ANALYTICAL ELECTRON MICROSCOPY IN CLAYS AND OTHER ...clays.org/journal/archive/volume 46/46-3-301.pdf · 3 Electron Microscopy Unit, Australian National University, Canberra, ACT

Clays and Clay Minerals, Vol. 46, No. 3, 301-316, 1998.

ANALYTICAL ELECTRON MICROSCOPY IN CLAYS A N D OTHER PHYLLOSILICATES: LOSS OF ELEMENTS FROM A 90-nm STATIONARY

B E A M OF 300-keV ELECTRONS

CHI M A , t t JOHN D. FITZGERALD, 2 RICHARD A. EGGLETON 1 AND DAVID J. LLEWELLYN 3

Cooperative Research Center for Landscape Evolution and Mineral Exploration, Department of Geology, Australian National University, Canberra, ACT 0200, Australia

2 Research School of Earth Sciences, Australian National University, Canberra, ACT 0200, Australia

3 Electron Microscopy Unit, Australian National University, Canberra, ACT 0200, Australia

Abstract--Diffusion of alkali and low-atomic-number elements during the microbeam analysis of some silicates by analytical electron microscopy (AEM) has been known for some time. Our repeated analyses at 300 kV of kaolinite, halloysite, smectite, biotite, muscovite and pyrophyllite, however, showed differential loss (relative to Si) of not only alkali elements (such as K, Na, Mg) and low-atomic-number elements (such as A1) but also higher-atomic-number elements (such as Fe, Ti). For AEM of these phyUosilicates, a Philips EM430/EDAX facility with a tungsten filament was used to provide a current of 0.3 nA in a stationary beam of nominal diameter 90 nm. The loss of A1 in kaolin minerals during analysis is particularly severe. Kaolin crystals can be damaged by the electron irradiation over several seconds, making it the most sensitive clay to the electron beam; in general, relative phyllosilicate stabilities are kaolin < smectite < pyrophyllite < mica. A clear dependence of element loss on crystallographic orientation has been observed for layer silicates in our study; a greater element loss occurred when the plane of the specimen foil was perpendicular to the basal planes of the phyliosilicate crystals than when the foil was parallel to the basal planes. Lower beam current, larger beam diameter and thicker specimens all reduce the loss of elements. The initial stage of irradiation produces highest rates of element loss and the rate of loss can be fitted by an exponential decay law. The analyses at low temperature of phyllosilicates showed that element loss remains serious in our analytical conditions. Since the element loss appears to be instrument- and method-dependent, one should use closely related, well-characterized phyllosilicates as compositional standards to calibrate any AEM instrument that is to be used to analyze unknown phyllosilicates, and the standards and unknowns should be analyzed under identical conditions.

Key Words----Analytical Electron Microscopy, Clay Minerals, Diffusion, Element Loss, Low Tempera- ture, Phyllosilicates, Sputtering, Stationary Beam.

I N T R O D U C T I O N

Preferent ia l loss of alkal i e l emen t s and low-a tomic- n u m b e r e lements , such as K, Na and AI, by A E M gives rise to anoma lous ly low coun t rates for those e l emen t s and m a y cause s ignif icant p rob lems in the analys is of geologica l spec imens ( M a c k i n n o n and Ka- ser 1987; van der P lu i jm et al. 1988; Peacor 1992). In a c o n v e n t i o n a l t r a n s m i s s i o n e l e c t r o n m i c r o s c o p e ( T E M ) at acce lera t ing vol tages o f 1 0 0 - 3 0 0 kV, elec- t ron b e a m d a m a g e for the c o m m o n clays is ve ry rapid ( < 20 s). Wi th a focused p robe for X- ray analyses , the d a m a g e rate is exace rba ted by a h igh cur ren t dens i ty and usual ly resul ts in the loss of some e lement s dur ing X- ray co l lec t ion and consequen t analy t ica l errors. Losses can of ten be avo ided by sacrif icing some spatial reso lu t ion by b e a m scann ing ( scann ing t ransmis- s ion e lec t ron mic roscopy [STEM], Peacor 1992) or b e a m defocusing.

In our s tudy us ing a s ta t ionary b e a m of 300-keV electrons, the loss o f e l ement s in clay minera l s was

t Plesent adress: Division of Geological and Planetary Sci- ences, California Institute of Technology, Pasadena, Califor- nia 91125.

found to be m u c h more ser ious than in o ther sheet s i l icates ana lyzed in the same way. The loss o f A1 in kao l in minera l s is par t icular ly severe, the s t ructure o f kao l in be ing d a m a g e d b y our e lec t ron b e a m wi th in 10 s. A m o n g clays, kao l in minera l s are mos t sens i t ive to the e lec t ron b e a m (Rober t son and Egg le ton 1991; M a and Egg le ton 1994; M a 1996).

In this paper, the p rob l ems associa ted wi th A E M analys is in T E M m o d e for A1, K, Mg, Ca and Fe in kaol in and o ther phyl los i l ica tes are p resen ted and compared . In particular, the var ia t ion of resul ts as a func t ion o f the e lec t ron b e a m diameter , the spec imen th ickness and the c rys ta l lograph ic o r ien ta t ion o f the sample p lane are discussed. A prefer red p rocedure for ob ta in ing re l iable analyses is also presented .

E X P E R I M E N T A L

Ana lyses were ob ta ined in a Phi l ips E M 4 3 0 T E M at an accelera t ing vol tage o f 300 k V wi th an a t tached sol id-state , low- takeof f -ang le (15 ~ ) ene rgy-d i spe r s ive X- ray (EDS) detector.

For analysis , the sample was t i l ted 30 ~ toward the X- ray detector. A tungs ten f i lament was used for all spectra co l lec t ion and p rov ided 1.95 x 109 e lec t rons

Copyright �9 1998, The Clay Minerals Society 301

302 Ma et al. Clays and Clay Minerals

<

0.82

0.8

0.78

0.76

0.74

0.72

0.7

0.68 0

i

m

B. - I(AI)/I(Si) = 0.779 + 0.029*e^(-t/51.9), R=0.985

- _ ~ 0 " 6 9 7 + 0 " 1 0 3 * e ^ ( ' t / 1 1 1 ) ,R=0.999

- 3 ~ 8 8 o cps

i i i i i i

50 100 150 200 250 300 350 Seconds

Figure 1. An example (kaolinite) showing the decreasing ratio of A1 versus Si X-rays detected for increasing live time of analysis for constant probe current and sample position. The error bar for each analysis point has been calculated using simple square-root counting statistics for total integrated counts of both peak and background in the energy region of interest. The exponential correlation lines (fitted parameters and correlation coefficient shown) are well within the uncertainty range.

per second (a beam current o f 0.3 nA) at the sample for a nominal beam diameter of 90 nm. In addition, some data at a defocused beam diameter of 350 nm were obtained for comparison.

