Crystal Structure Refinement of Huntite, Camgr(Cor)4, with X-Ray Powder Data
Total Page:16
File Type:pdf, Size:1020Kb
American Mineralogist, Volume 71, pages 163-166, 1986 Crystal structure refinement of huntite, CaMgr(COr)4, with X-ray powder data W. A. Doness Department of Earth and Space Sciences Univer s ity of Ca I ifo r nia Los Angeles,Caffirnia 90024 .q,NoR. J. Rpsoen Department of Earth & Space Sciences State University of New York at Stony Brook Stony Brook, New York I 1794 Abstract Peak-fitting of CuKa,,, X-ray powder diffraction data to sin d/tr : 0.57 gave 86 integrated (Bragg)intensitieswhich have beenused to refine the crystal structure ofhuntite, CaMgr(COr)., to R(4 : 5.3o/o.A March distribution model was used as the basis of a correction for the effectsof preferred orientation of the platey, micrometer-sized crystals. The atomic ar- rangementin spacegroup R32, a : 9.5027(Q A, c : 7.8212(6)A, proposedby Graf and Bradley (1962), is confirmed. Ca and Mg atoms occur within the same cation plane but are ordered with Ca's in trigonal prism sites (Ca occupancy : 0.99(2) and Mg's in edge- sharing octahedral sites. Interatomic distancesare nearly identical to those of dolomite, CaMg(CO,),. Introduction The present study of the huntite structure was under- Huntite, CaMgr(COr)o,is a naturally occurring rhom- taken to: (l) confirm and refine the huntite model using bohedral carbonatewhose composition lies half way be- a more extensive data set, (2) determine the degree of Cal tween dolomite, CaMg(COr)r, and magresite, MgCO3. Mg ordering, and (3) provide a test case for some recently Despite structural similarity, the polyhedral connectivity developed crystal structure analysis methods for use with of huntite is not the same as found in these latter, more X-ray powder diffraction data showing overlapped peaks common carbonates. andlor preferred orientation effects. Huntite is produced in low temperaturesurface or near surfaceenvironments. The mineral seemsto form either Experirnental methods by direct precipitation from Mg-rich, aqueous solutions The sample from Tea Tree Gully, Australia (Skinner, 1958), or as the result of interaction of such solutions with pre- sTasfightly crushed,loaded into 25 x 12 x I mm recessof plastic existing carbonates(for a summary see Kinsman, 1967; holder, moistened, compacted and smoothed with a spatula. Cole and Lancucki, 1975).Grain size of the natural sam- Data collection used a Picker horizontal powder diftactometer, ples is uniformly small (lessthan about two micrometers CuKar,2radiation, 40 kV , 20 mA, 4" take of angle,collimation by 4" Soller slits, l" divergenceand antiscatter slits, 0.005 inch in diameter), and therefore structural studies have been receiving aperture; exit beam monochromator and scintillation carried out using powder diffraction methods. detector.Intensities were measuredfor 10 secondsat each0.01' A model for the huntite structurewas proposedby Graf step from 20 : l22 to 15. and Bradley (1962) on the basisof similarity of the X-ray Integrated intensities were obtained by least-squarespeak fit- powder difraction pattern of huntite with that of the com- ting of step-scannedintensity measurements using starting mon rhombohedral carbonates,such as calcite, which in parameters obtained from second derivatives ofbackground cor- turn is relatedto the NaCl structure.From the relationship rected, smoothed,Ka2 stripped intensities (Snyder, I 983). Peak- of the unit cells, and using a NaCl-like arrangement of shapefunction is the surn of Lorentzian and Gaussianterms with : : metal atoms and CO3 groups, these authors adjusted the 5006Lorentzian for 20 15-30', 60010Lorentzian for 20 3O- 75', 75o/oLorentzian for 20 : 75-120. Fitted peak variable positional parameters to give interatomic dis- and widths varied, but generallyincreased from 0.16'to 0.34'FWHM from tancesappropriate for the chemical speciesinvolved. This lowest to highest24. Weights were assigned as 1/1o.The goodness model was then tested by comparing calculated X-ray offit, defined as: nns llo - I"l2/Io, is 1.19. 86 peaks were fitted intensitieswith 4l low-angle observedX-ray powder dif- ranging over three orders of magnitude in intensity, and were fraction intensities.The reasonableresidual, R(F) : I l0l0, corrected for Iorentz and polarization factors and for a mono- attestedto the essentialcorrectness of their model. chromator polarization ratio of 0.80. 