Kaolinitization of Biotite: Rem Data and Implications for an Alteration Mechanismt

American Mineralogist, Volume 72, pages 353-356' 1987

Kaolinitization of biotite: rEM data and implications for an alteration mechanismt

JuNc Ho AHNr2 DoN.ql-o R. PBlcon Department of Geological Sciences,The University of Michigan, 1006 C. C. Little Building, Ann Arbor, Michigan 48109,U.S.A.

Ansrucr Partially kaolinitized biotite from Brighton, New Znaland, was studied by using transmission and analytical electron microscopy (rErraand aer',r).Kaolinite occursin two distinct modes: (l) as 50-300-A-thick packetsof layersinterstratified within biotite and (2) as two- layer units (14 A) irregularly interlayered within biotite. Some two-layer kaolinite units terminate at single biotite layers (10 A), implying a reaction of one biotite layer to two kaolinite layers. Structural and chemical considerationssuggest an alteration mechanism consisting of complete dissolution of biotite and crystallization of kaolinite at linear boundaries. Tunnel-like structures inferred at reaction boundaries may serve as conduits for diffusion of reactant and product ions. Such linear imperfections may be significant in mass transport during the alteration of large rock volumes.

yield images of the interfaces where biottte alters INrnooucrroN direct to kaolinite. craw et al. (1982) indicated that kaoliniza- phases, Biotite can react to form a variety ofdifferent tion in thesesamples appears to have been accompanied depending on alteration and weathering environments, by hrgh activity of COr, resulting in partial loss of epidote (Walker, including vermiculite 1949), montmorillonite and producing rutile and/or anatase at the expense of (MacEwan, (Veblen Ferry, 1983;Yau 1954),chlorite and sphene, without oxidation of siderite. Alteration was gibbsite(Wil- et al., 1984;Eggleton and Banfield, 1985), therefore not a simple weathering process. son, 1966),or kaolinite (Wilson, 1966;Stoch and Sikora, 1976;Muenier and Velde, 1979; Craw et d., 1982).Al- ExpnnrunNTAL METHoDS though kaolinitization is a common biotite alteration Sampleswere first preparedas thin sectionsoriented approx- phenomenon that has been investigated by optical mi- imatelyperpendicular to (001)of the partiallyaltered biotites. croscopy and powder X-ray diffraction (xno) (e.g., Wil- Followingoptical examination, selected areas were ion-milled son, 1966; Stoch and Sikora, 1976:'Craw et al., 1982), and thencarbon-coated. Samples were observed at 100kV in a little is known about the textures,chemistries, and struc- JEoLJEM-loocx scanning-transmission electron microscope (sreIr'r) tures involved in the alteration process.We therefore have fitted with a solid-statedetector for X-ray energy-dispersive carried out a rEM and aev study at high resolution in analysis.One-dimensional lattice fringe images were obtained in Iijima order to directly characteize the mechanism of alteration rErr,robservations by using only 00/ reflectionsfollowing (1978). of biotite to kaolinite. andBuseck partially The samples used in this study are kaolini- TEM OBSERYATIONS tized biotites from the lower limit of the biotite zone of Kaolinite occurs in two modes. The first consists of the Otago schist at Brighton, approximately 16 km south packetsofkaolinite layersinterstratified with biotite (Fig. of Dunedin, New Zealand. Craw et al. (1982) described ofthe packetsvaries from 50 to 300 A similar samplesfrom the samelocality as containing bio- i). fn. thickness less than the spatial resolution of nEru tite having a curious greenishcolor and low KrO contents and is therefore (=300 aeu analysesofinterstratified packets (4.4-8.5 wto/o).Using xno and electron-microprobedata, analysis A). yield high Al and Si contents characteristic Craw et al. (1982) showed that the "green biotite" ac- of 7-A hyers low K, Fe, and Mg contents due to con- tually consistsofan intimate intergrowth of metamorphic of kaolinite, and tamination by in the analyzedarea (Fig. 2a). This biotite and secondarykaolinite that is not fully resolvable "biotite with the electron-microprobe data of Craw optically or with the electron microprobe. Their data im- is consistent (1982) the possibility that the specific plied that rEM-AEMstudies of such intergrowths might et al. and excludes 7-A hyers imaged by reu are trioctahedral minerals such as serpentine or berthierine. Interfaces between biotite 'Contribution no. 428, Mineralogical Laboratory, Depart- and kaolinite along (001) are parallel, and they display ment of Geological Sciences,University of Michigan, Ann Ar- no strain contrast. Parallel intergrowth featuresofkaolin- bor, Michigan 48109,U.S.A. '?Present address: Department of Geology, Arizona StateUni- ite and wonesite were previously reported in a reu study versity, Tempe, Arizona 85287, U.S.A. of alteredNa-rich biotite by Veblen (1983). 0003404x/87/03044353$02.00 353 354 AHN AND PEACOR: KAOLINITIZATION OF BIOTITE

