Canadian MineralogistVol. 13, pp. 22-26 (1975)

X-RAY CRYSTALLOGRAPHY OF WELOGANITE

T. T. CHEN AND G. Y. CHAODepartment of Geology, Caileton ()nLversity, Ottawa, Ontario KIS 5B6

ArsrRAc"rSingle-crystal r-ray studies of weloganite showed

tlat the mineral is triclinic, PL ot pi, with a -q:9-8q( .1) , , : 8 .988(1) , c : 6 .730( t )A, a :l0-2.84(r), p : tt6.42e) and ? - 59.99(1)..Weloganite has a pronounced rhombohedral

'sub-

cell which resembles in geometry that of rhombo-hedral carbonates. It also displays a pseudo-mono-clinic cell that, corresponds to the cell reported byother authors for the monoclinic polytypo of welo-ganite. Twinning by [103]r:oo is very common, withtwin obliquity @ = 0 and twin index r - 3. Thetwin cell is trigonal and is identical to the cell orig-inally assigned to weloganite. The ideal chemicalformula for weloganite is proposed as Narsrlr(COs)B.3Hro with Z - 1 for the triclinic cell.

-

IutnopuctroN

Weloganite, a hydratcd carbonate of Na. Srand Zt from St. Michel, Montreal Island, eue-!ec, was origrnally described by Sabina er a/.(1968). The symmslry of weloganite was con-sidered to b-e trigonal P3r,, with a = 8.96 andc : 18.06A. The cell contained two units ofSrZrzGHeOil. The authors noted the discre-pancy that, between crossed nicols, nearly allcleavage (basal) fragments gave an off-centredbiaxial interference figure with 2V about 15",The mineral was later restudied bv Gait &Grice (1971) who reported a monociinic poly-type with C2/c, CZ or Cd symmetry and a =8,95, b = 15.54, c = 37.094 and I = LO3'4S,.The r-ray powder diffraction pait"rns of thetwo polytypes were indistinguishable. The rela-tionships betw,een the two polytypes were given4J au= er, bu=2arSln60o, and Cu =2c, co-sec(l8oo-B). Based on a new analysis the chem-ical formula of weloganite was revised to (Srr.seNar.oZto.asCao.r)(COs)s " 2HzO with Z = 24 ftortle monoclinic cell and Z = 6 for the trisonalcell.

Our inlerests in weloganite began when anunknown mineral from Mont St. Hilaire wasbrought to our attention and shown to be an vt-trium analogue of weloganite. The crystals "oftl,is yttrium mineraL were composed of parallelintergrowths of a large numGr of individuatsthat were too small to be separated for single-

cvystal .r-ray studies. It was hoped that once tlecell geometry of weloganite was established thecell parameters of the unknown yttrium mineralmight be derived from tle powder diffrastiondata by its isomorphous relationship to weio-ganite.

The problems of weloganite may be outlinedao follows:(1) The space group P3r, s is one of the few inwhich no minerals and only a few chemical com-pounds are found, as Sabina et al. (7968) cor-rectly pointed out.(2) If the published chemical analyses and spacegroup symmetry (Sabina et al. L968; Gait &Grice 1971) are correct they would require thepositional disorder of Sr, Na andZr atoms whichcrystal-chemically are drastically different.(3) The .r-ray precession photographs of the so-called trigonal polytype showed peculiar extra-space-group systematic extinctions with reflec-tions of the type h-k = 3n absent when hIl andkll arc not equal to 3n.(4) The relattve intensities of many weak reflec-lons o.n the precession photographs appeared tobe variable from crystal to crystal.(5) The deviation of the asute bisectrix by morethan 5o from the normal to the basal cliavage,consistently observed on grains lying on thecleavage plane, suggests that the symmetrv ofweloganite cannot be higher than monoclinic.

Oprrcar OssnnverroNs

Prior to.r-ray single-crystal studies weloganitewas examined optically using both immirsionmounts of crushed grains and thin sections oflarge crystals. Both colourless and yellow vari-eties were examined. The biaxial characteristicsof weloganite r,eported by Sabina et at. (L96g)were confirmed.

In the immersion mounts two types of grainswere noted. Type I grains, although free of lndnlamellae, either did not show shim extinotionsor showed no extinctions at all between crossednicols.^For the type I grains 2V ranged, from 10to 15o. Type II grains were charicterized bvsharp and homogeneous extinction b€twee;crossed nicols and considerably larger 2V (about20"). Type I was by far the morJco,mrnon.

22

X.RAY CRYSTALLOGRAP}IY OF WELOGANITE 23

One thin section of a large crystal cut parallelto tle basal cleavage showed a dbtinct core inthe form of a nearly perfect equilateral trianglesurrgunded by material showing relatively coarsetwin lamellae and patchy extinctions. Along thewelldefined boundaries there was, at places, athin yensel of very fine-grained pyrite. The corematerial gave extinctions that were sharper butnot homogeneous over the whole area. 2V wasvery small, about 2 to 5o, or nearly uniaxial incontrast to the 2V of 10 to 15' for the surround-

ing material. Both the rim and the core, how-ever, gave identical powder diffraction patterns.

