Canodian Mineralogist VoL 22".pp. 681-688(1984)

THE CRYSTALSTRUCTURE OF TUNGSTTTE,WO3.H2O

JAN T. SZYMANSKI SciencesLaboratories, CANMET, Departmmt of Energ,, Mines & Resources, 555Booth Street,Ottawa, Ontario KIA ^Gt ANDREW C. ROBERTS GeologicalSuney of Canada,601 Booth Street,Ottawa, Ontorio KIA 1Eg

ABsrRAcr lite and other primary of . It has been found in small ilnounts at various localities The crystal structure of tungstite, WO,.grg, has been throughout the world, but apparently not in suffi- solved from single-crystal X-ray-diffractometer data col- cient quantity nor as crystals large enough for com- lected with MoKa radiation (248 unique reflections, 405 prehensivemineralogical study. Literature references observed),and has beenrefined to an R index of 4.3V0. to this mineral date back over 160years (Palacheer Tungstite is orthorhombiE,Pmnb (pnmo with bad), with ol. 194), including studies by (1908) a 5.249,b l0.7ll, c 5.133A, Z:4. Thestructure is charac- Walker and Kerr (1944), terized by distorted octahedral units of tungsten atoms co- & Young but only recently have crys- ordinated by five oxygen atoms and a water molecule: the tallographic unit-cell parameters,possible space- octahedrashare four cornersin the equatorialplane to form group and a fully indexed X-ray powder-diffraction sheets. The structure is very similal to that found in pattern been reported (Roberts l98l). Sahama MoOy2H2O (Krebs 1972).The sheetsare held together by (1981),in his review of secondarytungsten minerals, hydrogen-bondingbetween the water moleculeand the oxy- concluded: a mineral speciestungstite is gen "As still atom in the axial position in the adjacent layer. The inadequatelydescribed and needsto be studiedwith W-O bond lengths are 1.69(2) Iaxiall, 2 x 1.83(3), modern laboratory techniques." Since crystalline 2 x 1.93(3)lequatorial] afi2.34(2) A waterl. Tne t*iut, massesof tungstite from the Kootenay Belle published powder-pattern for tungstite is corrected and mine, indexed. Salmo, B.C. are available (National Mineral Collec- tion of Canada, SystematicReference Series, Geo- Keywords: tungstite, aquatungsten(Vl) oxide, crystal- logical Survey of Canada, specimennumber 16549), structure determination, corrected powder-pattern. the crystal-structure analysis of this somewhat enig- matic mineral was undertaken as part of a study of SOtvttr{aIRS the mineralsof that region. Prior to this study, the only crystallographic information available was that grislailins On 6tablit la structure de la tungstite, tungstite probably is orthorhombic on the basisof WO3.HrQ, partir d de donn€esde diffraction X obtinues its optical properties (Sahamal98l), sur monocristal but eventhe en rayonnementMoKa e48 riflexions uni- publishedpowder-pattern (PDF ques,405 observ€es)affinement jusqu'au 18-1418)was unin- r€sidu R de 4.390. dexed. La tungstite est orthorhombiqte lmnb (pnma avecbad), aveca 5.249,b l0.7ll, c 5.133A et Z = 4. La strucrure contient des octaddres d6form6s d'atomes de tungst&ne coordonn6spar cinq atomesd'oxyg&ne et une moldcule CovposrnoN d'eau; chaqueoctaCdre parlage quatressommets dans le plan dquatorial pour former des feuillets. La structure res- The mineral specimenfrom which the crystalsfor semblebeaucoup d celle du MoOr.!!1r6 (IGebs 1972).Les structure analysis were extracted consists of a fine- feuillets sont reliespar desliaisons hydrogbneentre la mol6- grained intergrowth of tungstite and ferritungstite. cule d'eau et l'atome d'oxygbne en position axiale dans le It is not possible to isolate from it enough pure feuillet adjacent. Les longueurs des liaisons W-O sonr material for a microprobe analysis, or for a bulk 1.69(2)[positionaxi+], 2x 1.83(3), 2x 1.93(3) [plan analysisfor HrO. However, our powder-diffraction equatoriall et2.34(2)A [positionaxiale, eau]. Le clichdde poudre de la tungstite est corrigd et index6. data fit syntheticWO3.H2O (PDF 18-1418),and the synthetic material has been extensively analyzed Mots-cbs: tungstite, oryde de tungstbne(Vl)hydrat6, d6ter- (Freedman1959, Freedman & Leber 1964).Natural mination de la structure cristalline, correction du cl! material from various sources,identifed as tungstite ch6 de poudre. by the powder pattern (Kerr & Young 1944),has also been analyzed and found to agreewith the formula INTRoDUCTIoN WOr.Hrg. In the latter paper, the poorest agree- ment was found for tungstite from the Salno mine _ Tungstite, WO3. H2O, is a secondary mineral, (analysisof Walker 1908),the major lmpurity being fbrmed as an oxidation product of , schee- formulated as FqO3 and FeO. This is consistent 681 682 THE CANADIAN MINERALOCIST

