This website uses cookies to ensure you get the best experience on our website. Further info

Contact us

Tel. +358 29 503 0000
- Register office:
- Personal mail:

Offices and contact information

Management and contact persons

Mineralogy and Geochemistry of Kimberlites and Related Rocks from Finland 

Hugh E. O'Brien and Matti Tyni

(printed version published in the Proceedings of the 7th International Kimberlite Conference, 1999, p.625-636.)

Geological Survey of Finland, P.O. Box 96, FIN-02151 Espoo, Finland


More than twenty-four intrusive bodies, mostly kimberlitic diatremes, but also including hypabyssal kimberlites, olivine lamproites and ultramafic lamprophyres have been discovered within the Archean Karelian Craton in Finland. A subset of these, twelve pipes and dikes, have been studied in detail. The kimberlites occur in two clusters at the edge of the Archean Craton, near the towns of Kuopio and Kaavi in northeastern Finland. Their mineralogy is typical of Group I kimberlite including abundant macrocrysts of olivine, picroilmenite, Cr-diopside, pyrope garnet; phenocrysts of olivine and microphenocrysts of monticellite, perovskite, kinoshitalite mica and spinel in a calcite + serpentine matrix. Notably phlogopite is not abundant, and there are no indications of Group II kimberlite (orangeite) affinities in these rocks. Major elements indicate that despite the great care that was taken in whole rock powder preparation, there is a crustal component in some of the samples. Nevertheless, trace elements show that this assimilation has had mostly a dilutive effect and that the Finnish kimberlites in this study form a compositionally cohesive group of rocks more akin geochemically to the Koidu type kimberlites of West Africa than the Group I kimberlites from South Africa. Sr-Nd isotopic compositions either reflect derivation from a near bulk-earth mantle or represent assimilation and reaction of radiogenic Karelian Craton lithospheric mantle by kimberlite magmas that originally had Group I isotopic signatures.

Farther into the Karelian Craton occur dikes of carbonate-rich ultramafic lamprophyre (aillikite) and a rock which appears to have both olivine lamproite and Group II kimberlite affinities. The latter contains most of the typomorphic minerals of lamproite including phlogopite, K-richterite, low-Al diopside and has a major element composition equivalent to average olivine lamproite. Despite this the rock has tetraferriphlogopite zoning profiles and an accessory Ca-Zr silicate that differ from other reported lamproites and are in fact more typical of those seen in Group II kimberlites (orangeites).

Keywords: kimberlite, lamproite, orangeite, aillikite, Finland, Nd-Sr isotopes, petrography, mineralogy, geochemistry.


In January 1997 the Geological Survey of Finland initiated a study of Finnish kimberlitic intrusions discovered by Malmikaivos Oy as a result of their diamond exploration activities (Griffin et al., 1995; Tyni, 1997). These Paleozoic intrusions, two pipes have K-Ar ages of 590 and 430 Ma (Tyni, 1997), form the Eastern Finland Kimberlite Province which includes the Kaavi and Kuopio clusters. The pipes and dikes intruded into 3.1-2.6 Ga gneiss complexes of the Archean Karelian craton and allochthonous 1.9 - 1.8 Ga metasedimentary cover rocks thrust onto the craton during the Svecofennian orogeny (Kontinen et al., 1992). The intrusions are dominated by textural and mineralogical variants of archetypal kimberlites and range from purely hypabyssal kimberlite dikes to multiphase pipes of diatreme facies rocks. The Finnish kimberlites generally contain abundant diamond indicator minerals and almost all contain at least trace amounts of microdiamonds. In addition to the kimberlites, a spatially distinct and presumably older group of dikes located elsewhere in the Karelian Craton comprise ultramafic lamprophyre and rocks that have both olivine lamproite and Group II kimberlite affinities. Of the 24 intrusive bodies discovered by Malmikaivos, drill core samples have been taken from 12 from which 30 samples have been analyzed for major and trace elements and 12 for Sr and Nd isotopic compositions.

1.1 Historical Background

The first kimberlite in Finland was discovered in 1964 by Malmikaivos Oy, a private prospecting company, based in Luikonlahti in eastern Finland (Tyni, 1997). The pipe was found during regional prospecting for base metals in the vicinity of the Luikonlahti copper mine, which was under development at the time. In ground magnetic mapping the body appeared as a strong circular anomaly which was interpreted to represent a vertical pipe of about one hectare in size. Till overburden was only two meters thick and in the trenching and drilling a strange dark rock was exposed. This unmetamorphosed rock looked very odd macroscopically as well as in thin section. Because there are a great number of lamprophyric dikes in the area, the rock was connected to them although the possibility of it being a kimberlite was also considered. Since this body had no copper potential, the prospecting objective at the time, the discovery of this pipe was nearly forgotten.

