Document Xzxrbv6g83zrNpmERZMZ2DVgK

Ens Ultrasonic oise support films reda R. 1974. High py by using BeO Microscopy Proc 8th Canberra Australia dchild D. J. eds E. 1974. The direct molecules in the scope In Electron Electron Microscopy anders J. V. and 231 ter program system -beam dynamical Thesis Arizona State electron microscopy of small holes and k in press Micron 1977 Vol 8 47 5P5ergamon Press Printed in Great Britain Practical limitations of selected for identifying mineral area particles electron diffraction techniques R. E. FERRELL JR and G. G. * PAULSON Department of Geology Louisiana State Baton Rouge Louisiana 70808 U.S.A. University and Materials Evaluation Laboratory Inc. Baton Rouge Louisiana 70808 U.S.A. Manuscript received in final form November 25 1976 Selected area electron diffraction techniques are being suggested and characterization of mineral particulates in studies as potentially useful ones for the identification of these studies it has proven to be of industrial and environmental health hazards In make the interpretation of great value However there are serious limitations in the method many required to of diffraction patterns ambiguous or drastically increase the which confirm the identification identification of each individual particle in a complex sample amount of operator time The problems mostfrequently encountered arise because their degree of crystal perfection and the presence of ofvariations in the thickness of individual crystallites also eliminate diagnostic reflections The polycrystalline aggregates developed cleavages standardization and even then the results precise measurement of interplanar distances requires intemrnaayl Electron diffraction patterns from different minerals are so similar that thay cannot be are difficult to index and interpret without some differentiated the particles prior knowledge of the identity of techniques The results obtained from a variety offibrous minerals mineral identification in polyminerallic samples It indicate that the technique is not well suited for but may be useful in a qualitative cannot recomended for quantitative identification identification or a complex aggregate Electron sense to tell whether a particle was crystalline diffraction must be supplemented with other fibrous twinned unambiguous identification information in order to obtain an INTRODUCTION frequently Electron microscopy is an extremely valuable faonralytical technique which has been employed many years to study the morphology of finely divided particulate materials The electron microscope can also be used to obtain a selected area electron diffraction SAED pattern which contains information about the of the material in the electron becarymstaanldstirnusctoumree cases can be used to identify the crystallites The two techniques are especially useful because the observations are unique enough to provide an adequate identification In the field of procedure mineralogy the combined ob- servations are useful in many cases where indi- vidual small crystals must be identified The Also Department of Engineering Science Louis- iana State 70808 University Baton Rouge Louisiana technique has been used to determine crystal structures identify traces of minerals in rocks and characterize the phases formed during thermal decomposition or nucleation of micron- sized particles There is an increased interest in the mineralogical applications of electron diffraction results as explained by Ross and Christ 1958 because the method can be used to identify small particulpaotteesntiinalaliry and water samples For example Clark and Rudd 1974 presented electron micrographs SAED patterns and ray diffraction data for asbestos standards supplied by the IUCC International Union Against Cancer and have ways that the results might be useful isnugsgteusdtieeds of industrial and environmental health hazards The SAED technique is extremely difficult to use with some minerals and its application to others should not be attempted without an understanding of the potential sources of error Among the most serious limitations are the 47 JNJTALC0000708 ~- 48 R. E. Ferrell, Jr. and CG. G. Paulson s4 ample8 thicknessa,ritshieng from R. Ferrell in from variations variations high energy energy beam SS electromagneticeelelctromeagnetcic tromagnetic specimen demonstrated demonstrated that symmetry altered diffraction diffraction pattern by minor - changes in objective current current The thermal instability some minerals vacuum them thermal standards difficult examine examine Internal ) (b) must standards interplanar spacings spacings employed review interplanar interplanar following of the prin- ciples brief review review prin- prin- following diffraction and discussion ciples