K. Toyoda et al., DUPAL ANOMALY OF BRAZILIAN CARBONATITES - GEOCHEMICAL CORRELATIONS WITH HOTSPOTS IN THE SOUTH-ATLANTIC AND IMPLICATIONS FOR THE MANTLE SOURCE, Earth and planetary science letters, 126(4), 1994, pp. 315-331
Geochemical and Sr, Pb, O and C isotopic data are reported for carbona
tite samples from five locations in southeast Brazil. Elemental abunda
nces and delta(13)C(PDB) data (between -5.8 and -7.2 parts per thousan
d) prove that all the samples are derived from the mantle. Dupal isoto
pic characteristics in the all carbonatite samples from five locations
in southeast Brazil are found in this study, characteristics that hav
e not previously been recognized in carbonatites. Brazilian carbonatit
es possess average Delta 8/4 values of between 101 and 145, average De
lta 7/4 values of between 5.2 and 10.3, and initial Sr isotopic compos
itions of between 0.7046 and 0.7062. The Brazilian carbonatites compri
se two groups: The northern group is coincident with the passage of th
e Trindade non-Dupal hotspot at ca. 80 Ma, while the southern group ma
inly corresponds to the passage of the Tristan de Cunha Dupal hotspot
at ca. 130 Ma. Although we expected a geochemical correlation between
the Brazilian carbonatites and the South Atlantic hotspots, the enrich
ed isotopic signature (EM1) of all the carbonatite samples is similar
to that of alkali basalts on Tristan de Cunha. The combined O-Sr isoto
pic diagram indicates that the southern group carbonatites have neglig
ible or only slight crustal contamination. The northern group samples
show significantly higher delta(18)O(SMOW) values of 9-14 parts per th
ousand, more radiogenic Pb isotopic ratios, and a Sr-87/Sr-86 value of
0.705. Even if these signatures are derived from the contamination of
a lower crustal component with mantle sources, it is clear that the p
arental magma also has inherent EM1 isotopic characteristics. The inte
rpretation of the origin of EM1 in the Brazilian carbonatites (subcont
inental lithospheric mantle vs. asthenosphere) is dependent on the mod
el of Parana volcanism at ca. 130 Ma, which remains controversial. One
possibility is that both the northern and southern carbonatites came
from enriched SCLM under a part of Gondwanaland. In this case, a hotsp
ot would provide the thermal energy to melt the lithospheric source re
gion for both the Parana flood basalts and the alkali carbonatitic vol
canism. Another possibility is that the source of the northern carbona
tites is also the Tristan plume carbonate-rich material which had once
been trapped under the crust and reactivated by the Trindade hotspot,
on the assumption that Parana volcanism at ca. 130 Ma was mainly trig
gered by the huge Tristan plume activity. If anything, we favour the l
atter and believe an asthenospheric mantle plume origin for bath the u
ltimate carbonate source and the Dupal anomaly in the Brazilian carbon
atites.