0.00 1.00 2.00 3.00 4.00 5.00 5.00 7.00 ENEBGY(KEV} Fig. l. Latticefringe image of biotite and interstratifiedpack- etsofkaolinite layers.(001) interfaces between biotite and kaolinite areparallel.

A second mode of kaolinite occunence that provides direct insight into the alteration processeswas also observed (Figs. 3, 4). This mode invariably occurs as pairs of kaolinite layers that are randomly interlayered within biotite. Electron-difraction pattems from such areasex- hibit intense streaking along c* due to random mixed layering (e.g.,Fig. 3). However, severalrepeats of ordered sequencesconsisting oftwo-layer kaolinite units and single biotite layers occur locally (Fig. 3). AEManalyses of biotite containing such pairs oflayers show a significant 0.00 t.00 2.00 3.00 4.00 5.00 6.00 7.00 decreasein K, Fe, and Mg, and increase in Al and Si ENEBGYKEV) content compared to biotite that is free of interlayering, indicating that the interlayered 7-A hyers are kaolinite (Figs.2a, 2b). In addition, twoJayer kaolinite units frequently are ob- servedto terminate at a singlebiotite layer (Fig. 4), where the termination is inferred to extend as a linear feature at some angle to the lattice fringe images. As shown in Figure 4, there is some image contrast associatedwith these terminations, apparently reflecting strain in re- sponseto the replacementof a single biotite layer (10 A) by two kaolinite layers (14 A).

DrscussroN The layer transition boundariesbetween a singlebiotite Iayer and two kaolinite layers representa reaction front 0.00 2.00 lnn 4.00 Enn 6.00 7.00 and imply that one T-O-T unit of biotite must changeto ENEBGYKEV) two T-O units of kaolinite, with the addition of a gibbs- Fig. 2. Energy-dispersiveX-ray spectraof (a) interstratified iteJike octahedral sheet and the reversal in orientation packet ofkaolinite layers, (b) biotite free ofinterlayering ofka- of one tetrahedral sheet of a T-O-T unit. Concomitant olinite layers, and (c) biotite containing interlayered kaolinite chemical changesinclude replacement of the single Fe- layers. Small Mg, K, and Fe peaks in spectrum (a) are due to and Mg-rich trioctahedral sheet by two aluminous di- contamination by biotite in the analyzed area. Low intensities octahedral sheets.Furthermore, all Al in the tetrahedral ofpeaks ofI! Mg, and Fe in spectrum (c) relative to spectrum sheetsmust be replacedby Si, assumingthat the kaolinite (b) indicate that the interlayered 7-A hyers are kaolinite. layers are compositionally ideal. If the biotite has approximately equal fractions of Fe and Mg, as shown by the electron-microprobe data of Craw et al. (1982), the l.5Fe+2 * l.5Mg+'z + 6H+, implying a net gain of HrO, reaction of one biotite layer to two kaolinite layers can Al, and Si and loss of K, Mg, and Fe (if other biotite/ be approximatedby K(Fe, rMg,,)SirAlOr0(OH), + kaolinite ratios are involved in the reaction, specific ele- 3Al+3 + Si*o + 6HrO : 2AlrSirO5(OH)4+ K+ + ments can be eliminated from the reaction). AHN AND PEACOR: KAOLINITIZATION OF BIOTITE 355

