| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Journal of Petrology | Volume 43 | Number 3 | Pages 535-556 | 2002
© Oxford University Press 2002
The Skaergaard Layered Series. Part VI. Excluded Trace Elements
DEPARTMENT OF GEOLOGICAL SCIENCES, UNIVERSITY OF OREGON, EUGENE, OR 97403, USA
Received July 22, 1999; Revised typescript accepted September 21, 2001
 |
ABSTRACT |
|---|
TOP
ABSTRACT
INTRODUCTIONIn contrast to the smooth trends of major elements and mineral compositions, the excluded trace elements in the Skaergaard Layered Series have an irregular distribution that does not conform to the normal trends of Rayleigh-type fractionation. Their concentrations are about constant or even decline through the Lower and Middle Zones before increasing sharply to reach maximum concentrations 100–200 m above the Sandwich Horizon. As in the case of included elements, the relative concentrations of excluded elements in coexisting phases deviate widely from those predicted by experimentally determined partition coefficients under presumed magmatic conditions. This is seen most clearly in the immiscible melanogranophyres and conjugate ferrogabbros. Although the major elements conform to the experimentally determined relations for immiscible liquids, the trace elements do not; they follow a totally independent trend. The abrupt increase in the concentrations of excluded elements in the upper part of the intrusion could plausibly be attributed to an addition of new magma or to a density inversion that resulted in upward migration of a late liquid or fluid, but these possibilities are inconsistent with the compositional and spatial relations of the upper parts of the intrusion. Although a late residual liquid certainly migrated upward, the most likely explanation for the observed distribution of excluded elements is that the partition coefficients were altered by volatile components, which gradually increased during the early stages of crystallization then began to exsolve near the top of the Middle Zone.
KEY WORDS: igneous differentiation; Skaergaard intrusion
|
INTRODUCTION |
|---|
Part V of this series (McBirney, 1998a
) dealt with the distribution of included trace elements in the rocks and principal minerals of the Skaergaard Layered Series. My purpose here is to examine the corresponding excluded elements and their relations to progressive differentiation of the Skaergaard magma. The principal conclusion that will emerge is that the trends of enrichment of these elements differ markedly from those of the major elements, possibly because their distribution was affected by a build-up and eventual exsolution of volatile components in the magma. It will also be shown that the observed partition coefficients differ widely from those predicted from experimental studies.
Several earlier studies addressed various aspects of the Skaergaard trace elements. Haskin and his co-workers (Haskin & Haskin, 1968; Paster et al., 1974) were the first to attempt to reconcile the distribution of excluded trace elements with earlier interpretations of trace-element behavior during crystallization and differentiation of the Layered Series. Other previous work has been summarized by Wager & Brown (1968) and more recently by McBirney (1996, 1998a).
The structural form of the intrusion and its principal units, illustrated in simplified form in Fig. 1, have already been described at length (McBirney, 1989, 1995, 1996; Boudreau & McBirney, 1997; McBirney & Nicolas, 1997; Sonnenthal & McBirney, 1998) and need not be repeated. Apart from minor differences noted in the Appendix, the distribution and selection of samples, as well as the methods by which they have been analyzed, are essentially the same as those described in Part V of this series (McBirney, 1998a).
|
, UBSß, and UBS
, correspond mineralogically to the three main divisions of the Layered Series, the Lower, Middle, and Upper Zones. The Upper Border and Layered Series converge on the Sandwich Horizon (SH). The Marginal Border Series (MBS) is also divided into units (LZa*, LZb*, etc.) corresponding to those of the Layered Series, but no equivalent to Upper Zone C has been found. Granophyre is an important interstitial component of all rocks above the horizon marked GR. A dashed line labeled SHs is the level at which excluded elements reach their maximum concentration. UBST is the Upper Border Series equivalent of the hidden part of the Layered Series.
Petrologic relations
Earlier studies (McBirney, 1995, 1998a) led to the conclusion that the compositional relations of the principal units of the intrusion are consistent with the scheme illustrated in Fig. 2. As the magma crystallized on the steep walls to form the Marginal Border Series, the remaining liquid was enriched in iron and, being denser than the main magma, flowed down the walls to pond on the floor. Similar conditions prevailed as the Upper Border Series crystallized at the roof; part of the dense, evolved liquid, as well as some of the heavy mafic crystals it contained, drained away from the zone of crystallization. Much of this material entered directly into the main mass of magma, but part of the crystalline fraction was deposited on the floor close to the walls. Thus, most of the Layered Series crystallized from liquids that had already evolved to some extent by crystallization at the walls and roof. The chief mechanism of differentiation of the Layered Series is thought to have been compaction augmented by compositionally driven convection (McBirney, 1995), but so long as the differentiated liquid was denser than the main magma, the Layered Series must have been largely isolated from the overlying magma (Sonnenthal, work in progress).
|
By this reasoning, only the rocks of the Marginal Border Series preserve bulk-rock compositions inherited directly from the central body of magma. It is regrettable (but unavoidable) that only a single complete section through these rocks is represented by the samples used here. Because they crystallized from a magma that was evolving as it descended along the wall, the rocks of any given unit within the sequence vary vertically with structural height (Hoover, 1989). The section used here is from the uppermost level where the magma began its descent and is therefore most representative of the composition of the main body. This is also the most complete section and has the added advantage that the basaltic wall rocks at this level contributed less contamination than the underlying Archean basement.
