Charge Contrast IMAGING APPLICATIONS TO the Bayer Process
Gerald I. D. Roach,* John B. Cornell,* Brendan J. Griffin** and Travis C. Baroni***
* Research and Development, Alcoa of Australia Limited
**Western Australian Centre for Microscopy, The University of Western Australia
***Department of Chemistry, The University of Western Australia
The new and exciting technique of charge contrast imaging developed at the Centre for Microscopy and Microanalysis in conjunction with Alcoa of Australia Limited utilises the Environmental Scanning Electron Microscope. The technique gives valid and detailed information on the growth of a variety of materials in the Bayer process. Imaging of gibbsite particles from the precipitation process gives information on their growth history such as the number of times a particle cycles through precipitation, at what stage particles agglomerated together and relative growth rates on different crystal faces. Hydrates with varying impurity levels were grown to determine if these impurities were the cause of the observed growth rings. Similar structures were observed with and without the impurities. Improved imaging with the atomic force microscope revealed topographical features at the ring positions.
The technique has also been applied to a range of scale samples including gibbsite and DSP scales and can give useful insights into their growth. For alumina, improved imaging of cracks is obtained as well as the appearance of textural differences which are still awaiting an explanation. The technique is now finding wide application in areas such as medicine (examination of kidney stones), semi-conductors, mineralogy and ceramics.
Bayer, crystallisation, charge contrast, electron microscopy, gibbsite
Charge Contrast IMAGING APPLICATIONS TO the Bayer Process
Gerald I. D. Roach, John B. Cornell, Brendan J. Griffin and Travis C. Baroni
Bayer gibbsite particles are grown by recirculating the gibbsite seed through the precipitation process several times, either in batch or continuous reactors, until it has grown (or agglomerated) to the required size. Hence growth rings similar to those in trees are expected but had not hitherto been able to be imaged by any known technique. A technique was developed at the Centre for Microscopy and Microanalysis at the University of Western Australia that enabled such information to be obtained. The technique images uncoated polished sections of hydrate in the Environmental Scanning Electron Microscope (ESEM). The type of information that can be obtained has been detailed elsewhere (Roach et al, 1998) and includes information on growth rate, residence time and agglomeration. Such information can help in understanding the growth and agglomeration processes in the precipitation circuit and the effect of impurities on particle growth. Knowing the growth mechanism of the particles compliments the data from techniques such as atomic force microscopy and cathodoluminescence.
Since the initial work, further work has not only been related to utilising the technique to better understand the precipitation process, but has also been directed at determining the role of impurities in image production, improving image quality, determining if the fine detail observed is real and how it relates to crystal growth, qualitatively and quantitatively describing the features and contrast and extending the technique to three dimensional imaging. Charge contrast imaging has been applied to other Bayer materials such as process scale samples to enable their growth mechanism to be understood. The technique is also useful in examining alumina samples, as it is extremely successful at distinguishing cracks.
2.0 Charge Contrast Imaging (CCI)
Since its publication the technique is being widely studied and used, (Griffin, 1997). However, the underlying physics behind the technique is still not fully understood and is currently undergoing further investigation (with sponsorship by Alcoa). Readers will need to keep abreast with the literature if they wish to understand the technique more fully. A brief but incomplete summary of the technique is presented below.
The technique essentially monitors the discharge of the surface charge build-up in uncoated specimens in an ESEM. In a conventional SEM the charge build-up results in a flood of secondary electrons, which normally causes saturation of the secondary electron detector preventing any imaging. However the ESEM, with its lower and variable vacuum in the specimen chamber, enables the charge dissipation to be controlled. Further, under very specific operating conditions, the emission from the sample is directly related to the trapped charge that in turn relates to the defect density. The defect density is a function of the type of crystal growth (eg initial nucleation as opposed to uniform deposition) and the amount of contaminants. Consequently the image shows high contrast across growth layers and where line defects, such as incipient or developed fractures, are present. It is now also considered that the impacting positive ions, which result from electron-chamber gas interaction, play a significant role in this imaging process and that the data is from the near-surface region of the sample.
