Structure Refinement of Zirkelite from Kaiserstuhl, West Germany

American Mineralogist, Volume 67, pages 615420, 1982

Structure refinementof zirkelite from Kaiserstuhl, West Germany

Wllrrllra SrNcrnrn Research School of Earth Sciences

eNo Rrcueno A. EccrEroN

Department of Geologl The Austrqliqn National University P.O. Box 4, Canberra, A.C.T. 2600, Australia

Abstract

The crystal structure of zirkelite from Kaiserstuhl, West Germany has been refined by the full-matrix least-squaresmethod using 949 independentreflections to a final R value of 0.052.The data were collected on a four-circle diffractometerusing graphite-monochromat- ed MoKa radiation. The structure is monoclinic, spacegrottp C2lc with cell dimensionsa : 12.431(l),b :7.224(l), c : 11.483(3)4,P: 100.33(lf andZ:8. The resultsfrom this study show that the crystal lattice has sufferedonly minor damagedue to an accumulated radiation dose of 1 2 x l0ts alphasper gram from the decay of natural uranium and thorium contained within the mineral. The most significantchange is the increase(2-3 times) in the temperature factors. The high values of these parametersare interpreted to be the direct result of the action of alpha-recoil.

Introduction million years, during which time a large number of atomic displacementsoccur in the host lattice. Pyatenko and Pudovkina (1964) first proposed Natural zirkelites from the Kaiserstuhl carbona- that zirkelite (:zirconollte), CaZtTi2OT,was close- tite W. Germany, contain -5 weight per- ly related to the fluorite structure. They established complex cent UO2 + ThO2. Oversby and Ringwood (1981) that it had a superstructure based on a C-centered have calculated an accumulated radiation dose of monoclinic lattice which is eight times the fluorite - - 1.2 x 1018alphas per gram for theseminerals, using subcell such that d"rperst*.t re -a1 qz * 2a3,bt an age determined by the K-Ar method (Wimmen- : ar-a2,c":312a1 * 312a2* ca,where a1, a2, a3 auer, 1966) 16 million years. refer to the cubic cell. Recently, the crystal struc- of In this report, a single crystal X-ray analysis has ture has been solved and refined using both powder Kaiserstuhl zirkelite in (Rossell, 1980)and single crystal X-ray techniques been carried out on the alpha-recoil damage (Gatehouseet a1.,1981).The close relationshipto order to examine the effects of the lattice. The results from the single fluorite was confirmed by these experiments.In the on crystal refinement of a synthetic zirkelite (Gate- latter study, the site of the M4 titanium atom was crystal house et a1.,1981)are used as a standardto which found to be displacedoff specialposition 4(e) at (.5, the natural zirkelite data can be compared. y, .25) to the general position 8(fl on either side of the 2-fold axis; decreasingthe coordination from 6 Crystal structure analYsis to 5. Zirkelite comprises about 35 percent of sYNRoc, The grains of zirkelite used in this study were a synthetic rock designedfor the immobilization of separated from the Kaiserstuhl carbonatite (Wim- high level nuclear wastes (Ringwood et al., 1979a, menauer, 1966), and provided by Dr. J. Keller. b). The highly radioactive actinide elementswithin Many of the crystals exhibited polysynthetic twin- these wastes, are partitioned into the zirkelite and ning. Preliminary Weissenberg and precession X- perovskite phaseswhen hot-pressedat I 100-1200"C ray photographs showed that reflections (hkt) and with the syNRoc mixture. The actinides decay by a (ft0l) were absent for h + k : 2n and I : 2n series of alpha-decays at a significant rate for a respectively.These restrictionssuggested the pos- 0003-004)v82l0506-06I 5$02.00 615 616 SINCLAIR AND EGGLETON: ZIRKELITE

