THE AMERICAN MINERALOGIST, VOL. 53, JULY_AUGUS'I, 1968

THE OF AUGELITE 'r.An.trr, Deportmentof Geologyond Geophys'ics,Uni,aersity o1Minnesota, Minneapolis, Minmesota J.J.FtNNev,Department oJ Geology, Colorado School oJ Mines, Golden, Colorad,o AND '1. Zorrx, Department of Geology and. Geophysics, Llnirersitt o;f M innesota, M i,nneab olis, M innes ota.

Aesrnncr

An augelite, Al:(POr) (OH),,, from Mono County, California, has unit cell o:13.124 +0.006.b:7988+0.005,c:5.066+0.0C8]\,8:112.25+0.02'anditsspacegroul>isC2/ ru, in agreementwith Peacock and Moddle (1941). Patterson maps were calcuJatedtrom three-dimensional intensities and solved by the minimum function method. The structure rvas refined by a sequence of Fourier syntheses and least-squarestechniques to an R value of 6.6 percent for all and 5 0 percent for all but the eight strongest reflections. Al has two drfferent coordinations in this structure. One is coordinated to four OH and trvo O and the other fo three OH and two O. The P atom is coordinated to four O. Four Al polyhedra, two of each t),pe, are linked together by sharing OH-OH edges. These groups in turn are connected to P-tetrahedra sharing corner O atoms.

ExprntlrnNIAL PRocEDURE

A fragment from a large crystal from Mono County, California, measttring 0.65,0.65 and 1.0 mm along the a-,b- and c-axes respectively, was mounted on the goni- ometer with the c-axis as rotation axis. The precession photographs ofiered a choice between the pseudoorthorhombic body centeredunitcellwitho:12 14+0.02,b:7.98+O.02,c:5.07+0.024and9:903+0.1' and the monoclinic C-centeredunit cell with o:13.13+0.02, b:7.98+0.02, c:5.07+ 0.02 A and B:112.4+0.15" These cell parameters were refined from powder diffractometry data with least-squarestechniques and the following parameters rvere obtained for the C-centeredmonoclinic unit cell:a: 13.124+O006 A, b:7.988 + 0.005A,c:5.066+ 0 003 A, 9:11225+O.O2",2:4. The difiraction symmetry of the crystal and the systematic ab- sences of difiraction spot indicated the possible space groups of C2/m, Cm, or C2. AII the above data are in agreement with those determined by Peacock and Moddle (1941). Intensities of 1474 reflections were collected on an equi-inclination, single-crystal difiractometer using molybdenum radiation and,Zr-Y balanced filters withinihe upper half of the reciprocal lattice limited to sin 0:0.503. Three reflections (202), (404) and (606), however, were not scanned because these reciprocai points are too close to the rotation axis. The usual corrections were applied for absorption and Lorentz-polarization with the aid of Charles W. Burnham's (1966) computer program written for_an arbitrary-shaped crystal, but no attempt rras made to correct for extinction effects at this stage. For the 26 unob- served reflections, the value of the standard deviation of the intensity measurements was used as the intensities, giving a total of 782 independent observed reflections.

1096 STIIUCTUREOF AUGELIT']J IO97

Srnucrunp DrrBnurN.q.uoN The,N(z) test of Howells, Phillips and Rogers (1950),using all reflec- tions, disclosedthat the structure is centric. Accordingly the three di- mensional Patterson maps were solved for the centric spacegroup of C2/m. The tracing of image lines and the judgment of the heights of the prominent peaks on the Patterson maps led to the locationsof four AI atoms on two-fold axes,and another four Al and the P atoms on mirror planes.Accordingly the coordinatesof the Al and P atoms were obtained directlv from the Pattersonmaps. The coordinatesof the atoms and the hydroxyl groups,however, could not be determined,because of peak broadeningin areaswhere they could be expected.Using the ob- servedAI-AI and P-P inversionvectors M6 maps wereprepared which re-

