- Email: [email protected]
Mineral inclusions in Brazilian diamonds
50 MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS HENRY O. A. MEYER and DARCY P. SVISERO* Department of Geosciences. Purdue UniversIty. West Lafayette, Indiana 47907. U.S.A.
ABSTRACT Mmerals mcluded m diamonds from several localIties m Brazil have been exammed. The mmerals observed are olIvme, enstatIte, chrome-pyrope and pyrope-almandme. These minerals have simIlar compositions to those previously reported for mcluslOns from other worldwide localItIes. In addItIOn rutIle, Ilmemte, Zircon, quartz and pyrrhotite occur as mcluslOns. This IS the first record of Ilmenite and zircon m natural diamond The ilmenite IS umque compared to most kImberlIte Ilmemtes m that It contains no appreciable content of magnesIUm. The presence of quartz as an mcluslOn IS unusual.
INTRODUCTION Natural diamonds are generally considered to be the products of stable growth (e.g. KENNEDY and NORDLIE, 1968). On this assumption it has been argued (MEYER and BOYD, 1972) that many crystallme inclusions in diamond represent original phases which formed in equilibrium with diamond. Detailed chemical examination of these minerals should provide valuable data pertinent to the genesis of diamond, to kimberlite, and to the associated xenoliths. Recent studies of mineral inclusions in diamonds from South Africa, West Africa, Venezuela and Thailand (MEYER, 1971; MEYER and BOYD, 1969, 1972) and from Siberia and the Urals (SOBOLEV et al., 1969a, 1970, 1971, 1972, 1973) have shown a remarkable similarity in the compositions of the various minerals. Furthermore, the inclusions (olivine, enstatite, dlOpside, garnet, chromite) although generally similar to their counterparts in kimberlite, and associate xenoliths, have elemental abundances that are distinct in detail. In an attempt to extend the present geographic coverage of diamonds and their inclusions we have examined specimens from Brazil. In general, these new results are in agreement with earlier observations. In addition we have analyzed rutile, ilmenite, pyrrhotite, zircon and quartz, all of which occur as inclusions.
GEOLOGIC OCCURRENCE Although diamonds have been found in Brazil since about 1725 no kimberlite has yet been recorded. Mostly the diamonds are found in river placer deposits throughout several * Permanent address' Instltuto de Geosclenclas, Umversldade de Sao Paulo, Sao Paulo, BraZIl. 785
786
H. O. A. MEYER AND D. P. SVISERO
areas, mainly in the states of Minas Gerais, Mato Grosso, Goias, Bahia, Parana and Para. In east-central Brazil (Diamantina in Minas Gerais and Chapada Diamantina in Bahia) the diamonds are associated with metasediments, conglomerates, arenites, phyllites and quartzite, of upper Precambrian age (Fig. 1).
Morro do
GUlmaraes
....
Rondonopol1s
,
.
Alto,Araguala A
Bolivia
L.
".>,
"'Itulutaba
<
... frlltalo::..~
IDEV
Paraguay
..
Dlamantlna
COlli"'.
PARANA
Ch~peu
len~ols
IPRE~CAMB I
Datn OCEAN
1
BASIN (RET!
..
Jaguariaiva llbilJI.
-
""
,-
Argentina
.D,amond Local,t,es
FIG I. LocalItIes in BrazIl from whIch diamonds were obtamed
In
thIs study
In contrast, the other major region of diamond production is associated with the large sedimentary Parana basin in south central Brazil (Fig. I). Devonian and Carboniferous arenaceous sediments form the rim of the basin and are unconformable on Precambrian basement rocks. The diamond placer deposits are predominantly found in the rivers draining the Devonian strata, but placers do occur in rivers on the Carboniferous rocks. It is generally thought that the Devonian and Carboniferous sediments are the secondary matrix for the diamonds which were previously derived from Precambrian rocks. Currently, several alkalic intrusions of Cretaceous age have been found bordering the Parana basin. This has led some to suggest the possibility of Cretaceous kimberlites as a source of diamonds.
SAMPLE DESCRIPTION AND TECHNIQUES The specimens exammed in this study are from diamonds obtained in both of the above regions. All the diamonds were collected from placer deposits in the most productive areas (Fig. 1). Morphologically, the diamonds are complex but most are characterized by the predominance of dodecahedral forms and a large number of contact twins. Infra-red studies
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
787
showed that about 95~~ of the diamonds examined are of Type I. The specific gravity of the diamonds varied between 3.500 and 3.530 g/cm 3 , and trace elements detected by spectrographic emission analysis include AI, Ca, Mg, Si, Fe, Cu and Cr (SVISERO, 1971). The majority of the diamonds studied were colorless and were between 2 and 10 mm in size. They were cleaned in HCl and HF for several hours and examined for lack of fractures reaching from the surface. The inclusions were all released by cracking the diamond host in a small enclosed container. Following optical examination the inclusions were mounted for scanning electron microscope observations, X-ray diffraction and electron-microprobe analysis. The analyses were obtained using a M.A.C. 500 electron microprobe. Standards were predominantly glasses or pure metals. The analytical techniques and data reduction procedures used are similar to those described by BoYD (1969). Some initial characterization of the inclusions was done using a qualitative energy dispersive system on the scanning electron microscope.
MINERALOGY The majority of minerals found as inclusions in the Brazilian diamonds are very similar in composition to those previously recorded (MEYER, 1968; MEYER and BOYD, 1972; SOBOLEV et al., 1973). These inclusions are, namely, olivine, enstatite, chrome-pyrope, almandine-pyrope and clinopyroxene. In addition we have found zircon, ilmenite, rutile, pyrrhotite and quartz. To avoid repetition in the following sections reference to "previous results" will refer specifically to MEYER and BOYD (1972) and SOBOLEV et al. (1969a, 1970, 1971. 1972, 1973).
OLIVINE Olivine was first identified as an inclusion in Brazilian diamonds by CAMARGO and LEITE (1968) using X-ray techniques. It is the most common inclusion observed. In composition
the olivines are similar to previously reported analyses (Table 1). They are characterized by a small range in composition (F092 to F0 9S ), relatively low contents of the minor elements AI, Ca, Mn and Ni. and high values (up to 0.10 wt.°/~) ofCr 2 0 J . These Cr 2 0 J values are an order of magnitude greater than for other terrestrial forsteritic olivines.
ENSTATITE Enstatite appears to be more common in Brazilian diamonds than in others previously examined. The present analyses, however, agree closely with those of enstatite i!1clusions from other localities. The enstatite molecule (MgSiO J ) accounts for 93-94~;; of the total mineraL there being very little solid solution towards diopside. Both Cr 2 0 J and Al 2 0 3 contents are low (Table 2). Enstatite coexists with both olivine and chrome-pyrope.
I> \( () rl
\,\1 \)
HB
788
1
H. O. A. MEYER AND D. P. SVISERO TABLE
1.
REPRESENTATIVE ANALYSES OF OLIVINE INCLUSIONS IN DIAMOND
BA35A
Si0 2 Ti0 2 Al 2 0 3 Cr 2 0 3 FeO MgO CaO MnO NIO
41.3
MG26A
-PR38A--
-----
--------
40.7
40.9 0.02 0.02 0.10 7.17 52.4 0.04 0.10 0.28
< 0.01 < 0.01
< 0.01 0.02
0.05 7.98 50.1 0.03 0.11 040
0.10 7.84 504 0.03 0.10 028
100.0
99.5
MT15B*
MTI2Bt
+----~
41.6
42.1
< 0.01
< 0.01
0.04 0.10 7.06 49.0 0.07 0.11
0.04 010 542 52.2 0.03 0.10
2_]__~_0.4~~
03
101.0
98.4 "~--
100.4 ~--
No of cations ( x 1000) for 4 oxygen
l
r-----~~
SI Al Cr Fe+ 2 Mg Ca Mn Ni °0 Fa
1006
996 I
I 162 1814 1 2
160 1836 I
8
983 1 2 144 1877
1022 I 2 145 1794
1 2
2
6
91.8
2 108 1860 I
2 7
5
92.0~ 92.9 _
1008 I
2 8 94.5
92_.5_
L
* CoexIsts wIth chrome-pyrope (MTI5A). t CoexIsts with enstatite (MTI2A)
TABLE
SI0 2 TI0 2
AI 2 0 3 Cr 2 0 3 FeO MgO CaO MnO NiO
2.
