sodic, potassic, acid and silica metasomatism.
»
Q-Si/3-(K*Na*2Ca/3)
F«K-(Na+Ca) in millications
Quartz
00
Albite
++ o ++
JL—xMica
111
ii -200 -100 100 200Microcline
ll.Il!...di... ■ 300 400A
▼ w
A ▼ ▼o
♦ biotite granitex
peralkaline graniteo albitised biotite granite
+ albitised peralkaline granite -$■ albitite
O microclinised biotite granite
® altered basement
This highlights changes in silica and alkali element ratios without regard to how the cations are distributed throughout the sample. Not only can salic components be displayed but mica analyses can be plotted into the diagram to define the greisenisation process between quartz-rich greisens and mica-rich greisens. The data presented on Figs 4.3 & 4.4, have been chosen from complexes throughout Nigeria. By calculating the millicationic components in terms of the parameters:
Q = l/3Si-(K+Na+2/3Ca) and F = (K-(Na+Ca))
for any granitic rock suites which have undergone varying degrees of subsolidus reactions, the degree of disturbance from granite minimum compositions can be clearly demonstrated. Furthermore, the dominant process of sodic, potash or acid metasomatism can be defined.
Samples have been selected which are unequivocal in interpretation in order to determine major element trends most clearly. The majority of samples however come from the Ririwai Complex and are discussed in Kinnaird et al (1985a) and Bowden and Kinnaird (1984a,b and c).
The granites that show little subsolidus modification cluster within a fairly small area of the diagram. The peralkaline granites shown as crosses on Fig 4.3, plot slightly nearer the F axis than the biotite granites shown as closed circles.
Sodic metasomatism facies have compositions which trend towards the albite pole. The peralkaline granites, shown as a plus sign, are arfvedsonite albite granites with pyrochlore. The biotite granites of the L13 core in Ririwai (Fig 3.1), show the greatest effect within the biotite granite facies, although the zinnwaldite and lepidolite albite granites of the Afu Complex also display this trend strongly. Most of the open circles on Fig 4.3, are sections of the L13 drill-core taken at different depths with varied textural and mineralogical facies. Samples L13-411 and L13-440 which are albitites, shown as a circle within a cross on Fig 4.3, are the most albitised facies of the biotite granite spectrum (two samples from the Afu Complex have negative Q values in addition to high negative F values).
Potash metasomatised facies have compositions which define a trend towards the microcline pole and these are represented by open diamonds in Fig 4.3. Samples selected to represent potash metasomatism have come from the reddened wallrock of the Ririwai lode, from microclinites and from microcline pegmatites within alteration zones of the Saiya Shokobo Complex. Potash metasomatised granite gneissic basement samples from the margin of the Gindi Akwati ring dyke in the Ropp Complex, show a distinctive but different trend. Compositions shown by a cross within a circle, plot in a trend from the granite field, towards the F axis, reflecting the nature of the basement. The bulk chemistry of the basement in Nigeria is variable, but the selected sanples come from modified calc-alkaline compositions. Thus, because of the chemical parameters Q and F, any enrichment in Ca will displace compositions towards the base of Fig 4.3. The value of this multicationic diagram becomes apparent since other diagrams do not separate these altered basement rocks into a separate field.
During acid metasomatism in contrast, compositions plot towards the Q pole
(Fig 4.3). However, there are several distinct trends towards the Q pole since
the geochemical characteristics of acid metasomatism in Nigeria depends upon the intensity and retention of earlier stages of sodic and potash metasomatism, the bulk chemistry of the host rock, the composition pH, Eh, fS, fF, fCO?, fO? and the petrological characteristics and extent of greisen development; Since acid metasomatism may affect country rock, volcanics, perthite granite and metasomatised variants the chemical signature differs for each set of variables. The trends have been designated (1), (2), (3), and (4) in Fig,SA 4. Trend 1 represents compositional trends during acid metasomatism of basement country rocks; trend (2) represents compositional variations during acid metasomatism developed in biotite granites close to the volcanic cover, such compositional variations reflect the acid metasomatism of albite-rich rocks; trend (3) represents compositional variations during acid metasomatism of microcline rich rocks and trend (4) represents compositional variations during pervasive acid metasomatism of previously unaltered perthite granites together with those granites that had already been affected by both sodic and potassic metasomatism.
