Magmas volatile elements

Figure 16 Comparison of observed (open) and calculated (solid) depletions of phosphorus, tungsten, cobalt, nickel, molybdenum, and rhenium (circles) together with those for gallium, tin, and copper (inverted triangles) (sources Righter and Drake, 1997, 1999, 2000). The calculated depletions utilize the partitioning expressions of Righter and Drake (1999) for conditions of 2,250 ( 300) K (1,973 °C), 27 ( 6) GPa, AIW = — 0.4 ( 0.3) between a hydrous peridotite (NBO/t = 2.65) magma ocean and metallic liquid. The observed depletions are those of McDonough and Sun (1995), but volatility corrected as described by Newsom and Sims (1991), where the correction is made based on comparisons to trends of lithophile volatile element depletions.

It is likely that the earliest events in the Earth s mantle were not the product of "normal" mantle convection but rather related to planetary processes such as fractionation within a magma ocean. In this way we can explain the very early differentiation of the Earth (pre-4.5 Ga), proposed on the basis of short-lived Nd-isotopes. Similarly, the extreme volatile element loss from the Earth might be explained in this way. These early processes are thought to have ceased within the first 100 Ma of Earth history (Yokochi Marty, 2005). 

In most respects, asteroid 4 Vesta is geochemically similar to the Moon. As judged from howardite-eucrite-diogenite (HED) meteorites (see Chapter 6), Vesta is an ancient, basalt-covered world (Keil, 2002). Its rocks are highly reduced, and its depletions in volatile and siderophile element abundances resemble those of lunar basalts. And like the Moon, Vesta is hypothesized to have had an early magma ocean. The exploration of Vesta is now in progress, and within a few years we may have enough data to discuss it in a similar way that we have considered the Moon. 

Weathering, erosion, and sedimentation are also important in concentrating arsenic in continental crusts. Togashi et al. (2000) even argue that weathering, erosion, and sedimentation are more responsible for enriching arsenic in the upper crust of Japan than contributions from magmas and lavas. Additionally, as discussed in the next section, the moderate volatility of arsenic, its sufficient solubility in hot fluids, and its reluctance to enter and accumulate in the mantle should preferentially concentrate the element in the crust. 

REE patterns are fractionated for all the rocks, but tholeiites show lower La/Yb ratios than alkaline products (Fig. 8.5a). Incompatible element patterns normalised to primordial mantle compositions for mafic rocks are very different from the Aeolian arc and central-southern Italian peninsula. Both tholeiitic and alkaline basalts show a marked upward convexity, with negative spikes of K (Fig. 8.5b). Note, however, that there are also negative anomalies for Hf and Ti, which are uncommon in most Na-alkaline basalts from intraplate settings (e.g. Wilson 1989). Overall, the Etna magmas have been found to be more enriched in volatile components than common intraplate magmas, and water contents up to 3-4 wt % have been found by melt inclusion studies (Corsaro and Pompilio 2004 Pompilio, personal communication). 

There are two major issues of concern with the approach of using C02/ He and N2/ Ar (or N2rHe) ratios in combination with and values to constrain the sources of volatiles at arcs. The first issue is the selection of representative end-member isotopic and relative elemental abundances—this factor has a profound effect on the deduced provenance of the volatile of interest. The second is the assumption that various elemental (and isotopic) ratios observed in the volcanic products are representative of the magma source. Both have the potential to compromise the accuracy of the output flux estimates. 

---