There have been two intertwined main threads to my research: petrology of alkaline igneous rocks and processes in evolved magma chambers, and the feldspar group of minerals, particularly the alkali feldspars. The launching pad for both these interests was my PhD work on a small syenite pluton, the Loch Ailsh intrusion, in Assynt, which happens to be the type locality of the monomineralic alkali feldspar rock perthosite. Perforce, I became interested in feldspars.
|Fig. 1: Members of a Nordic Research Workshop held in 2003 admiring an 'inversely graded' layer in the Klokken intrusion. Feldspathic syenite gives way gradually upward via pyroxene rich syenite to pyroxenite. The student in red has her hand on a dune-like structure. The smiling man to her right is a Latvian sedimentologist.|
The Proterozoic Gardar Province in SW Greenland includes some of the world's best exposed and preserved alkaline intrusions, many showing igneous layering with a whole variety of styles. A succession of research students (see Career Outline) have worked on individual centres.
Klokken is a small but strikingly layered syenite intrusion, which I mapped for the Geological Survey of Greenland in 1970, and have visited many times since. It has proved to be an Alladin's cave not only for understanding processes in evolved magma chambers but also for understanding alkali feldspars. Klokken provides a wonderfully exposed cross section of a magma chamber with a gabbroic marginal border group, a series of syenogabbroic to syenitic sidewall cumulates with beautifully regular cryptic variation, a syenite core with a well-developed upper border group and a layered series with a unique style of inversely graded macrorhythmic modal layering (Fig. 1) interleaved with textural layering. Spectacular load structures are developed at the junctions (Fig. 2). Klokken has become one of the most intensively studied evolved intrusions, its many facets described in 35 papers with numerous coworkers.
Aspects of the very large Nunarssuit intrusion, involving several pulses of syenite and alkali granite, have been studied by several students, Jim Anderson, Andrew Butterfield, Adrian Finch and Mark Hodson. The most spectacular feature of the syenites, which are chemically almost identical to those of Klokken, is a zone of beautifully developed normally graded modal layering, with cross-bedding and unconformities, zones of slumping and reworked breccias, and striking cross-cutting washout channels with some similarites to the celebrated trough bands of the Skaergaard intrusion, but often containing mafic syenite breccias at their base. If anyone still doubts that sedimentary processes operate in magma chambers they should visit Nunarssuit. Adrian Finch and I have made use of biotite in Gardar intrusions and their envelopes to obtain information on halogens in associated fluids. It is generally believed that the exceptional development of igneous layering in Gardar intrusions reflects universally high magmatic fluorine contents.
|Fig. 2: Spectacular load structures at the interface between two texturally distinct, but chemically closely similar types of syenite. The load pouches and load balls have formed at the base of a layer of brown, fine-grained granular syenite. The flae structues are composed of white, coarse grained laminated syenite.|
A great range of poorly understood processes occur in magma chambers some of them leading to the development of layering. I was co-organizer, with Henning Sørensen, of a NATO Research Workshop, involving many of the best workers in the field, which was held in S Greenland in 1986, and led to a book 'Origins of Igneous Layering' (Reidel, 1987), that combined both theoretical and field-based approaches to the problems. We have subsequently held two similar workshops funded by the Nordic Council of Ministers, the latest in collaboration with Tom Anderson from Oslo.
Brian Upton and I are studying the marginal gabbroic rocks at Klokken and the anorthosite xenoliths and plagioclase megacrysts they contain. We are also working on a syenite centre called Syenitknold, on the edge of the Inland Ice, which is chemically similar to Klokken but is enclosed in the NE extension of the predominantly troctolitic Younger Giant Dyke Complex of Tugtutôq.
In 1980 I fulfilled a boyhood dream and took part in the Geological Survey of Greenland Peary Land-Kronprins Christian Land Expedition to the extreme North of Greenland. Peter Brown and I mapped the Kap Washington peralkaline volcanic sequence on Gertrude Rask Land to latitude 83°33'N. This is the most northerly land in the world.
Other work in Greenland has involved the Ketilidian rapakivi granites around Kap Farvel, the southern tip of Greenland, and two visits to the North Atlantic Tertiary province in E Greenland, to the Lilloise layered intrusion in 1974 and the Skaergaard intrusion in 1990.
