American Mineralogist, Volume 61, pages 334-336, 1976

X-ray difrractionpowder data

Prren Beylrss Departmentof Geology,Uniuersity of Calgary Alberta T2N I N4. Canada

Abstract

The presentstandard of documentationof X-ray diffraction powder data is generallynot adequate' Therefore suggestions for adequate documentation of X-ray diffraction powder data are tabulated.

Introduction Discussion When X-ray diffraction powder data are used for Following the mineral's name and ideal formula routine mineral identification,a referenceto the unit (Table l, la), cite its locality if it occurrednaturally, cell constantsand errors calculatedtherefrom, and to then give its chemical analysisand the formula de- published comparative data-if available-is suf_ rived therefrom so as to indicate how far it deviates ficient. On the other hand, powder data better than from its end-member composition. For synthetic that available should be published. The suggestions specimens state the source of the constituents, that minor improvements in such data should be method of preparation, and any subsequenttreat- published only by the Joint Committee of powder ment. Diffraction Standards (often still incorrectly called The Debye-Scherreror Gandolfi camera is useful ASTM is ) unacceptable;this would restrict the pub- for routine mineral identification;however, it is bet- lished data to a copyrighted format on cards, which ter to use the Guinier camera or diffractometer to are not acceptable to most libraries,and would also record data for publication, becausethe resolutionis incur a long publication delay. Therefore the data three to five times better. Poor data (e.g., Debye- would not be exposedto critical evaluation as in a Scherrerwith visually estimatedintensities) may gen- mineralogical journal, and mineralogists would be erally be usedeasily for identificationagainst a good discouragedfrom attemptingto producebetter data. standard (e.g., Guinier with densitometermeasured Selected Powder Diffraction Data for Minerals intensities),but not vice versa.If insufficientquantity (1974), which was produced by the Joint Committee of the mineral is availableand a Debye-Scherreror of Powder Diffraction Standards,contains X-ray dif- Gandolfi camera is used (more than one Gandolfi fraction powder data for about 1850minerals. Future pattern is measured,because Gandolfi may miss re- editionswill include both additional and replacement flectionsof low multiplicity or give falseintensities), data. The reasonsfor replacementdata include (a) a then a minimum diameter of 114.6mm is requiredto larger 20 range, (b) additional reflections with weak give adequateresolution- intensities,(c) greater accuracy in measurementof For filtered radiation, only the target elementand intensity and interplanar spacing,(d) better resolu- composition of the filter (e.g.,Cu radiation/Ni filter) tion, (e) improved indexing,and (f) additional miner- are stated. For monochromatizedradiation, the tar- alogical properties. get element with spectral line and monochromator A standard for X-ray diffraction powder data has crystal with diffraction plane (e.g., CuKa radiation been recommended by Kennard et al. (1971) in the with graphite 002 monochromator) are stated. International Union of Crystallographycommission In order that the data may be corrected if more report. Since their report is neither widely used nor precise constants become available, the International very specific,the suggestionsfor adequatedocumen- Crystallographic Tables (1974) recommend the con- tation of mineralsare listedin Table l. tinued publication of the wavelength value (e.g., 334 X.RAY DIFFRACTION POWDER DATA 335