REPRESENTATIVE ANALYSES OF ENSTATITE INCLUSIONS IN DIAMOND
MG27A
G055A
57.3 < 0.01 051 0.34 4.75 36.2 0.18
57.5 < 0.01 060 0.47 4.68 37.8 031
o.IOj 0.39 -
-
-
--B~:~B - ---=~G2IB*
0.10 0.11
57.6 < 0.01 0.87
57.6 < 0.01 078
062 4.27 36.6 0.14 0.13 Ql1
0.47 436 36.2 047 0.11
1- MTI2At~ 57.6
< 0.01
I
051 0.11 4.95 35.3 0.54 034 Q02
--~---
99.8
100.6
--~----
100_.3
---
1000_
J __ 99.4 _ _
No. of catIOns I x 1000) for 6 oxygen
SI Al Cr Fe Mg Ca Mn NI ° En 0
1972 21
9
[37 1855 7 3 [I 93.1
1960--TI9M--- 24 13 133 1874
1~
35 17 122 1861
1_~
---=~
93~__J
1971 31 13 125 1846 [7 3
1987 21 3 143 1818 20 10
~~ _~_9~
* CoexIsts wIth chrome-pyrope (MG21A) t CoexIsts wIth olIVIne (MTI2B).
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
789
CLINOPYROXENE In keeping with earlier observations, clinopyroxenes are comparatively rare as inclusions
in diamond. In this study two inclusions identified by energy dispersive technique as possible diopside (high Ca, Mg, and Si) were found. Unfortunately, both inclusions were very small, < 30l1m, and were lost during polishing.
GARNET Garnet appears to be the second most common inclusion in Brazilian diamonds. Two distinctive suites occur: one, Cr-rich, Ca-poor pyrope; and two, Fe-rich, Cr-poor almandine-pyrope garnet (Table 3). The latter group is minor. SOBOLEV et al. (1973) discuss in
TABLE 3. REPRESENTATIVE ANALYSES OF GARNET INCLUSIONS IN DIAMOND
P2~-' 1
'.IT;;Al
-_._-
MG2IA*
Si0 2 TIO z AlzO J CrzO, FeO MgO CaO MnO
< 0.01 10.4 17.3 5.84 24.6 0.93
41.2 < 0.01 14.8 10.9 6.30 23.1 243 0.35
99.5
99.1
--_._---- - -
BA43A
G08A
MG25A
41.1 < 0.01 15.6 10.3 6.23 23.5 2.48 0.33
41.5 < 0.01 16.7 9.42 6.07 23.4 230 0.33
99.5
997
409 < 0.01 16.2 9.34 6.51 21.5 5.09 0.25
- - -----_.-
' - -
Si Ti Ai Cr Fe Mg Ca Mn
2984 l}O9 1014 361 2711 74
-I'
99.8
---~-
-
MT22A
G057A
1---- .
.--
40.5 0.86 19.6 0.20 16.0 13.1 8.39 0.34
39.8 1.05 20.3 011 16.4 11.5 8.46 0.56
39.9 0.85 20.3 0.09 17.0 12.7 8.32 0.50
99.0
98.2
99.7
-
-----
No. of catIOns ( x 1000) for 12 oxygen 3012 1275 630 386 2519 191 21
T
I
2989 1334 590 379 2544 193 20
- -I 2990
1420 537 366 2518 178 20
::77 1394 2977
r-"30~~
1739 538 12 397 1007 ~38 1470 398 677 ~_-----=-_
-
---
3029 60 1821 7 1046 1310 690 36
3001 48 1800 6 1069 1421 670 32
* Coexists With enstatite (MG2IB).
t Coexists With olivine (MT15B).
detail the parageneses of Cr-rich garnets from kimberlites in Yatkutia. They conclude that the Cr-pyropes which occur as inclusions in diamond can be differentiated from other Cr-rich garnets on the basis of their low content ofCaO (see also SOBOLEV et ai., 1969a, 1971). The analyses presented here of garnets from Brazilian diamonds are similar to those previously examined from other world wide localities (Fig. 2). Recently, chrome-pyrope garnets, similar in composition to those occurring as inclusions in diamond, have been found to be present in concentrates in the Finsch Pipe, South Africa (GURNEY and SWITZER, 1973).
790
H. O. A. MEYER AND D. P. SVISERO
' 'f I..
j
1--GARNETS
I I I
80r I I I 0 0
c
Inclusions
In
Diamond
80
Peridotite Xenoliths
and KimberlIte
In
Kimberlite
,:,
•
~I~:
'.'. iI
I. c"2JI:
l'J
~
~
1~l'J
••
60
0
o
o
..
® Eclogite Xenoliths
60
x
100
40
'.'
i
§g.,6
.
~ ...
J
•
0
6t
....
I
20
~
,
Q)6!J0
40
60
·fl
0
I
80
, ~
I
100
0
0
r
0
0
0
0
20 0
.....
~
40
"
J
.0 0
•
0
eli·
20
0
°
0
C
\5:, ~' 40
20
MgO X100 [MgO + FeOl
CoO X100 [CoO +MgO J
FI
xenolIth,. from kimberlIte and from eclogite xenolIths In kimberlIte BrazilIan mcluslOns-solId trIangle,; SiberIan Inclusions· solId squares, other mcluslOns (M~.YER and BOYD, 1972)- ,olId Circles (data as for MEYER and BOYD, 1970 plus SOBOL~.V I!f al.. 1969a, 1973 and GURNEY and SWITZER, 1973)
RUTILE Rutile has been observed as an inclusion in diamond by HARRIS (1968) and GURNEY et al. (1969) using X-ray diffraction methods. We have found and analyzed rutile in Brazilian diamonds (Table 4). The rutile inclusions were generally elongated parallel to c, euhedraL reddish-brown in color and moderately transparent. From single crystal diffraction data TABLE 4 Zircon MT30A
Sl02 TI0 2 AI 2
°-'
Cr 2 0
FcO MgO
CaO MnO Zr0 2
J
31.1 0.03
om
-
-
-l
RFPR~.sI,'HATlVF
R~~e
GO~IA +~TI6~ 006 99 X 003 0.\6
ANALYSES OF INCLUSIONS IN DIAMOND
- MTIOA
BAl9A
G049A
il.11
0.18 50.5 022 003
< 0.01
0.73
012 519 0.21 < 0.01 47.6 0.14 0.03 064
011> 514 03(, 004 482 014 < 001 0.6X
I
00') 99 6
I
Ion
I 'I
~:~~ ~:~~~I 4~~~
001 0.02 69.7
< < 0.01
1009
100.5
I
Pyrrhotite
Ilmenite
<
Fe S NI Cr Cu
100 I
1010
MT36B
(,2.4 37.8 0.48 0.02 001
61.2 38.2
100.6 Fe.
}t00-4-._"- 992
MT36A
0.4
1.26 0.03 0.01
100.7 091
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
791
the cell dimensions were determined to be a = 4.59 A and c = 2.95 A. Analyses show the rutile to be extremely pure with less than 0.4 wt.~,;, of other oxides. These results are in agreement with those of SOBOLEV et al. (1972b) who examined rutile from eclogitic xenoliths in kimberlite.
ILMENITE Ilmenite was extracted from three diamonds which all satisfied the basic criteria by which we judge inclusions to be primary (i.e. inclUSIOn totally within diamond and no visible fracture from inclusion to diamond surface). The ilmenites appear to be close to stoichometry (Table 4), manganese being the major impurity «0.7 wt.~o MnO). This composition of FeTi0 3 contrasts markedly with that of ilmenites that occur as xenocrysts and in xenoliths in kimberlIte. These latter ilmenites all have appreciable contents of MgO (e.g. MCCALLISTER, MEYER and BROOKINS, this vol.; MITCHELL, 1973) which may reach up to 20 wt. 0;,. Ilmenite has been recorded previously by HARRIS (1968) as a syngenetic inclusion. Harris observed the ilmenite to be powdered and to be commonly exposed in holes on diamond surfaces.