The material to be examined is mounted in a special resin to minimise edge charging effects and sectioned and polished to give a very flat surface, but with minimum distortion. The sectioned and polished samples are placed in the ESEM and illuminated by the electron beam which causes the particles to charge. The ESEM used has a specially modified detector to improve the secondary electron imaging. The atmosphere in the chamber and other operating parameters are adjusted until the required image is observed.
3.0 CCI of GIBBSITE
The technique was initially verified utilising specially prepared hydrate samples which had several growth layers added under standard laboratory growth conditions (Roach et al 1998). A section through a particle from such a sample utilising both conventional SEM and the new CCI technique is shown in Figure 1. It is clear that the internal structure of the hydrate is revealed with the CCI technique and in particular the new layers that have been grown are readily distinguished. The measured thickness of the growth layers corresponded directly to that calculated from the amount of alumina precipitated.
Standard SEM and Charge Contrast images of laboratory produced hydrate
Typical plant samples are shown in Figure 2. For particle (a) the regular growth rings reveal that this particle had been grown in a batch precipitation circuit. For particle (b) there are well defined growth rings but of varying width indicating that it is from a continuous system. For particle (c) the varying band width indicates it is again from a continuous system but also that the particle has formed through agglomeration of a few large particles. For particle (d) the sample has formed predominantly by agglomeration and then had a small layer of growth.
a - Regular spaced growth rings
b - Irregular spaced growth rings
c - Agglomerated core
d - Highly agglomerated
3.1 Effect of impurities
Hydrate samples were grown in a variety of liquors with different impurities to determine whether the impurities were the cause of the observed rings. Impurities such as iron, silica and calcia are known to precipitate either initially (iron and calcia) in the precipitation process or at the end in the classification circuit (silica) (Roach et al 1986). Synthetic liquors were depleted to less than 1 ppm calcia or iron and to less than 0.1 g/L silica, or increased to high values.
A particle is shown in Figure 3 in which the layers were:
CCI image of a four-layered batch precipitated hydrate crystal
The growth rings were formed in all the liquors indicating that such minor impurities were not responsible for the observed rings. Also a sample was grown in a continuous reactor and impurities added at specific times. Such additions did not create rings. This technique does not, by itself, indicate the presence of impurities although by knowing the structure, the location of such impurities can be more readily pinpointed. Further work on impurity location is being pursued using other techniques in conjunction with the structural information given by this technique.
3.2 Characterising the Structure
The images have been qualitatively and quantitatively assessed by describing various features and with grey scale evaluation. An example of the information obtained is given in Figure 4. Such information enables more detailed comparisons between samples. The variation in contrast is not unlike what might be expected in terms of soda variation across such growth bands (see Roach et al, 1988). In that paper it was also demonstrated that soda can migrate under the influence of the electron beam and that may play a role in the imaging.
Line profile across growth layers in a Charge Contrast image
3.3 Fine structure information
The information visible between the bands was assessed to determine if it was real or possibly an artefact from the imaging process or sample preparation. Samples were grown as previously but during the growth sub-samples were extracted from the reactor and the growing surface examined using the FESEM. Comparison of these FESEM images with CCI images of sections through the final particles were made, (Figure 5). The first FESEM image shows the development of new small crystal growth areas on the main faces of the seed. These continue to grow and finally one smooth growth surface is formed as seen in the second micrograph. The boundary between the seed crystal and the new growth is clearly revealed in the CCI images. However the CCI images could not resolve to the level of detail required to image the formation of the new growth centres (a resolution of about 100nm appears to be required). Further development of the technique is required if such information is to be obtained. The origin and physical significance of the fine lines within the bands has still to be explained.
