Lactose is the major carbohydrate in milk. The presence of lactose in whey constitutes a significant pollution problem for dairy factories. At the same time, there is an increasing market for high quality crystalline lactose. The main problem of lactose crystallisation, compared to sucrose, which is also a disaccharide, is that it is very slow, unpredictable and cannot easily be controlled. Compared to sucrose crystallisation, which has been extensively studied, lactose crystallisation lacks the fundamental research to identify the mechanisms of growth and effect of additives. An important difference from most other crystal growth systems is that ([alpha]-lactose hydrate crystals never grow from a pure environment; their growth environment always contains beta lactose. [alpha]-lactose monohydrate crystallises much more slowly because of the presence of [beta]- lactose in all solutions. Although there have been some studies on growth rates and the effect of additives, there has not been any reported work on the fundamentals of lactose crystallisation and the mechanisms that operate on the molecular level. The aim of this thesis is to gain a greater understanding at the fundamental processes, which occur at the molecular level during the crystallisation of lactose, in order to improve control at a macroscopic level. / The growth rates of the dominant crystallographic faces have been measured in situ, at three temperatures and over a wide range of supersaturation. The mean growth rates of faces were proportional to the power of between 2.5-3.1 of the relative supersaturation. The rate constants and the activation energies were calculated for four faces. The [alpha]-lactose monohydrate crystals grown in aqueous solutions exhibited growth rate dispersion. Crystals of similar size displayed almost 10 fold difference in the growth rate grown under identical conditions for all the faces. Growth rate dispersion increases with increasing growth rate and supersaturation for all the faces. The variance in the GRD for the (0 10) face is twice the variance of the GRD of the (110) and (100) faces and ten times higher than the (0 11) face at different supersaturations and temperatures. The influence of [beta]-lactose on the morphology of [alpha]-lactose monohydrate crystals has been investigated by crystallising [alpha]-lactose monohydrate from supersaturated DMSO ethanol solutions. The slowness of mutarotation in DMSO allowed preparation of saturated solutions with a fixed, chosen [beta]-lactose content. It was found that [beta]-lactose significantly influences the morphology of [alpha]- lactose monohydrate crystals grown from DMSO solution. At low concentrations of [beta]-lactose, the fastest growing face is the (011) face resulting in long thin prismatic crystals. At higher [beta]-lactose concentrations, the main growth occurs in the b direction and the (020) face becomes the fastest growing face (since the (011) face is blocked by [beta]-lactose), producing pyramid and tomahawk shaped crystals. / Molecular modeling was used to calculate morphologies of lactose crystals, thereby defining the surface energies of specific faces, and to calculate the energies of interactions between these faces and [beta]-lactose molecules. It was found that as the replacement energy of [beta]-lactose increased, the likelihood of [beta]-lactose to dock onto faces decreased and therefore the growth rate increased. The attachment energy of a new layer of [alpha]-lactose monohydrate to the faces containing [beta]-lactose was calculated for the (010) and (011) faces. For the (0 10) face, the attachment energy of a new layer was found to be lower than the attachment energy onto a pure lactose surface, meaning slower growth rates when [beta]-lactose was incorporated into the surface. For the (011) face, attachment energy calculations failed to predict the slower growth rates of this face in the presence of [beta]-lactose. AFM investigation of [alpha]-lactose monohydrate crystals produced very useful information about the surface characteristics of the different faces of the [alpha]-lactose monohydrate crystal. The growth of the (010) face of the crystal occurs by the lateral addition of growth layers. Steps are 2 nm high (unit cell height in the b direction) and emanate from double spirals, which usually occurred at the centre of the face. Double spirals rotate clockwise on the (010) face, while the direction of spirals is counterclockwise on the (010) face. A polygonised double spiral, showing anisotropy in the velocity of stepswas observed at the centre of the prism-shaped a-lactose monohydrate crystals grown in the presence of 5 and 10 % [beta]-lactose. / The mean spacing of the steps parallel to the (011) face is larger than those parallel to the (100) face, indicating higher growth rates of the (011 )face. The edge free energy of the (011) face is 6.6 times larger than the (100) face in the presence of 5% [beta]-lactose. Increase of [beta]-lactose content from 5% to 10 % decreases the edge free energy of the growth unit on a step parallel to the (011) face by 10 %. Tomahawk-shaped [alpha]-lactose monohydrate crystals