TABLE1. CRYSTAIDATA X-ray diffractometer, and precisecell-dimensions and orientation matrix were determined from the Tungstlte: f,Os.H20 source: KootenayBelle nlne' salmo, Brltlsh columbia. least-squaresrefinement of the observed%),x,6 and (Natlonal Mlneral collectlon of canada,systemtic ReferenceSerles, 0slogical Surveyof Canada, orvalues of 59 reflectionsin the range20' .20 < 34" SpeclnenN0. 16549). using MoKa radiation = 0.71073 A) (Busine Fomula l{elght: 249.86 O : 0rthorhonblc, 7 - 4. 1970).At this stage,although it was apparentthat Systemtic absences:L&2, Oenerally present' the spacegroup was ot(Z' generally Present' the crystalsare orthorhombic, h0!'h+L"?n+f still uncertain. t&0, k _ 24+7 spacegroup: P@b (062, Poa cllh bdi). A positive octant of reciprocal spacewas defined celf dlmenslons:a = 5,249(2),b = 10.711(5)'a = 5.133(2)A. spots was best, Denslty: 0""]"-5.75 l,lgm-3, Dobsrct neasured,but Ualker (1908) where the shapeof the diffraction glves 5.517 for tungstlte fron the sane locallty, the lou and one octant of data was collected three times to value probably causedby lntergrorth with ferrltungstlte, of 20 = 72o using monochromated (Dobs=5.02 l,,lgn-3). a maximum radiation. All reflectionswithin this 20limil Absorptlon: u(Moeo)- 406.9 m-l MoKa Intenslty Data, one octant of, data collsted with l4ofdradlatlon were measured without reference to systematic to 20-72"t neasuredthree tlnes and averaged: o/min 748unlque reflectlons,405 wlth I > 1.65xo(I). absences.A scanrate of I was usedwith a base peak-widthof 2.1" , increasingwith dispersion(0.69 tand). Background counts were measuredfor 80 with the above descriptionof the mineralogy, and secondson either side of the peak. Three standard with our premiseabout the impossibility of doing reflections were monitored in order to check crystal precise analysis on material from the Salmo mine alignment and instrument stability. No systematic becauseof the intergrowth with ferritungstite. Thus variation was observed, and the minor variations it seemed necessary and justifiable, initially, to were ignored. Absorption corrections were applied assumea formula WOr,Hrg for the single crystals using a Gaussian integration procedure (Gabe & of the natural material, and to check the final O'Byrne 1970)and a erid of l0 x 10 x 10points. The difference-synthesisfor evidence of compositional correctionfactors varied between 1.85 and 7.05.The differences. No significant peaks were found at the three segmentsof data were averaged and reduced end of the refinement; though this doesnot preclude to structure amplitudes with the application of possibleminor substitution for tungsten, no such Lorentz and polarization corrections. Reflections substitution has been reported in the literature. were consideredas "observed" on the criterion that Iobs> 1.65 x o(I), where o(I) is obtained from ExPERIMENTAL counting statistics.

