Incomplete Outgassing of Very Gas Rich MORB from the Mid-Atlantic Ridge at 14°N

Abstract

The samples dredged along the Mid-Atlantic Ridge between 8 and 16~ by the R/V Akademic Boris Petrov (Vernadsky Institute for Geochemistry), listed in Staudacher et al. (1989) and chemically analysed by Bougault et al. (1988) and Dosso et al. (1991) resemble very much normal MORB. The reason for fame of these samples is that among them were the samples from one dredge haul close to 14~ known as 2IID43, which contained extra ordinary quantities of volatiles reported by Standacher et al. (1989), Sarda and Graham (1990), Javoy and Pineau (1991), Pineau and Javoy (1994), Burnhard et al. (1997), and Moreira et al. (1998). Besides the porosities of 17 and 5% of 2IID43 samples indicated that they contained substantially more gas when they erupted than has been measured by the here fore cited authors. The volatile contents of samples from the other dredge hauls were MORB like with porosities varying from 1 to 3% (Staudacher et al.). The enrichments of rare gas isotopic concentrations in 2IID43, with respect to the average concentration of the other BP samples amount to factors of 5, 50, 100 and 140, for both helium isotopes, for 4~ for 2~ 36Ar, and 13~ and for 84Kr, respectively. The 21-ID43 samples were dredged from 3510 m below sea level. The temperature of and the pressure on the erupting 2I-ID43 lava were larger than those for the overlying deep sea water, and its rare gas, C Q and N2 concentrations are significantly larger than those found in deep sea water. These conditions stand in the way of contamination, implying always a diffusion step, by atmospheric gases in deep sea water. The inferred absence of important atmospheric contamination of the lava at eruption time or while it flowed upwards towards sea bottom is also indicated by the results of the step crushing experiments of Staudacher et al. (1989) and Moreira et al. (1998) on 2IID43. Their data show that most small bubbles had smaller 36Ar/4~ ratios than the larger ones. If important contamination had taken place the the opposite should have been observed because: (i) 36Ar diffuses faster in the melt than 4~ (ii) the mean distance between the small bubbles is smaller than between the large ones, (iii) the surface/volume ratio of a given volume of small bubbles is larger than that of the same volume of large bubbles, and (iv) the atmospheric and deep sea water 36Ar/4~ ratios are larger than those of 217D43. Usually, oceanic tholeiites, on their way up to the sea floor, stay for a certain time interval in a magma chanther. During this pause in their ascension, these lavas exsolve volatiles in bubbles. The great majority of these bubbles escape before and during expulsion of lava from the magma chamber. Therefore, upon eruption the total CO2, N2 and rare gas contents of the lavas correspond roughly to the dissolved equilibrium concentrations under magma chamber conditions. We think that the 217D43 lava, on its way up, bypassed a magma chamber stage, as is also inferred for xenoliths-bearing alkali-rich oceanic island lavas. As a result, the 2IID43 lava degassed very incompletely in the relatively narrow conduit leading to the sea floor. The degassing caused rheological changes, very turbulent two phase flow, and the occurrence of coalescence and fission of bubbles. These conditions give rise to a spatially varying bubble distribution and the observed local lack of correlation between the small 36Ar and the large 4~ concentrations. The faster diffusing helium isotopes manage to maintain a homogeneous distribution, hence the linear correlation between 3He and 4He concentrations remains indistinguishable from 3He and 4He concentration correlations shown by MORB and the other BP samples. This incomplete degassing process, at depth greater than sea bottom, can explane the difference between 2IID43 and BP samples. The latter ones degassed to an approximate equilibrium level under magma chamber conditions. They were quenched shortly after being expelled from the magma