I. isotopic composition of hydrogen and oxygen in the mantle
(1) oxygen
Facts have proved that it is much more difficult than expected to evaluate the oxygen isotope composition of the mantle, especially the possible change degree of oxygen isotope composition. One method is to use basalt as a representative of mantle samples. Isotopic fractionation is small during partial melting, so the oxygen isotopic composition of basalt should be similar to that of its mantle source region. However, the exchange of oxygen isotopes during the assimilation and weathering of magma and crustal rocks complicates the situation. Another method is to use direct mantle samples, such as occasional inclusions in basalt. However, inclusions are much rarer than basalts.
Fig. 12- 1 1 illustrates the oxygen isotopic compositions of olivine and clinopyroxene in 76 peridotite xenoliths analyzed by Mattey et al. (1994) by laser fluorination technique. The observed total value range is only twice the expected single analysis error, indicating that the isotopic composition of the mantle is quite uniform. The average difference between olivine and clinopyroxene is about 0.5‰, which is consistent with the expected fractionation between minerals at mantle temperature. Mattey et al. estimated that the total composition of these samples was about +5.5‰.
Figure 12- 12 shows the δ 18O distribution of four different groups of basalts. In order to avoid weathering, Harmon and Hoefs( 1995) only choose basalt glass with water content less than 0.75% or basalt with known eruption history.
Fig. 12- 1 1 oxygen isotopic composition range of olivine and clinopyroxene in mantle peridotite xenoliths.
Fig. 12- 12 oxygen isotope composition of young fresh basalt.
The average δ 18OSMOW of MORB is +5.7%, and the change around this average is quite small. Therefore, the depleted upper mantle seems to be a fairly uniform oxygen source. The uniformity observed in MORB is consistent with that observed in mantle-derived inclusions. The slight difference (0.2‰) between the two average values can well reflect the fractionation during partial melting.
Other rock formations show greater heterogeneity than MORB rocks or peridotite xenoliths. It is generally believed that the average value of ocean island basalt, which represents the mantle plume, is the same as that of peridotite xenoliths, but the change is greater. Basalt related to subduction (i.e. island arc basalt and continental basalt) has corrected δ 18O value, which is similar to volcanic rocks in continental plate.
It is not clear whether the changes of these groups reflect the inhomogeneity, melting process, assimilation or just weathering of the mantle. Hawaii is a place where basalt and inclusions have high quality data, which can be compared. The δ 18O values of olivine in Hawaiian basalt and Hawaiian inclusions are lower than those of similar substances in other places. This at least shows that the variation of oceanic basalt partly reflects the inhomogeneity of mantle. However, other studies are needed to determine its significance.
(2) Hydrogen
The estimation of isotopic composition of hydrogen, carbon, nitrogen and sulfur in the mantle is more problematic. They are volatile elements and exist in mantle materials at low concentrations. They are partially or completely distributed in the magma gas phase during eruption. This gas phase disappears unless it erupts at the bottom of the deep sea. In addition, there are several oxidation states of carbon, nitrogen and sulfur, and isotopic fractionation occurs among various components formed by these elements (such as carbon dioxide, carbon monoxide, methane, H2 and H2O). This raises two problems: first, in the degassing process, even at magma temperature, obvious fractionation can occur; Secondly, the concentration of these elements is very low due to the loss of gas phase. Other problems are that their isotopic ratios are easily disturbed by pollution. Therefore, the isotopic composition of these elements in igneous rocks does not necessarily reflect the composition of magma or its mantle source.
Hydrogen mainly exists in the form of water, but also in the form of H2, H2S and CH4, which can be lost from magma during degassing. However, basalt erupting in the ocean at 1000 meters or deeper retains most of the dissolved water. Therefore, the mid-ocean ridge basalt and the basalt erupted from seamounts are important sources of information on the isotopic composition and abundance of hydrogen in the mantle.