Although one of the final products of differentiation was a granophyric liquid rich in silica and alkalis, the magma did not evolve to this composition until a very late stage, probably after the fronts of crystallization had met at the Sandwich Horizon. There can be no question that the main trend of differentiation was one of prolonged iron enrichment. The essential field and petrologic evidence for this interpretation (Brooks & Nielsen, 1990; McBirney & Naslund, 1990; Morse, 1990) has recently been substantiated by studies of the trace-element compositions of plagioclase (Tegner, 1997; McBirney, 1998b). The iron content of this mineral increases continuously from the base of the Layered Series to the uppermost levels of Upper Zone C. The change to a silica-rich, iron-poor composition occurred only in the last remaining interstitial liquids.
The product of this final stage of differentiation is represented by patches of interstitial granophyre found in most of the upper part of the intrusion. A dotted line in Fig. 1 defines the lowest level at which this granophyre is a conspicuous interstitial phase. It is important to note that it does not follow a single horizon but crosses the boundaries of lithologic units. In the central part of the intrusion it divides the Upper Zone B into two parts, UZb-1 below the appearance of abundant interstitial granophyre and UZb-2 above, but toward the margins it descends well into the Middle Zone.
When the magma reached an advanced stage of differentiation it separated into two immiscible liquids, one rich in iron and a smaller fraction rich in silica (McBirney & Nakamura, 1973; McBirney, 1975). The latter, referred to by Wager & Brown (1968) as melanogranophyre, differs from the felsic granophyres of dikes or from the interstitial phase that permeates much of the upper part of the intrusion. It takes the distinctive form of pods and schlieren that make up 
15% of the volume of Upper Zone C. The interpretation of the melanogranophyres as a second liquid is consistent with other evidence that the Skaergaard magma continued to follow a course of iron enrichment, even after abundant Fe–Ti oxides began to crystallize.
|
ANALYZED ELEMENTS |
|---|
Using the methods described in the Appendix, about 20 excluded elements can be measured with enough precision to define their bulk-rock abundances throughout the entire sequence of differentiation. Although a few of these elements may have partition coefficients >1·0 for certain phases, their bulk-rock values are <1·0 for the intrusion as a whole (Table 1). Compositions of the immiscible pairs and mafic–felsic segregations are given in Table 2. The concentrations of most rare earth elements (REE), as well as Sr, Ba, and Zr, have been determined in one or more of the principal minerals by ion probe (Table 3).
|
|
|
|
The analyzed elements can be divided into three major groups, each with distinctive geochemical properties.
Large ion lithophile elements (LILE; Ba, Rb, and Sr)
Ba and Rb have small partition coefficients for all the principal minerals. Although Sr enters plagioclase, Ca-rich pyroxene, and apatite, its bulk partition coefficient is <1·0 at all levels (Table 4). As pointed out in Part V (McBirney, 1998a), Sr is of special interest, because it is one of the few trace elements other than REE that enters both plagioclase and pyroxene.
|
High field strength elements (HFSE; Zr, Hf, Ta, Nb, Y, and Th)
None of these elements is totally excluded from all essential minerals of the intrusion. Zr, Hf, Ta, and Nb have partition coefficients >1·0 for ilmenite, and Y and Th are strongly partitioned into apatite. In the discussion that follows I have selected Nb, Y, and Hf as representative of this group, mainly because they are well determined and have the greatest variation for the entire series and least scatter for a single unit.
Rare earth elements
The REE are most strongly concentrated in apatite but are also taken into calcium-rich pyroxene and, to a lesser extent, plagioclase (Table 5). Nd is well suited to serve as representative of the REE, because it is normally one of the most abundant and, therefore, the best determined members of the series. Even when apatite crystallizes in abundance, the bulk distribution coefficients for the REE are substantially less than one.
|
|
BULK-ROCK VARIATIONS |
|---|
The trends of enrichment of elements representing these three groups (Ba, Rb, Nb, Y, Hf, and Nd) are illustrated in Fig. 3. Average bulk-rock abundances are plotted according to the structural height of units of the Layered Series or in equivalent parts of the two Border Series. The spatial distribution of these same elements is shown in contoured profiles for the intrusion as a whole (Fig. 4).
|
|
The most striking aspect of these elements is that their average bulk-rock abundances vary only slightly through the Lower and Middle Zones and do not increase to any notable degree until the Upper Zone began to crystallize. The average concentrations of several excluded elements, such as Nd, actually decline upward through the Lower and Middle Zones with the result that, after 60–70% crystallization, they are less abundant than they were in the earliest rocks of the Layered Series! This upward decline through the lower part of the Layered Series has been explained as the result of an increasing efficiency of compaction as the rate of cooling declined and the zone of crystallization became thicker (McBirney, 1995).
All excluded elements increase exponentially above the Middle Zone and reach maximum concentrations 100–200 m above the Sandwich Horizon (Fig. 5a). The rates of enrichment are almost identical for all excluded elements except P and Ti, both of which change from excluded to included elements above Middle Zone. This is best seen in a multi-element diagram for the rocks of the Upper Zone (Fig. 6a). As soon as apatite appears at the base of Upper Zone B, the abundances of P begin to decline. The reversal of Ti comes somewhat earlier where ilmenite becomes an important primary phase.
|
|
|
The rates of increase are similar for the REE (Fig. 6b). The most pronounced increase is between UZa and UZb when the modal abundance of apatite increases sharply. The most significant feature, however, is the lack of any abrupt change with the appearance of abundant granophyre between UZb-1 and UZb-2. As will be shown in greater detail in a later section, the concentrations of REE in the granophyric component are very small, and the earlier interpretation of the abrupt increase of excluded elements as a result of upward migration of a late silica-rich liquid (McBirney, 1995) cannot be correct. Similarly, the lack of any marked change of slope in the REE patterns seems to rule out the possibility that these elements were carried in a volatile-rich fluid.
