Magmatic Ore Deposits

Magmatic ore deposits are deposits where the metal is found in the primary magma, and
as such is the most basic (least evolved) type of deposit. The most common elements
mined in magmatic deposits are nickel (Ni), chromium (Cr), and the PGEs (Platinum
Group Elements): platinum (Pt), iridium (Ir), palladium (Pd), osmium (Os), ruthenium
(Ru) and rhodium (Rh). Of these, Pt, Ir and Pd are the major money makers. Nickel and
the PGEs are chalcophile elements, that is, they are sulfur-loving elements. All of these
elements are hosted in mafic and ultramafic rocks such as basalts, gabbros and
komatiites. Granites contain almost no PGEs (certainly not in exploitable quantities)
and sediments do not contain any at all.
The metals
Platinum. The main use for platinum is in catalytic converters in cars, bonding with CO
and thereby purifying the air. It is extremely insoluble in room temperature fluids (like
gold), malleable, and very resistant. Due to its high melting temperature (~1700oC),
platinum is also used to make high temperature containers, such as crucibles, in which
other materials can be heated.
Chrome. Found in chromite, a spinel with a composition FeCr2O4 . Chrome is a very
important industrial metal. It insoluble, doesn't rust, and is only found in magmatic
deposits, being a 100% mantle component.
Mantle Melting
Ultramafic rocks are the result of mantle plumes or hot spots (e.g. Hawaii). Mantle
plumes occur as heat rises from the core (this is still debated) through the mantle until it
reaches a pressure where melting is able to occur.. As the plume rises it is able to melt
more and more of the mantle, with melt travelling up through permeable unmelted
mantle. Diffusional heat loss is very minimal compared to ascendancy. Note that mantle
melting is not batch melting - you don't take a chunk of the mantle and melt it 100%.
Only a few percent melts at a time, so that the composition is constantly changing.
Some of the crust is assimilated into the melt as it rises, changing its composition. The
Sr87/86 ratio can be used to determine the age of the intrusion as well as how much of the
crust has been incorporated into it. As the melt ascends through the crust, the ratio
increases.

A very high degree of mantle melting is required to form PGE deposits. The amount of
MgO in a melt can tell you the how much melting took place since the it is directly
correlated with temperature. Mg-rich rocks indicate high melting temperatures and a
high degree of melting (see olivine phase diagram). PGEs are found in olivine and
sulfide minerals. Platinum sulfides, unlike many sulfide minerals, melt at very high
temperatures.
Some Important Magmatic Ore Deposits around the World...
1. Bushveld: The Bushveld Complex in South Africa covers an area of 67 000 km2
(the size of Ireland!). It is the largest LMI deposit and also provides most of the
world's platinum. Besides being the best PGE deposit, it also contains a lot of
chromite.
2. Kambalda: nickel deposit, Australia.
3. Sudbury: nickel and PGE deposit, Canada.
The Distribution Coefficient
D is the distribution coefficient, a ratio of how many x there are a to b. The notation is:
Dxa/b .
For example, if x = rocks, and Anne has 2 rocks while Bob has 1, then Drocksanne/bob =
Drocks2/1 = 2/1 = 2. It is important to define what you are considering and in what ratio.
For example, DPtsulfide melt/silicate melt = 105 means that platinum prefers the sulfide melt
many, many times to the silicate! So in the sulfide melt it is 105 while in the silicate it is
only 1. To put it in perspective, this means that you would find 10 ppm (parts per
million) Pt in the sulfide to every 1 ppb (pp billion) in the silicate. Iridium is similar DIrsulfide melt/silicate melt = 104. This shows that you don't need much sulfide to strip the PGEs
out of a silicate melt.
Solubility
Solubility is the ability of a substance to dissolve / the amount it dissolves. It is a
function of temperature (T), pressure (P),
composition (X), and oxygen and sulfur fugacity (fO2, fS2). For example, there is a
specific amount of salt you can dissolve in a given amount of water at a specific set of

conditions. At room temperature, for example, the composition may be 25% salt, while
at much higher temperature you can have up to 90% salt (see diagram). Saturation is
reached when no more salt will dissolve in the water. Like salt, sulfur has a solubility
which is a very important factor in forming deposits. Once a silicate melt is saturated in
sulfur, any added sulfide cannot dissolve in the silicate, and so you end up with a
silicate melt and sulfide.