Work on the igneous rocks of Assynt, led to the publication, in 1980, with Mike Johnson, of the Geological Excursion Guide to the Assynt District of Sutherland (Edinburgh Geological Society, 1979). The most exciting development in recent years was the discovery, in 1988, by one of my research students, Barry Young, of a small body of carbonatite, the first and only example in the British Isles. It is emplaced slightly outside the strongly alkaline Loch Borralan intrusion, and has all the isotopic and trace element signatures of a true carbonatite. Completely unexpected discoveries can still be made by observant people even in a very well-trodden region like Assynt! I am currently collaborating with Kathryn Goodenough and others in BGS on a new look at the minor intrusive rocks.
Two trips to West Africa, with French expeditions have taken me to alkaline complexes on the Cameroon Line, and to ring complexes in the Aïr region of Niger, in the central Sahara.
|Fig. 3: Transmission electron micrograph of a braid cryptoperthite from the Coldwell syenite intrusion, Ontario. Zig-zag lamellae are microcline, diamond shaped lozenges are albite, with straight-sided Albite twins. The width of the micrograph is about 3 micrometers, and all features of this beautiful strain-controlled intergrowth are suboptical. The crystal appears featureless in a light microscope.|
Feldspars, the most abundant minerals in the crust of the Earth. Their study has taken me down many truly interdisciplinary pathways, often at the interface between geochemistry and mineralogy. What can studies of feldspars from well-understood geological settings tell us about the origin of the microstructures and microtextures that feldspars contain in abundance, and, conversely, how can we make deductions about the geological history of rocks from the feldspars that they contain?
In Aberdeen i built a small hydrothermal high-pressure apparatus, with pressures of up to 4 kilobars reached using a very ordinary car jack. My most important experimental work, with my first research student, Peter Smith, was the reversed determination of the alkali feldspar solvus at 1 kilobar (1974), which has survived several in-depth thermodynamic analyses and remains the definitive study of this important phase equilibrium boundary. Roger Mason did interesting work on the effect of different fluids on the ordering rate of albite. Bill Brown and I (1981) produced the first two-feldspar geothermometer that was theoretically sound because it treated feldspars as ternary solid solutions, rather than two independent binaries. Our thermometer was graphical, but more recent computerized versions use the same basic principles. More recently, in Edinburgh, Pauline Thomson has studied the stability of the high pressure, relatively low temperature phase sanidine hydrate.
Up to the early 1980's the bulk of the work done on feldspar variation in rocks was chemical or based on X-ray diffraction, both methods that provide spatial averages that conceal most of the information contained in the feldspar. My PhD study of the Loch Ailsh syenite included single-crystal X-ray diffraction work on the feldspars, a state-of-the-art approach for the time, and this introduced me to a problem of which some aspects remain unsolved to this day. How was it that two adjacent units of syenite, in intimate contact, could contain different polymorphs of K-feldspar, orthoclase in the earlier unit, microcline in the later? Textbooks told me (as many still do) that microcline grew at low temperatures, or formed from orthoclase during very slow cooling, but neither of these ideas fitted the bill, because the solidus of both syenites was well above the temperature at which microcline was the stable form, and the two rocks clearly cooled together. My explanation appealed to the magical effects of water in the subsolidus on feldspar ordering, and this has been a recurrent theme as understanding of feldspars, and indeed crustal processes in general, has increased. A survey of the distribution of the two polymorphs in intrusive sequences, with Ron Boyd, showed that in composite intrusions, microcline rather than orthoclase always tended to form in the more evolved members, attributed to the build up of water with igneous fractionation, but carrying the implication that the fluid remained trapped in the individual units of the intrusion as they cooled into the microcline stability field. This was the theme of my 1977 Hallimond Lecture to the Mineralogical Society 'Feldspars and fluids in cooling plutons' which, looking back, was an early entry in the field of fluid-rock interaction.
|Fig. 4: SEM image of a cleavage surface of a fully coherent strain-controlled braid cryptoperthite from Klokken that has been etched in HF vapour. A vein of deuterically coarsened ('unzipped') perthite runs across the centre of the image, composed of discrete incoherent albite and microcline subgrains. Note the numerous micropores in this band, completely absent from the pristine braid perthite.|
Apart from a few pioneering studies using transmission electron microscopy (TEM), none in a geological context, almost all knowledge of feldspar microtextures, such as exsolution lamellae and fine-scale twins, was based on light microscopy. In 1977 Bill Brown and I resolved to use TEM to tackle problems of feldspar mineralogy in a geological context. TEM and SEM have proved extraordinarly rewarding, providing surprises and robust conclusions in equal measure. I am not an expert microscopist: Bill Brown, Richard Worden, Kim Waldron, Martin Lee and most recently, John Fitz Gerald have been brilliant expert collaborators.
