Alkali Feldspar Microstructures as Provenance Indicator_921

Journal of Sedimentary Research, 2005, v. 75, 921942

DOI: 10.2110/jsr.2005.071

ALKALI FELDSPAR MICROTEXTURES AS PROVENANCE INDICATORS IN SILICICLASTIC ROCKS ANDTHEIR ROLE IN FELDSPAR DISSOLUTION DURING TRANSPORT AND DIAGENESIS

IAN PARSONS,1 PAULINE THOMPSON,1 MARTIN R. LEE,2 AND NICOLA CAYZER11Grant Institute of Earth Science, University of Edinburgh, West Mains Road, Edinburgh EH9 3JW, U.K.

2Division of Earth Sciences, Centre for Geosciences, University of Glasgow, Lilybank Gardens, Glasgow G12 8QQ, U.K.

ABSTRACT: Alkali feldspars usually exhibit optical or sub-optical inter-growths of albite and K-feldspar (known generically as perthite) which havemorphologies and crystallographic characteristics that are distinctive of theigneous or metamorphic environment in which they grew and cooled tosurface temperatures. We review the current terminology and understandingof how the microtextures form and discuss their role in feldspar dissolutionduring weathering, transport, and diagenesis. We show how microtexturescan be used as provenance indicators in arkosic siliciclastic rocks, usingexamples from the Fulmar Formation, a reservoir rock in the UpperJurassic Humber group in the Central North Sea, in which feldspardissolution is a major source of secondary porosity. We describe the mosteffective techniques for routine characterization, of which the most practicalis back-scattered electron imaging in a scanning electron microscope, oftencoupled with cathodoluminescence. We provide an atlas of the main types ofperthitic intergrowth likely to be encountered in igneous and metamorphicrocks worldwide, and show how many of the microtextures can be matchedin clastic grains in the Fulmar Formation. Grains which are preferentiallypreserved have microtextures that make them relatively unreactive inaqueous fluids at low temperatures, and we conclude that provenance hasa major impact on reservoir quality. Detrital grains in the Fulmar exhibitauthigenic albitization, K-feldspar overgrowths, and replacement, and wediscuss how diagenetic features can be distinguished from replacive featuresdeveloped in the protolith.

INTRODUCTION

We here describe how intracrystal microtextures in detrital alkalifeldspars can be used to infer the provenance of siliciclastic rocks andreliably distinguish detrital feldspars from authigenic grains and over-growths. We discuss how these microtextures influence the rate at whichdetrital alkali feldspar grains degrade and dissolve during weathering (Leeand Parsons 1995; Lee et al. 1998) and undergo dissolution andreplacement during diagenesis (Lee and Parsons 1998) and hence controlwhich grains, themselves usually fragments of crystals of igneous ormetamorphic origin, are likely to survive or to dissolve to createsecondary porosity in potential reservoir rocks. From our previousexperience of intergrowths in igneous or metamorphic rocks we cansuggest a variety of ultimate provenances for detrital alkali feldspars withconsiderable confidence. The use of microtextures to deduce provenance,and their role in the reactivity of feldspars during diagenesis, was brieflydiscussed by Lee and Parsons (2003) and the present paper develops theseideas. The proportion of feldspars relative to other minerals hasfrequently been used as a provenance indicator (e.g., Arribas et al.2000), but as far as we know our use of intracrystal feldspar microtexturehas not been suggested before.

We begin this paper with a review of the current understanding ofalkali feldspars relevant to the characterization of detrital grains in clasticsedimentary rocks, and to their relative stability during burial and

diagenesis, and suggest the most effective techniques for routine work.We illustrate the method with sedimentary examples from the FulmarFormation, an important reservoir rock in the Upper Jurassic Humbergroup in the Central North Sea, and with examples from our earlier workon igneous and metamorphic rocks. The Fulmar oilfield, from which thesamples in the present study come, lies in a small, steep-flanked anticlineon the southwestern edge of the Central Graben in the U.K. sector of theNorth Sea. Our observations were made on a total of 24 core fragmentsfrom 5 wells from measured depths ranging from 3273 to 5483 m. TheFulmar Formation sandstones and sandbodies are Upper Jurassic(Oxfordian to Kimmeridgian) in age (Stockbridge and Gray 1991) andcomprise fine- to medium-grained, heavily bioturbated and largelystructureless, shallow marine arkosic sandstones that possibly werederived from Triassic sedimentary rocks. The reservoir is overpressuredand currently at a temperature of , 130uC (Saigal et al. 1992).In the Fulmar Formation feldspar dissolution is believed to be the main

cause of secondary porosity and it has been suggested that the amount ofdetrital alkali feldspar decreases systematically with burial depth(Wilkinson and Haszeldine 1996; Wilkinson et al. 1997). Nevertheless,we have found considerable amounts of detrital alkali feldspar (. 10 vol.% of solids) in deep Fulmar wells. We emphasize that the survival of anyparticular feldspar grain, in pore fluids which can be expected to be at orvery close to local equilibrium with the mineral assemblage in the clasticrock, is strongly dependent on intra-grain microtexture, rather thanchemistry, so that provenance is directly relevant to secondary porositydevelopment, and may have predictive potential. By comparing thepopulations of distinctive feldspars to those in earlier sedimentaryformations in the same basin, it should be possible, in principle, toestablish whether they have passed through previous sedimentary cycles,for example, in the case of the Fulmar, to establish whether grains havecome from reworked Devonian or Permo-Triassic sandstones, assuggested by Stewart (1986).

UNDERLYING PRINCIPLES

Nomenclature and Polymorphism of Alkali Feldspars

The terminology used here is that of Smith and Brown (1988) and asdeveloped in reviews by Parsons and Brown (1984), Brown and Parsons(1989), Brown and Parsons (1994), and Parsons and Lee (2000). Manyillustrations of perthitic textures can be found in Smith (1974, Chapter 19)and Smith and Brown (1988, Chapter 19). A convention to be noted isthat the word Albite is capitalized when it refers to the twin law,whereas albite refers to the mineral phase. Similarly, Adularia refersto the characteristic {110} habit of the low-temperature variety ofK-feldspar known as adularia.

At igneous and metamorphic temperatures alkali feldspars crystallizeas homogeneous ternary solid solutions of the three end-membercomponents NaAlSi3O8 (albiteAb), KAlSi3O8 (orthoclaseOr) andCaAl2Si2O8 (anorthiteAn). (They may also contain significant amounts

Copyright E 2005, SEPM (Society for Sedimentary Geology) 1527-1404/05/075-921/$03.00

of the component BaAl2Si2O8 [celsianCsn]). As temperature decreases,the originally homogeneous crystals usually exsolve (unmix is analternative term) in the solid state to form intergrowths of two feldsparphases known generically as perthitic intergrowths. One phase is rich in theOr component and is best called the Or-rich phase; the other is rich in theAb component and is best called the Ab-rich phase or the plagioclasephase. Although petrographers often refer to albite exsolution lamellaein perthitic crystals, this is not necessarily correct; they often containsignificant An and may be oligoclase (Ab90An10Ab70An30) orexceptionally andesine (Ab70An30Ab50An50). The term perthite shouldproperly be reserved for intergrowths in which Ab-rich plagioclaselamellae or patches are enclosed by Or-rich feldspar, antiperthite is usedwhen Or-rich feldspar is enclosed by plagioclase, and mesoperthite whenthe two phases are present in approximately equal proportions. Althoughmany perthites form by exsolution it is certainly the case that some formby replacement and many single grains contain both exsolved andreplacive intergrowths (e.g., Lee et al. 1995). An important point is thatintergrowths which have formed by exsolution must have developedinitially in homogeneous single crystals at an elevated temperature andwhen found in sedimentary rocks must therefore be detrital. Intergrowthsformed by replacement, on the other hand, can form both during coolingfrom igneous and metamorphic temperatures and during diagenesis.Criteria for differentiating the different types of intergrowth are givenlater.

Sub-optical intergrowths (i.e., , 1 mm in scale) are called crypto-perthites (or cryptomesoperthites, cryptoantiperthites), optically visibleintergrowths are given the prefix micro-. Modern electron microscopework has shown that many crystals contain both cryptoperthite andmicroperthite. Textbooks frequently assert that cryptoperthite is charac-teristic of alkali feldspars from volcanic rocks and is a result of rapidcooling, but this is correct only in a very general way, and it often occurs,together with coarser intergrowths, in relatively slowly cooled igneous(Brown et al. 1983; Lee et al. 1995) and metamorphic rocks (Evangela-kakis et al. 1993; Cayzer 2002). This is a point to consider whenidentifying fragmentary clastic grains.

Replacement by Na-rich feldspar is often called albitization, althoughthe replacive phase is not always pure albite (Lee and Parsons 1997b).Albitization may occur both in cooling igneous rocks and duringdiagenesis. When encountered in clastic grains it is not self-evidentlya product of diagenesis. Replacement by K-rich feldspar (Morad et al.1989; Lee and Parsons 1998) has been called microclinization, but thisterm has also been used for the process by which orthoclase changes intomicrocline, and is best avoided. Furthermore, replacive K-feldspar formedin diagenesis is not usually microcline.

The high-temperature monoclinic form of K-feldspar, sanidine, in-cluding fully disordered (with respect to Si and Al in the Si,AlOframework) high sanidine and partially ordered low sanidine, may growmetastably (Baskin 1956; Glover and Hoseman 1970) during diagenesis.Orthoclase has a distinctive microtexture, visible only at the transmissionelectron microscope (TEM) scale known as tweed. This is composed ofalternating twin-like domains only a few unit cells thick (Eggleton andBuseck 1980) and develops from low sanidine by a continuous process inwhich alternating left-handed and right-handed triclinic domainsform. Orthoclase retains average overall monoclinic symmetry andbecomes stranded and unable to order further to produce fullyordered, twinned triclinic microcline. The word orthoclase will be usedhere in its modern sense of an Or-rich K-feldspar with a well developedtweed microtexture. The symbol Or refers to the KAlSi3O8component only, and carries no crystal structural implications.

