American Mineralogist, Volume 58, pages 889-900, 1973
The CrystalStructures of One-LayerPhlogopite and Annite
RosEnr M. HezEN, lNo CnnnrEs W. BunNnlvr
Department ol Geological Sciences,HarDqrd Uniuersity, C ambridge, M ass achusen s 0 2 I 3 8
Abstract
The crystal structures of natural fluorophlogopite and annite have been refined by least- squares analysis of three-dimensional X-ray diffraction data to weighted R-values of 4.1 percent and 4.4 percent respectively. Both of these lM trioctahedral micas possess monoclinic spacegroup symmetry C2/m. Averagebond lengths for phlogopite are (Al,Si)-O 1.649,Mg-(O,F,OH) 2.064,innerK-O 2.969 and outer K-O 3.312A; while those of annite are (Al,Si)-O 1.659,(Fe,Mg,Ti)-(O,OH) 2.108,inner K-O 3.144 and outer K-O 3.216A. Atomic fractional coordinates closely conform to the "ideal" mica structure as defined by Pabst (1955). The only significant deviations from this ideal model are tetrahedral '7.5" rotation (o) and octahedral flattening (p). For phlogopite a : and 'y' - 58.95', while for annite d: 1.5' and,l, - 58.40'. These values agree with distortionspredicted by Hazen and Wones (1972). Anisotropies in apparent atomic vibrations may be bbrrelated with positional disorder of aluminum and silicon in the single tetrahedral site. Magnitudes of these apparent vibrations are greater in annite, reflecting the more extensive cation substitutional disorder in that specimen. No ordering was detected in the octahedral sheet of annite. At least 6 percent octahedral vacancies are indicated from both least-squaresrefinement and chemical analysis. These vacancies are evenly distributed between Ml and, M2, and thus do not represent solid solution between trioctahedral micas and the dioctahedral variety in which only the M2 sites are occupied.
Introduction and illustrated by several authors (e.g., Bragg et al a The first researcher to describe an X-ray diffraction 1965, p. 254 et seq). ln the ideal mica structure, fit between examination of mica was Mauguit (1927,1928), who sheetofregular octahedrais constrained to rings of determined the symmetry, unit cell dimensions, and two sheets of interconnected hexagonal oxygens cell content of several natural specimens. Shortly tetrahedra. Furthermore, basal Ol and 02 oxygen and thereafter the basic crystal structure of mica was are coplanar in (001), as are apical 03 the proposed almost simultaneouslyby Pauling (1930)and hydroxyl groups. The above conditions imply that site Jackson and West (1930, 1933). While several sub- interlayer cation is in the twelve-coordinated A sequent studies established the existence and geometry formed by a hexagonal prism of basal oxygens. geometrical con- of complex twinning and stacking arrangements in further consequence of ideal mica these layer silicates (e.9., Hendricks and Jefferson, straints is the orthohexagonal unit-cell relationship, y'\a. 1939; Smith and Yoder 1956; Ross, Takeda, and b: micas is most Wones, 1966), it was not until 1959 that a three- The structure of real trioctahedral from dimensional structure analysis of a trioctahedral mica conveniently discussed in terms of deviations the was carried out by Takeuchi and Sadanaga(1959). A the ideal model. Such deviations result from list of several recent studies on the oneJayer triocta- differing sizes of ideal octahedral and tetrahedral it hedral mica crystal structure is given in Table l sheets,which share the 03 oxygens. For example, Pabst (1955) presented theoretical atomic co- is well known that the ideal AlSi, tetrahedral sheet ordinates for an "ideal" C2f m, one-layer, trioctahedral of phlogopite is larger than the regular Mg3 octa- mica; and this model has been subsequently described hedral sheet. Thus, these two sheets must coincide 889 890 R. M. HAZEN, AND C, W, BURNHAM
Tesrr. 1. Previous Studies of One-Layer Trioctahedral difficulty of obtaining undeformed cleavage frag- Mica Crvstal Structures ments of layer silicates, the specimen was carefully nenen! Authors (Date) Speclnen examined for single crystals of a size appropriate Condltlons/Results for X-ray diftraction studies. Eventually, a suitable Takeuchl & ^4r | ! | rvPrrJ f rr u c T.^fr^ht. ?n fllm Sadanaga Naturaf R=131 euhedral hexagonal plate 60p thick, and 400p in (r959,1966) maximum width, was selected. The structural for- Zvyagln & PhLogopl te 2D Electronography Franklin phlogopite, calculated from a wet lillshchenko Nat ura I R . I7t mula of (1962) chemical analysis (seeTable 2), is (Ko 77Na616Bae.,65) Stelnflnk Fe rr1- Ph logop 1 t e Yc^tF^ht^ ?n ftln Mg3 which closely (1952) Natural R - 13i n,,Al1n5Si2 e5Olo,iro(F1 311OH6 6), approximates the ideal phlogopite formula, KMgr Donnay et. -aI. Ferrl-Annlte lc^tn^^l ^ ?n f{ lh ( 19'64tt Synthetlc R-2lfi AlSiBOlo(F,OH)r. Optical properties of the crystal Franzlnl -et. -al. 51x Blotltes z atomlc fractlonal selectedate a: 1.530,F : y = 1.558and 2V = 0". (1953) NatUral coordlnates fron 001 data The indices of refraction are slightly higher than
Takeda & Llthlum FIuor-M1ca lhl c^tF^hl ^ those of pure synthetic fluor-phlogopite (y = 1.550, Donnay Synthetl c 3D counter ( 1966) R = 4.81 Hatch et al, 7957), and considerably lower than the
