Download Isotopic Composition of Nitrogen and Oxygen in Rocks and Minerals: An Overview and more Study notes Geochemistry in PDF only on Docsity!
Chapter 9
erupted beneath a kilometer of more of wa- ter probably retain most of their dissolved water. However, basalts, particularly submarine basalts, are far more readily contaminated with hydrogen (i.e., with wa- ter) than with carbon. Furthermore, the ef- fect on hydrogen isotopic composition de- pends on the mode of contamination, as Figure 9.09 indicates. Direct addition of water or hydrothermal reactions will raise δD (because there is little fractionation dur- ing these processes), while low tempera- ture weathering and hydration will lower δD, because hydrogen, rather than deute- rium, is preferentially incorporated into al- teration phases. Loss of H 2 and CH 4 , which may partition into a CO 2 gas phase when it forms, could also affect the hydrogen iso- topic composition of basalts. However, the available evidence suggests that these spe- cies constitute only a small fraction of the hydrogen in basalts, so this effect is likely to be minor. As Figure 9.10 indicates, MORB have mean δD (^) SMOW of about –67.5‰ and a stan- dard deviation of ±14‰. How much of this variability reflects the processes shown in Figure 9.09 and how much reflects true heterogeneity in the mantle is unclear. Kyser (1986) has argued that mantle hy- drogen is homogeneous with δD (^) SMOW of –80‰. The generally heavier isotopic composition of MORB, he argues, reflects H 2 O loss and other processes. However, Poreda, et al., (1986) found that δD in ba- salts from the Reykjanes Ridge south of Iceland correlated significantly with La/Sm and other trace element ratios, suggesting at least some of the isotopic variation of hydrogen in basalts reflects real variations in the mantle. Submarine basalts from Kilauea’s East Rift Zone have higher δD than MORB. Kyser and O’Neil (1984) argued that these higher values result from the addition of water to the magma in the rift zone. Hawaiian submarine basalts analyzed by Garcia et al. (1989) have δD very similar to MORB. Hydrous minerals in xenoliths also provide a sample of mantle hydrogen. As Figure 9.10 shows, phlogopites have δD that is generally similar to that of MORB, though some lighter values also occur. Amphiboles have much more variable δD and have heavier hydrogen on average. Part of this differ- ence probably reflects equilibrium fractionation. The fractionation between water and phlogopite is close to 0‰ in the temperature range 800°-1000°C, whereas the fractionation between water and am- phibole is about –15‰. However, equilibrium fractionation alone cannot explain either the variability of amphiboles or the deference between the mean δD of phlogopites and amphiboles. Complex pro- cesses involving in amphibole formation that might include Rayleigh distillation may be involved in the formation of mantle amphiboles. This would be consistent with the more variable water content of amphiboles compared to phlogopites.
Nitrogen
Figure 9.11 summarizes the existing data on the nitrogen ratios in the crust and mantle. There is far less data than for other stable isotope ratios because of the low concentrations and pervasive con-
Figure 9 .10. δD in MORB and in mantle phlogopites and amphiboles. The MORB and phlogopite data sug- gest the mantle has δDSMOW of about – 60 to – 90.
