A COESITE-SANIDINE GROSPYDITE FROM THE ROBERTS VICTOR KIMBERLITE JOSEPH R. SMYTH Los Alamos Scientific Laboratory, Los Alamos, N.M. 87545 (USA),

and C.J HATTON Department of Geochemistry, University of Cape Town, Rondebosch 7700 (South Africa) Received November 10, 1976 Revised version received December 30, 1976

Earth and Planetary Science Letters 34 (1977) 284-290

284

Earth and Planetary Science Letters, 34 (1977) 284-290

A COESITE-SANIDINE GROSPYDITE FROM THE ROBERTS VICTOR KIMBERLITE JOSEPH R. SMYTH Los Alamos Scientific Laboratory, Los Alamos, N.M. 87545 (USA),

and C.J HATTON Department of Geochemistry, University of Cape Town, Rondebosch 7700 (South Africa) Received November 10, 1976 Revised version received December 30, 1976

Primary crystals of coesite and high sanidine (Or98Ab2) up to 3 mm in greatest dimension have been identified in a grospydite inclusion from tire Roberts Victor kimberlite pipe, South Africa. Microprobe analyses show that the coesite is nearly pure silica with less than 0.05 wt.% each of Al2O3, FeO and Na2O. Cell edges determined by single crystal X-ray diffraction techniques are: a = 7.143(2) Å, b = 12.383(3) Å, c = 7.143(2)Å, β = 120.00(3)º ; and the space group is C2/c. The feldspar has cell dimensions: a 8.618(3) Å, b = 13.044(3) Å, c = 7.184(3) Å, b = 116.10 (3)º ; and the space group is C2/m indicating a very high degree of Al-Si disorder. Coesite, previously considered diagnostic of impact processes, is reported for the first time from a natural static pressure environment. This simple provides evidence that sanidine can be a stable phase in the upper mantle and that free SiO2 may also exist. The structural state of the feldspar indicates equilibration above 900ºC, and the presence of coesite indicates equilibration pressures greater than 29 kbars. The preservation of the coesite indicates that the samples never spent more than a few hours above 700ºC since crossing the quartz-coesite transition.

1. Introduction Eclogites, kyanite eclogites, and grospydites from the Robert Victor Mine near Kimberley, South Africa, have been described by several authors: Kushiro and Aoki [1], Mathias et al. [2] MacGregor and Carter [3] , Sobolev et al. [4] , Lappin and Dawson [5] , Harte and Gurnev [6] , and Reid et al. [7]. These rocks are considered to have equilibrated within the upper mantle. The principal pleases are pyrope to grossular-rich garnet, sodic clinopyroxene, kyanite, and rarely corundum and diamond. Free SiO2 has not been reported from these rocks, although coesite has been identified in synthetic diamonds, and possibly natural diamonds [8,9]. Reid et al. [7] report a small amount of a phase of orthoclase composition in a diamondiferous eclogite, however the crystalline state of this material was not characterized, and it was

considered to be possibly of' secondary origin. Prinz et al. [10] report sanidine as an inclusion in diamond. Coesite was first synthesized and described by Coes [11]. The crystal structure was determined by Zoltai and Buerger [12] and recently refined by Gibbs et al. [13] and optical properties described by Sclar et al. [14]. The stability field and inversion kinetics have been investigated by MacDonald [15], Boyd and England [16], and Bell and Boyd [17]. The first natural occurrence was reported by Chao et al. [18] from Meteor-Crater, Arizona. Since then there have been many reports from impact structures, and the presence of coesite has come to be taken as characteristic impact-metamorphosed rocks. The pressure-temperature stability field of' coesite precludes its occurrence in crustal rocks [16] except at impact sites. Because of the rarity of silicasaturated compositions in the mantle and the rapid quenching required to preserve coesite, it does not commonly occur in mantle-derived rocks.