method's practical practical limitations discussion fying limitations samples identi- particulates particulates fying cases diffraction and samples In several specimens collected collected am- phibole asbestos asbestos mcuimmninegtroniatel cummingtonite cummingtonite an as examples examples the encountered and results obtained problems encountered encountered made obtained diffraction among the electron comparison data from diffraction and habit simple device is the radial radial displacement the the diffraction spot ring ring is the diffraction constant and is diffraction camera the Bragg Bragg diffraction angle the specimen to screen distance and the the angle ) such that L these tan 20 ======== or sin sin cos 0 cos = = dTehveiceesneirs gy of the electron beams very small small values sufficiently of 20 great that only very small are observed and monitoring monitoring changes monominerallic than identifying identifying present monomineralic poly- discussion discusion similar to ones minerallic minerallic This ones Alderson Halliday 1965 and PRINCIPLES PRINCIPLES ELECTRON DIFFRACTION cos cos 201 ~ ] Thus for all practical purposes the diffraction condition is Bragg sin 0 = L Substituting the derived which which an expression is direct can used lattice spacing ( from from Land Thus known Thus d= the constant constant difficult measure measure by obtaining obtaining diffraction evaluated directions standard with diffraction pattern pattern from known spacings can then Unknown simple simple obtained simple simple ratio can then be obtained fluorescent by a simple and and the viewing screen In the the displacement displacement undeviated otnehfuenctdiiffroactned beam beam from undeviated undeviated one a function of Variations in can quite an an i results to to k used the for are are thi diffraction pa Electron d cleavage cleavage ba irregular irregular star arrays in caused th caused close The pattern pattern pattern pattern tl rows of close The pattern JNJTALC000070860 gy eam electromagnetic lens specimen viewing screen lectron diffraction diffraction rings or of the specimen and of diffraction in a isplacement of the diffraction camera action angle nce L and the cos 20 = R ams used in these at only very small 1 poses the Bragg e Bragg equation h can be used to spacing d from casuring R when ffraction camera e It is evaluated pattern from a ies Unknown d- ned by a simple Rstandard quite pronounced ation in a number f precise measure- ectron diffraction Practical Limitations of SAED for Identifying Mineral Particles 49 patterns an internal standard is a necessity In all results to be discussed below a thin gold film was used for calibration The gold diffraction effects are the continuous rings in the electron diffraction patterns Fig 2 Electron diffraction phenomena are fairly easy to interpret because the pattern produced on the fluorescent screen is a direct image of a section of the reciprocal lattice and the recipro- cal lattice is directly related to the mineral's atomic structure For example the arrangement of atoms in cummingtonite can be reduced to a Bae Ome TR Ge sad Se 0205 td re") YRS 2 OU 4 Ree ByBes 3 : eet SMA ae OB OS Fig 2. Typical electron micrograph and electron diffraction patterns produced by fibrous minerals in air or water samples The micrograph a illustrates the straight sides produced by cleavage in crystals of cummingtonite and the tendency for the particles to agglomerate The irregular background is caused by the thin discontinuous layer of gold applied to the sample for internal standardization of the diffraction effects Some crystals produce clear hexagonal arrays of diffraction spots b while others are more complicated c The continuous rings are caused by the gold standard A typical pattern of cummingtonite is illustrated by the latter The rows of closely spaced diffraction dots are perpendicular to * and the rows are separated by 001 The pattern is complicated by a second row of dots from another crystal in the electron beam which lies at an angle to the major crystal ^ 9,000 a al JNJTALC0000708 aad 50 R. E. Ferrell Jr. and G. G. Paulson * dhki Origin 100 .951 .951 100 1001) 01 a 200 primitive set of equivalent points which form lattice with monoclinic a points in the * and * symmetry The vectors defining the actual unit cell 001 001 101 plane such as 100 100 b = 1.819nm c = 0.533nm are a = 0.951nm A 101 etc. The and B = 101.90 complete planar array of reciprocal lattice relationship between an actual lattice points could be constructed as in the plane and the corresponding reciprocal lattice fashion for other planes in the above network is shown in Fig 3. This is a hil or h21 and crystal such as tion of the h0 plane of cummingtonrietpereasenndtai-t a