Fluidfeldspar reactions are usually required for the transition fromorthoclase to triclinic microcline via what has been called an unzippingreaction (Brown and Parsons 1989; Putnis 2002). Microcline usuallyexhibits the well known cross-hatched tartan twinning in the optical

microscope, although twins may be visible only using TEM (see, e.g.,Fig. 7 for a detrital example). Much microcline in igneous andmetamorphic rocks is very irregularly twinned at the TEM scale and isusually called irregular microcline. Within single K-feldspar grains it isoften mixed with volumes of tweed orthoclase on the scale of a few mm(see Fitz Gerald and McLaren 1982, their fig.11). Many authigenic K-feldspars are the variety known as adularia, which is distinguished by its{110} habit and structurally by partial ordering and a weak departurefrom monoclinic symmetry (see Lee and Parsons 2003 for examples fromthe Fulmar Formation, and also Worden and Rushton 1992). Althoughmany diagenetic adularias have weakly developed tweed microtexture,well developed tweed is probably diagnostic of detrital grains. Welldeveloped tartan twinning is a reliable indicator of a detrital origin.Ab-rich feldspars also undergo a monoclinic R triclinic phase

transition on cooling, from high albite and high Na-sanidine to triclinicanorthoclase. This phase transition leads to tartan twinning in anortho-clase, but the twins are usually finer-scale, straighter, and more regularthan in microcline, with angular intersections. In microcline the tartantexture is visible only in sections nearly parallel to (001), on which onlyone cleavage, (010), is visible, whereas in anorthoclase the tartan twinningis visible on (100) and two cleavages, (010) and (001), are visible.Anorthoclase cannot grow during diagenesis, and discovery of anortho-clase in a sedimentary rock would clearly indicate a detrital origin.

Solvus Relationships

The compositions of pairs of feldspar phases coexisting in equilibrium,whether in perthitic intergrowth or as separate crystals, becomeprogressively more different as temperature decreases, and are definedby a temperatureomposition surface, the ternary feldspar solvus (Fig. 1).The significance of the various curves is discussed in the caption. WhenAn is not present the composition of coexisting feldspars is defined by thebinary alkali feldspar solvus (Smith and Parsons 1974), and this is oftenused as an approximation of solvus relationships. However, althoughmany alkali feldspars have low bulk An contents (12 mol%), smallamounts of the third component, An, have a large effect on thetemperature at which exsolution begins and exercise a strong control onthe microscopic appearance of perthitic intergrowths (Fig. 2). AbOrAnfeldspars with more than a few percent of the least abundant componentare often loosely called ternary feldspars, although strictly speaking allnatural feldspars are at least ternary.The strain-free ternary feldspar solvus (the term strain-free is

explained below) defines the minimum temperature at which a feldsparphase of a given bulk composition could have grown from magma, ina metamorphic rock, or from a diagenetic fluid. Igneous rocks in whicha pair of discrete feldspar phases crystallize from magma (e.g., tie line PBAF in Fig. 1) are called subsolvus rocks. Magmatic rocks in whicha single feldspar with a bulk composition such as B crystallizes are calledhypersolvus rocks. Small amounts of An in an alkali feldspar and of Or ina plagioclase feldspar have an extremely large effect on minimum possiblecrystallization temperature (Fig. 1). At diagenetic temperatures stablefeldspars are nearly pure Ab and Or, and the amount of An that candissolve in an alkali feldspar at diagenetic temperatures is very small (%0.5 mol%). This provides a way to distinguish authigenic from detritalfeldspars using back-scattered electron imaging (BSE) because of thehigher mean atomic number when Ca is present. Similarly, the amount ofOr that can dissolve in a plagioclase (AbAn) feldspar growing at lowtemperature is small.

Determination of Bulk Composition

When stating a chemical analysis of an alkali feldspar, it is crucial tomake clear, particularly when it has been analyzed using electron-probemicroanalysis (EPMA) or energy-dispersive X-ray microanalysis (EDX)

922 I. PARSONS ET AL. J S R

in the scanning electron microscope (SEM), whether it is a bulkcomposition or a phase composition that has been obtained. Phasecompositions can be obtained by point analysis of coarse perthiticintergrowths or by measurement of peak positions using X-ray diffraction,but they have little value in provenance studies because compositions ofintergrown phases usually remain in equilibrium down to low tempera-ture and therefore approach pure albite and almost pure K-feldspar(. Or90) on the ternary solvus (Fig. 1). The only exceptions are somequickly cooled volcanic rocks (e.g., MacKenzie and Smith 1956) and raregranulite facies xenoliths in volcanic rocks (Hayob et al. 1989).Bulk compositions can provide useful provenance information; for

a compilation of feldspar bulk analyses (both major and trace elements)see numerous figures in Smith and Brown (1988), Chapter 14. The bulkcomposition of completely homogeneous crystals can be obtained bysimple EPMA point analysis, but such crystals are rare and most naturalalkali feldspars are perthitic. If the exsolution textures are on a , mmscale an estimate of bulk composition can be obtained using a defocussedbeam of . 10 mm diameter or a rastered beam, but to obtain the bulkcomposition of a coarsely exsolved microperthite requires the integrationof counts over a large area of a section. A combination of EPMA andBSE image analysis has recently been applied successfully (Marks and

Markl 2001). Even if this is done, it is often uncertain that the bulkcomposition obtained is that with which the crystal initially grew, becausereplacement reactions involving addition of both Ab or Or can affectalkali feldspars during cooling from igneous temperatures and duringdiagenesis (e.g., Lee et al. 1995; Lee and Parsons 1998). The replacivefeldspars may differ in composition from those which have exsolved (Leeet al. 1995).

Exsolution Microtextures

Truly homogeneous alkali feldspar crystals are extremely rare,restricted to sanidines quenched very rapidly from volcanic temperaturesand not subsequently subject to fluidrock interactions, and to crystals orovergrowths growing at very low temperature during diagenesis.Petrographic descriptions based on optical microscopy (OM) oftendescribe optically featureless alkali feldspars as unexsolved, but exceptin the case of authigenesis the chances are high that the crystal in questionis cryptoperthitic. BSE imaging in an SEM can reveal cryptoperthite,except for the finest intergrowths, which require TEM. The mostconvenient way to distinguish detrital alkali feldspars that are featurelessusing OM and BSE imaging, from authigenic crystals, is cathodolumi-

FIG. 1.Temperaturecomposition prism for the ternary feldspar system AbOrAn. The Strain Free Solvus surface (SFS, purple) defines the compositions of pairs offeldspar phases (P: plagioclase, AF: alkali feldspar) growing simultaneously from magma, during metamorphism, or during diagenesis. P and AF must lie on the ends ofa tie line passing through the bulk composition B. The position of this tie line depends on temperature, and to a lesser extent on pressure. This relationship is the basis oftwo-feldspar geothermometry. The SFS also defines the compositions of the phases coexisting in perthitic intergrowths provided that they do not share a common Si,AlO framework (i.e., they are incoherent). When they share a common framework (are coherent), the phase compositions are depicted by the Coherent Solvus (CS, blue). Ifan alkali feldspar crystal of bulk composition AF (typical of a calc-alkaline granite) is cooled, together with a coexisting plagioclase P, the two phases should, in principle,react continuously with one another and maintain compositions on the SFS. In practice this usually does not occur. P and AF behave as separate systems and begin toexsolve when they encounter the CS, each unmixing into two phases. For the alkali feldspar these will have compositions N and K, on the blue surface, in proportionswhich reflect the bulk composition of AF, and the intergrowth is a perthite. P will also exsolve, forming an antiperthite. As cooling proceeds the compositions of N and Kbecome progressively more different, but unless the crystals undergo dissolutionreprecipitation reactions in an aqueous fluid they will remain on the CS. When a fluid ispresent, all or part of the crystal my recrystallize to a strain-free assemblage on the SFS. The driving force for this recrystallization is coherency strain energy. Strain-freeassemblages at low temperature are close to end-member feldspars, so feldspars growing, or actively replacing detrital feldspars during diagenesis, are nearly pure phases(N and K on the tie line passing through B 5 AF, which is drawn at the present-day temperature of the Fulmar Formation). The SFS and CS shown are for equilibriumSi,Al ordering (Brown and Parsons 1989). The 750uC isotherm is from Fuhrman and Lindsley (1988) and is for disordered feldspars.

ALKALI FELDSPAR MICROTEXTURES AS PROVENANCE INDICATORS 923J S R

nescence (CL). This is likely to be bright in the volcanic case but absentfrom authigenic grains. Note that in the present paper all CL images wereobtained in an SEM with a detector with a spectral range of 275700 nm.Microperthitic and cryptoperthitic crystals fall into two broad

categories: finer scale (usually # 1 mm), regular strain-controlled inter-growths and coarser, much more irregular, deuterically coarsened inter-growths (Parsons and Brown 1984). The former are the result ofa continuous process in which Na and K ions diffuse through anunbroken SiAlO framework (Ca will also diffuse, as discussed below).The latter are the result of a discontinuous process involving dissolution

and reprecipitation in an aqueous fluid, while the crystal retains itsexternal shape. The microtextural differences between the two types areextremely profound, and have a strong influence on the rate ofdissolution of feldspars during weathering, transport, and diagenesis.The most obvious difference in OM is that in plane-polarized lightcrystals with strain-controlled intergrowths are glass-clear (apartperhaps from isolated fluid or mineral inclusions) whereas deutericallycoarsened intergrowths are in various degrees turbid. Most feldspars fromcommon types of igneous and metamorphic rocks are to some extentturbid.

FIG. 2.Schematic diagram showing coherent exsolution microtextures, as they appear viewed down the Z axis, developed as a function of composition in slowlycooled feldspars in the system AbOrAn. Vertical lines are Albite twins in plagioclase and in the Ab-rich phase in perthitic intergrowths; cross hatching at the rightrepresents tweed orthoclase. Unornamented areas are Or-rich feldspar, low sanidine where intergrown albite lamellae are straight, microcline where zigzag or whereenclosing lozenge-shaped areas of albite. Note that these textures apply to crystals that have not been externally deformed. From Brown and Parsons (1988).