The clay minerals kaolinite, halloysite and smectite, and other phyllosi l icates pyrophylli te, muscovi te and biotite were studied. Their chemical composi t ions and sample locations are g iven in Table 1. Samples were prepared in 1 of 3 ways: 1) particles dispersed on a carbon film supported by a copper grid so most (001) planes o f platy phyllosi l icates lay parallel to the film, 2) sectioning with a d iamond micro tome to produce slices normal to the (001) plane or 3) ion mill ing, which al lows the study of any specimen thickness effect on the analysis of clay minerals by AEM. The role of the crystal lographic orientation in the analytical procedure was examined using platy kaolinite and biotite.

Cl i f f and Lor imer (1975) established the fol lowing equat ion for " th in - f i lm" A E M in which atomic number and X-ray absorption and f luorescence effects are insignificant:

Ca/C b = kab(IJIb) [1]

where Ca and Cb are the weight concentrat ions of elements a and b, la and I b are the intensities of emission lines o f those elements, and /cab is a proportionali ty factor that is a function of parameters concerned with beam/spec imen interaction and detector efficiency.

In an alternative formulation:

C'a/C'b = k'ab(IJlb) [2]

where C'a and C' b are the atomic concentrat ions of

e lements a and b. Ideally, k'ab is a constant indepen- dent of measur ing t ime and can be routinely calibrated using standards o f known composi t ion for the accelerating voltage of interest.

Because silicates are the most abundant group of minerals, intensity for X-ray emissions o f a g iven element is ratioed to that of Si (Lorimer and Cl i f f 1976; Peacor 1992). Thus for AI:

k ' A I , S i = (C'Al]C'si)(Isi[IAl) [3]

Ratios of intensity of AI, K, Fe, Mg, Ti relat ive to that of Si were measured in this study and plotted against the l ive t ime (that is, counting t ime corrected for detector dead time). In a similar study on clay minerals, Mackinnon and Kaser (1987) min imized uncertainty by averaging data f rom about 15 separate analyses. In our study, we were unable to average due to slight chemical and structural heterogeneity. Figure 1 shows the uncertainty (~/n counting statistics for total integrated peak count and for background) associated with our single analyses, and demonstrates that the curve fitted to the decay over t ime (described in the Discussion section) locates well within the error bounds. Since this was true for all our trends, error bars are not drawn on Figures 2 to 11.

RESULTS

Kaolinite and Haltoysite

Changes in AI/Si intensity ratios were measured as a function of counting t ime (Figure 2) at 3 locations on an ion-mil led kaolinite specimen. Different specimen thicknesses produced different count rates (that is, 860, 1600, 3500 counts per second [cps]) f rom a

Vol. 46, No. 3, 1998 AEM: Element loss in clays and other phyllosilicates 303

0.82

0.8

0.78

~ 0.76

0.74

0.72

0.7

0.68

I(AI)/l(8i) = 0.779 + 0.029*e~(-t/51.9), R=0.985

~, ~ -o_. o Rate=3500eps

l ( ~ S i ) = 0.720 + 0.105*eA(-t/89.5), R-0.997

I(AI)/l(S~eA(-t/111), R = 0 . 9 ~ 8 6 0 c p s

I I I I I I

0 50 100 150 200 250 300 350

Seconds

Figure 2. AI/Si intensity ratio for increasing live time of analysis for constant probe current and stationary beam at 3 different points of an ion-milled kaolinite specimen. The count rate is related to the thickness. The sample is highly ordered kaolinite (Kao I), specimen plane perpendicular to (001).

constant incident beam current. The highest rate cor- responds to the thickest area (but with some spurious counts arising from the copper support grid). Figure 2 demonstrates that the relative loss of A1 is greater for thinner areas.

Figure 3 shows the change in A1/Si, K/Si and Fe/Si intensity ratios with increasing counting time in a microtomed kaolinite specimen. It is evident that inten- sifies for K and Fe decrease much more than A1 (67% for K, 14% for Fe and 7% for AI). The presence of K implies that the kaolinite contains interstratified mica layers (Ma 1996).

Figure 4 shows the differential trend of AI/Si intensity ratio with changing irradiation area for a constant beam current of 0.3 nA focused to 90-nm diameter (full width, one-tenth maximum) or spread to 350-nm diameter. The loss of A1 decreases when analyzing the larger area. As clearly shown in Figure 5, the crystallographic orientation of the specimen plane does affect the loss of elements in kaolinite. Aluminium loss is higher for the specimen cut perpendicular to (001).

Tubular halloysite also undergoes differential A1 loss during analysis (Figure 6).

Smectite

Two types of smectite were studied. For saponite when the specimen plane was parallel to (001), the Mg/Si, Ca/Si and Fe/Si intensity ratios decreased with the counting time whereas the A1/Si ratio changed little. However, when the specimen plane was perpendicular to (001), the A1/Si ratio decreased also (Figure 7a), although direct comparison is complicated by the analyzed domains having slightly different compositions (which are Sm I and Sm II as listed in Table 1).

The lower plot of Figure 7a shows the effect of thickness: the loss of Mg is greater in the thinner area. Wyoming montmorillonite (Sm III) shows only a slight differential element loss (Figure 7b).

Muscovite

Figure 8 shows that the K loss in muscovite is larger than those for A1 and Fe when the (001) planes are placed perpendicular to the beam. The beam damage rate monitored from SAED patterns is much lower than that for kaolin.

Biotite

The loss of elements in biotite is orientation-dependent. Figure 9 shows that the K loss is significant in both orientations, whereas the loss of Fe, A1, (Ti and Mg) is only apparent when the specimen plane is perpendicular to (001).

Pyrophyllite

The loss of AI is evident in pyrophyllite (Figure 10), although its structure is not damaged as rapidly as that of kaolinite.