dimensional network would result This is the reciprocal lattice but can be used to predict the arrangement of all of the points in it would when the electron beam is spots diffraction not appear in the axis parallel to the zone 010 Several planes in the real lattice are pattern Only those points which intersect the Ewald sphere are observed This shown The indices of these are enclosed in sphere which limits diffraction is drawn regular parentheses The reciprocal lattice points are along lines drawn through the origin the origin of the reciprocal lattice its through in the electron beam and its centre is radius is 1/2 and perpendicular to a A given reflecting plane 1/2 In electron diffraction a number of reciprocal lattice point is in effect a point which represents a real plane in the lattice points are usually observed berceacuisperotchael reciprocal lattice structure The lengths of the crystal Ewald vectors are very small and the vectors which locate the reciprocal lattice sphere is large A section through the reci- related to the distance points are inversely procal lattice and the limiting sphere helps and are arbitrarily scalbeedtween the lattice planes explain this situation Fig 4 As shown the in the illustration The lattice point 100 is on the line represent- reciprocal lattice points are elongate spikes because the crystals yielding good diffraction ing the reciprocal lattice direction a which is perpendicular to ( 100 and the point 001 is in are usually thin and therefore are not triperiodic The perfectly the direction of * which is resulting diffraction pattern perpendicular to would contain the points 000 001 001 002 013 and 013 vectors 100 and respectively The points aalsssoocsihatoewdnwiItfhthtehe planes 101 and 200 are half that of spacing of one plane is one- will be another the reciprocal lattice point twice as far from the origin i.e. 100 and 200. The resulting pattern would contain all Single crystal electron diffraction can be interpreted in a fairly patterns straightforward manner because with a little knowledge of crystallography one can predict the kind of image to be formed by certain minerals Close agreement between observed and theoretical patterns identifies the particles However in ewald sphere Fig 4. A hypother sphere which limit lattice of cummingt intersecting the Ew appear in the practice electron rather ambiguous of the problems are in more detail late 1. Crystals may measurement of i internal standardi many minerals ar be distinguished by 4. developed diagnostic reflectio planar spacings are because the zone may not be coincid 6. Polycrystalline c may produce spec may be more diffic MATERIAI After a careful diffraction pheno standpoint several Philips EM tr scope and actual d tained The minera properties are listed were selected becau commonly occurrin ratios greater than test the applicabilit the identification of The samples were the following mann en eee ee JNJTALC000070862 | T e and the re- and of; is the e such as 100 100 f reciprocal lattice 1 as in the above _he crystal such as mensional network iprocal lattice but not appear in the hose points which are observed This n is drawn through Lattice its centre is radius is 1/2 umber of reciprocal "servedbecause the very small and the -on through the reciiting sphere helps 4 As shown the re elongate spikes ng good diffraction re are not perfectly diffraction pattern 000 001 001 002 diffraction patterns irly straightforward little knowledge of redict the kind of tain minerals Close ed and_ theoretical ticles However in Practical Limitations of SAED for Identifying Mineral Particles 51 ewald sphere electron beam i by 023 }reciprocal reciprocal lattice plancs 023 h21 Fig 4. A hypothetical section through the Ewald sphere which limits diffraction and the reciprocal lattice of cummingtonite The elongate lattice nodes intersecting the Ewald sphere are the points which appear in the electron diffraction pattern practice electron diffraction patterns can be of the problems are outlined below and in more detail later discussed 1. Crystals may be imperfect 2. The precise measurement of interplanar spacings requires internal standardization 3. The spacings of many minerals are so similar that they cannot be distinguished by electron diffraction methods 4. developed cleavages may eliminate diagnostic reflections 5. The measured inter- planar spacings are usually longer than expected because the zone axis of the diffracting plane may not be coincident with the electron beam 6. Polycrystalline or fibrous mineral aggregates may produce special diffraction effects which may be more difficult to analyse gold film was applied to a collodion 200 mesh electron microscope grid The gold film forms the grainy background of Fig 2. 