Strain-Controlled Intergrowths.The Si,AlO framework passes con-tinuously, or nearly so, between the Or-rich and Ab-rich phases. Suchintergrowths are said to be coherent, or, if dislocations are present on theinterfaces, semicoherent. Because of the difference in cell dimensionsbetween Na- and K-rich feldspar the framework is elastically strainedclose to the interface of the two phases. The chemical driving force whichleads to exsolution has two opposing forces: surface and strain energy.The compositions of the exsolving phases are defined by a coherent solvuswhich is at lower temperature than the strain-free solvus (Fig. 1).Coherent exsolution is a continuous process which begins during coolingat or just below the coherent solvus. It is incorrect to speak of oneexsolution temperature and preferable to describe this temperature asthe beginning of exsolution. The coherent solvus defines the minimumtemperature at which a feldspar of any particular bulk composition canpersist as a homogeneous crystal during cooling unless cooling rates areextremely high, for example in ash deposits. The mechanisms by whichexsolution occurs in alkali feldspars are discussed by Parsons and Brown(1991).

As exsolution proceeds, surface energy is minimized by coarsening ofthe intergrowths and lamellae initially a few nm thick become visible

optically (. 1 mm). Structural strain is minimized by migration of theinterfaces of the two phases into orientations where strain energy is least(Fig. 2). In perthites sensu stricto coherent lamellar interfaces are usuallyapproximately straight and initially between (601) and (801) of themonoclinic host. The overall shape of the albite lamellae is that ofextremely extended flat lenses (e.g., Lee and Parsons 1995, their figs. 14)when viewed on (010) and rather shorter lenses on (001), the overall shapebeing like an aircraft wing (Fig. 3; also Figs. 1214). Such intergrowthsare called film perthites. In cryptoantiperthites, lenses of K-feldspar occurin a plagioclase host. When the proportion of K-feldspar is low, the lensesresemble coffee beans (Brown and Parsons 1988) and the lenses bendwhere they intersect Albite twin lamellae (Fig. 2, area 1).

The orientation and shape of coherent intergrowths formed duringslow cooling, in plutonic igneous and metamorphic rocks, depend on thebulk composition of the crystal. The final products are summarized inFig. 2. Coherent lamellae in perthites and antiperthites remain essentiallyplanar, but in coherent mesoperthites the film lamellar interfaces bendwhen framework ordering and associated twinning occur (Brown andParsons 1984), giving zigzag perthite (area 5 in Fig. 2) and formesoperthites with low An contents a so-called braid perthite develops,

FIG. 3.Generalized sketch of an , 100-mm-long cleavage fragment of an alkali feldspar phenocryst from the Shap granite, Cumbria, U.K., a typical calc-alkalinegranite, as it might occur as a detrital fragment. The upper surface is the (001) cleavage, the right-hand is the (010) cleavage, and the front is the Murchison plane. Thickerlamellae and subgrains of albite are filled in gray. K-feldspar, orthoclase in the areas with lamellar perthites and microcline in areas of patch perthite, is unfilled. Alkalifeldspars from many granites and alkali feldsparbearing gneisses have similar microtextures, so that, on an Earth-wide basis, this is the most common type of detritalfeldspar likely to be encountered in a siliciclastic rock. Note how the exsolution textures vary in appearance depending on viewing direction, and the large variety oftextures that might be found in detrital grains formed from the further disintegration of such a fragment. (Modified from Parsons et al. 1999).


in which lozenges of albite are separated by sheets of microcline (area 4).Braid perthites are visible under high power using OM and are easilyvisible using BSE in an SEM (e.g., Fig 10). These intergrowths, and thefine straight intergrowths which precede them, are fully coherent.Coherency strains are minimized by the complex shapes of the interfacesbetween the phases. However, in other parts of Figure 2, where lamellaeare straight or nearly so (in fact the lamellae have much higher aspectratios than shown; see Figs. 3, 1214) spaced dislocations form and theintergrowths become semicoherent (next section).

The coarseness of coherent exsolution textures has been shown to bea function of cooling rate by experimental work (Yund et al. 1974; Yundand Davidson 1978) and by inference in both volcanic rocks (Yund andChapple 1980; Snow and Yund 1988) and relatively rapidly cooledplutonic igneous rocks (Brown et al. 1983; Waldron and Parsons 1992).The coarsest regular coherent intergrowths occur in feldspars in granulitefacies metamorphic rocks (Evangelakakis et al. 1993; Cayzer 2002) thathave been cooled extremely slowly or maintained for long periods at sub-

solvus temperatures. However, the relationship between coarseness andcooling history in plutonic rocks is not a simple one, except in exceptionalcircumstances (e.g., in the Klokken syenite intrusion; Brown et al. 1983),and many individual grains of K-feldspar from granitic rocks containcoherent and semicoherent exsolution textures at a range of scales fromseveral mm to a few nm, with variation in wavelength and scale occurringover short distances (e.g., Lee et al. 1995).

The presence of the An component (and possibly Csn) is probably animportant control on coarsening rate, because diffusion of divalent Ca2+

(and Ba2+) is coupled, to preserve charge balance, with that of Al3+ in theframework of the feldspar. During exsolution Ca goes into the Ab-richphase (Mason 1982) so that the rate-limiting step in coarsening isdiffusion of framework Al, which is much slower than that of the alkaliions (see Smith and Brown 1988, Sec. 16.2.4 and references therein). Thisis explains the presence, in many normally microperthitic graniticfeldspars (e.g., Shap figs. 3 and 4a), of zones in which the lamellae arecryptoperthitic (Lee and Parsons 1997a). The zones represent chemical

FIG. 4.Micrographs illustrating the role of microtexture in alkali feldspar during weathering and diagenesis. Parts A, B, and D are secondary electron SEM images,Part C is a BSE image. A) (001) cleavage surface of an alkali feldspar phenocryst from the Shap granite that has been briefly etched (, 50 secs) in HF vapor in thelaboratory. Albite has dissolved slightly less rapidly than orthoclase so that the perthitic lamellae stand out as low ridges. The black dots on thicker lamellae aredissolution etch pits on edge dislocations, showing that thicker lamellae are semicoherent. Thinner lamellae lack dislocations and are coherent. The band in the center is anigneous growth zone with slightly higher An content and contains only very fine cryptoperthite. The absence of dislocations makes such zones particularly resistant to HFetching and to weathering (modified from Lee and Parsons 1997a). B) Naturally weathered (001) surface of an alkali feldspar fragment from a peat soil overlying Shapgranite, with semicoherent albite lamellae. The albite has been almost completely dissolved because adjacent etch pits have combined, leaving lines of pits resembling theperforations on tear-off labels. K-feldspar between lamellae is breaking off as thin flakes (from Lee et al. 1998). C) Detrital grain of perthitic alkali feldspar from theFulmar Formation. In parts of the grain, mainly near the edges, albite lamellae (medium gray) have been partly or completely dissolved. Weakened by partial dissolution,the grain has broken during compaction. Subsequently, authigenic K-feldspar (light gray) has formed a partial overgrowth, and formed a lining on both surfaces of thefracture between the two largest grains. Dark grains at the extreme right are quartz and the medium gray grains forming a cement around much of the feldspar aredolomite. D) Same origin as Part B. Weathered (001) surface on which a band of patch perthite cuts across semicoherent film perthite. The irregular areas mark zoneswhere preferential dissolution has occurred along subgrain boundaries within the patch perthite (from Lee et al. 1998).


zoning in An and Csn inherited from magmatic growth and are lesssusceptible to weathering and diagenetic dissolution than the bulk of thecrystal (see Lee et al. 1998 and below).

Semicoherent Strain-controlled Intergrowths.As film perthites andsome nearly straight film mesoperthites and antiperthites coarsen,intergrowths often become semicoherent by developing spaced edgedislocations which can be made visible by HF etching (Fig. 4A; seeWaldron et al. 1994 for practical details). These dislocations are commonin feldspars from subsolvus granites (Fig. 4) and have a crucial role intheir low-temperature reactions during igneous cooling (Lee and Parsons1997b, 1998), weathering (Fig. 4B; see Lee and Parsons 1995; Lee et al.1998), and diagenesis. They also contribute to mechanical degradationduring compaction (Fig. 4C). Each edge dislocation is a line representing

the end of a single 020 lattice plane in the Ab-rich phase where itterminates at the lamellar interface. The dislocations develop as lamellarcoarsening occurs because it is impossible for the Si,AlO framework ofboth phases to remain continuous while sustaining the strains whichdevelop because of the difference in cell dimensions. The strains areminimized by the formation of regularly spaced dislocations on thesurface of the perthite lamellae, which is between (601) and (801), one setparallel to the crystal b axis, the other at right angles in the ac plane. Thedislocations form very extended lens-shaped loops around lamellae(Fig. 3) and therefore appear in pairs straddling lamellae on fragmentsurfaces. A simplified diagram illustrating the relationship is given asFigure 2 in Lee and Parsons (1998). This crystallographically irrationalplane was known as the Murchison plane of easy fracture by earlymineralogists and, with the perfect (010) and (001) cleavages, defines theshape of many detrital alkali feldspar grains. The angles between (010)and (001), and (010) and (601)(801) are 90u in monoclinic feldspars, andthe angles between (601) and (801) and (001) are 95u and 101u,respectively. Thus broken fragments of feldspar tend to be approximately

FIG. 5.Optical micrograph (crossed polarizers) of perthitic microtextures inan alkali feldspar phenocryst in a sample of fresh granite from a quarry in theShap granite. In places the feldspar has been unaffected by deuteric fluids, and isglass-clear, with nearly straight film albite lamellae. Elsewhere turbidity has begunto spread along lamellae, as dissolution and replacement take place on and arounddislocation outcrops. This is the beginning of deuteric coarsening. Veins ofdeuterically coarsened patch perthite also cut across the crystal, probably partlyfollowing fractures.