Analysis with a Cold Stage

A Gatan 636N liquid-N2 double-tilting holder was used for low-temperature analysis. Before analyses, the holder was cooled by liquid nitrogen to a steady temperature of - 170 ~ measured by a thermocouple in the solid tip of the side-entry holder. From other studies (such as Mackinnon and Kaser 1987), much lower element loss was expected during these X-ray analyses of kaolinite, smectite, muscovite and biotite. The representative results are shown in Figure 11. The


0.8

0.7

0.6

0.5

0.4

0.3

,-, 0.2

~ , ~ r~'" ~ 0.1

v o o o o

Rate = 1750 cps

I(Al)/l(Si) = 0 .712 + 0.068*eA(-t/62.1), R=0 .947

I(Fe)/l(Si) = 0.044 + 0.0077*e^(-t /94.0), R=0.931

I(K)/I(Si) = 0 .0064 + 0.0171*e^(-t /58.5), R=0 .993

~ - ~ ~ ~ ~ O - - ~ - - ~ - i ~ - - - - @ - - - - @

G--El , -E~---FI m R ,m [] m

50 100 150 200 250 300 Seconds

350

Figure 3. A1/Si, K/Si and Fe/Si intensity ratios for increasing live time of analysis for constant probe current and sample position in a microtomed specimen of medium-ordered kaolinite (Kao III), specimen plane perpendicular to (001).

loss rates o f K, A1, Mg, Fe and Ca in these minera l s r e m a i n s imi lar to those m e a s u r e d at r o o m tempera ture . For Pi t tong kaol in i te (Kao I) (Figure 1 la) , the loss of A1 may even h a v e b e e n h igher at low tempera ture , t h o u g h the lower coun t rate indica tes tha t a th inner reg ion was be ing analyzed.

The t empera tu re of the ana lyzed reg ion can obvi- ous ly n e v e r be measured , bu t an a t tempt was made to get a be t t e r ind ica t ion than p rov ided by the the rmo- couple in the t ip o f the 636N spec imen holder. A Type K the rmocoup le (n i cke l - ch romium/n i cke l - a lumin ium)

wi th smal l the rmal mass was m a d e f rom 0 . 2 - m m di- amete r wires and the j unc t i on soldered to a 3 - ram Cu disc of th ickness 0.15 mm. This disc cou ld be sc rewed into the 636N ho lder jus t l ike the s tandard ho ley-C or ion-mi l led spec imens . The the rmocoup le /d i sc was cal- ibra ted c rude ly by measu r ing the e lec t romagne t ic field ( e m 0 genera ted dur ing i m m e r s i o n in 1) l iquid N2, 2) a ba th of e thanol + sol id CO2 and 3) a ba th of wate r + ice. The ca l ibra ted t he rmocoup l e was screwed in to the 636N holder, the ho lder inser ted in to a h igh vac- u u m c h a m b e r (10 -5 torr) and the t he rmocoup le e m f

0.84

0.82

0.8

-- 0.78

~ 0.76

0.74

0.72

0.7

0.68

O Analysis spot 0f350 nm in diameter

I(AI)/I(Si) = 0 .778 + 0.016*e^(-t /33.4), R=0 .608

[] [] [] o - - [] --'[3

o I(AI)/I(Si) = 0 .712 + 0.110*e^(-t /82.8), R=0.981

I I I I I I

0 50 100 150 200 250 300 350

Seconds

Figure 4. AI/Si intensity ratio for increasing live time of analysis for constant probe current and sample but for 2 different beam diameters in a microtomed specimen of highly ordered kaolinite (Kao II), specimen plane perpendicular to (00l).

Vol. 46, No. 3, 1998 AEM: Element loss in clays and other phyllosilicates

0.82 . . . .

305

0.8

�9 - 0.78

0.76

0.74

0 [ ]

0 [ ]

0

[] []

o o

[] [] [] []

s p e c i m e n / / ( 0 0 1 )

O

O specimen 2- (oo 1) o o

o 0.72 ' ' ' ~ ' '

0 50 100 150 200 250 300 350

Seconds

Figure 5. AI/Si intensity ratio for increasing live time of analysis for constant probe current and sample position but with 2 different orientations of specimen plane in highly ordered kaolinite (Kao I).

m e a s u r e d at r o o m tempera tu re to con f i rm that extra f eed th rough connec t ions did not affect vol tages . Fi- nally, w i th the 636N dewar filled wi th l iquid n i t rogen, the t empera tu re of the spec imen cup was m e a s u r e d to be - 1 3 0 • 5 ~ even t h o u g h the 6 3 6 N cont ro l le r in- d ica ted - 1 7 0 ~ Clearly, samples cons i s t ing of iso- la ted phyl los i l ica te c rys ta ls suppor ted on th in C films and Cu grids are un l ike ly to be coo led even to - 1 3 0 ~ t h o u g h the th icker and con t inuous ion-mi l l ed specimens could be. Never the less , all samples mus t be sig- ni f icant ly coo led in the 6 3 6 N holder. For s impl ic i ty in

the fo l lowing, we refer to coo led t empera tu res o f - 1 7 0 ~ even t h o u g h we k n o w that the spec imen t empera tu re is at leas t 40 ~ above this indicat ion.

D I S C U S S I O N

Loss o f Al l E lemen t s

E lec t ron i r radiat ion, w h i c h turns c lay minera l s amorphous b y d a m a g i n g the i r c rys ta l s tructures, is also r e spons ib le for loss o f e l emen t s dur ing A E M analysis . F igure 12 is a typical b r igh t field image

0.75

0.74

" = . . - = . 9 8 1

.'~. 0.73

g ~ 0.72

0.71

0.7 , , , ,

0 50 100 150 200 250

Seconds

Figure 6. Al/Si intensity ratio for increasing time of analysis for constant probe current and sample position in a dispersed halloysitc specimen. The sample is tubular halloysite with tubes parallel to the plane of the Cu grid.


"" 0.35

0.3

0.25 ,._., <___ 0.2

r./3 "-~ 0.15

~.) 0.1

0.05

o

Sm I: s p e c i m e n / / b e a m Rate= 1000cps

0 �9

0 �9

0 0

[] [] [] [] [] [] []

O <> r O <) 0

X X X ,~ X X I I

50 100 150

O

[]

<> <> <>

200 250 300 350

Sm II: s p e c i m e n k (001)

0.25

0.2

0.15

0.1

0.05

0

8

0

8 8 ~ m B B

0 <>

X X X

r (>

[] O

Rate=960eps

I(A1)/I(Si) [] []

o o I (Mg)/ I (Si )

<> C,

X X X X X X X

I I I I I I

50 100 150 200 250 300 350

0.38

0.36

0.34

0.32

0.3

0.28

0.26

0.24

0.22 0

Sm I: s p e c i m e n / / ( 0 0 1 )

[ ] [ ] [ ] [ ]

[ ] o [ ]

[ ]

o [ ] Rate = 4200eps

0 [ ]

0 [ ]

0

0 o Rate = t O00cps

o o 9 50 100 150 200 250 300 350

Seconds

Figure 7. Mg/Si, A1/Si, Ca/Si and Fe/Si intensity ratios for increasing time of analysis for constant probe current and sample position in dispersed smectite specimens, a) Saponite (Sm I & S m II).