McCauley(1968) Fluor-Phlogoplte lhl <^fF^h{. indices of synthetic hydroxy-phlogopite (y = 1.587, and Synthetlc 3D counter McCauleyet. and R = 6.1.1 Hazen and Wones, 1972). a1. (r973r Ba- Ll -Fluor- Ph logop 1 te Synthetlc Sp ecimen D escri ption- A nnite taJ hLl ; af -.1 Iron Blotlte Tc^fF^n{. in fl th ( 1959)- Natural R.14i Specimensof an iron-rich, igneous biotite from the Takeda & F luor- Poly I1 thlonl t e Isotroplc, generously Burnhan Synthetlc 3D counter Pikes Peak Granite, Colorado, were ( 1959) R = 5.1' provided by Dr. F. Barker of the United States Geological Survey. After opticat and X-ray examina- tion of more than thirty separated annite grains, an either by expansion of the octahedral sheet or by undeformed, approximately rectangular plate, 250p.X contraction of the tetrahedral sheet in the (001) 400p X l00p thick, was found and selectedfor further plane. Radoslovich and Norrish (1962) suggestfour study. Wet chemical analysisof Pikes Peak annite was mechanisms to accomplish this: 1) bond length provided by Barker (Table 2). The resultant structural alteration within the ideal sheets,2) tetrahedral sheet formula, calculated on the assumption that (O * tilting or corrugation, 3) octahedral sheet flattening, oH+F+Cl):12,is and 4) tetrahedral sheet rotation. Octahedral flatten- ing and tetrahedral rotation are believed (Ko orCaoo.)(Fell, Fe3]rnTio ,rMgo ,rMno ou to be the "rNao most important distortions in trioctahedral micas Alo onno,,)(Al, ,osi,,,XOro 3sOHl'sFO.2rClo ou). (Donnay, Donnay, and Takeda, 1964a). A gamma-ray resonant absorption (Mossbauer) The present study is an attempt to determine the u'Fe spectrum of in Pikes Peak annite was measured extent of deviations from Pabst's ideal mica struc- by Professor Roger Burns of the Massachusetts ture for two natural specimensand to compare these Institute of Technology, and confirmed the ratio deviations with values predicted by Hazen and Fe'*:Fe'* : 10:l in octahedralcoordination. Due Wones (1972). In addition. careful examinations of to the possibility of local variations in biotite com- anisotropic atomic thermal vibration ellipsoids and position, the singlecrystal under investigationwas also octahedral occupanciesand cation distributions have analyzed by electron microprobe (Table 2). No zoning been made to aid our understandine of trioctahedral was observed, and the composition determined by mica crystal chemistry. microprobe analysis is in close agreement with wet chemical data. Optical properties of Pikes Peak annite Experimenfal areot: 1.624,P: "y : l.672and 2Z: 0'. These Spe cimen D ercri ption-P hlogopit e values are similar to those obtained by Wones (1963) for a biotite with Fel(Fe + MC) : 0.75. Specimens of colorless euhedral phlogopite, em- bedded in calcite from the marbles of Franklin. New X-ray Diffraction Jersey (Harvard Museum #99276), were kindly Unit cell parameters of phlogopite and annite supplied by Professor Clifford Frondel. Due to the were determined using a precision back-reflection ONE.LAYER PHLOGOPITE AND ANNITE 891
'Il^B.rB Weissenbergcamera and measurementsof both Cu 2. Chemical Analvses Ka1 and Ka2 reflections. These data were refined using a least-squares program LCLse (Burnham, oxlde Franklln PlkcE P.ak Annltc Phlogopl te1 l{et Analysls2 Mlcroprobe3 1962) that corrects for systematic errors due to film ut-l lr* ta uj---L ltea-tq u!-L Alea_La shrinkage, specimen absorption, and camera ec- 919r- \2.0 2.95o 33.96 2.81 33.92 2.84 centricity. Results of these measurements are listed lAr20r r2,7 1.050 r3.to 1.19 rr.6z t.t6 r (rv) in Table 3. Careful examination of a- and b-axis 4 .000 4.00 lr.oo Olevel Weissenberg photographs revealed no twin- tAl2O,(VI) O.OOo 0,09 0.0r lFe20! 0.00 0.000 3.06 t ning or complex stacking arrangements. Precession Feo 9.25 0.010 32.0/l Z.iZ 35.\2 2'.'rg M8O 28.6 2.980 0.97 0.12 0.89 0.11 photographs confirmed the diffraction symbol MnO 0.08 0.000 0,65 0.05 0.611 0.05 as N10 0.02 0.000 0.00 0.00 2/mC-/-, resulting in C2, Cm and C2/m as possible zno 0.01 0.000 0.00 0 ,00 T1O2 0'2? 0.008 3.55 0,22 0.23 spacegroups 3.61 . C2 / m was selectedas the trioctahedral r (vr) 2.998 2,89 2,93 mica space group, based on the discussion of Pabst |xao 8. !5 0.759 8.117 0.88 8.9! 0.95 (1955) and the experience of previous investigators lNaro 1.19 0,160 0.119 0.07 0.10 0, 02 IrLlrO 0.01 0.000 0.00 0,00 (seeTable 1). lRb20 0.0r 0.000 0.00 0.00 Bao I.70 0.046 0. 00 0 ,00 Intensity measurements were made with a four- cao 0.04 0.001 0.q2 0.04 0,03 0.00 circle computer-controlled Picker diffractometer, [ (ark) 0,957 1.02 o.97 using proceduresdescribed by Burnham et al (1971). LH2O 1.5r 0.70 2,5O 1.38 1.40 F 5.85 1.30 0.82 0.22 ).3.35 o.22 A1l non-equivalent reflections with h * k : 2n in cl 0,00 0.00 0.31 0.05 0.05 o - - 10.00 - - 10.35 10.33 one hemisphere of reciprocal space within the range [ (antons) 12.00 12.00 12.00 0. 1 to 1.0 of sin d/I were measured;totals were 2244 r (h'r. t) 100.32 100.07 99.77 for phlogopite and 2239 for annite. Integrated in- tensities were corrected for Lorcntz and polarization 1. Wet chenlcal. ana]ysls by Jun lto, Harvard Unlver- slty, Canbrldge, Massachusetts, effects, and absorption corrections were computed by 2. Wet chenlcaL analysls by V, C. Snlth, U. S. Gcologl- cal Survey. numerical integration (Burnham, 1966). Transmis- 3. Mlcroprobe analysls by Steven Robert8, Harvard Unlverslty. sion factors varied from 0.66 to 0.92 for phlogopite, l. Calculated on basls of (O + F + Cl + OH) = 12, and assunlng sun and from 0.25 toO.62for annite. of tetrahedral catlons cquaLs four.