ZIRCON Zircon is not uncommon as a mineral in kimberlite (KHARKIV et aI., 1972) but this is believed to be the first substantiated record of its occurrence as an inclusion in diamond. The zircon appears very pale brown in color and is slightly elongated and its analysis (Table 4) indicates that the minor elements content is low.
QUARTZ Quartz and coesite have at various times been considered as inclusions in natural diamonds (MILLEDGE, 1961; HARRIS, 1968; ORLOV, 1959). In this study we have obtained, apparently as a primary inclusion in terms of identification criteria, an inclusion of quartz. The specimen was colorless and had well-developed faces. The identification was substantiated by both single crystal X-ray diffraction methods and by electron-microprobe analysis. Some caution in accepting this as a primary inclusion is perhaps warranted since quartz is known to occur as an intergrowth in Brazilian diamonds (CORRENS, 1931; BOYD and MEYER, unpubl.).
SULFIDES Examples of sulfides llCcurring as inclusions in diamonds are well known and SHARP (1966) has described troilite and pentlandite as relatively common (among opaque inclusions). HARRIS (1968) also mentions the identification, using X-ray techniques, of pyrrhotite and pentlandite. In Brazilian diamonds inclusions are opaque, small and irregular. Analyses confirmed the sulfide as pyrrhotite (Table 4).
792
H. O. A. MEYER AND D. P. SVISERO
DISCUSSION The results presented above agree with earlier studies of mineral inclusions in diamond. For example, many of the silicates are remarkably uniform in composition irrespective of provenance and age. Olivine and chrome-pyrope are the most abundant inclusions. Furthermore, several inclusions are essentially similar in composition to their counterparts in ultramafic rocks whereas others are similar to the constituent minerals of eclogite. It is usual to find only those minerals belongmg to the same suite coexisting in a single diamond. For example, in this study we have observed the following pairs of inclusions: olivine + chrome-pyrope, olivine + enstatite, enstatite + chrome-pyrope. This last pair has not previously been recorded. The above minerals could all be considered as members of an ultramafic suite. In contrast MEYER and BoYD (1972) have described clinopyroxene + almandine garnet from a single diamond, both minerals being in some respects similar to minerals from eclogite. Also, SOBOLEV et al. (1970) have discussed the assemblage olivine + garnet + chrome-diopside in one diamond. The garnet in this assemblage is unique as it contains approximately 8 wt./~ Cr 2 0 3 but unlike the normal chrome-pyrope inclusion it has an extremely high content of CaO (13 wt. ~<,). Sobolev points out that the garnet is close in composition to some rare pyrope-uvarovites from kimberlite (SOBOLEV et aI., 1968). The rarity of clinopyroxene inclusions is surprising in view of the abundance of this mineral in kimberlite and associated xenoliths. In the Brazilian diamonds the overall assemblage olivine + garnet + enstatite is present. MACGREGOR (1974) has examined the solubility of AI 2 0 3 in enstatite for garnet and spinel peridotite composition. Using his data plus the AI 2 0 3 contents of the enstatite inclusions and assuming the assemblage consists of garnet + orthopyroxene + clinopyroxene, a minimum of 40 kb at lO00°C is obtained for possible equilibration temperatures. Their values are approximate since no account has been taken of the other trivalent elements (e.g. Cr and Fe + 3). According to MACGREGOR, consideration of all the trivalent elements would generally increase the above values of P and T to the region of approximately 130n C and 50kb. BoYD and NIXON (1973) have plotted Ca/(Ca + Mg) ratios of enstatites against temperature of equilibration for coexisting diopsides. Using their data, it appears enstatite inclusions should have formed at less than 800'C. This is in sharp contrast to the values obtained from MACGREGOR'S (1974) data. It is thus very likely that the assemblage enstatite + olivine + chrome-pyrope did not coexist with diopside in diamond. This conclusion is in agreement with that of SOBOLEV et al. (1973) and SOBOLEV et al. (l969a, b). The presence of rutile as an inclusion in diamond has been noted previously (HARRIS, 1968; GURNEY et al., 1969). It is noteworthy that the diamond which contained the rutile inclusions observed by GURNEY et al. was found within an eclogite xenolith. The listing of rutile inclusions in the suite of minerals with eclogite affinities is thus fairly well established. The occurrence of pure ilmenite as an inclusion is puzzling, especially when one considers that all kimberlitic ilmenites have relatively high contents of MgO. It is unlikely that these ilmenites were adhering to cavities within the diamond surface because of the cleaning procedures used prior to careful optical examination of the diamonds. MITCHELL (1973) comments that magnesian ilmenites might perhaps be phenocrystal in origin and unrelated to kimberlite and presumably diamond genesis. If this comment is valid, and considering the hypothesis of BOYD and NIXON (1973) regarding the origin of magnesian ilmenite-silicate intergrowths, then it could be suggested that diamond has formed at some other region G
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
793
than at the top of the low velocity zone. Experimental work on the role of magnesium in ilmenite at pressure should clarify some of the problems. The results of this study are in general agreement with the earlier conclusions of MEYER and BoYD (1972) and SOBOLEV et al. (1969a), that many of the inclusions formed in equilibrium with diamond. Undoubtedly, the inclusions have been armoured by their envelopment in diamond. This fact does not oppose the views of SOBOLEV et al. (1969a) regarding the genesis of these minerals as suggested by GURNEY and SWITZER (1973). It is obvious that detailed mineralogical investigations of kimberlites are a necessary requisite to interpreting inclusion data. It is important that the division of inclusions into an "ultramafic suite" and an "eclogite suite" be not overlooked. The presence of diamond as a constituent in eclogite xenoliths is fairly common (SOBOLEV et al., 1966a, b, 1972; GURNEY et al., 1969) but diamond in ultramafic xenoliths is rare (V. S. SOBOLEV et al., 1969b). This is surprising in view of the commonness of inclusions of the "ultramafic suite" compared with those resembling eclogitic minerals. Diamond is stable in both the eclogite and peridotite stability fields (Fig. 3). The region of possible diamond growth suggested by MEYER (1968) coincides very closely with the region in which BoYD'S (1973) granular xenoliths equilibrate.
16
/
I
14 •SHIELD
GEOTHERM
8
6L......~-IJ----=-----olL---4-h0--~!!!--"6;1;;-O
I
100
--....1...--'801;--"--...........- - - '
200
FIG 3 POSSible diamond growth region (MEYER, 1968) combmed with other pertment data. Shield geotherm-RINGWOOD et at (1964), Peridotite sohdus--ITo and KENNEDY (\ 967), basalt to eclogite --ITO and KENNEDY (1970); graphite to dlamond- -BERMAN (\965); granular and sheared xenohths--BoYD (1973), chrome-pyrope (Cr -py) synthesis-BYKOVA and GENSCIIAFT (1972); coeslte to stishovite-AKIMoTO and SYONO (1971).
There is a large variation in the detailed mineralogy and xenolith content of different kimberlites. Such variation can obscure important local consistences. At present there is a wide geographic coverage of data on inclusions in diamond from many localities. In order
794
H.
O. A. MEYER AND D.
P.
SVISERO
to remove world-wide variation we intend to examine in detail the mineralogy and chemistry of specific diamond-bearing kimberlites, including the inclusions in diamond. Due to lack of pertinent information on the mineralogy of kimberlites, and also to lack of experimental data, the role of many inclusions, for example quartz, is enigmatic. HARRIS (1968) stresses the fact that quartz is an inclusion and states that "...no obvious fractures emanate from the specimen to the diamond surface". However, in spite of this Harris considers quartz to be epigenetic on the basis of his criteria that many included minerals do not have stability fields at the conditions envisaged for diamond synthesis. Undoubtedly the time has come to more critically appraise the criteria by which we decide what are primary or secondary inclusions.
ACKNOWLEDGEMENTS This study was supported by National Science Foundation contract GA-36141. Advanced Research Projects Agency contract DAHC-0213 provided partial support for purchase of the microprobe. We thank Dr. R. H. McCallister for helpful comments regarding this study. Svisero gratefully acknowledges the financial support by Fundacao de Amparo a Pesquisa do Estado de Sao Paulo, Sao Paulo, Brazil.
