Regime change: convection and crystallisation of magma

Lead Research Organisation: University of Cambridge
Department Name: Earth Sciences


The cooling of poly-component liquids, such as magma (and also ice-cream and salt- or sea-water), can drive solidification in a bewildering array of styles. Often the solid that forms is of a different composition from the liquid (e.g. pure ice from salt water). This means that the composition and temperature of the residual liquid is always changing during cooling, causing changes in the density of the liquid. These density changes can drive convection in the liquid, and can have profound effects on the way in which mass and heat are transported within the crystallising system. When cooling rates are gentle solidification occurs from the cold boundaries as when ice forms on the pond on a still winter's day. In contrast, when cooling rates are very high, vigorous convection in the liquid can drive crystallization away from the cold boundaries, forming a flurry of crystallization in the swirling interior.
In the context of bodies of molten rock (magma) the way convective motion can re-distribute mass has significant effects on the way the residual liquid changes composition. This plays a vital role in determining the final composition (and hence the explosivity) of any erupted lava flows. The style of crystallization also affects how quickly a magma conduit feeding a surface eruption will freeze sufficiently to prevent more magma travelling along it. A further important reason to understand how convection controls the way magmas evolve in crustal magma chambers is because the only way we can make deductions about processes occurring in the inaccessible deep Earth is by an examination of the composition of erupted lavas.
The project will involve creating small-scale, bench-top analogues for real magma bodies using salt-water solutions. We will be able to control the cooling and solidification rates in our tanks and watch directly what happens and where the crystals are forming - something that is not possible in real magmas. We will compare our experimental results with natural examples of basaltic, magmatic intrusions by taking advantage of some recent new discoveries that mean we can decode the record of crystallization style left in fully-solidified basaltic intrusions and flows using details of grain shape, internal compositional variations and the spatial distribution of dense minerals. These microstructural markers will enable us to work out whether the liquid in the magma bodies convected or was static during solidification. These discoveries provide an exciting opportunity to make real progress in understanding the fundamental processes at work as these bodies cooled.

Planned Impact

The results of this project will have impact all industries dealing with convective crystallising systems such as the mining, food processing, and the chemical industries.
In particular our work will have a direct impact on mineral exploration. Our South African Project Partner, Dixon, already has strong links with the mining industry and has expressed willingness to help foster informal links between the companies with which he works (Anglo American, Lonmin and Impala Platinum) and the Cambridge group. We anticipate that building bridges in this way will lead to more concrete future collaborations on problems relating to crystallization and fluid flow in magmatic systems, via CASE studentships (with companies such as Lonmin, with UK links), and application to the Newton Fund to support secondments for Masters students working with mining companies.
Our work will be of immediate interest to the part of the food science industry concerned with crystallization and particle-bearing systems. We already have close links with Unilever and are collaborating on a project examining the rheological behaviour of ice-glucose slurries (NE/K007661/1). Our Unilever colleagues have expressed interest in the potential of the proposed work to expand their understanding of, and expertise in, food processing, particularly ice-cream and other frozen confections. The proposed work will have direct economic implications for Unilever through our collaboration.
We will engage in public outreach, using video, podcasts, talks, and articles in popular magazines to target a general audience. The PI has extensive experience of film-making and radio work, and regularly gives talks to amateur geological societies and the Cambridge Science Festival. The rich possibilities provided by the proposed work for engaging the public means the Cambridge group is extremely well-placed for further outreach work, putting the fundamental problems concerning solidification and magma behavior in an exciting and dynamic geological context.
The fundamentally inter-disciplinary nature of the proposed research will be of benefit to the PDRA who will acquire a skill set that could be deployed in a wide range of industries (food processing, mining, detailed geochemical analytical work). The PDRA will benefit from our existing links with industry, and from the links already developed by Project Partner Dixon, and will be well placed for employment in industry.


10 25 50

publication icon
Holness Marian (2016) Microstructural indicators of convection in sills and dykes in EGU General Assembly Conference Abstracts

publication icon
Holness MB (2017) The thickness of the crystal mush on the floor of the Bushveld magma chamber. in Contributions to mineralogy and petrology. Beitrage zur Mineralogie und Petrologie

publication icon
Holness MB (2017) Crystal settling and convection in the Shiant Isles Main Sill. in Contributions to mineralogy and petrology. Beitrage zur Mineralogie und Petrologie

publication icon
Holness MB (2018) How deceptive are microstructures in granitic rocks? Answers from integrated physical theory, phase equilibrium, and direct observations. in Contributions to mineralogy and petrology. Beitrage zur Mineralogie und Petrologie

Description We have discovered that the microstructure of a solidified igneous rock records whether or not it solidified inwards from the margins or whether crystallisation took place by the growth of isolated crystals suspended in a vigorously convecting magma. We have found that the crystallisation regime is determined by the orientation of tabular intrusions, with three distinct patterns of behaviour recorded by spatial variations in plagioclase grain shape. Sills display an M-shaped variation of plagioclase aspect ratio while dykes display either univariant plagioclase grain shape across their width or a U-shaped variation. The criterion determining whether a dyke displays a flat or U-shaped variation is the distance the magma has propagated laterally from its source. We argue that the marginal reversals in plagioclase shape are related to a transient stage during which the magma in the sill convects sufficiently to bring hot magma to the margins of the body and thus result in relatively equant plagioclase - this is likely to resolve a major dispute in the literature concerning the longevity of convection in horizonal tabular intrusions. We have also demonstrated that it is possible to detect the convection of magma by examining closely the grain size variations within a bed of crystals that settled from the magma. We have developed a large dataset for olivine-phyric sills that has resulted in a regime diagram showing convective and crystallisation behaviour as a function of sill thickness and original crystal load. This work provides an elegant observational approach to addressing the controversy over whether magma convects or not.
Exploitation Route They will be of general use in decoding the crystallisation regime of magma bodies in the crust
Sectors Manufacturing, including Industrial Biotechology

Description Talk 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach Regional
Primary Audience Public/other audiences
Results and Impact A talk to the Oxford Geology Group
Year(s) Of Engagement Activity 2017