Thermobarometry is the art of inferring Pressure and Temperature conditions by relating the composition of a metamorphic rock to a thermodynamic model, assuming chemical equilibrium was achieved. That’s the single sentence summary, but as usual, things are more complicated if you look in more detail. I’ve discussed elsewhere reasons why chemical equilibrium may not be achieved – in summary, where the rocks are not hot for long enough, or where reaction kinetics (the details of how minerals grow)  control metamorphic mineral growth. Can we at least be certain of the composition of the rock? That must stay the same, surely?

Nope.

A major assumption when making thermobarometric calculations is that you are dealing with a closed system, that you have the same bunch of atoms at the end as you do in the beginning. Let’s think about this assumption of a ‘closed system’ in more detail. A first complication to mention is that metamorphic petrologists tend to make an exception for fluids. We don’t expect a rock that’s been through chlorite breakdown to leave a puddle on the desk – we know the water produced by the reaction is long gone (but we do account for it in our thermodynamic models).

Consider a closed system – how big is it? What scale are we talking about? Metamorphic mineral growth usually requires the movement of atoms, from a phase that is breaking down into one that is growing. So on the scale of a mineral grain we are not dealing with a closed system at all. To do metamorphic petrology though, we assume that the system is closed on the scale of a rock sample, or maybe thin-section. This seems reasonable if you think about metamorphosed sediments. Some sediments can’t make their minds up, are they are sandstone, or a mudstone? They end up being a bit of both, with alternating layers of dramatically different composition. When such rocks get metamorphosed, they still have layers, the compositional differences haven’t been completely homogenised. In a gneiss, centimetre scale layers in the sediment protolith aren’t preserved but metre scale layers often are. This is despite the intense ‘mixing’ that comes from the multiple phases of deformation these rocks have often enjoyed.

I’ve not presented any real proof here of course, just some hand-waving. Evidence that metamorphic rocks are not completely homogenised on a metre scale doesn’t mean that they are completely closed systems. Maybe only quartz doesn’t move, but other atoms zip around like billy-o over great distances?

Let’s think about things from the other side. What mechanisms could drive open system behaviour over >1m scale? How do you shift atoms around within solid rock?

Firstly, let’s get granite out of the way. If you heat rocks enough they melt and you get migmatites. Sometimes the pockets of melt manage to meet up and the melt is able to move up and away. Granite plutons can be huge, so this is an important mechanism of mass transfer on a crustal scale. Like with fluids earlier, this is something metamorphic petrologists have a handle on. The whole concept of restite in migmatites is an acknowledgement that the composition of the rock has changed. The reactions that create the melt can themselves be modelled, and so on.

Metasomatism is the chemical alteration of a rock. It is applied to cases where the composition has clearly been altered, which is what distinguishes it from metamorphism. Classic examples are skarn, or greisen, where rocks in contact metamorphic aureoles have their composition dramatically altered by hot fluids carrying other elements. Cases where elements such as gold are dissolved in fluids and then transported and concentrated are of obvious interest to economic geologists.

Around granite plutons the effects can be dramatic, but fluids are known to be present in many types of metamorphism. Indeed the presence of a fluid phase is an important factor in reaction kinetics, as a means of moving atoms around to help drive metamorphic reactions. This being the case, why should this only occur on scales of <1m? What if metasomatism was the norm in metamorphism, only with effects too subtle to spot?

Its an interesting possibility and some recent research is finding evidence that some textures previously interpreted as metamorphic are actually metasomatic. This is worrying for thermobarometry – if mineral growth is driven by fluid-induced compositional change, rather than by changing P-T conditions, then standard estimates of P-T may well be wrong.

Is this is a big problem? Well, thermobarometry is proven to be a useful tool for understanding mountain belts, but it’s a reminder that our understanding is incomplete. I think of my own PhD field area. The syntectonic gabbros had large areas of ‘Acid Gneiss’. This rock-type was rich on quartz, but also contained amphibole and calcic plagioclase. It is most likely a metasomatic rock, with silica being added to a gabbro host. An area of ‘andesine porphyroblastic gneiss’ in the aureole was suspiciously homogeneous and hard to link to a sedimentary protolith. Hmmmm. There may yet be a fantastic case-study of metasomatism to be made in these rocks, yet in my thesis I merely describe and move on. I didn’t have any conceptual or analytical tools available to me to make anything more of them.  My feeling is that one day someone will come up with a technique (stable isotopes? fluid inclusions?) that will revolutionise our understanding of hot rocks and show quite how far from being closed systems they really are.

One last quirk. Even if a rock sample is a closed system, can we simply take its bulk composition and plug it into THERMOCALC? Think of the classic Mn profile in Garnet, with lots in the core. Once it is in there, it is ‘locked away’ and doesn’t take part in metamorphic reactions, (until the garnet is broken down or higher temperatures allow the Manganese to diffuse away). Until this happens, metamorphic reactions are taking place in a system with low-Mn, compared to the bulk composition. This is a general problem. We know that systematic compositional variations in metamorphic mineral grains are common. This causes an effect analogous to fractional crystallisation in cooling magma – the composition active in metamorphic reactions (‘effective bulk composition’) changes over time. What’s a metamorphic petrologist to do? Dave Waters has a fine summary of these issues of effective bulk composition. In summary, study your rocks carefully and think carefully.