REFERENCES AKIMOTO. A. and SYONO, Y (1971) The coesite-stishovlte tranSItIOn. In: Accurate charactertzatlOn of the hlghpressure environment (edItor LLOYD), N.B.S. Spec. Publ 326, pp 273-7 BERMAN, R. (1965) Thermal properties. In: Physical Properlies of Diamond (editor BERMAN), Clarendon Press, Oxford. BOYD, F R. (1969) Electron probe study ofdlOpside inclUSIOns In kImberlIte Amer J Sci 267A. 50-69 BOYD. F. R. (1973) The pyroxene geotherm. Geoclllm. Cosmochlm. Acta 37, 2533-46. BOYD, F. R and NIXON, P. H. (1973) Ongin of the Ilmenite-sIlicate nodules In kimberlites from Lesotho and South Afnca. In: Lesotho Klmberlites (editor P. H. NIXON), pp. 254-68. Lesotho National Development Corp, Maseru. BYKOVA. Yu. M and GENSHAFT, Yu S (1972) Synthesis of chrome garnets In the pyropeknornnglte senes (In RUSSIan) Geokhlmlya 10, 1291-3. CAMARGO, W. G. R and LEITE, C. R. (1968) OlIVIne: InclUSIOns In dIamonds from Brazil (In Portugese.) Anal,l Acad Bras Ctenc., Supp!. Spec, pp 89-92. CORRENS, C. W. (1931) Diamond wIth quartz InclUSIOns. (In German.) Zeit Krist. 80, 37-44. GURNEY, J J., SIEBERT, J C. and WHITfIELD, G G (1969) A dIamondlferous eclogIte from the Roberts Victor mme. Geol. Soc S Ajr., Spec. Pub!. No 2, pp. 351-7. GURNEY, J J and SWITZER, G. S. (1973) The discovery of garnets closely related to diamonds m the Fmsch PIpe, South Africa. Contr Mineral. Petrol 39, 103 16 HARRIS, J. W (1968) The recognItIOn of diamond mcluslOns. PI. I SyngenetIc mmeral mcluslOns Pt II: Epigenetic mineral mcluslOns Indust Dw. Rev., London, pp 402-10,458 61. ITO, K. and KENNEDY, G. C (1967) Meltmg and phase relations m a natural pendotlte to 40 kiiobars. Amer J. SCI.
265, 519-39.
ITO, K and KENNEDY, G. C. (1970) The fine structure of the basalt-eclogite transitIOn. Minerul Soc. Amer Spec Pap. 3, nH KENNLDY, G. C. and NORDLIE. B. E. (1968) The genesIs of dIamond depOSit, Econ. Geol. 63. 495-503. KHARKIV, AD, SOBOLEV, N V. and CHUMIR1N, K. G (1972) InclUSIOns of chrome dlOpslde In Zircon from kimberlite of Malo-Botuohy dlstnct (In Rus<;Ian ) Zay V~es Mtneralol/. Oh"lullc'lUI 101. 431 3 MACGREGOR, I D. (1974) The system MgO-AI 2 0,- SI0 2 . the solubility of AI 2 0 3 m enstatIte for spmel and garnet pendotlte compOSitions. Amer. Mmerul 59, 110--19. MCCALLISTER, R. H., MEYER, H. 0. A. and BROOKINS, D C. (1974) "Pyroxene" -Ilmenite xenoliths from the Stockdale pipe, Kansas: Chemistry, crystallography and origin (thiS volume). MEYER. H. 0. A. (1968) Inclusions in diamonds Carnegie Inst. Wash. Yearh 66.446 50
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
795
MEYER, H. O. A. (\968) Chrome pyrope: an inclusion m natural diamond. Science 160, 1446--7. MEYER, H. O. A. and BOYD, F. R. (\969) Mineral inclusions in diamond. Carnegie lnst. Wash. Yearb.67, 130-5. MEYER, H O. A. and BoYD, F. R. (\ 970) Inclusion in diamonds. CarnegIe lnst. Wash. Yearb. 68, 315-22. MEYER, H. 0 A. and BOYD, F R (\ 972) CompOSitIOn and origin of crystalline inclusions in natural diamonds. Geochim. Cosmochim Acta 36, 1255~73. MILLEDGE, H. J. (\961) Coeslte as an inclusion m G.E.C. synthetic diamonds. Nature 190, 118\. MITCHELL, R. H. (1973) Magnesian ilmenite and its role in kimberlite petrogenesis. 1. Geol. 81, 301~1\. ORLOV, Yu. L. (\ 959) Synthetic and epigenetic inclUSIOns in the crystals of diamond. (In Russian.) Mineral. Mus. Acad. Sci. SSSR, Trans. 10, I 03~ 20. RINGWOOD, A. E., MACGREGOR, I D. and BOYD, F. R. (1964) Petrological constitution of the upper mantle. Carnegie lnst. Wash. Yearb.63, 147-52. SHARP, W. E. (1966) Pyrrhotite: a common inclUSIOn in South Afncan diamonds. Nature 211, 402-3. SOBOLEV, N. V., BARTOSHINSKY. Z. V., YEFIMOVA, E. S., LAVRENT'YEV, Yu. G. and POSPELOVA, L N. (1970) Association of olivine, garnet and chrome diopside in a Yakutsk diamond. (In Russian.) Dok!. Akad. Nauk SSSR 192, 1349~53. SOBOLEV, N. V., BOTKuNov, A I., KUZNETSOVA, I K. and KHARKIV, A. D. (\966a) A new discovery of diamondiferous ecologlte in "Mir" pipe, Yakutia. (In Russian.) Geol. Geophys. 11, 114~16 SOBOLEV, N. V., BOTKuNov, A. I., LAVRENT'YEV, Yu G. and POSPELOVA, L. N. (1971) Peculiarities of the composition of the minerals coexlstmg With diamonds from the Mir pipe, YakutIa. (In Russian.) Zap. Vses. Mineralog. Obshchestt'a 100, 558~64. SOBOLEV, N, V. and KUZNETSOVA. I. K. (l966b) Mineralogy of diamondiferous eclogites (In RUSSian.) Dok!. Akad. Nauk SSSR 167, 1365-8. SOBOLEV, N. V" KUZNETSOVA, I. K and ZYUZIN, N. I. (1968) The petrology of grospydlte zenoliths from the Zagadochnaya kimberlite pipe in YakutIa. J. Petrol. 9, 253-80. SOBOLEV, N, V. and LAVRENT'YEV, Yu. G. (1971) Isomorphic sodIUm admixture in garnets formed at high pressures. Contr. Mineral. Petrol. 31, 1~ 12. SOBOLEV, N, V" LAVRENT'YEV, Yu. G. and USOVA, L. V. (1972) Mmor elements m rutiles from eclogites. (In Russian.) Geol. Geophys. 11, 108~12. SOBOLEV, N. V., LAVRENT'YEV, Yu. G., POKHILENKO, N. P. and USOVA, L. V. (1973) Chrome-nch garnets from the kimberhtes of Yakutia and their parageneses. Contr. Mineral. Petrol. 40. 39~52. SOBOLEV, N. V., LAVRENT'YEV, Yu. G., POSPELOVA, L. N. and SOBOLEV, E. V. (1969a) Chrome pyropes from diamonds ofYakutia. (In Russian.) Dok!. Akad. Nauk SSSR 189, 162-5. SOBOLEV, V. S., NAI, B. S., SOBOLEV, N. V.. LAVRENT'YEV, Yu. G. and POSPELOVA, L. N. (l969b) Xenoliths of diamond beanng pyroplc serpent mites from Aykhal pipe, Yakutia. (In RUSSian.) Dok!. Akad Nauk SSSR 188, 1141~3
SOBOLEV, V. S., SOBOLEV, N. V. and LAVRENT'YEV, Yu. G. (1972) InclUSIOns in diamonds from diamondiferous eclogite. (In Russian.) Dokl Akad. Nauk SSSR 207,164-7. SVISERO, D. P (\ 971) Mineralogy of diamonds from the region of Alto Araguaia. (In Portugese.) Doctoral TheSIS, Inst. of Geoscience, Univ. Sao Paulo, Brazil, 137 pp.