The greater average bulk-rock abundances of excluded elements in rocks that crystallized under the roof (Fig. 3) was previously attributed to contamination with xenoliths of felsic gneiss that floated to the upper levels of the magma. This interpretation is now ruled out by recent work showing that the isotopic character of strontium in these rocks does not differ significantly from that of the Layered Series (A. R. McBirney & R. A. Creaser, in preparation).
It is evident from Fig. 4 that the values used to construct Fig. 3 are averages with a large degree of variance. They fail to express the wide variations within individual units, which, in some instances, exceed those of the intrusion as a whole. The magnitude of these variations is shown by the relative standard deviations tabulated in Table 6.
|
Correlations
A large set of analyses of this kind should be amenable to statistical analyses designed to clarify their petrologic relations. To this end, the data for 135 rocks have been examined statistically using their raw bulk-rock abundances. I have calculated correlation coefficients for pairs of elements according to the method described in a preceding paper dealing with included elements (McBirney, 1998a). The results shown in Table 7 provide a crude measure of how the behaviors of the elements in each group resemble one another. The correlations between like elements are not as clear as one might expect. For example, two LILE, Ba and Rb, have a better correlation with the HFSE, Zr and Hf, than they have with one another. Relations such as these could reflect the strong dependence of bulk distribution coefficients on the modal assemblages. Y and Nd show this especially well. Although geochemically different, both are strongly concentrated in apatite.
|
|
RESIDENCE OF EXCLUDED ELEMENTS |
|---|
Strongly excluded elements are normally assumed to reside in accessory phases and in the rims of large, zoned crystals. If one could measure the precise concentrations of these elements in all the minerals, it would be a simple matter to compare the total quantity contained in the mineral assemblage with the bulk-rock abundance. This approach has proved useful for some of the included elements, particularly Ni (McBirney, 1998a
). Several key elements can be measured with reasonable accuracy by ion microprobe, but the total concentration in a strongly zoned mineral is difficult to determine. Many excluded elements tend to be more abundant in the rims of plagioclase and pyroxene than in the cores, but the relative values differ widely according to the identity of the adjacent grain. REE, for example, have smaller concentrations in plagioclase rims next to augite than they do next to oxides or other crystals of plagioclase, no doubt because the larger partition coefficients for pyroxene give that mineral an advantage in competing for these elements in a thin layer of liquid separating the two grains.
Bearing these limitations in mind, one can obtain a reasonable balance using the REE in two samples in which their concentrations are unusually large (Table 8). One sample (UB-313) is below and the other (UB-323) above the level at which interstitial granophyre becomes abundant. The difference between the bulk-rock abundances and the sum of mineral contents represents the amount of an element that cannot be accounted for in the minerals. In sample UB-323 any excess would logically be in accessory or occult phases of the granophyre, which is an abundant interstitial component of this rock. For the REE, however, the calculated difference between the bulk-rock and sum of mineral contents is small, and for some elements it is even slightly negative. This is true of both rocks. It is surprising that the balance is this good, because large proportions of the REE are contained in apatite and a small error in estimating the modal abundance of that mineral has a large effect on the total bulk-rock abundance.
Attempts to apply this test to other elements give mixed results, mainly because of the limitations of the analytical methods. In the case of the REE, however, the results in Table 8 show that essentially all of these elements reside in the principal minerals and that the interstitial phases contain no significant amounts of the excluded components.
|
RELATIONS TO LATE-STAGE LIQUIDS |
|---|
The concentrations of strongly excluded elements in gabbros of differentiated intrusions have long been taken as a measure of the amount of ‘trapped liquid’ remaining after crystallization of the primary minerals (e.g. Henderson, 1970
; Barnes, 1986; Chalokwu & Grant, 1990). By this interpretation, the absolute concentrations of a given element would be a function of the amount of this liquid, and the proportions of elements with very small partition coefficients should preserve a nearly constant ratio inherited from the parental magma. Barnes (1986) showed that the major-element compositions of minerals that have grown from a trapped liquid would be shifted to more evolved compositions, and Cawthorn (1996) showed that this effect is even more pronounced for excluded trace elements. In both cases, the compositions of minerals should be shifted by an amount that is an inverse function of the modal proportions of the mineral into which an element is partitioned.
A number of elements have been used to estimate the amounts of these ‘trapped liquids’. In the case of the Skaergaard rocks, Henderson (1970) selected phosphorus, which enters a minor phase, apatite, that does not come on the liquidus until late in the course of crystallization. Grant & Chalokwu (1998) used TiO2, but the utility of such an element is doubtful for the simple reason that it enters mafic minerals in substantial amounts and will vary with the abundance of those minerals regardless of the amount of ‘trapped liquid’.
Earlier interpretations assumed that the residual liquid responsible for a compositional shift was trapped in the interstices between primary phases, but the same effect could result during compaction when liquid expelled from deeper levels flows through the interstices and equilibrates with minerals that crystallized initially from a less evolved liquid. This is a reasonable explanation for the varied concentrations of strongly excluded trace elements illustrated in Figs 3 and 4; the residual liquid was largely expelled from the lower part of the intrusion and was concentrated at higher levels where it enriched the primary minerals in excluded elements.