0003404x/86/01024 l 63S02.00 r63 t64 DOLLASE AND REEDER: STRUCTURE OF HUNTITE Table l. Comparison of observed and calculated (Lorentz- Table 3. Cation-oxygen distances (in Angstroms) for huntite polarization corrected)intensities for huntite in order ofincreasing and ordered dolomite after Reeder and Wenk (1983) diffraction angle. An ampersand indicates that the several preceding reflections are overlapped and contribute to one huntlte dolo!lt€ observed peak Ca O(3) = 2.4O9(11) 6x O - 2.340 C. O(2) = 3.154(13) 6x hkll I(o) I(c) hkll I(o) I(c) Is-O(1)=2.704<16> 2*. llg-O(2).2.LO7<74> 2X !tg 2.O44 1011 84 L562 64 70 5491 93 92 I 120 43 0660 r47 L42 3690 17 27 ilg-O(3)=2.O31(13) 2X o221 3254 63 66 3SA4 - L7 OLL2 56 3363 A7 110 7074 - 252 C(1)-O(1)=1.27O(15) 3X 2L3L 202 209 437L 55 59 3365 - 256 C(2r-O<2>.1.292<22, 1X c-o-1.283 2022 1403 1397 2570 1A5 1a9 1743 - L73 C(2) - O(3) .1.27A<2L> 2X o330 30 26 ood6 3o7 349 zra3 - 1as ooo3 145 L42 31as 29 s2 1317 - 33 1232 164 155 3472 69 a5 e 957 927 2240 149 161 0554 aa 101 aoa2 34 2a t123 10a L2L o44s Llo 145 62a4 43 40 1341 152 141 t6lr L44 160 4047 32 30 404L 237 r92 2464 101 lO9 6173 139 139 positional parameters,overall isotopic temperaturefactor, Ca- 3r42 ?I4 674 2355 33 49 7292 311 269 tol4 32 33 6t72 26 31 3693 - 3l site occupancyand preferredorientation parameter).Without o333 - 16 s164 - 42 6353 - 22 : 3033 -29 0335 - 15 & 64 53 preferredorientation correction, R(F) 8o/0,and someC-O dis- & 51 45 3036 - a 3237 34 4I tanceswere as short as L24 A. A preferredorientation correction 3251 109 103 5273 - 27 A191 3A 3A o442 78 76 2573 - 33 2025 234 202 similarto that of Rietveld(1969), except that a Marchfunction, 1450 93 100 E 106 L26 0666 - 66 : o224 35S1 - 145 6066 - 63 p(a) [P-%cosza+ Pbsin2a]-3zz(March, 1932), was substituted 2243 377 367 707L-9&123L29 p(a) : resulted 2352 109 114 & t4a 154 5365 - 65 for Rietveld'sexponential function, exp(-Ga2), 2734 23 3l 2246 135 L49 0775 - 4 in a lowerR-factor. With thepreferred orientation plane (101l), o551 5055 48 43 E Lr? 90 3360 405 395 53A2 - 29 4601 4L 35 P refinedto 2.00(11) resulting in correctionfactors ranging from o 115 10 t4 0772 - 293 123a 30 36 : : 246r tt2 114 & 325 323 3494 - 39 0.8 to 1.3.Final R-factors are: R(f') 0.080,R(F) 0.053, 5052 54 52 4374 23 26 2576 - SA wR(F) : 6.966.All calculationsused unpublished programs writ- 1344 255 309 42es ?4 zg s216 - 131 1453 - 114 26e2 63 64 & 222 25a ten by the authors. 4153 - 165 4483 - 5A 7301 5a 4L & 270 2AO 0227 - 26 6402 46 57 2023 A 76 94 0990 195 164 Results and discussion s 161 139 150 ls6s ?7 93 os64 14 rs 4262 162 159 1456 - 32 2467 33 54 4044 r37 L37 4156 - 32 3tea ro7 1lo To test the ordering of Ca and Mg atoms between the 1235 148 L7? E 74 64 2794 !74 141 5503 19A LL7 alkaline-earth sites, the occupancy of the "Ca" site was treated as a least-squaresvariable and refined to 0.99(2) Ca per site. This occupancy,plus the near stoichiometric chemical formula reported for the Tea Tree Gully huntite Unit cell dimensionswere obtained by a least-squaresfit to all (Skinner, 1958),indicates that the alkaline-earth sitesare 86 d-values obtained from peak-fitting with 2d calibration by Table I lists external quartz standard, refined sample displacement from fo- fully orderedwithin the accuracyof the data. cussingcircle 0.12(l) mm, unit weights,a:9.5027(6), c: observedand calculated (Lp-corrected)intensities. 7.8212(6)A. Spacegroup is R32 consistentwith X-ray powder Table 2 lists the refined positional and thermal param- data and isomorphous borates (Hong and Dwight, 1974). TEM eters. Cation-oxygen distances resulting from these pa- observations ofgirains on holey carbon film show diffraction sym- rameters are given in Table 3 where they are compared metry consistent with this unit cell and space group; there were with the correspondingdistances found in the single-crys- no indications seenoftwinning, which might be expected,ifthe tal refinement of dolomite by Reeder and Wenk (1983). symmetry were lower. The inequant, platey grain shapes match The huntite bond lengths are not significantly different previous SEM observationsofFaust (1953). from those of dolomite. The crystal structure was refined by least-squaresfitting of 86 A major differencebetween huntite and dolomite is the 12, or, a sum ofF for peaksspaced less than 0.5 FWHM (Cooper polyhedron. In the more et d., l98l). Weights were assignedas 1/(e2* s2)where e is esd shape of the Ca coordination polyhedron from peak fitting and s is an estimate of all other sources of common rhombohedral carbonatesthe Ca is measurementand model uncertainties.In this case,s was chosen a nearly regular octahedron,whereas in huntite this poly- as 1.0, a value lessthan about 850/oof all e-values.Neutral atom hedron is quite closeto being a right trigonal prism. Though scatteringfactors were taken from International Tables (1974) a trigonal prism is somewhat uncommon for Vl-coordi- and there was a total of I I variables (scale factor.