X-Rev Cnvsrer,rocneluv

Many crystal fragments were selected fromthe immersion mounts, after careful examina-tion under the polarizing microscope, for single-crystal r-ray diffraction studies. Precession pho-tographs of type II grains consistently showedtriclinic diffraction symmetry corresponding to

.:,:l:11i

,iFro. 1. Precession photographs (0-1eve1, !r. : 30o, MoKa radiation) of a single crystal of weloganite

showing (a) the a*-b* net and (b) the D*-cx netwith a pseudo-mirror plane normal to 08.

Frc. 2. [103] - precession photographs (O-level, p = 30o, MoKa radiation) of weloganite (a) single crys-tal and (b) crystal twinned by [103]rzo" resulFng in trigonal symmetry. Pseudo-rhombohedral symmetryis evident in both (a) and (b) trom the strong reflestions.

u THE CANADIAN MINERALOGIST

spac€i goups Pt or pI (Figs. 1 and 2). The pa-tameters of ,the triclinic cell presented in TableI are results of least-squares refinement usingpowder diffraction data. The cell, although nota reduced one, \tras chosen because the relatiou-shiq between the present cell and the trigonal cellof Sabina et al. (L968) may be describe-d in sim-ple terms (ot r" = arro, bao : -b*u and c,ryo = -3ctr"r") and it brings out clearly the pseu-do.symmetry of the mineral (a-b i_6O; andc8=90o). Due to the high pseudo-symmetry ofthe lattice there are two equally acceptablechoices for the reduced cell, which mav 6e de-::u"d j.9t the p5esenr_ce-ll by the transforma_rions 101/ I 10/001 anO iOiZOtO,/001. The para-meters of both reduced cells are compared inTable 1 with those of the chosen cell.^

with materials used in this study are identicalto that given by Sabina et al. (t968). The in-dexing of the powder pattern based on thepresent triclinic cell is presented in Table 2,

hkr, dcarc(R) aots(R) r hkr, dcalc(A) dobs(ff) r

'I

'100

001r0r0 I1orT't20

ziTlT0127tTr210'l0l' I I I

121rTlztz002211'tIZ

122337c3030rJJU3ozmr031300a t 51721222'T?474272202403I0rIr212

I

8

'I

5

1 0

7

2.012 2.012

Y

a*b*c.qt

8.987O . / J U

102,870'108.74

1 10 .63^ "421.624"

o. r:sBF-10. i 2850 .1696

6 7 a " o

64 .01

8.988O . / J U

102.8401 08.74l l 0 . 67^242't.624'

o. r sggl-l0. I 2850. 1 695

67.74064.0I

420 2.012003 2.010335 1.966305 'l

.96603e 1 .966032 1.966301 I . 966331 1.966241 1.909427 1.9082n I . 90854T r .691FlZ r ,691]40 1.69r150 t . 6914Tr 1.69145r 1.691214 1.677]T3 1.677515 r . 5e0141 't

,590'l5l I .590

4m 1.590033 t .58903t '1.588

302 1.588304 t .s88540 1.5384Tt I .537'tET

I .537152 t .537510 I . 53733T I.49836T 1.498632 1.498363 I .34133t 1.34r

'I .907

1 .676

' I .497

6 Z

iii':'?ffi il3io[?':ll ii"hio:'i3t3i;)oli. ""o,ced ce,,

The weloganite lattice has a nearly idealrhombohedral sub-cell with a,r._ = Vzi*r,/sit_60" - Vzatu/pin6}o - 5.161A and c"h;_ =c*rg = 17.9014. This sub-cell is eviclent fromthe fact that reflections that do not conform tot}te requirement of rhombohedral symrnetry areeither absent or very weak (Fig. 2a). Tlne g"onr"-try of the rhombohedral sub-cell resembl6s tharof the cells of the rhombohedral carbonates suchas calcite (a - 4.9898 and c - 17.060A: An-drews 1950) and bariun calcium--"uit,l*r".(Cao.sBao.r)_ CO', (a =. 5.02 and c = lT.ggAlDonnay 1963, p.'801). This--Gg*t" tiut"lilestructure of weloganite is probably derivablerrom the structure of the rhombohedral carbo_nates.

The weloganite lattice also possesses a pseudo-T1rro{

plane-perpendicular to 64.r" @igs.'lb andza), gtvrng fise to a pseudo_monoclinic symme_TI._-Ihg pseudo-monoclinic cell has a : a.td" =8.988, , = 2btau sin60o = 15.567, c* =" r/zc*torc of c = 36.865A and B = 1b3.8o. cor-responding exactly to the monoclinic

""U ,"-po$d by Gair & Grice (1.971).

The powder patterns of weloganite obtained

Plus many more l lnes

which includes indices only of those reflectionstlat make substantial contributions to the pow-der diffraction lines as judged frona preceisionand Weissenberg x.-ray photographs.

