Preparation of powder specimen for quantitative XRD Reinhard Kleeberg TU Bergakademie Freiberg Mineralogical Institute kleeberg@mineral.tu-freiberg.de References Moore, D.M. & Reynolds, R.C.: X-Ray Diffraction and the Identification and Analysis of Clay Minerals. Oxford Univ. Press/CMS 1989, 1997. Bish, D.L. and Reynolds, R.C., Jr. (1989) Sample preparation for X-ray diffraction. Pp. 73-99 in: Modern Powder Diffraction (D.L. Bish and J.E. Post, editors). Reviews in Mineralogy 20, Mineralogical Society of America, Washington, D.C. Środoń, J., Drits, V.A., McCarty, D.K., Hsieh, J.C.C. and Eberl, D.D. (2001) Quantitative X-ray diffraction analysis of clay-bearing rocks from random preparations. Clays and Clay Minerals, 49, 514-528. Hillier, S. (2003) Quantitative analysis of clay and other minerals in sandstones by X-ray powder diffraction (XRPD). International Association of Sedimentologists Special Publication, 34, 213-251. Kleeberg, R., Monecke, T., Hillier, S. (2008) Preferred orientation of mineral grains in sample mounts for quantitative XRD measurements: How random are powder samples? Clays and Clay Minerals, 56 (4), 404-415. 2 1
Crystallite statistics 1 Impact of crystallite statistics on the diffraction signal depends on: sample volume contributing to diffraction (limited by the primary slit system, sample holder, absorption) section of the diffraction rings seen by the detector (limited by Soller slits, receiving slit, detector area) number of particles per volume unit (limited by particle size) sections seen by a point detector sections seen by an area detector coarse material ideal powder - only the particle size may be optimized by sample preparation - ideal case: infinite number of crystallites = continuous cones of diffraction 3 Crystallite statistics 2 Practical consideration: How many particles do contribute to our diffraction signal? Exemplary calculation at: http://epswww.unm.edu/xrd/xrdclass/07-errors-sample-prep.pdf Quartz in different particle diameter in a conventional Bragg-Brentano diffractometer Particle diameter 40 µm 10 µm 1 µm Diffracting particles 12 760 38 000 To achieve a standard uncertainty of < 1%, > 52900 particles would be needed! Experimental test see: Klug, H.P. & Alexander, L.E. X-Ray Diffraction Procedures. John Wiley, New York, 1954 Quartz in different particle diameter in a conventional Bragg-Brentano diffractometer Conclusions: Particle size fraction /microns 15-50 5-50 5-15 <5 Standard deviation of I quartz 101 /% 18.2 10.1 2.1 1.2 -If we want to measure single peak intensities of minerals correctly, we should try to mill any rock samples to < 5 µm. -This is necessary to get reliable results of Rietveld structure analysis, too. 4 2
Microabsorption 1 Microabsorption causes loss of intensity inside a particle. If this loss is different between two phases, the beam is interacting with different phase volumina! Thus, the QPA results will be falsified. High absorption Lower absorption 5 The theory of BRINDLEY (1945): Microabsorption 2 - depends on grain diameter d and linear attenuation coefficient µ - intensity loss / correction factor can not be determined from the powder data! - can be ignored if the product µd is equal for all phases - intensity resp. scale factor can be corrected if µd < ~0.5, and if both values are known for each phase! Approximated BRINDLEY correction in the BGMN structure file: GOAL:forsterite=GEWICHT*exp(my*d*3/4) GEWICHT my d Rietveldscale factor, mass weighted linear attenuation coefficient in 1/µm, provided by BGMN estimated particle diameter in µm, to be set by the user in structure or control file 6 3
Conclusions: Microabsorption 3 When does microabsorption occur in practice? Example: QPA of sulphide bearing rocks Absorption contrast between quartz and pyrite coarseness according to BRINDLEY (1945) phase grain size/µm µd Cu K µd Co K quartz 10 0.093 medium 0.146 coarse pyrite 10 1.012 0.516 very coarse coarse - No problem for fine clay fractions and silicate minerals having similar (low) µ. - But, if we want to quantify rock samples containing high absorbing materials, we should try to mill the samples to < 5 µm. 7 What we have For example, sedimentary rocks: large crystals (>mm, e.g. quartz), together with fine grained material (clay) hard and soft phases together differences in density may occur (pyrite, hematite, rutile, silicates ) mechanically or thermally sensitive phases (e.g. clay minerals, zeolites, sulphates) What we want to get mean particle size below 5 µm, optimum 1-3 µm narrow particle size distribution, no amorphization, no rocks-in-the-dust no additional disorder ( e.g. by shifting of layers) no contamination (e.g. by grinding elements) no loss of material (e.g. by dusting or dissolution) no phase transformation (e.g. by dissolution-precipitation, hydration-dehydration ) no phase separation (e.g. by hardness, density, primary particle size, electrostatic ) 8 4
General principles: Golden rules for milling - If a sample is not homogenous or not representative, any further efforts (for phase quantification in general) are unnecessary. - The maximum grinding energy has to be adapted to the most sensitive phase. - Some overmuch large grains are better than any destroyed phase(s). - Working in closed containers to avoid loss of material by dusting, checking for losses by sedimentation, dissolution, electrostatic adhesion... - Quick working for minimum contact with dissolution agents and air to avoid any phase alteration. - Samples should not be exposed any enhanced temperature. - Note (or keep in mind) what treatments and changes your sample was exposed between sampling and XRD measurement. 9 Grinding methods useful for clay-bearing rocks 1 Hand grinding in agate mortar - easy, but strenuous - allows a stepwise sieving of < 20 µm, very mild - resulting powder is mostly to coarse for high absorbing materials - danger of loss by dusting Foto: Detlev Müller Foto: Detlev Müller 10 5
