Canadian Mineralogist VoL 17, pp.469-482 (1979) THEROLE OF GOLLECTORSIN THE FORMATIONOF THE PLATINUMDEPOSITS IN THEBUSHVELD COMPLEX S.A. HIEMSTRA Natianal Institute tor Metallurgy, Prtvate Bag X3015, Ratrdburg 2125, Soutlt Alrica ABsrRAcr INTRoDUcTIoN We consider the formation of the platinum de- Most of the sulfide occu,rrences of the posits in the Merensky and UG-2 Reefs in connec- Bushveld complex are associated with enriclr- tion with the action of collectors settling under ments in platinum-group metals. However, there the influence of gravity. The sulfides in the UG-2 the Reef, too small to have settled on their own, were are only three known occurrences where carried down by larger chromite grains in pieey- concentrations of these elements reach value$ back fashion. The sulfide droplets adhered to the of economic significance. These are the Merens- chromite grains either because they could wet the ky Reef (under which, for present puryroses chromite more readily than the silica.tes, or nu- is included the Plat Reef), the Upper Chromitite cleation of sulfide liquid droplets could have taken Layer (or UG-2 Reef), and the three dunite place preferentialli otr the surfaces of chromite pipes Driekop, Onverwacht and Mooihoek in grains. An iron alloy, stable only under exceptionally the eastern Bushveld. In this paper, attention Iow oxygen fugacity, could have acted as an effi- is given to one aspect of platinum deposits: the cient collector of platinum-group metals, which processes gave to the con- could have caused exceptional enrichment of some collection that rise magma fractions. Some of the mineral intergrowths centration of the platinum-group metals. in the Merensky Reef may be relics of iron-plati- num grains that reacted with sulfur, to form intergrowths of iron-platinum alloy with base Gsot-octclr- CoNsroEnatloNs metal sulfides that are pseudomorphous after the original crystals. As shown by von Gruenewaldt (1977) the occurrencesof platinum-group metals of econo- mic significance in the Bushveld complex, the SoMMATRE Merensky Reef and the UG-2 Reef follou, each other fairly olosely.In the western Bushveld, On considdre la formation des gisements de the Merensky Reef is situated only about 150 platine des bancs Merensky et UG-2 et le r6le m a.bove the UG-2 layer (von Gruenewaldt d'agents collecteurs dans un champ gravifique. Les 1977), but in the eastern Bushveld at the Atok grains petits pour de sulfures du banc UG-2, trop mine, this distance is increased to 300 m pouvoir se d6poser d'eux-memes, auraient 6t6 trans- (Schwellnus er a/. 1976). Idealized sections port6s "i cheval" sur des cristaux de chromite" plus gros. Les gouttelettes de sulfure auraient adh6r6 through a portion of the Merensky suite of aux cristaux de chromite i cause soit d'une tendan- rocks are given in Figures I and 2. Cousins ce i mouiller Ia chromite plut6t que les silicates, & Feringa (1964) provided sections through soit de la nucl6ation pr6fdrentielle de gouttelettes the UG-2 suite of rocks. Descriptions of the de sulfure liquide sur la surface des cristaux de three dunite pipes can be found in Wagner chromite. Un alliage de fer, stable seulement sous (1973, pp. 50-92). fugacit6 d'oxygdne exceptionnellement basse" aurait Bushveld com- servi de collecteur pour les platinoides, Most workers agree that the enrichis- .form sant ainsi fortement certaines frac.tions du magma. plex was emplaced in the of a magrna Certaines intercroissances min6rales du banc Me- from which the different rock units were formed rensky semblent Ctre des reliques de cristaux d'un by fractional crystallization and sedimentation alliage fer-platine qui, par r6action avec le soufre, of the crystals into cumulates (Willemse 1969, auraient donn6 une intercroissance d'un nouvel Lombaard 7934, and many others). More than alliage fer-platine avec des sulfures de m6taux one heave of magrna is postulated, but some vils, pseudomorphe des cristaux originels. workers (Coertze 1970, Schwellnus 1956) think (Traduit par la R€daction) that most different rock units are separate in- 469 470 THE CANADIAN MINERALOGIST trusions of magmas that underwent differentia- tion at depth. Usuolthickness. The origin of the mafic magma is seen as the result of a process of differential melting of the upper mantle, but there was $ome degree of assimilation of the rocks that came into contact with the magma. In some instances, this assimilation could have played an impor- Frqn < I metre tant role in tle formation of sulfide deposits