FIG. 6.Detrital alkali feldspar fragment from the Fulmar Formation showing ambiguous late albitization. A) BSE image. The lenticular lamellae of albite (dark gray)indicate an igneous origin, but the irregular patches of albite are of uncertain origin. B) CL image. The K-feldspar is brilliantly luminescent, and the regular albite lamellaeshow weak luminescence. Albite patches are completely non-luminescent, but their origin is ambiguous. Their presence in the fragment interior suggests that theyoriginated by replacement while in the igneous rock. Had they formed a rim or partial rim to the detrital grain, an authigenic origin would have been inferred.

FIG. 7. Bright-field TEM image of a detrital microcline from the FulmarFormation with fine (sub-optical) tartan twinning and cryptoperthitic albitelamellae (lower right). The grain has a diagenetic overgrowth of adularia in theform of tightly packed subgrains (upper left). Despite the wealth of microtexture inthis image, the grain and overgrowth would look clear and featureless in OM.However, the overgrowth does not exhibit CL, in contrast with the detrital core.(From Lee and Parsons 2003).


rectangular from whichever crystallographic axial direction they areviewed.Dislocation outcrops on grain surfaces or internal fractures are

extremely reactive in aqueous fluids, and the role of various defects infeldspar weathering has been recognized for many years (e.g., Berner andHoldren 1977, 1979). Dislocations are strongly reactive because of Gibbsfree energy contributed by undirected bonds in the core and because ofelastic strain energy in the disturbed crystal structure surrounding thecore. Blum and Lasaga (1987) used Monte Carlo simulations todemonstrate the importance of dislocations in feldspar and quartzdissolution. TEM work currently in progress (Fitz Gerald and Parsons2003) on alkali feldspar phenocrysts from the Shap granite shows thateven in what appear optically to be the freshest parts of grains fromfresh igneous rock the dislocation cores have been dissolved away,leading to tiny (, 2 nm) tubes. A most important concept is that ofchemical affinity (Velbel 1989), which implies that the relative dissolutionrates of structure close to dislocations and normal crystal structure

differ most when the aqueous solution is close to chemical equilibriumwith the solids. This situation is likely to pertain in confined spaces inigneous rocks, in sedimentary basins, or in established soil waters.Figure 4B shows the naturally weathered surface of an alkali feldsparfrom a peat soil, in which etch pits on dislocation outcrops havecombined near the grain surface, leading to dissolution of albite andmechanical degradation (see Lee et al. 1998 for a detailed discussion).

Deuteric Intergrowths.These form subsequent to coherent andsemicoherent exsolution and replace strain-controlled microtextures.They are up to 103 times coarser than coherent intergrowths and areeasily visible using OM. The microtexture is profoundly different tostrain-controlled intergrowths (e.g., Figs. 4D, 5, 8, 10 (right), 11, and 12)and is composed of a composite of subgrains on a range of scales froma few tens of nm to several mm. TEM micrographs of such intergrowthscan be found in Worden et al. (1990), Guthrie and Veblen (1991),Waldron and Parsons (1992), Walker et al. (1995), and Lee et al. (1995).Coherency between the Ab- and Or-rich phases has been lost (they areincoherent), the intergrowths are essentially strain-free, and phasecompositions will be defined by the strain-free solvus (SFS in Fig. 1). Thedriving force for this recrystallization is loss of coherency strain energy(Brown and Parsons 1993) during reactions between the feldspar anddeuteric or hydrothermal fluids. Alkali feldspars which have escaped suchreactions are rare in plutonic rocks, although volcanic sanidines are morecommonly unaffected. In some rocks the alkali feldspars have beenrecrystallized in their entirety (e.g., Fig. 11), but in many cases (such asthe Shap feldspars illustrated in Figs. 3, 4, and 12) parts of the originalcrystal have strain-controlled microtexture while elsewhere they aredeuterically coarsened.Deuteric intergrowths are turbid in plane-polarized light. It is usually,

although not invariably, the Or-rich phase which is the most turbid.Although frequently ascribed to clay minerals, turbidity is overwhelm-ingly caused by myriads of sub-mm micropores (Worden et al. 1990;Walker et al. 1995), usually developed at the junctions between subgrains.Very turbid detrital alkali feldspars in sedimentary rocks probably owetheir appearance to enlargement of preexisting pores and growth of clayminerals, although some clay particles are imported from the protolithigneous or metamorphic rock (see e.g., Worden et al. 1990, their fig.8).The intergrowths are irregular, and various descriptive terms are useful:

patch perthite, when each phase forms shapeless patches (Fig. 10, right);and vein perthite, when the lamellae form sinuous bands (Figs. 11, 12).Note, however, that at a scale smaller than the patches and veins

FIG. 8.BSE image of Fulmar sandstone showing four detrital feldspar grainswith three distinct provenances. The two bright featureless grains at the left areboth volcanic sanidines. The right hand grain has regular lenticular albite lamellae,and is likely to have come from a granitic protolith, and the central grain isa deuterically coarsened perthite, with a few more regular lamellae, also froma granitic protolith. In this case the grain has been subject to fluidrock reactionsbelow , 450uC.

FIG. 9.A) BSE and B) CL image of a detrital sanidine (left) from the Fulmar Formation, with strong BSE and CL signals. A strongly resorbed authigenic K-feldsparovergrowth has a strong BSE signal but no CL, and a larger area of authigenic feldspar occurs at the right. Although authigenic feldspars may appear optically defect-free, TEM shows that they can be composed of fine-scale subgrain mosaics (Fig. 7; see also Worden and Rushton 1992, Lee and Parsons 2003). Dissolution has affectedthe reactive authigenic overgrowth, leaving the comparatively defect-poor sanidine unaffected, even though they differ little in chemical composition.


themselves the deuteric perthite is actually a mosaic of subgrains. Like thedislocations discussed above, the subgrain boundaries have a stronginfluence on the dissolution of alkali feldspars, and weathered surfacesmay be very irregular where preferential dissolution has occurred alongthem (Fig. 4D).

Deuteric coarsening can occur in a number of ways which reflect theunderlying strain-controlled microtexture and/or external deformation.Changes in the bulk composition of the crystal (giving replacementperthite) may also occur (next section). Because replacement perthite mayform during diagenesis, it is important to be aware of the variety whichcan arise from processes occurring before transport and sedimentation.All types of deuteric coarsening involve interaction between a fluid filmand the feldspar which leaves a microporous microtexture behind anadvancing recrystallization front. Putnis (2002) discusses such processesin minerals in general. In alkali feldspars three broad types of process canbe inferred:

(i) Fully coherent lamellar and braid mesoperthites recrystallizepervasively along a sharp interface (Fig. 10). A fluid film movesforward along a broad front into the crystal, releasing (unzip-ping) the elastic strain in the coherent intergrowth and pre-cipitating a strain-free, coarsened and microporous mosaic in itswake. The exact details are complex (Lee et al. 1997; Brown et al.1997).

(ii) In originally semicoherent intergrowths deuteric coarsening may beguided by dislocations on lamellae. Figure 5 shows a typical filmmicroperthite from the Shap granite. Areas of fresh film perthiteare glass-clear, and the lamellae can be seen because of thedifference in refractive index or birefringence. Deuteric alterationleads at first to faint turbidity which extends along lamellae. Thetiny dusty particles can be shown using SEM to be pores(Walker et al. 1995) or minute grains of new albite (Lee andParsons 1997b, their figs. 24). Both features nucleate ondislocations. These small albites increase in size and coalesceleading to areas of patch perthite.

(iii) Deuteric coarsening also occurs along fractures through crystals(Figs. 5, 12); this feature, where coarse irregular perthite forms

FIG. 10. BSE image of braid cryptoperthite (bottom left) that has partlyrecrystallized to a patch perthite during deuteric reactions in the cooling igneousrock. From the Klokken syenite, South Greenland. The surface is viewed ina direction approximately parallel to the Z crystallographic axis; X is horizontaland Y vertical. The low albite phase is dark gray, the low microcline phase is lightgray. Black dots are micropores, rare in areas with the braid texture. The patchperthite areas are irregular mosaics of Ab- and Or-rich subgrains at a scaleconsiderably smaller than the patches themselves.

FIG. 11.A) BSE image of crystals of coarse vein perthite in a quartz syenite sheet cutting the Klokken syenite (compare Fig. 10). Hypersolvus alkali granites usuallyhave feldspars with similar microtextures. Note the low magnification. Black crystals are quartz; brightly emitting grains are mafic silicates. B) Higher-magnificationimage of small area of Part A, showing a relic of braid texture in vein perthite. The black dots are micropores, rare in braid areas where they are aligned and follow healedcracks. (From Lee et al. 1997).


veins across grains, is common in alkali feldspars from granites,and usually coexists with, and grades into, more pervasivecoarsening of type (ii). It perhaps forms because of fluid flowalong fractures which develop in response to external stresses. Inthe case of the Shap feldspar illustrated, EPMA analysis showedthat the albite in these veins is an oligoclase of compositiondifferent from the exsolved Ab-rich phase, so this is strictlya replacement perthite (Lee and Parsons 1997b).

Replacement Textures.In the case of regular coherent and semi-coherent intergrowths we can be confident that the microtexture formedby exsolution which occurred during cooling following crystal growth athigh igneous or metamorphic temperatures. Such intergrowths cannot beproduced by any known low-temperature process. At the scale of

individual crystals the process is isochemical. In contrast, irregularintergrowths introduce the possibility that they form non-isochemically,by a replacement process. Non-isochemical replacement is well known tooccur during diagenesis, replacement by albite (often called albitization)being most common (Ogunyomi et al. 1981; Walker 1984; Morad 1986;Saigal et al. 1988) although introduction of K-feldspar also occurs(Milliken 1989; Morad et al. 1989; Lee and Parsons 1998; Lee andParsons 2003). However, both these processes have also been shown tooccur in cooling igneous (Ferry 1985; Aldahan et al. 1987; Ramsayer et al.1992; Lee and Parsons 1997b) and metamorphic rocks (Pryer and Robin1996; Cayzer 2002).