Str uctur e Refine ment-P hlo gopi t e ference map was computed, on which no significant All structure refinements were carried out using features were found. the full-matrix least-squares refinement program RFINE (Finger, 1969), with a weighting scheme Struc t ur e Refineme rut- A nni t e based on counting statistics, atomic scattering fac- phlogopite tors, and rejection of unobserved reflections as Annite refinement was initialized using described by Burnham et el (1971). Phlogopite atomic coordinates and isotropic temperature factors refinement was initialized using the positional pa- as determined above. Octahedral Ml and M2 sites by Fe'*, rameters and isotropic temperature factors of fluoro- were initially assumed to be fully occupied position as 100 percent phlogopite (McCauley, 1968), and, after four cycles while the fnterlayer was refined - potassium. After five cycles (R : 12.4%) anisotropic of refinement (R 8.3Vo), convergence was ob- temperature factors and real and imaginary anomalous tained. A three-dimensional Fourier difference syn- dispersion terms (International X-ray thesis was generated using a Fourier summation Tables for Yol. p. 215) were included. Con- program by L. Guggenberger to detect principal Crystallography, 3, vergence was achieved in an additional five cycles sources of model error. Analysis of the Fourier map (R : 9.3%), and a three-dimensional difference map indicated that major errors were due to the presence was generated (as described above) to determine of strongly anisotropic apparent thermal vibrations, possible sourcesof error. Major positive anomalies in and thus subsequent cycles of refinement included pobe- p"n1"w€re detected at Ml and M2 octahedral anisotropic vibration parameters. After six addi- sites, indicating that the 100 perpent Fe'* model tional cycles (R : 6.6% ), observations for which AF > 5.0 were rejected. 34 of 1725 observed data 'Tabulated observedand calculatedstructure factors for were rejected on this basis after convergence on the phlogopiteand annite are available from the authors on twelfth cycle (R = 4.lVo ).1 A second Fourier dif- request. 892 R. M. HAZEN, AND C. W. BURNHAM
Tlsr-e 3. Unit-Cell Parametersof Micas Studied tributions for the above elements. Since scattering power is proportional to the number of electrons, Z, Paralnete! Phlogopite Annite a mean Z for the octahedral sites was computed by . (Al s .3078(4 ) s.3860(9) the relation: Z -- ZnZo'.xn,where Zn is the number of u tAl 9.190I(s) 9 .324L(7 ) electrons in the fth cation, and xo is the fraction ofthat (A) (8) (9) c I0. I5{7 r0.2683 cation in the octahedral layer. The shape of scattering (o) r00.08(I) 100.63(r) B curves is affected by the ionic charge, and thus a mean vor. (it ) 487-7 12' s06.8(2) cation charge e : 2oC,'xi was similarly 0 octahedral (A) q 1q V3 a 9.32 calculated.Resultantvalues of Q : 23.2,e : +2.25) * obs. 68 closely approximate those of divalent manganese * Parenthesized figures represent the eatinated (Z : 23, C : +2). To test the validity of modelling standard deviation (esd) in terms of least units cited for the Vefue to the imtnediate the Pikes Peak annite octahedral sites as Mn'*, the left, thus 5.3078(4) indicates an ead of from 0.0 to 0.0004. scattering curve of divalent manganese 1.5 sin 0/), (International Tables for X-ray Crystal- lography, Vol. 3, pp. 202-205) was compared with a composite curve for (Fe3lr' Fefl'orMno.orTi6.6rMgoo+ represented too much scattering power for those sites. Alo or).Nearly perfect agreement(+0.3 electrons)was A moderate positive anomaly at the interlayer K found for all values of sin 0/\. Thus, Ml and M2 position signaled the need for including a sodium occupancies were refined using (Mnlt!'-") as a contribution to that site. No other significant errors composition model. After five cycles of refinement, were detected. with independent Ml and M2 occupancy variables, Manganese, titanium, and iron have similar scat- convergencewas reached wilh Ml : M2 : (Mn3]nru tering factors. Thus, the contributions of these octa- no our) I : 4.4%). On the final cycle, 69 of 1586 hedral cations (Fefr*rrFefitorMnoorTio.or) in annite observed data were reiected with AF > 5.0. were conveniently modeled as Fe3lrr. Similarly, aluminum and magnesium were treated as Mgo oo.On Discussion the basis of wet chemical data and Ap"u, magnitudes, a new composition model of (Fel*rrMgo oono oo)was Atomic Coordinates selected for Ml and M2. The interlayer site was redefinedas (Ko n.Naonr), consistent with the chemical The atomic coordinatesand anisotropic temperature analysis. These occupancies were held constant for factors for phlogopite and annite are presented in three refinement cycles (R: 5.3%),but Ml and M2 Table 4 along with the "ideal" fractional coordinates compositionswere then allowed to vary independently, of Pabst (1955) for comparison. Principal bond with total occupancyofeach site equal to 94 percentas distances,mean coordination polyhedra bond lengths, the only chemical constraint. After three additional and bond anglesappear in Table 5. Refined values of cycles (R : 4.4%) convergence was reached with atomic coordinates closely approximate those of Ml (Fefi*rruMgoouuno oo) and M2 (Fe3lrr. Pabst's ideal model, and in no instance deviate more Mgo.ourno.ou).These values are equivalentwithin the than 0.02 fractional units