Immiscibility
Immiscible liquids do not mix. A classic example is oil and water (see diagram above).
Cooling from the one liquid field, two liquids appear once you hit the solvice. These
become more pure as you cool further.

Layered Mafic Intrusions (LMI)

There are only two LMIs in Australia. LMIs form as a result of rifts or hot spots (mantle
plumes). Layered mafic intrusions are formed when minerals crystallise (fractional
crystallisation) out and settle in compositionally distinct layers in a magma chamber.
Crystals that form on chamber walls slump and form cross bedding in the chamber. This
is useful because it is easy to see which way is up! Another feature of LMIs is cryptic
layering, where you can see layering but not compositional changes (except in thin
section). The cumulate nature of the deposit means that there is a lot of crystal-magma
interaction. Melt is trapped in the spaces left between crystals. Sometimes this has the
effect of homogenising the surrounding minerals or forming zones on the minerals. In
other cases, very primitive melt is trapped and undergoes its own little fractionation.
PGEs are only incorporated into olivine (e.g. in the mantle) at high pressure, and at
low pressure are incompatible. However, they are chalcophile elements and prefer to
form sulfides. As the melt is fractionated it is enriched in sulfur, and two immiscible
liquids form - liquid silicate and liquid sulfide. The platinum sulfides then drop out and
are concentrated.

Sometimes the system is a little more complicated. The normal Sr87/86 trend (below left)
observed in a magma chamber is sometimes repeated. This is an indication that there
were many pulses of magma into the magma chamber.

Several minerals change composition from the bottom to the top of the magma chamber.
One way to note the compositional variation in the magma chamber is the Fe/Mg ratio.
As crystallisation progresses, minerals (e.g. olivine) become more Fe-rich (forsterite)
(those that crystallised first were more Mg-rich (fayalite), having formed at a higher
temperature). The final minerals are sometimes so iron rich that rust forms at the top of
the chamber. The composition of plagioclase feldspars also changes with crystallisation.
As you go up in the layers in the chamber, less and less anorthite (Ca-rich) is formed,
and more albite (Na-rich) crystallises.

Steps in the formation of magmatic sulfide deposits:

1.

Generation of mafic / ultramafic magma in mantle (PGE content in melt is

proportional to the degree of partial melting).
2. Ascent and emplacement of magma in crust:
a. thermal and / or chemical interaction with country rocks
b. crystallisation preceded, accompanied or followed by

c. precipitation of immiscible sulfide liquid droplets (mainly Fe-S),

collector for Na, Cu, PGE
d. concentration of sulfide mainly gravitational (p (density) sulfide liquid ~
4, p silicate liquid ~ 2.7 - 3.1) - pyrrhotite (Fe1-xS, pentlandite (Ni,Fe)9S8
(main Ni mineral), chalcopyrite (CuFeS2), pyrite (FeS2, worthless),
discrete PGE sulfides... P+S, PdS (very rare).
Forming chromite deposits
To form a chromite deposit you have to saturate in chrome. Basalts usually have some
chrome in them but are not saturated. The way to do this is to mix melts. The first
(primitive) melt is fractionated (A to D), then a new magma pulse introduces more
primitive melt A to mix with D and so chrome saturation occurs. Chromite crystallises
first out of the melt, and after that the melt is depleted in chrome. If you then add
another Cr-rich melt to the remaining melt, you crystallise chromite again. Chrome
doesn't go into sulfides.

In the magma chamber

In the magma chamber, things are layered by density (composition) before they
crystallise, forming layered magma. After a while, with heat from the bottom, each layer
convects. This is called Double Diffusive Convection. There is no mixing of layers.
Each layer that convects is called a double diffusive convection cell.

So there is both chemical stratification and temperature stratification within the

chamber. Normally, when you heat something up, you would expect it to become less
dense and rise (for example, through the layer boundaries). However, the chemical /
compositional influence on density is far greater than the effect of temperature in this
case, and the magma remains stratified. In addition, each layer can fractionate within
itself. Any new pulses of magma are first added to the middle of the chamber, being of
average density. They are the hottest but are not yet stratified. They are later added to
the layer they are most suited to by density.