Numerous crystalline fragments of tungstite were SvsrsMAtrc AsspNcss AND SPAcEGRoUP extractedfrom mineral specimen16549. Of these, only two very small crystalsshowed no grossevidence The lattice is primitive, with reflections hkl gen' of twinning, and both were used for extensive erally present; reflections Okl are generally present, preliminary precession-photo$aphy.Exposures of although reflections not obeying the/-centring con- up to 200 hours were used in order to try to observe dttion (hkl, k + I = 2n'1are very weak. Reflections the weak reflections from thesevery small crystals' h\l were observedonly for h = 2n and I : 2n, with and henceto differentiate weak reflections from sys- the exceptionof the (102)reflection, which appeared tematically absentreflections, so €tsto determinethe to be presentbut weak, and reflections (104) and possiblespace-group. (10O, which were at the extreme limit of observa- It was apparent from the preliminary work that bility on the 1.65 o(I) criterion. Reflections hlfi ate crystals of tungstite belong to the orthorhombic sys- present for k = 2n. The above absencesindicate tem, with the dimensionsquoted in Table 1. The spacegloup Pmcb or P2rb. However, the Patterson larger of the two crystals is a flat hexagonal tablet, function and the vector map (E2-1) could not be whose faceswere indexed from the precessionwork. interpreted on the basis of thesespace groups. The crystalis 53 pm between(100) and (100),61 pm Finally, a CAD4 diffractometer was used to col- between(101) and (T01),66 pm between(101) and lect a new data-set (MoKcr radiation, hemisphereof (101)and about 9 pm between(010) and (010).The data with k positive, all reflections collected to spotsobtained on the precessionphotographs from 20 - 60'). It was found that none ofthe four reflec- this crystal show minor splitting in certain regions tions of the {102}, {lM}, {106} forms wereobserved of reciprocalspace and somedisplaced "tails", but in this seconddata-set. In the latter, the "FLAT" as the other crystal is no better in terms of shapeand option in the CAD4 data-collectionroutine was used. singularity of diffraction spots, and in addition is The crystal was automatically rotated about the scat- lessthan half the sizeof this crystal, structural work tering vector (ry')to the position of minimum absorp- proceededwith this less-than-satisfactoryspecimen. tion for each reflection. Consequently, the diffrac- The crystal was mounted on a Picker four-circle tion geometry was totally different from that on the THE CRYSTAL STRUCTURE OF TUNGSTITE 68.3=

TABLE 2. ATOMICPOSITIONAL AND THERMALPARAMETERS AtomPosltion s urr uzz Uas utz Urs Uzg w 4c -0.0037( I/4 0.22090(8) 3) 0.72(3) 1.54(3) 0.83(3) o.o 0.0 -0.28(s) o(1) 4e I/4 0.436(2) 0.075(4) 3.4(1.2)2.4(1.0) 1.5(0.8) o.o 0.0 -0.2(8) 0(2) 4e L/4 0.066(2) -0.064(4) 1.7(0.8)2.0(0.8) r.5(0.8) 0.0 0.0 0.2(6) 0(3) 8d 0.4e5(8) 0.227(2) 0.24e(5) 6.1(r.2)2.0(o.e) 7.4(1.3)-0.6(1.6) -6.0(1.1)-0.6(1.3)

H(1) 4o L/4 0.479 0.200 3.3 3.3 J.J H(2)4c L/4 0.478 -0.055 3.3 3.3 ??

The anisotropic temperature-factorsare expressedin the fonn: = T exp[-2n2(Ul1a*2[2..+2Urza*b*hk..)1, and the va]uesquoted are xl00.