Chapter 9
tamination problems. The solubility of N 2 in basalts is very limited, though much of the nitrogen may be present as NH 4
, which is somewhat more soluble. Hence of volcanic rocks, once again only submarine basalts provide useful sam- ples of mantle N. There are both con- tamination and analytical problems with determining nitrogen in basalts, which, combined with its low abundance (gen- erally less than a ppm), mean that accu- rate measurements are difficult to make. Measurements of δ^15 N (^) ATM in MORB range from about –2 to +12‰. The few avail- able analyses of Hawaiian basalts range up to +20. At present, it is very difficult to decide to what degree this variation reflects contamination (particularly by organic matter), fractionation during de- gassing, or real mantle heterogeneity. Perhaps all that can be said is that nitro- gen in basalts appears to have positive δ^15 N on average. Diamonds can contain up to 2000 ppm of N and hence provide an excellent sample of mantle N. As can be seen in Figure 21.11, high δ^13 C diamonds (most common, and usually of peridotitic par- agenesis) have δ^15 N that range from -12 to +5 and average about -3‰, which contrasts with the gener- ally positive values observed in basalts. Low δ^13 C diamonds have generally positive δ^15 N. Since or- ganic matter and ammonia in crustal rocks generally have positive δ^15 N, this characteristic is consistent with the hypothesis that this group of diamonds are derived from subducted crustal material. How- ever, since basalts appear to have generally positive δ^15 N, other interpretations are also possible. Fi- brous diamonds, whose growth may be directly related to the kimberlite eruptions that carry them to the surface (Boyd et al., 1994), have more uniform δ^15 N, with a mean of about -5‰. Since there can be significant isotopic fractionations involved in the incorporation of nitrogen into diamond, the meaning of the diamond data is also uncertain, and the question of the nitrogen isotopic composition of the man- tle remains an open one.
Sulfur
There are also relatively few sulfur isotope measurements on basalts, in part because sulfur is lost during degassing, except for those basalts erupted deeper than 1 km below sealevel. In the mantle, sul- fur is probably predominantly or exclusively in the form of sulfide, but in basalts, which tend to be somewhat more oxidized, some of it may be present as SO 2 or sulfate. Equilibrium fractionation should lead to SO 2 being a few per mil lighter than sulfate. If H 2 S is lost during degassing the re- maining sulfur would become heavier; if SO 2 or SO 4 2 − is lost, the remaining sulfur would become lighter. Total sulfur in MORB has δ^34 S (^) CDT in the range of +1.3 to –1‰, with most values in the range 0 to +1‰. Sakai et al. (1984) found that sulfate in MORB, which constitutes 10-20% of total sulfur, was 3.
Figure 9 .11. Isotopic composition of nitrogen in rocks and minerals of the crust and mantle. Modified from Boyd and Pillinger (1994).
Chapter 9 Spring 2011
general effect of fractional crystallization on magma is generally to increase δ^18 O slightly, generally not more than a few per mil. On the other hand, water-rock interaction at low to moderate temperatures at the surface of the Earth, produces much larger O isotopic changes, generally enriching the rock in 18 O. Consequently, crustal rocks show significant variability in δ^18 O. As we have seen, the range of δ^18 OSMOW in the fresh, young basalts and other mantle materials is about +4.5 to +7. Because this range is narrow, and the range of those crustal materials is much greater, O iso- tope ratios are a sensitive indicator of crustal assimilation. Isotope ratios outside this range suggest, but do not necessarily prove, the magmas have assimilated crust (or that post-eruptional isotopic exchange has occurred).
Oxygen Isotope Changes during Crystallization
The variation in O isotope composition produced by crystallization of magma will depend on the manner in which crystallization proceeds. The simplest, and most unlikely, case is equilibrium crys- tallization. In this situation, the crystallizing minerals remain in isotopic equilibrium with the melt un- til crystallization is complete. At any stage during crystallization, the isotopic composition of a mineral and the melt will be related by the fractionation factor, α. Upon complete crystallization, the rock will have precisely the same isotopic composition as the melt initially had. At any time during the crystalli- zation, the isotope ratio in the remaining melt will be related to the original isotope ratio as:
R
R 0
f + α( 1 − f ) ;
Rs
R
where Rl is the ratio in the liquid, Rs is the isotope ratio of the solid, R 0 is the isotope ratio of the original magma, ƒ is the fraction of melt remaining. This equation is readily derived from mass balance, the definition of α, and the assumption that the O concentration in the magma is equal to that in the crys- tals; an assumption valid to about 10%. Since we generally do not work with absolute ratios of stable isotopes, it is more convenient to express 9.04, in terms of δ:
Δ = δ melt − δ 0 ≅
⎥ ×^1000 9.
