3.4 Volatiles in Magmatic Systems 119
Carbon. Isotopic fractionation between CO
2
and dissolved carbon in melts has
been estimated by various authors to vary between 2 and 4‰ (as summarized by
Holloway and Blank 1994), the vapor being enriched in
13
C relative to the melt. This
fractionation can be used to interpret the carbon isotope composition of glasses and
CO
2
in volcanic gases and to estimate the initial carbon concentration of undegassed
basaltic melts.
Reported δ
13
C-values for basaltic glass vary from −30 to about −3‰ that repre-
sent isotopically distinct carbon extracted at different temperatures by stepwise heat-
ing (Pineau et al. 1976; Pineau and Javoy 1983; Des Marais and Moore 1984; Mattey
et al. 1984). A “low-temperature” component of carbon is extractable below 600
◦
C,
whereas a “high-temperature” fraction of carbon is liberated above 600
◦
C. There
are two different interpretations regarding the origins of these two different types
of carbon. While Pineau et al. (1976) and Pineau and Javoy (1983) consider that
the whole range of carbon isotope variation observed to represent primary dissolved
carbon, which becomes increasingly
13
C depleted during multistage degassing of
CO
2
, Des Marais and Moore 1984) and Mattey et al. (1984) suggest that the “low-
temperature” carbon originates from surface contamination. For MORB glasses, the
“high-temperature” carbon has an isotopic composition typical for that of man-
tle values. Island arc glasses have lower δ
13
C-values, which might be explained
by mixing two different carbon compounds in the source regions: an MORB –
like carbon and an organic carbon component from subducted pelagic sediments
(Mattey et al. 1984).
Nitrogen. The determination of nitrogen isotopes in basaltic glasses is severely
complicated by its low concentration, which makes nitrogen sensitive to atmo-
spheric contamination and to addition of surface-derived materials i.e., organic mat-
ter. Nitrogen in basaltic glasses has been determined by Exley et al. (1987), Marty
and Humbert (1997), and Marty and Zimmermann (1999). The recent studies by
Marty and coworkers indicate that nitrogen in MORB and OIB glasses has an av-
erage δ
15
N-value of around −4 ±1‰ (see Fig. 3.6). The major factors affecting
its isotopic composition appear to be magma degassing and assimilation of surface-
derived matter.
Sulfur. The behavior of sulfur in magmatic systems is particularly complex: sul-
fur can exist as both sulfate and sulfide species in four different forms: dissolved
in the melt, as an immiscible sulfide melt, in a separate gas phase, and in various
sulfide and sulfate minerals. MORB glasses and submarine Hawaiian basalts have a
very narrow range in sulfur isotope composition, with δ
34
S-values clustering around
zero (Sakai et al. 1982, 1984). In subaerial basalts, the variation of δ
34
S-values is
larger and generally shifted towards positive values. One reason for this larger vari-
ation is the loss of a sulfur-bearing phase during magmatic degassing. The effect
of this process on the sulfur isotope composition depends on the ratio of sulfate to
sulfide in the magma which is directly proportional to the fugacity of oxygen (Sakai
et al. 1982). Arc volcanic rocks are particularly enriched in
34
S, with δ
34
S-values
up to +20‰ (Ueda and Sakai 1984; Harmon and Hoefs 1986) which is considered
to be mainly a product of recycling of marine sulfate during subduction.