The 'R' Factor

The R factor is the ratio of silicate to sulfide, specifically, the ratio of the mass of
silicate to the mass of sulfide. To make a decent sulfide deposit, you need to strip heaps
of Pt from the silicate component. For example, if you have an amount of melt, half of
which is basalt and half which is sulfide melt, and all the Pt goes into the sulfide from
the basalt, you still wouldn't have a very good deposit, because there's not enough Pt in
the basalt for that amount of sulfide. If, however, you have a lot of basalt, and a little bit
of sulfide, then you can sufficiently enrich the sulfide in Pt to make a reasonable
deposit. Thus high R's are needed to make good deposits (i.e. lots of silicate and little
sulfide). Repeated bursts of silicate around the same sulfide also enriches the sulfide. As
you get to higher D's, you need higher R's to enrich the sulfide. Here is an example
situation:

DPtsulf/sil = 106; initial concentrationPt = 10 ppb

Amount of metal in sulfide = D x initial concentration x (R+1) / R
Ptsulfide = 106 x 10 ppb x (R+1) / R
So you need about 1000x more silicate than sulfide to begin stripping Pt from it. At
around 107x silicate, you reach saturation of the sulfide. If you have very low initial
concentrations, and very high D's, you need very high R's. So there are some important
factors to consider: the initial concentration; the distribution coefficient and the
eagerness of the metal to go into the sulfide. To form nickel deposits, for example, you
don't need very high R's , because the initial concentration of the metal in the silicate is
much higher. DNisulf/sil = 250-750.
Kambalda
Kambalda is a nickel deposit formed from a komatiite, a very ultramafic (MgO > 25% 35%) magma that melts at about 1600oC (1 atm). In comparison, basalt contains ~ 12%
MgO and melts at 1200oC (1 atm). These magmas existed over 2Ga, when the mantle
was hotter (they do not form today, indicating that the earth is cooling). They are found
in Archean Shields and are all between 2.4 and 3.2Gyrs old. Bowen (N. L. Bowen) said
that these sorts of melts could not exist at the surface (because then everything would
melt) and called them cumulates. However, he did not know about plumes, i.e., that the
whole mantle doesn't have to be that hot! Komatiites start deep in the mantle and ascend
adiabatically through it. These high MgO melts do not quench to a glass like basalt,
however, they form a spinifex texture. They are not viscous and flow like water over the
earth's surface. Because they are so hot, they melt channels into the ground rather than
forming flows on the surface:

The channels are important. Once they are established, all the magma flows through
them like a stream (lava pipes also form, with magma cooling around the outside to
make the pipe while the rest flows through the centre). Assimilation and contamination
are particularly important with respect to a komatiite, because it melts everything it
comes in contact with (everything melts at this temperature!).
Sulfur Saturation
If you bring up a komatiite, it is initially under saturated in sulfur. There is a little sulfur
in the sediments which is assimilated into the melt as it rises, enriching the melt in
sulfur. The sulfur then drops out as sulfides. As subsequent flows go by, these sulfides
strip out metals from the magma (very high R value - lots of silicate, constantly
replenished, to a small amount of sulfur) and form deposits. High MgO melts have a
high nickel content (100's ppm) because a lot of Ni-rich olivine is melted. Once melted,
however, Ni doesn't return to olivine and is easily stripped out of the melt into sulfides
at the surface.

Sulfur saturation is a major factor in forming ore deposits. When the melt is saturated in
sulfur, sulfides can drop out to form the deposit. At depth, where a melt forms in the
mantle, there is a limit to how much sulfur can be dissolved in it. The melt may be
saturated in sulfur at high pressure with 300 ppm S. At the surface, however, the melt
may need 1000 ppm to be sulfur saturated. So as the melt rises it becomes hugely under
saturated in sulfur due to the decrease in pressure (saturation is a function of pressure,
among other things).
Ways to saturate in sulfur. The amount of sulfur in the melt depends on iron, fO2, fS2 :
FeO + 1/2S2 = FeS + 1/2O2
fO2 is a function of the activity of oxygen as a function of temperature. The oxidation
states of iron, in increasing oxidation state, are as follows: Fe (iron) - FeO (wustite) Fe3O4 (magnetite) - Fe2O3 (hematite).