Taelr 1.Arorrrc PaneunrrnsrN Aucnr,nr Standard deviations at lower rows, f, y, a's multiplied by 10a,p's by l0r

Number of atoms in Atom the unit cell

Al(1) 1989 lJo Jt+ IM 2 18 43 1A

A1(2) 0 186 238 64 0 18 +2 l4

P 0 126 108 47 0 l.) 35 16 o(1) 0 464 627 r73 0 49 101 37 o(2) 0 t26 1027 267 78 0 0 79 65 96 36 0 o(3) 42t5 1563 280 +69 2054 -230 510 - 650 2 3 16 93 2l 33 49 oH(1) 0853 0 126 284 JO/ 0 0 2 0 18 44 96 0 34 0 oH(2) 107| 1813 8431 loJ 273 1080 .)l 266 t 4.) 2 J it IJ 33 78 18 26 40 1098 ARAKI, FINNEY AND ZOLTAI

Telrn 2. Elr,rpsoros on Tlrn'rlu.lt Vmnarroxs

(")(A)

0.0640.004 1400 6" 900 0 28" 6" .066 .004 900 180 0 900 .089 .003 506 90 0 626

.049 .005 108 7 90 0 .)/ .075 .004 162 7 90 0 8s7 .o77 .004 900 0 0 900

.032 .006 104 5 900 dJ .064 .004 900 180 0 900 .065 .004 145 900 985

.072.007 126 10 900 t4 10 .103 .007 lM 10 900 104 10 .r22.007 900 00 900

052 .010 100 8 900 128 .098 .007 r70 8 900 788 .t82.006 900 00 900

.074.007 t.t I 342 707 .107 .004 305 91 6 r43 5 .190 .004 662 124 2 592

.064 .008 91 9 900 219 .096 .007 900 180 0 900 .101 .007 t9 900 111 9

.084 .005 128 59 116 69 33 28 .087 .006 t27 60 5l JO 79 66 .126.004 605 665 595

B : equivalent isotropic temperature factors. r:rms displacement along the principal axes, 1, 2, or 3. d:angles between the principal axes and the crystallographic axes.

vealed the positions of the missing oxygensand hydroxyls. A single run of Foulier synthesis using the coordinates thus obtained, confirmed the location of all the atoms. At this stage,the R-factor was 18.7percent. Using these atomic coordinates and all reflections a cycle of least squares refinement was calculated which gave negative temperature factors for P and Al (2) and unreasonably low values for other atoms. After removing 81 strong reflections, two cycles of least-squaresrefine- ment were calculated. This refinement yielded reasonableisotropic tem- STRUCTURE OF AUGELITE 1099

1

F ;

a'? ; "o ,9

D-axis projection of the augelite structure. perature factors. In the next two cycles anisotropic temperature factors were calculated.Although the data obtained appearedto be reasonable, - the negative If, I | F" I differencesof the strong reflectionsindicated the presenceof strong extinction effects.In order to remedy that, the scaleco- efficients of secondary extinction corrections (Zachariasen, 1963) were plotted for the strong reflections with sin I as the abscissaof the chart. The results showed that the strong reflections are dispersedin the lower angleregion, especially below sin 0:0.12 which can be attributed to the effect of extinction. In order to correct for that a weighted mean of scale coefficientwas obtained from the 70 strong high angle reflections and was applied for all reflections using Zachariasen'sformula, with C (the co- efficient for secondary extinction correction): (1.756+0.132)X 10-6. After applying this correction and using all reflections, a cycle of least- squares refinement was calculated which resulted in slightly lowered temperaturefactors. 1100 ARAKI, FINNEY AND ZOLTAI

o o F

q 3rnp

Frc. 2. c-axis projection of the augelite structure.