ABSTRACT Mmerals mcluded m diamonds from several localIties m Brazil have been exammed. The mmerals observed are olIvme, enstatIte, chrome-pyrope and pyrope-almandme. These minerals have simIlar compositions to those previously reported for mcluslOns from other worldwide localItIes. In addItIOn rutIle, Ilmemte, Zircon, quartz and pyrrhotite occur as mcluslOns. This IS the first record of Ilmenite and zircon m natural diamond The ilmenite IS umque compared to most kImberlIte Ilmemtes m that It contains no appreciable content of magnesIUm. The presence of quartz as an mcluslOn IS unusual.
INTRODUCTION Natural diamonds are generally considered to be the products of stable growth (e.g. KENNEDY and NORDLIE, 1968). On this assumption it has been argued (MEYER and BOYD, 1972) that many crystallme inclusions in diamond represent original phases which formed in equilibrium with diamond. Detailed chemical examination of these minerals should provide valuable data pertinent to the genesis of diamond, to kimberlite, and to the associated xenoliths. Recent studies of mineral inclusions in diamonds from South Africa, West Africa, Venezuela and Thailand (MEYER, 1971; MEYER and BOYD, 1969, 1972) and from Siberia and the Urals (SOBOLEV et al., 1969a, 1970, 1971, 1972, 1973) have shown a remarkable similarity in the compositions of the various minerals. Furthermore, the inclusions (olivine, enstatite, dlOpside, garnet, chromite) although generally similar to their counterparts in kimberlite, and associate xenoliths, have elemental abundances that are distinct in detail. In an attempt to extend the present geographic coverage of diamonds and their inclusions we have examined specimens from Brazil. In general, these new results are in agreement with earlier observations. In addition we have analyzed rutile, ilmenite, pyrrhotite, zircon and quartz, all of which occur as inclusions.
GEOLOGIC OCCURRENCE Although diamonds have been found in Brazil since about 1725 no kimberlite has yet been recorded. Mostly the diamonds are found in river placer deposits throughout several * Permanent address' Instltuto de Geosclenclas, Umversldade de Sao Paulo, Sao Paulo, BraZIl. 785
786
H. O. A. MEYER AND D. P. SVISERO
areas, mainly in the states of Minas Gerais, Mato Grosso, Goias, Bahia, Parana and Para. In east-central Brazil (Diamantina in Minas Gerais and Chapada Diamantina in Bahia) the diamonds are associated with metasediments, conglomerates, arenites, phyllites and quartzite, of upper Precambrian age (Fig. 1).
Morro do
GUlmaraes
....
Rondonopol1s
,
.
Alto,Araguala A
Bolivia
L.
".>,
"'Itulutaba
<
... frlltalo::..~
IDEV
Paraguay
..
Dlamantlna
COlli"'.
PARANA
Ch~peu
len~ols
IPRE~CAMB I
Datn OCEAN
1
BASIN (RET!
..
Jaguariaiva llbilJI.
-
""
,-
Argentina
.D,amond Local,t,es
FIG I. LocalItIes in BrazIl from whIch diamonds were obtamed
In
thIs study
In contrast, the other major region of diamond production is associated with the large sedimentary Parana basin in south central Brazil (Fig. I). Devonian and Carboniferous arenaceous sediments form the rim of the basin and are unconformable on Precambrian basement rocks. The diamond placer deposits are predominantly found in the rivers draining the Devonian strata, but placers do occur in rivers on the Carboniferous rocks. It is generally thought that the Devonian and Carboniferous sediments are the secondary matrix for the diamonds which were previously derived from Precambrian rocks. Currently, several alkalic intrusions of Cretaceous age have been found bordering the Parana basin. This has led some to suggest the possibility of Cretaceous kimberlites as a source of diamonds.
SAMPLE DESCRIPTION AND TECHNIQUES The specimens exammed in this study are from diamonds obtained in both of the above regions. All the diamonds were collected from placer deposits in the most productive areas (Fig. 1). Morphologically, the diamonds are complex but most are characterized by the predominance of dodecahedral forms and a large number of contact twins. Infra-red studies
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
787
showed that about 95~~ of the diamonds examined are of Type I. The specific gravity of the diamonds varied between 3.500 and 3.530 g/cm 3 , and trace elements detected by spectrographic emission analysis include AI, Ca, Mg, Si, Fe, Cu and Cr (SVISERO, 1971). The majority of the diamonds studied were colorless and were between 2 and 10 mm in size. They were cleaned in HCl and HF for several hours and examined for lack of fractures reaching from the surface. The inclusions were all released by cracking the diamond host in a small enclosed container. Following optical examination the inclusions were mounted for scanning electron microscope observations, X-ray diffraction and electron-microprobe analysis. The analyses were obtained using a M.A.C. 500 electron microprobe. Standards were predominantly glasses or pure metals. The analytical techniques and data reduction procedures used are similar to those described by BoYD (1969). Some initial characterization of the inclusions was done using a qualitative energy dispersive system on the scanning electron microscope.
MINERALOGY The majority of minerals found as inclusions in the Brazilian diamonds are very similar in composition to those previously recorded (MEYER, 1968; MEYER and BOYD, 1972; SOBOLEV et al., 1973). These inclusions are, namely, olivine, enstatite, chrome-pyrope, almandine-pyrope and clinopyroxene. In addition we have found zircon, ilmenite, rutile, pyrrhotite and quartz. To avoid repetition in the following sections reference to "previous results" will refer specifically to MEYER and BOYD (1972) and SOBOLEV et al. (1969a, 1970, 1971. 1972, 1973).
OLIVINE Olivine was first identified as an inclusion in Brazilian diamonds by CAMARGO and LEITE (1968) using X-ray techniques. It is the most common inclusion observed. In composition
the olivines are similar to previously reported analyses (Table 1). They are characterized by a small range in composition (F092 to F0 9S ), relatively low contents of the minor elements AI, Ca, Mn and Ni. and high values (up to 0.10 wt.°/~) ofCr 2 0 J . These Cr 2 0 J values are an order of magnitude greater than for other terrestrial forsteritic olivines.
ENSTATITE Enstatite appears to be more common in Brazilian diamonds than in others previously examined. The present analyses, however, agree closely with those of enstatite i!1clusions from other localities. The enstatite molecule (MgSiO J ) accounts for 93-94~;; of the total mineraL there being very little solid solution towards diopside. Both Cr 2 0 J and Al 2 0 3 contents are low (Table 2). Enstatite coexists with both olivine and chrome-pyrope.
I> \( () rl
\,\1 \)
HB
788
1
H. O. A. MEYER AND D. P. SVISERO TABLE
1.
REPRESENTATIVE ANALYSES OF OLIVINE INCLUSIONS IN DIAMOND
BA35A
Si0 2 Ti0 2 Al 2 0 3 Cr 2 0 3 FeO MgO CaO MnO NIO
41.3
MG26A
-PR38A--
-----
--------
40.7
40.9 0.02 0.02 0.10 7.17 52.4 0.04 0.10 0.28
< 0.01 < 0.01
< 0.01 0.02
0.05 7.98 50.1 0.03 0.11 040
0.10 7.84 504 0.03 0.10 028
100.0
99.5
MT15B*
MTI2Bt
+----~
41.6
42.1
< 0.01
< 0.01
0.04 0.10 7.06 49.0 0.07 0.11
0.04 010 542 52.2 0.03 0.10
2_]__~_0.4~~
03
101.0
98.4 "~--
100.4 ~--
No of cations ( x 1000) for 4 oxygen
l
r-----~~
SI Al Cr Fe+ 2 Mg Ca Mn Ni °0 Fa
1006
996 I
I 162 1814 1 2
160 1836 I
8
983 1 2 144 1877
1022 I 2 145 1794
1 2
2
6
91.8
2 108 1860 I
2 7
5
92.0~ 92.9 _
1008 I
2 8 94.5
92_.5_
L
* CoexIsts wIth chrome-pyrope (MTI5A). t CoexIsts with enstatite (MTI2A)
TABLE
SI0 2 TI0 2
AI 2 0 3 Cr 2 0 3 FeO MgO CaO MnO NiO
2.