Picker diffractometer. Re-examinationof the (102) Examination of the environment of the water reflection with rotation about the scattering vector moleculeshowed it to be within 3 A of two neigh- produced a minor peak in certain orientations, boring O(2) oxygen atoms, and the geometry of this indicating it to be a Renningerreflection. trio was such that a hydrogen-bondingscheme could The structurewas treatedas triclinic, and the struc- be proposed. Using an idqalizedwater moleculervith ture determinationproceeded in spacegroupl'l, then an O-H distanceof 0.8 A and an H-O-H angleof Pl until the symmetry became apparent, and the ll0o, the proposedpositions ofhydrogen atoms were spacegroup was identified asPmnb (#62,pnmovith examinedin a difference synthesis.Two peaks were ba?). This requiresthat reflectionsftOl be observed found clearly correspondingto thesepositions, but for h + l:2n, whereasall h\l reflectionswith odd refinement of thesewas not found possiblein the = indices(h 2n * l, I = 2n + l) weretoo weakto be least-squaresprocess. The hydrogen atoms were observed.Furthermore, the (102), (104) and (106) included in the latter stagesof refinement, with their are systematically absent in this space gxoup, and calculated positions and isotropic temperature- indeed they were not observed in the CAD4 data- factors averagedfrom O(2) and O(1). The Residual set. Evidently, the differencesbetween the two geom- Index improved marginally, tlough significant posi- etries of diffraction were sufficient to eliminate these tional shifts were observedin the water orygen O(1). Renninger reflections. However, as the picker data- ths idsalizedpositions of the hydrogenatoms were set had been corrected for absorption, the above then re-calculated,and the final cycle of anisotropic three offending reflections were removed, and the refinementyielded an overall Residual of 4.30/0. Picker data-setwas usedin the structure determina- No attempt was made at the end to releasethe con- tion and refinement. straints of the mirror plane in spacegroup Pmnb, i.e., to decreasethe symmetry to space group SoLUTToNoF THE Stnucrunr F2pb. Duringan earlier stage of the refinement in spacegroup Pl, atoms W, O(l) and O(2) had all As explainedabove, the structure was solvedby refined to within one standarddeviation of x = V+, Fourier and Patterson methodsin the triclinic space- and O(3) and its subsequentlymirror-related O(3,) gfoup Pl, and, later, PI. In the latter space-group, had refined to correspondingco-ordinates xx.Va, thq spatial relationship between the four atoms ;f again to within one standard deviation. tungstenwas identified*, and the mirror planesat In the structure-factor and least-squares r/+,3A calcula- x: becameevident. Finally, the spatial sym- tions, the atomic scatteringfactors usedwere those metry of the elementsnot on the mirror planeswas of the neutral speciesW, O, and were calculated identified, and the spacegroup was fixed. The tung- from the coefficients given by Cromer & Mann sten atom was found to be octahedrally co-ordinated (1968).The scatterinefactor for hydrogen was taken by five orygen atoms and one water^moleculelidenti- from Stewart et al, (1965). The anomalous scatter- fied by the long axial bond, 234 A, es opposedto ing.contributions for W and O were taken from the short axial bond, W-O(2) : 1.69A, by analogy Cromer & Liberman (1970).All the crystallographic to the structure of MoO:.2HzO (freUs 1972)1. calculationswere carried out using the XRAY-26 programs (Stewart *The systemof et al. 97A, relationshipbetween an equivalentposition,.r, The final fractional co-ordinates and thermal = /, Z for x % and an equivalentposition with x : 3A parametersare listed in Table 2. The observedand is not unambiguous.The result can be achievedyja calculated structure-factors €ue given in Table 3, Vz+ x or l-x, and theseare not equivalent.gimilarly, available at a nominal charge from the Depository - with e 0, confusion occursbetween e and -2, and of Unpublished Data, CISTI, National Research t/z t/z-2. between + Z and. Council of Canada, Ottawa, Ontario KIA 0S2. 684 T}IE CANADIAN MINERALOGIST

Frc. 2. The co-ordination octahedron of oxygen atoms around tungsten.Atoms W, O(l), O(2), H(l) and H(2) lie on the mirror planesat x : %' %. The thermal ellip- soids are drawn at 5090 probability (Johnson 1965).