- The eight strong r eflectionswhich still had negative F, I i F" I differ- ences were removed and three more cycles of least-squaresrefi.nement were calculated. Reasonable temperature factors and significantly smallerstandard deviations for atomic parameterswere then obtained. All the above computations were performed with the University of Minnesota's CDC 6600 computer using L. W. Finger's program.l Throughout the calculations,neutral atomic scatteringfactors (Cromer and Waber, 1965)were used, and the contribution of the hydrogenatoms to the structurefactors was neglected.Weights were set as the reciprocals ol o(F2).The final R factors were 5.0 and 6.6 percentrespectively for the unrejected and for all reflections. The final atomic parametersare tabulated in Table 1 and the ellipsoids of thermal vibration ih Table 2. Lists of observedand calculatedstruc- ture factorsare availablefrom the authorson request.

DBscnrprroN ol rHE Srnucrune The structure of augeliteprojected along the D-and c-axesis illustrated in Figures I and 2, and a stereoscopicview of the bonding polyhedra is

1 Leest-squaresancl Fourier synthesis programs (1955) STRUCTU R]], OF AUGELITE 1101

Talr-n 3. Snmcron BorqoDrsrnNcrs (r-rss urex 2 95 A)

POr-tetrahedron P -o(1) 1 s29sA + 0.0029A P -o(2) 1.5356 0.0028 P -o(3) 1.5059 0.0028 (xz) o(1) -o(2) 2 4819 0.0039 o(1) -o(3) 2.5129 0.0035 (x2) o(2) -o(3) 2.4353 0.0040 (xz) o(3) -o(s;: 2 4977 0 0051

Al (1)-octahedron A1(1) -o(r;s 1.8264A +0 0ffi04 (xz1 Ar(1) oH(1) 1 9830 0 0029 (xz;t Al(1) -oH(2) 1 8639 0.0021 (x2) o(3)u -o(s;, 2.8277 0 0047 o(3)u -oH(1) 2.7853 0.0030 (xz7 o(3)u -oH(2) 2.7244 0.0031 (x2) o(3)u --OH(2)4 2 6168 0.0038 (x2) oH(1) -oH1t;+ 2.37s2 0.0056 oH(1) -oH(2) 2.4087 0.0031 (xzS oH(1) -OII(2)4 2 8491 0.0037 (x2)

A1(2)-polyhedron Ar(2) -o(1) r.7984A +0.0028A A1(2) -o(2) 1.7504 0.0032 Al(2) -oH(1) 2.0544 0.0029 Lt(2) -oH(2) 1.7788 0.0029 (x2) o(1) -o(2) 2.7076 0.0037 o(1) -oH(2) 2.6750 0.0035 (xz) o(2) -oH(1) 2.7152 0.0045 oH(1) -oH(2) 2.4087 0.0031 8z) oH(2) -oH(2)3 2.8965 0.0046

Superscripts identify symmetrical equivalent positions (superscript I is omitted in the table): 1. r, y, z z. -!, A 3. r, z f, s. i*x, i-ly, z 6. 4-x, i-y, -z 7. ilr' i-y, z 8. i-r, tlt, -z shown in Figure 3. The most significantbond distancesand anglesare tabulated in Tables 3 and 4 respectively. In the augelitestructure there are two difierent types of coordination polyhedra of the Al atoms. Al (1) is octahedrallycoordinated with four OH groups and two O atoms,the -cation-aniondistances within a poly- hedronrange from 1.826to 1.983A and average1.891 A. The otheralu- minum, Al(2), is surrounded by five neighbors,two O atoms and one OH group locatedwithin a plane and two other OH groupsnearly perpen- tr02 ARAKI, FINNEY AND ZOLTAI

Frc 3. Stereoscopic drawing of the polyhedral model of the augelite structure. r Shaded polyhedra represent five-fold coordinated AI(2).