REPRESENTATIVE ANALYSES OF ENSTATITE INCLUSIONS IN DIAMOND
MG27A
G055A
57.3 < 0.01 051 0.34 4.75 36.2 0.18
57.5 < 0.01 060 0.47 4.68 37.8 031
o.IOj 0.39 -
-
-
--B~:~B - ---=~G2IB*
0.10 0.11
57.6 < 0.01 0.87
57.6 < 0.01 078
062 4.27 36.6 0.14 0.13 Ql1
0.47 436 36.2 047 0.11
1- MTI2At~ 57.6
< 0.01
I
051 0.11 4.95 35.3 0.54 034 Q02
--~---
99.8
100.6
--~----
100_.3
---
1000_
J __ 99.4 _ _
No. of catIOns I x 1000) for 6 oxygen
SI Al Cr Fe Mg Ca Mn NI ° En 0
1972 21
9
[37 1855 7 3 [I 93.1
1960--TI9M--- 24 13 133 1874
1~
35 17 122 1861
1_~
---=~
93~__J
1971 31 13 125 1846 [7 3
1987 21 3 143 1818 20 10
~~ _~_9~
* CoexIsts wIth chrome-pyrope (MG21A) t CoexIsts wIth olIVIne (MTI2B).
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
789
CLINOPYROXENE In keeping with earlier observations, clinopyroxenes are comparatively rare as inclusions
in diamond. In this study two inclusions identified by energy dispersive technique as possible diopside (high Ca, Mg, and Si) were found. Unfortunately, both inclusions were very small, < 30l1m, and were lost during polishing.
GARNET Garnet appears to be the second most common inclusion in Brazilian diamonds. Two distinctive suites occur: one, Cr-rich, Ca-poor pyrope; and two, Fe-rich, Cr-poor almandine-pyrope garnet (Table 3). The latter group is minor. SOBOLEV et al. (1973) discuss in
TABLE 3. REPRESENTATIVE ANALYSES OF GARNET INCLUSIONS IN DIAMOND
P2~-' 1
'.IT;;Al
-_._-
MG2IA*
Si0 2 TIO z AlzO J CrzO, FeO MgO CaO MnO
< 0.01 10.4 17.3 5.84 24.6 0.93
41.2 < 0.01 14.8 10.9 6.30 23.1 243 0.35
99.5
99.1
--_._---- - -
BA43A
G08A
MG25A
41.1 < 0.01 15.6 10.3 6.23 23.5 2.48 0.33
41.5 < 0.01 16.7 9.42 6.07 23.4 230 0.33
99.5
997
409 < 0.01 16.2 9.34 6.51 21.5 5.09 0.25
- - -----_.-
' - -
Si Ti Ai Cr Fe Mg Ca Mn
2984 l}O9 1014 361 2711 74
-I'
99.8
---~-
-
MT22A
G057A
1---- .
.--
40.5 0.86 19.6 0.20 16.0 13.1 8.39 0.34
39.8 1.05 20.3 011 16.4 11.5 8.46 0.56
39.9 0.85 20.3 0.09 17.0 12.7 8.32 0.50
99.0
98.2
99.7
-
-----
No. of catIOns ( x 1000) for 12 oxygen 3012 1275 630 386 2519 191 21
T
I
2989 1334 590 379 2544 193 20
- -I 2990
1420 537 366 2518 178 20
::77 1394 2977
r-"30~~
1739 538 12 397 1007 ~38 1470 398 677 ~_-----=-_
-
---
3029 60 1821 7 1046 1310 690 36
3001 48 1800 6 1069 1421 670 32
* Coexists With enstatite (MG2IB).
t Coexists With olivine (MT15B).
detail the parageneses of Cr-rich garnets from kimberlites in Yatkutia. They conclude that the Cr-pyropes which occur as inclusions in diamond can be differentiated from other Cr-rich garnets on the basis of their low content ofCaO (see also SOBOLEV et ai., 1969a, 1971). The analyses presented here of garnets from Brazilian diamonds are similar to those previously examined from other world wide localities (Fig. 2). Recently, chrome-pyrope garnets, similar in composition to those occurring as inclusions in diamond, have been found to be present in concentrates in the Finsch Pipe, South Africa (GURNEY and SWITZER, 1973).
790
H. O. A. MEYER AND D. P. SVISERO
' 'f I..
j
1--GARNETS
I I I
80r I I I 0 0
c
Inclusions
In
Diamond
80
Peridotite Xenoliths
and KimberlIte
In
Kimberlite
,:,
•
~I~:
'.'. iI
I. c"2JI:
l'J
~
~
1~l'J
••
60
0
o
o
..
® Eclogite Xenoliths
60
x
100
40
'.'
i
§g.,6
.
~ ...
J
•
0
6t
....
I
20
~
,
Q)6!J0
40
60
·fl
0
I
80
, ~
I
100
0
0
r
0
0
0
0
20 0
.....
~
40
"
J
.0 0
•
0
eli·
20
0
°
0
C
\5:, ~' 40
20
MgO X100 [MgO + FeOl
CoO X100 [CoO +MgO J
FI
xenolIth,. from kimberlIte and from eclogite xenolIths In kimberlIte BrazilIan mcluslOns-solId trIangle,; SiberIan Inclusions· solId squares, other mcluslOns (M~.YER and BOYD, 1972)- ,olId Circles (data as for MEYER and BOYD, 1970 plus SOBOL~.V I!f al.. 1969a, 1973 and GURNEY and SWITZER, 1973)
RUTILE Rutile has been observed as an inclusion in diamond by HARRIS (1968) and GURNEY et al. (1969) using X-ray diffraction methods. We have found and analyzed rutile in Brazilian diamonds (Table 4). The rutile inclusions were generally elongated parallel to c, euhedraL reddish-brown in color and moderately transparent. From single crystal diffraction data TABLE 4 Zircon MT30A
Sl02 TI0 2 AI 2
°-'
Cr 2 0
FcO MgO
CaO MnO Zr0 2
J
31.1 0.03
om
-
-
-l
RFPR~.sI,'HATlVF
R~~e
GO~IA +~TI6~ 006 99 X 003 0.\6
ANALYSES OF INCLUSIONS IN DIAMOND
- MTIOA
BAl9A
G049A
il.11
0.18 50.5 022 003
< 0.01
0.73
012 519 0.21 < 0.01 47.6 0.14 0.03 064
011> 514 03(, 004 482 014 < 001 0.6X
I
00') 99 6
I
Ion
I 'I
~:~~ ~:~~~I 4~~~
001 0.02 69.7
< < 0.01
1009
100.5
I
Pyrrhotite
Ilmenite
<
Fe S NI Cr Cu
100 I
1010
MT36B
(,2.4 37.8 0.48 0.02 001
61.2 38.2
100.6 Fe.
}t00-4-._"- 992
MT36A
0.4
1.26 0.03 0.01
100.7 091
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
791
the cell dimensions were determined to be a = 4.59 A and c = 2.95 A. Analyses show the rutile to be extremely pure with less than 0.4 wt.~,;, of other oxides. These results are in agreement with those of SOBOLEV et al. (1972b) who examined rutile from eclogitic xenoliths in kimberlite.
ILMENITE Ilmenite was extracted from three diamonds which all satisfied the basic criteria by which we judge inclusions to be primary (i.e. inclUSIOn totally within diamond and no visible fracture from inclusion to diamond surface). The ilmenites appear to be close to stoichometry (Table 4), manganese being the major impurity «0.7 wt.~o MnO). This composition of FeTi0 3 contrasts markedly with that of ilmenites that occur as xenocrysts and in xenoliths in kimberlIte. These latter ilmenites all have appreciable contents of MgO (e.g. MCCALLISTER, MEYER and BROOKINS, this vol.; MITCHELL, 1973) which may reach up to 20 wt. 0;,. Ilmenite has been recorded previously by HARRIS (1968) as a syngenetic inclusion. Harris observed the ilmenite to be powdered and to be commonly exposed in holes on diamond surfaces.