Unambiguous criteria for distinguishing diagenetic replacement fromnon-isochemical replacement in an igneous or metamorphic protolith arenot easy to establish. In an igneous rock deuteric microtextures are likelyto be isochemical if a large number of individual alkali feldspar crystalshave the same bulk composition (see Brown et al. 1983). However, thisapproach is unworkable in a siliciclastic rock. Areas or veins of Ab-richfeldspar which are not pure Ab but which contain a few percent of An arelikely to be due to non-isochemical replacement of late igneous origin. Ina study of low-temperature replacement in the Shap granite, Lee andParsons (1997b) found that replacive plagioclase forming veins andshowing strong blue cathodoluminescence was Or1.5Ab89.8An8.7. Within, 1 mm of the crystal surface, turbid, non-cathodoluminescent replacivealbite Or0.7Ab99.1An0.2 also occurred, in part penetrating along preexist-ing semicoherent perthitic albite lamellae. However, it would be extremelydifficult to distinguish such albite from diagenetic albite once fragmentedand transported into a sedimentary rock. Figure 6 shows a detrital grainwith patches of non-cathodoluminescent albite, which is similar to thepatches of pure replacive albite seen in the Shap granite and thereforecould be of igneous origin. A continuous or semicontinuous rim of suchalbite would unambiguously indicate a diagenetic origin. It is possiblethat trace-element or isotopic criteria might exist to distinguish thesepossibilities.

PREPARATION, ANAYTICAL, AND IMAGING TECHNIQUES

The techniques adopted in the present study were designed to providea rapid survey of as many samples as possible by SEM, including EDX,which then could, if required, be followed by OM and TEM of selectedgrains. Optical microscopy has inadequate resolution to distinguish mostof the types of feldspar intergrowth encountered in siliciclastic rocks, andthe general turbidity of many detrital feldspar grains makes identificationdifficult. In our study of the Fulmar Formation samples were initiallymade into resin-impregnated polished blocks for SEM, which could thenbe turned over and made into double polished sections for OM and TEM.While TEM can be very informative in diagenetic and provenance work(Worden and Rushton 1992; Lee and Parsons 1998; Lee and Parsons2003; Lee et al. 2003) it is time consuming and unlikely to becomea routine method.

The polished blocks were characterized petrographically (i.e., grainsize, shape, mineralogy, and porosity) using BSE with a Philips XL30CPSEM at Edinburgh University operated at 20 kV. All varieties of feldsparobserved were recorded. Detrital grains were classified using the criteriaprovided and illustrated in the next section. Authigenic K- and Na-feldspar overgrowths and replacement, separate authigenic grains, andevidence of corrosion of authigenic overgrowths and detrital grains wererecorded. The general type of each feldspar grain encountered alonga regular stepped traverse in the SEM was estimated by qualitative EDXusing an Oxford ISIS or PGT Spirit EDX system. This point-counting didnot include all feldspar microtextural types in a sample, some of whichoccurred as only a single grain, but it provided a reasonably quantitativeguide to the main varieties of feldspar present. An advantage of the EDXpoint-counting technique is that it allows An-bearing plagioclase to be

FIG. 12. Collage of BSE images of a typical area of film, vein, and patchperthite in a phenocryst from the Shap granite, viewed approximately parallel tothe Z axis. Similar microtextures are found in most subsolvus granites, withvariation imposed by bulk composition, in this case , Ab25Or75, and the degree ofdeuteric alteration. A 50 mm detrital grain might sample regions with strikinglydifferent microtextures. Albite is dark gray, K-feldspar is medium gray, black dotsare micropores. Regular lamellae near MP are semicoherent films of albite inorthoclase, as seen on the etched surfaces in Figure 4. Elsewhere, ragged veins ofalbite, Ab, correlate with regions rich in micropores. Featureless K-feldspar in thevicinity of the albite veins is intermediate microcline, Im. This material and thealbite veins and patches are an incoherent mosaic of subgrains, and the irregulartexture, with its micropores, marks feldspar which has pervasively recrystallized ina deuteric fluid. (From Lee and Parsons 1997b).


distinguished from An-poor plagioclase by the presence of a Ca peak inEDX spectra. Lee et al. (2003) suggest a detection limit of , 1.5 mol%An. A disadvantage of this technique is that K-feldspar and quartzovergrowths could not be confidently distinguished from their detritalcores in the BSE images alone during the point-counting procedure.However, corresponding CL images obtained in the SEM could be usedto identify overgrowths from their low luminescence intensities. It was notpractical to use CL when point counting because of the time taken toacquire an image. An advantage of the SEM/EDX technique is that itdistinguishes clear, unexsolved, and untwinned grains of feldspar fromquartz, which is difficult during OM point counting. We have detectedsignificant amounts of feldspar in samples logged commercially asfeldspar-free, presumably because turbidity was the main criterion usedto differentiate feldspar. For each sample a total of over 450 points werecounted to ensure that abundances as low as 5% were meaningful, witha relative error of 20% at the 95% confidence level. We found thatporosities estimated by the SEMmethod were larger, by a factor of two inmany cases, than those obtained using OM on the same core bycommercial logging. For point counting by OM the rock samples arenormally impregnated with blue resin prior to making the thin section sothat holes produced subsequently by plucking will be obvious. Possiblyporosities may be overestimated in our point-counting statistics becauseof plucking of grains during sample preparation for the SEM. Converselyit is likely that in the SEM it is easier to detect smaller pores.

Where necessary the blocks were made into thin sections and feldsparswere analyzed quantitatively using a Cameca SX50 electron probe in theDivision of Earth Science, University of Glasgow and a CamecaCamebax electron probe in the Grant Institute of Earth Science,University of Edinburgh. Data were acquired using an acceleratingvoltage of 20 kV and a beam current of 10 nA, and the beam wasdefocused to 20 mm diameter. Where required, following SEM andelectron probe work, it is possible to do TEM using the polished thinsections, provided that they have been mounted using low-melting-temperature resin. For this purpose 3.05 mm diameter copper discs wereattached to the thin section with their central 0.4 mm diameter holepositioned over a grain of interest. The copper discs, with grains attached,were then extracted from the thin section and Ar ion beam thinned untilelectron-transparent using an Atom Tec 700 series thinner operated at, 6 kV (Edinburgh University) and a Gatan Duomill operated at 5 kV(Glasgow University). High-resolution TEM images and electrondiffraction patterns were acquired from the thinned foils using a PhilipsCM120 Biotwin TEM operated at 120 kV (Edinburgh University) anda Jeol 200FX TEM operated at 200 kV (Glasgow University).

A useful auxiliary SEM technique is to etch feldspar cleavage surfacesin HF vapor for short periods (typically , 50 s) followed by secondaryelectron (SE) imaging, as described by Waldron et al. (1994). An exampleis given as Figure 4A. The etching provides pronounced surfacetopography caused by rapid dissolution on dislocation outcrops, inregions of high coherency strain (along coherent or semicoherent perthitelamellar boundaries) and on subgrain boundaries in deuteric inter-growths. Less pronounced relief is produced by the different dissolutionrates of Ab- and Or-rich feldspar. The method has higher resolution thanBSE imaging (see e.g., fig. 2B in Waldron et al. 1994, which showsfeatures , 50 nm across) but requires that natural fresh cleavage surfacesare available. An example of this technique applied to a detrital grain isgiven by Lee and Parsons (1998, their fig. 6c).

RECOGNITION OF ALKALI FELDSPAR MICROTEXTURES

Two factors must be borne in mind when applying our approach:firstly, the strong effect of orientation on the appearance of a perthiticintergrowth when sliced, and secondly that the relatively small fragmentslikely to be found in many siliciclastic rocks may be untypical of the

original grain as a whole. In general, strain-controlled microtextures andestimates of grain bulk composition are the most informative forprovenance purposes (see Fig. 2 and photomicrographs below).Fragments of deuterically coarsened perthite are relatively uninformativeexcept in cases where they are large enough to estimate bulk composition,at least roughly. While some of the strain-controlled intergrowths aresufficiently distinctive that provenance can be suggested from a singlegrain, in other cases it is best to deduce provenance from families ofgrains within the sedimentary rock.

The appearance of a perthitic intergrowth varies with viewing direction(Fig. 3). Regular lamellar intergrowths appear diffuse and irregular oreven patchy if viewed from a direction nearly normal to the lamellarinterface. As a routine method we recommend looking for the mostregular exsolution features visible using BSE imaging. Experiment withmagnification and with the BSE contrast. The presence of any sort ofregular lamellar intergrowth is secure evidence that a grain is detrital. Themorphology of the intergrowths, the relative proportions of the Ab- andOr-rich phases, a CL image, and, if necessary, a qualitative chemicalanalysis can all be obtained in the SEM.