Trend 1- Acid metasomatism of the basement has been studied in the Gindi Akwati area on the western perimeter of the Ropp Complex (Bowden and Kinnaird 1984b) and in the Dutsen Rishi area of the Saiya Shokobo Complex where acid metasomatism has also affected the volcanic cover. Quartz-poor and quartz-rich types of vein-controlled mineralisation are developed by fluid interaction on basement rocks. Much depends on the distance from the contact. Close to the contact quartz-poor veins rich in fluorite, sphalerite, chalcopyrite and other sulphides, are developed. These quartz-poor veins reflect the bulk chemical composition of the basement and its calcium-enrichment which displaces compositions towards the base of Fig 4.4. The trend of such samples (trend 1), is shown by a cross within a circle. Veins which penetrate well out into the basement are wolframite-bearing quartz veins. Obviously such quartz-enriched samples, regarded as monomineralic quartz greisens by Aubert (1969), will plot towards the Q pole along trend 4.
Trend 2- The apical region of a biotite granite may have undergone a process of sodic metasomatism already described. Thus acid metasomatism of such rocks will reflect this early sodic imprint. Similarly the volcanic cover to most biotite granites consists of rhyolitic ignimbrites some of which are mildly peralkaline. Acid metasomatism of such rocks will have been dominated by host rock chemical controls. Thus any greisens developed against or within the volcanic cover may show the sodic effect - with increasing silica effects as acid metasomatism gave way to silica metasomatism. This trend (trend 2) is well developed in the Tibchi rocks. Thus acid metasomatism of albite-rich roof rocks, which is usually a pervasive effect, plots parallel to the albite-quartz join in the qiartz-albite-mica-microcline tetrahedra. This trend reflects the high initial alkali element enrichment of the altered facies.
Trend 3 -Vein-controlled acid metasomatism produced a greisen composed of quartz +mica+topaz+fluorite. Micaceous greisens and monomineralic mica rocks, developed from the breakdown of microclinite in the roof zones during acid metasomatism of microcline-rich facies (trend 3 on Fig 4.4). Such mica-rich greisens plot on the quartz-mica join of the tetrahedra. Silica was released during microcline breakdown and compositions trend towards the quartz pole as silica metasomatism is superimposed on acid metasomatism. This trend has been
Fig 4.4-. Summary Q-F diagram of the major trends of acid metasomatism and ore deposition for the Nigerian anorogenic
province. Trends 1, 2, 3, and 4 are described in the text.
F
Trend 4 -The final trend is the pervasive acid metasomatism of perthitic granites that have not been affected by earlier metasomatism. This trend is characterised by thin marginal greisen veins from the Kudaru, Fagam, Jos Bukuru and Banke Complexes. Trend 4 is also the compositional trend for facies, such as that of the Uwar Gida area of the Ririwai Complex, where acid metasomatism had been superimposed both on remnant sodic and potassic metasomatism close to the volcanic cover. In this case the trend - as would be expected - bisects trends 2 and 3.
In conclusion therefore, it is clear that chemical data obtained from hydrothermally altered rocks, when plotted on the QF multicationic diagram, generate definite zones which can be correlated with known structural environments and known sequences of alteration. The QF diagram therefore, since it defines a series of trends for samples that have undergone acid metasomatism, is both a more useful and more valid method for the interpretation of hydrothermal effects than the Q-Ab-Or diagram.
Minor and trace element variations during hydrothermal alteration
Although it has been assumed in earlier literature that certain trace elements, especially the rare earths, are immobile under postmagmatic condition, recent research has shown that many trace elements have considerably different crystal-fluid partition coefficients compared with crystal-liquid values. Trace element variations therefore together with petrographic studies illustrate the geochemical modifications that have taken place in the subsolidus.
The chondrite-normalised plots of trace elements are grouped ranging from compatible to incompatible and distribution coefficients show distinctive trends during the alteration processes (Fig 4.5). Rock types like the quartz porphyries, fayalite granite and syenite show element distribution patterns that are governed by crystal-liquid partition coefficients. In contrast, all the chondrite-normalised plots show element distribution patterns that are governed by crystal-fluid distribution coefficients.
The granitoids display flat of HREE-enriched trends on chondrite-normalised diagrams with attendant strong Eu depletion. During hydrothermal alteration there was clearly REE mobility although the pattern of depletion or enrichment varies from that described for S.W. England (Alderton, Pearce and Potts 1980).
