-
Some questions of basalt petrogenesis
1. How are tholeiites produced? Experimental petrology, trace
element modeling and isotope data indicate that tholeiitic magmas
are produced by relatively high degrees of melting (20-25 %) of
normal mantle, i.e., mantle that can range from depleted (MORB) to
slightly enriched (some ocean island and island arc
tholeiites).
2. How can we explain the range of compns from picritic
tholeiites to quartz tholeiites to high alumina tholeiites? After
making corrections for the effects of fractionation, i.e.,
comparing only primary magmas, experimental studies show that
increasing pressure increases the MgO content of the melt produced.
At 25 Kb basalts with MgO contents of 16 wt. % are formed, 14.5 wt.
% MgO at 20 kb and 12.5 wt%. MgO at 15 kb Quartz tholeiites are
common in the CRB province and appear to be the result of
fractionation.
3. Can alkalic basalts be formed by fractionation of tholeiites
or vice versa? Certainly not at low pressure because of the thermal
divide, i.e., cant go from nenormative to hyp or qz normative or
vice versa
1 (1058C): L1 (Fo, Ne, Ab)2 (1118C): L2 (Fo, Ab): thermal
maximum3 (1070C): L3 (Fo, En, Ab). Reaction at 3? 4 (1062C): L4
(En, Qz, Ab)
What happens when Ca is added to the system?
However, the olivine-cpx-plag join remains a thermal barrier
OK, but what happens at P is increased? Look at point 3.
Fo
Ne Ab Qz
En
QzEn
Fo
sp
Ne Ab1 2 3
4
Fo-Ne-SiO2system at P = 1 atm At 3: Fo + L En + Ab
(peritectic)
Diopside (cpx) becomes stable and feldspar is now plag ss
-
Somewhere between 1 atm and 1 GPa (10 Kb) the invariant reaction
equilibrium crosses the En-Ab join and then the Fo-Abjoin. So
what?
The Fo-Ab join is no longer a thermal barrier so, in theory, at
higher pressure subalkalicmelts could fractionate to alkalic
melts.
As P is increased further, what happens?
Liquids in equilibrium with Fo (Ol) En (Opx) Ab (Plag or
Spinel/Gnt at higher P) become increasingly alkalic (ne
normative)
Tentative conclusions: (1) melting of mantle at higher pressures
favors the formation of alkalic basalts (2) melting at low P favors
melts richer in SiO2
What is the effect of adding volatiles to this system?
CO2 at 20 Kbdry at 20 Kb
H2O at 20 KbAddition of H2O tends to produce more silica-rich
basalts Addition of CO2 tends to produce more alkali basalts.
Basalt petrogenesis (cont.)
Effect of pressure (dry) on the Fo-Ne-SiO2 system
FoEn
-
As discussed earlier, pressure also has an effect on the MgO
content of partial melts.
Blue: L (Fo, Diss, Enss) at P = 1 atm
Red: L (Fo, Diss, Enss) at P = 3 GPa
Some general conclusions:
Smaller degrees of partial melting favor the formation of
alkalic basalts because Na behaves primarily as an incompatible
element (DNa
-
Trace elements provide a test of fractionation models. For
example, is it likely (or even possible) to derive alkalic basalts
from a subalkalic parent? We concluded that its at least
theoretically possible at higher pressures. Lets assume that the
data presented below represent magmas that we think might be
related to each other by crystal fractionation. Could an alkalic
OIB be derived from a tholeiitic MORB. As an example, lets use the
Rayleighequation to model Rb. Reasonable value for DRb for MORB
phenocryst assemblages is ~0.03. For a Rb increase from ~2 to ~90,
FL ~ 2%. In other words, ~98% crystallization required.
Basalt petrogenesis (cont)
)1( = iDLOi
Li FCC 98% crystallization of a parental basalt would produce
a
highly evolved liquid with extreme Fe-enrichment. It is
possible, in principle, to produce such a liquid but this would not
be an alkali basalt.Rayleigh equation
-
Basalt petrogenesis (cont)
Q: If tholeiitic basalts and alkalic basalts are not related via
fractionation at any pressure, could they be derived by different
degrees of partial melting of a common mantle source?
A: Perhaps, but each case must be considered on its own merits.