Appreciation of the orientation of exsolution lamellae relative to thetwo perfect cleavages of feldspar is of considerable help in interpreting therandom sections of grains encountered in a sedimentary rock. If OM isused, extinction parallel to the (010) cleavage and the (010) compositionplanes of Albite twins (which often can be seen in the Ab-rich phase ofperthites) are useful guides to orientation. Albite twins are sometimesfaintly visible using SE imaging in the SEM. Cleavages are often visible inBSE images (see e.g., Figs. 13 and 14), and there will be a tendency fordetrital fragments to be bounded by either the (001) and (010) cleavagesor the less regular Murchison plane between (601) and (801) (Fig. 3).Regular lamellar intergrowths imaged approximately normal to theMurchison plane will appear irregular and patchy but will show two well-developed cleavages at right angles. True patch perthites are lessinformative than strain-controlled intergrowths, although their bulkcomposition is some guide to potential source rocks (Fig. 2). They oftencoexist with strain-controlled textures within individual original crystalsor detrital fragments (e.g., Figs. 10, 11, and 12). However, patch perthitescontain large concentrations of subgrain boundaries and micropores, arerelatively easily weathered (Fig. 4D; Lee et al. 1998), and are likely todissolve relatively rapidly during burial. There is a natural tendency forpatch perthites to disappear relatively quickly from the inventory ofclastic feldspars in a given sequence of weathering, transport, and burial.Using BSE imaging alone, alkali feldspar grains free of visible

microtextures may be difficult to distinguish from discrete authigeniccrystals, unless the latter have grown into pore spaces and developed theAdularia habit. The most likely source for detrital fragments of this typeare acid and alkaline volcanic igneous rocks, and bulk composition canprovide evidence of provenance (see diagrams in Chapter 14 of Smith andBrown 1988). Lack of an exsolution microtexture makes such fragmentsresistant to dissolution because of the absence of either dislocations orsubgrain boundaries (Lee et al. 1998), and this probably accounts for therelative abundance of such grains in the Fulmar Formation (see below).Absence of cathodoluminescence and compositions close to end-memberAb and Or (Ab.99 or Or.99) are the best guides to an authigenic origin. IfTEM is used, the presence of cryptoperthite would rule out an authigenicorigin, as would the regular well-developed tartan AlbitePericlinetwinning characteristic of microcline or, less reliably, the tweed textureof orthoclase. It is sometimes possible to see tartan twinning inmicrocline using SEM-CL, because the small variation in the orientationof the twins affects luminescence intensity; this observation would ruleout an authigenic origin. However, Figure 7 shows an example ofa detrital grain and its overgrowth in which all microtextures are sub-optical and which would be indistinguishable using BSE or CL. At theTEM scale the authigenic K-feldspar overgrowth exhibits the diffuse


modulated microtexture characteristic of adularia and the contrastbetween the detrital K-feldspar core and overgrowth is obvious (Fig. 7;see also Lee and Parsons 2003, their fig. 7a, for examples of adulariaveining detrital K-feldspar and replacing perthitic albite).Detrital fragments that may be untypical of a larger original grain as

a whole pose problems best overcome by considering families of grains inany one slice. The problem is most acute in the case of grains whichoriginally had coarse deuteric microtextures but which have beenfragmented. Thus a single grain of Ab-rich plagioclase in a sedimentaryrock might have at least six sources:

1. It originally formed as an Ab-rich rich patch in an alkali feldspar byexsolution

2. It originally formed during relatively high-temperature replacement(albitization) reactions with an alkali feldspar (as shown by Lee andParsons 1997b).

3. It originally formed during low-temperature igneous albitization insuch a parent (Lee and Parsons 1997b).

4. It grew as a discrete plagioclase crystal in a sub-solvus granite ormetamorphic rock

5. It formed by albitization of basaltic calcic plagioclase in an oceanfloor environment (spilitization).

6. It grew as an authigenic grain or overgrowth (since fragmented)during diagenesis.

Fragments of types (1), (2), and (4) will luminesce, but types (3) and (6)will not. Type (5) is uncertain. In general An contents increase from type(1) to type (3), although there is much overlap. Thus Ab-rich phasecompositions in coarse microperthites in the Shap granite (Lee andParsons 1997b) are, An1, in cross-cutting patch perthites, An10, and incoexisting discrete plagioclase , An27. Replacive albite of type (5) is, An0.2, potentially the same as potential authigenic grains (type 6).There are no obvious criteria for distinguishing types (3) and (6). In theHumber Group sandstones of the Fulmar Formation, Lee et al. (2003)showed that detrital plagioclase grains ranged up to An18, and that manywere lamellar peristerite intergrowths. The presence of peristerite inter-growths rules out an authigenic origin, and a source in low-grademetamorphic rocks is implied.The Or-rich phase in patch perthites from granitic source rocks usually

has an appreciable Ab content; in Shap the range is from Or90 to Or97.

FIG. 13.BSE images of fine film perthite grains from the Fulmar Formation, with largely coherent lamellar microperthitic intergrowths. They are most likely to haveoriginated in subsolvus granitic protoliths, or possibly from granulite facies metamorphic rocks (see Fig. 19). If quantitative analysis indicated bulk An contents . 3% thelatter would be strongly suggested. Further examples are given in Figures 6A and 8. A) Fragment with irregularly distributed lamellae of variable size and well defined(001) cleavage (compare Fig. 3). The detrital grain has a partially resorbed K-feldspar overgrowth. B) Similar grain with variable texture such that intergrowths appear tobe absent at the top center. This may be because the intergrowths in this part of the grain are of very fine scale, rather than of variable bulk composition (as in the case ofthe zone of fine intergrowth shown in Fig. 3). Straight edge at the left is the (001) cleavage. The grain has a strongly resorbed K-feldspar overgrowth. C) Grain withmainly straight film lamellae. A few lamellae are kinked, perhaps because of deformation. At the right the grain has fractured along the plane of the exsolution lamellae,the Murchison plane. A partially resorbed K-feldspar overgrowth is mainly developed at the bottom left of the grain. A second alkali feldspar grain on the left of theimage has been partially dissolved, leaving a skeletal relic (see also Fig. 13). D) Grain with fine, relatively evenly developed texture, resembling the platelet zones shownin Figures 3 and 4A, but on a coarser scale. This grain probably has a more albitic bulk composition than those shown in Parts AC. The pale lines (Or-rich) withmicropores could be healed fractures along which deuteric coarsening has occurred, or a diagenetic feature. A K-feldspar overgrowth has been almost completelyresorbed.


The non-endmember composition and luminescence would serve todistinguish fragments from this source from authigenic grains. Further-more, as noted above, optically visible tartan or, when TEM is used, sub-optical tartan and tweed textures are almost universal in detrital K-feldspar from plutonic sources (Fig. 7). Or-rich patches in patch perthitesare usually microcline, often in the irregularly twinned variant known asirregular microcline.

ATLAS OF MICROTEXTURES

In this section we illustrate our method with BSE and CL images (in theSEM) of microtextures in alkali feldspars from the Fulmar Formation,comparing them with examples from our work on igneous andmetamorphic rocks. We have included a few images illustrating distinctivemicrotextures which we have not yet seen in the Fulmar. The worldwiderange of chemical variation in igneous rocks is very large, so we haveconcentrated on the feldspar types that are most likely to be encounteredas detrital grains. Figure 2 can be used as a guide to coherent andsemicoherent microtextures in alkali feldspars from most plutonic igneousrocks. The images begin with feldspars that originated in volcanic rocksand lead through relatively rapidly cooled plutonic igneous rocks to deep-seated igneous rocks and finally to very slowly cooled metamorphic

terranes. Deuteric coarsening is illustrated in these contexts. Some imagesshow grains with several types of microtexture and are cross referencedfrom the relevant subsections. In addition, we comment in separatesubsections on the role of the microtextures in the chemical reactivity andpreservation of the grains, and we show how microtextures facilitatedissolution and replacement reactions. General points are made in themain text; details of interpretation of the micrographs are given inextended captions.

Acid and Alkaline Volcanic Rocks.Provided that grains have escapeddeuteric or hydrothermal reactions, volcanic Or-rich feldspars will befeatureless in OM or show very fine-scale, approximately straightexsolution lamellae. Ab-rich volcanic alkali feldspars (anorthoclase)may in addition exhibit regular cross-hatched tartan twinning, similarto microcline (for making the distinction using OM, see nomenclaturesection). Volcanic alkali feldspars may likewise be featureless using BSEin the SEM (Figs. 8, 9A), or have faint straight perthitic lamellae. In theabsence of visible microtexture at the SEM scale, the best way todistinguish such grains from authigenic feldspars is CL (Fig. 9B), becausethe latter will not luminesce. Alternatively, qualitative analysis in theSEM, showing appreciable Ab content, is a good guide to a detrital originas is a resolvable Ca peak in the EDX spectrum. A high Ab content

FIG. 14.Detrital coarse film perthite grains from the Fulmar Formation, probably from granitic protoliths. These intergrowths are probably largely semicoherent.Parts A, B, and D are BSE images; Part C is a CL image. All the grains have authigenic K-feldspar overgrowths, but those in Parts AC appear to be stable while that inPart D is strongly resorbed. A) Lenticular microperthite with a euhedral adularia overgrowth. The (001) cleavage is visible in the detrital core. Coarse albite lamellae (darkgray) have dissolved, in some cases leaving open pores (black), but neither the detrital K-feldspar nor the overgrowth have dissolved (From Lee and Parsons 2003). B)Similar grain in which some albite lamellae have dissolved while others have survived, with a nearly continuous K-feldspar overgrowth. C) Same grain as Part B. Theimage shows the complete lack of CL from the authigenic overgrowth, which appears as dark as the voids left by the dissolution of albite lamellae. Undissolved albitelamellae show weak CL. D) Lamellae in this fragment have partially dissolved, particularly around the thicker parts of albite lenses, where dislocations are concentratedin semicoherent intergrowths (see Fig. 4A). Areas where the grain is particularly corroded may correspond with patch perthite. The trace of the (010) cleavage can be seenacross the center of the grain.


(Ab.60Or,40), combined with tartan twinning, indicates volcanicanorthoclase. At the TEM scale even rapidly chilled volcanic crystalsare usually exsolved (see Parsons and Brown 1984, their fig. 5, for anexample), particularly if the bulk composition is in the middle region ofthe feldspar solvus. Presence of exsolution textures is unequivocal proofthat a grain is detrital rather than authigenic. However, observing fineexsolution microtextures using TEM in randomly oriented grains isdifficult, particularly when bulk compositions are close to the Or endmember. A more practical technique in the case of volcanic K-feldspars iscareful EPMA analysis for major and trace elements. The compositionalrange of volcanic alkali feldspars can extend continuously from nearlypure sanidine through ternary compositions to the plagioclase field, andthe bulk composition can be a guide to possible parental rocks. Werecommend comparing analyses with the compilations given by Smithand Brown (1988, Chapter 14). However, it should be noted thatphenocryst and groundmass assemblages in individual volcanic rocksmay show a considerable range in composition and strong chemicalzoning (see Brown and Parsons 1994 for a brief treatment, and Brown1993 for details of this complex subject), so that a varied population ofvolcanic feldspar fragments in a sedimentary rock need not imply a variedprovenance.

