18O/16O and
13
C/12C secular trends in sedimentary carbonate
rocks. Geochimica et Cosmochimica Acta 40,
1387–95.
ZHANG R. Y., LIOU J. G., ERNST W. G., COLEMAN R. G.,
SOBOLEV N. V. & SHATSKY V. S. 1997. Metamorphic
evolution of diamond-bearing and associated rocks
from the Kokchetav Massif, northern Kazakhstan.
Journal of Metamorphic Geology 13, 479–96.