ZIRCON Zircon is not uncommon as a mineral in kimberlite (KHARKIV et aI., 1972) but this is believed to be the first substantiated record of its occurrence as an inclusion in diamond. The zircon appears very pale brown in color and is slightly elongated and its analysis (Table 4) indicates that the minor elements content is low.
QUARTZ Quartz and coesite have at various times been considered as inclusions in natural diamonds (MILLEDGE, 1961; HARRIS, 1968; ORLOV, 1959). In this study we have obtained, apparently as a primary inclusion in terms of identification criteria, an inclusion of quartz. The specimen was colorless and had well-developed faces. The identification was substantiated by both single crystal X-ray diffraction methods and by electron-microprobe analysis. Some caution in accepting this as a primary inclusion is perhaps warranted since quartz is known to occur as an intergrowth in Brazilian diamonds (CORRENS, 1931; BOYD and MEYER, unpubl.).
SULFIDES Examples of sulfides llCcurring as inclusions in diamonds are well known and SHARP (1966) has described troilite and pentlandite as relatively common (among opaque inclusions). HARRIS (1968) also mentions the identification, using X-ray techniques, of pyrrhotite and pentlandite. In Brazilian diamonds inclusions are opaque, small and irregular. Analyses confirmed the sulfide as pyrrhotite (Table 4).
792
H. O. A. MEYER AND D. P. SVISERO
DISCUSSION The results presented above agree with earlier studies of mineral inclusions in diamond. For example, many of the silicates are remarkably uniform in composition irrespective of provenance and age. Olivine and chrome-pyrope are the most abundant inclusions. Furthermore, several inclusions are essentially similar in composition to their counterparts in ultramafic rocks whereas others are similar to the constituent minerals of eclogite. It is usual to find only those minerals belongmg to the same suite coexisting in a single diamond. For example, in this study we have observed the following pairs of inclusions: olivine + chrome-pyrope, olivine + enstatite, enstatite + chrome-pyrope. This last pair has not previously been recorded. The above minerals could all be considered as members of an ultramafic suite. In contrast MEYER and BoYD (1972) have described clinopyroxene + almandine garnet from a single diamond, both minerals being in some respects similar to minerals from eclogite. Also, SOBOLEV et al. (1970) have discussed the assemblage olivine + garnet + chrome-diopside in one diamond. The garnet in this assemblage is unique as it contains approximately 8 wt./~ Cr 2 0 3 but unlike the normal chrome-pyrope inclusion it has an extremely high content of CaO (13 wt. ~<,). Sobolev points out that the garnet is close in composition to some rare pyrope-uvarovites from kimberlite (SOBOLEV et aI., 1968). The rarity of clinopyroxene inclusions is surprising in view of the abundance of this mineral in kimberlite and associated xenoliths. In the Brazilian diamonds the overall assemblage olivine + garnet + enstatite is present. MACGREGOR (1974) has examined the solubility of AI 2 0 3 in enstatite for garnet and spinel peridotite composition. Using his data plus the AI 2 0 3 contents of the enstatite inclusions and assuming the assemblage consists of garnet + orthopyroxene + clinopyroxene, a minimum of 40 kb at lO00°C is obtained for possible equilibration temperatures. Their values are approximate since no account has been taken of the other trivalent elements (e.g. Cr and Fe + 3). According to MACGREGOR, consideration of all the trivalent elements would generally increase the above values of P and T to the region of approximately 130n C and 50kb. BoYD and NIXON (1973) have plotted Ca/(Ca + Mg) ratios of enstatites against temperature of equilibration for coexisting diopsides. Using their data, it appears enstatite inclusions should have formed at less than 800'C. This is in sharp contrast to the values obtained from MACGREGOR'S (1974) data. It is thus very likely that the assemblage enstatite + olivine + chrome-pyrope did not coexist with diopside in diamond. This conclusion is in agreement with that of SOBOLEV et al. (1973) and SOBOLEV et al. (l969a, b). The presence of rutile as an inclusion in diamond has been noted previously (HARRIS, 1968; GURNEY et al., 1969). It is noteworthy that the diamond which contained the rutile inclusions observed by GURNEY et al. was found within an eclogite xenolith. The listing of rutile inclusions in the suite of minerals with eclogite affinities is thus fairly well established. The occurrence of pure ilmenite as an inclusion is puzzling, especially when one considers that all kimberlitic ilmenites have relatively high contents of MgO. It is unlikely that these ilmenites were adhering to cavities within the diamond surface because of the cleaning procedures used prior to careful optical examination of the diamonds. MITCHELL (1973) comments that magnesian ilmenites might perhaps be phenocrystal in origin and unrelated to kimberlite and presumably diamond genesis. If this comment is valid, and considering the hypothesis of BOYD and NIXON (1973) regarding the origin of magnesian ilmenite-silicate intergrowths, then it could be suggested that diamond has formed at some other region G
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
793
than at the top of the low velocity zone. Experimental work on the role of magnesium in ilmenite at pressure should clarify some of the problems. The results of this study are in general agreement with the earlier conclusions of MEYER and BoYD (1972) and SOBOLEV et al. (1969a), that many of the inclusions formed in equilibrium with diamond. Undoubtedly, the inclusions have been armoured by their envelopment in diamond. This fact does not oppose the views of SOBOLEV et al. (1969a) regarding the genesis of these minerals as suggested by GURNEY and SWITZER (1973). It is obvious that detailed mineralogical investigations of kimberlites are a necessary requisite to interpreting inclusion data. It is important that the division of inclusions into an "ultramafic suite" and an "eclogite suite" be not overlooked. The presence of diamond as a constituent in eclogite xenoliths is fairly common (SOBOLEV et al., 1966a, b, 1972; GURNEY et al., 1969) but diamond in ultramafic xenoliths is rare (V. S. SOBOLEV et al., 1969b). This is surprising in view of the commonness of inclusions of the "ultramafic suite" compared with those resembling eclogitic minerals. Diamond is stable in both the eclogite and peridotite stability fields (Fig. 3). The region of possible diamond growth suggested by MEYER (1968) coincides very closely with the region in which BoYD'S (1973) granular xenoliths equilibrate.
16
/
I
14 •SHIELD
GEOTHERM
8
6L......~-IJ----=-----olL---4-h0--~!!!--"6;1;;-O
I
100
--....1...--'801;--"--...........- - - '
200
FIG 3 POSSible diamond growth region (MEYER, 1968) combmed with other pertment data. Shield geotherm-RINGWOOD et at (1964), Peridotite sohdus--ITo and KENNEDY (\ 967), basalt to eclogite --ITO and KENNEDY (1970); graphite to dlamond- -BERMAN (\965); granular and sheared xenohths--BoYD (1973), chrome-pyrope (Cr -py) synthesis-BYKOVA and GENSCIIAFT (1972); coeslte to stishovite-AKIMoTO and SYONO (1971).
There is a large variation in the detailed mineralogy and xenolith content of different kimberlites. Such variation can obscure important local consistences. At present there is a wide geographic coverage of data on inclusions in diamond from many localities. In order
794
H.
O. A. MEYER AND D.
P.
SVISERO
to remove world-wide variation we intend to examine in detail the mineralogy and chemistry of specific diamond-bearing kimberlites, including the inclusions in diamond. Due to lack of pertinent information on the mineralogy of kimberlites, and also to lack of experimental data, the role of many inclusions, for example quartz, is enigmatic. HARRIS (1968) stresses the fact that quartz is an inclusion and states that "...no obvious fractures emanate from the specimen to the diamond surface". However, in spite of this Harris considers quartz to be epigenetic on the basis of his criteria that many included minerals do not have stability fields at the conditions envisaged for diamond synthesis. Undoubtedly the time has come to more critically appraise the criteria by which we decide what are primary or secondary inclusions.
ACKNOWLEDGEMENTS This study was supported by National Science Foundation contract GA-36141. Advanced Research Projects Agency contract DAHC-0213 provided partial support for purchase of the microprobe. We thank Dr. R. H. McCallister for helpful comments regarding this study. Svisero gratefully acknowledges the financial support by Fundacao de Amparo a Pesquisa do Estado de Sao Paulo, Sao Paulo, Brazil.