O Q

<

0.4

0.35

0.3

0.25

0.2

0.15

%" 0.1

0.05

---" 0

Sm HI: specimen//(001) Rate=1200cps

i i J

5~0 ' ~ ~ ~ 100 150 200 250 300 Seconds

Figure 7. Continued. b) Montmorillonite.

350

showing a hole "dri l led" in 250 s by an electron beam within a kaolinite particle. Most original components within the hole (Si, A1, O) are presumed to have been sputtered away in the form of ions, atoms and com- plexes though some atoms/ions could diffuse into the surrounding region. For the mechanism of element loss, the "diffusion" accompanying irradiation by 100-keV electrons described by van der Pluijm et al. (1988) may not be an adequate description under our stationary 300-kV beam, since the loss of elements involves sputtering and perhaps other phenomena.

One should be aware that during analysis all elements are differentially lost to some degree, especially when the sample is sputtered away (Figure 12). In silicates, the loss of Si seems to be lowest compared with the loss of other elements; therefore, the change of i/ Si intensity ratios (i: elements other than Si) will be reflected in chemical composition calculated via the Cliff-Lorimer approach.

In the AEM analyses of Ahn et al. (1986) at 100 kV, only the alkali element concentrations of micas changed significantly during exposure to the electron beam at room temperature. The effect was particularly marked in Na- and K-muscovites and gave rise to characteristic mottled or fissured textures. Similar mi- crostructures were shown by Knipe (1979) and Peacor (1992). Loss of low-atomic-number (low-Z) elements by AEM at 200 kV for some clays (kaolinite and smectite) was reported by Mackinnon and Kaser (1987). In our study, we found that even Fe (and possibly Ti) concentrations also change significantly in some phyllosilicates under certain conditions at 300 kV (see Figure 9). It seems that element loss increases with accelerating voltage, though it is impossible to

make a quantitative comparison since the other experimental conditions used in this study and those pre- viously are different. However, it is evident that sputtering damage is significant using a stationary beam at 300 kV, as well as diffusion. This is of importance in our understanding of "element loss".

Selection of larger analysis areas also reduces the loss of elements for stationary probes. For example, in kaolinite (see Figure 4), the A1/Si ratio changes only slightly throughout the analysis using the defocused beam of 350-nm diameter, but clearly decreases for the focused beam 90 nm in diameter. Therefore, to use analysis areas as large as possible is particularly significant and important for clay minerals. When the beam area is set to totally enclose the individual analyzed grain, the element loss still exists in some of our AEM analyses on isolated kaolin particles. This might imply that diffusion is not the only primary loss mechanism under our experimental conditions.

Crystallographic Orientation Effect

The element loss from kaolinite studied by Mack- innon and Kaser (1987) was based on samples dispersed over holey carbon film on Be grids. In our study, both microtomed sections and ion-milled specimens were used in addition to dispersed samples for studying any crystallographic orientation effect. Fig- ure 5 shows that, in kaolinite, loss of A1 is dependent on crystal orientation; greater A1 loss occurs when the specimen plane is perpendicular to (001) compared to when the specimen is parallel to (001).

K is lost similarly during analysis of mica specimens, for specimen planes both parallel and perpendicular to (001). In biotite, significant losses of Fe and

3 0 8 M a et al. Clays and Clay Minerals

Tab le I. C h e m i c a l c o m p o s i t i o n o f kao l in i t e , h a l l o y s i t e , smec t i t e , b io t i te , m u s c o v i t e a n d p y r o p h y l l i t e .

t,:~ K a o 1 Kao I[ Kao III Hal l S m l S m II S m III M u s B i o Pyr

SlOe 4 7 . 4 9 4 8 . 9 4 4 5 . 7 0 48 .31 5 0 . 9 2 5 2 . 3 9 6 3 . 2 4 4 5 . 8 5 3 7 . 1 3 6 6 . 6 0 A1203 3 8 . 3 4 3 6 . 7 3 3 8 . 3 7 3 6 . 9 2 13 .62 2 0 . 8 5 2 3 . 4 4 3 2 . 3 5 17 .93 2 8 . 8 4 TiO2 0 . 0 0 0 . 0 0 - - - - 0 .03 - - - - - - 1 .62 - - F e O 0 .21 0 . 3 4 2 .54 0 .78 0 . 5 0 2 . 1 0 3 .98 3 .40 2 0 . 7 4 - - M g O . . . . 2 5 . 8 5 11 .64 2 .30 - - 7 . 1 9 - - M n O . . . . . . . . 0 . 37 C a O . . . . 2 . 76 2 .08 . . . . N a 2 0 . . . . . . 2 . 30 0 . 7 6 - - - - K 2 0 . . . . . . . 10 .83 10 .33 - - Tota l 8 6 . 0 3 8 6 . 0 0 8 6 . 6 0 8 6 . 0 0 9 3 . 6 7 8 9 . 0 5 9 5 . 7 5 9 3 . 1 9 95 .31 9 5 . 4 4

O , O H w 7 7 7 7 11 11 11 11 11 11 Si 1.97 2 . 1 0 1.98 2 .08 3 .34 3 .54 3 .91 3 .16 2 .86 3 .98 AI ( IV) 0 .03 - - 0 . 02 - - 0 . 66 0 . 4 6 0 . 0 9 0 . 8 4 1 .14 0 . 0 2 AI (VI ) 1.99 1 .86 1.94 1.87 0 . 3 9 1 .20 1.62 1 .80 0 .48 2 .01 Ti . . . . . . . 0 . 0 0 0 .08 - - F e 0 .01 0 . 0 2 0 . 0 9 0 . 0 4 0 .03 0 . 1 2 0 .21 0 . 1 9 1.35 - - M g . . . . 2 . 52 1.17 0 . 2 4 - - 0 .83 - - M n . . . . . . . . . . C a . . . . 0 . 19 0 .15 . . . . N a . . . . . . 0 .28 0 .11 - - - - K . . . . . . . 0 . 9 6 0 .83 - -

T h e s e E D S m e a s u r e m e n t s w e r e m a d e u s i n g a J E O L 6 4 0 0 S E M w i t h a d e f o c u s e d b e a m o f 1 5 - k V e l ec t rons . } K a o = kao l in i t e ; Ha l l = h a l l o y s i t e ; S m = s m e c t i t e ; M u s = m u s c o v i t e ; B i o = b io t i te ; P y r = p y r o p h y l l i t e .