Bill Brown and I began a long-standing collaboration with a study of alkali feldspars from the Klokken syenite with its perfectly exposed layered sequence. In one of the first of more than 20 co-authored papers involving TEM we showed that the lamellar periodicity of regular crypto- and micro-perthitic braid-perthite intergrowths (Fig. 3) plotted on a log scale, had a straight linear relationship with stratigraphic height, decreasing at the top to values very similar to volcanic rocks. We could safely conclude that the uppermost rocks of the layered series were very close to the unexposed roof and that feldspar microtextures could provide a relative cooling-rate-meter for these plutonic rocks. Because Klokken contains a continuous fractionated sequence of rocks from alkali gabbro through syenodiorites to syenite, and contains many rocks that have been little affected by deuteric fluids, we were able to study the variation with composition of regular perthitic intergrowths over a very large range of compositions and provide explanations of their diversity.
Perhaps the most important outcome of our TEM work on Klokken, which led to a conclusion applicable to almost all feldspars of plutonic origin, came as an answer to a simple question. Why do the two compositionally similar types of syenite on Fig. 2 look so different? The feldspars of the dark variety are transparent in small fragments and glass clear in thin section, those in the white are opaque in fragments and turbid in thin section. The clear feldspars have very fine-scale regular perthitic textures like Fig. 3, while the tubid ones are irregular patch perthite a thousand times coarser in scale. The turbidity is in caused by myriads of tiny micropores and at the TEM scale the crystals are mosaics of incoherent albite and microcline subgrains (Fig. 4). The turbid feldspar had been pervasively recrystallized by an aqueous fluid, new feldspar growing isochemically by dissolution and reprecipitation. In some rocks, particularly alkali granites, the entire volume of the alkali feldspar has been recrystallized.
|Fig. 5: Diagram summarizing the stages of development of microtextures in alkali feldspar phenocrysts from the Shap granite. The left hand diagram shows features prior to fluid-feldspar reactions, including igneous zoning which controlse the coarseness of exsolution textures, that on the right textures after three episodes of deuteric reaction.|
We now classify perthitic intergrowths into two types: 'strain-controlled' intergrowths have a continuous (coherent) or nearly continuous Si,Al-O framework and the orientation of the lamellae is controlled by the minimization of spontaneous coherency strain energy. 'Deuteric' intergrowths are much coarser and irregular, have a discontinuous (incoherent) framework with many micropores and subgrain boundaries. This 'dual microtexture' is the rule in alkali feldspars from plutonic rocks including those from ordinary granites and gneisses.
We called the process of dissolution and reprecipitation leading to coarse patch perthite an 'unzipping' reaction, because it was driven by minimization of elastic coherency strain energy in the strain controlled intergrowths. Another important unzipping reaction is the transformation of orthoclase to microcline, driven in this case by coherency strain energy in the tweed domains in orthoclase and by free energy lost by Si-Al ordering. We studied the very complex mechanism of unzipping through the stages preserved in features we called 'pleated rims', which form along crystal margins and cleavages. We suspect that unzipping reactions are more common in many minerals than is commonly supposed.
A 1988 40Ar/39Ar study of the Klokken feldspars with Alex Halliday showed others that the apparent ages of the pristine strain-controlled perthites were very close to the emplacement age of 1166 Ma, while the apparent age of the adjacent deuteric perthites was about half this. They have clearly shared the same thermal history since the Proterozoic and the role of microtexture in controlling the retention of radiogenic Ar is unequivocally demonstrated. This has led to a lively and continuing controversy with workers who apply 40Ar/39Ar 'thermochronology' without taking account of observable microtextural details. Ray Burgess, Simon Kelley and I used an early laser extraction technique to explore the distribution of Ar in samples from Klokken with respect to microtexture, and Ray and I used an on-line crushing technique coupled with the 40Ar/39Ar method to explore the origin of fluid inclusions in the famously altered feldspars in the Red Hills of Skye.
|Fig. 6: SEM image of a naturally weathered surface of lamellar perthite from Shap. The albite has almost completely dissolved, leaving thin intervening walls of K-feldspar which break off in thin flakes. The holes represent the outcrops of dislocation etch pits.|
Surprisingly, as they are one of the most abundant minerals in Earth's crust, it was not until 1995 that K-feldspar from a typical two-feldspar granite was described at the TEM scale. Martin Lee, Kim Waldron and myself explored the pink phenocrysts from the granite at Shap, Cumbria, known to generations of British geology students. They have dual microtextures (the mixing of turbid and clear feldspar can be seen in hand specimen), but the fluid-rock interaction history is more complex than Klokken, and seems to have involved at least three stages (Fig. 5).