from them. In both struc- estimatedstandard deviations (+0.005), and, thus, no tures, basal Ol and 02 oxygensare coplanar in (001), preference of Fe for Ml or M2 was indicated. as are the apical 03 oxygen and (OH, F) positions. One of the principal objectives of this investigation Note that the orthohexagonal cell parameter relation- was to detect the existenceand distribution of octa- ship, b : tE a, also applies to both specimens (see hedral vacancies. In an effort to refine accurately Ml Table 3). Therefore, deviations from the ideal mica and M2 site occupancies,a secondcomposition model structure may be discussed solely in terms of tetra- was attempted. Octahedral cation chemistry of the hedral rotation and octahedral flattening. Pikes Peak annite is complex, with Fe'*, Fet*, Mg, Octahedral flattening 'y' is defined as the angle Mn, Ti, and Al in significant amounts. Furthermore, between a line drawn perpendicular to two parallel least-squaresrefinement procedures can permit only faces, and a line joining opposite apices of an octa- one occupancy variable per atom in the asymmetric hedron (Donnay et al,1964a). {, is 54.73"in a regular unit. Thus, to refine total occupancy. of the octa- octahedron, and will exceed that value in a flattened hedral site, a suitable scattering factor model was form. In the ideal model a single value of ry'is sufficient needed to approximate the sum of scattering con- since all octahedra are assumed to be of equal size ONE-LAYER PHLOGOPITE AND ANNITE E93 and shape.However, in real micas separatevalues of ry' 1, plotted values of observed octahedral flattening may be required for the Ml and M2 sites. The octa- angle os octahedral cation radius show close agree- hedral flattening angle rt' may be easily derived from ment to those predicted by Hazen and Wones refined structural parameters as follows: (1972). Since basal oxygens are coplanar, the tetrahedral , t' cos : rotation angle o may be simply calculated from the Y 2.d^' ye1 fractional coordinate (see Fig. 2) by the relation: where - / 2(zor) * zo,r\ 2b(0.25 v\ h r,,: 2\0.5 sJc.sin g tand : - - : a-Q.25 - y) 0J0, (McCauley, 1968), wherc do is the mean cation-to- This may be further simplified to anion octahedral bond distance, to is the octahedral sheet thickneSSlZo3 and zorr are atomic fractional co- tana: 4\/3(0.25 - !o) ordinates, and c and B are unit-cell parameters. For phlogopite's Ml and M2 octahedra, 12 angles are for those micas in which the orthohexagonal lattice both 58.96', while for annite they are 58.55. and relation is valid. Values of o for phlogopite and an- 58.20" respectively. Hazen and Wones (1972) pre- nite thus calculatedarc 7.5" and 1.5' respectively, dicted values of,2 as a function of octahedral cation and agree almost precisely with magnitudes of tetra- radius. The ionic radius of Mgr. in phlogopite is hedral rotation oJ mean octahedral cation radius of 0.72 A (Shannon and Prewitt, 1970), while the ob- Hazen and Wones (1972). Predicted and refined servedcation radii of annite MI and M2 are 0.765 A values of a aS a function of cation radius are illus- and 0.75 A respectively, assuming the radius of Or- trated in Figure 3. is the same in both annite and phlogopite. In Figure Aluminum and silicon are disordered in the tetra-
TreI-E 4. Atomic Coordinates, Anisotropic Temperature Factors, and Equivalent Isotropic Temperature Factors
h Atomxvz Bzz Brr Brr grr q2r lequ1v
Pabst ( 1955) ldeal HodeI
T c .57': r/3 0.220 !41 0 r/2 T/? lti2 0 5/6 1/? K 0 0 0 01 0.8c5 I,/ 4 0.160 O2 0.055 0 0.160 03 0.630 r/6 0.380 0li c.130 0 0.380
Franklln PhloDoolte
r 0..)7rz(r)c 0.1.668(5) 0.2254(1) o.o0q1(1) 0.0011(1) o.oo22(.r) -0.0001(1) 0.0004(1) -0.00007(4)0.57(1) llr 0 I/2 I/2 0.0017(4) 0.0009(1) 0.0028(1) 0 o.c00B(2) 0 0.6r(2) "4? 0 0.831511) r/2 0.0017(2) 0.0009(I) 0.0029(1) 0 0.0007(r.) 0 0.62(1) K000 0.0205(5) 0.0056(2) 0.0071(1) 0 0 0013(2) 0 2.45(3) or 0.B248(3) 0.2307(2) a.1677(2) 0.0104(rt) 0.0038(2) 0.0038(2) -0.00r9(2) -0.0002(1) ()2 0.0013(2) 1.31(l) 0.5180(lr) 0 O.)"67E()) c.0149(7) 0.0026(2) 0.0028(2) 0 -0.0001(3) 0 1.25( 4 o3 0.6297(?) ) 0.L66rr(t) 0.3902(1) 0.0047(3) 0.00r.1(r) 0.0026(1) -0.0000(2) 0.0008(2) -o.ooo2(1) 0.64(2) orr o.r.3l0(3) 0 0.4008(2) o.oo6r(5) o.oo17(2) o.oo29(2) 0 0.0008(2) 0 0.80(2)
Plkes Peak Annlte
0.5703(r) 0.16{55(1) 0.2246(1) o.oo75(2) o.oo24(1) 0.0036(r) -0.000i(r) 0.00r6(1) r.03(r) 111 0 r/2 )./2 0.0070(3) 0.00r8(1) 0.00q0(r) 0 0.0022(r) 0.98(2) tj2 0 0.8332(1) I/2 0.0061(2) O.oo28(r) 0.0039(1) 0 0.0014(1) I 06(1) K 00 ) o.0277(7\ 0.0100(3) 0.0116(3) 0 0.0014(3) 0t- 3.65(5) 0.8031(rl) 0.2u57(3\ 0.1670(3) 0.0r40(7) o.oo70(3) 0.0050(2) -o.oo26(4) 0.0025(3) 1.98(4) c2 a.5\27(6) 0 0.1681(rr) 0.02t3(12) O,OolOlrrl 0.0055(4) 0 0.0022(5) 1.96(9) o3 0.6291(4) 0.1674(3) 0.3894(2) 0.0109(6) o.oo35(2) o.oo4o(2) -o.ooo4(4) 0.00t7(3) 1.34(3) ofi 0.1239(6) o 0.393r(/r) 0.0138(t.0) 0.0037(4) 0.0040(3) o 0.0017(5) 1.48(6) a) Equlvalent lsotroplc tenperature factors are computed accordlng to D-,., .. = \/31t.1,.8,,a..a.), and are related 1n the standar'l wav to a vlbratlon sphere that approxlmates r4u'uro..iiliu.tt-,o"orl i.li"io"* by tne anrsotroprc temperature factor tensor (Ham11ton, 1959). b) The anlsotroplc temperature factor forn 1s exp[-tltJBlJhlhJ]. c) Parentheslzed flgures represent the estlnated standard devlatlon (esd) ln terms of Least unlts clted for the value to the tmnedlate teft, thus 0.5703(1) tnatcites ""."J-.f'ffiooi- &94 R. M. HAZEN, AND C. W. BURNHAM