K a o I: P O N G 4 , P i t t o n g , Vie. ; K a o II: E2 , C M S ; K a o III: G 1 5 0 0 , W e i p a , Q ld . ; Hal l : Mt . M o r g a n , Q ld . ; S m I & S m II: s a p o n i t e , C o o m a , N S W ; S m III: m o n t m o r i l l o n i t e , W y o m i n g , U S A ; M u s : P h i l a d e l p h i a , U S A ; Bio : H a r t s R a n g e , NT; Pyr : f r o m A N U G e o l o g y D e p a r t m e n t M u s e u m .

w N u m b e r s o f e q u i v a l e n t 2 + a n i o n s [ b a s e d o n O5(OH)4 f o r K a o a n d Hal l , a n d OI0(OH)2 f o r Sin , M u s , B i o a n d Py r ] u s e d as the b a s i s f o r c a l c u l a t i n g the m i n e r a l f o r m u l a e e x p r e s s e d as c a t i o n n u m b e r s b e l o w .

0.7

0.6

0.5

0.4

0.3

0.2

b 0.1

Muscovite: specimen//(001)

G O G O O C~ O 0 0 0

I I I I

50 100 150 200

S e c o n d s

R a t e = 5 0 0 e p s

I q

250 300 350

F i g u r e 8. K/S i , A1/Si a n d Fe /S i i n t ens i t y r a t ios f o r i n c r e a s i n g t i m e o f a n a l y s i s fo r c o n s t a n t p r o b e c u r r e n t a n d s a m p l e p o s i t i o n in a d i s p e r s e d m u s c o v i t e s p e c i m e n , s p e c i m e n p l a n e pa ra l l e l to (001) .


(a)

~ 0.7

~ 0.6

0.5

r ~ 0.4

~ '0 .3

0.2 Gr

Biotite : spec imen/ / (001) J

~o ~ u 0 0 0

Rate = 2600 cps i i

U U U U

I(Al)/I(Si) -- -- ~ - - -- o ~ .- ;~ r :~.- - - ) tT-- :g-- :" ~ - Y ~ - - - - ' ~ . . . . ~ : - - 2"7 7 ~

I(Mg)/I(Si)

[] [] [] rl D D [] [] [] I I / I I I

5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0

Seconds

350

(b)

0.7 O

0.6

0.5

~ 0 . 4

0.3

0.2

~ 0.1

Biotite: specimen_.J._ (001) Rate -- 2600 cps

)~ - X ' -X- -X- - ) ~ - - 9 ( - - - )~ - - - -X . . . . . X . . . . . X

~ ~ ~ ~ - - ~ - - ~ _ _ ~ __ _ _ ~ _ _ _ _

I(A1)/I(Si) [] [ ] [ ] D 0 0 0 C

I(Mg)/I(Si)

+ - + . -+- - + - + . - . + . . . + . . . -+. . . . . + . . . . . +

I I I I I I

5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0

Seconds

Figure 9. K/Si, A1/Si and Fe/Si intensity ratios for increasing time of analysis for constant probe current and sample position in a dispersed biotite specimen, a) specimen plane parallel to (001), b) specimen plane perpendicular to (001). Rate = 2600 cps.

A1 occur only in specimens cut perpendicular to (001) (see Figure 9). This indicates some dependence upon the crystal structure.

We must highl ight that the orientat ion effect cannot have its origins in channe l l ing-enhanced X-ray emission. Whi l e this ef fec t has been wel l known (Wil- l iams and Carter 1996), it can be d iscounted as a major phenomenon affect ing our measurements . This is because for analysis all samples were t i l ted 30 ~

about the sample rod, an axis with a r andom orientation re la t ive to the individual crysta l ' s axes. Thus, for samples cut paral lel to (001), the b e a m direct ion during analysis is 30 ~ r e m o v e d f rom [001], but in an unspecif ied direct ion. Likewise , for samples cut perpendicular to (001), after t i l t ing r andomly by 30 ~ only v e r y few would show strong (OOl) diffract ion. Therefore , in no crystal wou ld the e lec t ron b e a m have been paral le l to a major zone axis during anal-

310

0.42

Ma et al.

Pyrophyllite: specimen//(001)

Clays and Clay Minerals

0.41

~ 0.4 o ~ , , ~ O

0 . 3 9 ~ = 0.332 + 0.073*e (̂-t/361), R=0.971

0.38 O

0.37

0.36 , , , , , 50 100 150 200 250 300 350

Seconds

Figure 10. A1/Si intensity ratio for increasing time of analysis for constant probe current and sample position in a dispersed pyrophyllite specimen, specimen plane parallel to (001).

ysis, and only a few would have been oriented for strong Bragg diffraction.

Under reproduced TEM conditions including similar specimen thickness, a crystal with its specimen plane parallel to (001) is less beam sensitive (more resistant) than one that is perpendicular to (001). We also observed generally that hkO diffractions disappeared more slowly than did hkl diffractions. This broadly agrees with the specimen orientation effect on the measured loss of elements. A reason for this could be that less energy is needed to disrupt truncated, weakly bonded layers whereas more energy is needed to remove part of the crystal in which extensive sheets are strongly bonded, as illustrated in Figure 13. In addition, possibly the beam energy disrupts the interlayer, allowing K loss, destroying the periodicity parallel to [001], but leaving the hk0 plane less disordered. Our measurements on the orientation effect are opposite to the observations of van der Pluijm et al. (1988).

Element loss from kaolinite shows an equivalent dependence on orientation with higher AI loss when the specimen plane is perpendicular to (001). Kaolinite in either orientation is damaged much faster (in less than 10 s in general) than all other layer silicates.

Finally, the effect of crystallographic orientation in the analysis of clay minerals cannot be evaluated when the analysis area is very thin because rapid damage masks the effect.