285

2. Description

The specimen of coesite-sanidine grospydite was collected by the authors at the Roberts Victor Mine in October, 1975. The roughly tabular nodule, weighing about 5 kg, was separated from the mined kimberlite before crushing. The rock has a slightly layered appearance resulting from bands rich in orange

garnet in a matrix of grey-white clinopyroxene. The average grain size is about 5 mm. A point count on one thin section cut across several of the garnet-rich bands yielded a mode of 56% clinopyroxene, 28% garnet, 9% kyanite, 6% coesite and 0.6% K-feldspar. The clinopyroxene occurs as grey-white anhedral grains up to 6 mm in greatest dimension. In thin section, the pyroxene appears to be nearly opaque because of

TABLE 1 Microprobe analyses, calculated bulk composition and CIPW norm of coesite-sanidine grospydite SRV1 ___________________________________________________________________________________ Clinopyroxene(7) Garnet(10) Kyanite(14) Coesite(4) Sanidine(8) ____________________________________________________________________________________ Percent in rock (mode) 56 SiO2 56.63 TiO2 0.07 A12O3 17.04 Cr2O3 0.04 FeO 1.56 MgO 6.24 CaO 11.57 MnO 0.00 Na2O 7.23 K 20 0.196 P205 0.00 Total 100.59

O Si AIIV AIVI Ti Fe Mg Ca

28 40.31 0.17 21.94 0.14 10.25 7.57 18.90 0.29 0.07 0.001 0.049 '99.69

9 36.82 0.03 62.63 0.06 0.15 0.05 0.00 0.00 0.00 0.00

6 99.56 0.00 0.05 0.00 0.03 0.00 0.00 0.00 0.04 0.00

0.6 64.97 0.00 18.35 0.02 0.05 0.00 0.00 0.00 0.12 16.59

99.74

99.68

100.11

6.000 2.011

12.000 3.016

5.000 0.993

0.713 0.002 0.045 0.329 0.440

1.935 0.008 0.641 0.845 1.516

1.990 0.001 0.004 0.002 0.000

4.000 1.997 0.001

8.000 3.001 0.999

0.000 0.001 0.000 0.000

0.000 0.002 0.000 0.000

52.69 0.09 21.43 0.04 3.76 5.62 11.77 0.08 4.08 0.24 0.01 99.56 CIPWnorm(wt.) Or Ab An Ne Wo Di En Fs

Ol Fo Mn 0.000 0.017 0.000 0.000 0.000 Fa Cr 0.002 0.008 0.000 0.000 0.000 11 Na 0.447 0.013 0.000 0.001 0.011 K 0.011 0.000 0.000 0.000 0.978 ____________________________________________________________________________________ Figure in parentheses is number of spot analyses included in average.

1.42 32.57 39.45 1.06 7.91 4.97 2.45

3.44 0.17

All elements except P were analysed on an ARL-EMX electron microprobe at the Max-Planck-Institut fair Kernphysik at Heidelberg. Matrix corrections were performed in the automation program GEOLAB using correction factors of Albee and Ray [26]. Counting period for each element was 30 seconds on 40,000 counts and an excitation potential of 15 kV with a specimen current of approximately 0.015 µA was used. Standards used were: Si, Ca, Mg-diopside; Na-clinopyroxene Jd35Di65 or pyrope.- K-andularia; Ti-pseudobrookite; pyroxmangite; Fe-synthetic fayalite. P, K and Na analyses of garnet and clinopyroxene were c@ecked using a Cambridge electron microprobe at the University of Cape Town.

286

extensive alteration; however, optically transparent areas up to 0.05 mm persist and are evenly distributed through the grains. The average chemical analysis (Table 1) is similar to that of omphacites from grospydites reported by Sobolev et al. [4]. The com. position is quite varied, and microprobe analysis showed the more altered areas were strongly enriched in CaTschermaks component. All clinopyroxene analyses show an amount of AlVI in excess of Na + K + AlIV indicating a significant apparent charge imbalance or cation vacancy in the pyroxene ' X-ray and electron diffraction studies currently in progress may help to explain the alteration reactions in this pyroxene. The garnet occurs as orange subhedral grains up to 5 mm in greatest dimension. The garnet shows very little alteration, however, a few grains contain less than 1% of apparently secondary amphibole alteration along fractures.