As shown in figure 12- 13, the average standard deviation of MORB is-67.5 ‰ To what extent this change reflects the degassing, fractionation and pollution in the process, it is still unclear. Kyser( 1986) proposed that the isotope of hydrogen in the mantle is a unified value of -80‰. He believes that the heavy isotopic composition of MORB generally reflects the loss of H2O and other processes. Others, such as Chaussidon et al. (199 1), have observed the correlation between δD and Sr and nd isotopic ratios, and put forward that these are clear evidence of the heterogeneity of hydrogen isotopes in the mantle. Chaussidon et al. proposed that the δD value of the depleted upper mantle is about -55‰.
Fig. 12- 13 MORB (mid-ocean ridge basalt) and hydrogen isotope composition range of mantle phlogopite and amphibole.
Sheppard and Epstein( 1970) first tried to estimate the hydrogen isotopic composition of mantle materials. They analyzed the water-bearing minerals in the inclusions and concluded that the δD of the mantle is changing. Since then, a lot of research on these substances has been going on. As shown in figure 12- 13, the δD of phlogopite is generally similar to that of MORB, although a heavier value appears. The δD value of amphibole varies greatly, and the average hydrogen content is heavier. In the temperature range of 800 ~1000℃, the fractionation between water and phlogopite is close to zero, while that between water and amphibole is about-15‰. However, equilibrium fractionation itself cannot explain the change of amphibole or the difference of average δD between phlogopite and amphibole. The formation of mantle amphibole may involve a complex process of Rayleigh fractionation. Compared with phlogopite, this is also consistent with the more variable water content of amphibole. There is also a large-scale heterogeneity in δD of inclusions. For example, Deloule et al. (199 1) found that the amphibole δD system of inclusions in basalt in central France has a very high value (-59 ‰ ~-28 ‰), while the amphibole δD system from Hawaii has a very low value (-125‰). On the contrary, Deloule et al. also observed the isotope heterogeneity in single crystals, which they attributed to the incomplete equilibrium with magma or fluid.
Second, stable isotopes in crystalline magma.
The change of stable isotopic composition produced by magmatic crystallization depends on the crystallization method. Generally, the crystallization is balanced in the simplest and most impossible situation. In this case, the crystallized mineral and the melt remain in isotopic balance until the crystallization is completed. At any stage in the crystallization process, the isotopic composition of minerals and melts is related to the fractionation coefficient (α). Once the crystallization is completed, the rock will have the same isotopic composition as the initial molten state. In the process of crystallization, at any time, the relationship between the isotope ratio in the residual melt and the original isotope ratio is
Isotopic chronology and geochemistry.
Where: R 1 is the ratio in the liquid phase; Rs is the isotope ratio in the solid phase; R0 is the isotope ratio of primitive magma; F is residual magma's score. This equation can be easily derived from the mass balance, the definition of α and the assumption that the concentration of oxygen in magma is equal to that in crystal. It is assumed to be effective in the range of 10‰. Equation (12- 1 1) is more conveniently expressed by δ:
Isotopic chronology and geochemistry.
Where: δ melt is the value1-f after partial crystallization of magma; δ0 is the value of primitive magma.
For silicate, α should not be much less than 0.998 (i.e. δ = δ18 oxals-δ18 omelt ≥-2). For α=0.999, that is, after 99% crystallization, the isotope ratio in the remaining magma will only change by 1‰.
The treatment of separating crystals is similar to Rayleigh distillation. In fact, they are all described by the same equation:
Isotopic chronology and geochemistry.
The key of these processes is that the reaction products (gas phase in distillation and crystalline phase in crystallization) are only in instantaneous equilibrium with the original phase. Once it is produced, it loses the opportunity to further balance with the original and separates. This process is more effective in generating isotopic changes in igneous rocks, but its effect is limited, because α is generally not much different from 1. The figure 12- 14 shows that the δ value (≈ 1000(α- 1)) has undergone the separation crystallization and the calculation change of the melt isotope composition. In fact, during the crystallization process, δ will change due to ① temperature change, ② mineral crystallization change and ③ liquid phase composition change. This change usually means that the effective δ will increase with crystallization. We can predict that the non-silicate crystallized in the melt, such as magnetite, has the highest isotope fractionation with the melt crystallized at low temperature, such as rhyolite; Melts that crystallize at the highest temperature, such as basalt, have the smallest fractionation.
