INVESTIGATION OF THE GROWTH OF GIBBSITE CRYSTALS BY ATOMIC FORCE MICROSCOPY AND OPTICAL MICROSCOPY

Sawsan Freij, Mei-yin Lee, Manijeh Reyhani and Gordon Parkinson

A.J.Parker Cooperative Research Centre for Hydrometallurgy, School of Applied Chemistry, Curtin University of Technology, GPO Box U1987, Perth 6845, Western Australia

ABSTRACT

In situ optical microscopy and ex situ atomic force microscopy have been used to investigate the kinetics and surface evolution during the growth of single crystals of gibbsite at 80 C as a function of supersaturation. The results of in situ optical microscopy reveal that growth on the basal face follows a spiral growth mechanism up to a relative supersaturation of 0.81 (0.65 A/C). Above that level, the growth of the basal face was found to follow the birth and spread mechanism. The edge free energy, calculated for a polynuclear birth and spread mechanism, is 6.0 1.5 kJ/mole.

Similar studies have been carried out in a series of ex situ experiments in which the same crystal area was repeatedly located and imaged after immersion in synthetic liquor, using atomic force microscopy (AFM). The results of these experiments are in agreement with those of the optical microscopy, and show that growth on the basal face at a low relative supersaturation of 0.43 (0.50 A/C) occurs by a step flow mechanism, while growth was found to occur by continuous birth and spread mechanism at a higher supersaturation of 1.0 (0.70 A/C). Growth by continuous birth and spread mechanism and step flow were both observed to occur simultaneously on the basal face of different seed crystals at a relative supersaturation of 0.71 (0.60 A/C).

KEY WORDS:

gibbsite, growth mechanism, AFM, optical microscopy

INVESTIGATION OF THE GROWTH OF GIBBSITE CRYSTALS BY ATOMIC FORCE MICROSCOPY AND OPTICAL MICROSCOPY

Sawsan Freij, Mei-yin Lee, Manijeh Reyhani and Gordon Parkinson

1.0 INTRODUCTION

The growth of gibbsite via the Bayer Process has been widely studied through bulk experiments, and these have shown that the growth rate of gibbsite has a second order dependence on the supersaturation of the solution, suggesting a spiral growth mechanism consistent with the BCF model (Misra and White, 1971; Brown, 1972, King, 1973, Veesler and Boistelle, 1994). However, in contrast, the dependence of growth rate on the surface area and the negligible growth reported at low supersaturation suggest that growth occurs not through a screw dislocation mechanism, but via a birth and spread model (Brown, 1972).

In situ optical microscopy is a relatively straightforward technique, available for the study of growth rates of individual crystal faces; however, the resolution of this method is insufficient for direct observation of the mechanism of growth. Recently, AFM has been used to investigate a variety of crystal surfaces in contact with aqueous solution (Hillner, 1992., Bosbach, 1995., Campbell, 1996). These studies have provided a wealth of qualitative information concerning surface morphological evolution and growth dynamics.

The complementary nature of optical microscopy and AFM is used to advantage in the present study, which combines the results of in situ optical microscopy studies of the kinetics of growth for the basal face of gibbsite and ex situ atomic force microscopy investigations of the growth mechanism of gibbsite crystals over a range of supersaturations.

2.0 NOMENCLATURE

Following the North American alumina industry terminologies, the aluminate solution compositions are expressed as follows: aluminium concentration (A), as g/L Al2O3; sodium hydroxide or caustic concentration (C) expressed as g/L Na2CO3; and A/C refers to the alumina to caustic ratio.

Relative supersaturation, s, is defined by Equation 1(Veesler and Boistelle, 1994), and the solubility model used to determine equilibrium concentrations is given by Rosenberg and Healy (1996).

(1)

3.0 EXPERIMENTAL

3.1 Liquor Preparation

The methods used to prepare the liquors have been described in a previous paper (Lee et al., 1996). All liquors were made with a caustic concentration of 200 C.

3.2 Optical Microscopy - Seed Preparation and Experimental Set-Up

The method used to prepare the seed and set up the experiments for in situ observations of gibbsite crystal growth have also been described previously (Lee et al., 1996; Lee and Parkinson, 1998).

