Magmatic deposits
Some ores form straight out of molten rock. As magma cools, heavy crystals sink, metal-hungry sulphide droplets separate, and the last dregs of melt collect the elements no crystal wanted.
- Explain compatible and incompatible elements from first principles
- Describe how crystal settling builds chromite and platinum layers (Bushveld, Sukinda)
- Explain why sulphide liquid separates from magma and scavenges nickel, copper and PGEs (Sudbury, Norilsk)
- Describe how kimberlites carry diamonds up from more than 150 km down (Panna)
- Explain why lithium, tantalum and rare earths end up in pegmatites and carbonatites
In the source–transport–trap framework from What makes an ore deposit, magmatic deposits are the simplest case. The source is the mantle or deep crust that melted. The transport is the magma itself. The trap is a physical or chemical separation that happens as the magma cools: something heavy sinks, something immiscible separates, or something unwanted is left behind.
Compatible and incompatible elements
A crystal is a regular three-dimensional framework with sites, or "slots", of particular sizes and electrical charges. In olivine, the slots are built for magnesium and iron ions, which carry a charge of 2+ and have a particular size. As a crystal grows from magma, it takes in only the ions that fit those slots well. Think of a car park built for compact cars: small cars park easily, buses are turned away.
That simple rule sorts elements into two groups:
Compatible elements
Ions that fit the sites in early-forming minerals are pulled out of the melt quickly. Nickel (2+, about the size of magnesium) slips into olivine, which ends up holding roughly ten times more nickel than the melt around it. Chromium (3+) goes into chromite and pyroxene. The remaining melt is rapidly stripped of them.
Incompatible elements
Ions that are too big, or carry the wrong charge, don't fit the common minerals and stay in the liquid. Potassium, rubidium and caesium are too large. Rare earths, zirconium, niobium, tantalum, uranium and thorium have awkward size and charge combinations. Lithium is small, but its single positive charge makes it a poor swap for magnesium.
Geologists measure this with a partition coefficient (written D): the concentration of an element in the crystal divided by its concentration in the melt. Above 1 means compatible, below 1 means incompatible. Follow the logic through a whole cooling magma and you get a powerful result. The more of the magma has crystallised, the more the incompatible elements are crammed into the shrinking volume of liquid that remains.
Pick how strongly crystals take up an element (its partition coefficient, D), then crystallise the magma. The plot shows the element's concentration in the melt that is left; the box shows the magma chamber, with each dot one "packet" of the element.
Below 1, the element is incompatible and piles up in the shrinking melt; above 1 it is compatible and is stripped out early. Push crystallisation past 90% with a low D and watch the runaway: that last, enriched melt is what makes a pegmatite. An idealised calculation (called Rayleigh fractionation): crystals are removed from the melt as they form, and D stays the same throughout. Real D values change with the minerals crystallising and the temperature; nickel's D in olivine is anywhere from about 5 to 20.
Crystal settling: chromite and platinum layers
In a large chamber of mafic or ultramafic magma, the first minerals to crystallise (see Bowen's series in Minerals & lithology) are olivine, pyroxene and chromite, an iron–chromium oxide. These crystals are denser than the liquid around them. Chromite is about 4.5–4.8 g/cm³; a basaltic melt is closer to 2.7. So they sink, like sand settling in a glass of water, and pile up on the chamber floor. This process is called fractional crystallisation with crystal settling, and the resulting crystal piles are cumulates.
Repeat this over thousands of years, with fresh pulses of magma being injected from time to time, and you build a layered intrusion: a body of igneous rock stacked like a layer cake, with each layer dominated by one or two minerals. Some layers are almost pure chromite, called chromitite. Others, often only a metre or less thick, are enriched in the platinum-group elements (PGEs): platinum, palladium, rhodium and their relatives.
- Remaining magma. Still liquid; crystals and droplets sink through it.
- Plagioclase-rich layer. Pale feldspar, which crystallises later.
- PGE reef (Merensky-type). A thin layer where sulphide droplets carrying platinum-group elements were caught among the crystals.
- Chromitite seam. Almost pure chromite, laid down after a pulse of new magma.
- Olivine–pyroxene cumulate. The first, densest crystals to settle.
- Sulphide pool: Ni, Cu, PGEs. Early droplets that sank to the very bottom and gathered in a hollow in the floor.
Bushveld Complex
The world's largest layered intrusion, about 2 billion years old and covering an area roughly the size of Sri Lanka. It holds most of the world's known platinum and chromium resources, plus large vanadium reserves. Two thin layers matter most: the Merensky Reef, a PGE-rich layer often under a metre thick, and the UG2 chromitite, which carries PGEs as well as chromite.
Sukinda valley
Ancient ultramafic rocks in the Sukinda valley host most of India's chromite reserves, mined from chromitite bands in open pits. Tropical weathering of the same ultramafic rocks has also produced a nickel-bearing laterite on top, a link to Sedimentary & surface deposits.
Sulphide liquid: oil and vinegar
A mafic magma can only dissolve a small amount of sulphur, roughly a tenth of a percent. If it takes in more, for example by melting sulphur-rich sediments it passes through, or if cooling and crystallisation lower how much it can hold, the excess sulphur does not form crystals. It separates as tiny droplets of a different liquid: a molten iron sulphide. The two liquids don't mix, just like oil and vinegar in a salad dressing. Liquids that won't mix are called immiscible.