Effect of Crystal Ordering and Defects

Most common clay minerals, such as kaolinite, smectite and illite, occur as submicrometer crystallites with varying degrees of structural order and may be

as thin as 1 or 2 unit cell dimensions along [001]. The present study considered the effect of overall crystal defect state that may also be significant in controlling the loss of elements at the early stage of an analysis in clay. It is known that clay minerals are the most electron-beam-sensitive phyllosilicate, and kaolin is the most sensitive clay mineral (Mackinnon and Kaser 1987; Robertson and Eggleton 1991; Ma and Eggleton 1994; Ma 1996). Kaolinite and halloysite can only "surv ive" for less than 10 s in normal image mode under TEM before their electron diffraction patterns disappear. Highly ordered kaolinite can survive for a little longer than poorly ordered kaolinite (Ma 1996). For an analysis by AEM, a counting time of 50 s or greater (to achieve acceptable counting statistics) is required. Therefore, the majority of X-ray emissions collected for analysis using a stationary electron probe probably arise from a phase that has been highly disordered in situ relative to the original clay crystal. It is apparent that the AI/Si intensity ratios of the highly ordered kaolinite from Pittong are higher than those of the less-ordered kaolinite from Weipa (see Figures 2 and 3, A1/Si data for count rate --1600 cps) even though their compositions are very similar (Kao I and Kao IH, Table 1). This implies that differential element loss is more serious in less-ordered crystals. Therefore, crystal ordering and defects do play a significant role in element loss, particularly at the earlier stage of an analysis. However, the effects of disorder in AEM analyses of clay may be masked to some degree by rapid damage rates.

Peacor (1992; personal communications) showed that the grade of diagenesis is reflected by the rate of


r,t)

0.82

0.8

0.78

0.76

0.~

0.72

0.7

0.68

0.66 0

Figure 11.

(AI)/l(Si) = 0.720 + 0.105*e^(-t/89.5). R=0.997

O , , , , . ~ At Room Temp " ~ ~ Rate= 1600ep s

at - 170~ Rate=1300 cps

I ~ ~ ~ 257), R=0.998 " " " " , O

I I I I I I

50 100 150 200 250 300

Seconds

Loss of elements when using a cold stage, a) Kaolinite (Kao I).

350

element loss: K loss is negligible in 2M muscovite but becomes increasingly greater as one goes to illite and ultimately to smectite with decreasing grade. Selected area electron diffraction (SAED) patterns were not reported to change for those minerals under AEM conditions of 100 (or 120) kV accelerating voltage and scanning probe mode. Peacor attributed the change in K-loss to be due to change in degree of disorder as evidenced in part by variations in lattice fringes and in dislocation densities. We have observed the same in kaolin group minerals analyzed at 300 kV with stationary probe, although the damage rates for highly ordered crystal may be so rapid that the initial state of disorder is rapidly overprinted.

Morphology Effect

It is possible that morphology affects the degree of element loss at the very beginning of an analysis on a kaolin particle (within 10 s), although the rapid damage by a focused beam may mask any such morpho- logical effect. An analysis spot of 90 nm in diameter damaged both the platy and tubular kaolins' structure in about 3-5 s, similar to the loss of A1 in kaolin with a different morphology (Figures 2 and 6).

Effects of Different Analytical Conditions and Techniques

LOW TEMPERATURE. Mackinnon and Kaser (1987) and Mackinnon (1990) demonstrated that low temperature reduces the rate of element loss on AEM of clay minerals when a 200-kV stationary beam was used. Al- though we used a cold stage similar to that used by Mackiunon and Kaser (1987), we could not reproduce any strong effect (Figure 11). We took care to verify

that our Gatan 636N stage was indeed cooling the sample region by directly measuring a specimen cup temperature around - 1 3 0 ~ We suppose that the higher accelerating voltage (300 kV) we used could be responsible for overwhelming the effect of low temperature.

OTHER CONDITIONS. Comparison of the analytical results on phyllosilicates by Peacor (1992, scanning beam analyses at 100 or 120 kV), Mackinnon and Kaser (1987, stationary beam, 200 kV) and our own (stationary beam, 300 kV) illustrate the complexity of AEM analyses of these sensitive minerals. Results range from minimal or zero loss of elements by Peacor to serious loss of elements in our measurements. Part of the explanation will lie in the use of a scanning beam by Peacor, although we must point out that the beam selected for analyses in our EM430 (LaB 6 filament, low beam current 0.3 nA and microprobe-mode beam diameter 90 nm) is not particularly severe relative to microbeams used in most modern AEM studies. However, the fact that the 350-nm beam diameter re- duced our element loss shows that current densities were still excessive. However, X-ray counting rates on the order of 1000 cps are the minimum required to give acceptable analytical precision over 100-s counting times, so we could not sensitively analyze with any lower beam current using our detector with a solid angle of collection of 0.13 steradians. The alternative, to defocus the 300 kV beam to several hundred nano- meters in diameter, is also undesirable since it defeats the purpose of turning from the electron microprobe (thick specimen) to the AEM (thin film) for improved spatial resolution. We submit that our results, put in


r ~ 0.4

0.35

0.3

?,~ 0.25

~ 0.2

0 . 1 5

0 . 1

I(Ca)/I(Si) 0.05

I(Fe)/I(Si) 0

-170~ Cooma saponite (Sm I) Rate = 1320 eps

~ ~ - - - 0 ~ 0

~ - O - - O - . .O. O - - - O - - - - ~ _ - - - O . . . . . ~ . . . . . r

X - - ) 6 - , - ) 4 - - ~ - - - ~ - - --~ - - n - N - - - - - ~ - - - - ~ - - - - - h(

50 100 150 200 250 300

Seconds

350

�9 - 0 . 4

0.35

0.3

0.25

0.2

"--" 0.15

0.1

�9 - 0.05

o

- 170~ Wyoming montmori l loni te (Sm Ill)

i t i

%-.o o o 0 0 0 0 O - - " 0

Rate = 900 eps

X - - X - -X- - X - - - , ; < - - - X - - X - __ - )< __ _ _ .X_ _ _ _ _ x

', I I I I I

i i i i i i

50 100 150 200 250 300

Seconds

Figure 11. Continued. b) Smectite (Sm I and Sm III).

350

the con tex t of o ther studies, emphas i ze the need for phyUosi l icate minera log i s t s to careful ly select thei r analyt ical cond i t ions w h e n e m b a r k i n g on A E M studies.