To determine the effect of supersaturation, the growth of the crystals was recorded under standard conditions (eg; s = 0.81 at 80oC) for two hours, before changing the liquor to one with a different supersaturation value. The growth rates obtained are therefore relative to the standard conditions. This method was adopted because the single crystals of gibbsite that were studied were found to exhibit considerable growth rate dispersion.

3.3 Atomic Force Microscopy-Seed Preparation and Experimental Set-Up

For the preparation of seed crystals in a form suitable for AFM growth studies, stainless steel disks (1.5 cm diameter) were covered with a 0.60 A/C liquor (s = 0.71) and held without stirring in a water bath at 80o C for 24 hours. The disks were then removed from solution, rinsed with deionised water and air dried at room temperature.

To observe the growth of the basal face, synthetic liquors were prepared and allowed to stand at 80o C in a water bath for 30 minutes to ensure that the liquor was hot when the seed (on disk) was added. The basal faces of seed crystal were imaged and then grown in 30 ml of synthetic liquor at 80o C for specific time intervals.

After each growth period, the stainless steel disk was rinsed with deionised water, air dried and the previously imaged area was located and imaged. Location of the same area was assisted by the use of a co-axial optical microscope.

AFM investigations were carried out in air with a Digital Instrument Nanoscope E AFM, using a 12m m scanner; all the AFM images were collected in the contact mode. Wide triangular shaped 100 m m cantilevers made of gold coated Si3N4, with a spring constant of 0.58 Nm-1 were used. All the images shown are unfiltered.

4.0 RESULTS AND DISCUSSION

The initial results from the in situ optical microscopy studies of gibbsite crystal growth showed that although the growth rates of the crystal faces of individual crystals are constant for the duration of the experiment, where the liquor supersaturation is fixed, variations in the growth rates of individual crystals were observed from experiment to experiment (under identical conditions), as well as from crystal to crystal within the same experiment. An experimental regime was therefore adopted to obtain the relative growth rates of the crystal faces, in order to study the effect of supersaturation in the presence of growth rate dispersion.

In order to determine the growth mechanism, plots of versus were made. For growth rates following the birth and spread mechanism, where , such plots will be linear, whereas for growth rates following the BCF model (GR = K s 2) , they will not (Liu et al., 1997). This method is used to distinguish between these two growth mechanisms of the crystal face, with the aid of curve fitting software.

The dependence of relative growth rate obtained as a function of supersaturation is given in Figure 1. Curve fitting calculations showed that the data fitted a quadratic equation (correlation coefficient = 0.964) better than a linear function (correlation coefficient = 0.925), implying that the basal face grows via a spiral growth mechanism, which is not consistent with visual observations. Figure 2 shows some optical microscopy images from an extended sequence, where the nucleation and lateral growth of features are seen on the basal face.

 

Figure 1

A plot of relative growth rate (GR) versus supersaturation(s ) at 80 C

a                                         b                                         c

Figure 2(a-c)

Optical microscopy images showing nucleation and lateral growth of features on the basal face after the following time intervals: 0, 73 and 135 minutes, respectively

Figure 3 shows the plot of versus , showing a linear relationship above a relative supersaturation of 0.81 (0.65 A/C) at 80oC, implying that above this level of relative supersaturation, growth occurs by a birth and spread mechanism. Below this level of supersaturation, the spiral growth model better describes the crystal growth. This is in agreement with visual observations by in situ optical microscopy, where the development of nuclei and their subsequent growth were only observed above a relative supersaturation of 0.81 at 80oC. Although the existence of spirals was not observed, their presence cannot be dismissed due to the insufficient resolution of the optical microscope, as the vertical resolution of a typical optical microscope ranges from ca 0.1 to 1 m m, whereas it is less than 1 nm for AFM. Consequently, the optical observations are limited to macrosteps. Similarly, growth by a birth and spread mechanism below a relative supersaturation of 0.81 (0.65 A/C) at 80C cannot be dismissed. It is, however, believed that the growth mechanism is a combination of the birth and spread and spiral growth processes, and that growth by the birth and spread mechanism dominates at a relative supersaturation of 0.81 at 80C.

 

Figure 3

A plot of versus at 80 C

From the kinetic results, the edge free energy may be calculated, and this may be used to investigate whether the birth and spread mechanism is plausible. Visual observations of the growth process by optical microscopy showed the formation of multiple nuclei, but not the formation of nuclei on top of existing ones; hence, only the edge free energy for a polynuclear birth and spread process was considered. To calculate the edge free energy (f ), the appropriate natural logarithmic plot of

was used (Veesler and Boistelle, 1994). The edge free energy was calculated to be 6.0 1.5 kJ/mole, which is comparable with values reported for other ionic solids (Gindt and Kern, 1968).