Fig. 12- 14 the relationship between oxygen isotope composition and crystallinity during Rayleigh fractional crystallization.
Fig. 12- 15 Relationship between oxygen isotope composition and silica content of tholeiite and alkaline basalt.
Figure 12- 15 shows that δ 18O observed in two groups of rocks is a function of temperature: one group comes from the rift developed in the Galapagos spreading center, and the other group comes from Ascension Island. In Galapagos, the net variation of δ 18O of the most differentiated and undifferentiated rocks is about 1.3‰. The ascension kit has only a net change of 0.5‰. These rocks and other rocks show that the effective δ is generally very small, about 0. 1 ‰ ~ 0.3 ‰. Consistent with this, the similarity between peridotite and MORBδδ 18O indicates that the fractionation is generally below 0.2‰ in the melting process.
We can conclude that the temperature dependence of oxygen isotope fractionation is: low temperature (that is, the surface environmental temperature is 300 ~ 400℃ to the hydrothermal system temperature), and the oxygen isotope ratio changes through chemical processes. Variation can be used to represent the nature of the process and the temperature at which the process occurs under equilibrium conditions. At high temperature (the temperature inside the earth or magma), the oxygen isotope ratio is least affected by chemical processes and can play the role of tracer like radioactive isotopes.
This generalization leads to a famous saying: Igneous rocks whose oxygen isotope composition is obviously different from the original value (6‰) must be influenced by the low temperature process or must contain components that once existed in the earth's crust (Taylor &; Shepard, 1986).
Thirdly, separation crystallization and assimilation mixed dyeing process.
Because the oxygen isotope ratio of mantle-derived magma is reasonable and uniform ((5.6 1) ‰), and it is generally different from rocks with surface and water balance, oxygen isotope is a useful tool to identify and study the assimilation of intrusive magma to surrounding rocks. We can think of this process as a simple mixture of two components: magma and surrounding rock. In fact, it is always a question of at least three components, including surrounding rock, magma and mineral crystallization in magma. Magma is basically never overheated, so the heat for melting and assimilating surrounding rocks can only come from the latent heat of magma crystallization. It takes about 1kJ g- 1 to heat the rock from 150℃ to150℃, and another 300J g- 1 can be melted. If the latent heat of crystallization is 400 J g-1,the crystallization of 3.25g magma can assimilate 1g surrounding rock. Because a part of the heat is lost through simple conduction to the surface, it can be concluded that the amount of crystallization is definitely greater than that of assimilation amount (the limit is that the crystallization quality of rocks in the deep crust is equal to the assimilation quality only when the melting point begins to assimilate). After separation crystallization and assimilation (AFC), the change of melt isotope composition is given by the following formula:
Isotopic chronology and geochemistry.
Where: r is the mass ratio of crystalline substance to assimilated substance; δ is the difference of isotope ratio between magma and crystal (δmagma-δ crystal), f is the fraction of residual melt, δ m is δ 18O of magma, δ0 is the initial δ 18O of magma, and δa is δ 18O of assimilate. It is assumed that the oxygen content in crystal, magma and assimilate is the same. When R= 1, the equation does not hold, but as mentioned above, r is usually much larger than 1. Figure 12- 16 illustrates the change of magma δ 18O with initial δ 18O=5.7‰ with the progress of crystallization assimilation.
The combination of stable isotope ratio and radioactive isotope ratio provides a more powerful tool for testing assimilation.
Fig. 12- 16 Relationship between oxygen isotope and crystallization fraction during assimilation and separation crystallization.