Late in the 1970's, during further base metal prospecting in the area surrounding the copper mine, glacial boulders of well preserved "almond rocks" were found. The name referred to the numerous white almond-shaped inclusions that were later proven to be altered peridotite xenoliths. Samples of both of these strange rocks were sent to diamond companies where they were correctly identified and found to contain microdiamonds. This naturally raised considerable interest toward Finland.

Malmikaivos located kimberlite Pipe 2 in 1984 under a small swamp, about one kilometer up-ice from the previously mentioned glacial boulders and Pipe 3 in 1985 under a small lake only 500 meters away from Pipe 2. Because of insufficient expertise specific to diamond exploration and diamond prospect evaluation, Malmikaivos decided in 1985 to approach Ashton Mining Limited (AML) from Australia for assistance. The technical cooperation and funding agreement was signed in 1986 with the aim to prospect for and develop economic kimberlites in Finland. On January 1, 1994 Ashton acquired 100% of Malmikaivos. By 1996 Malmikaivos/Ashton had discovered 24 intrusive bodies, most of them kimberlitic in composition and most of them diamondiferous, twelve of which form the basis of this study.



The main features of each of the intrusions, the rock types and mineralogies are given in Table 1 and summarized briefly below, further details are available in O'Brien and Tyni, 1997.

2.1 Intrusion Morphology and Facies

The kimberlites from northeastern Finland range from <1 hectare to almost 4 hectares. Surface plan maps of the intrusives included in this study are shown in Fig. 1 and the general characteristics of the intrusions are given in Table 1. The intrusions range from circular to elliptical pipes to dikes up to 0.5 km in length. Surface exposure of the pipes is poor, however, Pipes 1 & 2 are presently exposed due to recent exploration and diamond evaluation activities. Tuffisitic diatreme facies rocks are dominant, although several of the intrusions are composed entirely of hypabyssal material (Table 1). Evidence for possible crater facies material is limited to pipe 5 which contains volcaniclastic rocks that vary from breccias to ash-rich sandstones (photo). It is unclear whether the sandstones represent downrafted blocks of crater facies material or simply more completely comminuted diatreme material.


2.2 Xenoliths

Ultramafic to mafic mantle xenoliths are relatively common in pipes 5, 9, 10 and 14 and comprise a suite ranging from garnet websterite to lherzolite, harzburgite and eclogite (see Peltonen et al., in press, for details). Pipe 10 (photo) contains a particularly rich variety of eclogites (including amphibole, apatite and carbonate-bearing types), phlogopite-bearing peridotites and Cr-rich to Cr-poor lherzolites. Abundant xenoliths of lower and upper crustal wallrocks also occur in the kimberlites, ranging from granites to sulfide-rich black schists, quartzites and two pyroxene granulites. Typically the largest population of crustal xenoliths in any given pipe represents the highest level wallrocks encountered, but in all cases xenoliths of felsic gneisses derived from Archean gneiss basement occur. Spectacular tuffisitic breccias from pipe 3 contain abundant xenoliths of quartzite, gneisses and granitoids. However, the xenolith population of the tuffisitic rocks varies dramatically among drill cores located only a few tens of meters apart, and the quartzite xenoliths do not occur in any other drill core intercepts.

2.3 Kimberlite Samples

Pipe 1, the "strange rock" discovered in 1964, is the only pipe that does not contain breccia, is almost xenolith-free, and represents the least contaminated kimberlite in this suite of samples. As it is also by far the least altered, it makes an ideal endmember with which to compare the remaining samples. Containing up to 40% olivine as macrocrysts (photo) and phenocrysts in a matrix of monticellite microphenocrysts, perovskite, chromite zoned to magnesian ulvöspinel-magnetite (MUM), calcite, serpentine, and kinoshitalite mica, the main phase of the pipe is mostly massive but about 10% has a carbonate segregation-texture. The marginal phase of the pipe is quite distinct and is formed of autoliths of the kimberlite magma cemented into a carbonate matrix (25% by volume; see photo). These autoliths contain abundant spheres of calcite that likely record the separation of an immiscible carbonate magma that coalesced near the margin of the pipe to form the carbonate matrix.

The freshest samples of the other nine kimberlites have the same matrix minerals as those in Pipe 1, but in most cases only pseudomorphs of monticellite remain, the bulk of olivine has been serpentinized and there is more sphene as opposed to MUM. An additional important component in the other kimberlites is varying amounts of xenoliths (described above), xenocrystal peridotite detritus (esp. olivine, chromite, chrome pyrope, chrome diopside), megacryst suite garnet, ilmenite, pyroxene and abundant crustal detritus in the diatreme breccias. Some parts of pipes 5 and 14 in particular are crystallinoclastic, comprising essentially disaggregated peridotite minerals in minimal kimberlite matrix.

