American Mineralogist, Volume 67, pages 1195-1205, 1982

Electron microscopy of a muscovitFbiotite interface

Surr,tro Iutuer exu JtNc ZHu2

Center for Solid State ScienceArizonq State University

Tempe, Arizona 85281

Abstract

An intergrowth of mucovite and biotite from Mitchell Creek Mine, Upson County,Georgia, was studied with a high resolution transmission electron microscope. Atomicarrangements of the interface between muscovite and biotite have been derived, revealingfar more complexity for the interface structure than that proposed previously based onoptical and X-ray analysis (Gresens and Stensrud, l97l).

It was found that the muscovite has an almost perfectly ordered 2M1 structure, while thebiotite is highly disordered. Despite such disorder in biotite, both octahedral layers and Kion interlayers of the two minerals are connected perfectly at the interface, while aboutffi% of the tetrahedral sheets are discontinued across the interface.

Occasional occurrences of coherent intergrowths of chlorite-like structure in biotite wererecognized in the vicinity of the interface.

Introduction

Interface structures between two minerals, whichresult from precipitation or exsolution involvingdiffusion of particular atoms in the system, havebeen studied to determine their crystallization his-tories (see, for example, Wenk, 1976; Veblen andBuseck, 1981). Atomic structures or morphology ofsuch boundaries are influenced by thermal historyand evolution of the minerals. The configurationenergy of the interfaces, for which misfit disloca-tions or other types of lattice relaxation may bepresent, tends to be minimized. Therefore, elucida-tion of the atomic structures of interfaces is impor-tant in the study of minerals.

The specimen of intergrown crystals of musco-vite and biotite examined in the present study isfrom the Mitchell Creek Mine, Upson County,Georgia (pegmatite). In this single mica book apyramidal crystal of biotite has intergrown in mus-covite. The same crystal has been studied formerlyby Gresens and Stensrud (1971), and its detaileddescription has been given in their paper and alsoby Lester (1946). Under the optical microscope, the

rPresent address: Research and Development Corporation ofJapan, Department of Physics, Meijo University, Yagoto-Urayama Tepaku-ku, Nagoya, Japan.

2On leave from Central Iron and Steel Research Institute,Beijing, People's Republic of China.

0003-004)o82ll r l2-l 195$02.00

interfaces are not always sharp and straight in the(001) plane projection. The book containing the twomicas can be cleaved easily through the interfacewithout breaking it, suggesting a good continuity ofthe two minerals across the interface. The bookcontains small inclusions of other minerals near thebiotite, such as pyrite and apatite.

From the optical properties and the X-ray diffrac-tion pattern of this intergrowth, Gresens and Stens-rud (1971) proposed that the basic tetrahedralsheets of the micas continue across the interfacebetween the two minerals. They also suspected thata slight difference in the d661 values between the twominerals might be accommodated by edge disloca-tions formed parallel to (001) planes. In the presentstudy, we used the high resolution transmissionelectron microscopy technique to investigate atom-ic structures of the interface between the muscoviteand biotite. The technique provides more informa-tion on the interface structures that has been previ-ously obtained (Gresens and Stensrud,l97l).

Technique

Useful information on the interface between thetwo micas can be obtained by examining its crosssection in a direction parallel to the layers' Edges ofthe biotite intergrown in muscovite are parallel tothe (100) or (110) directions of 2Mr muscovite. Inprojections along those directions of the crystal,

I t96 IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

Fig. l. Low magnification electron micrograph ofan interfacebetween biotite (B) and muscovite (M). The crystal is orientedwith ulOl parallel to the electron beam. Note that a slab ofmuscovite having a width of 3004 penetrates into biotrte.

positions of the alkali ions in the interlayers areseen on end. When going from one layer to the next,these positions can be utilized for describing astacking sequence of the mica layers, as reportedpreviously (Iijima and Buseck, 1978).

The method for the preparation of the specimenby sectioning a mica crystal with a diamond knife inan ultramicrotome (Iijima and Buseck, 1978) is notsuitable for the present purpose and thus the speci-men was prepared by an ordinary ion beam thinningmethod. However, the easy cleavage of mica crys-tals and the necessity of pin-pointing an interface inthe crystal required very careful specimen prepara-tion.

A thin specimen was prepared by argon ionmilling, in which great care was taken not to com-pletely erode the interface region. The specimenwas lightly carbon coated. A rnor 200CX transmis-sion electron microscope operated at 200 kV and

equipped with a top-entry goniometer stage wasused. The tilting specimen cartridge accepting astandard specimen of 3 mm diameter was built inour laboratory.