The lines in the diagram above are buffers and are much more important than the fields.
Both minerals on the buffer are produced. For any given temperature, adding oxygen
increases oxidation. If you are in the Fe field, for example, and keep adding oxygen,
you will keep making FeO until the Fe runs out. Then you will start to make Fe3O4 .
All these melts are contaminated by crustal material. Contamination changes the
composition of the melt. FeO is vital to this and to sulfur saturation. Sulfur saturation is
affected by:
1.
2.
3.
4.

contamination
cooling
fractionation
pressure

Simple fractionation decreases FeO in the melt because you crystallise olivine - (Mg,
Fe)2SiO4 - which in turn increases sulfur saturation as you move to a cotectic where an
immiscible sulfur liquid exists. As FeO decreases, the solubility of S decreases i.e. as
you increase FeO in the melt you increase S in the melt. As fO2 increase, sulfur
solubility decreases. You can decrease FeO by forming Fe2O3. SiO2 also has an effect
on sulfur saturation. As SiO2 increases, the solubility of S decreases (hence sulfur
saturation increases). The more silica there is in a melt, the more polymerised the melt
(viscosity is a function of how polymerised the melt is).

If the melt is polymerised, there are fewer places for sulfur to attach to, and so sulfur
saturation occurs. In addition, adding silica means that FeO and MgO (and others) are
diluted. At Kambalda, the assimilated material is both S and SiO2 rich, which enables
sulfur saturation. An average basalt contains 200 ppm S, while a komatiite contains 600
ppm S. You can saturate any melt is sulfur not by adding more sulfur but by
fractionating the melt, that is, crystallising sulfur-free phases. As you do this, sulfur (and
sulfur-loving metals) becomes concentrated in the melt.
Lever Rule

The Lever rule allows you to find the compositions of melts on ternary phase diagrams.

The composition of the melt at x:

the amount of A is b/a+b
the amount of B is a/a+b
Sudbury
The Sudbury nickel deposit is in Ontario, Canada. Canada is most like Australia in that
both countries have very old rocks, Canada having the oldest dated rocks (~4Ga). The
Sudbury Igneous Complex (SIC) is a 50 x 80 km ellipse. The whole structure, the result
of a meteorite impact, is about 1.8Gyrs old. 'Shatter cones', radiating outward blast
marks and very high pressure quartz polymorphs ('shot quartz' - stichovite) are evidence
of this.

Sudbury is the largest Ni deposit (not originating from the meteorite), containing 1
billion tonnes of ore. The Onaping Tuff, the top layers of rock, looks like an ordinary
volcanic tuff but it is actually the fallback breccia from the meteorite impact. The
Sudbury breccia, under the Ni deposit, is made up of pulverised gneiss (further evidence
of high energy events). The nickel deposit formed as a result of the impact, which had
an effect all the way to the upper mantle. The reason the structure is oval shaped and not
round, as you would expect, partly because the crust has stretched in the last 1.8Gyrs
and partly because it was probably a glancing blow in the first place (and so slightly
elliptical).
The impact, reaching to the upper mantle, caused peridotite melting by the sudden
pressure loss and adiabatic ascension. The ultramafic, PGE, Cu, Ni - rich melt rose
towards the surface, assimilating crustal material. As the melt rose, it decompressed and

expanded, forming the funnel or tulip shape. It is likely that not only is mantle melt
present, but that a chunk of the upper mantle was involved in forming the deposit
(fractionated). This could be one of the reasons why the deposit is so Ni-rich. Another is
its sheer size. Pods of ore exist with the sulfides at the bottom of each pod. Most
sulfides are at the sides of the deposit (along the walls of the tulip structure) rather than
in the centre. The main ore consists of pyrrhotite, pentlandite, and chalcopyrite. It is ~
1% Cu, 1% Ni, 10 ppm PGE.
Factors controlling PGE content of a sulfide melt
1. Percentage partial melting of the mantle (a few % forms a basalt, 50% forma a
komatiite). Ni and PGEs live in olivine so a lot of melting is necessary to
acquire these metals. Cu, Pd and Pt in the mantle are also in sulfides which are
very hard to melt. Pt and Pd linger in residual sulfides. Ultramafic magmas are
high in Mg, Ni, PGE, mirrored in the composition of associated sulfide ores.
MORBs average < 0.7 ppb Pd, < 0.06 ppb Ir, while continental basalts average ~
6 ppb Pd, 0.1 ppb Ir. MORBs are low in incompatible LILs and LREE because
they are generated from previously depleted mantle.
2. Silicate melt - sulfide liquid equilibrium.