Mantle Migmatites

Abstract

Textbooks on igneous petrology explain a range of primitive basalts and more alkaline volcanic rocks such as basanites, nephelinites and melilitites as products of various degrees of partial melting of mantle peridotite – a supposedly uniform rock consisting of four minerals olivine, ortho- and clinopyroxene and either spinel or garnet, depending on depth. However, all is not well in igneous petrogenesis: many more alkaline rocks such as lamproites, lamprophyres, kimberlites and carbonatites do not fit this homogeneous peridotite mantle model. They require other ultramafic rocks or mixed rock source regions; however, they are volumetrically relatively insignificant and so the discrepancy is often ignored by most people who use the standard peridotite model. This simple peridotite model is widely used in neighbouring specialities that do not consider complex source regions.

The solution is that the mantle source regions of magmas are more complex, containing veins and dykes with abundant hydrous minerals that have accumulated over the eons. The results of episodic melt infusions are mantle migmatites, the details of which have been hidden behind the shroud of vague labels such as mantle metasomatism and mantle heterogeneity that have resisted attempts to specify what they mean for decades.

Recently a role for mixed rock source regions containing pyroxene derived from deeply recycled eclogites has become accepted, but the need for hydrous and carbonate-bearing minerals in rocks distinct from peridotite has only recently emerged as a necessity. Recent experimental work shows that these hydrous ultramafic rocks melt before peridotite (including hydrous peridotite) in mixed sources under most geodynamic conditions and so will be components, however diluted, in almost all mantle derived melts. This includes voluminous types such as picrites and komatiites that are mostly derived from peridotites by further melting. The veins and dykes – the neosomes of mantle migmatites – are minor components of the mantle but nevertheless important because they always melt first and so cannot be ignored. Many voluminous basaltic rocks are thus a kind of geological palimpsest; they still bear evidence of events that are swamped by later melting of peridotite.

My career has overlapped to a large degree with the development of the mixed source concept and so this review tracks the history of relevant areas that I have worked on: mantle derived alkaline rocks, mantle petrology, mantle metasomatism, and experiments on peridotites and hydrous ultramafic rocks.

After a brief career history (Section 2) I review the different types of alkaline rocks that may originate by melting of hydrous ultramafic rocks and carbonates in the mantle (Section 3). They cannot be simple melting products of peridotite alone, not even if the peridotite is “metasomatised”. I point out similarities and differences between these alkaline rocks for the non-specialist and to illuminate some of the semantic problems surrounding these rocks, some of which have changed their names more than once. This semantic confusion can bewilder the non-specialist, so it is not surprising that many would rather ignore them and retain the simple peridotite model.

Section 4 reviews the evidence for the variety of mantle source rocks, mostly from mantle xenoliths. Amphiboles and micas are common, usually with clinopyroxene. Other minerals such as the Ti oxides (ilmenite, rutile and perovskite), as well as apatite and carbonate may be abundant in mantle dykes despite usually being thought of as accessory phases.

The history of experiments at mantle pressures (Section 5) shows that the concept of mantle metasomatism followed the first experiments on peridotite remarkably quickly in the 1970s. This may be part of the reason that alkaline melt origins were sought simply by adding H₂O and CO₂ to peridotite melting experiments. Liquidus experiments on alkaline rocks were common, but failed to find the desired melting conditions of volatile-rich peridotites. Melting of volatile-rich peridotite never worked well as an explanation, but most were satisfied with ad hoc modifications, leaving only few people to continue working on explaining the strongly alkaline rocks.

Experiments on the melting of hydrous ultramafic rocks (Section 6) – the mantle migmatite neosomes – have mostly been published since 2010. The results we now have show that rocks with amphibole or phlogopite will usually melt to a substantial degree before peridotites begin to melt and so will dominate the early melt compositions. Comparing experimental melts to natural rocks (Section 7) shows that melting of amphibole-rich dykes produces sodic alkaline melts akin to nephelinites and melilitites, whereas mica melts more slowly, resulting in high K₂O/Na₂O in melts that resemble lamproites.

There is a huge variety of potential dyke rocks in the mantle, but few of these have yet been subjected to melting experiments. In the future, more experimental work should consider carbonate-rich assemblages, Ti oxides and sulfides, and rocks with both calcic amphibole and phlogopite. The mineralogy of reaction zones between infiltrating melts and peridotite wall rocks has come into focus, but the range of melts produced by re-melting of these reaction assemblages have not yet been studied in detail.

Using the analogy of thin section microscopy, it is time to switch lenses to a more powerful objective and appreciate changes in scale. For example, we need to move away from global averages such as the oxidation state of the mantle and from experiments on global average sediments that do not exist anywhere. Melting processes depend on conditions and rock compositions that vary on a small scale. Many melts emanate from an oxidised beneath the cratons in which carbonate melts can exist in places, but C is present only in reduced form at both shallower and deeper levels. Care must be taken when referring to low degree melting in mixed source regions as this depends on spatial scale.

Progress will rely on demystifying metasomatism and heterogeneity by being more specific about what they are supposed to mean and about the abundances of volatile species involved. Explanations of melt origins can be expressed as a continuum between melting of peridotite alone and melting of pure neosome assemblages. Between them is a transition zone where melts such as nephelinites, basanites, melilitites, and alkali basalts have alternative explanations: their sources may consist either of volatile-bearing peridotite or mixtures of hydrous dykes and peridotite. The hydrous and carbonate melts produced from mantle migmatite neosomes are important carriers of critical metals, making study of mantle migmatite melting process societally topical and important.