Ideally, one could calculate the trace-element composition of such a liquid by using experimentally determined partition coefficients to invert the concentrations in minerals. When one attempts to do this, however, it is immediately apparent that the partitioning between phases does not conform to the patterns expected for magmatic conditions, at least so far as they can be inferred from laboratory measurements.
|
PARTITIONING BETWEEN COEXISTING PHASES |
|---|
Partitioning between minerals
One of the most surprising results to emerge from the study of included elements examined in Part V was the wide divergence between the predicted and observed partitioning of these elements between coexisting phases. This is seen most clearly in the distribution of Ni between mafic minerals, but it is also conspicuous in the partitioning of Sr between plagioclase, pyroxene, and apatite (McBirney, 1998a
). Of the excluded elements considered here, only the REE and Sr enter both plagioclase and Ca-rich pyroxene in sufficient amounts to permit a comparison of their relative concentrations. Taking Nd as typical of the former, the ratios of its concentrations in pyroxene to those in plagioclase are at least an order of magnitude greater than would be predicted by the experimentally determined partition coefficients in Table 5 (Fig. 7a).
Partitioning between pyroxene and plagioclase is a function of several factors, the most important of which are the composition and temperature of the liquid. At lower temperatures greater proportions of the REE enter pyroxene. An even greater difference might be expected if the compositions of the liquids differed in some critical way from those used to obtain the values in Table 5. Hence the discrepancy seen in Fig. 7a could be attributed to equilibration with a late-stage liquid or fluid at temperatures below normal magmatic conditions.
In the case of Sr, the partitioning between plagioclase and pyroxene changes abruptly above the Middle Zone (Fig. 7b). In the lower part of the Layered Series, the observed relative concentrations in pyroxene are much less than the experimentally determined partition coefficients would indicate, but above the Middle Zone the observed ratio of concentrations in plagioclase to those in pyroxene decreases sharply, even though the experimentally determined partition coefficients would predict no change. The tendency for the observed ratio to be greater in the rims than in cores of crystals suggests that at least part of the difference could be due to lower temperatures and a more evolved, late-stage liquid. It is worth noting that the abrupt change in the partitioning of Sr at the Middle Zone is not seen in Nd. This difference may reflect the much greater mobility of Sr in hydrous fluids.
Partitioning between two immiscible liquids
As noted earlier, the uppermost parts of the Layered Series contain numerous small segregations of melanogranophyres formed from an immiscible second liquid that separated from the main iron-rich magma. The distribution of elements between two immiscible silicate liquids has been studied in considerable detail, first by Watson (1976) and more recently by Vicenzi et al. (1994, 1995) and Shearer et al. (2001). When the measured partition coefficients are compared with the observed distribution of major elements between a Skaergaard melanogranophyre and its host ferrogabbro the correlation is fairly good (Fig. 8a), but the excluded trace elements show no correlation whatever. The REE (Fig. 8b) have a remarkably regular pattern of distribution in which their partitioning into the mafic phase increases with increasing atomic number. (The only element that does not fit this pattern is europium, which, owing to its different valence, it more strongly partitioned into the mafic phase.)
|
Again, the observation that major elements are distributed in the predicted way but excluded trace elements are not implies that the abundances of the latter were established under conditions that differed from those used in laboratory measurements. These conditions are not reflected in the major elements, which are governed by phase equilibria and have concentrations beyond the limits imposed by Henry’s Law. A comparison of the observed pattern with those measured experimentally at different temperatures shows that the former could result from equilibration at lower temperatures (Fig. 9a). The temperature at which the melanogranophyre crystallized, about 1000°C, was well below that of the experiments (1160° and 1240°C).
|
Partitioning between anorthosites and pyroxenites
Pods and schlieren of anorthosite and closely associated olivine pyroxenite are common in Lower Zone A and in certain parts of the Marginal Border Series. Although they share several chemical features with the melanogranophyre–ferrogabbro association described above, they are thought to be products of late-stage metasomatic replacement of the original gabbro (McBirney & Sonnenthal, 1990). Their isotopic compositions appear to have been affected by the same processes that were responsible for some of the trace-element relations (A.R. McBirney & R.A. Creaser, in preparation).
The evidence against immiscibility is primarily in the major-element composition of this part of the Layered Series where the gabbros have not yet reached the iron-rich compositions of the two-liquid field. Nevertheless, it is possible that the strong polarization of mafic and felsic components that produced the anorthosite and pyroxenite could be driven by a physical-chemical relationship akin to that of immiscibility. Unfortunately, excluded trace elements offer no evidence in this regard. Their partitioning bears no resemblance to the pattern predicted for immiscible liquids, and the pattern of REE abundances is distinctly different from that in the melanogranophyre–ferrogabbro association (Fig. 8b). There is no sign that the heavier elements are more strongly partitioned into the mafic rocks.
|
ENRICHMENT AND DIFFERENTIATION |
|---|
Abrupt changes such as the increased rate of enrichment above the Middle Zone (Fig. 5a) are conventionally attributed to one or more of four possible causes:
- an introduction of new magma;
- decreased crystal–liquid fractionation;
- redistribution in a late-stage liquid or fluid; or
- changes of partition coefficients.
Each of these possibilities can be considered in terms of the field and compositional relations described in the preceding sections.
Introduction of new magma
Although most studies of the Skaergaard intrusion have accepted the long-standing premise that the body was formed from a more or less continuous injection of homogeneous magma, this assumption is difficult to reconcile with volcanic eruptions of basaltic magma that normally discharge small volumes of varied compositions in repeated eruptions over periods of decades or centuries. The isotopic variations found in the Skaergaard rocks have shown that the magma was indeed inhomogeneous (Stewart & DePaolo, 1990; A.R. McBirney & R.A. Creaser, in preparation). Moreover, several unusual features suggest that an injection of new magma did in fact take place when the Layered Series had crystallized as far as the Middle Zone. One of these is an abrupt decrease of oxygen fugacity (Morse et al., 1980); another is the precipitation of sulfides (Bird et al., 1991). These changes occurred about the same time as the sudden increase in the rate of enrichment of the included elements that concerns us here.
