GEMOC ARC National Key Centre

Jackson, S.E¹, Davies, R.M.¹, Griffin, W.L.^1,2, O'Reilly, S.Y¹., Doyle, B.³

¹ GEMOC Macquarie

²CSIRO Exploration and Mining

³Kennecott Canada Inc.

Natural diamonds erupted by kimberlites and lamproites are unmodified samples of the deepest upper mantle. Trace element analysis of their impurities provides information on the composition of their source environment. Questions to be addressed through the trace element study of diamonds are: (1) Do diamonds of different parageneses crystallise from different fluids? (2) What are the origins of these fluids? (3) Are diamond-forming processes similar worldwide or depend on mantle age and composition? (4) Can we "fingerprint" diamonds from a particular locality?

A laser ablation microprobe (LAM)-ICP-MS technique, developed for trace element analysis of carbonaceous substances [2], has been used to analyse trace elements in fibrous and non-fibrous diamonds. The custom-built LAM employed in this work uses a Continuum Surelite I-20, Q-switched Nd:YAG laser frequency quadrupled to 266 nm wavelength and operated at a frequency of 10 Hz. The ICPMS is a Perkin-Elmer Sciex ELAN 6000. Calibration was performed by direct ablation of a certified oil reference material (S-21 (50 ppm); Conostan) contained in a capillary tube. A bitumen standard reference material (SRM1632b; NIST) was used to extend list of the elements calibrated and crosscheck the data. Carbon was used as an internal standard to correct for differing ablation yields. Laser pulse energy used was 0.15 mJ/pulse for the oil and -bitumen standards, and 2.0-2.2 mJ/pulse for diamond. For each sample, optimal background (ca. 60 replicates counted on the Ar carrier gas) and ablation (ca. 120 replicates) regions were selected from the time-resolved spectra and integrated to determine the net count rates and concentrations for each element [2]. 2-4 analyses were averaged for each diamond. Relative deviations for replicate analyses of fibrous diamond samples ranged between 2% and 81% for different elements (median 20%), indicating significant heterogeneity from spot to spot.

To test the accuracy of the results, we have analysed fibrous diamonds from the Jwaneng kimberlite (Jwn 110 and Jwn 115) previously analysed by NAA [5], EMP [4] and PIXE [6]. Good agreement was found between the methods, considering sample heterogeneity and the different volumes sampled by the different analytical methods (Table 1). The results demonstrate that the LAM-ICPMS can rapidly and accurately measure a wide suite of trace elements with high spatial resolution. Several improvements are being made to the technique. They include development of a frequency quintupled laser wavelength option (k = 213nm) to improve ablation efficiency of the diamonds, and development of a synthetic aqueous standard(s) to extend the list of elements that can be quantified (halogens, S, etc).

Table 1. Comparative data on trace element abundances in diamond Jwn110 from LAM-ICPMS, PIXE and EMP+NAA analyses. Values are in ppm.
 
 

To compare diamonds of different paragenesis, we are analysing non-fibrous diamonds from the DO27 kimberlite, Slave Craton, whose parageneses (peridotitic, eclogitic and superdeep (ferropericlase-bearing)) were defined from mineral inclusion studies [1]. Preliminary results show that diamonds from the three parageneses are significantly different. Relative to peridotitic diamonds, Na, Mg, Al, K, Rb, Sr, Y, Sn and Ce are high in eclogitic stones and low in superdeep ones. The differences suggest that meaningful trace-element "fingerprinting" to identify a diamond as being from a single locality is unlikely to be successful.

References
[1] Davies, R.M. et al. (in press) Proc. 7^th Internat. Kimberlite Conf..
[2] Longerich H.P. et al. JAAS 11, 899-904.
[3] Mossman D.J. et al. (submitted) Econ. Geol.
[4] Schrauder, M. and Navon, O. (1994) Geochim. Cosmochim. Acta, 58, 761-771.
[5] Schrauder, M. et al. (1996) Geochim. Cosmochim. Acta, 60, 4711-4724.
[6] Griffin, W.L. and Ryan, C.G. unpubl data.