Results and interpretation

Morphology of the interface

Lester (1946) noted that in the thick book of micaexamined presently the biotite is intergrown inmuscovite as pyramidal shapes. Gresens and Stens-rud (1971), however, found that the contact of thetwo minerals has occasional irregularities, andsmall, irregular patches of biotite were recognizedin the muscovite. They also speculated that similarsmall patches of muscovite might occur in biotite.These observations suggest inhomogeneous distri-bution of (Fe, Mg) ions and Al ions near theinterface.

A low magnification electron micrograph (Fig. 1)of an interface from our sample provides a view ofthe overall microscopic morphology of the interfaceand immediately answers some of the questionsraised from the light microscope examination.

A curving line running approximately from thetop right to the bottom left partitions two regions,the dark region which is biotite and the lighter onewhich is muscovite. This was confirmed by examin-ing peak heights of Fe in the X-ray energy disper-sive spectra from these two regions. The measure-ment was conducted with a Philips 400T micro-scope equipped with a Tracor Northern TN2000EDS spectrometer. The darker contrast in biotitethan in muscovite is attributed to the Fe ions, sincethey have a higher scattering cross section for theelectrons than the Al ions in muscovite. The crystalwas oriented so that the tTT0l or [tT0] direction ofthe 2Mr of muscovite was parallel to the direction ofthe incident electron beam. An electron diffractionpattern from the muscovite is shown in Figure 2a.

Vertical white lines in Figure I are images of gapsin the crystal due to the cleavage introduced duringthe sample preparation. It is noted that a narrowband of muscovite about 3004 wide penetrates intoa region of biotite (marked M in the middle of thephotograph). We could not confirm whether or notthe band is terminated in biotite. As we describelater, two types of interface between muscovite andbiotite can be charactenzed, one parallel to themica layers and one perpendicular to the layers.The boundary inclined to the layers can be consid-ered as a mixture of these two types. The penetrat-

IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

MUSCOV I TE B I O T I T EFig. 2. A pair of electron diffraction patterns, (a) from a region of muscovite and (b) from a region of biotite, which lie next to each

other at the interface. The muscovite shows sharp spots of the (3k k l) reciprocal lattice section, while the biotite shows ditruse

streaking at the spots with k I 3n.

ing slab of muscovite has an interface perfectlyparallel to the mica layers as seen in Figure 1.

Figure 3 shows a low magnification image ofanother portion of the interface. The dark portion isbiotite, an electron diffraction pattern (Fig. 2b) fromwhich displays diffuse streaks at the reciprocallattice points with ft I 3n. The origin of the diffusestreaks has been explained by the occurrence ofrandom 120n'layer rotations. The disordered stack-ing of the layers can be recognized in the image ofbiotite (Fig. 3), while the muscovite is almost per-fectly ordered as demonstrated by the sharp diffrac-tion spots in Figure 2a.

Atomic structure of the perpendicular interface

The dark contrast appearing along the interfacecould be caused by dislocations formed by discon-tinuities of the mica layers at the interface. A highresolution two-dimensional lattice image of the in-terface corresponding to the small area enclosed inFigure 3 is shown in Figure 4a. The crystal was notsufficiently thin and its orientation was not alignedprecisely enough at the zone axis orientation, sothat the image intensity of this image can not beinterpreted directly in terms of the projection of thestructure.

However, it was demonstrated in our previous

study on disordered micas that two-dimensionallattice images viewed along the (100) or (110) direc-tions are adequate to derive a stacking sequence ofthe layers (Iijima and Buseck, 1978)' It was experi-mentally shown that in any one image the sameportion of every unit cell of a perfect crystal alwaysproduces an identical image intensity, which isindependent of the crystal diffraction parametersand the instrumental parameters. This is due to thefact that the spread of the electron wave travelingthrough the crystal is confined within a small area ofseveral flngstrcims depending on the crystal thick-ness. The usefulness of such high resolution latticeimages was demonstrated recently by the structureanalysis of the sheet silicate mineral mcGillite byone of the present authors (Iiiima, l982a,b).