A dike of 4 m width in the lower part of the Layered Series has the structural relations one would expect if it had introduced new magma at this stage (Fig. 10). Its north-trending orientation is characteristic of dikes that were approximately contemporaneous with the Skaergaard intrusion (Nielsen, 1978). Where it transects the base and Lower Zone A it is relatively fine grained, but its texture becomes coarser upward until it merges with the gabbros of the Middle Zone. Because it cannot be traced higher in the section, one could reasonably postulate that the dike entered the Skaergaard magma at that level. Compared with the chilled margin, its composition was exceptionally primitive (Table 9), and addition of such a liquid should have had a conspicuous effect on the main magma. Because no significant change is seen in the abundance of included elements (McBirney, 1998a) even though the dike is much richer in Ni, Cr, and other transition elements than the Layered Series (Table 9), one would conclude that the volume of injected magma, if any, was not great enough to have a notable effect on the composition of the main magma. Thus, on the basis of included elements alone, the evidence for a substantial influx through the dike is at best neutral.
|
|
Comparing the absolute abundances of excluded elements in the dike with those of the Layered Series offers no clear evidence, because, as seen in Table 6, the latter differ widely, even at a single horizon. Nevertheless, the dike is unusually rich in excluded elements, such as Ba and Zr, and introduction of substantial amounts of such a liquid should have had a notable effect on the enrichment of these elements. Ba does in fact increase in the Upper Border Series at this stage (Fig. 3a) but less so in the Layered Series. If, as seems likely, the density of incoming new magma was less than that of the iron-rich magma of the main body (about 2·73 vs 2·84), it would have risen to the roof where it would contribute its signature to the Upper Border Series, and the Layered Series would show less effect. This is precisely what is seen.
A further test is provided by the isotopic ratios of Sr and Nd (A.R. McBirney & R.A. Creaser, in preparation). The initial Sr isotopic ratio of the dike (0·70641) is substantially greater than that of the Skaergaard gabbros (0·7044 at this stage), and yet we find no detectable change in the latter that can be correlated with an addition of magma at the time the Middle Zone was crystallizing. The increased Sr isotopic ratio that should appear in the Upper Border Series is not seen. As noted earlier, the Sr in the Upper Border Series is no more radiogenic than that of the Layered Series (A.R. McBirney & R.A. Creaser, in preparation).
One cannot, of course, rule out the possibility that a gradual addition of magma was thoroughly mixed with the original one or that new magma was introduced from another unseen source, but in view of the conflicting evidence it is difficult to attribute the changes that occurred after the Middle Zone to a major injection of new magma.
Changing effectiveness of crystal–liquid fractionation
A second possible explanation could be that the mechanism of crystal–liquid fractionation was altered in such a way that the excluded elements were not expelled but retained in all the units above the Middle Zone. If we have interpreted the density relations in the Layered Series correctly, it seems more likely that the conditions would have the opposite effect. As long as the evolving liquid was becoming denser than the overlying magma, the efficiency of compaction on the floor would steadily decline, but at some point the density relations would be reversed. Once the evolving liquid passed the point of maximum iron enrichment it would become less dense than the overlying magma and increasingly likely to be expelled. This would be especially true if the interstitial liquid were enriched in volatile components.
The parallel trends of differentiation in the Layered Series and two Border Series require that any change in the effectiveness of fractionation affected all three units in the same manner despite their differing orientations with respect to gravity. Even if one could conceive of some combination of mechanisms that could achieve this, it would be fortuitous if the effects were simultaneous.
Redistribution in a late-stage liquid or fluid
Any attempt to explain the larger concentrations of excluded elements in the Upper Zone as a result of extensive redistribution in a late-stage liquid or fluid entails two basic questions. First, what was the nature of the transporting medium, and, second, how did it produce the observed spatial distribution of excluded elements?
The observation that excluded elements reach their maximum concentrations one or two hundred meters above the Sandwich Horizon leaves little room for doubt that some sort of late interstitial liquid continued to rise after the two fronts of crystallization had converged (McBirney, 1995). An obvious candidate for this liquid is the interstitial granophyre that permeates the same rocks. In detail, however, the concordance is far from perfect. In the central part of the intrusion the rapid enrichment begins near the base of the Upper Zone, whereas abundant granophyre does not appear until two or three hundred meters higher. Moreover, we have already noted in the inventory of trace elements summarized in Table 8 that the interstitial granophyre accounts for little if any of the bulk-rock abundances of REE, and gabbros with abundant granophyre are not notably richer in these elements than nearby rocks lacking this interstitial component.
Another possible transporting agent could be a hydrous fluid exsolved from the evolving magma. The beginning of enrichment of excluded elements corresponds closely to a similar increase in the F/Cl ratio of apatites (Fig. 5a and b). This remarkable correlation suggests that the two phenomena could be linked in some way to exsolution of a vapor phase. Because Cl is preferentially partitioned into a hydrous vapor while F remains concentrated in the silicate melt (Candela, 1986; Boudreau & McCallum, 1989), the F/Cl ratio would increase after the melt became saturated with volatiles and exsolved a hydrous phase that escaped and left behind a Cl-depleted melt. The steady upward increase of the F/Cl ratio seen in Fig. 5b indicates that this was not a single event but a continuing process concurrent with differentiation.