Reactivity and Preservation.The Fulmar Formation contains a highproportion of grains of sanidine which are devoid of microtextures at theSEM scale (Figs. 8, 9), but which show rare fine Albite-twinned albitelamellae at the TEM scale. The bulk composition of these grains isunusually Or-rich (typically Or88 to Or96), and some have high Bacontents (up to Csn4.5) (Mark Wilkinson, personal communication 2003).The most likely source is ultrapotassic volcanic or hypabyssal rocks,although where these extremely rare rock types might be located in theNorth Sea catchment is not known. In principle, a perfect crystal ofsanidine, the structurally disordered form of K-feldspar stable at hightemperatures, should dissolve slightly more rapidly in aqueous solutionsat reservoir temperatures than a perfect crystal of microcline, the orderedform stable at low temperatures. However, the effect of microtexture ondissolution kinetics far outweighs the effect of structural state, and thelack of microtexture in the sanidine grains probably accounts for theirrelative abundance throughout the Fulmar. We would expect freshvolcanic feldspars to be considerably more resistant to degradation anddissolution during weathering, transport, and diagenesis than potentiallymuch more common feldspars from granitic rocks (see Lee and Parsons1995, 1998; Lee et al. 1998), because of the absence of dislocations onperthite lamellae and of boundaries to subgrains formed by deutericcoarsening in the latter (see Fig. 4). Milliken et al. (1989) showed that inthe Frio Formation of South Texas, where the feldspar provenance isknown, Na-bearing volcanic feldspars dissolved more rapidly with burialthan more Or-rich grains. This might be accounted for by the abundanceof exsolution textures in the former. Figure 9 shows that in the Fulmarthe volcanic grains are less reactive than chemically similar authigenic K-feldspar overgrowths, because the latter are made of subgrains and aremicroporous (Fig. 7, top left).

Hypersolvus Acid and Alkaline Plutonic Igneous Rocks.Hyper-solvus igneous rocks crystallize at temperatures above the strain-freeternary feldspar solvus (Fig. 1) and therefore contain a single feldsparphase prior to exsolution. They characteristically form relatively small,anorogenic plutons of alkali granite, syenite or nepheline syenite. Thealkali feldspars show a wide range of coherent and semicoherentexsolution microtextures depending on bulk composition (Fig. 2) andcooling rate, modified in various degrees by deuteric coarsening. Forphase-equilibrium reasons (Tuttle and Bowen 1958) the bulk composi-tions of the alkali feldspars are strongly concentrated in restricted ranges,, (Ab6555Or3545)10097An03 in syenites and nepheline syenites, and

(Ab6050Or4050)10098An02 in alkali granites, around 4 in Figure 2. Theanorthite content is low (, An3) in most examples, but in syeniticmagmas formed by fractionation, feldspar bulk compositions or chemicalzonation within crystals vary continuously from the plagioclase join alongcurved paths roughly parallel to the path 1234 in Figure 2 (see Parsonsand Brown 1988, and Brown and Parsons 1988, for examples). However,detrital fragments representing a sample of hypersolvus igneous proto-liths would overwhelmingly lie in the restricted ranges given above. Theseare comparatively uncommon rocks, so that the feldspars have consider-able potential as provenance indicators.In syenites, provided that deuteric feldsparfluid reactions have not

occurred, braid perthites (Fig. 2 around 4, and Fig. 10) are common. Theaverage repeat distance of the threads of microcline ranges from, 40 nm to . 500 nm (Brown et al. 1983) and depends on the coolingrate of the igneous body and An content (Brown et al. 1983; Brown andParsons 1988; Waldron and Parsons 1992). Grains may have zigzag,wavy, or straight lamellae (Fig. 2, areas 5, 6, 7), but these types are lesscommon than the braid intergrowth. The coarsest textures are just visiblein OM, but all but the finest are readily imaged using BSE (Fig. 10, left).The finest intergrowths require TEM or use of HF etching and SE SEM(see, e.g., Waldron et al. 1994, fig. 2a). Deuteric reactions lead tocoarsening of intergrowths by up 103, giving irregular patch perthites(Fig. 10, right). The coarsened areas are characterized by micropores(black dots in Fig. 10) at the boundaries of sub-mm subgrains (see Wordenet al. 1990 for detail of the coarsening process).Although the feldspars in hypersolvus alkali granites have bulk

compositions similar to those in syenites, they are usually characterizedby vein perthites much coarser than braid textures (Fig. 11A). Rare relicsof braid texture can sometimes be found (Fig. 11B). The example given isfrom a sheet of quartz syenite, an evolved late member of the Klokkenintrusion with a higher water content than the syenite magmas. Lee et al.(1997) concluded that the vein perthites formed by almost completelypervasive deuteric recrystallization of braid perthites, reflecting the highwater content.

Reactivity and Preservation.We have not seen feldspars of these typesin the Fulmar, although the Permian alkaline rocks of the Oslo graben,which are potential provenance rocks for the Fulmar, contain braidperthites (Muir and Smith 1956). Because braid perthites are fullycoherent we would expect them to be considerably more resistant todissolution than associated patch and film perthites, in which microporesand subgrain boundaries will be rapidly attacked by fluids (e.g., Fig. 4D).

Subsolvus Granitic Rocks Including Acid Gneisses.Alkali feldspars insub-solvus rocks crystallize in equilibrium with a plagioclase feldspar onan isothermal tieline such as PBAF in Figure 1. The majority of alkalifeldspars in subsolvus igneous granites and granitic gneisses are in therange Ab25Or75Ab10Or90, usually with , 3% An (region 8 in Fig. 2)although a few may lie outside this range. Figure 12 shows a typical areaof a phenocryst from the Shap granite. Regular strain-controlled perthitescharacteristically form nearly straight film lamellae which are initiallyfully coherent but which develop edge dislocations (become semicoherent)as they coarsen (Fig. 4A). When viewed parallel to Z the exsolutionlamellae form short lenses; they are straighter and less markedly lenticularviewed parallel to the Y axis. In more Or-rich bulk compositions lamellaeare less conspicuous than in relatively Ab-rich feldspars like those fromShap (near Ab25Or75). The thickest lamellae are a few mm thick (easilyvisible using OM and BSE in an SEM), but lamellae only a few nm thickmay occur (Lee et al. 1995; also Fig. 4), below the resolution of the SEM.Images of perthitic feldspars found in the Fulmar Formation which aremost probably from granitic protoliths are shown in Figures 13 to 15.Deuteric coarsening, giving either ragged vein perthites or patch

perthites (Figs. 12 and 3), is common in granitic K-feldspars. If the bulk


composition of the fragment is in the Ab25Or75Ab10Or90 range it isreasonable to assume that these textures formed isochemically duringigneous cooling of a granitic protolith. Figure 14A shows a grain of thistype. However, as we noted above (Fig. 6), non-isochemical replacivealbitization can occur during both igneous cooling and diagenesis. Thusthe regular film intergrowths and grain bulk composition are the mostreliable provenance indicators.

Reactivity and Preservation.Figures 13 and 14 illustrate the effect ofdislocations on the reactivity of perthitic feldspars. The lamellae inFigure 13 are relatively thin, and the grains are relatively Or-rich. From thelack of dissolution features we infer that they are largely coherentintergrowths. There is no sign of dissolution preferentially affectingexsolution lamellae or their boundaries. Lamellae in Figure 14 are coarser,with thicker lenticular cross sections, and some of have been preferentiallydissolved while others have been preserved. We infer that these thickerlamellae have dislocations along their boundaries and those that wereexposed at the surface of the grains have dissolved leaving lenticular voids,while those not reaching the grain surface have been preserved. Figure 14Dshows an early stage in this process. Similar behavior was noted ina Carboniferous conglomerate at Shap (Lee and Parsons 1998).

Coarse patch and vein perthites in subsolvus granitic feldspars havecomplex subgrain microtextures and are very reactive during weathering(Fig. 4D) and are easily corroded during diagenesis (Fig. 15BD). Theyare likely to be the first grains to be lost from the alkali feldspar inventoryof a siliciclastic sediment, and a preponderance of fine intergrowths(Fig. 13) is likely to be encountered in suites of detrital alkali feldsparsfrom a granitic protolith.

Low-Grade Metamorphism of Granitic Rocks.Some alkali feldsparsexhibit a distinctive texture know as flame perthite. Detrital examplesare shown in Figure 16. The Ab-rich phase in flame perthites adoptsdistinctive sinuous shapes, which sometimes bifurcate. The lamellae arerather coarse and are irregularly developed throughout the K-feldsparhost and are often concentrated around the margins of grains. The genesisof flame perthite has been discussed by Pryer et al. (1995) and Pryer andRobin (1995, 1996). They suggest that the flames represent a coherentreplacement perthite although they do not provide electron-microscopeevidence of the character of the interfaces. Flame perthite occurs ingranitic rocks sensu lato which have been subject to brittle deformation inthe greenschist facies, much less commonly in the lower amphibolitefacies. The texture does not occur in cataclastic rocks below the

FIG. 15.BSE images of detrital patch perthite grains from the Fulmar Formation, Part A largely undissolved, Parts BD showing different stages in dissolution. A)Patch perthite grain. The sub-regular distribution of albite (darker gray) throughout the entire grain, the edges of the albite patches parallel to faint relics of film lamellae,and the bulk composition of the grain all suggest that this is an original patch perthite from a granitic source. The grain has a partially resorbed authigenic K-feldsparovergrowth along its lower margin. B) Corroded grain in which much of the left-hand end is patch perthite (dark gray), much of which has dissolved, giving elongatepores (black). The right-hand end is devoid of visible microtexture and has not dissolved. The grain has a partially resorbed overgrowth (right lower corner). C) Highlycorroded patch perthite in which only a few relics of perthitic albite have survived so that the grain is largely skeletal. A large patch of albite right of center is probably theresult of diagenetic albitization. D) A feldspar grain at the right is very heavily corroded and unrecognizable. The grain at center left, a film perthite that has beenalbitized, has survived.


greenschist facies. Thus flame perthite seems to have a relatively limitedmetamorphic provenance.