Tesln 4 Srlrcrno Bom ANcms .q.NoTnrrr SraNoeru DrvrnrroNs

o(3)u _Al/1\ -o(3)' 101.5' +n ?o o(3)u -oH(1) 93.9 0.1 (xz1 o(3)u 4rI(2)4 90.3 0.1 (xz1 o(3)u -oH(2) 95.2 0.1 (xz) oH(1) -oH(1)4 /J. J 0.2 oH(1) -nT{rrt\ 77.5 0.1 (x2) oH(1) -oH(2)4 95.5 0.1 o(3)u - {H(1)4 161.2 0.4 (xz) oH(2) -oH(2)4 171.4 v.l

o(1) -Alr'r\ -o(2) 99.40 +0.2 o(1) -oH(2) 96.8 0.1 (x2) o(2) oH(1) 90.7 0.1 oH(1) -oH(2) 77.5 0.1 (x2) oH(2) -oH(2)3 109.0 0.2 o(1) -oH(1) 170.0 0.6 o(2) -oH(2) 122.8 0.1 (x2)

o(1) -o(2) 108.1" +0.2 o(1) -o(3) 111.8 0.1 (x2) o(2) -o(3) 1064 0.1 (xz1 o(3) -o(3)' tlz.r 0.2

P -o(1) _A ! 1r\ 158.6' +0.3 P -o(2) _At /t\ 149.9 0.3 P --o(3) -Al/1) 148.9 0.2 (x2) Al(1) -oH(1) _Al/1\3 106.5 0.2 (X2) Ar(1) -oH(1) -Al i/?\ 95.1 0.1 (xz) Ar(1) _oH(2) -Al /?) 109 8 0.1 (xz) STRUCTURE OF AUGELITE 1103

dicular to that plane. The cation-anion distancesin the Al(2) polyhedron rangefrom 1.750to 2.054L, and average1.S33 A. Two Al(1) and two Al(2) polyhedra are linked together by sharing OH-OH edgesand are concentrated in layers parallel with the 001plane and are separatedby d(001). Thesegroups of aluminum polyhedra are con- nected with each other by phosphate tetrahedra sharing corner oxygen atoms thus linking the polyhedra into a continuous network. There are no similar crystal structures known in mineralogy. The five- fold coordination of aluminum has been observedin andalusite by Taylor (1929) and by Burnham and Buerger (1960). As in andalusite, thesefi.ve- fold AI coordination polyhedra are linked to other Al polyhedra by shar- ing edges.In a remote way the augelite structure resemblesthat of lazu- lite (Lindberg and Christ, 1959).That is, in both structureshigh density groups of Al- (and Mg) polyhedra are connectedby P-tetrahedra. That group in is composedof two Al- and one Mg-octahedra. The perfect (110) cleavageof augelitecan be explainedby the relatively weak AI-OH bonds acrossthat plane. The Iess perfect cleavagesin the (201) and,(001) planes on the other hand require the breakage of several AI-OH and Al-O bonds. The authors wish to acknowledgethe support of the National Science Foundation for this investisation.

Rrrnnnrcns Bunwnlu, c. w. (1966)computation of absorptioncorrections and the significanceof end effectA mer.M inera.l,.,51, 159-167. M. J. Bunncrn (1961)Refinement of the crystalstructure of andalusite.Z. Kr istallo gr., | 15, 269-290. Cnoutn, D. T., enn J. T. Warrn (1965)Scattering factors computed from relativistic Dirac-Slaterwave functions.Acta. Crystallogr.,18, 104-109. Howrr,ls, E. R., D. C. Pnrr,r,rrs,ano D. RocBns (1950)The probability distribution of X-ray intensity. II. Experimentalinvestigation and the X-ray detectionof center of symmetry. A cta. Cry stal,l,o gr., 3, 210-214. Lrnnnrnc, Menre L., alvo c. L. cnnrsr (1959) crystal structures of the isostructural mineralslazulite, scorzaliteand barbosalite.Acta. Crystallogr.,12, 695-697. Pracocr, M. A., awo D. A. Mootr,r (1941) On a crystal of augelite from California. Mineral,.Mag., 26, 105-115. Tavton, W. H. (1929)The structureof andalusite,AhSiO5. Z. Kristdlogr.,Tl,2OS-219. Z,rcrrauasnN, W. H. (1963) The secondaryextinction correction.Acta. Crystaltrogr.,16, r139-tt&.

Manuscri.ptruei.red., December 26, 1967; accepted.Jor publ,ication,Febru.ary 3, 196S.