SL,..what is it?
http://www.cg.nrcan.gc.ca/slave-kaapvaal-workshop/abstracts/agashev.pdf
A unique kimberlite-carbonatite primary association in the Snap Lake dyke
system, Slave Craton: evidence from geochemical and isotopic studies.
A.M. Agashev 2,4 , N.P. Pokhilenko 1,2 , J.A. McDonald 1 , E. Takazawa 3 , M. A. Vavilov 2 , N.V.
Sobolev 2 , T. Watanabe 4
1. Diamondex Resources Ltd. Vancouver B.C. Canada.
2. Institute of Mineralogy and Petrography, Novosibirsk, Russia.
3. Niigata University, Niigata, Japan.
4. Hokkaido University, Sapporo, Japan
Results are presented of a geochemical study of the Snap Lake (SL) kimberlite dyke located
within the Slave province. The dyke mainly comprises hypabyssal facies macrocrystic kimberlite with
small but variable amounts of wall-rock fragments together with minor amounts of aphanitic
carbonatite rocks that may also contain wall-rock fragments. These rocks provide a unique opportunity
to study the primary continuum that accompanies a low degree of partial melting in the mantle.
The close spatial and temporal association of kimberlites and carbonatites has long been
recognized and has led to the proposal that these rocks are genetically related (e.g. Dawson, 1966, and
many later works). Eggler’s melting experiments in the CO2-bearing peridotite system (1974)
demonstrated the possibility of obtaining low SiO2 melts. More detailed experiments showing the
relationship of carbonatite, kimberlite, and carbonated lherzolite (Wyllie and Huang, 1975; Wyllie,
1980) provided further arguments for the genetic relationship of these rocks. Radiogenic isotope
studies of carbonatites and kimberlites revealed that they share a common source with OIB (Ocean
Island Basalts) assuming a convectively mixed asthenospheric mantle (e.g. Bell et al, 1973; Smith,
1983).
Kimberlite of the SL dyke has been dated by Rb-Sr method. using phlogopite. Three phlogopite
fractions separated from sample (SL 6) yielded a broad range of 87 Rb/86 Sr (17 to 65) and 87 Sr/86 Sr
ratios (0.798-1.039). The isochron defined by the three phlogopite fractions and a whole-rock
kimberlite gives an age of 535±11Ma (MSWD=9). This corresponds closely with an isochron obtained
independently by Geospec Consultants Ltd. (internal report for Winspear Resources Ltd., 1999) using
five phlogopite analyses that produced an age of 522.9+/- 6.9 Ma (MSWD=0.18).
Major element compositions of the studied rocks range from magnesiocarbonatites with SiO2
contents of~ 3 wt% through to typical kimberlite with a composition approaching that of mantle
peridotites. All major element data are plotted on the MgO/CaO –MgO/SiO2 ratios plot (Fig. 1)
together with experimental data of Dalton and Presnall (1998) for melting in a CaO- Mg0-Al203- SiO2-CO2
system at pressure of 6 Gpa. Rocks with CaO contents higher than 12% plot on the carbonated
peridotite melting line and will be referred to in the text as carbonatites. Typical kimberlites define a
separate field trending to olivine composition. It is generally accepted that the major element
composition of kimberlites is modified by addition of desegregated mantle minerals, mostly olivine
(Mitchell, 1995; Price et al, 2000). To asses this possibility, we have calculated the mixing line
between compositions of 12 wt % of CO2 and olivine with Mg#=92. The observed kimberlite
compositions require a 30%-80% dilution by mantle olivine. Dilution processes must lead to a decrease
in the concentration of trace elements in the resulting kimberlite. There is no systematic decrease of any
trace elements shown by our data with the increase of MgO contents. Therefore we conclude that
hypabyssal kimberlite comprising the SL dyke represents a primary kimberlite melt with very little or
no addition of desegregated mantle material.
Kimberlite from the Snap Lake dykes is enriched in most incompatible trace elements and in
elements of ultramafic affinity (Cr and Ni). These features are typical for kimberlites. The micaceous
nature of the Snap Lake kimberlite together with its Sr isotope ratio of 0.71515, point to geochemical
similarities with Group I kimberlites from South Africa and Siberia. The plot of trace element
abundances of SL dyke kimberlites normalized to primitive mantle (Fig. 2) has a shape generally
similar to Group I kimberlites with a positive Nb anomaly and negative K anomaly. Trace elements
ratios such as Nb/Zr >1 and La/Nb <1 are typical of Group I kimberlites. An initial Sr isotope ratio of
0.7045 and å Nd = 0.3-0.8 are also in the range of Group I kimberlites. However, the composition of SL
dyke kimberlites has several distinct features. They are enriched in the most incompatible elements (Pb,
Rb, Ba, Th, Ce, La), slightly depleted in TiO2 contents, and significantly depleted in Sr compared to
averages of South African Group I and Siberian kimberlites. The Rb/Sr ratio is higher and the Sm/Nd
ratio is lower than that in Siberian kimberlites and Groups 1 and 2 kimberlites from South Africa.