Tasr.B 5. Interatomic Distancesand Bond Angles
o (A) Itngte ( iJond Dlstance (A) Atons angte (o) Ilond Dlstance )
Phlogoolte Tetrahedra Annlte Tetrahe(lra 0t-T-o1 rc8.52(6) r-or 1.655(2) 0l-T-o1 10u.5(I) T-01 1.6rrij(2)1 -T-02 -0r. 1.65r(1) .)1-T-0 2 r0il.29 (c ) -or r.666(2) 0l 108.7(2) -a2 r.6503(q) o1-T-02 1oii.50(! ) -oz 1.653(2) o 1-'r-o 2 108.3(r) _oj 1.6119(I) 0 1-T-0 3 1ro.6rr(7) -03 1.661(2) o I-T-O 3 110.2(1) ol-T-0 3 110.rr3(7) 0 1-T-O3 t4ean T-o 1 6ll9 t,leanT-o I.559 0 2-T-O3 110.rro(B) 02-T-01 l,1ean 0-T-O 109.rr6 )4eanC-T-o 109.5 or-01 2.517r(5) oL-or ?.(,9\(a) o1-02 2.679(2) 01-02 2.687(3) ot-02 2.673(2) 01-o1- 0 2 59.s8(7) o1-02 2-699(3) 01-01-02 59,8(1) o1-03 2.749t2) 01-o1-o2 60.a9(6) o1-o3 2.719(3) o1-or-o2 60.2(t) cr-03 2.7r\(2) 0 1-01-0 3 60.371111 o1-03 2.731(t) o1-or-03 60.5(1) a2-o3 2.709(2) 0 t-o 1-o 3 60.42(1) a2-o3 2 .72r( Lt) or-c]-o3 60.2(1) o1-02-01 60.03(o) o1-o2-ol 6o.o(1) MeanO-0 2.693 l4eanO-0 2.142 r)1-o2-o3 60.38(t) or-02-o3 €'0.r1(1) o 1-o2-0 3 60.rt8(5) D1-02-03 60.4(1) r-r t2l2 3.0634(6) 0 1-0 3-o 1 59.21,(4) r-r [2] 3.111(2) 0r-o3-o1 59.3(I ) r-r 3.065(1) 0 1-o 3-o2 59.12(5) T-T t. Lorr( 2) o1-03-02 59.211) o2 27 o1-03-02 59.3(1) 0 1-0 3- t9 lt) l,leanT-T 3.1CN l'(eanT-'r 3'o6ll o2-ol-03 6a 35(6) o2-01-o3 60.2(r) o2-or-o3 60.40(6) 02-o1-03 60./r(1) l4ean 0-0-0 6o.oo Yean o-o-o 60.0
Phlocooite 1410ctahedra Annlte Yl i)ctaheC.a
t41-o3[r]l 2.oBo(r) o3-r,11-o3[2] 85.3?(1) ll1-o3[rrl 2.]:3(:t) o3-irr1-o3[2] 35.3(1) l'11-0ir[2] 2.0-?0(2) o3-lr1-03[2]r 9lt.68(1) l41-ori[2] 21r9(3) o3-li1-03 [2] 9rr.7(r) o3-).r]-oilfrrl 83.50(5) Ol-tl1-oll Irr] 84.3(r) Mean:.{1-O ?.A63 l4ean111-l 2 .12L ]r (7) o3-tll-oH Irr] 96.50(5) - o3-111-olt [ l 95.7 l4ean 0-:'11-0 90 .00 l.lean0-M1-0 90,0 o3-03 [2] 2.8\5('t) 03-03[2 3 067(t) (adI acent) ( ad.Jacent ) o3-03i2 2.Br.B(2) o3-03 f2l t.12t(5) o3-0il [ ]r 3.a67(2) 03-t41-03[2] 180.00 O3-olr; t1; 2.0rr7(3 ) o3-r11-O3[2] r80.00 o3-orr[4] 2.137(2) ot{-i.ll-oit 180. 00 a3-orifrrl 3.r02(2) oii-llt-oii 180.00 Mean 0-0 2.915 ( opDoslte) liean o-O 2,9 /l ( opposlle ) ( adJacent ) ( ad,jacent )
Phlogoplte M2 Octahedra Annlte l'120ctahcdra
1'42-03[2] 2.071(1) 03-^42-01 8t.78(3) "i2-03[2] 2.109(2) o3-i42-03 115.9(1) 142-03[2] 2.0E3(1) o3-M2-03[2] 35 6),(t) r12-o3l2) 2.112(2) ca-v2-o3[2] 85.2(1) l'12-oir[2] 2.039(1) n?-rr)-nt f )l 95.18(5) r2-ori [2] ,2.033(?) o3-M2-03[2] 9!.5(1) 03-r12-0ii o3-r.12-0ir[2] 95.4(1) l4eanM2-0 2.064 [2] 96.55(5) l'lean i42-0 2 .101 03-112-oil[2] 33.20(6) o3-M2-OH[2] 85.6(1) o3-M2-0rr[2] q5.99(8) o3-.42-0ri[2] 9q.7(r) 03-03 t2l 2.8r.3(2) 0ti- M2-0t i 8r.r3(t) tl-o3 [2] 2.'8r5(',t) oir-l42-oir 83.4(1) c3-03 r)3-o3 : .377(rr) 2.i)22(2) l4ean 0-Y2-0 l4ean 0-1"12-O 90 .0 o3-o3 90.0c ca-03 t2 l 3.lca(2) 12) 3.059(2) Tadl..anfl ( adJacent ) o3-oii [2] 2.7'17(2) o3-o'i [2] 2.8!1(3) o3-crr [2] 3 a63(2) (.r3-oll[2] 3.0f12(3) (I) r.3-ori[2] 3.068(1) 0 3-)42-03 178.9r1(q) o3-0lr[2] 3.lrt5(3) 03-r,12-O3 179,7 0I-01r 2.65)1(3) o3-1.12-o:r[2] 177.01,(6) cir-oii [2] 2.'t72(i) o3-112-oil[2] 173.8( 1) llean 0-C 2.916 l4ean 0-in2-O \77.63 liean O-O 2.916 Iean 0-142-0 ).79.25 ( adJaccnt ) I ^hn^e1 ral ( ad,lacent ) (opposlte)
fo Tnr6Fl.t,6r tl fa Annlte InLerlayer :l1te
K-c1 [/]l 2,959t.r) t{-o1[rr] 3.162(3) K-02 [2] 2.969(2) Y-a? l2l 3.lrC(lr) l4eanji-o 2.969 i4ean !i-o 3 ' 1l{il ( lnner) ( lnner)
ri-o1 [rr] 3.312(2) i(-01 [rrl 3.194(3) r(-02[2] 3.311(2) x-02 [2] _Ljr6_al_l_L l{ean ii-O 3.3115 i/iean li-o ?,.?16 (outer) ( outer)
^ (K-0) 0.343 ^ (i{-0) 0.071
(esd) l. parentheslzed flFures represent the estlnatecl slandard deviatl.on 1n terms of least unlts clted for the value to the lmmedlate left, thus 1.648(2) lndlcates ai esd of 0.002. 2. BrackeLed flfures represent nultirllclty of the bonl or anrle to LllP lnnedlale 1efL.
hedral sheetof a one-layerC2/m trioctahedralmica, (1963) noted that mean tetrahedralbond length is and, thus, there is only one crystallographicallydis- a sensitivefunction of Al/Si ratio, due to the differ- tinct tetrahedral site. Consequently,the tetrahedral ing sizesof thesecations. Similarly, in this study, the site compositionof such a mica varies directly with mean f-O : 1.659 A of (A11.21Siz.zg)in annite is the aluminum-to-siliconratio. Smith and Bailey greater than that of 1.649 A for (Alr osS'iz.gs)in ONE-LAYER PHLOGOPITE AND ANNITE 895
2(025-Yq lD
-c- o 74
+ TE
o E
O o7o = o rona =31"?i!'* = 46(0 25-Yor) J( ThisSiudy Ftc. 2. Calculation of the tetrahedral rotation angle a Ho..n I Wones(t972) f from the y fractional coordinate of atom Ol. It is assumed . Wones(1963) that Ol and 02 (the basal oxygens) are coplanar. In the ideal structure (Pabst, 1955) ye, - O.25, and tetrahedral rotation is thus zero.
58" V = Arc sin (bm/ 3,/T do')
Ftc. 1. Octahedralflattening anglelt os ionic radius of the octahedral cation for trioctahedral micas of the form KRs*lAlSLO,o(OH)r. Black dots represent predicted values of octahedral flattening for synthetic biotites on the join phlogopite-annite by Wones (196j), while diamonds are predicted values for synthetic micas by Hazen and Wones (1972). .s N + E phlogopite. Figure 4 illustrates mean tetrahedral o bond length as a function of xa1/(xa1 * xsr) for severalpublished mica structures.Linear regression E o analysisof this data yieldeda best-fittingline of : z.