where δmelt is the value of the magma after a fraction ƒ-1 has crystallized and δo is the value of the original magma. For silicates, α is not likely to be much less than 0.998 (i.e., ∆ = δ^18 Omelt – δ^18 Oxtals ≥ 2). For α = 0.999, even after 99% crystallization, the isotope ratio in the remaining melt will change by only 1 per mil. Fractional crystallization is a process analogous to Rayleigh distillation. Indeed, it is governed by the same equation ( 8 .69), which we can rewrite as:
α− 1
The key to the operation of either of these processes is that the product of the reaction (vapor in the case of distillation, crystals in the case of
Figure 9.13. δ^18 O as a function of SiO 2 in a tholeiitic suite from the Galapagos Spreading Center (GSC) (Muehlenbachs and Byerly, 1982) and an alkaline suite from Ascension Island (Sheppard and Harris, 1985). Dashed line shows model calcu- lation for the Ascension suite.
Chapter 9 Spring 2011
crystallization) is only instantaneously in equilibrium with the original phase. Once it is produced, it is removed from further opportunity to attain equilibrium with original phase. This process is more effi- cient at producing isotopic variations in igneous rocks, but its effect remains limited because α is gener- ally not greatly different from 1. Figure 9.12 shows calculated changes in the isotopic composition of melt undergoing fractional crystallization for various values of ∆ (≈ 1000(α–1)). In reality, ∆ will change during crystallization because of (1) changes in temperature (2) changes in the minerals crystallizing, and (3) changes in the liquid composition. The changes will generally mean that the effective ∆ will in- crease as crystallization proceeds. We would expect the greatest isotopic fractionation in melts crys- tallizing non-silicates such as magnetite and melts crystallizing at low temperature, such as rhyolites, and the least fractionation for melts crystallizing at highest temperature, such as basalts. Figure 9.13 shows observed δ^18 O as a function of temperature in two suites: one from a propagating rift on the Galapagos Spreading Center, the other from the island of Ascension. There is a net change in δ^18 O between the most and least differentiated rocks in the Galapagos of about 1.3‰; the change in the Ascension suite is only about 0.5‰. These, and other suites, indicate the effective ∆ is generally small, on the order of 0.1 to 0.3‰. We can generalize the temperature dependence of stable isotopes by saying that at low temperature (ambient temperatures at the surface of the Earth up to the temperature of hydrothermal systems, 300- 400° C), stable isotope ratios are changed by chemical processes. The amount of change can be used as an indication the nature of the process involved, and, under equilibrium conditions, of the temperature at which the process occurred. At high temperatures (temperatures of the interior of the Earth or mag- matic temperatures), stable isotope ratios are only minimally affected by chemical processes; they can therefore be used as tracers much as radiogenic isotope ratios are. These generalizations lead to a final axiom: igneous rocks whose oxygen isotopic compositions show signifi- cant variations from the primordial value (~6) must either have been affected by low temperature processes, or must contain a component that was at one time at the surface of the earth (Taylor and Sheppard, 1986). Rocks that have equilibrated with water at the surface of the Earth, e.g., sediments, tend to have δ^18 O values significantly higher than +6 (generally greater than +10). Waters on the surface of the Earth typically have δ^18 O significantly lower than +6 (0 for seawater, generally lower for meteoric, or fresh, waters). Interestingly, the δD values of sedimentary rocks and mantle-derived igneous rocks are rather similar (- 50 to - 85). This may be coincidental since the δD of sediments are controlled by frac- tionation between minerals and water, whereas δD of igneous rocks reflects the isotopic composition of mantle hydrogen (mantle water). One the other hand, it is possible that it is not coincidental. Plate tectonics results in the return of water- being rocks from the surface of the Earth to the mantle (i.e., via subduction). It may well be that after 4.5 billion years, the subduction process essentially con- trols the isotopic composition of H in the upper mantle.
Combined Fractional
Crystallization and Assimilation
Because oxygen isotope ratios of man-
Figure 9.14. Variation in δ^18 O of a magma undergoing AFC vs. amount crystallized. Initial δ^18 O of the magma is +5.7. After Taylor (1980).