During sodic metasomatism there was a substantial enrichment in all REE's, particularly in the peralkaline facies (Bowden et al 1979). In the Nigerian province it was the albitised granites that have the highest lithium and uranium enrichment. In the peralkaline facies there was an increase in all trace elements including the high field strength elements Nb, Zr and Hf (Fig 4.6). In the biotite granites there was substantial enrichment in some of the LIL elements, an enrichment in Zn and particularly Sn and sometimes a depletion in Zr and Y.
During potash metasomatism, there was a depletion of the whole rare earth spectrum and compared with all other processes the pot ash-met asomatised wallrocks on Fig 4.6e, shown as open diamonds are significantly the most depleted. This was combined with a similar decrease in most of the trace element
Fig 4.5. Chondrite-normalised trace element patterns arranged to show fluid-rock element partitioning. Elements to the left of each diagram have inferred bulk K/s partitioned to the rock
while elements to the right of each figure have k/s partitioned
to the fluid phase.
© j O> j w © 5 r / ^ \ / ■ ....—... ...-...«--- // 0 © © c <0 e© y . O) © CO © / < / >» w S \ X tu JS 5 s' J ■ <0 - * ... l siijpuoqo / >|oou R b T b N b L a Ce H f Zr Y R b T h N b La C e H I Zr Y R b T h N b L a Ce H I Zr Y R b T h N b L a Ce HI Zr Y R b T h N b L a C e H I Zr PAGE 78
Fig 4.6. Chondrite-normalised rare earth element patterns for Ririwai rocks
(a) volcanic rocks: (b) peralkaline granite: (c) arfvedsonite albite granite: (d) +biotite granite: (e) H /K+ metasomatised rocks, H =triangles, K =diamonds: (f) "parental" syenite from Kila Warji (Bowden and van Breemen 1972, Bowden and Whitley 1974, Kinnaird et al 1984a and b).
chondrite-normalised curves potash metasomatised wallrocks of the Ririwai lode
show the most REE depletion (Kinnaird et al 1985a).
During acid metasomatism, the chondrite normalised plot of Fig 4.5, shows that there was a depletion in Zr relative to biotite granite and an increase in some of the trace element populations compared with potash metasomatism particularly in Th, Ce and Y. However, trace element patterns in greisens a^e very variable, since they are affected by earlier processes. Thus where H+ metasomatism was superimposed on an earlier sodic effect, the chondrite normalised pattern (Fig 4.5g), shows a close similarity to that for sodic metasomatism and where superimposed on a potash effect (Fig 4.5h), the potash pattern is reflected. Where H metasomatism was superimposed on an unaltered biotite perthite granite or basement, the chondrite normalised trace element patterns (Fig 4.5j), show a depletion most notably in Nb in the basement. The rare earths patterns (Fig 4.6), show that there was partitioning of the LREE to the greisen mineral assemblage, a slight enrichment of Eu coupled with increasing Yb and Lu with the interpolated HREE spectra of Dy, Ho,Er and Tm, partitioned slightly to the fluid. There was a marked increase of Li, which was accommodated in the new mica of the greisen and there was a substantial increase in the ore elements notably Sn,Pb,Zn,W,Cu and Fe - the latter element occurring within both the micas and the ore minerals.
No data exists for the rare earth behaviour during the silicification process but it is assumed that there was an overall decrease in REE's. The chondrite normalised plot (Fig 4.5), shows that whilst there was an overall decrease in the spectrum of many trace elements compared with biotite granite, there was a selective dilution, most notably in zr. There was an increase in the ore elements during silicification particularly in Sn,Zn,W,Bi,Cu,Mo and Pb sometimes in appreciable quantities.
Minor and trace element indicators of hydrothermal alteration
A number of trace element patterns can be used to examine the ftprocesses of
hydrothermal alteration. Several element pairs - K/Rb, Rb/Sr, y/Sr/ybSr, and Th/U have already been considered by previous workers in addition to a discussion of rare earth element behaviour. In addition to these, the Ba/Rb and the Zr/Hf ratio can be considered, as can the concentrations of Li and F and the F:C1 and Ba:Rb. A study of such trace element assemblages shows that certain element pairs clearly indicate that fluid reactions have taken place even before any significant mineralogical changes can be detected.
The elements K, Rb, Sr and Ba
The element pairs K-Rb, K-Ba and Ca-Sr respectively have similar ionic properties. Therefore Rb and Ba are usually incorporated in potash-bearing minerals such as K-feldspar and biotite,and Sr is usually found in Ca-bearing minerals such as plagioclase. Since the CaO content of all Nigerian anorogenic granitoids is lower than the average for granitic rocks, strontium values will obviously also be lower. This is the case not only for fresh rocks but also the altered variants. More surprisingly however in view of the enriched potash levels in many of these granites, all the Nigerian granitoids are also low in Ba when compared with other low Ca-bearing granites.