2.4 Dike with Olivine lamproite/Group II kimberlite (Orangeite) affinities

In hand sample this dike rock has the light brown color of phlogopite because microphenocrysts of this mineral comprise nearly 70% of the rock. Approximately 20% serpentinized olivine macrocrysts/phenocrysts occur with microphenocrysts of phlogopite, K-richterite, diopside, apatite and perovskite in a serpentine and calcite matrix. This rock is intriguing because it has some affinities to Group II kimberlites (orangeites in the terminology of Mitchell, 1995). It contains an accessory calcium zirconium silicate (Ca-catapleite?) rather than the lamproite typomorphic mineral wadeite and zoning in micas that follow the extreme Al depletion evolutionary trend seen in orangeite micas (see discussion below). Our drill core sample from this dike contains a 3 cm wide, partially serpentinized garnet lherzolite xenolith.

2.5 Ultramafic Lamprophyre Dike

Roughly 1 km long and 5 m wide, this dike rock is composed mostly of dark brown phlogopite crystals and subordinate serpentine pseudomorphs (after olivine?), in a olive green matrix of serpentine and carbonate containing accessory ilmenite, titanomagnetite, rutile and apatite. Compositions and zoning profiles of phlogopite (see discussion below) and oxides indicate this is not a kimberlite or lamproite but rather belongs to the ultramafic lamprophyre group. Its high CO 2 content helps classify it as an aillikite (Rock, 1990). So far no diamond indicator minerals have been found in this sample.


3.1 Analytical Conditions

Microprobe analyses were carried out with the CAMECA SX-50 at the Geological Survey of Finland. For minerals containing Na 2O (phlogopite, clinopyoxene, amphibole), analytical conditions were 15 kV accelerating voltage, 20 nA cup current, 3-5 µm beam. In order to decrease detection limits for trace elements, the more robust minerals (garnet, pyroxene, ilmenite, olivine) were analyzed at 20 kV accelerating voltage, 48 nA cup current, and a fully focused beam. Data were reduced using the PAP program supplied by Cameca.

3.2 Kimberlite minerals

3.2(a) Olivine

Fresh olivine exists only in pipes 1, 5, 9 and 14. The forsterite-rich olivines from pipes 1 and 14 have a restricted compositional range, from Fo 87 to Fo 92 which includes olivine from peridotite xenoliths, rounded macrocrysts and smaller euhedral to subhedral phenocrysts that presumably crystallized from the kimberlite magmas. The larger olivine grains are typically unzoned but it is common for the smaller phenocrysts to show core Fo 90 to rim Fo 88 variations. Olivines from pipe 5 and 9 appear to have a bimodal compositional distribution with a Fo 93-89 population like those described above and a Fo 86-83 population representing megacryst-suite olivine.

3.2(b) Monticellite

Monticellite is common in Pipe 1, rare in Pipe 14 and found only as inclusions in mica and sphene in pipe 10. It is absent from the remaining pipes which are all in less pristine condition. Euhedral monticellite grains, mostly from 10 to 50 micrometers in size, form the main component of the groundmass in the Pipe 1 samples. Those that occur in Pipes 10 and 14 are partially altered microphenocrysts, quite similar in size and composition to those in Pipe 1.


3.2(c) Mica

Ba-rich mica (kinoshitalite) occurs in the groundmass of virtually all of the Finnish kimberlite samples. Its abundance ranges from very sparse as in Pipe 9, to as much as 10% of the matrix, as in Pipe 10. Relative to the kinoshitalite micas reported from the Iron Mt. kimberlite (Mitchell, 1995), the Finnish examples range to extremely Ba-rich compositions (up to 17.8 wt% BaO), especially those from Pipe 1, and also contain a large amount of fluorine (Table 2).