Here is the trick. Nickel, copper, cobalt and especially the PGEs are chalcophile ("copper-loving"): chemically, they far prefer to bond with sulphur than to stay in the silicate melt. PGEs can be tens of thousands of times more concentrated in the sulphide droplets than in the magma around them. So each droplet acts as a sponge. As it sinks through the magma it soaks up these metals, and the more magma it meets, the richer it gets. Because sulphide liquid is much denser than silicate melt, the droplets collect in hollows in the floor of the chamber or in the narrow conduits magma flows through, forming magmatic Ni–Cu–PGE sulphide deposits.
How rich a droplet gets depends on how much magma it meets. Geologists call this the R factor: the mass of magma that a given mass of sulphide liquid has exchanged metals with. A droplet that meets little magma can only collect a little metal, however much it "wants" it. The stronger an element's preference for sulphide, the more magma the droplet must meet before it is full.
Slide the droplet down through the magma. The further it travels, the more magma it meets (R), and the more metal it soaks up. Bars show how close the sulphide is to the richest it could ever get for each metal.
Nickel and copper fill up once a droplet has met a few thousand times its own mass of magma. Platinum, which likes sulphide far more, keeps climbing until R is close to 100,000. That is why the richest PGE ores need droplets that "saw" huge volumes of flowing magma, as in a conduit or a big chamber. Illustrative starting magma: 200 ppm Ni, 100 ppm Cu, 10 ppb Pt. Uses the standard equilibrium formula C = C₀ × D(R + 1) ÷ (R + D) with fixed D values; real partition coefficients vary widely, especially for the PGEs.
Sudbury
About 1.85 billion years ago, a large meteorite struck what is now Ontario and melted a huge volume of crust. The melt sheet became saturated in sulphide, and the droplets sank to its base. Sudbury has been one of the world's great nickel camps for over a century: a rare case where the "magma" was made by an impact.
Norilsk–Talnakh
Linked to the Siberian Traps, enormous flood basalts erupted about 250 million years ago. The magmas flowed through sill-like channels and are thought to have picked up extra sulphur from evaporite sediments they cut through. Norilsk is one of the world's largest sources of nickel and palladium.
Kimberlites: express lifts for diamonds
Diamond is pure carbon arranged in a very dense crystal structure. It is only stable at enormous pressures, which under the old, cold cores of continents (cratons) means depths of more than about 150 km. At shallower depth, carbon's stable form is graphite. Diamonds near the surface survive only because converting back to graphite is extremely slow at low temperature.
Kimberlite is a rare, gas-rich ultramafic magma that forms deeper still, at 150–250 km or more. It rises explosively fast, probably in hours to days, ripping off pieces of the mantle on the way. Some of those pieces contain diamonds that formed long before, often billions of years earlier. Near the surface, the expanding gas shatters the rock and blasts out a carrot-shaped pipe. The diamonds are passengers, not products, of the kimberlite. Even a rich pipe contains under about one part per million of diamond by weight, and only a small fraction of known kimberlites carry enough to mine.
Diamond is stable only below the dashed line. Switch between an old craton and ordinary, younger lithosphere, and see whether the cold, rigid plate reaches deep enough to store diamonds.
Schematic depths. The graphite–diamond boundary sits deeper where the mantle is hotter, because diamond needs more pressure to stay stable at higher temperature. Kimberlite melt forms below the plate either way; whether it carries diamonds depends on what it passes through on the way up. That is why diamond explorers focus on cratons.
India's diamonds have a long history. For centuries the world's diamonds came from river gravels along the Krishna valley in the Golconda region. Today India's only large mechanised diamond mine is at Majhgawan, near Panna in Madhya Pradesh, run by the state miner NMDC. The Majhgawan pipe is usually classed as a lamproite, a close cousin of kimberlite that carries diamonds up in the same way.
The last dregs: pegmatites and carbonatites
Go back to the explorer at the top of this module, set D near zero and push crystallisation past 90%: the element is 10× enriched at 90% and 20× at 95%. When a granite magma is almost entirely crystallised, the last few percent of liquid is loaded with everything the crystals turned away: lithium, beryllium, tantalum, niobium, tin, caesium, rare earths. It is also rich in water, fluorine and boron. These make the melt runny and let atoms move fast, so the crystals that finally grow can be enormous, centimetres to metres across. The resulting very coarse-grained rock is a pegmatite, usually found as dikes and pods near the top of a granite or in the rock around it.
A carbonatite is something stranger: an igneous rock made mostly of carbonate minerals, like a molten limestone. Carbonatites form from very small degrees of melting deep in the mantle, which concentrates incompatible elements from the start, especially the rare earths and niobium. Only one volcano on Earth erupts carbonatite lava today, Ol Doinyo Lengai in Tanzania.
Greenbushes
A huge ancient pegmatite in Western Australia, mined first for tin and tantalum and now one of the world's largest hard-rock lithium mines. The lithium is in spodumene, a lithium–aluminium pyroxene. India's Koderma belt in Jharkhand is a classic pegmatite province, long famous for mica.
Mountain Pass and Bayan Obo
Mountain Pass in California is a carbonatite rich in the rare-earth mineral bastnäsite. Bayan Obo in Inner Mongolia, the world's largest rare-earth deposit, is linked to carbonatite magmas, though exactly how it formed is still debated.
Magmatic deposits supply most of the world's chromium and platinum-group metals and much of its nickel, used in stainless steel, batteries and catalytic converters. Exploring for them means finding the right mafic and ultramafic intrusions, and the ideas above explain where inside those intrusions the ore settles.
The rocks that host magmatic ores are mostly mafic and ultramafic: olivine- and pyroxene-rich rocks that weather fast and are full of magnesium and calcium. That makes them interesting for carbon removal too. Chromite and nickel mines produce large volumes of crushed ultramafic waste, and the compatible-element logic above explains why: those feedstocks often carry nickel and chromium, which need to be measured and managed when spreading rock on farmland.