During fractional crystallisation or partial melting, Rb was concentrated in the liquid phase. Such characteristics can be explained by the large ion size of Rb and its electronegativity which helps to define the differences in bond energy between K and 0 than between Rb and 0. Thus K was preferentially removed whilst Rb tends to concentrate in the residual fluids. During late magmatic processes, Sr - which has a slightly greater ionic radius than Ca - was depleted relative to Ca in late stage processes. Thus, the geochemical behaviour of Rb and Sr show a greater sensitivity to late magmatic processes than either K or Ca. Thus Rb, Sr and Ba concentrations are initially controlled by processes of fractional crystallisation with Sr and Ba decreasing and Rb increasing in residual fluids with increasing crystallisation.
Therefore, the K/Rb, Rb/Sr, and Ba/Rb ratios are good indicators of the hydrothermal alteration processes:-
K/Rb
Rubidium was enriched in many Nigerian anorogenic granites but the amount of potash does not vary greatly unless there has been widespread potash metasomatism. Thus the K/Rb ratio or a plot of K against Rb reflects the degree of rubidium enrichment in the granitic rocks. The range in K:Rb ratio varies from complex to complex. There is a distinct negative correlation between K and Rb for all the Nigerian ring complex granites (Bowden and Kinnaird 1984c). Early explanations of this phenomena sought to explain the variations by extreme differentiation of a parental fayalite granite magma although the major element compositions do not vary widely even between the peralkaline, peraluminous and metaluminous variants. More recently, Bowden and van Breemen (1972) and Bowden and Turner (1974) suggested that Nigerian granites containing high Rb up to 1500ppm and low K/Rb ratios down to 35, had undergone substantial recrystallisation in the subsolidus during albitisation. Such a process may be recognised by a negative correlation of K against Rb (Bowden and Kinnaird 1984c). Thus the trend of K/Rb ratios to low values in Nigerian granites is believed to represent the degree of post magmatic adjustment in response to mineralising albite-rich fluids (Bowden and Kinnaird 1984c).
Rb/Sr
The Rb content of the anorogenic granites has been described by various authors but Bowden 1961, and Butler et al 1962 provided the first comprehensive study. Values for hornblende-bearing granites and porphyries are in the range 120-31Oppm with peralkaline granites in the range 190-43Oppm and biotite granites in the range 180-860ppm. A comparison of Rb-Sr ratios shows that for hornblende bearing granites and porphyries the average ratio of Rb to Sr is 200, for peralkaline granites the average ratio is 300 and for biotite granites the average ratio is 400:1. Rb/Sr ratios are high for the biotite granites of Banke and Ririwai compared to the mineralised granites of the Afu complex (Imeokparia 1981) The ratios for the altered facies of the Ririwai complex are even higher. Such high values recorded for the hydrothermally altered rocks are due to an increase in Rb associated with microclinisation and greisenisation. The ratio for unaltered Ririwai biotite granite is 268, for potash metasomatised rocks is 851 and for greisenised variants is 674 whilst even the sodic metasomatised albite arfvedsonite granite has a ratio of 310.
Rubidium-strontium isotopic studies on Ririwai rocks that have been carried out in Britain (van Breemen et al 1975) and France (Bonin et al 1979) suggests that the Rb/Sr system remained open until the circulation of mineralising hydrothermal fluids had ceased. Bonin et al (1979), suggest that the proportion of Rb-Sr in anorogenic alkali feldspars was controlled by fractional crystallisation of alkali feldspars in the magma chamber. After the emplacement of the granites, hydrothermal fluids mobilised Rb relative to Sr and therefore modified the relative proportions of Rb-Sr without any notable modification of the ° Sr/DDSr ratio.
A consideration of the Rb-Sr systematics therefore could give a guide to the type of hydrothermal process that might have taken place.
Strontium ratios
87 86
The wide variation in initial Sr/ Sr for the Nigerian anorogenic granites has been discussed by Bowden et al (1976) and more recently by Vidal et al (1979). The values range from 0.705 to 0.752 even from rock types from the same ring complex. In the past, this wide variation of initial strontium isotopic ratio has been used as an argument to suggest that there was a significant