3.2(d) Spinel

Pipe 1 samples contain up to 5% spinel, which uniformly have spectacular atoll structures. In thin section translucent dark red chromite cores are surrounded by an amorphous serpentine-like material which is in turn mantled by magnesian ulvöspinel and surrounded by an outer thin rim of magnetite (Fig. 2a). However, in rare cases it is possible to find remnant pleonaste spinel that originally filled the "lagoon" in all of these atoll spinel grains (Fig. 2b). In detail the chromite cores are typical kimberlite spinel magmatic trend 1 (Pasteris, 1983, Mitchell, 1986) titanian aluminous magnesian chromite (TIMAC) but are abruptly zoned to titanium-bearing, nearly chrome-free pleonaste spinel (Table 3), quite away from the normal zoning in Trend 1 spinels (op.cit.). The MUM mantles are practically unzoned (Table 3) and a few, in addition to an outer rim of magnetite, have an inner, discontinuous magnetite rim of the same composition (see Fig. 2b, c). Pipes 2, 3 and 5 also contain atoll spinels analogous to those described above although they are less abundant and suffer from the relatively altered condition of the host rocks. Somewhat similar atoll spinels with considerably thinner pleonaste rims and titanomagnetite cores instead of chromite were described from the De Beers kimberlite (Pasteris, 1983).


Additional information on the growth of the pleonaste spinels comes from mantles developed on large olivine macrocrysts and rounded olivine-monticellite-spinel cumulate fragments that form what we have termed "orbicules" in several Pipe 1 samples. In many respects these are similar to the "spheroids" described by Reid et al. (1975) from the Igwisi Hills. Figure 3a shows a portion of a large olivine crystal, several atoll spinels and phenocrysts of olivine (grains at center and right of image) rimmed by monticellite. An enlargement (Fig. 3b.) shows that crystallization of the complexly zoned mantle followed exactly that in the atoll spinels, with initial chromite, followed by rapid zoning to pleonaste followed by an abrupt zoning to MUM and a final discontinuous magnetite rim. Other orbicules display mantles composed only of pleonaste, only of MUM or are complexly zoned with pleonaste mantles rimmed by MUM.


3.2(e) Ilmenite

Magnesium-rich ilmenite (picroilmenite), a typical mineral found in kimberlites and representing one of the most important diamond indicator minerals, is abundant in pipes 4, 5, 6, 9, 10, 14 and 23. In a similar fashion to the spinels, ilmenite compositional zoning patterns in the Finnish kimberlites are complex. Thus far three main types of zoning have been identified (see Fig. 4). Representative analyses of ilmenite from each type of zoning are given in Table 4.

a. Decreasing Cr with increasing Mg toward rim. High Cr (ca. 5 wt% Cr 2O 3) picroilmenite macrocrysts from Pipe 14 (triangles in Fig. 4) have homogeneous cores, but show decreasing Cr within 20 microns of the crystal edge (Fig 5a). The lowest Cr, highest Mg picroilmenites occur just inside the very edge of the crystals, adjacent to a few micrometer wide, outer MUM rim.


b. Increasing Cr and Mg toward rim. This is the predominant type of zoning developed in picroilmenite macrocrysts with low Cr cores (circles in Fig. 4). Well documented in kimberlite picroilmenites (Mitchell, 1986), it is manifested as diffuse zoning across most of the crystal at a relatively constant Cr content and from 10 to 12 wt% MgO. However, near the crystal edge a much more abrupt zoning pattern is developed, with a large increase in MgO and Cr 2O 3 over distances of tens of micrometers. At the very edge of the crystal a ca. 5 μm reaction rim of MUM spinel occurs on these grains. The high Cr ilmenite reaction zone mimics the shape of the MUM rim and where there are embayments of MUM farther into the ilmenite crystal, the high Cr picroilmenite reaction zone rather uniformly outlines the embayment (BSE image). However, in some cases the core composition occurs right up to the present crystal edge, which probably reflects fracturing of the crystal shortly before pipe emplacement.


In the same sample, another version of zoning with an increase in Cr and Mg toward the rim occurs in picroilmenites from peridotite microxenoliths, such as that shown in Fig. 5b. The microxenolith consists of a large Fo 89olivine, a very magnesian (ca. 14 wt% MgO) ilmenite (black squares, Fig. 4) and between them a patch of chalcopyrite and associated djerfisherite. Olivine included in the picroilmenite is also Fo 89, and this, coupled with the rounded shape of the microxenolith, indicates that the fragment represents a mantle sample. Developed only at the outer edge of the microxenolith, where the ilmenite was exposed to kimberlite magma, is distinct zoning to more MgO-rich compositions (yellowish orange, Fig. 5b.) and a sharply defined MUM outer rim (light yellow). There is no zoning in the ilmenite toward the area of serpentinized olivine, indicating that this late stage process had no effect on ilmenite compositions. Chrome and Mg-rich picroilmenite inclusions are relatively common in olivine macrocrysts in this sample and have very similar compositions (pink squares, Fig. 4) to the microxenolith ilmenite. There is no difference between those inclusions that occur in fresh olivine and those that occur in serpentinized olivine further demonstrating that late stage serpentinizing fluids had no affect on the ilmenites and in particular did not contribute to the development of the spinel reaction rims.