TABLE4. BOIIDLENG'T1{5 AIID ANGLES I{Ifi STANDARDDEVIANOilS

1. Coordlnatlonof l: Atom d(h o(2) 0(3) o(3)1 0(3)5 0(3)6 o(1) 2.34(2)\7e.3(e ) 80.e(7) ao.9(7) 80.1(7) 80.1(7)' 0(2) 1.6e(2) 99.6(7) ee.6(7) ee.4(7) e9.4(7) 0(3) 1.83(3) 98.4(1.s)88.2(1.5) i61.0(8) 0(3)r 1.83(3) i61.0(8)88.2(1.s) 0(3)51.e3(3) 88.1(1.5) 0(3)6 1.93(3) = Frc. l. The unit cell of tungstite.The atomsof the asym- 2. Coordlnatlonof 0(1), (llater): metricunit arelabeled. The doubled circles for tlte O(3) o(r)- x 2.33s(23)l r,r - o(1)...0(2)r2e.o(1.0)" : gonding adja- o(1)...0(2)2.87(31 l{ - 0(1)...0(2) 108.1(e) atomsnear r % indicatethat ths from 0(r)...0(2) 2.e7(3) o(2),..0(1)...(2) 122.e(1.0) centatoms of tungstenis directedto two O(3) atoms separatedbye=1. 3. Coordlnatlonof 0(2): o(2)- r r.os8(zs)i l,t - o(2)...0(1)7 108.s(1.0)' rd - o(2)...0(1)8 128.5(i.1) o(1)7...0(2)...0(1lE 122.s(s) DEScRIPTIoNoF TI{E STRUcTURE 3. Coordlnatlonof 0(3): - o(3)-lr r.ezg(ss)i $ - o(3) r 165.2(1.0)" The unit cell of tungstite is illustrated in Figure 0(3) - 14 1.925(34) sheetsof W 1, which clearly shows the two distinct 4.-' unreflned hydrcgen-bonddlstances and angles'o and O(3) atoms near/ : Vr, 3A,Figure 2 showsthe (Basedon liealized watet mlecule, o-H ; 0'8I, H-0-H'110"): distorted octahedronof oxygenatoms that surrounds Donor- H. . . . .Acceptor d(H..0)i Angle at H follows: o(1)- H(1).....0(2) z.oe 171' the W atom. The octahedra are distorted as 0(1) - H(2).....0(2) 2.17 172 one axial^bond, to O(2), is significantly sho{er Al than the other axial bond Al, Superscrlpts used abov€, in the text and the dlagrar6 refer t0 11.69(2\ 12.34Q) the follodng equivalent positlons:- water oxygen O(l). The equatorlal which is to the 1. l-o , "!, .a bondsto O(3) are 2 x 1.83(3)and 2 x 1.93(3)A, 2. e , |+u , t-a tta ,'?'8 and thesefour are displacedaway from O(2), such 4. 1-r , ,-s , ,ta that the O(2)-W-O(3) anglesare all closeto l@". TT{E CRYSTAL STRUCTURE OF TUNGSTITE 685

tA, Frc. 3. (a), O). The W-O corner-sharing octahedra stacked in the layers near / = %, respeciively. The hydrogen atoms :ue located at the extremities of the tapered bonds at the axial vertex of each octahedron, In the structure, each sheetoverlaps the otier, so that a three-dimensional checkerboard network of octahedra and holes is formed.