FESEM image showing nucleation
on seed in early stages
FESEM image showing final
smooth growth surface
CCI images of grown crystal with original seed surface indicated
FESEM and CCI images of nucleation and growth
3.4 Comparison with Atomic Force Microscopy Images (AFM)
Previous work utilising the AFM had indicated the polished samples were flat and there were no major topographical variations related to the observed structure. With further refinement of both the CCI and AFM imaging, this has been re-evaluated. AFM images are shown in Figure 6 for a sample with well characterised growth rings; the rings are visible utilising tapping mode phase contrast AFM. There is a topographical feature associated with the ring boundary that might be related to some etching at this location by the polishing medium. The level of detail obtained with CCI (and the speed and ease at which it is obtained) is significantly better than that obtained currently with the AFM.
Tapping mode AFM image of polished section showing 4 growth layers together
with the topographical information
The calcination of gibbsite particles to alumina results in a series of phase changes as well as producing fine pores and fissures in the particles. These fissures can contribute to the breakdown characteristics of the alumina.
CCI can be used to highlight the fissures in the material due the enhanced edge contrast. See Figure 7. Also contrast differences are observed that are not seen using other imaging techniques. Those contrast differences might be related to phase differences and hopefully Kikuchi analysis of the areas will determine if that is the case. The crack and texture information is important in assessing the fragility of such material.
CCI for calcined alumina
Scales grow by two basic mechanisms and have been previously termed settled scales and growth scales (Roach and Cornell, 1985). In the first the slurry particles settle and are cemented together and in the second the scale grows directly from solution. The CCI technique should enable a better distinction between the slurry particles and the cement especially where both are the same phase such as in certain precipitator scales, (Figure 8). Surprisingly such imaging has been difficult to accomplish and is leading to further understanding of some of the subtleties of the technique. Growth bands revealed by CCI for a particle from a process sand filter on green liquor is shown in Figure 9. The filter is backwashed with spent liquor. By counting the number of layers the number of cycles the filter has been on line can be deduced.
Scale from a precipitator showing hydrate particles and cementing agent
Growth bands on sand from sand filter
Also in growth scales CCI imaging can reveal banding which is not seen in conventional SEM (but often can be seen optically). Interestingly the reverse can apply. In a scale sample from a thickener launder multiple banding is observed, but not by CCI. Whether this is because the sample has grown continuously such that impurity distribution is homogeneous or because of other effects is unknown. Certainly there is still much to learn about the technique and also more information is expected to be gained when such anomalies are understood.
Charge contrast imaging is a powerful technique for studying gibbsite growth. The technique gives valid information on the growth history of gibbsite particles. It enables growth rates to be directly measured in the plant process. The technique should be an important adjunct to any studies related to gibbsite precipitation or studies relating gibbsite morphology to its properties. Indeed it is proving to be an exciting and valuable tool in the study of any non- or semi-conducting material which have been formed via a crystallisation process. As examples it is now being used to examine a range of materials including: development and growth of kidney stones, zircon mineral growth, semiconductors and other minerals and ceramics, with ever increasing applications being found.
Griffin, B.J. (1997a). A new mechanism for the imaging of crystal structure in non-conductive materials: an application of charge-induced contrast in the environmental scanning electron microscope (ESEM). Microscopy and Microanalysis, 3 (s2), pp 1197-8.
Griffin, B.J. (1997b). Novel and advanced applications of the low vacuum and environmental scanning electron microscope (ESEM). Microscopy and Microanalysis,
3 (s2), pp 385-6.
Roach G.I.D. and Cornell J.B. (1985). Scaling in Bayer Plants. CHEMECA 1985. The 13th Australian Chemical Engineering Conference, Perth. pp 217-222.
Roach G.I.D., Cornell J.B. and Antonovsky A. (1988). Hydrate and Alumina - Unmasking Their Mysteries. Proceedings First International Alumina Quality Workshop, Gladstone, Queensland, Australia. pp 137-148.
Roach G.I.D. and Cornell J.B., (1996). Alumina and Hydrate - Unmasking Their Mysteries, Eight Years On. Proceedings Fourth International Alumina Quality Workshop, Darwin, Northern Territory, Australia. pp 1-8.
Roach G.I.D., Cornell J.B. and Griffin B.J. (1998). Gibbsite growth history revelations of a new scanning electron microscope technique. Light Metals. (The Metallurgical Society of AIME). pp 153-158.