Carbonatites of the SL dyke have clearly different trace element concentrations compared to SL
kimberlites. Although the four analysed carbonatite samples show a range in composition, they are
systematically enriched or depleted in particular elements (Fig. 2). They are depleted in Cs, Rb, K, Ta
and Ti, and enriched in U, Sr, P, Zr, Hf, as well as in middle and heavy REEs relative to SL
kimberlites. On the normalized primitive mantle plot, average SL carbonatite shows distinctive U, Y
and heavy REE positive anomalies and a negative anomaly for Ti. Normalized to chondrites, the REE
pattern is less fractionated in carbonatites than that of kimberlites and is similar to carbonatites from
Yacupiranga (Nelson et al, 1988). However, these observed compositional differences in the natural
system are not compatible with experimental data demonstrating the distribution of trace elements
between immiscible carbonate and silicate liquids. The only obvious similarities are depletion of Ti,
fractionation of Hf from Zr, enrichment of Sr and P in carbonatite liquids, and the neutral behavior of
Nb and LREE (Green et al, 1994; Veksler et al, 1998). In the experimental systems, heavy REE, Y, and
Zr are preferably incorporated into silicate liquid, whereas Rb, K, and Ba preferentially partition into
the carbonate-rich liquid. This is in contrast to the observed natural carbonatite-kimberlite system from
SL. Consequently, we suggest that the composition of the SL magmas reflects a primary partial melting
sequence of carbonated mantle peridotite in which the first liquid was magnesiocarbonatite. At that
time, most, if not all, carbonate was consumed in the source. With elevation of temperature, the degree
of partial melting exceeded 1% and kimberlite liquid was formed.
The Nb/Ta and Zr/Hf ratios of carbonatites are particularly interesting because these twin
element pairs should not fractionate during partial melting. Both these ratios are significantly
superchondritic in carbonatites and display a negative correlation with SiO2 contents, being highest in
sample with the lowest SiO2 content. The Nb/Ta ratios of SL carbonatite range from 81 (sample with
SiO2= 3.7%) to slightly superchondritic (sample with 17 wt% SiO2), whereas Nb/Ta ratio in kimberlite
scatters around the chondritic average of 18. Fractionation of Zr/Hf in carbonatite melts has previously
been reported in metasomatised xenolith suites from Australia (Yaxley et al, 1998) and Tanzania
(Rudnick et al, 1993). In the SL carbonatite, Zr/Hf ratios (56-120) and kimberlite (56 in average) are
substantially higher than the chondritic value of 36. The concentration of U and the Th/U ratio are
important distinctive features of SL carbonatites, which contain 7-41 ppm U and have a Th/U ratio of
0.6-3.5. SL kimberlites contain only 2.3-4.5ppm U and have significantly higher Th/U ratios of 7-16.
References:
A.M. Agashev., T Watanabe., K. Maehara., D.A. Budaev., N.P. Pokhilenko., A.S. Fomin and J. Maeda. (2001).
Geology, v3, p 267-270.
Dalton J.A and Presnall D.C.(1998). J.of Petrology, v 39: 1953-1964.
Dawson J.B. (1966). In Naidu P.R.J., ed., Papers and Proceedings of the 4 th General Meetings, Intl. Mineral.
Assoc., IMA Volume, Mineral.Soc.India, 1-4.
Green T. H. (1994). Chemical Geology, v 117: 1-36.
McDonough W.F., Sun S.-s.[1995]. Chemical Geology, v 120,.pp223-253.
Nelson D.R., Chivas A. R., Chappel B.W., and McCulloch M.T. (1988). Geochem. Et Cosmochem. Acta, v 52:
1-17.
Rudnick R.L, McDonough W. F., and Chappel B.W. (1993). EPSL, v 144, 463-476.
Wyllie P. J. and Huang W.-L, (1975).. Geology v 3: 621-624.
Wyllie P. J.,(1980). J.G. Res. v 85: 6902-6910.
Yaxley M.G., Green D.H., and Kamenetsky V. (1998). J. of Petrology, v 39:1917-1930. |