Chapter 9 Spring 2011
at nearly the same concentration, making calculation of the mass assimilated fairly straightforward. Also, it is easier to uniquely characterize the assimilant using both radiogenic and stable isotope ratios, as suggested in Figure 9.15. The equation governing a radiogenic isotope ratio in a magma during AFC is different from 9.07 be- cause we cannot assume the concentration of the element is the same in all the components. On the other hand, there is no fractionation between crystals and melt. The general equation describing the variation of the concentration of an element in a magma during AFC is:
Cm
Cm^0
= f − z^ +
R
R − 1
Ca
zCm^0
( 1 − f z^ ) 9.
where Cm and the concentration of the element in the magma, Co^ is the original magma concentration, ƒ are as defined for equation 9.07 above, and z is as:
z =
R + D − 1
R − 1
where D is the solid/liquid partition coefficient The isotopic composition of the magma is the given by (DePaolo, 1981):
ε m =
R
R − 1
Ca
z
( 1 − f − z^ ) ε a + Cm^0 f − z ε^0
R
R − 1
Ca
z
( 1 − f
− z
) + Cm
0
f
− z
where ε is the isotope ratio with subscripts m , a and 0 denoting the magma, the assimilant, and the original magma respectively and the other parameters are defined as in equation 9.08. Fig- ure 9.16 shows calculated AFC curves on a plot of δ^18 O vs. 87 Sr/^86 Sr. Note that except in the case where β = 1, where the problem simplifies to one of simple mixing, the mixing lines end at ƒ = 0. (99% crystallized). Figure 9.1 7 shows an actual case, the Adamello Massif in Italy. Actual analyses are plotted as dots with the concentrations of Sr in the sample shown adjacent to the dot. AFC lines are com- puted assuming the original magma had δ^18 O = 5.6, 87 Sr/^86 Sr = 0.704, and 750 ppm Sr, and the country rocks have 87 Sr/^86 Sr = 0.736, δ^18 O = +13.6, and 150 ppm Sr. Dashed lines are contours of cal- culated Sr concentrations in the magma. There is reasonably good agreement between the calcu- lated model and the actual data, if we assume the bulk partition coefficient varied a bit (which would certainly be the case).
Sediment Subduction vs. Assimilation
Magmas erupted and emplaced in subduction zones have unique geochemistry that is apparent in both their trace element and radiogenic iso-
Figure 9.16. Variation of δ^18 O with 87 Sr/^86 Sr during AFC for a magma with an initial δ^18 O = 5.7 and (^87) Sr/ (^86) Sr = 0.703, and an assimilant with 87 Sr/ (^86) Sr = 0.735 and δ^18 O = +19. All curves are for R = 0. (5:1), except for one with D = 2 for which R = 0. (labeled 9:1). Dashed red lines are calculated Sr concentrations (ppm) assuming an initial concentra- tion of Sr in the magma of 500 ppm. Where D = 1, the problem simplifies to one of simple mixing. After Taylor (1980).
Chapter 9 Spring 2011
topic compositions. Most often, this unique geochemistry is taken as evidence that the sources of these magmas contain subducted sediments. However, in many instances, this interpretation is non-unique. The geochemi- cal signature of subducted sediment is, in many respects, simply the signature of conti- nental crust. In general, it is just as plausible that magmas have acquired this signature through assimilation of continental crust as by partial melting of a mantle containing subducted sediment. Magaritz et al. (1978) pointed out that by combining radiogenic isotope and oxygen isotope analyses, it is possible to distinguish between these two possibilities. Magaritz et al. reasoned as follows. First, many continental materials have oxygen iso- tope ratios that are quite different from man- tle values (e.g., Figure 9.18). Then they noted that all rocks have similar oxygen concentra- tions. Because of this, addition of 10% sedi- ment with δ^18 O of +20 to the mantle (+6) fol- lowed by melting of that mantle produces a magma with the same δ^18 O (+8.4), as adding 10% sediment directly to the magma (i.e., in this case the magma also ends up with δ^18 O = +8.4). However, addition of 10% sediment to the mantle, followed by melting of that mantle produces a magma with a very different Sr iso- tope ratio than does adding 10% sediment di- rectly to the magma. The reason for this differ- ence is that the Sr concentration of the magma is much higher (an order of magnitude or more higher) than that of the mantle. Thus addition of sediment to magma affects the Sr isotope ra- tio less than addition of sediment to the mantle. Indeed, most subduction-related magmas are richer in Sr (~ 400 ppm) than most continental materials and sediments; but sediments in- variably have more Sr than does the mantle (10-40 ppm). This point is illustrated in a more quantita- tive fashion in Figure 9.18. We discussed mix- ing trajectories earlier and noted that when two materials are mixed in varying proportions, a curve re- sults on a plot of one ratio against another. The degree and sense of curvature depended on the ratio of the ratio of the two denominators in the two end-members. Since the concentration of oxygen is ap- proximately the same in most rocks, the ratio of 16 O in the two end-members can be taken as 1. So, for example, in plot of 87 Sr/^86 Sr vs. oxygen isotope ratio, the degree of curvature depends only on the ratio
Figure 9.18. Mixing curves on a plot of δ^18 O vs. (^87) Sr/ (^86) Sr. The labels on the lines refer to different ratios of Sr concentration in the two end-members. After James (1981).