First requirement would be that they have identical initial Sr, Nd
and Pb isotopic ratios (assuming no assimilation). Second, major
elements must be consistent with experimental petrology results.
Third, trace elements must also be consistent. Example: Is it
possible that the OIB and MORB shown on previousslide could be
derived from a common mantle source?
MORB rare earth element pattern clearly shows that MORB mantle
source was depleted. Why?
Careful trace element modeling could tell us if it was possible
to derive the OIB from the same depleted source. To do this one
would have to (1) assume a melting model, (2) know all the relevant
distribution coefficients, (3) decide what pressures to use in the
model, (4) know (or assume) the stoichiometry of the melting
reaction
Quick feasibility test: Can we model La abundances? Using the
batch melting equation, and assuming DLa ~ 0.01, we can reproduce
the ratio of La in OIB to that in MORB (~12) with ~20% melting to
produce MORB and 1% melting to produce OIB. Is this reasonable?
More detailed modeling involving a variety of trace elements might
provide a more definitive answer. However, many petrologists would
argue that it is difficult (maybe impossible) to segregate such a
small melt fraction from its crystalline residue.
))1((1
iiLOi
Li
DDFCC
+=
-
Basalt petrogenesis [summary of comprehensive model proposed by
Green and Ringwood (1967)]
Valiant effort but this model, which was developed before we had
abundant trace element and isotopic data, has problems.
-
Partial melting experiments on depleted lherzolites. Dashed
lines represent percent partial melt produced. Curved lines
shownormative olivine content of the melt. Opx out and Cpx
outrepresent the T&P at which these phases are completely
melted. After Jaques and Green (1980). CMP 73, 287-310.
Results: (1) Tholeiite: ~10 to ~30% partial melting with more
silica-rich types at lower P
(2) Difficult (maybe even impossible) to get alkalic basalt from
a depleted mantle source
While it seems possible, at least in principle, to get a wide
range of basalt types from a uniform mantle source over a range of
P, T, and melt fractions, we already know from xenoliths that the
mantle is inhomogeneous, albeit over a fairly narrow major element
compositional interval.
Basalt petrogenesis (cont)
We have fertile (?primordial) mantle, depletedmantle and
depleted mantle later enriched.
Caveats: 1. Its the depth (pressure) of magma segregation from
its crystalline residue that is the important parameter.2.
Experiments cited assume some sort of equilibrium (batch) melting
process. Recent work indicate that fractional melting may be the
norm.
In a series of classic experiments, Jaques and Green (1980)
carried out extensive high P/T experiments on depleted and enriched
peridotites.
-
Partial melting experiments on enriched lherzolites. Dashed
lines represent percent partial melt produced. Curved lines
shownormative olivine content of the melt. Opx out and Cpx
outrepresent the T&P at which these phases are completely
melted. Shaded field outlines PT conditions for alkali basalt
generation. After Jaques and Green (1980). CMP 73, 287-310.
Basalt petrogenesis (cont.)
(1) Tholeiites are generated over a wider pressure range than in
the case of depleted lherzolite
Partial melting experiments on enriched (metasomatized)
lherzolites
(2) Alkalic basalts and basanites are generated over a
relatively limited range of pressures and at significantly lower
degrees of partial melting (near solidus melts).
Results:
900 pound gorilla: What is the influence of volatiles on mantle
melting and how do we produce weird magmas like kimberlites,
alnoites, carbonatites, lamproites, to name just a few.
-
Experiments on MORB thought to be a primary magma [after Fujii
& Kushiro (1977). Carg. Inst. Wash Yearbook].
Basalt petrogenesis (cont.)
Experiments on a MORB primary magma aimed at determining the
pressure and temperature at which it segregated from the
crystalline residue. Inverse modeling.
A magma is said to be multiply saturated if several minerals (at
least two and better still three or four) crystallize from the
magma at the same T and P (or at least within a few degrees of each
other).
The inference drawn is that this magma was in equilibrium with
these minerals at this T and P and therefore this is the T and P at
which the melt segregated from the crystalline residue. For mantle
melts the minerals in the residue would be olivine + opx cpx
garnet/spinel/plag.
If the magma is not primary or if the melting process is
fractional rather than equilibrium then these experiments dont mean
much
Multiple saturation point