Table l. Crystal data for Kaiserstuhl zirkelite (20) andrange from 0.8"below the Ko1 to 0.8"above the Ka2 peak for the reflection concerned. Station- (Ca. s 5REE.o zltn. orTh. o aU.o z) (Zr. crTi. r s) ary background counts of 20 secondsduration were (Ti ' gsZr. r ,Nb. q sFe?TaFells)0t made at the extreme of the scan range. Three standardreflections (800),(040), (008), showed little

l,lonoclinic, C2lc (No. 15) variation in intensity throughout the course of the data collection. 1,710reflections having 20 values between 3 and 60' were collected from the octants = 12.431(1)8 9 = 11.as3(3)R hH and,hH except those which were absent for the b = 7.224(1)8 6 = 100.33(1)o C centering condition. The data were sorted and v = 1014.5(5)8' L =6 averagedand reduced to structure amplitudesin the = Dc 4.9Mg m-3 p(MoKc) = 9.6m-t usual way. Data reduction proceduresare described by Fergusonet al. (1979).Because of the small size = FR 0.29 (effective pR for spherical crystal of the crystal (p.R : 0.29) an absorption correction with same volme as crysEal used; was not applied. A partial compensation was R = radius of spherical crystal) achieved by averaging the intensities of equivalent reflections, (0ft1)and (0ft1),in the monoclinic cell. groups with sible space Cc or CZlc in agreement (Interscale R factor for equivalent reflections : Gatehouse (1981). et al. 0.019). After discarding reflections for which 1/o(4 For the structure determination a crystal measur- < 3.0, a total of 949 were available for subsequent ing approximately x x 0.05 0.03 0.08 mm was structural analysis. The statistical discrepancy val- mounted on a Picker FAcs-l four-circle diffractom- ue for this data set R" (: IoJFo)DlFol, where y parallel eter with the unique axis approximately to o"(Fo) : o(I) (Lp-rl2lFol is the error contribution the parameters @axis of the diffractometer. Lattice to lFol from counting statistics alone) is 0.026. and crystal orientation matrix were obtained by least-squares refinement of the setting angles of twelve carefully centered reflections having 20 val- Structure solution and refinement ues between 49 and 60o,using MoKal radiation ()t : Scattering factors for neutral atoms were taken 0.709264) reflected from a graphite-crystal mono- from International Tables for X-ray Crystallogra- chromator (20m : 12.12).Crystal data are given in phy (1974) and were corrected for both real and Table l. The chemicalanalysis is given in Table 2. imaginary anomalous components. Intensitieswere measuredby the 0 - 20 continu- The refinement was carried out using the full ous-scantechnique with a scan speedof 2" min-l matrix least-squaresprogram snls (Prewitt, 1966) which minimizes the function lw(lFol-ktFcl)z, Table 2. Chemical analysis of Kaiserstuhl zirkelite. Structural where k is an overall scalefactor and w is the weight formula based on 7 oxygens of an observation taken as unity in the initial stages. The atomic coordinates from the synthetic zirke- szos L5.7 0.45 lite refinement in the space group CZlc (Gatehouse Ti02 1.08 et al., 1981)were used in the starting model for ZtO2 34.8 108 Kaiserstuhl zirkelite. Initially, Ca, Th and U were Tho2 4.r .06 assignedto the Ml site,Zr to the M2 site, Ti, Nb to uoz 1.4 .02 the M3 and M5 sites and Fe and Ti to the M4 site. Fe0 2.28 .12 The other minor cations were ignored. After several Fero, 5.32 cycles of refinement and converting to anistropic temperature factors of the form, Ifrr0 o.2 .01 Ca0 12.5 .6) exp l- 2# (U nhza*2 + Urrl? b*2 + u rrPc*2 REE 0.9 .02 IZU phka*b* I2U shla* c* *2Uzzklb*c*)l suM 99.90 3.94 for all atoms, the conventional R factor decreased to 0.063. The equivalent isotropic temperaturefac- ElecEron microprobe analyses by N.G.Ware; tors for individual atoms at convergencewere about Iron deterninations by E. Kiss. 2-3 times as large as those reported by Gatehouseer SINCLAIR AND EGGLETON: ZIRKELITE 617 cr. (19E1).This behaviorpersisted in all subsequent were restricted to occupy the M3 * M4 sites, Zt the refinements. M2, M3 and M4 sites (but mostly the M2 and M3 The significanceof splitting the M4 site into two sites) and Ti all cation sites except for the Ml site. individual half atoms which occupy equivalent sites Refinement of the Zr population parameter in the on either side of the two-fold axis as reported by M2 site resulted in an immediate decrease, and Gatehouseet aI. (1981),was checkedby calculating suggestedthe presenceof the lighter element,Ti. In a difference Fourier map without the M4 atoms. keeping with previous single crystal results, the This showed a large region of electron density (24 e excess Zr from the M2 site was transferred to the A-3) centeredon 4 (e) at (0.5,y,0.25) where y : M3 site. Other refinementsindicated that the heavy 0.14. Peakson either sideof the two-fold axis were elements Zr + Nb (40 + 4l electrons respectively) not found. This is not surprising, since the expected did not occupy the M4 site, but were found to favor separationof the M4 sites-0.5-0.84 is closeto the the M3 and M5 sites. resolutionof our data set (0.43A).Alternatively, the Refinementof the Nb population parameterin the M4 atoms may occupy a range of positions. The M5 site was carried out; at the same time adjust- structure was therefore refined with M4 placed on ments were made to the amount of Ti in M5 and Nb specialposition 4(e). However, at convergencethe and Ti in M3 to maintain both the total occupancy R factor remained at 0.088, a value considerably of these sitesto be 1.0 (or 0.5 in the caseof the M4 higher than that obtained for the split site model. site) and the unit-cell composition. The population The M4 position for this model was (0.5, 0.121(2), parameter of Fe in M4 was then released and the 0.25) and the equivalent isotropic thermal parame- corresponding changesto Ti and Fe were made in ter B, was 8.142. Thus, the split site model was the M4 and M3 sites. During these final cycles it readopted and manual adjustmentand refinementof became necessary to refine separately the z posi- the population parametersfollowed. tional parameter and the beta (1,3) value of M4 At first, the placement of elementswas guided by becauseof a strong correlation. The errors quoted the previous results of synthetic zirkelite and by for these two parameters in Table 3, were derived crystal chemical reasoning. Thus, Fe2+ + Fe3+ from the inverted full-matrix at the end of each