c. Increase in Mg, no increase in Cr toward rim. A completely different zoning trend from b. is developed in many low Cr picroilmenite cores and derived fragments; one of increasing Mg at nearly constant Cr toward the grain rim (trend c, Fig. 4). The example shown in Fig. 5c has a relatively wide reaction zone and a correspondingly wide rim of MUM spinel. Remnant patches of ilmenite within the MUM rim attest to the formation of the latter by the reaction ilmenite + liquid = MUM. Supporting this is the fact that the spinel rim compositions reflect their ilmenite precursors systematically (Fig. 4), i.e., the highest Cr ilmenite has the highest Cr spinel reaction rim and so on.


Distinct from the ilmenites described above are the extremely Cr-rich examples from Pipe 9. These contain up to 12 wt% Cr 2O 3 (Table 4) and consist of an ilmenite host with submicron Cr-rich planar domains that appear to be chromite exsolution lamellae.

3.2(f) Apatite and Perovskite

Abundant apatite occurs as acicular grains commonly in radiating stellate clusters in the groundmass and as larger more prismatic grains grown primarily within calcite segregations. They are relatively Si-rich (0.7-1.1 wt% SiO 2) and Sr-poor (<1 wt% SrO), characteristic of Group I kimberlite apatite (Mitchell, 1995).

The majority of the perovskite occurs as euhedral to subhedral discrete grains (rarely as aggregates) that are 0.02 to 0.1 mm across. Our limited data on perovskite shows them also to be typical of Group I kimberlites (op. cit.) with ca. 1.3 wt% FeO, 1.5 wt% Nb 2O 5 and 0.1-0.3 wt% SrO.

3.2 Olivine Lamproite/Group II kimberlite (Orangeite)

Although giving a name to this dike rock seems to be problematic, its petrography and mineralogy is relatively simple. The more common texture is porphyritic with pseudomorphs of olivine rimmed by perovskite in a matrix of subhedral phlogopite and lesser K-richterite with a mesostasis of calcite and serpentine (right hand side of Fig. 6). However the sample also contains apatite-rich residual pockets of late-stage melt (probably representing devitrified glass) which contain crystals of phlogopite, tetraferriphlogopite, low-Al diopside and K-richterite crystallized at the edges (left side of Fig. 6). Phlogopite microphenocrysts occur as brown to reddish brown, tabular plates that are weakly pleochroic and zoned to reversely pleochroic tetraferriphlogopite. It has been suggested that the compositions of phlogopite may be diagnostic vis-à-vis distinguishing among Group I kimberlite, Group II kimberlite (orangeite), lamproite and lamprophyre (Mitchell, 1995; Mitchell and Bergman, 1991). Phlogopite microphenocrysts from this dike rock (Table 2; Fig. 7) have core compositions that plot within the lamproite field, but core to rim zoning trends are not toward TiO 2-enrichment as seen in lamproites but rather toward extreme TiO 2-depletion as is observed in phlogopite from orangeites (Mitchell, 1995). Analyses of other minerals from this dike including K-richterite, low Al diopside, perovskite, spinel, apatite and a relatively abundant Ca-Zr silicate accessory mineral show them to be quite typical of olivine lamproites and more evolved orangeites except for the last mineral, which may be Ca-catapleite. It has not been reported from lamproites (Mitchell, 1995).


3.2 Aillikite

Compositions of phlogopite, ilmenite, titanomagnetite, apatite and perovskite in this sample all fit within the ultramafic lamprophyre spectrum of minerals. The TiO 2 and Al 2O 3-rich phlogopite compositions in particular are diagnostic of lamprophyre (Fig. 7).


4.1 Sample Preparation and Analysis

Preparation of the thirty samples analyzed in this study involved crushing approximately 500g of each into small chips with hammer, hand-picking under binocular microscope approximately 200g of chips least contaminated by crustal xenoliths and powdering the picked material in an trace element-free iron grinding vessel. Major elements were measured on glass fusion disks using a Philips PW 1480 X-ray fluorescence (XRF) spectrometer at the Geological Survey of Finland and pressed-powder pellets were used to measure the trace elements Ba, Sr, Ni, Cr, V, Cu, Zn, S, and Cl. LECO analyzers were used for CO 2 and H 2O. The remaining trace elements reported in Table 5 were analyzed by ICP-MS following dissolution by mixed hydrofluoric-perchloric acid and lithium metaborate/sodium perborate fusion of the insoluble residue. Sr and Nd isotopic compositions and Sr, Rb, Nd, and Sm concentrations by isotope dilution were analyzed on a VG Sector 54 mass spectrometer at the Geological Survey of Finland.