Reactivity and Preservation.Flame perthites are uncommon and havea limited paragenesis. At first sight it is surprising that they have beenfound in the Fulmar in abundance similar to film and patch perthitesfrom common granitic rocks. However, the two examples shown from theFulmar (Fig. 16) are devoid of the selective dissolution seen in othercoarse perthites (Figs. 14, 15), which supports the contention of Pryer andcoworkers that the flames are coherent. Thus the relative abundance ofthe flame perthites in the Fulmar reflects microtexture that is resistant todegradation rather than a widespread source.

Granulite Facies Metamorphic Rocks and Charnockites.High-grademetamorphism in the granulite facies may involve crystallization of alkalifeldspar at very high temperature and pressure, sometimes . 1050uC and1.2 GPa. At these conditions, because of the shape of the ternary feldsparsolvus (Fig. 1) it is possible for ternary alkali feldspars with high Ancontents to grow provided that water activities are low enough topreclude melting. The exsolution microtextures are varied and here weconcentrate on the commonest and most distinctive types using examplesfrom Antarctic granulite facies suites studied by Cayzer (2002). Otherillustrations or discussions of granulite feldspar microtextures can befound in Hubbard (1965), Carstens (1967), Machado (1970), Kay (1977),Yund et al. (1980), Hayob et al. (1989), Waldron et al. (1993),Evangelakakis et al. (1993), and Raase (1998, 2000). Charnockites aregranitic rocks emplaced under dry granulite-facies crustal conditions;their feldspar microtextures are similar to those in metamorphicgranulites.

The range of ternary feldspar bulk compositions in granulites is large,varying almost continuously from antiperthites in the vicinity of 1 (Fig. 2)through mesoperthites near 3 and 5 and perthites near 8 and extending tonear the AbOr join, essentially in the field of granites (Fig. 12).Feldspars with large amounts of all three feldspar components are oftensaid to be strongly ternary and this characteristic of granulitic feldsparsis matched only by those in some uncommon syenitic rocks (seehypersolvus igneous rocks, above; Parsons and Brown 1988). Fourmicrotextural features are distinctive of granulite-facies alkali feldspars:(1) the microtextures have become relatively coarse (. or& 1 mm) whileremaining coherent or semicoherent; (2) there is often evidence for at leasttwo stages of coherent or semicoherent exsolution; (3) micropores and

therefore turbidity are rare or absent; (4) patch perthites are uncommon.All these features are consistent with growth at high temperature followedby relatively slow cooling under dry conditions.

Micromesoperthites (Fig. 17) are one of the most distinctive granulitefacies feldspar types; they are common in the Fulmar Formation (Fig. 18).The film lamellae are sinuous and the intergrowths are coarse comparedwith the lamellae in syenites, yet dislocations are rare or absent (Fig.17A,B). There are superficial similarities to the vein mesoperthites found inhypersolvus granites (compare with Fig. 11), but the lamellae in Figure 11are ragged and pervaded by micropores (Fig. 11B), which are almostcompletely absent from the granulite feldspars. Granulite mesoperthiteswith coarse, stubby lenses of oligoclase (Fig.17C) and complex inter-growths (Fig. 17D), with two lamellar orientations, are also encountered.The origin of the latter texture is uncertain but is probably the result ofreorganization of an original mesoperthite under the influence ofdeformation (Cayzer 2002). A similar texture was illustrated by Hokada(2001, his fig. 2c). As far as we are aware such textures are distinctive ofthe granulite facies.

A special case of granulite facies mesoperthite was reported by Hayobet al. (1989) in xenoliths in Quaternary volcanic rocks from Mexico.Unusually, they found that the intergrown phases in these mesoperthiteswere far from the Ab and Or end-members, and they deduced that thecooling history had been abruptly terminated at 800950uC, when theyascended rapidly from hot lower crust. Discovery of coarse mesoperthiticintergrowths with non-end-member phase compositions in detrital grainswould be an extremely distinctive provenance indicator.

Perthites (sensu stricto) in the granulite facies (Fig. 19A) may formstraight lamellae very similar to those in feldspars from subsolvus granites(Figs. 12, 13), and it may be difficult to assign a granulite-facies originunequivocally to a detrital fragment of such a feldspar. However, veryregular lamellae, with few dislocations and micropores, without deutericcoarsening, and relatively highAn +Csn,would strongly suggest a granulitesource. Two-stage perthites (Fig. 19BD) are very distinctive of thegranulite facies and can also be matched in the Fulmar (Fig. 20A, B),although care must be taken to differentiate possible patchy diageneticalbitization (Fig. 20C), using CL (Fig. 20D). Two-stage perthites containboth fine (usually sub-optical) lamellae and much coarser rounded blebswhich often have coarse Albite twinning (Fig. 19C). The blebs aresemicoherent when small, incoherent when large, and often form trailsacross grains (Fig. 19D). Two-stage perthites have been described by manyauthors (Eskola 1952; Machado 1970; Yund and Ackermand 1979; Yund

FIG. 16.BSE images of detrital flame perthite grains from the Fulmar Formation. There is no sign of corrosion guided by microtextures, suggesting that theseintergrowths are coherent. A) The large grain is a well preserved fragment of coarse flame perthite. Note the distinctive shape of the albite lamellae, their tendency tobifurcate, and their heterogeneous distribution. The two smaller feldspar grains are sanidines, devoid of microtexture at this scale. That on the right has an authigenic K-feldspar overgrowth. B) Flame perthite fragment with areas of film perthite, particularly at the lower left corner of the grain.


et al. 1980; Waldron et al. 1993; Evangelakakis et al. 1993; Raase 1998),and they occurred in all four Antarctic terranes studied by Cayzer (2002).Antiperthites also occur in granulite facies rocks (Fig. 21). They are

unusually coarse. Although in principle they should be common,antiperthites are not often reported in the petrological literature, possiblybecause they are usually suboptical in scale (cryptoantiperthites; seeBrown and Parsons 1988). The textures shown in Figure 21 consist ofblebs (rods in three dimensions) of a range of coarseness and a variety ofshapes, and they arise by a combination of exsolution followed byreplacement reactions. Aligned rods with a distinctive rhombic crosssection (Fig. 21B) appear to have nucleated on Albite twin planes in theplagioclase phase (Cayzer 2002). The strongly ternary bulk composition(caption to Fig. 21) would be highly suggestive of a granulite-faciesorigin. We have not seen antiperthites of this type in the FulmarFormation.

Reactivity and Preservation.Considering that high-grade granulitefacies rocks are not abundant, and often occur as relatively small enclavesin regions of lower-grade metamorphism, granulite-facies alkali feldspars

are remarkably abundant in the Fulmar. This is likely to be because of thegeneral lack of dislocations on regular lamellae, and lack of deutericrecrystallization in feldspars from these very dry rocks. Figure 18Dillustrates a small but well-preserved detrital fragment of granulite-faciesmesoperthite in contact with a much more corroded alkali feldspar,perhaps originally a granitic patch perthite.

DISCUSSION: PROVENANCE AND RESERVOIR QUALITY

The micrographs above (Figs. 821) illustrate the ability of exsolutionmicrotextures in alkali feldspars to provide provenance information insiliciclastic rocks. They cover the most common textural types that occurworldwide, many of which have been encountered in the FulmarFormation. The alkali feldspar inventory of the Fulmar Formationcontains a preponderance of types which on a North Sea catchment-widescale would be comparatively rare. This is particularly true of the volcanicsanidines and the mesoperthites and two-stage perthites fromgranulite-facies rocks. Feldspars from subsolvus granites, which shouldon the basis of areal extent at outcrop be common, do not predominate

FIG. 17.Granulite facies mesoperthites, from the Napier Complex, Antarctica (Cayzer 2002). A) SE image of an etched (001) cleavage surface of a film mesoperthite,bulk composition Ab35.3An3.2Cn1Or60.6. The Ab-rich lamellae, which stand out in relief, are relatively coarse but almost fully coherent, the whole image being nearlydevoid of etch pits on dislocations. There are three pairs of etch pits, one at top right and two halfway up at the left. This is a true mesoperthite in which both phases arecontinuous in three dimensions (contrast the more Or-rich granitic feldspars in Figs. 1214, where there is a clear Or-rich host phase and the albite is discontinuous). B)BSE image of a granulite-facies gneiss showing many mesoperthitic feldspar crystals. Pale gray is the Or-rich phase, mid gray is the Ab-rich phase. The feldspar bulkcomposition is strongly ternary, Ab49.2An19.5Cn0.1Or30.8. The high An content, the smoothly sinuous character of the lamellae, and the absence of micropores arediagnostic of the granulite facies. Some of the grains have continuous plagioclase rims, another distinctively granulite feature. C) SE image of an etched (001) cleavagesurface of lenticular mesoperthite, bulk composition Ab32.0An9.5Cn0.2Or58.4. There are some extended lamellae, as in Part A, but the majority have formed stubby lensesand appear to be discontinuous in the K-feldspar host. Although the lamellae are coarse there are few dislocation etch pits, although some pairs are distributed over themicrograph. Other parts of this sample have more etch pits and are more semicoherent. D) A complex, very coarse ternary mesoperthite, bulk compositionAb55.4An17.0Cn0.6Or26.9, which has two, almost perpendicular lamellar orientations. Micropores are almost completely absent.


and are often strongly corroded. The relative abundance of detrital alkalifeldspar types is not a simple crustal average, but reflects the relative easeof destruction of the original feldspars during weathering, transport, anddiagenesis. The feldspar types that degrade most rapidly do so notbecause of their chemistry but because of the defect inventory of thegrains, particularly the presence of edge dislocations on semicoherentalbite lamellae or subgrain boundaries and micropores in deutericallyrecrystallized feldspars. Both of these microtextural features are absent orrare in the feldspar types which are most frequently preserved.Dissolution is strongly texture-sensitive, and

Alkali Feldspar Microstructures as Provenance Indicator_921

Documents

alkali feldspar microtextures

detrital grains

detrital feldspars

intracrystal microtextures

microtextures form

use of microtextures

feldspar overgrowths

provenance indicators