(r - o): 0.163(--4-r--)+ r.ooaA )+ This Study \Xa1 f Xs;/ ,or.n I wones(1972) ) Both annite and phlogopite tetrahedra are remark- ably regular, with bond distances deviatine bv less "Uona than 0.5 percent of the mean, and O-f]O angles varying less than 1.25" from the ideal value of 109.47.. to. Ml and M2 octahedral positions are crystallo- or = Arc cos(b^/4,/2.dt) graphically distinct and difier in nearest neighbor Frc. 3. Tetrahedral rotation angle a us ionic radius of arrangementsof oxygensand hydroxyls (or fluorines). the octahedral cation for trioctahedral micas of the form KR"*'AISLO''(OH)". Black dots represent predicted values Ml is located at a center of symmetry, with two of tetrahedral rotation for synthetic biotites on the join coordinating hydroxyls at opposite apices of the phlogopite-annite by Wones (1963), while diamonds are octahedron, and four coordinating 03 oxygens.M2 predicted values for synthetic micas by Hazen and Wones on the other hand, is non-centric, with two adiacent (re72). 896 R. M. HAZEN, AND C. W. BURNHAM
-l (1969) .72 x Tokodo A Burnhom - I TokduchiI Sodonogo (1966) F (9 (1967) z o GuvonA Burnhom UJ .L ol. (1968) J + Tcipkin llccouley(196E) 2l o o (1967) (D + Gwon o+ J . This Study E. o 8a_ l Xonihophyllrte lr,l ,t' 2 3T Muscoute (T1 ) r 164 6 3 lron Eiolite E Annrte F 5 Phloq opr i e UJ 6 Fluor-Phlogoprle F f 2M l\4uscovrie -Phlogoprte 1r-O2=o.ros(;$J+r6oe 8 Bo-Li- trluor z Fluor-Po I yl i thionite t0 3T Muscwite (T2) uJ t6o 2
rE r RAHEDRAL coMpostr toN (*-^t-*)
Frc. 4. Tetrahedral composition os mean tetrahedral bond distance for micas. Regression analvsis of data from this and other studies yielded a best-fitting line of (?-o): o ror(;ff[) * t.uo'.
coordinating hydroxyls. There are two M2 octahedra In terms of octahedral flattening, tetrahedral rota- for each Ml in the trioctahedral layer. As previously tion, and interlayer cation coordination, phlogopite described, octahedra are distorted through flattening displays significantly greater deviations from the in the c* direction. Mean (OH, F)-M bond lengths ideal model than does annite. This implies that the average0.04 A less than O3-M bond distances,thus misfit of ideal tetrahedral and octahedral sheets is introducing additional octahedral distortions. Mean substantially less in annite, due to its larger mean bond lengths of phlogopite, Mt-O and M2-O (both octahedral cation radius. Furthermore, Pikes Peak 2.063 A), are shorter than those of annite, Ml-O annite is extremely close to the ideal structure, with (2J22 A) andMz-O (2.102A). a tetrahedral rotation of only 1.5o' Thus, micas with The coordination polyhedron of the interlayer an AlSi, tetrahedral sheet are not expected to be cation is sensitive to deviations from the ideal mica stable if the octahedral mean cation radius is signifi- structure. In the ideal model, potassium is twelve' cantly larger than that of this annite, i.e', 0.755 A' coordinated in a regular hexagonal prism. Holvevet, This is consistent with the prediction that end-mem- is as ditrigonal distortion (i.e., tetrahedral rotation) of ber manganous phlogopite KMn'?-3AlSisOro(OH)z the tetrahedral sheet increases,six of these coordi- not a stable phase (Hazen and Wones, 1972) ' nating oxygens move closer to, and six farther away A p p arent T hermal V ibratio ns from, the potassium. Thus with increasing rotation angle o, potassium approaches six coordination in an Magnitudes and orientations of apparent atomic octahedron. In annite, potassium is in nearly perfect vibration ellipsoids are presented in Table 6, and twelve-coordination, with K-O distances ranging are illustrated in Figure 5. These apparent thermal from 3.1I to 3.26 A. However, in phlogopite six motions are striking in both their size and aniso' "inner" K-O bonds are 2.97 A. while the other tropies, when compared with those of other silicate six "outer" K-O bonds are 3.31 A in length, and minerals (Bumham, 1965). Of eight atoms in the potassium may be consideredto be in approximately trioctahedral mica asymmetric unit, six (i.e., T, K, six-coordination. Ml, M2, 03, and OH) have rms displacements ONE.LAYER PHLOGOPITE AND ANNITE 897
TasLB 6. Magnitudes and Orientation of Thermal Ellipsoids
rns (A) Atom Axls wlth respect to: rns (A) Ie wlth respect to: Dlsp lacemen t a(') !.(.) c(") D1sp 1ac enent a (o) q (") c (o) T rr 0,059(t)l 79(I2) 12(12) B8(2) 0.r00(2) 175(5) 89(31) 12 0.075(t) 169(9) 7\(2) 78(12) 93(2) 0.103(2) 9t(31) t77(7) 11 0,IoZ(t) 100(1) 93(9) B9(2) 3\2) 0.r37(1) 85(2) 93(3) 16(2) r1 0.062(4) 90 0 90 0.089(2) 12 90 o 90 0.071(3) r77(3) 90 B3(3) 0.092(2) r3 0.120(2) Lo(r) 90 110(1) 93(3) 90 7(3) 0.145(2) 80(r) 90 20(1) P1 0.062(3) 90 0 90 0.09r(1) 12 0(r) 90 r01(1) 0.071(2) 175(2) 90 85(2) 0.1r1(r) r! 0.122(2) 90 o 90 90 5(2) 0.141(1) 90(1) 90 r1(r) 0.167(2) 20(4) 9A Bo(4) 0.179(3) 15(2) 0.168(2) 90 85(2) 90 0 90 0.210(3) 90 o a.192(2) rlo(3) 90 90 10(3) 0.2rr8(3) r05(2) 90 5(2) 0l o.lo2(3) q0(3) 50(3) 94(3) 0.r22(t1) 153(3) 0.138(2) 114(3) 59(r) 55(13) 115(r7) 1rr9(23) o.16r(4) 99(6) 159(q) o.trr2(3) 6r(12) 9B(7\ 130(r3) 6o( 23) 0.1u5(4) 64(3) r5!(3) 92(8) 0.106(4) 90 0 90 0.132(6) 90 0 90 0.r16(4) 108(5) 90 r52(5) 0.167(5) 0.150(rr) 7o(35) 90 171(35) 162(5) 90 62(5) 0.171r(5) 20(35) 90 81(35) 03 0.070(3) 94(15) 8(7) 83(3) o.r2o(4) ]q3(25) 0.080(3) r27(25) 79(10) 176(11) 93(15) 82(t4) 0.r26(q) 5u(25) r\2(26) ft 0.115(2) r.o6(12) 92(3) 97(3) 11(3) 0.144(3) 82(7) 96(1r) 20(8) 0.085(4) 900 90 0.128(5) 90 o 0.092(4) 175(6) 90 90 85(6) 0.138(6) 2r(39) 90 121(39) 0.r20(3) 95(6) 90 5(6) 0.145(5) 69(39) 90 32(39) 1. Parentheslzed flSures represent the estlnated standard devLatlon (esd) ln terns of least unlts clted for rhe vafue ro the lnnedlate reri, t;;" o:06tilt i"a-il.i""-in or 0.001. ""o
elongated by as much as 2:l:l approximately imposedon all atomsof the trioctahedralmica asym- parallel to c* in both phlogopite and annite. The metricunit. basal 01 and 02 oxygensare less elongatethan the An alternative explanation of observed thermal other atoms,but the magnitudesof their equivalent2 ellipsoid elongation along c* has been suggestedby isotropic temperature factors average 85 percent David R. Wones of the United States Geological greaterthan the apical 03 oxygens.It is improbable Survey. Wones (personal communication) noted that suchstrong apparentlyanisotropic vibrations are that "if F and OH are orderedin the sensethat all F due entirelyto thermalmotion (Burnham, 1965). is in a single layer and these are randomly mixed Observedanisotropic vibration ellipsoids may be with OH layers, then (one might observe) elongate correlatedwith positional disorder of the tetrahedral temperature factors along c*." Similar arguments site. As previously discussed,aluminum and silicon might be extended to ordering of cations btween are disordered in the single