The average garnet reported in Table 1 is quite homogeneous within the section studied; both within individual grains and from grain to grain. The grossular-rich composition (Gr50Py28Al22) together with the pyroxene and kyanite qualifies the specimen to be termed a grospydite (grossular-pyroxene-disthene). The small amounts of Na and P are consistent with its high pressure of origin [6,19]. Kyanite occurs as light blue euhedral grains up to 3 mm across and shows a slight preference for the garnet-rich bands. Chemical analysis (Table 1) indicates Less than 0.2% of components other than Al2O3 and SiO2. Accessory sulfide grains up to 0.3 mm in greatest dimension appear to be polycrystalline intergrowths of Fe-Ni sulfides similar to those described by Desborough and Czemanske [20]. Coesite occurs as colourless subhedral grains up to 3 mm greatest dimension which are

287

Fig. 1. Crossed nicols (A) and plane-polarized light (B) photographs of coesite grain surrounded by a rim of polycrystalline quartz. Grain is about 2 mm across.

invariably surrounded by inversion rims of polycrystalline quartz. Fig. 1 is a photo-micrograph in crossed nicols and plane-polarized light of one such grain. In Fig. I B, the higher index of refraction of the coesite is clearly visible. The crossed nicols micrograph (Fig. 1A) shows the quartz rim to be finely polycrystalline and randomly oriented with respect to the coesite. Several, such coesite grains were separated from the crushed mineral separates of the rock and studied by single crystal X-ray techniques. Precession photography confirmed the identification as coesite and gave the space group as C2/c [12]. Cell edges were determined, using a Phillips automated fourcircle diffractometer, by least squares refinement from automatic centering of 20 strong diffraction maxima, as: a = 7.143(2) Å, b = 12.383(3) Å, c = 7.143(2) Å; β = 120.00(3)º. The specimen studied showed apparent cleavages on (010) and (012) being elongate

parallel to a*. A few grains in the thin section show fine lamellae, however, none Of the specimens X-rayed was twinned. K-feldspar occurs as untwinned subhedral grains up to 3 mm across, generally in close proximity to other feldspar grains and kyanite. Two such grains are illustrated in Fig. 2. The chemical composition in Table 1 is an average of eight analyses of five grains in the polishedthin section studied; there is little variability from Or98Ab2. Although there is considerable Ca in both the clinopyroxene and garnet, the feldspar appears to be completely Ca-free. A portion of one of the analysed grains was removed from a thin section for single crystal X-ray study. Precession photographs indicate space group C2/m and cell edges determined on the single crystal diffractometer were: a = 8.618(3) A, b = 13.044(3) Å, c = 7.184(3)Å; α = 90.00(2)', β = 116.10(2)º, γ = 90.00(2)º consistent with sanidine in a very high

288

Fig. 2. Crossed nicols (A) and plane-polarized light (B) photographs showing garnet (G), coesite (C), kyanite (K), and sanidine (S) grains in a matrix of altered clinopyroxene. The area of the section photographed is about 8mm across.

state of Al-Si disorder. The b-axis is slightly larger than previously reported for natural or synthetic high sanidines. The high structural

state indicates equilibration above about 900ºC [21,22].