Let us consider the image details of the region ofmuscovite marked M in Figure 4a' The image showsthree different intensity distributions for individualslabs having a 10A width, which matches the spac-ing of the basic mica layer. They correspond to thethree different projections of the basic mica layeralong the [100], [110], and [lTO] directions. Theposition symbols, A, B, and C, designated byZvyagin (1962), can be used to describe the stackingsequence of the layers. We assume that in thepresent mica crystal there is no stacking disorder

t 198 IIJIMA AND ZHU: MUSCOVITEBIOTITE INTERFACE

Fig. 3. A low magnification electron micrograph showing an interface between biotite (B) and muscovite (M). The dark contrastappearing along the boundary may be caused by lattice strain. The regions enclosed by the small and large rectangles are shown inFigs. 4 and 8 respectively.

due to 60o, 180o, or 300" rotations of the layers,which are uncommon in muscovite and biotite(Ross et al., 1966).

Using these symbols we obtained the stackingsequence shown at the top of Figure 4a. Similarly,the stacking sequence labelled in Figure 4a is illus-trated in Figure 5a by using the vector symbols ofSmith and Yoder (1956). These vectors define theslip direction from the bottom tetrahedral sheet tothe top one of the T-O-T layer projected on the(001) plane. The sequential numbers shown on thevectors correspond to layer numbers between I and36.

These illustrations demonstrate that the crystalhas the two-layer periodicity -BC- of the 2M1structure, viewed down the [110] direction. Thisagrees with the electron diffraction pattern of Fig-ure 2a. However, there are occasional stackingfaults, as indicated by -CA- at the layer number 4-5, -AC- at the layer number 15-16, and -CAC- atthe numbers 22-24.

Similarly, a stacking sequence in the region ofbiotite marked B in Figure 4a was analyzed. Theresult is shown at the bottom of the image, and thevector representation of the sequence is given in

Figure 5b. As expected from the electron diffractionpattern (Fig. 2b), the stacking sequence is almostrandom, except for the region corresponding to thenumbers 26-32. This region has the same -BC-stacking sequence as the muscovite, and the layersappear to be continuous across the interface. Disor-der of the stacking can be seen best by viewing themicrograph at a grazing angle in the direction per-pendicular to the layers.

Lattice fringes wittr tOA spacing corresponding toa single layer appear to be continuous across theinterface. This suggests that individual T-O-T micalayers and the K ion interlayers are registeredperfectly in their z axis coordinates on both sides ofthe interface. This may be the reason why the micabook cleaves smoothly through the interface. How-ever, it should be emphasized that the periodicstacking of the layers in muscovite becomes abrupt-ly disordered in crossing the interface into biotite.The spatial transformation from biotite to musco-vite takes place within a range of several unit cellsat the interface.

Figure 4b represents a dark field electron micro-graph of the same area of biotite as shown in Figure4a. This was imaged using only the (000 reflections

IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

cBc

Fig. 4. (a) A high resolution lattice image of the region enclosed by the small rectangle in Fig. 3. The capital letters indicate thestacking layer sequences represented by Zvyagin's position symbols obtained by the analysis of the image intensity. The muscovitehas a nearly ordered stacking sequence -BC- corresponding to the 2Mr polymorph, while the biotite is almost completely disordered.Continuities ofthe individual layers from layer numbers I to 36 have been analyzed in the text. (b) A dark field electron micrograph ofthe same region as for the bright field image of(a), showing variations in contrast which are consistent with the stacking sequenceobtained from (a).

32 -36i t

CBCBCBAC BCBCACBCBCBCECBCB C BCBC BC BBC

with / : 1,2,3. It is seen that the fringe imagesconsist of lines having three different contrast lev-els; namely, the darkest, dark and grey lines. If thedarkest line appears at the layer position A, thepositions of the darkest lines match well with thelavers A which were determined from Fisure 4a.

Similarly, the correspondence of the dark and greylines to the layers B and C respectively is excellent.The dark field image therefore supports our inter-pretation of the stacking sequence in biotite derivedfrom the bright field image. The region of muscovitehas been vitrified entirely due to electron beam

1200 IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

(b)Fig. 5' Vector representations of the stacking layer sequences; (a) muscovite of Figure 4a and (b) biotite. The numbers on the

vectors are the layer numbers. Note that the layers from numbers 26to32 are perfectly continuous across the interface.

,ta,tO

damage, while the damage to the biotite is muchless. A further point concerning this observationwill be discussed later.