Table 3. Atomic coordinatesand thermal parametersfor Kaiserstuhl zirkelite, with esd's in parentheses

Atm Site Occupancy xlA YIB reg. (92) M(1) 8(f) 0.87 Ca+ O.13(Th,U,REE) 0 .3752(1) 0.r24s(3) 0.497r ( r) L.78 M(2) 8(f) 0.85Zr + 0.15Ti 0.12r0(1) o.126r(2) -0.0239(l) 1.36 M(3) 8(f) 0.125Fe + 0.375Ti + 0.35Nb + O.I5 Zt 0.2500(1) o.r2s2(3) 0.74s0 ( r) 1.50 11(4) 8(f) 0.33Fe + 0.17Ii 0.4788(8) 0.0892 (6) 0. 2s09(5) M(5) 4 (e) 0,72 Ti + 0.28 Nb 0.0 o.1279(5) o.25 r.60 x(1) 8(f) 1.000 0.3087(7) 0.r243(r3) 0.2864 (6) 3. l0 8(f) 1.000 0. 4693 (6) 0.r373(r4) 0.089I ( 7) 2.75 x(3) 8(f) 1.00 0 0.2070(8) 0.08s7 ( r2) 0.s687(8) 2.84 x(4) 8(f) 1.000 0.3957 ( 8) 0.r66s(13) 0.71sr(7) x(s) 8(f) 1.000 0.712r(8) 0.1700(rr) 0.582r ( 7) 2.51 x(6) 8(f) r.00 0 -0.oo24(7) u. rr)/(1c, 0.4r89 (6) 2,42 x(7) 8(f) r.00 0 0.1086(8) o.06s3(13) o.79L7(7') 2.98 ull [22 U33 u12 u13 r23 11(1) 0.020( I ) 0.019(r) 0.029(1) 0.003( r ) 0.005(1) 0.004(1) M(2) 0.017 ( 1) 0.0r2(r) 0.023(1) 0.002(i) 0.004(r) 0.00r( 1) M(3) 0.019( l) 0.017(r) 0.02r(r) -0.00r ( 1) 0.00s(1) 0.000( 1) M(4) 0.056( 7) 0.062(3) o.oo9(r) -0. oo3( 3) - 0.002(3) -0.004(3) M(s) 0.020(r) 0.020(1) 0.020(r) 0.0 0.002(1) 0.0 x( 1) 0 .062( 6) 0.013(4) O.O42(4) -0.004 ( 6) 0.007 ( 4) -0.007 ( s) x( 2) 0 .028( 5) 0.024(s) 0.051(s) 0.002(6) 0.004(4) 0.002 ( 5) x(3) 0.030( s) 0.028(6) 0.048(s) -0.007 ( 4) -0 .001 (4) -0 .007 (4) x(4) 0.03s(6) 0.04r(7) 0.036(s) -0.0 10( 5) 0 .005(4) -0.006(4) x (5) 0.034(5) o.o27(6) o.o3r (4) 0.000(4 ) -0.004 ( 4) 0.00 7 (4) x(6) 0.031(4) 0.028(5) 0.03s(4) 0.00I (6) 0.010 ( 3) 0.005(4) x(7) 0.026(5) 0.043(6) 0.04r(s) -0.0 14( 5) -0.00r (4) 0.003(4) 618 SINCLAIR AND EGGLETON: ZIRKELITE