Trau l X-ray diffractionpowder data suggestions visual estimateof intensitiesfrom a film is sumcient. In order to record all weak reflections, the film is l.a. Mineral name and ideal tbrmula. the intensitiesof b. Locality, chemical analysis, and derived formula; or method commonly overexposed.Therefore of synthesis. the strong reflectionsare similar. Becausemany users 2.a. Type of instrument and diameter if camera. record quantitativeintensity data for routine mineral b. Radiation; and filter or monochromator. identification,this is the minimum publication stan- c Wavelength value, or name of standard, plus its unit cell con- dard. If equipmentis not availableto useeither coun- stant. a photographic strip d. Temperature 'C. ter methods or a densitometer, e. 2l range greater than 62 (d: l.5A forCuKa radiation). as describedby Klug and Alexander(1974) should be f. Quantitative intensity measurement. used. In the future, Kennard et al. (1971) expect that g. It/l"-intensity ratio of the strongest reflection of the mineral the observedintensities will be compared with calcu: to the l13 of corundum. lated intensities to quote a discrepancyindex Rr, List of dou d.,r. I/It, /".1" (if available), and hkl. 3. , , = - good where 4.a. Space group(s) consistent with indexing. where Rr >( I.^r")/2.I. Rr is only b. Unit cell constants with standard deviations. the structureis accuratelyknown. c. Z-the number of formula units per unit cell. The relative reflection intensity (peak height) d. Calculated density. method is used to obtain semiquantitativevalues of e. Reasons for unindexed reflections. the minerals in a rock. In order to use this method, 5.a. Physical properties; color, hardness, density. (4) of all b Optical properties. the intensity of the strongest reflection 6. References. mineralsmust be known relativeto the intensityof a standard(1"). The 113reflection of corundum (Alros) by the Joint Com- I .54I 8A ) o/ the name of the standardwith its unit cell has beenselected as this standard constant (e.g., silicon 5.430884). The wavelength mittee of Powder Diffraction Standardsbecause of value with only the name of a standard is unaccep- corundum's commercial availability as "Linde A" in table. lf a standardis used,it is the unit cell constant approximately 0.3 pm size, its purity, its chemical that is neededfor correctionrather than a wavelength stability, and its freedom from orientation due to value. A silicon standard is now commerciallyavail- particle shape,although one must checkthat it is well able from the U.S. National Bureau of Standards, crystallized.If enough of the mineral is available, and a standard to provide larger d values is being the ratio 4//" should be determinedfrom a l: I mix- sought. Becausethe unit cell constants vary with ture with corundum. temperature,an exacttemperature is needed.A state- The observedspacings (d.o") are listed only to the ment such as "room temperature" is unacceptably proper number of significantfigures; this meansthat vague. the number of significantfigures will increasesystem- The threeentries in the Hanawalt searchmanual of atically as d decreases,according to the precisionof SelectedPowder Dffiaction Datafor Minerals (1974) the 20 measurement. The calculated spacings (d"'1") are basedupon the three strongestreflections which are all listed out to the maximum observed2d so that occur before 90'm (d > 1.09A for CuKa radiation). the indexing may be evaluated. Only 30 percentof the data in the Powder Diffraction The reflection intensities (1) are quoted relative to File is recorded to 20 greater than 90o, whereas 70 the strongestreflection (l), which is taken arbitrarily percent is recorded to 20 greater than 62" (d < 1.54A as 100. The intensities are measured as peak in- for CuKa radiation). In order that published X-ray tensitiesunless otherwise stated. To insureaccurately diffraction powder data may be used successfullyin measuredintensities, special precautions are needed this searchmanual, d should be given for 2d anglesup to avoid preferredorientation in mineralswhich have to 90"?l; however, for many minerals little useful prominent cleavages.Intensity terms such as weak data is recorded with 20 greater than 62o, which and strong are useful for routine mineral identi- appearsto be the minimum limit. Such statementsas fication but are too vague for data publication. Al- "plus 15 lines down to 1.08A" are unacceptablebe- though methods are describedby Henry et al. (1961) causethey are useless,as the number of reflections for the indexingof powder X-ray diffraction data,the recorded depends on the ability of the method to indexingobtained has beenfound incorrectin a num- detectweak reflections. ber of cases.For instancethe mineral ferritungstiteis : Intensity measurementmay be made with counter describedas tetragonal(a = 10.28A,c 7.774),but methods, a densitometer,a photographic strip, or a recent single crystal X-ray diffraction indexing by visual estimate.For routine mineral identification,a Machin (personal communication) shows it to be 336 7ETER BAYLISS

: cubic (a 10.284). Therefore it is desirableto use Conclusions singlecrystal methods wheneverpossible. If approximate unit cell constants-preferably by Becausepublished X-ray diffraction powder data single crystal X-ray diffraction-have been recorded will be usedas standardsby other mineralogists,they by the investigatoror by previous investigators,the should meet certain minimum requirements which unit cell constantsshould be refinedfrom the powder are obtainable today in most laboratories. Such a procedure data by the least-squaresmethod. Many such least- will eliminatethe publication of an unnec- squarescomputer programs are available (e.g., Ap_ essaryseries of minor improvements. pleman et al., 1972).It is useful to state whether an absorption correction is applied or not, becausethe Acknowledgments absenceof an absorption correction may cause a Drs. D. E large error in both intensity and d (e.g., a sulfosalt Appleman, J D. Hanawalt, A. A Levinson, and D. K Smith are thanked for their critical suggestions. mineral with Cu radiation in a Debye-Scherrercam- era).The estimatedstandard deviation determined by References a least-squaresmethod is quoted in brackets to in- dicate the error in the last figure (e.g.,5.43099(4)A). AppLevnN,D. 8., H. T. EvnusJn eNn D. S HnNDwsnrsn 0972) An estimatederror should not be publishedbecause Job 9214: Indexing and least-squaresrefinement of powder dif- fraction data. U.^S. Dep. it is unreliable(Kennard et al., Commer., Nat Tec. Inf. Seru. l97l). PB2r6188. Calculated values such as density and Z (number HsNnv, N F. M., H. LresoN,AND W. A. Woosren (1961)The of formula units per unit cell) are useful, sincethey Interpretation of X-ray Difraction photographs MacMillan, provide a check.Other calculatedvalues such as axial London. ratios and volume are useful but not essential.A International Crystallographic Tables (1974) Vol. IV. Revised and supplementarytables. Kynoch press,Birmingham, discussionof unindexablereflections should indicate England KeNNnno,O., J. D HlNnwnlr, A J. C WrlsoN, p. M DrWolns. the probability that thesereflections representan im- nuo V. A. FneNr-KeueNrrsxv (1971) powder data. J. Appt. purity, a supercell,or an incorrectchoice ofunit cell. Crystallogr.4, 8l-86. All mineral speciesmay be distinguishedone from Kt-uc, H. P., nNo L. E. AuxnNorn (1974) X-ray Dffiaction pro_ another with a chemicalanalysis and a unit cell deter- cedures John Wiley & Sons, New York. Selected Powder Dffiaction Datafor Minerals (1974) mination. However, optical and physical properties Joint Com- mittee of Powder Diffraction Standards, Swarthmore. pennsvl- are given as useful additional information. If these vanla. properties are taken from the literature, adequate Manuscript receiued, March 17, I97S; accepted referencesshould given. be for publication, October 6, 1975.