Full details of bond lengths and angles are given in Leber are noted, and one line that should have been Table 4. The tungsten-oxygen octahedra, as indi observed(d L7n,I 5/lm) hasbeen included. Other cated above, are stackedin sheets,and shareall four than thesediscrepancies, the agreementbetween Io* corners in the equatorial plane, tlus producing a and IG1"and between do6.and d*1" is excellent. The checkerboardpattern of octahedra and holes in the indexed powder-pattern previously grvenby Roberts .rq plane (Fig. 3). The checkerboardlayers above and (1981)shows significantly poorer agreement,espe- below the first are displacedby D/2, so that the octa- hedra lie over holes. (Using the checkerboard analogy, black overlays white, and vice versa, in TABLE5. POTDERPATTERN (CuX6) FORTUNGSTITE three dimensions). &L dctlc dob" The sheets are held together by a network of 020 5.355 5.39 113 1.608 1.611 6 hydrogenbonds, such that the water moleculeO(l), ltt 4.629 4.66 22 251 1.579 l'583 5 H(l), H(2), which is in an axial position in a given 720 3.749 3.77 45 033 1.543 I Lt7 3.472 3.49 100 100 242 1.514 1.515 octahedron, bonds two axial O(2) atoms from octa- 031 2.931 2.95 56 331 1.502 1.504 040 2.67A 2.687 109 133 1.480 1.482 hedra in the adjacent layer. Each O(2) atom is 200 2.625 2.626 12 11 260 r.476 1 hydrogen-bondedto two O(l) oxygenatoms, sepa- 002 2.5661' Er^ ta^- "8 071 1.466I (2 t31 2.559'"-'," rla 062 1.465i 1.470 ? {l rated by z : l. Thus, a zig-zag pattern of hydrogen 140 2.385 2.387 45 340 1.465t ri bonds is formed, extendingin the direction, as illus- 220 2.357 2,3@ 86 162 I .411 1.414 e 022 2.314 2,322 oo 322 L.396 I trated in Figure 4. zL't 2.283 -1 nEllt?7tltt I tt2 2.254 2.260 2t 233 1.330 I ; 722 2.178 2.122 400 1.312," -.^ 2 t2 Pownen PetrEnN oF TuNGsrrrE 051 1 .977 1.979 45 351 l.3lo"''rt '2 231 1 .955 1.959 180 1.297 1.298 2 ?40 7.874 L.875 42 raJ r.zva 1 004 1.283 powder-pattern given ru'^ 17,2 ) (1 A revised for tungstite is in ffi l'!!& 1.854 271 1.280l.1.282 2 Table 5. This showsthe powder-diffraction intensi- 202 1.835 1.837 14 11 262 1.280' ll 060 1 .785 1.788 21, 420 L.275 1.276 ,_ 2 ties, calculated on the basis of the absorption- !42 L.747 342 7.272 2 cingle-crystal 222 7.736 7.739 10 72 360 1.250-" t'ttu --^ ,1 corrected intensitis, using the program 160 1.690 1.693 024 l.z4gt r2 POWGEN (HaU & Sarmafski 1975). For com- 320 1.663 -1 313 1.215 1.211 311 1.637 1.637 n l? parison, the observed X-ray powder-diffraction intensities and d valuesof Freedman& Leber (1964) I" rnd d"al" data fron PolcEN (Hall & szymiskl 1975), poyder- pattern generatlng progran basedon slnglearystal lntensltles. (PDF 18-1418)are included. Two gpuriouslines and Io dob" data frcn Freedmn& Leber(1964), (PDFN0.18-1418). "nd a thfud marked "@eta?)" observedby Freedman & Reflectlons at d-3.86, 3,29, 3.17 ln PDF18-1418 are not genuine. 686 THE CANADIAN MINERALOGIST

Ftc. 4. The layer-like structure of tungstite, illustrating the interlayer hydrogen bond- ing. Each octahedrallyco-ordinated tungsten is bonded to four O(3) oxygenatoms, forming W-O(3) sheetsin the x-e plane. Two such sheetsare illustrated, near y = V andy = %, The W-O axial bonds are to O(2) and to the water molecule, O(l), H(l), H(2). Each water moleculeis hydrogen-bondedto two O(2) atoms, and each O(2) bonds two water moleculesseparated by z= l. cially with regard to intensities. It must be presumed pattern$ of MoO3.2HrO and WO3.2HrO show these that the intensity valuesof Roberts (1981),visually compoundsto be isostructural,yet the powder pat- estimatedfrom Debye-Scherrerphotographs made terns of the monohydratesare different. This has not from very small quantities of specimenmaterial, are preventedthe disseminationof someconfu$ion, e.g., not reliablq. The indexed and correctedvalues of "The yellow modification (of MoO3.HrO) has been PDF 18-1418should now be usedas the standard mentioned to be isotypic with WO3.H2O (Freedman powder-datafor tungstite. 1959),the crystal structure of which is, however, unknown aswell" (Giinter 1972).Thecrystal struc- DIScUSSIoN ture describedin this paper fills half this gap, and the definitive powder-data should supplant the unin- The solution of the presentstructure fills an impor- dexed powder-patternfor tungstite in the Powder tant gap in the field of tungstenand molybdenum Diffraction File (No. l8-1418). oxide hydrates. The crystal structure of The structure presentedhere is very closelyrelated MoO3.2HrO has been examined (Lindqvist 1950, to that of MoO3.2H2O (Krebs 1972). The latter is 1956)and redeterrnined(Itebs 1972).The powder monoclinic, P21/n, and characterized by THE CRYSTAL STRUCTURE OF TUNGSTITE 687