Table 4. Interatomicdistances (A) and o-M-O angles(') of 3. Bond lengths and angles are shown in Table 4. Kaiserstuhlzirkelite, with esd'sin parentheses Table 5 gives a listing of Structure faCtOrsl The exact location and occupancy of the cations M3 Ocrahedron Mr -x2 in the M3, M4, and M5 sitesare opento debate.The 2.326(8) u3 -x4 r.927(r0) x4 -xl 9S.9(4) final distribution was based mainly on crystallo- xzr 2.371(g) xl r .960(9) -xrr 82.3(4) chemical reasoning combined with minimization of x3 2.39r(9) xlr r.970(9) -x5 95.2(4> the R value. x3I 2.396(10) 1,97s(9) -x3 86.r(4) The eNucnys Structure Determination Package xl 2.412(7) x5 1.983(E) xl -x7 8b. r(4) (Whimp et al.,1977), as implementedon the Univac x5 2.4E6<9) x3 2.019(9) -x5 92.7(3> ll00l42 computer at the Australian National Uni- x4 2.490(8) UBAN r.972 -x3 8s.6(3) versity, was used throughout x6 2.532<9' xr1 -x7 92.5(4) the structure solution.

wN 2.426 -x5 84.1(3) Description of the structure -x3 97.6(3) u2 -x3 2.051(S) x7 -x5 80.r(4) The structure of Kaiserstuhl zirkelite (Fig. 1) is x6 2.065(8) l(3 98.E(4) similar to that reportedby Gatehouseet al. (1981). xs 2.107(8) }|4 Trigoffil blpyrstd Calcium (M1) is coordinatedby eightoxygen atoms x2 2.10E(r0) M4 -x2 1.s57(ro) x2 -xzl r52.s(6) lying at the corners of a distorted cube. The large x7 2.14r(9) r.873(10) -x4 115.0(5) cations. Th and U substitutefor Ca on this site. x5l 2.3r4(ro) x4 2.120(rI) -xl 88.3(4) Zirconium (M2) is surrounded by seven oxygen x6r 2.34o(s> xL 2,239(11) -x4l 8j.3(4) atoms at the corners of a truncated cube and

ms 2.161 x4r 2.494(L2, r21 -x4 9r.6(4) contains about 15 percent of substituted titanium.