Grinding methods useful for clay-bearing rocks 2 McCrone micronising mill - wet grinding (in water or alcohol) produces optimum grain distribution - crushing the starting material < 0.4 mm is necessary - danger of contamination by grinding elements (corundum, quartz, ZrO 2 ) - dissolution and/or alteration by the grinding liquid (water, ethanol, hexane ) - filtering or drying of the slurry and additional homogenisation is necessary Milling time /min 48 24 12 6 6 Mean diameter /µm 1.8 2.6 5.8 11.4 5 4 Volume % 3 2 1 0 0.04 0.1 0.4 1 2 4 10 20 40 100 400 Particle diameter/µm Particle size distribution of McCrone milled quartz, from Hillier (2003) Foto: Detlev Müller 11 Homogenisation - is necessary for admixing of any standards as well as to overcome the (unavoidable) separation processes during grinding, sieving, and sedimentation - should be applied with minimum energy - can also used for destroying of aggregates - should be done immediately before filling the sample Useful methods: - admixing of standards by grinding together (to get similar grain size and intimate mixing) - larger amounts of easily flowing powders by overhead-shaking - small amounts by manual stirring - homogenisation and destroying of aggregates by swirling the powder with small steel balls ( Ardenne vibrating mill or Fritsch mini-mill pulverisette 23 ) 12 6
Problems Milling Filling Examples Making the powder mount for Bragg-Brentano XRD 1 are an old matter of debate, and sometimes treated like religion Possible pre-treatment of powder before filling: - No pre-treatment (filling the milled and homogenised powder as is ) - Forming irregular shaped aggregates - by freeze-drying - by spray-freeze-drying - by admixing glass powder, plastics or organics, e.g. cork powder - Forming spherical aggregates - by spray-drying with binder (cellulose-acetate, polystyrene, polyvinyl-alcohol) - by spray-drying without binder, application of heat - by mechanical aggregation in a shaking/sieving procedure ( sieving in ) Aspects for decision: time consumption, material consumption, stability and density of the samples, loss of intensity by dilution, sample transparency, alteration of minerals, reproducibility, labour protection 13 Problems Milling Filling Examples Making the powder mount for Bragg-Brentano XRD 2 Simple tricks may make the difference Front-loading (standard holders, very popular) - powder pressed using a glass slide: very easy, but extreme PO - powder roughened by emery paper: not so much PO, but badly reproducible and depending of the properties of the powder - surface roughened by a razor blade: as above, plus high roughness - sprinkling/sieving some powder on the surface: good randomness, but errors in sample height, high roughness, sometimes phase separation - possible to do with any kind of aggregates, perfect randomness possible 14 7
Problems Milling Filling Examples Making the powder mount for Bragg-Brentano XRD 3 Back-loading (special Philips/Panalytical equipment) - easy to do, but extreme PO - no chance to use for unstable aggregates like freeze-dried clays 15 Problems Milling Filling Examples Making the powder mount for Bragg-Brentano XRD 4 Side-loading (with special side-opened holders, but also for standard holders) - with normal powder: relatively easy, reproducible, acceptable PO - with irregular aggregates: perfect randomness possible - general problems are density, homogeneity, and stability, e.g. for rotated samples 16 8
Errors in sample preparation 1 Influence of overmuch large grains on the powder pattern Comparison of 2 Scans 300 < 63 µm plag 002/040 < 20 µm Intensity / cps 200 kfsp 002/040 preferred orientation! topaz 230 rocks in the dust 100 0 25 30 35 4 0 2 Theta / ( Scan Axis: 2:1 sym. const. area ) topaz bearing granite, samples ball-milled < 63 µm and hand-ground < 20 µm 17 Errors in sample preparation 2 Phase separation and loss of material during powder preparation mica + quartz lost by dusting during dry sieving halite dissolved in ethanol Mineral mixtures, ground by hand/dry sieving < 20 µm and McCrone milling/ethanol, filtered slurry 18 9
Errors in sample preparation 3 Sample contamination by abrasion from grinding elements corundum line positions Kaolin Spergau reference sample, McCrone ground using corundum grinding elements and using agate grinding elements 19 Errors in sample preparation 4 Preferred orientation of grains/aggregates dependent on filling technique Comparison of 2 Scans 400 bentonite, fraction < 0.2 µm 02,11 300 Intensity / cps 200 005 side-loading front-loading 100 06,33 0 20 30 40 50 60 2 Theta / ( Scan Axis: 2:1 sym. ) 20 10
Errors in sample preparation 5 Preferred orientation of grains/aggregates dependent on filling technique Ms 002 Dc 002 Ms+Dc 110, 020 Ms 004 Ms 060 Dc 060 Spray-dried front-loaded Powder side-loaded Powder front-loaded Mixture quartz/dickite/muscovite 40:30:30, McCrone milled, different filling techniques (from Kleeberg et al., 2008) 21 11