REFERENCES AKIMOTO. A. and SYONO, Y (1971) The coesite-stishovlte tranSItIOn. In: Accurate charactertzatlOn of the hlghpressure environment (edItor LLOYD), N.B.S. Spec. Publ 326, pp 273-7 BERMAN, R. (1965) Thermal properties. In: Physical Properlies of Diamond (editor BERMAN), Clarendon Press, Oxford. BOYD, F R. (1969) Electron probe study ofdlOpside inclUSIOns In kImberlIte Amer J Sci 267A. 50-69 BOYD. F. R. (1973) The pyroxene geotherm. Geoclllm. Cosmochlm. Acta 37, 2533-46. BOYD, F. R and NIXON, P. H. (1973) Ongin of the Ilmenite-sIlicate nodules In kimberlites from Lesotho and South Afnca. In: Lesotho Klmberlites (editor P. H. NIXON), pp. 254-68. Lesotho National Development Corp, Maseru. BYKOVA. Yu. M and GENSHAFT, Yu S (1972) Synthesis of chrome garnets In the pyropeknornnglte senes (In RUSSIan) Geokhlmlya 10, 1291-3. CAMARGO, W. G. R and LEITE, C. R. (1968) OlIVIne: InclUSIOns In dIamonds from Brazil (In Portugese.) Anal,l Acad Bras Ctenc., Supp!. Spec, pp 89-92. CORRENS, C. W. (1931) Diamond wIth quartz InclUSIOns. (In German.) Zeit Krist. 80, 37-44. GURNEY, J J., SIEBERT, J C. and WHITfIELD, G G (1969) A dIamondlferous eclogIte from the Roberts Victor mme. Geol. Soc S Ajr., Spec. Pub!. No 2, pp. 351-7. GURNEY, J J and SWITZER, G. S. (1973) The discovery of garnets closely related to diamonds m the Fmsch PIpe, South Africa. Contr Mineral. Petrol 39, 103 16 HARRIS, J. W (1968) The recognItIOn of diamond mcluslOns. PI. I SyngenetIc mmeral mcluslOns Pt II: Epigenetic mineral mcluslOns Indust Dw. Rev., London, pp 402-10,458 61. ITO, K. and KENNEDY, G. C (1967) Meltmg and phase relations m a natural pendotlte to 40 kiiobars. Amer J. SCI.
265, 519-39.
ITO, K and KENNEDY, G. C. (1970) The fine structure of the basalt-eclogite transitIOn. Minerul Soc. Amer Spec Pap. 3, nH KENNLDY, G. C. and NORDLIE. B. E. (1968) The genesIs of dIamond depOSit, Econ. Geol. 63. 495-503. KHARKIV, AD, SOBOLEV, N V. and CHUMIR1N, K. G (1972) InclUSIOns of chrome dlOpslde In Zircon from kimberlite of Malo-Botuohy dlstnct (In Rus<;Ian ) Zay V~es Mtneralol/. Oh"lullc'lUI 101. 431 3 MACGREGOR, I D. (1974) The system MgO-AI 2 0,- SI0 2 . the solubility of AI 2 0 3 m enstatIte for spmel and garnet pendotlte compOSitions. Amer. Mmerul 59, 110--19. MCCALLISTER, R. H., MEYER, H. 0. A. and BROOKINS, D C. (1974) "Pyroxene" -Ilmenite xenoliths from the Stockdale pipe, Kansas: Chemistry, crystallography and origin (thiS volume). MEYER. H. 0. A. (1968) Inclusions in diamonds Carnegie Inst. Wash. Yearh 66.446 50
MINERAL INCLUSIONS IN BRAZILIAN DIAMONDS
795
MEYER, H. O. A. (\968) Chrome pyrope: an inclusion m natural diamond. Science 160, 1446--7. MEYER, H. O. A. and BOYD, F. R. (\969) Mineral inclusions in diamond. Carnegie lnst. Wash. Yearb.67, 130-5. MEYER, H O. A. and BoYD, F. R. (\ 970) Inclusion in diamonds. CarnegIe lnst. Wash. Yearb. 68, 315-22. MEYER, H. 0 A. and BOYD, F R (\ 972) CompOSitIOn and origin of crystalline inclusions in natural diamonds. Geochim. Cosmochim Acta 36, 1255~73. MILLEDGE, H. J. (\961) Coeslte as an inclusion m G.E.C. synthetic diamonds. Nature 190, 118\. MITCHELL, R. H. (1973) Magnesian ilmenite and its role in kimberlite petrogenesis. 1. Geol. 81, 301~1\. ORLOV, Yu. L. (\ 959) Synthetic and epigenetic inclUSIOns in the crystals of diamond. (In Russian.) Mineral. Mus. Acad. Sci. SSSR, Trans. 10, I 03~ 20. RINGWOOD, A. E., MACGREGOR, I D. and BOYD, F. R. (1964) Petrological constitution of the upper mantle. Carnegie lnst. Wash. Yearb.63, 147-52. SHARP, W. E. (1966) Pyrrhotite: a common inclUSIOn in South Afncan diamonds. Nature 211, 402-3. SOBOLEV, N. V., BARTOSHINSKY. Z. V., YEFIMOVA, E. S., LAVRENT'YEV, Yu. G. and POSPELOVA, L N. (1970) Association of olivine, garnet and chrome diopside in a Yakutsk diamond. (In Russian.) Dok!. Akad. Nauk SSSR 192, 1349~53. SOBOLEV, N. V., BOTKuNov, A I., KUZNETSOVA, I K. and KHARKIV, A. D. (\966a) A new discovery of diamondiferous ecologlte in "Mir" pipe, Yakutia. (In Russian.) Geol. Geophys. 11, 114~16 SOBOLEV, N. V., BOTKuNov, A. I., LAVRENT'YEV, Yu G. and POSPELOVA, L. N. (1971) Peculiarities of the composition of the minerals coexlstmg With diamonds from the Mir pipe, YakutIa. (In Russian.) Zap. Vses. Mineralog. Obshchestt'a 100, 558~64. SOBOLEV, N, V. and KUZNETSOVA. I. K. (l966b) Mineralogy of diamondiferous eclogites (In RUSSian.) Dok!. Akad. Nauk SSSR 167, 1365-8. SOBOLEV, N. V" KUZNETSOVA, I. K and ZYUZIN, N. I. (1968) The petrology of grospydlte zenoliths from the Zagadochnaya kimberlite pipe in YakutIa. J. Petrol. 9, 253-80. SOBOLEV, N, V. and LAVRENT'YEV, Yu. G. (1971) Isomorphic sodIUm admixture in garnets formed at high pressures. Contr. Mineral. Petrol. 31, 1~ 12. SOBOLEV, N, V" LAVRENT'YEV, Yu. G. and USOVA, L. V. (1972) Mmor elements m rutiles from eclogites. (In Russian.) Geol. Geophys. 11, 108~12. SOBOLEV, N. V., LAVRENT'YEV, Yu. G., POKHILENKO, N. P. and USOVA, L. V. (1973) Chrome-nch garnets from the kimberhtes of Yakutia and their parageneses. Contr. Mineral. Petrol. 40. 39~52. SOBOLEV, N. V., LAVRENT'YEV, Yu. G., POSPELOVA, L. N. and SOBOLEV, E. V. (1969a) Chrome pyropes from diamonds ofYakutia. (In Russian.) Dok!. Akad. Nauk SSSR 189, 162-5. SOBOLEV, V. S., NAI, B. S., SOBOLEV, N. V.. LAVRENT'YEV, Yu. G. and POSPELOVA, L. N. (l969b) Xenoliths of diamond beanng pyroplc serpent mites from Aykhal pipe, Yakutia. (In RUSSian.) Dok!. Akad Nauk SSSR 188, 1141~3
SOBOLEV, V. S., SOBOLEV, N. V. and LAVRENT'YEV, Yu. G. (1972) InclUSIOns in diamonds from diamondiferous eclogite. (In Russian.) Dokl Akad. Nauk SSSR 207,164-7. SVISERO, D. P (\ 971) Mineralogy of diamonds from the region of Alto Araguaia. (In Portugese.) Doctoral TheSIS, Inst. of Geoscience, Univ. Sao Paulo, Brazil, 137 pp.