to >' 7 metres. (Coertze 1970, Liebenberg 1970, de Waal 1977, Naldrett & Cabri L976). The upper mantle also must be consideredto be the source platinum-group of the elements in mafic rocks. Fran 0 to obout 3 metres. CorvcsNTRAuoN AND Drposrrrox There are two concepts of concentration and deposition that are of interest in the present Ftc. 1. Idealized section through the Merensky paper: (1) crystal settling (Liebenberg 1970, Reef in the western Bushveld complex. Brynard et al, 1976), which suggeststhat plati- nlrm-group minerals settled out of a magma to become concentrated at the bottorn of the magma chamber, and (2) concentration in a col- from 25 meire lectot e,9., sulfide which settled out of the mag- Bose metol to O5 metre. Porphyritica ma, thus also concentrating the platinum-group sulfides Oi From0.35 metre elements. The rates of settling in both these cnd plotinum $ ryroxeniteT I to 0.5metre. instancesmust have been such that the platinum- groupminerol group minerals or sulfide droplets settled at a From030 metre rate different from that of the other crysta{s to l.52metre. that formed in the magma at the same time; no concentration would have been effected by the settling process alone. The conditions under whish particles or droplets can settle out From060 metre of a magma have been discussedby Wadsworth to 9.0 metres. (L973) among others. The grain sizes and densities measured by the author and his col- leagues, or collected from the literature, are given in Table 1, and these were used, following Wadsworth (1973), for the calculation of the settling velocities for various mineral consti- Fro. 2. Idealized section through the Merensky tuents (Table 2). Reef in the eastern Bushveld. As only relative settling speeds are of in- terest here, tfie sarne density (2.58 g.cm-s) and the same viscosity (1000 P) were used in all further factor to be taken into account is that the calculations, although the estimates for the crystails could have increased in size during these values given in the literature range from settling and after deposition. As will be shorrn lower to higher values. Stokes' for'mula then later, it is probable that many of the grains of simplifies to V : I7169.4D (p" - 2.58) m per platinum-group minerals attained their final year, where D is the particle diameter (cm) and sizesafter having been deposited. pr is the particle density (g.cm"). The size data reported in Table 1 were Strictly, the densities should have been cor- collected by different workers and are usual'ly rected for temperature, but in view of the larger the sizes as seen in polished sections without influence exerted on settling velocity by grain any allowance for sectioning. The averages sizes, it was decided that this correction would were calculated from a few tens to a few not have influenced the results undulv. A hundreds of grains and are usually close to the PLATINI'M DEPOSITS IN THE BUSHVELD COMPLBX 471 T^8[r l. GRArt{.STZES 0F SotECoI{STITUEI{TS til nE i€nSrSKYnEEF TlgL[ 2. SFtTUile lELoClTlEsmR S0r qilEnALSFnoil IllE I'ERt[Sfr IEEF tf,llslTl 0F liAOlA. 2.58 s,@-' VISOSIIY3 lo00 P Rock mtt/localtty Range Average,/Slze Yclslty. Hlneral lbnslry Slzs. @ q A.ttt ot Pl tanesDgtg 'l.6rD p$ yolr Chrwlio T@ lqyer 0.1-1.6m ChroElte EottoE layer l,0m 2'| (e.9. In p€g@told) 52gI (0.9. ln Fgstold) t3 u0 Pyrorene Porphyrlti c 2-3m 1.5@ 0.3 (s.9. ln porphyrltlcpyrcEnlto) I 189 pyrorenl ts 0.1 (e.9. In por?hyrltlcpyrcEnlte) 132 Pt-Fo 6llay Pegmtold 16 !m Chrgdlto 4.5 0.3 2 967 0.2 I 319 !.)quu oT rr tanesErg 0.1 w Orrorlte Top layor 0.3. 2.5m 1.0 m 0.05 v, Braggdte 3.5-47 !E 12.5 |lm Bnggltg 8,9 0.03 (u. val6 lablo l) 98 Laurl tE 5.5-27 un 13.25um 0.0012(dv.Atot ud S. of Pllmberg) 0.16 Sperrylits 6 -50 un 20.0 !n C6p€r{ta 10.2 0.01 (Mr. valE Tabl€ l) l3 0.0@8(!Y. fru rear R6tenburg) 0.08 Spedyllte 10.59 0.015 (Er, 6l@ Tlble l) 3l Brdgglte sl ll6te' rrca A 9,hn 0.0m (av. S. of Pll@3berg) Bragglte &rodt€, arca A 6. tllm Pt-F€ rIloy 15 0.0115(Er. valE Tlble l) 28.20 Coopsrlte Slllcats' arca A I 1.lum 0.0016([E of Pl lmesb€rg] 0.54 Cooperlt6 Chr@lte' arca A 8. lum 5Utn@ Co@erlte ChroElts' ar€a B 4.9un droplets 4.5 0.05 e. 0,C03 (UL'z @dlil) 0.s Laurl te Sllicater area.A 6'9um lm Ftrl 7.8 0,03 v Laurlte Sillcate, ar€a I 7.4um Laurite Chroullo, arca A 4.9!n Laur'lte B - chro0lte, ar€a 6.lum minerals from a u.tesEm Ptaunm mne (2, The settling of platinum-group Sperylite 6-150|ln crystallizing magma, on the assumptionthat they Braggilte 37-300!n in the sizes in whioh they are en- L.