Cor re la t ion b e t w e e n Io)/l(si) and M e a s u r i n g T ime

All our e lement - loss data cou ld be fitted by an ex- ponen t ia l cor re la t ion b e t w e e n the rat io l(JI(s 0 and measur ing t ime dur ing an analys is on a cons tan t spot (as

i l lus t ra ted in F igures 1 to 4, 6, 9, 10 and 11). The corre la t ion l ine locates wi th in the error r ange as dem- ons t ra ted in F igure 1. The genera l express ion o f the rat io l(iJl(s~ wi th t ime is:

I c i ) / l ( s i ) = K 3 + K1 elQt [ 4 ]

where K1, K2 and K 3 are cons tan ts and t is coun t ing time. Th i s equa t ion m a y sugges t tha t there are 2 parts

Vol. 46, No. 3, 1998

0.7

0.6

0.5

0.4

0.3

0.2

~ , 0.1

0

r,o 0.7

~-" 0.6

0.5

0.4

0.3

0.2

o.1

o

AEM: Element loss in clays and other phyllosilicates

-170~ Muscovite Rate = 800 cps

~176 o - - - o - - o - -o--_ o..__ _o__ ___o ~ o

i

50 100 1;0 2;0 ' ' 250 300 Seconds

-170~ Biotite

350

Rate = 1950 cps

Q . i i i

O - 4D- " 0 - O - . - 0 - - " 0 . . . . 0 . . . . 0 . . . . . 0

I(K)/I(Si)

-.N-

~ 1 7 6 0-----o ---o-- o

I(Mg)/I(Si)

X---

I I I I

~ - A - --~- -ZX-- -,'~ - - --~- - -/~ - - - "~ . . . . . ~ . . . . .

i i i I i i

50 100 150 200 250 300

Seconds

Figure 11. Continued. c) Muscovite. d) Biotite.

350

313

which can be attr ibuted to the intensi ty ratios; that is, one (KIeK:') that decays and another (K3) that does not.

Loss of material by electron b e a m damage gives rise to a reduct ion in X-ray ratio fo l lowing an exponent ia l decay curve. The effect o f e lectron distr ibution on the analysis area is likely to contr ibute to such a relation (Figure 14). Figure 14 shows that the main mecha- n i sms expec ted for the loss o f e lements are likely sputter ing and "d i f fu s ion" . The a tom d isp lacement within the analysis area is unlikely to affect e lement losses very much. However , h o w or why the exponent ia l relation occurs is not clear, a l though the (Kl eK2') part is suspected to relate to di f fus ion whereas the (/(3) part

is probably caused by sputtering. Perhaps the exponential relat ion could be used in future as an index o f structural stability in phyl losi l icate minerals .

Determinat ion o f Clay Compos i t ion

Because o f the differential loss of e lements during A E M analysis as men t ioned above, clay compos i t ions calculated using fixed k-factors (calibrated f rom beam- stable minerals other than clay, such as feldspars) generally show a small excess of Si compared to its ideal compos i t ion and a cor responding def ic iency in A1 and other e lements such as Fe. Two me thods were pro- posed and deve loped in this study to overcome this


Figure 12. TEM image showing a hole within a kaolinite particle, which was drilled by an electron beam during 1 analysis (250 s). The hole is 70 nm in diameter and the analysis spot marked by a dashed circle is 90 nm in diameter.

p rob l em and increase the rel iabi l i ty o f results. The me thods are descr ibed as fo l lows us ing kaol in i te as an example .

METHOD 1. " E f f e c t i v e " k-factors o f A1/Si and Fe/Si are ob ta ined us ing w e l l - k n o w n kaol ini tes ; these values will be ana lys i s - t ime dependent , Then the " e f f e c t i v e " k-factors are used accord ing ly to ca lcula te compos i - t ions of u n k n o w n kaol in i tes where these analyses are ob ta ined unde r exact ly the same condi t ions ( inc luding b e a m current , diameter , sample th ickness and crys ta l or ienta t ion) as were used for the s tandards. F igure 15, as an example , i l lustrates the " e f f e c t i v e " k-factors of A1/Si as a func t ion o f analys is t ime and at 3 d i f ferent th icknesses for we l l -o rdered kaol in i te wi th its specimen p lane pe rpend icu la r to (001).

METHOD 2. First, a series of m e a s u r e m e n t s should be m a d e o f the X- ray in tensi t ies as a func t ion o f t ime of an u n k n o w n kaol ini te , j u s t l ike the analyses in the Re- suits sect ion. T h e n the in tens i ty ra t ios ( such as A1/Si and Fe/Si ) of this u n k n o w n phase at t ime zero are es t imated us ing the i r exponent ia l cor re la t ion lines. The compos i t i on of the u n k n o w n kaol in i te is f inally cal- cula ted by us ing conven t iona l k-factors w h i c h can h a v e been de t e rmined f rom any compos i t iona l ly char- ac ter ized solid. However , uncer ta in ty does arise as s h o w n for example in F igure 2 where i t is c lear tha t the t ime-zero in tens i ty rat ios of A1/Si ( that is, K1 + K 0 va ry f rom 0.80 to 0.83, a l though this may be caused by or ig inal chemica l he te rogene i ty be tween kaol ini te crystals . A t o m i c AI /S i ra t ios conver t to 0 .936, 0.945 and 0 .936 for the 3 analyses w h e n us ing the conven t iona l k-factor (k = 1.17).

B o t h me thods are ve ry t ime-consuming . Table 2 shows the actual analys is of the Pi t tong kaol in i te (Kao I) and the A E M analys is us ing me thods 1 and 2. Meth - od 1 c lear ly p roduces be t te r results.

Figure 13. Sketch showing the possible electron disruption on sheet silicates, a) When the specimen plane is perpendicular to (001), less energy is needed to disrupt the truncated basal-plane layers, which leads to rapid sputtering, b) When the specimen plane is parallel to (001), more energy is needed to disrupt and remove extensive, strongly bonded sheets.

C O N C L U S I O N S

The effects o f direct a tom d isp lacement , ion iza t ion and hea t ing dur ing e lec t ron b o m b a r d m e n t of layer silicates g ive r ise to spec imen d a m a g e that can cause ser ious A E M analysis errors resu l t ing f rom the different ia l loss of e lements . 1) In our analysis o f phyl los i l ica tes , no t on ly alkali e l emen t s and low-Z e lement s bu t also Fe are significant ly lost ove r t ime. T h e degree o f loss o f A1 in kaolinite is greater than in o ther sheet si l icates, m u c h h igher even than the K loss in muscovi te , for example . 2) The rate of e l emen t loss is in par t a func t ion of s t ructure and compos i t ion . The e l emen t loss in c lay minera l s is sugges ted (main ly) to be due to d i f fus ion and sputter ing. Ear ly in the s t ructural d a m a g e period, d i f fus ion could be the m a i n cause for the e l emen t loss toge ther wi th sput ter ing, whereas dur ing the later stages (af ter the s t ructure has b e e n lost) sput ter ing could domina te . However , th is m i g h t be d i f ferent for o ther phyl los i l ica tes ( such as muscov i t e and biot i te) where d i f fus ion could be the ma jo r fac tor for m u c h longer because these minera l s are more res is tant to the h igh -ene rgy e lec t ron beam. 3) A n exponent ia l cor re la t ion b e t w een the rat io lt~/lcsi ~ and analyt ica l measu r ing t ime was found to adequa te ly descr ibe the measurement s : I~i/I~si~ = K3 + K1 eK2'- 4) Frac t ional e l emen t loss increases wi th decreas ing spec imen thickness .