4.2 Kimberlites

Major elements of the Finnish kimberlites show that the samples vary from being unaltered and uncontaminated (e.g., Pipe 1) to highly altered and/or contaminated (e.g., tuffisitic diatreme samples from Pipes 4 & 6) with contamination index values (C.I.; Clement, 1982) ranging from 1 to 2.4, respectively. In an Al 2O 3 - SiO 2 diagram (Fig. 8) only a few of the samples plot within the uncontaminated field suggested by Mitchell (1986), but do plot within a field of 38 wt% SiO 2 and 6 wt% Al 2O 3 which we believe demarcates a group with limited contamination. On this same diagram contamination vectors are shown as arrows pointing in the direction of increasing contamination for Pipes 3 and 14, pipes for which we have multiple samples. Not unexpectedly these indicate granitoid to be the most significant contaminant. Mixing with countryrock that occurs during diatreme formation amplifies the contamination process enormously, exemplified by data on tuffisitic diatreme samples from Pipes 4 and 6 which contain 44 and 46 wt% SiO 2 and 5.5 and 6 wt% Al 2O 3, respectively, and plot along an extension of the contamination vector for Pipe 14 (Fig. 8).


4.3 Olivine Lamproite/Group II kimberlite (Orangeite) and Aillikite

The Dike 16 lamproite has a bulk composition that is very near that of the average olivine lamproite given by Mitchell and Bergman (1991). Only the relatively high concentration of CO 2 is Dike 16 seems to be anomalous (Table 5). Considering that all olivine has been converted to serpentine, it is quite possible some of this carbonate is secondary. The ultramafic lamprophyre (aillikite) is also carbonate-rich, some of which may also be secondary.


5.1 Kimberlites

Despite the indication for contamination from major elements, incompatible trace elements show that this contamination mostly had a dilutive effect. As a baseline for this comparison we use the three samples from Pipe 1 that have nearly identical trace element compositions and we believe represent the closest example to an uncontaminated kimberlite. Relative to this Pipe 1 composition, the more contaminated samples have lower concentrations of nearly all of the incompatible trace elements except K and Rb (Fig. 9). This figure only includes trace element profiles from a few samples for clarity, but the observation holds true for the entire sample set. Importantly, there is no evidence that the shape of the trace element profile was changed significantly by the assimilation process. This is probably simply a consequence of the very high trace element concentrations in the original kimberlite magma relative to the crustal assimilant.


Compared to average Group Ia and Group Ib kimberlites (Smith et al., 1985), the Finnish examples have very similar incompatible element concentrations (Fig. 9). The only systematic differences are the considerably lower concentrations of Hf and Zr and higher Ba in the latter. This is shown clearly in a plot of Zr-Nb (Fig. 10) which compares kimberlites from various localities around the world. The Finnish kimberlites plot with the relatively low Zr Koidu kimberlites of W. Africa forming a vector pointing in the direction of the Aries kimberlite composition (Taylor et al., 1994). From this diagram, coupled with evidence from other trace elements (Fig. 9) and similarity in mineralogy, it is possible to infer that the Finnish kimberlites represent a relatively cohesive group that before dilution by crustal material represented a restricted compositional range.



Initial 87Sr/ 86Sr isotopic compositions of the Finnish kimberlites in this study range from 0.7033 to 0.7055 calculated for ages of 434 Ma (Pipe 1), 593 Ma (Pipes 2 &3) and 450 Ma for the remainder (Table 6). Initial Nd isotopic compositions expressed as epsilon values range from +1 to -2.4. The Sr compositions overlap with and extend to more radiogenic values than Group I kimberlites (Smith, 1983) but the ε Nd values, centered on bulk earth composition, are slightly lower (Fig. 11). It must be noted that the Finnish kimberlites are not yet well dated (only two K-Ar ages) leaving open the possibility that this field may shift slightly once more age dating has been completed. It is not likely however, that this shift will be significant, and certainly not sufficient to move the field to overlap with the S. African Group I kimberlite data in any case.


Dike 16 has a Sr and Nd isotopic composition, recalculated for 1100 Ma, that is very near the field for S. African Group II kimberlites (Fig. 11). The age estimate comes from somewhat similar dikes near Kostamuksha, in Russian Karelia (Proskuryakov et al., 1990) but may not be relevant to Dike 16. Clearly intrusion ages for all of these samples are necessary before firm initial isotopic compositions can be calculated and for this reason Ar-Ar and Pb dating of perovskite on Dike 16 and five kimberlites is presently being undertaken.