MoO3.IlrQ octahedra, corner-sharing to form betweenthese edges." Owing to the fact that the a sheets,exactly analogous to the sheetsfound in the and c dimensionsare very similar, the 4ngle between present structure. The additional molecule of water the facesof the {l0l} form is closeto a right angle in MoOr.2HrO fills one of the holes, and is utilized (88.72'). The bisectorsof theseangles are the axial in further hydrogen-bondingofthe octahedra.The directions. octahedra found in the two structure$are very com- Finally, a word about the role of the hydrate parable. In MoO3.2HrO, there are four different moleculeis warranted. The earlier literature on tung- Mo atoms, all with very similar geometry. The mean sten and molybdenum oxide hydratescontains much bond-lengthsare as follows, with the presentW-O discussionon the role of water in thesecompounds: distancesin squarebrackets for comparison:Mo-O are thesecompounds true hydrates containing crys- (axial): 1.694U.691;Mo-O (water,axial) =2.288 talline water? Are they acids (e.9., H2WO4, [2.34]; Mo-O (equatorial, shorter): 1.766, 1.800 H4WO5X At one stage,an Hod* ion was proposed [l.83]; JvIo-O (equatorial,longer) = 2.054,2.L56 to exist in MoO3.2H2O (Lindqvist 1950), though tl.93l A. As MoOr.2HrO appearsto be isostruc- this was later refuted (Lindqvist 1950. It is evident tural with hydrotungstite WO3.2H2Ofrom the very from the detailed structure-analysisof MoO3.2H2O close correspondenceof their powder patterns (IGebs 1972)and from the presentwork, that these @reedman& Leber l9@), i:t seemslogical to pre- compoundscontain water molecules,co-ordinated sume that the same type of metal co-ordination is orygen-to-metalin their crystal structuresand should found in the latter structure. be considered as aquametal(Vl) oxides. Fur- The only tungsten oxide hydrate on which struc- thermore, the molybdenum trioxide dihydrate con- tural work has been done is WO3.%H2Q(Gerand el tains both an aqua- water.(bondedto metal) and a al. l98l).Its structure has beendetermined from the hydrate water (not bonded to metal). It is strictly intensities obtained from powder diffraction. The aquamolybdenum(Vl) oxide hydrate. This nomen- rather rough geometryobtained from the refinement clature applied to the tungstencompound [aquatung- shows that the W-O corner-sharing octahedron is sten(VD oxidel is in agreementwith the proton mag- retained as the basic rrnit of the structure, but the netic resonanceand infrared-absorption spectra arrangementofthe octahedrais totally different; the (Mari6ie & Smith 1958, Schwaranann& Glemser voids in the sheetsof this structure are either trian- 1961). The earlier nomenclatureapplied to these eular (three octahedra)or hexagonal(six octahedra). compounds("tungstic acids", "molybdic acids") is The stackingofthe individual layersupon eachother definitely incorrect, and should not be used in a camot follow the simplecheckerboard pattern found chemical description. in tungstite. The structure determination of tungstite accounts physical properties well for the and optical of the RsFEREr,lcss crystallinsmaterial. Walker (1933),later quoted in ql. (194), gives Palache et the indices of refraction BusrNc,W.R. (1970):Least-squares refinement of lat- as 2.09 perpendicularto the cleavagedirection and tice and orientationparameters for usein automatic 2.24 and.2.?5(aI t 0.02) in the other two directions. diffractometry. "fu Crystallographic Computing The perfect (Walker 1908)is the crystallo- (F.R. Ahmed, ed.). Munksgaard,Copenhagen. graphic {010}, and occurswhere the layersof W-O octahedra,which are hydrogen-bondedin the ), direc- Cnorrann,D.T. & I-rseRMAN,D. (1970):Relativistic cal- tion, are separated.Within thesesheets, the struc- culation of anomalousscattering factors for X rays. tural packing is very comparablein the x and z direc- J. Chem.Phys. 53,1891-1898. tions, so that one can exp€c1the indices of refraction (1968);X-ray factors be & MANN,J.B. scattering to about the same.Within experimentalerror, this computedfrom numericalHartree-Fock wave func- is given exactly what is found. The other cleavage tiors. Acta Cryst.AA,32l-325. by Walker (1933)(reindexed on presentaxes) is {101} imFerfect. This can be explained in terms of a shear FnrBor4aN,M.L. (1959):The tungsticacids. J. Amer. acrossthe plane of O(3) atomsa't the cornersof the Chem, Soc.t1, 3834-3839. octahedra. However, as the shear plane in a given layer is displacedby a/4 + c/4 fromthat in the adja- & Lsspn,S. (1964):Identification of tungstic cent layer (owing to the checkerboard stacking of and molybdic "acids'f by X-ray powderdiffraction. octahedra), the cleavagewill be imperfect. We have J. Less-CommonMetak 7,427432. observed to {l0l} be one of the forias in the Genr,E.J. & O'BvnNs,T. (1970):An absorptioncor- hexagonal-shaped crystal used, the other being rection progxam for the PDP-8. Amer. Cryst. {100}. Walker (1933) states that "many of the Assoc. SummerMeet. Abstr, L4. (translatedinto cleavageplates show edges at right angles to one FORTRAN and adapted to CDC computers by another, with extinction directionsbisecting the an8le Gabe & Srymarlski). 633 THE CANADIAN MINERALOGIST