EN 2.to -xl ros.4(5) Of the remaining 3 cation sites, M3 and M5 are

H4-M4 0.53(2) -x4r 103.1(4) surrounded by six oxygens forming distorted octa-

x4 -xt 72.0(4, hedra while M4 is best described as occurring in

-x41 69,0(5) trigonal bipyramidal (S-fold) coordination. Titani- xl -xtl 131.6(4) um, Fe, Nb and Zr occapy the M3 site, Fe and Ti M5 Octahedron the M4 site and Ti and Nb the M5 site. The two H5 -X7 r.94r(9)x2 L7 -x71 gE.o(6) types of octahedra (M3 and M5) join together at r.947(7>x2 -x6 xz 94,5(4) their vertices to form three and six memberedrings, L.965(g>x2 -x61 x2 sr,7(4) similar to the planesof octahedraparallel to {111}in UEAN 1,951 -x4 x2 95,1(3) the pyrochlore structure. The M4 metal site is x6 -x4 x2 87-6(4) situated within the six memberedring and occupies -xql *Z 96.3(4) a split position displaced towards the X4 oxygen x4 -x+r s1.8(6) atoms on one side of the cavity. These two equivalent M4 sites are separatedby a distanceof 0.53 (2)A. The metal atoms occur in planes parallel to (001) in zirkelite and alternately contain M3, M4 independent cycle and are consideredto be under- and M5 or Ml and M2 cation sites. They are derived estimated. The least-squaresrefinement of all posi- from the metal only (111) planes of the fluorite tional and anistropic thermal parameters gave a structure. terminal R factor of 0.052 and a weighted R* 1: llw(lFol-lFct)2l2wtFot2l05)of 0.035, where the Discussion 'logo'1'. weighting scheme was taken as u, : The The isotopesof uraniumand thorium (U-238,235, standard deviation of an observation of unit weight and Th-232) contained in natural zirkelite decay to was 3.24. form stable isotopes of lead by the emission of The highest peak on the final difference Fourier alpha- and beta-particlesand gamma-rays.Damage map was 2.2 e A-3 located near the M2 (Zr) position. A refinement in the spacegroup Cc result- I To receive a copy ed in an increased R value and retention of the 2- of Table 5 order Document AM-82-203 from the Business Office, Mineralogical Society of America, fold symmetry about the y axis. Final atomic coor- 20fr) Florida Ave., N.W. Washington,D.C. 10009.Please remit dinates and site occupanciesare presentedin Table $1.00 in advance for the microfiche. SINCLAIR AND EGGLETON: ZIRKELITE 619 to the crystal structure is causedpredominantly by the recoil of the radioactive nucleus associatedwith alpha-decay. The recoil nucleus collides elastically with the surrounding atoms to a rangeof -200A and produces about 1500-2000 displacements/event (Reeveand Woolfiey, 1980;Fleischer et al., 1975; Roberts et a1.,1981).The alpha-particleis ejected with high energies(4-6 MeV) which it losesthrough ionization of the atoms along its path (up to 20 pm). Most of the damageoccurs near the end of its range producing 100-500displacements/event. The effec- tivenessof thesecollisions is dependenton proper- ties such as crystal structure and bond type. A more detailed discussionof this subject is given by Sin- clair and Ringwood (1981). Kaiserstuhl zirkelite has accumulateda radiation dose of 1.2 x 1018alphas per gram. This dose is equivalent to the ageof a svNnoc samplecontaining 10 percent high level nuclear waste for 1000years (Oversby and Ringwood, 1981).The compositionof the waste is calculated using U-only fuel with 3 percent enrichment of U-235, a burn-up of 33,000 MWd/MTHM, and a cooling time of 150 days before reprocessing (Cohen, 1977). ln a previous study, Sinclair and Ringwood (1981) reported a decrease in intensity and slight boardening of the high angle powder difraction maxima of Kaiser- stuhl zirkelite. They also recorded a small change Fig. |. Drawing of an (00t) ptane of zirkelite containing M3' (<1 percent) in the unit cell volume after heating M4, and M5 sites. The Ml and M2 atoms occur in sites above these samplesat 1200'C. At higher doses of alpha- and below the octahedral plane. M2 sites above the plane are radiation, single crystal X-ray photographsof zirke- marked. (After Gatehouseet al.,1981). lite from Jacupiranga, Brazil, displayed only the parent fluorite-type diffraction pattern which could be incorporated into the thermal motion and record- be indexed on a face-centeredcubic lattice with a : ed as large U values. The high values of Uii (Table 5.06(4)4.These observations are direct evidenceof 3) for Kaiserstuhl zirkelite are interpreted to be the displacement damageresulting from alpha-recoil. direct result of alpha-radiation damage. The major diference between the results of this Temperature factor values are also dependenton refinementand