Meurer & Boudreau (1998) found a zone in the Stillwater Complex where the changing halogen ratio of apatites is exactly the opposite of that in the Skaergaard, that is to say, chlorine has a marked increase relative to fluorine. They attributed the increase of chlorine to introduction of a volatile fluid below a horizon of low permeability. It is interesting to note, however, that they found no increase in the concentrations of excluded trace elements at this horizon. In contrast, the increasing F/Cl ratio in the upper levels of the Skaergaard intrusion is more probably the result of a loss rather than a gain of volatiles. These relations suggest that the increased concentrations of excluded elements that accompanied the exsolution of volatiles may reflect the effect of the loss of volatiles on partition coefficients.
Changing partition coefficients
It was apparent even in the earliest studies (Haskin & Haskin, 1968; Paster et al., 1974) that the distribution of many trace elements deviates widely from the behavior predicted by normally accepted partition coefficients. Several possible explanations have been considered but the most plausible cause is the changing composition of the evolving liquid. The effect is illustrated by the partitioning of Sr between coexisting pyroxene and plagioclase (Figs 5c and 7b). From the base up to the level where the F/Cl ratio begins to increase, the observed partitioning diverges steadily from the ratio predicted by partition coefficients measured in volatile-free experiments. It then reverses and the proportion of Sr going into plagioclase increases relative to the proportion that would be predicted from the ratio of partition coefficients. In other words, the proportion of Sr going into plagioclase returns toward values closer to the ratios of measured partition coefficients. If this interpretation is correct, volatile components, such as water, chlorine, and possibly sulfur, may have caused proportionately more Sr to go into plagioclase than into pyroxene. When these components began to exsolve and were lost from the liquid, this trend was reversed and took a new course toward more volatile-free compositions. The observed change of fO2 could have been an additional factor, particularly in the case of elements such as Eu and V that are present in two or more valence states.
Of the various possible explanations for the distribution of excluded elements, only a change of partition coefficients can account for the coincidence of these changes with what appears to have been an exsolution of volatile components. Steinberg (1960) showed many years ago that chlorine has a marked effect on partitioning of trace elements between melts and aqueous fluids, and Mahood (1981) has pointed out that changing concentrations and relative proportions of complexing ligands, such as H2O, Cl, F, and SO2, affect polymerization and the availability of cation sites in the melt. This explanation has been invoked to explain the changing rates of enrichment of excluded elements in several well-studied siliceous plutons and ignimbrites (Mahood & Hildreth, 1983; Bau, 1996). Beyond this, however, it would be pointless to speculate further until we have more experimental studies of the effects of volatile components on the behavior of trace elements.
|
SUMMARY AND CONCLUSIONS |
|---|
The variations of excluded trace elements in the Skaergaard Layered Series do not follow those of the major elements. The latter have regular trends consistent with crystal fractionation under magmatic conditions, but the excluded trace elements are distributed in a manner that is difficult to explain by Rayleigh fractionation using currently available values for partition coefficients.
This is seen most clearly in the distribution of elements between the immiscible melanogranophyres and ferrogabbro of Upper Zone C. Partitioning of the major elements between the conjugate phases corresponds closely to the pattern predicted by experimentally determined phase equilibria, but the excluded trace elements show no such relation; they are distinctly decoupled from elements that are present in concentrations exceeding the limits of Henry’s Law.
No clear evidence can be found to indicate that the abrupt increase in the rate of enrichment of excluded elements near the base of the Upper Zone resulted from an injection of new magma or upward migration of a granophyric liquid. The increase coincides with a similar increase of the F/Cl ratio that is interpreted as the result of exsolution of a vapor phase. The changing composition of the liquid, particularly the complexing volatile components, may have caused a corresponding change of the partition coefficients, as it is said to have done in several large bodies of siliceous magma. Experimental studies are needed to clarify the exact nature of this effect.
|
APPENDIX: ANALYTICAL METHODS |
|---|
The same sets of samples as used in Part V (McBirney, 1998a
) are the basis for the present study. From the primary set (Set I) of about 600 rocks analyzed by atomic absorption, a subset of about 150 rocks (Set II) has been analyzed by X-ray fluorescence (XRF) and neutron activation, and, of these, about 40 rocks (Set III) have been selected for analysis of their constituent minerals by ion-microprobe (Table 3). Representative samples from this third set were also analyzed for the complete set of REE by inductively coupled plasma source mass spectrometry (ICP-MS). Sets II and III contain additional samples that were not included in the previous study but illustrate special features, such as immiscible segregations and pegmatites. Sample locations and statistical data have been given in Part I of this series (McBirney, 1989, 1996). Bulk-rock abundances for the rocks of set II include:
- Ba, Co, Cr, Cu, Li, and Zn measured by atomic absorption spectroscopy at the University of Oregon;
- Ga, Ni, Nb, Pb, Rb, Sr, V, Y, and Zr by XRF in the laboratories of Professor J. M. Rhodes at the University of Massachusetts and Peter Hooper at Washington State University;
- REE, as well as Sc, Cs, Hf, Ta, Th, and U obtained by neutron activation at Oregon State University, and by ICP-MS in the laboratory of Peter Hooper at Washington State University.
