Impact volcanism and upper mantle melting; mega-melting
Adrian P Jones, Department of Earth Sciences,
2008
Vienna
University College London, Gower Street, London WC1E 6BT
email: adrian.jones@ucl.ac.uk
OLD STUFF
Abstract: It is now widely accepted that large-scale mantle melting initiated by energetic
bolide impacts must have been a common event during the early history of the Earth, and
igneous petrology can be used to address the fate of these mega-melts. In other words, the
same 1000 km scale impact craters seen on the Moon did not form on Earth, but instead are
envisaged to have formed gigantic melt-filled basins, of the order of >25 to 100 km thick,
which almost certainly differentiated during slow cooling, and may have helped to initiate the
crustal dichotomy between sialic and basaltic terrains (Therriault, Grieve et al LPSI 2001).
Different physical models for large impacts processes differ in detail, but converge on this
result that for impact events above some critical size (few hundred kilometres effective crater
diameter), the volume of impact melt exceeds the transient crater volume. In addition, the
contribution to substantially increasing volumes of impact melting (> 10e6 km3; Jones et al
EPSL 2002) derived from decompression of the underlying mantle remains an important but
largely unexplored concept.
The Sudbury impact melt sheet serves as a newly recognised example of differentiation
from super-liquidus melts of re-melted sialic crust, but what about similar events in purely
basaltic or peridotitic terrains (ie oceanic)? Igneous petrology provides the philosophical
criteria for predicting how such processes as immiscible liquid fractionation and subliquidus
phase fractionation will occur during thermodynamic crystallisation over the pressure range
of what are effectively discrete small magma oceans. However, there is a huge gap in our
knowledge of the high temperature data required,which extends well above the liquidus for
silicates (>> 1900 K) at relatively low pressures (< 5 GPa). Hence, a new era of
experimental igneous petrology is urgently required to understand these superheated events
in the Earth’s ancient mantle. To put this into context, we can imagine the instantaneous
formation of a gigantic superheated melt disc whose vertical height surpasses the total depth
of even the thickest crust. How did such a superhot lid influence the underlying mantle?
Lastly, such nonplume ideas (impact volcanism) need not be confined to the time period of
the early Earth, and have been proposed to operate during the Mesozoic (Ontong Java
Plateau; Jones Elements 2005) and on other planets, including Venus (Hansen JGR 2006).
NEW STUFF
NEW Igneous Petrology FRAMEWORK For Impact Magma Oceans
Hadean impact craters into basaltic and peridotitic lithosphere producing ~ million km 3
impact melts, as open or crusted small regional magma oceans.
Major melt
fractionation processes share revolutionary petrological concepts developed for (1) the
~200 km scale Sudbury impact melt sheet (Keays and Lightfoot, 2004) and (2) ~10,000
km (Earth-Moon) scale global magma ocean model (eg Solomatov 2000). Unlike a
magma ocean, the walls and floor of the magma body are relatively well confined to
~100-200 km depths. This model ignores large-scale secondary melting or “impact
plume” initiation beneath the impact site, and ignores seawater interaction.
Stage 1 (~10s s):
SUPERHEAT Impact melt volume rapidly increasing,
superheated liquids, vapour phase, violent phase disequilibrium, vesiculation from self
vapour possible (subsequently "void").
Stage 2 (~ hours):
SUPERLIQUIDUS Liquid phase, melt volume ~constant,
surface vapour condensed, all liquid phenomena; emulsions, immisicibility, liquid-liquid
frationation, gravity sensitive, low viscosity melt systems, cooling still above liquidus at
shallow depths. Strongly disequilibrium processes, strirring and reservoir segregations;
wall rock digestion. Subcrater melting persists perhaps ~doubling of melt volume.
Stage 3 (>>years):
SUPERSOLIDUS Crystallisation phase, melt volume
thermodynamically controlled, classical igneous petrology starts, P-T gradients
Approaching equilibrium processes. Increasing viscosity, slow cooling.
2
1
FIGURE 1. Impact model of melt distribution (left hand side) resulting from
large bolide impact into young, relatively hot oceanic lithosphere (uniform
lithostatic P-T gradients and parameterised peridotite melting model);
snapshots cover the period of about 10 minutes duration, during which time
approximately >106km3 melt forms, ranging from superheated liquid
peridotite (red - yellow: komatiite) to lower temperature melt (blue; basaltic).
After Jones et al (2005) Box is 700 km across by 150 km deep.
FIGURE 2. TOTAL MELT VOLUME produced, including shock melt (~1 million km3) and
post-shock melt ((~1 million km3) produced in an Ontong Java simulation (Jones et al
2005). Two apparent rates of melt production observed may be related to the rate of
energy to heat/ mass transfer (post shock) and are likely to reflect a transition from
vapour present (1) superheated melts and (2) superliquidus melts. A third and much
longer supersolidus phase (3) of volume reduction is thermodynamically constrained to
perhaps ~millions of years and has not yet been modelled, although all the likely
parameters involved are set out in a review of terrestrial magma ocean fluid dynamics by
Solomatov (2000). Crystal kinetics denies glass formation, even in the mantle, and
predcits grain sizes of ~mm scale (Solomatov 2000).
The large scale of the magma bodies result in long geological-scale time periods during
which conventional igneous petrology may be applied, including phase diagrams for
peridotite upper mantle appropriate to a range of ultrabasic to basaltic melt compositions
(from komatiite to basaltic). There is very little experimental data to quantify Stages 1
and 2, since most petrological experiments have been guided by the Earth geotherm,
and inhibited by convention to avoid conditions of superheat. With the recognition that
major sulphide ore-forming processes likely occur under large impact conditions, there is
an urgent need to design a new era of very high-T petrological experiments, in order to
derive new geochemical and mineralogical data appropriate to liquid-liquid fractionation,
and isotopic exchange within vapour-liquid-solid silicate systems.
FIGURE 3. Hadean MOHO A simple phase
diagram for pyrolite or chondritic mantle peridotite
shows the inescapable fact that magma oceans
may have triggered the first segregation of low
density feldspar-rich melts from higher density
feldspar-free peridotite, implying that the Moho
and feldspar-rich crust is a natural product of
impact magma oceans.
REFERENCES
Hansen V L (2006) JGR 111, E11010
Jones AP et al 2002 EPSL, 551-561
Jones A P et al 2005 Geol Soc Am, S Pap 388, 711-720
Jones A P 2005 Elements, 1, 277-281
Keays R R and Lightfoot P C 2004 Mineralogy and Petrology 82, 217-258
Solomatov V S (2000) Fluid dynamics in a Terrestrial magma ocean, in Origin of Earth and Moon,
Arizona Press, 323-338.