289

3. Discussion The existence of coesite and high sanidine as primary phases in this specimen permits estimates of minimum pressures and temperatures of equilibration of the rock. The quartz-coesite transition determined experimentally by Boyd and England [16] passes through 700ºC at 27.3 kbars and 1400ºC at 35.2 kbars (P = 19.5+ 0.0112T). The rock must therefore have equilibrated at depths in excess of 100 km in the upper mantle. The cell edges of this sanidine indicate a slightly larger cell (greater degree of Al-Si disorder) than those synthesized by Orville [22] at 700-800ºC which are presumed to have some small amount of AlSi ordering [21]. As this high-pressure sanidine apparently shows a greater degree of Al-Si disorder than sanidines synthesized at 700800ºC, and since increased pressure may be assumed to favour the slightly denser, more ordered states, this specimen cannot have equilibrated below approximately 900ºC. The structural state is certainly consistent with higher temperatures, and it is known that Kfeldspar is stable to at least 40 kbars and 14000C [23]. Recently, Hazen [24] has shown the monoclinic-triclinic inversion at 900ºC at 30 kbars, and in order to have preserved the high state of Al-Si disorder, this sanidine must have equilibrated on the monoclinic (hightemperature) side of this inversion. It therefore may be concluded that the rock equilibrated at temperatures greater than 900ºC and pressures greater than 29.6 kbars, consistent with previous studies on eclogites from the Roberts Victor pipe. Råheim and Green [25] present pressuretemperature equilibration curves for the KD of Fe-Mg distribution between coexisting garnet and clinopyroxene in rocks of basaltic composition. Although the current specimen is well outside the composition range studied by these authors, it is of interest to note that the KD for this rock is 6.04 corresponding to a temperature of 825ºC at 30 kbars. However, the effects of the higher Ca in the garnet and the apparent vacancy or charge imbalance in the pyroxene cannot currently be estimated.

The preservation of coesite in this rock allows a few constraints to be placed on the cooling history. Experimental studies [16,17] indicate that coesite breaks down rapidly when held outside its field of stability. In one experiment [16], coesite held at 705ºC and 26.1 kbars had 90% inverted to quartz in less than 7 hours in the presence of water. If water may be assumed to be present from the kimberlite during eruption, it seems likely that this specimen cooled below 7000C within a few hours of crossing the quartz-coesite boundary. However, it is known that shear stresses strongly affect the reaction kinetics of the quartz-coesite transition, and textures indicating strong shear effects are absent in this specimen. A few simple experiments utilizing the large coesite crystals from this rock would be useful to further constrain the cooling history. Reid et a]. [71 report small blebs (-20 /im) of a phase of K-feldspar composition associated with pyroxene in a diamondiferous eclogite; however, X-ray or optical characterization of this material was not readily possible. The current specimen contains large grains of high sanidine, which are apparently primary. No trace of a hydrous phase was observed in this specimen, so it is possible that sanidine is an important K-bearing phase in the upper mantle when insufficient water is available for the formation of phlogopite or amphibole. 4.

Conclusions

The petrology and mineralogy of a unique coesite-sanidine grospydite are described. The rock appears to have equilibrated at temperatures above 900"C and pressures in excess of 29 kbars, in the upper mantle. The preservation of coesite indicates rapid cooling to below 700ºC during eruption. The specimen provides evidence that free silica exists in the upper mantle and that sanidine may be an important K-bearing phase in the upper mantle, particularly when insufficient H20 is available for the formation of hydrous impact processes, the large grain size and primary habit of coesite from static pressure environments should easily distinguish it from impact-generated coesite.

290

Further studies on this specimen including Fe2+ Fe3+ determination and X-ray and electron diffraction study of the pyroxene breakdown are in progress. Acknowledgements The senior author is indebted to the Deutsche Forschungs Gemeinschaft for a temporary research position at the University of Marburg and for electron microprobe time at the MaxPlanck-Institut für Kernphysik in Heidelberg. Manuscript preparation was supported by the University of Cape Town and the Los Alamos Scientific Laboratory. The authors thank Prof. A.M. Reid and Dr. J.J. Gurney of the University of Cape Town, Prof. A, El Goresy of MaxPlanck-Institut für Kernphysik, Heidelberg and Prof. F.R. Boyd of the Carnegie Institution of Washington, Geophysical Laboratory for critical reviews of the manuscript and fruitful discussions. References 1 I. Kushiro and K. Aoki, Origin of some eclogite inclusions in kimberlite, Am. Mineral. 53 (1968) 1347-1367. 2. M. Mathias, J.C. Siebert and P.C. Rickwood, Some aspects of the mineralogy and petrology of ultramafic xenoliths in kimberlite, Contrib. Mineral. Perrot. 26 (1970) 75-123. 3 I.D. MacGregor and J.L. Carter, The chemistry of clinopyroxenes and garnets of eclogite and peridotite xenoliths from the Roberts Victor Mine, South Africa, Phys. Earth Planet. Inter. 3 (1970) 391 -397. 4. N.V. Sobolev, Jr., I.K. Kuznetsova and N.I. Zyuzin, The petrology of grospydite xenoliths front the Zagodochnaya Kimberlite Pipe in Yakutia, J. Petrol. 9 (1968) 25 3 -280. 5 M.A. Lappin and J.B. Dawson, Two Roberts Victor ctimulate eclogites and their re-equilibration, Phys. Chem. Earth 9 (1975) 351-365. 6 B. Harte and J.J. Gurney, Evolution of clinopyroxene and garnet in an eclogite nodule from the Roberts Victor Kiinberlite Pipe, South Africa, Phys. Chem. Earth 9 (1975) 367-387. 7 A.M. Reid, R.W. Brown, J.B. Dawson, G.G. Whitfield and J.C. Siebert, Garnet and pyroxene compositions in some diamondiferous eclogites, Contrib. Mineral. Petrol. (in press). 8 H.O.A. Meyer and F.R. Boyd, Composition and origin of crystalline inclusions in diamopds, Geochim. Cosmochim. Acta 36 (1972) 1255-1273. 9 H.J. Milledge, Coesite as an inclusion in G.E.C. synthetic diamonds, Nature 190 (1961) 1181.