Now we consider atomic arrangements of theinterface from the experimental results mentionedabove. The Zvyagin position symbols define therelative positions of neighboring layers and areinconvenient for comparing two layers at the samelevel, as for discontinuous tetrahedral sheets. Forinstance, the C layer of the layer number 4 inmuscovite in Figure 4a does not necessarily connectperfectly to the C layer of the layer number 4 of

biotite, and the type of connection will be depen-dent on their adjacent layers. For this reason, thevector representation symbols were used for analy-sis of the atomic structure of the interface.

As described above, the individual layers of micabetween the layer numbers 26 and 32 appear to beperfectly continuous across the interface. We takeone of these layers, namely number 32, as a refer-ence to the others. In other words, the two stackingvectors of number 32 shown in Figure 5a and 5b aresuperimposed. By doing so, we can find the posi-tions of the individual T-O-T lavers of the musco-

O

IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE t20l

(o ) ( b )

Fig. 6. Two possible idealized models for discontinuedtetrahedral sheets at the interface (arrowed) between muscovite(M) and biotite (B), which were obtained by the analysis of thestacking vectors in Fig. 5. Dark and hatched triangles representSi-O tetrahedra pointing down and octahedra in the T-O-T layerrespectively. The arrangement shown in (a) seems to be favoredover the one in (b).

vite relative to those of the biotite. Thus, we canexamine the continuity of the individual T-O-Tlayers.

The analysis of the vectors showed that there areonly two possible atom arrangements for the dis-continuous layers, assuming a perfectly sharp inter-face. They are schematically illustrated in Figures6a and 6b, where solid triangles represent the Si-Otetrahedral layers pointing down and hatched trian-gles the Al-O or Fe-O octahedral layers. In musco-vite the octahedral layers contain Al ion vacancies.M and B indicate muscovite and biotite. We foundthat among the 49 T-O-T layers labelled in Figure3a.37% of them were continuous. Another 37Vo oflayers are discontinuous and have the structure ofFigure 5a. The rest of the layers are also discontinu-ous, with the structure of Figure 5b.

Because the actual T-O-T layers contain twotetrahedral sheets, each of which has one of thethree structures described above, there are sixpossible structures for a T-GT layer. The mostcommon combination among the layers labelled inFigure 4a was such that one tetrahedral sheet iscontinuous and the other is that of Figure 6a. Theleast common one is such that both top and bottomtetrahedral sheets of a T-O-T layer have the struc-ture of Figure 6b. This suggests that the arrange-ment of atoms at the interface shown in Figure 6a isfavored over that of Figure 6b. The number oflayers examined here is, of course, too small todraw a general conclusion on the energetically mostfavorable atom configuration at the interface.

The actual tetrahedra near the interface may bereorganized to minimize the configuration energy,presumably by the incorporation of some foreignatoms or vacancies in octahedral layers. The con-tinuous 10A lattice fringe images across the inter-face suggest that the displacement of the T-O-Tlayers by one-halfofthe c axis perpendicular to thelayers, which occurs in the double or triple chainsilicate minerals (Veblen and Burnham, 1978), maynot be feasible for interfaces between muscoviteand biotite.

Reorganization of the atoms near the gap regionsmentioned above would not be accompanied bylong-range strain fields as for misfit dislocations,since individual octahedral layers and K ion sheetsare in registry in both micas. This may explain whywe did not observe image contrast characteristic ofdislocations at the interface.

Interfoces parallel to the mica layers

In general, the lattice parameters a and b ofbiotite are about 3Vo larger than those of muscovite.This will result in misfit on the interface when thetwo minerals meet epitaxially on the (001) plane.The difference in their alattice parameters might beaccommodated by the introduction of edge disloca-tions at an interval of every 30 unit cells, or 1504.Occurrences of such dislocations are not obvious inthe regions of the (001) interfaces, as shown inFigure L It may be that the relaxation of the strainfield near the interface results in fewer misfit dislo-cations than expected from a simple analysis of theunit cell dimensions.

A parallel interface is represented at a highermagnification in Figure 7. The boundary betweenbitotite (B) and muscovite (M) is indicated byalrows. In this instance, the interface is restrictedto a single (001) plane. The crystal has been alignedat the zone axis setting of biotite. The lattice fringeimages of the basic 10A layer have not been image4in most parts of the muscovite, but fringes of 20Aappear. This is a typical phenomenon for latticeimages formed when a crystal of muscovite is in aslight off-zone axis setting. In other words, thecrystal orientations of the two minerals are notexactly coincident. It is not clear whether such aslight misorientation between the two micas resultsfrom the difference between their lattice parame-ters.