Figure 9.17. Variation of δ^18 O with 87 Sr/^86 Sr in the Ada- mello Massif in Italy compared with model AFC process using equation 9.07- 9.10. After Taylor and Sheppard (1986).
Chapter 9 Spring 2011
The mean δ^18 O of Marianas lavas is, in contrast to Banda, about +6.2 (Ito and Stern, 1986; Woodhead et al., 1987). This implies the amount of sediment involved in the source is less than 1%, which is con- sistent with the amount inferred from radiogenic isotope studies. In contrast to this interpretation, James and Murcia (1984) have used radiogenic and stable isotope systematics to argue for extensive assimilation of crust by magmas in the northern Andes. As shown in Figure 9.20, andesites from Nevado del Ruiz and Galeras volcanoes in Columbia define a steep array on plots of δ^18 Ο vs. 87 Sr/^86 Sr and 143 Nd/^144 Nd. Comparing this with Figure 9.18, we see that such steep ar- rays imply mixing between magma and crust rather than mantle and sediment. As we observed in the last lecture, assimilation will inevitably be accompanied by fractionation crystallization. James and Murcia (1984) modeled this assimilation using the equations similar to 9.07- 9.10 and the assumed con- centrations listed in Table 9.05. The model fits the data reasonably well and implies assimilation of up to about 12% country rock. An R of 0.33 provided the best fit, though James and Murcia (1984) noted this parameter was not well constrained. A high value of R suggests the assimilation is occurring at moderate depth (< 10 km), because it suggests only moderate input of heat in melting the country rock. On a broader basis, Harmon et al. (1984) have used O isotope ratios to show that extent of crustal as- similation in Andean magmas varies regionally and correlates with crustal thickness. The Andes are divided into 3 distinct volcanic provinces. In the north, the subduction zone is steeply dipping, and the volcanoes are located approxi- mately 140 km over the Benioff zone. The crust in the Northern Volcanic Zone (NVZ), located in Co- lumbia and Ecuador, is approxi- mately 40 km thick and is mainly Cretaceous and Cenozoic in age. Volcanics are primarily basaltic andesites and andesites. In the Central Volcanic Zone (CVZ), which extends from southern Peru to northern Chile and Argentina, the crust is 50 to 70 km thick and Pre- cambrian to Paleozoic in age. The
Table 9.05: Parameter for Ruiz Assimilation Model
Original Magma Assimilant δ^18 O 6.5‰ 11‰ ∆ 0.5‰ (^87) Sr/ (^86) Sr 0.7041 0. βSr 0. Sr, ppm 650 100 (^143) Nd/ (^144) Nd 0.51288 0. Nd, ppm 20 20 βNd 0. R 0.
Figure 9.20. Sr, Nd, and O isotopes in lavas from Nevado del Ruiz Volcano in Columbia and an AFC model calculation. Dashed line shows AFC trajectory for another Colombian volcano, Galeras. After James and Murcia (1984).