tures at the reaction interfaces.The necessaryaddition of H2O, Al, and Si and loss of H, K, Fe, and Mg require a pathway for transport. The linear boundary is inferred to serve as such a conduit especiallyinsofar as there is an apparent absenceofany other structural defects.Similar pathwaysfor transport ofions and fluids were postulated at chain-silicate reaction interfaces (Veblen and Buseck, 1980) and at chlorite-phlogopite interfaces (Yau et al., 1984).Veblen (1985) has discussedthe importance of such extended defects in mineralogical processes.Once initiated at a crystal surface,such a linear reaction boundary would be self-perpetuatingand would causealteration of a single layer of biotite within an otherwiseunaffected sequenceofbiotite layers. Becausebiotite is altered pervasively over several tens of meters at Brighton, such defect-controlledtransport is inferred to be an important mechanism of ion and fluid transport within individual crystalsand to pervasively involve at leasttens of meters Fig. 3. Latticefringe image of biotite and interlayeredka- of rock. olinitelayers (inset diftaction pattern).Kaolinite invariably oc- cursas pairs oflayers. If the reaction mechanism for pairs of kaolinite layers also occurs for interstratified packets ofkaolinite layers, the number of kaolinite layers in the packets should be Such drastic structural and chemical changessuggest multiples of two. However, the number of layers in ka- that the biotite structure may be completely disarticulat- olinite packets cannot be accurately counted becauselarge ed; i.e., kaolinitization of biotite in these samples may portions of lattice fringe imagesof kaolinite are indistinct involve dissolution of biotite layersand crystallization of owing to the rapid beam damage that inevitably occurs. pairs of kaolinite layers. The discontinuity in structure It is therefore not clear if the same reaction mechanism along the linear reaction boundaries, coupled with the is applicable to all kaolinite. However, the observation large diflerence in layer thickness (10 vs. 14 A), implies of increasedvolume of kaolinite pseudomorphsafter bio- the existenceof extendeddefects similar to tunnel struc- tite (Stoch and Sikora, 1976:,Paruchoniak and Srodori,

Fig. 4. Lattice fringe image of kaolinite layers interlayered in biotite. Two-layer units of kaolinite occasionallyterminate at singlebiotite layers (seearrows). 356 AHN AND PEACOR: KAOLINITIZATION OF BIOTITE