Splits of the fringe images indicated by arrowspointing down are a secondary effect that takesplace under the electron beam during observation.

MisMls

1202 IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

t3101

IFig. 7. A high resolution lattice image sho^wing a region of the muscovite-biotite interface (arrowed) which is parallel to the basal

(001) plane. Two slabs having a spacing 144 are indicated and could be explained as intergrowths of single layers of chlorite-typestructure. Such intergrowth occurs frequently near the parallel interface. The convex lens shapes are the initial stages of splitting ofthe mica layers, which results from mechanical stress during the specimen preparation.

Such specimen damage will be discussed in thesection, "Damage due to electron beam irradia-tion."

I ntergrow t h of chlorit e - ty p e lay ers

In the biotite region shown in Figure 7 there aretwo vertical lines having lighter contrast than thosein the rest of the area. The widths of these narrowbands were measured to be about t+A (the distancebetween the two adjacent dark lines). They occuralways as isolated single bands. It is likely, there-

fore, that the K ion interlayers are replaced withslabs wider than that of the K ion interlayer of mica.

A possible layer would be a single layer of bruciteMg(OH)2, which would satisfy the l4A spacing andproduce a single unit cell of chlorite. Since theprimary difference between mica and chlorite is inthe interlayer, these two minerals can be intergrowncoherently. In fact, such a coherent intergrowth hasbeen foqnd in phyllosilicate by Page and Wenk(1979) and also Veblen and Buseck (1981).

The chlorite-type intergrowth layers occur al-

IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

Fig. E. A dark field image of a region corresponding to the area enclosed by the large rectangle of Fig. 3, showing the efect ofelectron beam irradiation. Some parts of the muscovite have been vitrified due to the beam damage. The dark regions betweenmuscovite and biotite where the vitrification has been completed indicate that the interface was a nucleation site for the vitrification.

t203

ways as isolated single slabs and are observed onlyin the vicinity of the biotite-muscovite interface.Two possible reasons for the chlorite-type inter-growths in biotite are as follows: (1) The inter-growth is associated with the lattice strain at theinterface, which has been accommodated by theoccasional intergrowth of chlorite-type layers, sincethey have smaller a and b lattice dimensions thanthose of muscovite. In going away from the inter-face. the misfit at the interface can be relaxed morerapidly by having chlorite-type structure. Thiswould explain why there are not chlorite-type lay-ers in regions very far from the interface. (2) Theiroccurrence could be associated with a compositionvariation of the biotite or could result from diffusionprocesses in the history of the mica book. Diffusionof the Mg ions, combined with deficiency of the Kions, would be of importance in producing thelayers.

In place of edge dislocations, the occurrence ofwhich has been proposed by Gresens and Stensrud(1971) in order to accommodate the difference in thec lattice parameters between muscovite and biotite,the chlorite-type layers could absorb the interfacemisfit. It is noted that the c axis of biotite is largerthan that of muscovite, so that the intergrowth ofchlorite in biotite would have the same effect asedge dislocations in muscovite.

Damage due to electron beam irradiation

In our ample experience with examining varioussilicate minerals under the electron microscope, ithas generally been true that resistance of the mate-rials to electron beam irradiation depends onamount of OH ions present. For instance, chlorite isless stable than micas. Similarly, clay mineralscontaining much more water are rapidly damagedby the beam.

tzm IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE

It was found in the present study that biotite ismuch more stable in the electron beam than musco-vite. This is demonstrated in Figure 8, which corre-sponds to a region enclosed by the large rectangle inFigure 3. The micrograph is a dark field image takenunder the same conditions as that of Figure 4b. Tworegions marked M are muscovite, but the rest of theoriginal crystal of muscovite has been vitrified. It isnoted that the vitrification of muscovite nucleatesalong the intertace. This can be recognized inFigure 8 by the enhanced vitrification betweenremaining muscovite and biotite. Nucleation at theinterface is quite reasonable because the damagecould start at gap regions at the boundaries.

The reason for the instability of muscovite maylie in the chemistry of the octahedral layers. Biotiteis richer in Fe ions than muscovite and is commonlyricher in F and poorer in OH ions; therefore it ismore resistant. It is also possible to explain differ-ences of behaviour of biotite and muscovite underthe beam due to slightly higher ionization energy forFe ions than Al ions.