those of Gatehouseet al. (1981)is parameters such as site occupancy and absorption. the increased values of the temperature factors. However, there is evidence to suggestthat these These parametersare found to be consistently 2 to 3 two effects have been kept to a minimum in our times as large as those reported for the synthetic analysis. First, the placement of elementswas well phase. The temperaturefactor correction is normal- defined from previous structure determinationsand ly made to allow for the oscillating thermal motions from crystal chemical considerations.The changes of the atoms. These motions efectively changethe in site occupancy made little difference to the high scattering curve and causethe calculated structure values of the temperature factors. Second, absorp- factors to be systematically greater than the ob- tion is considered to be low because of the small served ones with increasing angle. The effect of size of the data crystal and the excellent agreement alpha-recoil is to displace atoms or groups of atoms between equivalent reflections (interscale R from their lattice sites. In a single crystal X-ray 0.019). analysis, the displacementsassociated with the The results from this structure refinement are damagedareas averaged over the entire crystal, will most significant for the long term stability of zirke- 620 SINCLAIR AND EGGLETON: ZIRKELITE lite in syNRoc. Although Kaiserstuhl zirkelite has Prewitt, C. T. (l%6) SFLS. Report ORNL-TM-305. Oak Ridge received a dose of 1.2 x 1018alphas per gram, National Laboratory, Tennessee. equivalent to an age of syNnoc of 1000 years, a Pyatenko, Yu A. and Pudovkina, Z. V. (196/) The lattice metric of CaZrTi2Ol crystals(in successful single crystal X-ray structure Russian).Kristallografiya, 9,9E-100. solution Reeve, K. D. and Woolfrey, J. L. (1980)Accelerated was irradiation completed. The analysis shows that the crystal of syNnoc using fast neutrons. I. First resultson Ba-hollandite lattice has undergoneonly minor changesdue to the perovskite and undoped syNRocB. Australian Ceramic Socie- action of alpha-recoil. ty Journal, 16, 10-15. Ringwood,A. E., Kesson,S. E., Ware, N. G., Hibberson,W., Acknowledgments and Major, A. (1979a\The syNnoc process: A geochemical approach to nuclear waste immobilization. GeochemicalJour- The authors wish to thank Mr G. M. Mclaughlin of the na], 13, 141-165. ResearchSchool of Chemistry, The Australian National Univer- Ringwood,A. E., Kesson,S. E., Ware, N. G., Hibberson,W., sity for collection of the data and Dr J. Keller of the Albert- and Major, A. (1979b) Immobilization of high level nuclear Ludwigs University for providing the samplesof zirkelite. reactor wastes in syNRoc. Nature, 278,219-223. References Roberts, F. P., Turcotte, R. P. and Weber, W. J. (l9El) Materials characterizationcenter workshop on the irradiation Cohen, B. L. (1977) High-level radioactive waste from light effects in nuclear waste forms. PNL-3588, Pacific Northwest water reactors. Reviews of Modern Physics, 49, l-20. Laboratory Richland, Washington. Ferguson,J., Mau, A. W-H. and Whimp, P. O. (1979)photo- Rossell, H. J. (1980) Zirconolite-a fluorite-related superstruc- dimer of 9, l0-Dimethylanthraceneand tetracene.Crystal and ture. Nature, 283, 2E2-283. molecular structure, photophysics, and photochemistry. Sinclair, W. and Ringwood, A. E. (1981)Alpha-recoil damagein American Chemical Society, l0l, 2363-2369. natural zirconolite and perovskite. GeochemicalJournal, 15, Fleischer,R. L., Price,P. B. and Walker, R. M. (1975)Nuclear 229-243. Tracks in Solids. University of Califomia press, Berkeley. Whimp, P. O., Taylor, D., Mclaughlin, G. M. and Kelly, D. A. Gatehouse,B. M., Grey, I. E., Hill, R. J. and Rossell,H. J. (1977) The ANUcRys Structure Determination Package. Re- (l9El) Zircono\te, CaZr*Tb_xO7;structure refinements for search School of Chemistry, The Australian National Univer- near-end-membercompositions with x = 0.E5 and 1.30. Acta sity, P.O. Box 4 Canberra, 2600. Crystallographica, B.37, 306-312. Wimmenauer, W. (1966)The eruptive rocks and carbonatitesof International Tables for X-ray Crystallography (1974)Vol. IV. the Kaiserstuhl, Germany. In O. F. Tuttle and J. Gittins, Eds., Birmingham: Kynoch Press. Carbonatites, p. lE3-204. Wiley-Interscience,New York. Oversby, V. M. and Ringwood, A. E. (l9El) Lead isotopic studies of zirconolite and perovskite and their implications for long range syNRoc stability. RadioactiveWaste Management, Manuscript received, October 12, I98l; 1.289-307. acceptedfor publication, January 8, 19E2.