2. The mineral particles were dispersed in water and a droplet of each suspension was applied to separ- ate grids 3. Electron diffraction was produced in the electron microscope operated at 80kV 4 The patterns for the minerals with the internal standard were recorded on photographic film 5. A working copy of the diffraction pattern was prepared by placing it in an enlarger and pro- jecting the spots on a piece of plain white typing paper The positions of the dots were marked directly with a pencil Alternately one could measure from the projected image but we found a copy to be more desirable for reproducible measurements 6. Uniform enlargement of the diffraction patterns was obtained by adjusting the projector until the gold diffraction ring with a spacing of 0.0695nm was exactly coincident with a previously drawn standard circle having a radius of 2.56in 7. The patterns were indexed if possible and R was measured with a scale having fifty divisions to the inch Repeated measurements indicate this error of 0.001 The technique has an camera constant including the enlargement step was 890.12 This value was used to simplify the data hand- ling and reduction It is of course a unique value for each laboratory 8. The interplanar spacings were measured and recorded as the number of units on the scale then divided by 890.12 the camera constant in this procedure to convert to nm units DISCUSSION OF RESULTS MATERIALS AND METHODS After a careful consideration of electron diffraction phenomena from a theoretical standpoint several minerals were examined in a Philips EM transmission electron micro- scope and actual diffraction patterns were obtained The minerals studied and some of their properties are listed in Table 1. These minerals were selected because they are representative of commonly occurring ones with length to width ratios greater than 3 and are good ones to test the applicability of the SAED method for the identification of micron particles The samples were prepared and examined in the following manner 1. A thin discontinuous The electron diffraction patterns of cummingtonite are characterized by two rows of dots Fig 2 In one direction the spacing between the dots is much smaller than the other and the pattern has a layered appearance The dots form a regular planar array that is almost rectangular instead of the 102 parallelogram as illustrated previously Because of this difference one must assume that the crystals are in some orientation other than the one described and it is necessary to try and determine the orientation before proceeding Cummingtonite has a prominent cleavage 110 which creates elongate crystals and enhances the possibility of preferred orientation The crystals most likely will come to rest on 110 or some other face in the form The axis nn ae JNJTALC0000708 wal aterm rt 3 Table which may have fibrous habit and therefore may dimensions dimensioaWyyu,n Jo 70 techniques techniques Ajteau peuHoy Mineral Crystal s[B1sA39Cleavge sjeisA19 pamojs : AproximateAproximate dimensions Steisk49 Compsiton Compsiton & woxy aiqy sfeisk19 ayruo3tN ay-9[paou Cleavge ( ayjdwiour ay-pau fq snoaqy 7 aniq AyD01q uasid selnqey Compsiton Compsiton you sted aiqrxay sadyentd Actinolte Cleavge 0.9 2 you 0.53Ca Mg5Ca MgC5a Mg5 SiO2 OH Avie se Long remaioe arenbs Long 1.81 snoiqty Actinolite 110110 perfectsyley 0.9 1.81 0.53 Iron rich ue ey Anthophylite perfect ( Anthopylite0} imperfect 0.53 ( Monclinc CrocidlteMSYIP imperfect Na Diopside 010 imperfct 0.52 Mg aq Hednbrgite { ]*tvesy-zey 0.9750.891 (HO)["S]8w glaucophne Grunerite 0.5250.530.525 Fe OH 2 0.526 110 Aeur Mg SiO22 needle 24) Syy) 0.9850.985 ep [ SW 3W) com on aojsz9y) Hornblende Hednbergitepi 3 jou Monoclinic Ecg 110 good <S0 0.9 geo 9259 0.53 Na Kvariable amphibole Fe with gc'o generaly = ayeunxoidySUOIDULIP Tremolite Monoclinic Monoclinic []J29 ig1 110 good lel 0.99 0.814 1.278 0.716 NaAl Mg SiO2 SiO2 OH ) Usual y Enstaite Jun e 660 010 good } 1.823 1.823 0.81 0.519 Mg SiO3 Eze" zg 4620 99h Stuby SiO3 =X go d SiO3 01 good Peo Monclinc poos fibrous parlelparalel to 1.46 1.46 0.9240.533 Mg3Si SO5i O5OHOH ywaysad{701} Jaqferd roling { {p71} 110 goodO10} {901} 0.96 0.96 1.830.53 Mg Mg Mg Fe Si,O2 Si,O2 { OH OH 2 snoaqy Long uraysd [eyshary oTuPoUCTuIPouN|y a1quioysG STUPOUC dtulPpou|oy1]uTPOUCWy SuIpocydUTPOUA dTurPoupySUNIL 1quoys0I9aquioysmp219 TUNE dTUPoUO SPouW] T, = = an ayuoZrwumny tances with canreduces corespond observeTdhe reciprocalpane ddeevveellooppeeddFig saungesiasudaAypy SUOISE]FOM aqHoskayr 7) SRESRE "86-238aesiezPOGo-tygo29%Sad535 SSBes oO SE oEE 2.4 onet os geeE'sSG" og 80 3 | ' 38;g^' opegisapeosCgUEGS 3 >? ae Paulson 822 ae PaulsonPaulson S-& Ble ~an ss >00 - OEE 25 U8 S& 2 ae ageEs cokes PES ees on BOG