7.1 Magma Mixing

A basic tenet of kimberlite petrology is that kimberlites are very complex and represent mixes of various batches of magma and their phenocrysts along with abundant mantle xenocrysts (Mitchell, 1986). They therefore do not represent true magmatic liquids. There is abundant evidence for this kind of mixing in the kimberlites discussed here, best exemplified by spinels in Pipe 1 and ilmenites in Pipe 14.

The groundmass spinels in Pipe 1 record a pleonaste crystallization event that is also recorded on the rims of some olivines and in some orbicules where pleonaste forms mantles and partially replaces olivine. The change from chromite to pleonaste is rather abrupt but does not appear to represent a resorbed surface and may simply represent the consequences of fractional crystallization. Large amounts of olivine fractionation without contemporaneous crystallization of an aluminous phase may have caused the build up in Al 2O 3 that led to pleonaste saturation. This aluminum-rich magma is not the same one that crystallized the MUM mantles after the pleonaste. The most reasonable mechanism to explain this is the mixing of the chromite-pleonaste cores into a MUM saturated magma of a somewhat evolved, but more typical kimberlite composition.

In some atoll spinel grains late magnetite also occurs on the inside of the MUM mantle within the zone of altered pleonaste (Fig. 2b). This probably indicates reaction of the late-stage groundmass forming magma with the pleonaste and contemporaneous crystallization of magnetite. The nearly pervasive alteration of pleonaste was therefore not due to reaction with late stage magmatic fluids or post intrusion deuteric/hydrothermal fluids but instead required a reaction with magma, reflecting either the third magma recorded in these atoll spinels or substantially different conditions during the latest stages of crystallization.

A similar case for several transient liquids can be made concerning the zoning patterns in picroilmenite described above. It appears the high Mg, high Cr cores in trend a equilibrated with a liquid with similar Cr content as that which caused the trend b zonation (Fig. 4). The trend c zonation without an increase in Cr however indicates another transient liquid, one low in Cr. After this zonation occurred, all of these crystals were mixed into the Pipe 14 magma at which point the final spinel reaction rims formed.

7.2 Affinities to Koidu and Aries kimberlites

Group I kimberlites have Nd isotopic compositions that appear to reflect sources in the well-mixed asthenospheric mantle (Smith, 1983) and have radiogenic isotope compositions that imply a source with long term low Rb/Sr and low Sm/Nd. The Finnish kimberlites in terms of trace element and isotopic compositions do not match completely with the S. African examples but instead are closer to the Koidu kimberlites from W. Africa. The latter have, as described by Taylor et al. (1994), a significant Aries kimberlite (northwest Australia) component which is characterized by low P 2O 5/Ce, Sr/Ce and high Nb/Zr. The Koidu and Finland kimberlites overlap completely in all of these parameters. On the other hand, the Koidu and Aries kimberlites are mica-rich, and accordingly have low Ba/Rb (<32). Although the Finnish kimberlites contain matrix microphenocrysts of kinoshitalite-phlogopite, they contain almost no phenocrystic mica and reflect this with more typical Group I kimberlite Ba/Rb ratios(>32).


The kimberlites from Finland exhibit typical mineralogy of Group I kimberlites from elsewhere in the world including abundant macrocrysts of olivine, picroilmenite, Cr-diopside, pyrope garnet; phenocrysts of olivine and microphenocrysts of monticellite, perovskite and spinel in a calcite + serpentine matrix. Notably phlogopite phenocrysts are not abundant, and there are no indications of Group II kimberlite (orangeite) affinities in these rocks. Detailed studies of the complex zoning in constituent minerals reveal wide variations in the compositions of the magmas that mixed to form the kimberlites discussed here. Major elements indicate that there is a crustal component in some of the samples although trace elements show that this assimilation has had mostly a dilutive effect. The Finnish kimberlites form a compositionally cohesive group of magmas more akin geochemically to the Koidu type kimberlites of West Africa than the Group I kimberlites from South Africa. Sr-Nd isotopic compositions either reflect derivation from a near bulk-earth mantle or represent mixing of Group I kimberlite isotopic sources and a more radiogenic Karelian Craton lithospheric mantle.

Dike 16 has characteristics of both olivine lamproite and Group II kimberlite and may be an intermediate rock type that represents the Group II kimberlite equivalent from the Karelian Craton, i.e., a lithospheric mantle-dominated partial melt from a phlogopite and K-richterite modified peridotite (e.g., Skinner, 1989; Mitchell, 1995) triggered by the influx of an asthenospheric magma (O'Brien et al., 1995).