Vol. 46, No. 3, 1998 AEM: Element loss in clays and other phyllosilicates

~ p r o b e Electr~ Gaussian

315

Specimen

t = 0

7

Atom t l displacement

@ Sputtering

t2 'Diffusion.*'

(t2 > t l )

Figure 14. A diagram showing possible mechanisms for the loss of elements with time (t). The effect of electron distribution might contribute to the exponential relation between the l(i/l(si~ ratio and t. Sputtering and diffusion are likely to be the main causes for element losses.

5) A lower beam current and a larger analysis area reduce the loss of elements, which is particularly significant and important to clay minerals. 6) Most o f the e lement loss takes place at the beginning of spectral col lect ion, before statistically significant count totals can be obtained. 7) For phyllosil icates, including clay minerals, the loss of e lements is also related to the specimen habit, loss of e lements general ly being more severe for thin crystals having a specimen plane perpendicular to (001).

8) The mechan ism of "d i f fu s ion" used in some papers to explain the loss o f e lements in A E M analysis o f phyllosi l icates at 100 to 200 kV accelerating vol tages is probably accompanied at 300 kV by sputtering and there may be other factors. Since the specific mechanism of "d i f fu s ion" is still to be established, we rec- o m m e n d using the more general term "e l emen t loss" instead. 9) At low temperature (around - 1 3 0 ~ these analyses of phyllosil icates f rom electrons of 300-keV in-

1.45

1.4

1.35

1.3

1.25

1.2

Figure 15.

k = 1.24 + 0 . 1 8 9 " ~ 6 0 c p s

~ .f i~. ~ Rate = 1600cps

~ k = l 20 + 0 179"[1 e^( t/96 8)1

:7.:.'i::.: .7-:: J k = 1.23 + 0.046"[1 - e^(-t/52.7)]

. . . . ~ . . . . i . . . . i . . . . i . . . . i . . . . = . . . .

0 50 I00 150 200 250 300 350

Measuring time (second)

"Effective" k-factors from well-ordered kaolinite. Based on data from Figure 2.


Table 2. The actual analysis of Pittong kaolinite (Kao I) and the AEM analysis using Methods 1 and 2.

Atomic Actual Method Method Method Method Method Method ratio valuer 1 1 1 2 2 2

A1/Si 1.025 1.021 1.035 1.024 0.936 0.945 0.965

"~ From Table t.

c ident energy showed that the loss of e lements re- ma ined serious, in contrast to the results o f Mack innon and Kaser (1987) who also analyzed clays but used 200-keV electrons. 10) One should be aware that e lement loss during analysis is ins t rument-dependent . Before undertaking measurements o f unknown clay crystals, one should use s tandard clay minerals (which have a similar structural order ing to unknown crystals) to calibrate an A E M ins t rument for " e f f ec t i ve" k-factors o f all elements for each mineral and to establ ish acceptable condi t ions for analysis (Method 1). Analys is o f an un- k n o w n should be obtained under exact ly the same con- dit ions as were used for standards, including at least beam diameter, spec iment th ickness and total count ing t ime for analysis. Al though loss of e lements occurs in both the s tandard and unknown, it should occur ap- proximate ly equally in both and will be accounted for in the unknown ' s compos i t ion via the " e f f ec t i ve" k- factor es tabl ished f rom the standard.

A C K N O W L E D G M E N T S

We wish to thank D. R. Peacor and I. D. R. Mackinnon for their thoughtful and constructive reviews on the manu- script. C. Ma also would like to thank G. R. Rossman for his helpful comments.

R E F E R E N C E S

Ahn JH, Peacor PR, Essene EJ. 1986. Cation-diffusion-in- duced characteristic beam damage in transmission electron microscope images of micas. Ultramicroscopy 19:375-382.

Cliff G, Lorimer GW. 1975. The quantitative analysis of thin specimens. J of Microscopy 103:203-207.

Knipe ILl. 1979. Chemical analysis during slaty cleavage de- velopment. Bull Mineral 102:206-210.

Lorimer GW, Cliff G. 1976. Analytical electron microscopy of minerals. In: Wenk HR, editor. Electron microscopy in mineralogy. Berlin: Springer. p 506-519.

Ma C. 1996. The ultra-structure of kaolin [Ph.D. thesis]. Can- berra, Australia: Australian Nat Univ. 343 p.

Ma C, Eggleton RA. 1994. Structural characteristics of kaolin minerals from eastern Australian regolith. In: Pain CP, Craig MA, Campbell ID, editors. Abs Australian Regolith Conf; 1994; Broken Hill, Australia. p 40.

Mackinnon IDR. 1990. Low-temperature analyses in the analytical electron microscope. In: Mackinnon IDR, Mump- ton FA, editors. Electron-optical methods in clay science. Boulder, CO: Clay Miner Soc. p 90-106.

Mackinnon IDR, Kaser SA. 1987. Microanalysis of clays at low temperature. In: Geiss RH, editor. Microbeam Analy- sis. San Francisco Press Inc. p 332-333.

Peacor DR. 1992. Analytical electron microscopy: X-ray analysis. In: Buseck PR, editor. Minerals and reactions at the atomic scale: Transmission electron microscopy. Wash- ington, DC: Miner Soc Am. p 113-140.

Robertson IDM, Eggleton RA. 1991. Weathering of granitic muscovite to kaolinite and halloysite and of plagioclase- derived kaolinite to halloysite. Clays Clay Miner 39:113- 126.

van der Pluijm BA, Lee JH, Peacor DR. 1988. Analytical electron microscopy and the problem of potassium diffusion. Clays Clay Miner 36:498-504.

Williams DB, Carter CB. 1996. Transmission electron microscopy: A textbook for materials science. New York: Ple- num Pr. 729

(Received 4 August 1997; accepted 28 September 1997; Ms. 97-071).

Documents

ANALYTICAL ELECTRON MICROSCOPY IN CLAYS AND OTHER ...clays.org/journal/archive/volume 46/46-3-301.pdf · 3 Electron Microscopy Unit, Australian National University, Canberra, ACT