gee ESE FO aSBag Ses yefeu ass yB as oe e ESS mo ee Eso. PagrToas22a ot - RSEOGypgSoESPES easStegue22te ee SSC Sem oo er 2SsihHatos58be A GSE RAGzt2 EescesESE 80,58 58255ee#Ace aves ne me a G w3S ee82 US JNJTALC000070864 formed fibre plates crystals flexible up OT as aciular Del Ocurs roling Long at by Si 05 SMi,gO2 0.53 0.53 0.924 1.83 1.46 0.96 X to good 1} 02parlel good } 001 fibrous 110 Moncli Moncli Chrysotile Cumingtoe Practical Limitations of SAED for Identifying Mineral Particles 53 therefore will be parallel to the specimen support grid and the repeat distance along it should therefore appear in the diffraction pattern The other orientation represents some plane parallel to c and intersecting the a and b axis but its precise orientation cannot be determined The maximum spacing to be expected should not exceed about 2nm Fig 5 electron beam Cc Table 2. Results of cummingtonite spacing measurements by electron diffraction Average value for 001 0.5465nm Standard variation in 10 meas- urements Range of observed values Average 001 determined by ray diffraction Range of measured values in unidentified crystallographic direction 0.027nm 0.5267-0.6268nm 0.5215nm 1.0049 1.9074nm plane plane of viewing screen Fig 5. Preferential orientation of crystals on welldeveloped cleavages may make some electron diffrac- tion patterns difficult to interpret The departure from perpendicularity of a given reciprocal lattice plane to the electron beam causes the projected reciprocal lattice vector o to be different from ----- -- The above analysis is consistent with the observed patterns The closely spaced points represent the distances along the unidentified reciprocal lattice vector The other points correspond to multiples of 001. Because the points are in reciprocal space the shortest distances on the pattern correspond to the largest spacings in the crystal The prominent 110 cleavage helps in the interpretation of the pattern but drastically reduces the amount of useful information that can be obtained There is considerable variation in the spacings to be observed and the reciprocal lattice vectors are difficult to identify Because of these limitations only d 001 can be measured with any reliability The results of the measurement of d 001 for ten different specimens of cummingtonite are shown in Table 2. The average spacing is 0.5465nm as compared to 0.5215nm by ray methods This is approximately the amount of agreement that one would expect between the two methods but the deviation about the mean is very large for electron diffraction results For example the measured values ranged from 0.5267-0.6268nm This is because electron diffraction is a single crystal measurement for the most part and slight variations in the orientation of the crystal or mistakes in indexing the pattern will have pronounced effects In the unidentified reciprocal lattice direction the measured d values vary from 1.0049nm- 1.9074nm In some cases it is difficult to decide what to measure As suggested earlier these values have no practical significance for mineral identification because you would have to have prior knowledge of the minerals present in order to make the determination In other words the orientation of cummingtonite on its well developed cleavage restricts the number of actual spacing measurements made without assuming a priori and orientation that can be identification The indexing of the electron diffraction patterns can be difficult if systematic absences reduce the number of dots or if twinning crystal defects or polycrystalline particles increase the number of dots In Fig 2c the array of reciprocal lattice points has a pseudohexagonal ap- pearance because reflections with / = 1 are absent Twinning and rotation of the k reflecting plane about the electron beam produce a pattern which is typical of polycrystalline materials The cummingtonite pattern has begun to develop diffraction rings Unfortunately these rings are no more useful than the spots because you cannot identify the spacings they represent The number of rings produced by a polycrystalline aggregate may also be too great and closely spaced for effective measure- ment Electron diffraction methods are limited in the identification of cummingtonite In good single crystal patterns the appearance of the t JNJTALC00007086 1A etmeee nt () 54 R. E. Ferrell Jr. and G. G. Paulson reciprocal lattice points is characteristic in a is the repeat distance between qualitative sense because the reciprocal lattice rows of dots closely spaced points along * and this dimension is not unique of are readily observed Because of the enough to identify the minerals the good 110 cleavage quantitative measure- family or closely related mineral amphibole ments are restricted