tetrahedral position. the mica layers. Since aluminum-to-oxygentetrahedral bond lengths The magnitudesof apparentvibrations in annite, (=1.77 A, from Figure 4) arc considerablylarger as indicatedby equivalentisotropic temperaturefac- than those of four-coordinatedsilicon (=I.6I A), tors, average75 percentgreater than thoseof phlogo- oxygen positions will vary locally dependingon the pite. This may be due to the considerablesubstitu- coordinating tetrahedral cation. Apical 03 oxygens tional disorder in all of annite's cation sites. Not are coordinatedto one tetrahedron(see Figure 6), only is the content of tetrahedral aluminum greater and, thus, the apparent thermal ellipsoids for 03 in annite (A11.2?Sip.ze)than in phlogopite (Alr.ou show strong elongation parallel to the bond (i.e., Siz.so),but the octahedral and interlayer sites have parallel to c*). Basal 01 and OZ oxygensare co- complex cation occupanciesas well. Thus, ellipsoids ordinated to two tetrahedra each, thus explaining representingaveraged atom positionsare expectedto the significantlygreater apparentvibrations of basal be larger for annitethan for phlogopite. us apical anions.The positionsof Ml, M2, K, and (OH, F) respondto local variationsin oxygenposi- Annite OctahedralSite Occupancy tions, and large apparent vibrations are therefore Refinement of octahedral occupanciesusing the model Ml I [Feff.nn_",Mg,Eo.ou]and M2 S." nrrt tootnote to Table 4. [Felf.nn-,,Mg,Eo.ouJresulted in x : y within the 898 R. M. HAZEN, AND C. W. BURNHAM
PIKES PEAK ANNITE Fro. 5. Stereoscopicview of one unit cell of Pikes Peak annite, showing apparent thermal vibration ellipsoids. The structure is projected with the D-axis vertical, the c-axis vertical, and the a'axis normal to the page. Note the elongation of ellipsoids parallel to c. estimated standard deviation. However, the mean Foster, 1960). These vacancies have been assumed Ml-O bond distance (2.1224) is slightly larger than by several writers to represent solid solution between that of M2-O (2.|OZA). Thus, while there appears to trioctahedral micas and the dioctahedral variety in be no evidence for significant ordering of Mg and Fe which only M2 positions are occupied. Such a model, in the Pikes Peak annite, ordering of Fe, Ti, and Mn however, is not valid for the Pikes Peak specimen in may occur. It should be emphasized that the least- which equal fractions of both MI and M2 are vacant. squares refinement program nnNB allows only one Professor J. B. Thompson has proposed an alter- occupancy variable per atom in the asymmetric unit. native solid-solution model in which: Ti4- + n = Fe'*, Fet*, Mn'*, and Tint possesssimilar scattering 2(Fe2',Mg2*). The resultant end-member triocta- factor curves, and were thus conveniently modeled as hedral micas would be: Fe'* in the above occupancy refinement. However, K(Fe,Mg), (AlSi3) O1o(OH), these four cations have widely differing ionic radii (Mg,Fe Ti] (AlSia O1o(OH)2. (Fe'* : 0.78A, Fe'* : o.4gA, Mn'* : 0.83A, and - K[ ) ) Tin* : 0.61 A; Shannon and Prewitt, 1969, l97O). However, sinceK(FeTi) (AlSi3)O1o(OH), is known Thus, iron, manganese,and titanium ordering may to be a dioctahedral mica (Wones, personal com- occur, and could account for the small differences in munication), it is doubtful that complete solid- Ml and M2 octahedral bond distances. solution between these end-members is achieved. Refined values of octahedral occupancies indicate Note that in this model, the atomic fraction of approximately 7 percent vacancies in both Ml. and titanium equals the fraction of vacancies. In the M2 sites of this annite. Other authors have suggested Pikes Peak specimen studied, Ti = a = 0.2, and that as many as 15 percent of the octahedral sites similar results are obtained for other titanium-biotites are vacant in some biotites, based on calculations of from the Pikes Peak batholith (Barker et al, 1972). structural formulae from chemical analyses (e.9., Our experimental results, therefore, strongly sug-
ANNITE TETRAHEDRA Frc. 6. Stereoscopic view of a portion of the annite tetrahedral layer, showing apparent thermal vibration ellipsoids.Orientation is as describedin Figure 5. ONE.LAYER PHLOGOPITE AND ANNITE 899 gest that the presence of octahedral vacancies in of these parametersmay be predicted accurately many natural biotites does not necessarily imply from basic crystal data and plane geometry. Other solid-solution between trioctahedral and dioctahedral structural properties such as apparent anisotropic end-membermicas. thermal motion may be correlated to substitutional It is interesting to speculate on the effects of Tia* disorderwithin atom sheets,while polytypism and and vacancies on local charge balance in triocta- twinning in micas is most convenientlydiscussed in hedral micas. Each apical 03 oxygen is coordinated terms of stacking variations between these sheets. to one tetrahedron and three octahedra. Assuming Thus, simple geornetricalmodels may be used to that no 03 oxygen is coordinated to more than one understandmore fully the crystal chemistryof these Tia* or vacancy, there are eight possiblearrangements andother layer silicates. and electrostatic chargesfor the cations surrounding this oxygen: Acknowledgments The authors gratefully acknowledge the contributions of Neighborsto 03 Charge Neighborsto 03 Charge Professor J. B. Thompson of Harvard University and (Fe*FetFe)*Si 2 (Fe*Fe*Fe)*Al l-3/4 Dr. D. R. Wones of the United States Geological Survey. (Ti * Fe * Fe) * Si 2-1/3 (T1f Fe * Fe)* Al 2-1/12 Their stimulating discussions and careful review of the (Ti*Pe+!)+si 2 (TiiFe*tr)+at F3/4 manuscript have greatly improved the quality of this work. (Fe f Fe + fl) + si r-2/3 (Fe* Fe f E) + at Fs/12 Special thanks are also due to Professor Roger Burns of the MassachusettsInstitute of Technology, and to Dr. Jun Of these eight cation configurations, only three Ito and Mr. Stephen Roberts of Harvard University, for [(Fe3Si), (TiFerAl), and (TiFe!Si)] are close to aid in Mcissbauer, wet chemical, and rnicroprobe analysis, the ideal net charge contribution of 2-l for oxygen. respectively. We are indebted to Professor Clifford Frondel Oxygens associatedwith (FozESi) and (Fer[Al) are of the Harvard Museum and Dr. Fred Barker of the United significantly undersaturated, and vacancies in such States Geological Survey for providing specimensfor study. Finally, we wish to thank Dr. J. R. Smyth, Dr. Y. Ohashi, configurationswould causesignificant negative charge and Mr. T. L. Grove for their aid and advice in both labora- imbalances.Shortrange ordering associatingtitanium tory and computational aspectsof this study. with vacancies could minimize this unstable local Research on the crystal chemistry of silicates was sup- charge imbalance. ported by National Science Foundation Grant GA-12852.