Microprobe analyses of minerals in the main dataset were carried out at the Geophysical Laboratory of the Carnegie Institution of Washington, but all those in the selected subset II were obtained at the University of Oregon by means of a Cameca instrument. The ion-microprobe analyses were performed in the laboratory of Dr N. Shimizu at Woods Hole Oceanographic Institution according to established procedures [Shimizu & LeRoex (1986) and references therein] using a primary beam of O- with a diameter of 40 µm. Secondary intensities were measured for 139La, 140Ce, 146Nd, 147Sm, 151Eu, 163Dy, 166Er, and 174Yb, and molecular ion interferences were eliminated using an energy offset of 60 eV. Ion intensities were normalized to 30Si and quantified by reference to calibration curves determined from appropriate standards.
The ion-microprobe yields satisfactory analyses of most REE (La, Ce, Nd, Sm, Eu, Dy, Er, and Yb) in plagioclase, pyroxene, and apatite, but this list does not coincide in all respects with that for the whole-rock concentrations. Tb and Lu are included only with the latter, whereas Dy and Er are limited to the former.
|
ACKNOWLEDGEMENTS |
|---|
I have benefited greatly from helpful reviews and stimulating discussions with many friends and colleagues including Francis Albarède, Derek Bostok, Henry Dick, Gordon Goles, Peter Keleman, Edmund Mathez, Roger Nielsen, Eric Sonnenthal, and James Webster. Ion-probe analyses performed at Woods Hole Oceanographic Institution in the Northeast National Ion Microprobe Facility were supported by grants EAR-9305508 and EAR-9628749. The staff of the Radiation Center at Oregon State University provided many neutron-activation analyses. And finally, I am pleased to acknowledge the generous support of the National Science Foundation that has enabled me to carry on my thirty-year study of the world’s most instructive layered intrusion.
|
FOOTNOTES |
|---|
*Telephone: (541) 344-2539. Fax: (541) 346-4692. E-mail: McBirney{at}darkwing.uoregon.edu
|
REFERENCES |
|---|
Barnes, S. J. (1986). The effect of trapped liquid crystallization on cumulus mineral compositions in layered intrusions. Contributions to Mineralogy and Petrology 93, 524–531.[Web of Science]
Bau, M. (1996). Controls on the fractionation of iso-valent trace elements in magmatic and aqueous systems: evidence from Y/Ho, Zr/Hf, and lanthanide tetrad effect. Contributions to Mineralogy and Petrology 123, 323–333.[Web of Science]
Bird, D. K., Brooks, C. K., Gannicott, R. A. & Turner, P. A. (1991). A gold-bearing horizon in the Skaergaard Intrusion, East Greenland. Economic Geology 86, 1083–1092.
Boudreau, A. E. & McBirney, A. R. (1997). The Skaergaard Layered Series. Part III. Non-dynamic layering. Journal of Petrology 38, 1003–1020.
Boudreau, A. E. & McCallum, I. S. (1989). Investigations of the Stillwater Complex. Part V. Apatites as indicators of evolving fluid composition. Contributions to Mineralogy and Petrology 103, 138–153.
Brooks, C. K. & Nielsen, T. F. D. (1990). A discussion of Hunter and Sparks (Contrib Mineral Petrol 95: 451–461). Contributions to Mineralogy and Petrology 104, 244–247.
Candela, P. A. (1986). Toward a thermodynamic model for the halogens in magmatic systems: an application to melt–vapor–apatite equilibria. Chemical Geology 57, 289–301.
Cawthorn, R. G. (1996). Models for incompatible trace-element abundances in cumulus minerals and their application to plagioclase and pyroxenes in the Bushveld Complex. Contributions to Mineralogy and Petrology 123, 109–115.
Chalokwu, C. I. & Grant, N. K. (1990). Reequilibration of olivine with trapped liquid in the Duluth Complex, Minnesota. Geology 15, 71–74.
Grant, N. K. & Chalokwu, C. I. (1998). The cumulate paradigm affirmed. Journal of Geology 106, 641–644.
Haskin, L. A. & Haskin, M. A. (1969). Rare-earth elements in the Skaergaard intrusion. Geochimica et Cosmochimica Acta 32, 433–447.
Henderson, P. (1970). The significance of the mesostasis of basic layered igneous rocks. Journal of Petrology 11, 463–473.
Hoover, J. D. (1989). Petrology of the Marginal Border Series of the Skaergaard Intrusion. Journal of Petrology 30, 399–439.
Mahood, G. A. (1981). Chemical evolution of a Pleistocene rhyolitic center: Sierra La Primavera, Jalisco, Mexico. Contributions to Mineralogy and Petrology 77, 129–149.
Mahood, G. A. & Hildreth, W. (1983). Large partition coefficients for trace elements in high-silica rhyolites. Geochimica et Cosmochimica Acta 47, 11–30.[Web of Science]
McBirney, A. R. (1975). Differentiation of the Skaergaard Intrusion. Nature 253, 691–694.
McBirney, A. R. (1989). The Skaergaard Layered Series: I. Structure and average compositions. Journal of Petrology 30, 363–397.
McBirney, A. R. (1995). Mechanisms of differentiation of layered intrusions: evidence from the Skaergaard Intrusion. Journal of the Geological Society, London 152, 421–435.
McBirney, A. R. (1996). The Skaergaard Intrusion. In: Cawthorn, R. G. (ed.) Layered Intrusions. Amsterdam: Elsevier, pp. 147–180.
McBirney, A. R. (1998a). The Skaergaard Layered Series. Part V. Included trace elements. Journal of Petrology 39, 255–276.
McBirney, A. R. (1998b). Iron in plagioclase as a monitor of the differentiation of the Skaergaard intrusion: a discussion of Tegner (Contrib Mineral Petrol 128: 45–51). Contributions to Mineralogy and Petrology 132, 103–105.
