Atmospheric Contamination in Oceanic Samples

Only a few studies on oceanic basalts provide neon isotopic ratios with enough precision to allow a discrimination between their “neon B” ratios. Determining the mantle neon isotopic ratio is fundamental to understanding the mechanism of incorporation of noble gases in the silicate Earth as it will be discussed in Section 8. Solar-like 20_Ne/22_{Ne (~13.8) would suggest the dissolution of Ne in}

a molten Earth from a solar-like dense primordial atmosphere, whereas a ratio close to neon B would suggest that neon was incorporated into the parent bodies by solar wind irradiation during planetary accretion.

In order to discuss the actual 20_Ne/22_{Ne ratio of the mantle, it is essential}

to understand the physical process behind what has been called the atmos- pheric contamination and so be able to distinguish between a mantle-derived signature and any shallow contamination process. Step-crushing or step-heating often shows the presence of an atmospheric component in samples (Fig. 5.7), for both isotopic and elemental ratios, although sometimes the rare gases are elementally fractionated (Harrisson et al., 2003). I published one example of such step-crushing in 1998 on the popping rock sample 2πD43 (Moreira et al., 1998). Figure 5.7 shows the neon isotopic composition and concentration during step crushing. It can be easily seen that atmospheric neon is associated with high neon concentrations and is mainly found during the first crushing steps. Similar obser- vations have been seen in other studies where samples were step crushed (Trieloff

et al., 2000, 2002; Mukhopadhyay, 2012; Parai et al., 2012; Tucker et al., 2012). A few key remarks can be made about this phenomenon. The first one is that such an important atmospheric component is mainly observed in vesicle-rich samples (Fig. 5.8) (Ballentine and Barfod, 2000). The second is that the blanks measured prior to analyses of these gas-rich samples are generally much higher than for typical samples, suggesting (atmospheric) gas is leaking out of the sample. Blanks

decrease after the first crushing steps. Finally, the recent study of Raquin et al. (2008) on noble gas analyses of single vesicles has shown that atmospheric gases are at atmospheric pressure, not at crustal or mantle pressures.

Figure 5.7 Neon concentration and isotopic ratios in the 2πD43 popping rock sample (Moreira et al., 1998). This study shows that during step crushing, the atmo- spheric gases are liberated first.

All these observations suggest that the air component observed in oceanic basalts reflects atmospheric contamination and that this air fills empty vesicles during storage in the laboratory (Fig. 5.9). We have to suppose that some empty vesicles are connected to the atmosphere through thin chan- nels, formed by a fracture in the glass. Such a simple explanation accounts for all the observations. An empty vesicle filled with air can contain ~10–8 _cm3_STP

of 20_{Ne (for a radius of 500 µm), which is}

of the order of the measured value in the popping rock or other gas rich samples. Vacuum is reached prior to analysing samples (12 to 36 hours at ~100 °C), so the gas in these vesicles connected to surface should be pumped. High blanks observed before the analysis of these samples indeed suggest this is the case.

However, it takes time to pump such small volumes if the channels have small diameter (e.g., ~µm) and overnight appears to be not long enough to pump out all the atmospheric gases.

Figure 5.8 Figure showing the rela- t i o n s h i p b e t w e e n t h e vesicularity and the neon concentration, taken here as a proxy of atmospheric contamination. High vesicu- larity seems to entrain high atmospheric contamination (modified from Ballentine and Barfod, 2000).

Although contamination by hydrothermal alteration is a viable mechanism (e.g., Patterson et al., 1990), I am convinced that the atmospheric component observed in most samples reflects atmos- pheric contamination in the labo- ratory during storage, as suggested by (Ballentine and Barfod, 2000; Trieloff et al., 2003; Raquin et al., 2008). This is different from the interpretation of Sarda (2004) who proposed that this compo- nent instead reflects atmospheric subduction in the mantle, through the recycling of altered oceanic crust (see Section 7).

Atmospheric contamination is present in most of the samples and therefore the determination of the maximum 20_Ne/22_{Ne ratio in}

the mantle source of these samples is difficult. One can never be sure that the maximum ratio measured in a sample is really the mantle signa- ture free of atmospheric contamination. Step crushing or step heating is an analytical method that could minimise this uncertainty but does not provide an undis- puted mantle signa- ture. Laser ablation on single vesicles such as presented in Section 4 is also a powerful technique allowing the measurement of atmos- pheric-free noble gases, although the precision is less good that for bulk sample analyses.

Figure 5.9 Cartoon explaining the atmo- spheric composition observed in oceanic basalts. Air can enter empty vesicles during storage in the laboratory. High vesicularity samples can therefore be rich in atmospheric gases as observed by Ballentine and Barfod (2000) and baking overnight under vacuum might not be long enough to fully outgas these vesicles.

Figure 5.10 Method of correction for atmospheric contami-

nation using either a solar or neon B 20_Ne/22_Ne

ratio for the mantle. The 21_Ne/22_{Ne ratio can}

then be derived assuming a simple binary mixing between the uncontaminated magma and an atmospheric component. This method is slightly different from the one proposed by Honda et al. (1993).

5.3 Atmospheric Contamination Correction for Determination