Conclusion References B.ln
FINGER,L. (1969) Determination of cation distribution by Peurruc, L. (1930) The structure of micas and related least-squares refinement of single-crystal x-ray data. minerals. Proc. Nat. Acad. Sci. 16, 123-129. Carnegie Inst. llashington Year Book, 67, 216-217. Reoosrovrcs, E. W., exo K. NonnrsH (1962) The cell FosrER, M. D. (1960) Interpretation of the composition dimensionsand symmetry of layer lattice silicates.I. Some of trioctahedral micas. U. S. Geol. Suru. Prol. Pap. structural considerations.Amer. Mineral. 47, 59'9-616. 354-B,11-46. Ross, M., H. Terr,oe, ,rNo D. R. WoNss (1966) Mica Fne.NzrNI,M., eNr ScHrenrINo (1963) On the crystal struc- polytypes: systematic description and identification. ture of biotite. Z. Kristallogr. ll9, 297-309. Science,l5l, l9l-193. GuvEN, N. (1967) The crystal structure of 2M, phengite SneNNow, R. D., .rNo C. T. Pnnwrrr (1969) Effective and2M' muscovite.Carnegie Inst.14/ashington Year Book, ionic radii in oxides and fluorides.Acta Crystallogr. B2S, 66,487-492. 925-934. lNl C. W. BunNHnt (1967) The crystal structure AND - (1970) Revised values of effective of 3muscovite. Z. Kristallogr.l2S, 163-183. ionic radii. Acta Crystallogr. 826, 1046-1048. HelrrlroN, W. C. (1959') On the isotropic temperature SvrrH, J. V., eNn S. W. Berrr.v (1963) Second review of factor equivalent to a given anisotropic temperature fac- Al-O and Si-O tetrahedral distances.Acta Crystallogr. tor. Ac ta Cry stal Io gr. 12, 609-610. 16,801-811. Hlrcn, R. A., R. A. Hulrpnnrv, W. Etrlr, .ttto J. E. aNo H. S. Yoosn (1956) Experimental and theo- Colcrono (1957) Synthetic mica investigations. IX. retical studies of the mica polymorphs. Mineral. Mag. Review of progress from 1947-1955. U. S. Bur. Mines 3t,209-235. Rep. Inoest.5337, SrerNrrNr, H. (1962) Crystal structure of a trioctahedral HezEN, R. M., eNo D. R. WoNEs (19'72) The efiect of mica: phlogopite. Amer. Mineral. 47, 886-896. cation substitutions on the physical properties of tri- TlrEoe, H., eNr C. W. BunNrr,tnr(1969) Fluor-polylithion- octahedral micas. Anter. Mineral. 57, 103-125. ite: a lithium mica with nearly hexagonal(Si,O") ring. HrNonrcrs, S. 8., eNo M. JsruensoN(1939) Polymorphism Mineral. 1. 6, 102-109. of the micas.Amer. Mineral. 24,729-771. lNo J. D. H. DoNN,rv (1966) Trioctahedralone- International Tables f or X-ray Crystallogruphy (1962) layer micas. III. Crystal structure of a synthetic lithium Vol. III. Kynoch Press,Birmingham, England. fluor-mica. Acta Crystallogr. 20, 638-646. JecrsoN, W. W., eNr J. Wrsr (1930) The crystal structure Texeucnt, Y., eNo R. Slpexecl (1959) The crystal struc- of muscovite.Z. Kristallogr. 76, 211-227. ture of xanthophyllite. Acta. Crystallogr. 12, 945-946. AND- (1933) The crystal structure of musco- AND- ( 1966) The crystal structure of xantho- vite. Z. Kristallogr. 35, 160-164. phyllite revised.Mineral. 1. 4, 421-437. McC.ruI-Ev, I. W. (1968) Crystal stuctures of the micas TErpxtN, E. V., V. A. Dnrrs, eNo V. A. Arnx,tNmove KMg"AISi,OnF" and BaLiMgtAISi,Ou,F,. Ph.D. Thesis, (1969) Crystal structure of iron biotite and construction PennsylvaniaState University. of severalstructural models for trioctahedral micas. Proc. R. E. Ne,wNsAM,AND G. V. Gtsss (1973) Crystal Int. CIay Conl.,Yol. 1, Tokyo. structure analysis of synthetic fluorophlogopile. Amer. WoNrs, D. R. ( 1963) Physical properties of synthetic Mineral. 58,249-254. biotites on the join phlogopite-annite. Awer. Mineral' M,rucurN, M. Cn. (1927) Efide du mica muscovite au 48, 1300-1321. moyens des rayons X. C.R. Acad. Sci. Paris, 185' 288- Zw,tcrN. B. B., rNo K. S. MrsncnrNxo (1962) Electrono- 291. graphic data on the structure of phlogopite-biotite.Sor-r. (1928) Etudes des micas non fluores au moyens des Phys. C ry stallogr. 7, 5O2-505. rayonsX. C.R. Acad. Sci.Paris, 186, 879-881. PAssr, A. (1955) Redescriptionof the singlelayer structure Manuscript receiued,February 16, 1973;accepted of the micas.Amer. Mineral.40,967-974. lor publication, MaY 25, 1973'