1973) is consistentwith this alteration mechanism,in manuscript. We also would like to thank D. R. Veblen and R. J. Reeder that the alteration of one biotite layer (10 A) to two ka- for their reviews.We thank W C. Bigelow and the staffof the Univercity of Michigan Electron Microbeam Analysis Laboratory for providing srrrvr olinite layers (14 A) results in a significant increase in facilities. The analltical scanning-transmissionelectron microscopeused volume. in this study was acquired under NSF Grant DMR-77-09643, and this The occurrence of four repeats of 2:l ordered se- study was supporled by NSF Grant EAR-83-13236 to D. R. Peacor. quencesof kaolinite and biotite layers within a mixed- layered crystal (Fig. 3) is interesting, although the signif- Rrrrn-nNcns icance of such short-range sequencesis questionable according to the statistical test of Veblen and Buseck Bailey, S.W. (1982) Nomenclature for regular interstratifications.Amer- (1979).Although the occurrenceof various orderedmixed- ican Mineralogist, 67, 394-398. Craw, D., Coombs, D.S., and Kawachi, Y. (1982) Interlayered biotite- layer silicatesis well known (seeBailey, 1982),the genesis kaolinite and other altered biotites, and their relevanceto the biotite of such structwes is not well understood.They may orig- isograd in easternOtago, New Zealand. Mineralogical Magazine, 45, inate either through simultaneousprimary crystallization 79-85. or alteration after crystallization. There are many docu- Eggleton,R.A., and Banfield,J.F. (1985) The alteration ofgranitic biotite American Mineralogist, 70, 902-9 lO. mented casesof l: I ordered mixed-layered minerals to chlorite. oth- Iijima, S., and Buseck,P.R. (1978) Experimentalstudy ofdisordered mica er than phyllosilicates (e.g., members of the bastnaesite structuresby high-resolution electron microscopy. Acta Crystalloga- family, Van Landuyt and Amelinckx, 1975), where an phica,A34, 709-719. origin due to primary growth has been suggested.Al- Lee, J.H., and Peacor,D.R. (1985) Ordered 1:1 interstratification ofillite though it may seeminherently unlikely that the alteration and chlorite: A transmissionand analytical electron microscopystudy. Clays and Clay Minerals, 33, 463-467. of one layer could afect initiation of subsequentaltera- MacEwan, D.M.C. (1954) "Cardenite," a trioctahedral montmorillonoid tion in regularly spaced,adjacent layers, this appearsto derived from biotite. Clay Minerals Bulletin, 2,120-126. be the case in this instance, albeit over short repeats. Muenier, A, and Velde, B. (1979) Weathering mineral facies in altered Support for such an interpretation is given by Olives Ba- glanites: The importance of local small-scaleequilibria. Mineralogical flos and Amouric (1984) and Olives Baflos (1985) who Magazine, 43, 261-268. Olives Bafros,J. (1985) Biotites and chlorites as interlayeredbiotite-chlo- observedrepeats of tens of units of ordered sequencesof rite crystals.Bulletin de Min6ralogie, 108, 635-641. layers ofbiotite and chlorite that were interpreted to be Olives Bafros,J., and Amouric, M, (1984) Biotite chloritization by inter- replacing the biotite. Ordered mixed layering of phyllo- layer brucitization as seenby nnrelr. American Mineralogist, 69,869- silicates may therefore apparently be derived either 87l. Parachoniak,W., and Srodoi, J. (1973)The formation ofkaolinite, mont- through primary growth or replacement. morillonite and mixedJayer montmorillonite-illite during the altera- Interlayers between adjacent biotite and kaolinite lay- tion ofcarboniferous tuff(the Upper Silesiancoal basin). Mineralogia ers may be occupied by K in order to provide charge Polonica,4,37-56. balancefor the biotite layer. On the other hand, the pres- Stoch, L., and Sikora, W (1976) Transformation ofmicas in the process granites gneisses.Clays and Clay Minerals, interlayer is of kaolinitization of and enceof such cations not compatible with the 24, 156-162. neutral charge of a kaolinite layer of ideal composition. Van Landuyt, J., and Amelinckx, S. (1975) Multiple beam direct lattice Loss of at least some interlayer K as required by the pres- imaging of new mixed layer compounds of the bastnaesite-synchisite enceof the kaolinite layer may thereforebe accompanied series.American Mineralogist, 60, 351-358. by a charge-balancingprocess in biotite, such as oxida- Veblen, D.R. (1983) Microstructures and mixed layering in intergtown wonesite,chlorite, talc, biotite and kaolinite. American Mineralogist, tion of Fe+2to Fe+3or substitution of Sit" for AIIV. Al- 68, 566-580. ternatively, the interlayer K may be retained if there is a -(1985) Extended defectsand vacancy non-stoichiometry in rock- charge-balancingsubstitution in kaolinite, such as forming minerals.In R.N. Shock,Ed. Point defectsin minerals,p. 122- (Fe*2,Mg+2)substitution for Al*3. However,such details l3l. American GeophysicalUnion, GeophysicalMonographs 31. P.R. (1979) Chain-width order and disorder are beyond the resolution ofthe present study. Veblen, D.R., and Buseck, in biopyriboles. American Mineralogist, 64, 687-7 00. Finally, strain contrast was not observedin rnvr images -(1980) Microstructure and reaction mechanism in biopyriboles. of the (001) interfaces except where layer terminations American Mineralogist, 65, 599-623. occur, although structural misfit betweenthe dioctahedral Veblen, D.R., and Ferry, J.M. (1983) A rrIr.r study of biotite-chlorite re- and trioctahedral layers is expected as pointed out for action and comparison with petrologic observations.American Min- eralogist,68, I 160-l 168. mixed-layerillite-chlorite (ke and Peacor, 1985). The Walker, G F. (1949) The decomposition of biotite in the soils. Mineral- systemsof H bonding at the interfacebetween biotite and ogical Magazine,28, 693-703. kaolinite may therefore be flexible in order to accom- Wilson, M.J. (1966) The weathering of biotite in some Aberdeenshire modate the misfit at the interface. soils. Mineralogical Magazine, 35, 1080-1093. Yau, Y C , Anovitz, L.M., Essene,E.J., and Peacor,D.R. (1984) Phlog- AcrNowr-nrcMENTS opite-chlorite reaction mechanismsand physical conditions during re- trogmde reaction in the Marble Formation, Franklin, New Jersey.Con- We especiallythank D. S. Coombs and D. Craw of the University of tributions to Mineralogy and Petrology,88, 299-308. Otago, New Zealand, for their help in sampling and valuable comments on the manuscript-We are grateful to E. J. Essene,B. H. Wilkinson, and M,c.NUScnrPrRECEIvED Apnn- 2, 1986 B A van der Pluijm for reviewing and improving an early version of the M,c.NuscRrPrACCEPTED Novrrvrsen 7. 1986