10

11 12 13

14

15

16 17

18

19

20

21 22

23

24

25

26

M. Prinz, D.V. Manson, P.F. Hlava and K. Keil, Inclusions in diamonds: garnet Iherzolite and eclogite assemblages, Phys. Chem. Earth 9 (1975) 797-815, L. Coes, Jr., A new dense crystalline silica, Science 118 (195 3) 131-152. T. Zoltai and M.J. Buerger, The crystal structure of coesite, Z. Kristallogr. 111 (1959) 129-141. (;.V. Gibbs, C.T. Prewitt and K.J. Baldwin, A study of the structural chemistry of coesite, Z. Kristallogr. (in press). C.13. Sclar, L.C. Carrison and C.M. Schwartz, Optical crystallography of coesite, Am. Mineral. 47 (1962) 1292-1302. G.J.F. MacDonald, Quartz-coesit.e stability relations at high temperatures and pressures, Am. J. Sci. 254 (1956) 722-735, F.R. Boyd and J.L. England, The quartz-coesite transition, J. Geophys. Res. 65 (1960) 749-756. P.M. Bell and F.R. Boyd, Phase equilibrium data bearing on the pressure and temperature of shock metamorphisin, in: Shock Metamorphism of Natural Minerals, B. French and N.M. Short, eds. (Mono Book Co., Baltimore, Md., 1968)43-50. E.C.T. Chao, E.M. Shoemaker and B.M. Madsen, First natural occurrence of coesite, Science 132 (1960) 220-222. R.N. Thompson, Is upper mantle phosphorous contained in sodic garnet?, Earth Planet. Sci. Lett. 26 (1975) 417-424. G.A. Desborough and G.K. Czemanske, Sulfides in celogite nodules from a kimberlite pipe, South Africa, with comments on violarite stoichiometry, Am. Mineral. 58 (197 3) 195-202. J.V. Smith, The Feldspars, Vol. I (Springer Berlin, 1974) 255-276. P.M. Orville, Unit-cell parameters of the microcline4ow albite and the sanidine-high albite solid solution series, Am. Mineral. 52 (1967) 55-86. D.H. Lindsley, Melting relations of KAISi3O8: effect of pressure up to 40 kilobars, Am. Mineral. 51 (1966) 17931799. R.M. Hazen, Sanidine: predicted and observed monoclinicto-triclinic reversible transformations at high pressure, Science 194 (1976) 105-107. A. Råheim and D,H. Green, Experimental determination of the temperature and pressure dependence of the Fe-Mg partition coefficient for coexisting garnet and clinopyroxene, Contrib. Mineral. Petrol, 48 (1974) 179-203. A.L. Albee and L. Ray, Correction factors for electron probe microanalysis of silicates, oxides, carbonates, phosphates, and sulfates, Anal. Chem. 42 (1970) 1408-1414.