The authors gratefully acknowledge Ashton Mining Ltd. and Malmikaivos Oy, especially Peter Gregory and Juha Rissanen, for access to samples and exploration data and the right to publish this paper. Helpful reviews by E.M.W. Skinner and A. Fung are greatly appreciated.


Clark T.C., Smith C.B., Bristow J.W., Skinner E.M.W., & Viljoen K.S. 1991. Isotopic and geochemical variation in kimberlites from the southwestern Craton margin, Prieska area, South Africa. Fifth Int. Kimb. Conf. Ext. Abs., pp. 46-47.

Clement, C.R. 1982. A comparative geological study of some major kimberlite pipes in the northern Cape and Orange Free State. Unpubl. Ph.D. thesis, Univ. Cape Town.

Griffin B.J., Rissanen J.P., Pooley G.D., Lee D.C., Macdonald I., & Kinny P.D. 1995. A new diamondiferous eclogite-bearing kimberlitic occurrence from Finland. Sixth Int. Kimb. Conf. Ext. Abs., pp. 198-200.

Kontinen A., Paavola J., & Lukkarinen H. 1992. K-Ar ages of hornblende and biotite from Late Archean rocks of eastern Finland – interpretation and discussion of tectonic implications. Geol. Surv. Finland Bull. 365, 31 pp.

Mitchell R.H. 1986. Kimberlites- Mineralogy, geochemistry, and petrology. Plenum, New York, 442 pp.

Mitchell R.H. 1995. Kimberlites, orangeites, and related rocks. Plenum, New York, 410 pp.

Mitchell R.H. & Bergman S.C. 1991. Petrology of lamproites. Plenum, New York, 447 pp.

O'Brien H.E., Irving A.J., McCallum I.S. & Thirlwall M.F. 1995. Strontium, neodymium and lead isotopic evidence for the interaction of post-subduction asthenospheric potassic mafic magmas of the Highwood Mountains, Montana, USA, with ancient Wyoming craton lithospheric mantle. Geochim. Cosmochim. Acta 59, pp. 4539-4556.

O'Brien H.E. & Tyni M. 1997. Petrology and mineralogy of kimberlites and related rocks from Finland: Preliminary results, In Papunen, H., ed., Mineral Deposits: Research and Exploration, where do they meet? Balkema, Rotterdam. pp. 781-783.

Pasteris J.D. 1983. Spinel zonation in the De Beers kimberlite, South Africa: Possible role of phlogopite. Can. Mineral. 21, pp. 41-58.

Peltonen P., Huhma H., Tyni M., & Shimizu N. 1998. Garnet peridotite xenoliths from kimberlites of Finland: Nature of the lithospheric mantle at an Archaean craton-Palaeoproterozoic mobile belt transition. This volume.

Proskuryakov V.V., Uvad'yev L.I. and Voinova O.A. 1990. Lamproity Karelo-Kol'skogo regiona. Doklady Akademii Nauk SSSR 314, pp. 940-943.

Reid A.M., Donaldson C.H., Dawson J.B., Brown R.W. & Ridley W.I. 1975. The Igwisi hills extrusive "kimberlites". Phys. Chem. Earth 9, pp. 199-218.

Rock, N.M.S. 1990. Lamprophyres. Blackie & Son, Edinburgh, 285pp.

Skinner, E.W.S. 1989. Contrasting Group I and Group II kimberlite petrology: towards a genetic model for kimberlites. In Ross, J. et al., eds., Kimberlites and related Rocks, Vol. 1. Geol. Soc. Australia Spec. Pub. 14, pp. 528-544.

Smith, C.B. 1983. Pb, Sr, and Nd isotopic evidence for sources of African Cretaceous kimberlites. Nature 304, pp. 51-54.

Smith C.B., Gurney J.J., Skinner E.M.W., Clement C.R., Ebrahim N. 1985. Geochemical character of southern African kimberlites: a new approach based on isotopic constraints. Trans. Geol. Soc. S. Afr. 88, pp. 267-280.

SUN S.-S. & MCDONOUGH W.F. 1989. Chemical and isotopic systematics of oceanic basalts: implications for mantle composition and processes, In Saunders A.D. & Norry M.J., eds., Magmatism in the Ocean Basins., Spec. Publ. Geol. Soc. London 42, pp. 313-345.

Taylor W.R., Tompkins L.A., & Haggerty S.E. 1994. Comparative geochemistry of West African kimberlites: Evidence for a micaceous kimberlite endmember of sublithospheric origin. Geochem. Cosmo. Acta 58, pp. 4017-4037.

Tyni, M. 1997. Diamond prospecting in Finland – a review, In Papunen, H., ed., Mineral Deposits: Research and Exploration, where do they meet?, Balkema, Rotterdam. pp. 789-791.