HR: 11:50h
AN: V41D-07 [PDF]
TI: Constraints on Melting Processes Beneath Subduction Zones from U-Pa Disequilibria
AU: Regelous, M
EM: M.Regelous@bris.ac.uk
AF: Department of Earth Sciences, University of Bristol,
Wills Memorial Building, Bristol, BS8 1RJ
United Kingdom
AU: * Turner, S P
AF: GEMOC, Department of Earth and Planetary Sciences,
Macquarie University, Sydney, NSW 2109
Australia
AU: Hawkesworth, C J
AF: Department of Earth Sciences, University of Bristol,
Wills Memorial Building, Bristol, BS8 1RJ
United Kingdom
AU: Elliott, T R
AF: Department of Earth Sciences, University of Bristol,
Wills Memorial Building, Bristol, BS8 1RJ
United Kingdom
AU: Rostami, K
AF: Department of Earth Sciences, University of Bristol,
Wills Memorial Building, Bristol, BS8 1RJ
United Kingdom
AB:
$^{231}$Pa is the longest-lived intermediate daughter nuclide produced by the decay of $^{235}$U, and has a half-life (32,700
y) intermediate between those of $^{230}$Th and $^{226}$Ra. Pa$^{5+}$ is highly incompatible during mantle melting, and is
thought to be highly insoluble in aqueous fluids, compared to U, thus there is potential for large U-Pa fractionation during
fluid-present melting in subduction zones.
We have analysed $^{231}$Pa concentrations in a suite of 47 lavas from 7 different convergent margins, using isotope dilution
and MC-ICPMS. The samples are from plate margins with a wide range in subduction rate, lithosphere thickness, fluid flux,
subducting plate age and sediment thickness. All samples are of known eruption age, and have previously been analysed for
major and trace elements, radiogenic isotopes, and U-Th-Ra disequilibria.
Measured ($^{231}$Pa/$^{235}$U) ratios in these samples range from 0.82 to 2.42 with one lava from the Aleutians having a
ratio of 3.49. For the dataset as a whole, there is a broad positive correlation of ($^{231}$Pa/$^{235}$U) with
($^{230}$Th/$^{238}$U), and samples with higher ($^{231}$Pa/$^{235}$U) also tend to have lower ratios of Ba/Th and U/Nb
(smaller slab fluid input). There is no simple relationship between $^{231}$Pa excess and subduction rate, subducting plate
depth or sediment flux.
Although U/Nb ratios indicate that up to 98% of the U in these lavas is derived from the subducting plate, all but one of
the samples have ($^{231}$Pa/$^{235}$U) $>$1.0, and several have $>$100% $^{231}$Pa excess. This could indicate that a
period of several half-lives of $^{231}$Pa elapsed between the timing of fluid addition from the slab and the final melting
event. On the other hand, most samples have $^{226}$Ra excess, indicating that the last episode of Ra addition to the melting
zone occurred less than 8000 years ago. The large Pa excesses imply that significant fractionation of U and Pa occurs during
separation of melt from the mantle. As both U and Pa are highly incompatible in mantle minerals, melting must occur at low
porosity over a period of time that is significant relative to the half life of $^{231}$Pa to allow $^{231}$Pa in-growth in
the melting region, either by dynamic melting in the mantle wedge at low melting rate, or during melting of
continuously-fluxed mantle. More detailed modelling of the melting process is in progress.
DE: 1025 Composition of the mantle
DE: 1040 Isotopic composition/chemistry
DE: 3640 Igneous petrology
DE: 8434 Magma migration
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2003 Fall Meeting