The strain-controlled perthites in Shap (and other feldspars from ordinary granites) are planar. Because this leads to insupportable coherency strains, dislocations form in two directions at right angles on the surface of the albite lamellae. These defects have turned out to be very important, and are the main subject of current work. Because structure around dislocations is highly reactive and etch pits form readily, unzipping reactions tend to be guided along lamellae (upper part of Fig. 5, right). Non-isochemical replacement of Shap feldspars by pure albite can be demonstrated, and there was also a stage of cross-cutting patch-perthite development in which the plagioclase phase is oligoclase. Granite petrologists should not interpret the chemistry of their rocks in terms of purely igneous processes.
We studied the Shap feldspars after they had been transported as detrital fragments from the Lower Devonian granite and incorporated in a lower Carboniferous basal conglomerate. Many microtextures were still visible, but surprisingly almost all the albite lamellae had been replaced by diagenetic microcline, giving a 'perthitic' intergrowth of lamellar microcline in orthoclase.
Martin Lee, Pauline Thompson and myself have in recently turned our attention to detrital feldspars in potential North Sea reservoir rocks. Comprehensive SEM and TEM work on the fine-grained feldspathic sandstones of the Fulmar formation, in the central North Sea, suggests to us that alkali feldspars that survive weathering and transport into sandstones are those that lack defects that enhance dissolution rates. Thus feldspars from volcanic sources and granulite facies rocks are more abundant than those from granitoids, which one might expect to be prominent protoliths of the North Sea sedimentary rocks. With Nicola Cayzer, who has recently completed a PhD study of microtextures and two-feldspar geothermometry of granulite facies rocks, we have suggested that feldspar microtextures could be used as provenance indicators. Diagenetic dissolution and replacement of alkali feldspars is strongly dependent on microtextures inherited from the igneous or metamophic protoliths and these provide a guide to the likelihood of the development of secondary porosity, a major element in reservoir quality. In the Fulmar formation detrital grains were affected by several phases of diagenetic microcline growth and albitization, often only visible at the TEM scale. Another major discovery in the Fulmar rocks was an abundance of plagioclase with the very infrequently reported sub-optical albite-oligoclase intergrowth known as peristerite, which must point to a very particular provenance.
|Fig. 7: False colour SEM image of the weathered surface of a Shap feldspar with various types of bacterium coloured red. Is it fortuitous that these extremely abundant organisms have the same dimensions as the natural etch structures?|
Dislocations are of great importance during weathering and diagenesis. Martin Lee and I found that in natural peat soils overlying the Shap granite, dissolution of alkali feldspars is much more rapid in the vicinity of dislocations, leading to deep etch pits and to the mechanical disintegration of the feldspar surface (Fig. 6). The conventional wisdom in the experimental weathering field is that crystal defects play at best only a minor role in feldspar dissolution, quite contrary to what happens in nature. With Mark Hodson we carried out laboratory dissolution experiments on Shap feldspars and showed, using Atomic Force Microscopy, that in experiments lasting nearly a year etch pits were just starting to form. Amos Aikman has shown that labyrinths of intersecting etch pits in phenocrysts removed from weathered rock surfaces extend over 15 mm into crystals, enormously increasing their effective surface area. Joe Smith, Martin Lee and I have speculated that these complex surfaces would have provided ideal reactors for the first steps in the origin of life, encapsulating the self-organized biochemical reactions which are its essence (Fig. 7).
A high-spot of my feldspar work was the NATO Advanced Study Institute on feldspars which I directed in Edinburgh in 1993. This led to a book, 'Feldspars and their reactions' (Kluwer, 1994) which provides an overview on feldspar minerals written by most of the leading experts in the field. The development of modern mineralogy is tracked rather well by this book and its three NATO ASI predecessors. Feldspars might be considered a specialist interest, but they have, very satisfyingly, led me from igneous and metamorphic petrology, experimental mineralogy and aspects of mineral physics into weathering, transport, sedimentary diagenesis and provenance studies and even to speculate on the origins of life. It is fun learning about new fields and at the same time importing into them new specialist knowledge which may open new horizons.