to d 001 and even this species vary considerably may Wollastonite albite and chrysolite specimens Many of the other minerals in this produced electron diffraction patterns which study set were different from cummingtonite The wol- possess a good or perfect cleavage parallel to the lastonite reciprocal net was almost a with crystallographic axis The symmetry may be square edges of 0.7418nm and 0.7120nm This is slightly different but the cleavage forms are very close to what would be expected from a crystal usually 110 or 210 Therefore the same of wollastonite and the spacings represent * general prismatic forms and orientations would be expected and only axis spacings of the same and * respectively One of the albite patterns formed a magnitude as that of cummingtonite should be partial parallelogram and spacings of 0.8092nm and 0.7418nm were measured and observed Good complete single crystal electron identified diffraction patterns of these should be identical tentatively as * and * The results were not this good for all the specimens and within the limits of the technique therefore most of the incomplete patterns would Amosite anthophyllite hornblende crocidolite hedenbergite actinolite diopside tremo- not have been identified as wollastonite or albite or even lite and grunerite have the requisite cleavage The axis dimensions also fall within the distinguished from cumming- tonite Therefore even in cases where the minerals being examined have very different narrow limits 0.5251nm to 0.5348nm This range is narrower than the one observed for the crystal structures the inability to obtain a complete diffraction from all cummingtonite samples Enstatite and hypers- thene are also likely candidates for the above pattern particles makes specific identification of each small crystal impossible Thickness variations and category but their patterns should exhibit a crystal defects are thought to be the major truly orthogonal array causes of this problem As predicted most of the SAED patterns from One of the unique diffraction the above minerals were like those of cumming- observed chrysotile tonite When good single crystal patterns were tic was that of chrysotile The pat erns characteris- obtained they could be overlaid and the dis- rolled tube crystal morphology of this mineral creates a situation where all possible tances between the points along * were similar reciprocal lattice planes parallel to the fibre axis identical within experimental error and the are in a position to satisfy the condition for spacings in the other directions were variable diffraction As a result the pattern is a series of and difficult to relate to a specific crystallo- streaks perpendicular to the fibre axis Occasion- graphic direction Four hornblende diffraction ally spots may be superimposed on the streaks patterns provide good examples to illustrate the or small segments of arcs from fibres parallel to extreme varia- the electron beam are seen One does not find a bility in spacing measurements caused by the great deal of difference between the preferred orientation of crystallites Circles and crysotile were drawn through the points on incomplete diffraction patterns in order to determine how cummingtonite spacings because the axis about which the rolling takes place is the axis and its length is 0.534nm The streaked diffrac- many discrete reflections were present The tion lines are number of spacings observed varied from seven essentially equivalent to the pattern produced in a rotation ray camera and are to eighteen on the different specimens These quite distinctive for identification of a patterns were then placed on light table and the coincident circles marked on each The However all chrysotile chrysotile particles did not yield the same results and there are other maximum agreement between circles and therefore spacings was four In most cases the which may exist as rolled tubes i.e. halloysite meme} circles corresponding to a multiple of d 001 were the ones which overlapped perfectly This ~~ ldmake it very clear that the only consistent parameter to be determined for these the electron diffraction technique SUMMARY The results obtained during the study support the contentions of others Brindley and de Kimpe 1961 as well as Ross and Christ 1958 that electron diffra ray diffraction polyminerallic sam to identify single cr problems encount patterns Electron diffract obtain and only th crystal patterns | orientations can be interpreted quar cleavages in many the crystals from a tion and restrict s In the amphibole only repeat distan d 001 could be n there is more varia samples of the same separate species Other minerals su chrys tile may be boles in some case shape and the qu diffraction pattern ing