GSRAND,B., Nowocnocrr,G. & Frounz, M. (1981):A MARrblC,S. & Spnnr,J.A.S. (1958):A nnoher,'nagnqtic newtungsten trioxide hydrate,WOr.%11r9' Or.O- resonancestudy of the hydratesof molybdenumtri- aration,characterization, and crystallographicstudy. oxide. "/. Chem. Soc.,886-891. J. SolidState Chem.3t, 312-320. Per-acss,C., BnnueN,H. & FnoNoBt, C. (1944):The GUNrsR,J.R. (1972): Topotactic dehydration of Systemof Mineralogy. I. Elements,Suffides, Sul' molybdenumtrioxide-hydrates. J. Solid StateChem. fosahs, Ortdes. John Wiley & Sons, New York. 5, 354-359. Rossnrs,A.C. (1981):The X-ray crystallographyof Han, S.R. & Szvuarsn,J.T. (1975):Powder pattern tungstite.Geol. Sum. Can. PaperE1-1C' 82. generationfrom single-crystaldata. Amer. Cryst. Assoc, WinterMeet, A bstr. F13. Now incorporated Seneue, T.G. (1981):The secondarytungsten into the XRAY-76 systemof progftrms(Stewart e/ minerals,a review.Mineral. Rec.12, 8l-87. al. D7A. ScnwAnzrrrarw,E. & Gr.eMsBn,O. (1961):Zur Bindnng Jonrsor, C.K. (1965):ORTEP: A FORTRAN desWassers in denHydraten des Wolframtrioxids. thermal-ellipsoidplot program for crystalstructure Z. Anorg. allgetn.Chem. 312' 4549. illustrations. Rep. ORNL-3794, 2nd. Revision ORTEP-II additionl97l, OakRidge Nat.Lab., Oak Srrwanr, J.M., Mecutx, P.A., DrcrnsoN, C.W., Ridge,Tennessee. The Johnsonprogram has been Anauou,H.L., HBcr, H. & Fncr, H. (1976):The modified for usein the StewartXRAY-76 System X-RAY Systemof CrystallographicPrograms. Univ. by J.F. Gu6don,S.R. Hdl, P. Richardand S. Whit- Maryland Comp. Sci.Ctr. Tech.Rep. TR446, low, 1974. SrswARr,R.F., Davtnsorv,E.R. & SurapsoN,W.T. KnRR,P.F. & Yowc, F. (1944):Hydrotungstite, a new (1965):Coherent X-ray scatteringfor thehydrogen mineralfrom Oruro, Bolivia. Amer. Mineral.29, atom in the hydrogenmolecule. J. Chem.Phys. A' 192-210. 3175-3187. Knrss, B. (1972\: Die Kristallstruktur von Welren, T.L. (1908):A reviewof the mineralstung- MoO3.2ffrQ.Acta Cryst. E.28,2222-2231. stite and meymacite.Amer, J. Sci. 175' 305-308.

LrNoqvrsr,I. (1950):The crystalstructure of the yel- - (1933):Notes on tungstite.Univ. TorontoStud., low molybdicacid, MoO3.2H2O.On the existence Geol. Ser.35.13-14. of an HaO2+ion. Acta Chem.Scand. 4, 650-657.

(1956):A correctionof the crystalstructure of Received November 23, 1983, revised manwcript MoOr.211rg.Acta Chem.Scand. 10, 1362-13&. acceptedJanuary 20, 1984.