A feature with a convex lens shape similar to anintrinsic stacking fault is circled in Figure 7. This isan image showing the initial stage of splitting of thecrystal (indicated by larger arrows), which is acommon phenomenon due to mechanical stressduring specimen preparation. We can not tell con-vincingly from this image where the split starts.However, the split may have occurred at the K ioninterlayers because the T-O-T sheets are stronglybonded. If we assume so, then the dark lines in theregion of biotite would correspond to the T-O-Tlayers. This allows us to tell that the intergrowthlayer of 14A width results from replacing the K ionlayer by the Mg(OH)z layer as mentioned in thepreceding section.

Conclusion

Atomic arrangements of the interface of coexist-ing muscovite and biotite in a mica book fromMitchell Creek Mine, Georgia, have been inferredusing the high resolution transmission electron mi-croscope.

The crystallographic morphologies of the inter-face obtained by light microscope observation aremerely averaged structures (Gresens and Stensrud,1971). The electron microscope shows that theinterface is more complex. In the extreme case, anarrow band of muscovite 300A wide penetratesinto a biotite region.

The muscovite of the mica book examined in the

present study has a well ordered 2M1 polytypestructure, while the biotite is considerably disor-dered. The individual octahedral sheets were foundto be continuously connected from biotite to mus-covite at their interface. Discontinuity, however,occurs in the tetrahedral sheets. There are only twopossible atomic arrangements for the discontinuity,if the boundary is sharp, and one of them seems tobe energetically more favorable than the other.

Single layers of chlorite-like structure are com-monly intergrown in biotite near the (001) inter-faces. This intergrowth might promote relaxation ofthe lattice strain produced by the contrast of thetwo micas or might result from difusion processesinvolving the Mg ions in the present rnineral sys-tem.

Biotite is more resistant to electron beam irradia-tion than muscovite, which may be attributed to itsbeing poorer in OH ions. The electron beam dam-age of the specimen initiates at the interface. Thismay be caused by dislocations at the interface.

AcknowledgmentsWe would like to thank Dr. N. Guven of Texas Tech.

University for providing the mica specimen and Dr. D. Veblenfor improving the manuscript. We thank Mr. H: Kolar for hishelp in the X-ray micro-analysis. Financial support from NSFGrant DMR-8015785 is gratefully acknowledged. This researchwas also supported by the Nsr Hnrru facility at ASU (GrantcHE-7916098).

References

Gresens, R. L. and Stensrud, H. L. (1971) Chemical, optical,and x-ray analysis of an unusual muscovite-biotite inter-growth. Lithos, 4, 63-69.

Iijima, S. and Buseck, P. R. (1978) Experimental study ofdisordered mica structures by high-resolution electron micros-copy. Acta Crystallographica, A34, 7WJl9.

I{iima, S. (1982a) High-resolution electron microscopy of mcGil-lite: I. Structure determination. Acta Crystallographica, 38A,685-694.

Iijima, S. (1982b) High-resolution electron microscopy of mcGil-lite: II. Polytypism and disorder. Acta Crystallographica, 38A,694J02.

Lester, J. G. (1946) Inclusion in muscovite from Mitchell CreekMine, Upson County, Georgia. American Mineralogist, 31,77-81.

Page, R. and Wenk, H. R. (1979) Phyllosilicate alteration ofplagioclase studied by transmission electron microscopy. Ge-ology,7, 393-397.

Ross, M., Takeda, H. and Wones, D. R. (1966) Mica polytypes:systematic description and identification. Science, 15l, l9l-193.

Smith, J. V. and Yoder, H. S. (1956) Experimental and theoreti-cal studies of the mica polymorphs. Mineralogical Magazine,3t,209-215.

Veblen, D. R. and Buseck, P. R. (1981) Hydrous pyriboles and

IIJIMA AND ZHU: MUSCOVITE-BIOTITE INTERFACE I2O5

sheet silicates in pyroxenes and. uralites: Intergrowth micro- phibole-micareaction, American Mineralogibt,63, 1053-1073.structures and reaction meehanisms. American Mineralogist, Zvyagln, B. B, (1962) The theory of mica polyrnorphism. Soviet66, 1107-1134. Physics Crystallography,6,571-580.

Veblen, D. R. and Burnham, C. W. (197E) New biopyribolesfrom Chester, Verniontl II. The crystal chemistry of jim- Manuscript reeeived, January 5, 19E2;thompsonite, clinojimthompsonite, and chesterite and theam- acceptedfbr publication, June 21, 1982,