McBirney, A. R. & Nakamura, Y. (1973). Immiscibility in late-stage magmas of the Skaergaard Intrusion. Carnegie Institution of Washington Geophysical Laboratory Report 348–352.
McBirney, A. R. & Naslund, H. R. (1990). The differentiation of the Skaergaard Intrusion: a discussion of Hunter and Sparks (Contrib Mineral Petrol 95: 451–461). Contributions to Mineralogy and Petrology 104, 235–240.
McBirney, A. R. & Nicolas, A. (1997). The Skaergaard Layered Series. Part II. Magmatic flow and dynamic layering. Journal of Petrology 38, 569–580.
McBirney, A. R. & Sonnenthal, E. L. (1990). Metasomatic replacement in the Skaergaard Intrusion, East Greenland: preliminary observations. Chemical Geology 88, 245–260.
Meurer, W. P. & Boudreau, A. E. (1998). Compaction of igneous cumulates Part II: Compaction and the development of igneous foliations. Journal of Geology 106, 293–304.
Morse, S. A. (1990). A discussion of Hunter and Sparks (Contrib Mineral Petrol 95: 451–461). Contributions to Mineralogy and Petrology 104, 240–244.
Morse, S. A., Lindsley, D. H. & Williams, R. J. (1980). Concerning intensive parameters in the Skaergaard Intrusion. American Journal of Science 280-A, 159–170.
Nielsen, R. L. (1992). BIGD.FOR: a fortran program to calculate trace-element partition coefficients for natural mafic and intermediate composition magmas. Computers in Geosciences 18, 773–788.
Nielsen, T. F. D. (1978). The Tertiary dike swarms of the Kangerdlugssuaq area, East Greenland. An example of magmatic development during continental break-up. Contributions to Mineralogy and Petrology 67, 63–78.
Paster, T. P., Schauwecker, D. S. & Haskin, L. A. (1974). The behavior of some trace elements during solidification of the Skaergaard layered series. Geochimica et Cosmochimica Acta 38, 1549–1577.
Shearer, C. K., Papike, J. J. & Spilde, M. N. (2001). Trace-element partitioning between immiscible lunar melts: an example from naturally occurring lunar melt inclusions. American Mineralogist 86, 238–246.
Shimizu, N. & LeRoex, A. P. (1986). The chemical zoning of augite phenocrysts in alkaline basalts from Gough Island, South Atlantic. Journal of Volcanology and Geothermal Research 29, 159–188.
Sonnenthal, E. L. & McBirney, A. R. (1998). The Skaergaard Layered Series. Part IV. Reaction–transport simulations of foundered blocks. Journal of Petrology 39, 633–661.
Steinberg, E. P. (1960). The radiochemistry of zirconium and hafnium. National Academy of Science–National Research Council, Nuclear Science Series, 52 pp.
Stewart, B. W. & DePaolo, D. J. (1990). Isotopic studies of processes in mafic magma chambers: II. The Skaergaard Intrusion, East Greenland. Contributions to Mineralogy and Petrology 104, 125–141.
Tegner, C. (1997). Iron in plagioclase as a monitor of the differentiation of the Skaergaard Intrusion. Contributions to Mineralogy and Petrology 128, 45–51.
Vicenzi, E. T., Green, T. H. & Soey, S. (1994). Effect of oxygen fugacity on trace-element partitioning between immiscible silicate melts at atmospheric pressure: a proton and electron microprobe study. Chemical Geology 117, 355–360.
Vicenzi, E. P., Green, T. H. & Sie, S. H. (1995). Immiscible silicate liquids at high pressure: the influence of melt structure on elemental partitioning. Nuclear Instruments and Methods in Physics Research B 104, 470–475.
Watson, E. B. (1976). Two-liquid partition coefficients. Experimental data and geochemical implications. Contributions to Mineralogy and Petrology 56, 119–134.[Web of Science]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  C. Tegner, P. Thy, M. B. Holness, J. K. Jakobsen, and C. E. Lesher Differentiation and Compaction in the Skaergaard Intrusion J. Petrology, May 1, 2009; 50(5): 813 - 840. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. J. Hanley, J. E. Mungall, T. Pettke, E. T. C. Spooner, and C. J. Bray Fluid and Halide Melt Inclusions of Magmatic Origin in the Ultramafic and Lower Banded Series, Stillwater Complex, Montana, USA J. Petrology, June 1, 2008; 49(6): 1133 - 1160. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Whitney, J. H. Tepper, M. M. Hirschmann, and H. A. Hurlow Late orogenic mafic magmatism in the North Cascades, Washington: Petrology and tectonic setting of the Skymo layered intrusion Geological Society of America Bulletin, May 1, 2008; 120(5-6): 531 - 542. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. B. Holness, C. Tegner, T. F. D. Nielsen, G. Stripp, and S. A. Morse A Textural Record of Solidification and Cooling in the Skaergaard Intrusion, East Greenland J. Petrology, December 1, 2007; 48(12): 2359 - 2377. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. BOORMAN, A. BOUDREAU, and F. J. KRUGER The Lower Zone-Critical Zone Transition of the Bushveld Complex: a Quantitative Textural Study J. Petrology, June 1, 2004; 45(6): 1209 - 1235. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. F. D. NIELSEN The Shape and Volume of the Skaergaard Intrusion, Greenland: Implications for Mass Balance and Bulk Composition J. Petrology, March 1, 2004; 45(3): 507 - 530. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. McBIRNEY and R. A. CREASER The Skaergaard Layered Series, Part VII: Sr and Nd Isotopes J. Petrology, April 1, 2003; 44(4): 757 - 771. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||