TwtNNnIc

Weloganite may represent an ideal textbookillustration of twinning of crystals in the triclin-ic system. The twinning condition for tricliniccrystals requires that the following ratios mustapproach rational numbers (International Tabkslor X-Ray Crystallography, fr, 1959, p 106):

a2 a b2 : c2 : bccosot : cacxp : abcosy,

This condition is beautifully met for weloganiteas the required ratios are:

6.013 : 6.01.3 : tA X rc.Lt4 : 1.000 :2.003 : 3.O07.

In fact, all type I grains selected from the in-

8.988(16 .730 ( l

r02 .84 (1116.42s9. e9_fl

o. l 3ess( 2)f i- l0.12848(2)0. r 6592(4)

90 .01 ( l lo

X.RAY CRYSTALLOGRAPI{T OF WELOGANITE 25

m€r$ion mounts and fragments removed from thecoro material in the thin section were found to betwinned. The twin law .may be stated as [103]rzo"or [@1]*rzoo with twin obliqutty ctl:O and twinindex n=3. It falls in the class of twin bv twin-latticd symmetry (ILS) according to the;impli-fied clpssification of twins by Donnay & Donnay(1974), The complete twin symmetry is 3. Thetwinn€d cell corresponds exactly to the trigonalcell reported by Sabina et al. (1,968) with thesame type of peculiar extra-space-group syste-mafis extinctions wilh h-k=3n absent whenh*l and K+l are not equal to 3r. The 0-level[103] preces-ion photograph of a twinned crystal

Frc 3. D-precesrion photo€raphs (0-level, u, -30o, MoKa radiation) of weloganite (a) singlecrystal, (b) twinned crystal and (c) twinned crystaldisplaying diffuse streaks along c*.

is shown in Figure 2b which is to be comparedwith that of the untwinned crystal (Fig. 2a).

It is interesting to note that pronounced dif-fuse streaks along cE or the twin axis, [103],were observed on precession photographs ofsome twinned crystals (Fig. 3c). However, theywere not observed for all twinned crystals @ig.3b) and were never observed for untwinnedcrystals (Fig. 3a). Similar diffuse streaks wereobserved and described by Gait & Grice (1971)as "streaking" of reflections and were inter-preted as due to stacking disorder along the csdirection. Although stacking disorder could notbe completely ruled out, the development of thediffuse streaks may be adequately explained,without introducing stacking disorder, by twin-ning of individuals of submicroscopic thicknessrepeating along [103]. The adjacent twin indivi-duals, if small enough to diffract x-rays cohereat-ly, would act as antiphase domains and producediffuse streaks similar to the effect of stackingdisorder. The submicroscopic size of the twinindividuals is testified by the fact that manyfragments of weloganite crystals that showed notwin lamellae under polarizing microscope withhigh magnification (500x) were neverthelessfound by r-ray diffraction to be twinned.

C\nnarcar. Fonrrrwe

Recalculation of the chemical analysis ofweloganite reported by Gait & Grice (1971) on

26 THE CANADIAN MINERALOGIST

the basis of 6 oxygen atoms (excluding COz andHrO) per formula gave Naz.:il(o.orSrz.7eC4.r4o.*(COs)s.rrO0.se. 4.49FIIA or ideally NasSrZr(COs)r.4-5HrO. Several new analyses of the water con-tent by the Penfield method gave 6.78 and6.8OVo (Jambor, private communication, t974),6.9 and 6.9Vo (analyst: V. Boyko). The consis-tency of results of these independent analysessuggests that the water content (9.66Vo) re-ported by Gait & Grice is probably in error. Iftle new values are accepted, the ideal formulaof weloganite becomes NazSrZr(COs)u.3H2O.This formula is confirmed by results of a pre-liminary crystal structure analysis (Grice, privatecommunication, 1974). Assumns Z = 1, tledensity of weloganite calsulated from the idealformula using the volume of the triclinic cellis 3.208 g,/cms which compares favorably withthe measured values 3.22 (Sabina et al, 1968)and 3.20 g/cma (Gait & Grice 1971).

AcrNowreocEMENTs

We wish to thank Drs. J. L. Jambor and J.D. Grice for making available to us the singlecrystal r-ray diffraction photographs of weloga-

nite used in their studies and for their permissionto quote their unpublished materials. We wouldalso like to thank Mr. V. Boyko for his analysesof H0. This work is supported by researchgrant A5113, from the Natiohal Research Coun-cil of Canada (to G.Y.C.).

RnrsRENcss

ANDREwS, K. W. (1950): An x-ray examination ofa sample of pure calcite and of solid-solution ef-fects in some natural calcite. Mineral. Mae. 29,85.

DoNNey, G. & DoNNey, J. D. H. (L974): Classifi-cation of triperiodic twins. Can. Mineral. 12, 422-425.

DoNNal l. D, H, (Ed,) (1963): Crystal Data, De-terminative Tables, 2nd Edit. Amer. Crystall.Assoc.

GArr, R. I. & GHcs, J. D. (1971): Some observa-tions on welogariite. Can, Mineral, 10, 912(abstr.).

Sarnre, A. P., Jevson, I. L. & Pr,exr, A. G.(1968):Weloganite, a new strontium zirconiumcarbonate from Montreal Island, Catada. Can,Mineral,9, 468477.

Manuscript received August 1974, emended Septem-ber 1974.