characteristics patterns are avail circles from polyc good for identificat In conclusion e are difficult to inter not be recommende tion of mineral par It would be very dif tion results to dete amorphous an amp In a qualitative sen whether a fibre was | The difficulties in especially importan ing to identify the so JNJTALC000070866 reciprocal lattice on is not unique of the amphibole species solite specimens patterns which onite The wol sta square with am This is very d from a crystal gs represent * e albite patterns and spacings of measured and 1 * The results . specimens and e patterns would wollastonite or from cumming- ases where the | e very different ty to obtain a rom all particles of each small variations and o be the major Fraction patterns The characterisphology of this here all possible ] to the fibre axis ne condition for tern is a series of re axis Occasioned on the streaks fibres parallel to e does not find a een the crysotile because the axis place is the axis streaked diffracent to the pattern camera and are ion of chrysotile les did not yield e other minerals i.e. halloysite the study support Brindley and de and Christ 1958 Practical Limitations of SAED for Identifying Mineral Particles . 55 that electron diffraction is not as amenable as ray diffraction to mineral identification in polyminerallic samples It is virtually impossible crystal particles because imposible the to identify single problems encountered in interpreting SAED patterns _ Electron diffraction patterns are difficult to obtain and only those relatively complete single crystal patterns from specimens in special orientations can be unambiguously indexed and interpreted quantitatively unambiguosly developed cleavages in many monoclinic minerals prevent the crystals from assuming this special orienta- tion and restrict severely the methods utility distances the amphiboles only repeat and pyroxenes examined of 0.53nm equivalent to unfortunately d 001 could be measured and there is more variation in the measurement of samples of the same species than there is among separate species separate minerals such as wollastonite albite and chrys tile may be distinguished from amphiboles in some cases The cell dimensions its shape and the qualitative appearance of the diffraction pattern may be used as identify- ing characteristics when relatively complete patterns are available Rings or diffraction circles from polycrystalline particles are not good for identification purposes In conclusion electron diffraction analyses are difficult to interpret and the technique can- not be recommended for quantitative identifica- tion of mineral particles in composite samples It would be very difficult to use electron diffrac- tion results to determine whether a fibre was amorphous an amphibole or a amphibole In a qualitative sense it may be possible to tell a whether fibre was chrysotile or another mineral The difficulties inherent in the method are especially important to consider when attempt- ing to identify the source of particulate pollutants because electron diffraction techniques have to be supplemented with other information such as probe determinations in order to obtain unambiguous information It should not be used as the sole means for establishing the identity of mineral particles in complex samples In this paper we have attempted to call attention to some of the negative aspects or difficulties associated with the SAED technique because there are many electron microscopists entering this field who might not be aware of its limitations There are ways to overcome many of the problems which we have discussed and we cannot deny that electron diffraction is a very powerful analytical technique However the difficulties do exist and are very prevalent when dealing with natural samples composed of many mineral varieties Practical considerations such as time and effort do restrict the utility of the SAED technique REFERENCES Alderson R. H. and Halliday J. S. 1965. Electron diffraction In Techniques for Electron Microscopy Kay D. H. ed Blackwell Oxford England 478-524 Brindley G. W. and De Kimpe C. 1961. Identification of clay minerals by single crystal electron diffraction Amer Min 46 1005 1016 Clark R. L. and Rudd C. O. 1974. Transmission electron microscopy standards for asbestos Micron 5 83 88 G^...ven N. 1974. Factors affecting selected area electron diffraction patterns of micas Clays and Clay Minerals 22 97 106 McConnellJ. D. C. 1967. Electron microscopy and electron diffraction In Physical Methods in Determinative Mineralogy Zussman J. ed Academic Press London 335 370 Ross M. and Christ C. L. 1958. Mineralogical applications of electron diffraction I. Theory and Techniques Amer Min 46 1157-1178 nr i JNJTALC0000708