 

Figure 4 (a,b) contains AFM images showing growth on the basal face of a gibbsite crystal in a liquor with a relative supersaturation of 0.41 (0.50 A/C) at 80 C; the scan size is 5 m m. Initially, the surface contains polygonal growth hillocks at the middle of the image, and it is possible that they are emerging in the vicinity of a screw dislocation. After growth (90 minutes), steps form; note that the steps that spread over the hillock exhibit the same shape as those of the hillock.

a                                                       b

Figure 4

A growth sequence on the basal face of gibbsite in a solution with a relative supersaturation of 0.41 at 80 C after the following time intervals: 0 and 90 minutes, respectively. The relative height (Z range) of the features ranges from 0 to 500 nm, with the light colour indicating higher features and vice versa

Growth by a continuous birth and spread mechanism was observed on the basal face of another crystal grown in a liquor with a relative supersaturation of 0.71 at 80 C (0.60 A/C). Figure 5a shows the seed before the beginning of the experiments; the surface is relatively flat with an hexagonal feature, the edges of which are aligned with the bulk morphology of the parent crystal. After 20 minutes in the growth solution, approximately circular nuclei of nearly uniform size appear on the surface (Figure 5b). The average height of these nuclei is 10 nm, with lateral ‘diameters’ ranging from 200 nm to 300 nm, indicating that either nucleation occurred over a finite period, or the initial nuclei had a significant size distribution. Images of other areas on the same crystal and on other crystals on the same sample disk, taken at the same time, confirmed the typical nature of the nucleation. The image taken after 80 minutes of growth (Figure 5c) shows that the nuclei have grown laterally and vertically into elongated features, indicating that the growth onto the {100} is faster than the growth onto the {110} faces. This observation is commensurate with the observations of Rossiter et al. (1998) that roughening on the basal faces of gibbsite seeds in a stirred crystallizer is followed by the formation of diamond shaped features. Many of these features are subsequently released as secondary nuclei, but some remain and become intergrown with the parent crystal. With further growth time, the nuclei observed here continue to expand and coalesce to form a smooth surface (Figure 5d), which looks very similar to the surface of the original face shown in Figure 5a. This may be a consequence of the spacing between the nuclei being on average uniform, such that the coalescence of the nuclei on the surface would thus lead to the formation of the same number of layers overall; this is in agreement with the findings of Keller (1986) who studied the crystal growth of sodium chloride using electron microscopy and observed restoration of the cleavage step pattern by random nucleation. Comparison of Figure 5a and 5d shows that whilst the original surface topography has been restored, and the shape of the boundaries of the hexagonal feature are recognisable, the overall dimensions have increased, indicating that there has been lateral as well as vertical deposition of material during the growth cycle. The increase in thickness of the basal plane (as estimated from the heights of the nuclei) is ca 10 nm, while the growth along the prismatic ([100] and [110]) directions is 250 nm. The higher relative growth rate on the prismatic [(100) and (110)] faces compared to that on the basal (001) face is commensurate with the observed plate-like morphology of gibbsite.

Under the same experimental conditions, growth on the basal face of a different crystal was observed to occur by step growth at a growth hillock, and then nucleation events started, indicating that at this supersaturation level, step growth and nucleation both compete on the basal face of gibbsite (Figure 6). A comprehensive study of the mechanism of gibbsite growth and the effect of surface topography are discussed in separate papers (Freij et al a,b ).

a                                                            b

 

c                                                                d

Figure 5 (a-d)

A growth sequence on the basal face of gibbsite in a solution with a relative supersaturation of 0.71 at 80 C after the following time intervals: 0, 20, 80 and 300 minutes, respectively. The relative height (Z range) of the features ranges from 0 to 600 nm

a                                                                 b

c

Figure 6

(a-c). A growth sequence on the basal face of gibbsite in a solution with a relative supersaturation of 0.71 after the following time intervals: 0, 3 and 6 minutes, respectively. The relative height (Z range) of the features ranges from 0 to 100 nm

At a higher relative supersaturation of 1.0 (0.70 A/C), growth on a different crystal was observed to occur by continuous birth and spread as expected; Figure 7 is a growth sequence on the basal face of that crystal.

 

a                                                                  b

Figure 7 (a, b)

A growth sequence on the basal face of gibbsite in a solution with a relative supersaturation of 1.0 at 80 C after the following time intervals: 0 and 1 minutes, respectively. The relative height (Z range) of the features ranges from 0 to 400 nm

5.0 CONCLUSIONS

The results of both the kinetic study by in situ optical microscopy and the AFM observations indicate that there are two mechanisms operating during growth on the basal face of gibbsite crystals. The mechanism which dominates depends principally on the supersaturation of the liquor, but also on the seed surface structure, especially in the intermediate range of supersaturation, where there is some overlap between the two mechanisms. The mechanism of growth on the basal face of gibbsite can be summarised as follows: growth occurs by step growth at low relative supersaturations (, 0.50 A/C) and by polynuclear birth and spread mechanism at high relative superaturations (, 0.65 A/C), all at a constant C value of 200. In between these two levels of supersaturation, growth occurs by both step growth and birth and spread mechanisms, with the birth and spread mechanism becoming more dominant with increasing supersaturation.

These results apply to growth from solutions with caustic concentration of C 200 g/L. Changing the caustic concentration may affect the critical values of supersaturation for transition between step growth and birth and spread mechanism.

The observed birth and spread mechanism and anisotropic growth of nuclei on the basal plane are consistent with a proposed theory of secondary nucleation under industrially relevant conditions of supersaturation.

ACKNOWLEDGMENTS

This work has been supported by the Australian Government Cooperative Research Centre (CRC) and by the Australian Mineral Industries Research Association, and this support is gratefully acknowledged. We would like to thank Queensland Alumina Limited and Worsley Alumina Pty Ltd,, whose support made the purchase of the AFM possible.

REFERENCES

Bosbach, D, Jordan, G and Rammensee, W, (1995). Crystal growth and dissolution kinetics of gypsum and fluorite: An in situ scanning force microscope study. European Journal of Minerals , pp267-276.

Brown, N. (1972). Crystal growth and nucleation of aluminium trihydroxide from seeded caustic aluminate solutions, Journal of Crystal Growth, 12, pp39-45.

Campbell, P. A, Ester, G. R and Halfpenny, P. J. (1996). In situ studies of potassium hydrogen phthalate dissolution using scanning probe microscopy. Journal of Vacuum Science Technology B, 14(2),1373-1377.

Freij, S. J, Parkinson, G. M and Reyhani, M. M. (1998). Ex situ atomic force microscopy of gibbsite growth. To be submitted.

Freij, S. J, Parkinson, G. M and Reyhani, M. M. (1998) Direct Observation of the growth of gibbsite crystals by atomic force microscopy. To be submitted.

Gindt, R and Kern, R. (1968). Besichteder Bunsengellschaft. 72 (3), 459- .

Hillner, P. E, Manne, S, Gratz, A. J and Hansma, P. K. (1992). AFM images of dissolution and growth on a calcite crystal. Ultramicroscopy, 1387-1393.

Keller , K. W. (1986). Journal of Crystal Growth, 74, pp161-170.

King, W. R. (1973). Some studies in alumina trihydrate precipitation kinetics. Light Metals, 551-562.

Liu, X. Y., Maiwa, K and Tsukamoto, K. (1997). Journal of Chemical Physics, 106, 1870- .

Lee, M, Lincoln, F.J, Parkinson, G.M, and Smith, P. G. (1996). Microscopic and macroscopic observations of variabilities in the growth rates of gibbsite crystals. Proceedings of the Fourth Alumina Quality Workshop, Darwin, 217-226.

Lee, M and Parkinson, G.M. (1998). Growth rates of gibbsite single crystals determined using in situ optical microscopy. Journal of Crystal Growth, accepted for publication.

Misra, C and White, E. T. (1971). Kinetics of crystallisation of aluminium trihydroxide from seeded caustic aluminate solutions. Chemical Engineering Progress Symposium Series, 67(110), pp53-65.

Rossiter, D.S, Ilievski, D and Parkinson, G. M. (1999). Proceedings of the Fifth Alumina Quality Workshop, Bunbury.

Veesler, S and Boistelle, R, 1994. Growth kinetics of hydragillite Al(OH)3 from caustic soda solutions, Journal of Crystal Growth, 142, pp177-183.