produced from aqueous solutions where the [beta]-lactose content of the growth solution is about 60 % have shown clockwise double spirals as the source of unit cell high steps on the (010) face of the crystal. However , the spirals are more circular than polygonised, unlike the prism shaped crystals and the mean step spacing of the (011) face is less than the steps parallel to the (110) face, indicating the growth rate reducing effect of [beta]-lactose on the (011) face. The (100) face of the [alpha]-lactose monohydrate crystal grows by step advancement in relative supersaturations of up to 3.1. Steps are 0.8 nm high and parallel to the c rection. Above this supersaturation, rectangular shaped two-dimensional nuclei, 10 nm high, were observed. The (011) face of the crystal grown at low supersaturations (s= 2.1) displayed a very rough surface with no steps, covered by 4-10nm high and 100-200[micro]m wide formations. Triangular shaped macrosteps were observed when the crystal was grown in solutions with s=3.1. In situ AFM investigation of the (010) face (T = 20[degree]C and s = 1.18) has shown that growth occurs by lateral addition of growth units into steps emanated by double spirals. / The growth rate of the (010) face from in situ AFM growth experiments was calculated to be 1.25 gm/min. The growth rate of crystals grown in the in situ optical growth cell under identical conditions was 0.69 pm/min. The difference in growth rates can be attributed to the size difference of seed c stals used. The (010) face of a [alpha]-lactosemonohydrate crystal grown at 22.4 C and s=1.31 displayed triangular-shaped growth fronts parallel to the (011) face. The steps parallel to the (O11) face grow in a triangular shape, and spaces between triangles are filled by growth units until the end of the macrosteps is reached. No such formations were observed on steps parallel to the (110) face. Formation of macrosteps, 4-6 nm high, emanating from another spiral present on the surface was also observed on the (010) face of a crystal grown under these conditions.

Proteins are important biological products with unique functions, annually produced at the hundreds of millions of dollars value on a worldwide basis. The application of crystallization for these materials primarily was led by structural biologists and crystallographers who are keen on obtaining large and well-ordered crystals for protein structure determination via X-ray diffraction. Usually for this, crystallization is done on a small scale by vapor diffusion using a supersaturated solution of the material. In the past decades, production crystallization has slowly received increasing attention for the large-scale recovery of proteins. Among the numerous products, an industrial enzyme (alpha-amylase) that is extensively involved in food processing and laundry products was chosen for examination due to the lack of relevant data in the literature and the potential industrial interest in crystallizing this material. The chosen alpha-amylase is a product of Genencor International (the Danisco division) and is derived from a genetically modified Bacillus licheniformis. In parallel to the underlying principles that govern the bulk crystallization of small molecules, the broad topics of investigation for this macromolecular material included determination of solubility, studies of nucleation thresholds, and investigation of crystal growth kinetics and special phenomena accompanying the crystallization process. All these studies were preceded by a series of characterization tests conducted for the material. On the whole, this study aimed to extend the existing fundamental knowledge of bulk crystallization for biological macromolecules. Specifically, it intended to enrich the solubility and crystallization kinetic data for the alpha-amylase. The experimental data of this study were all obtained at conditions in line with industrial practice, which included the use of moderate temperatures, mild pH conditions and simple inorganic salts ((NH4)2SO4, Na2SO4 and NaCl) in order for the findings to be transferred to the industry directly. In a 20 mM sodium phosphate buffer (with no added salts), alpha-amylase solubility increased with solvent temperature and had a minimum at pH between 6.4 and 7.1. A generalized equation (as a function of pH and temperature) was obtained to correlate the data. The three inorganic salts examined affected the alpha-amylase solubility in a different manner, both qualitatively and quantitatively. Evidently, the interaction effect of a salt varied with solution pH. This confirms the importance of studying solubility with the two or more condition parameters at the same time. With relevance to crystal growth, the metastable region of the material was relatively wide at (NH4)2SO4 and Na2SO4 concentrations corresponding to maximum solubility. For example, σSNT was 1.2  0.2 in solutions with 5 wt% ammonium sulfate at pH 7.0 and 25oC. A wide metastability range is useful for the practical operation of batch crystallizers as nucleation can be minimized. This range, however, diminished as the salt concentration increased beyond the maximum solubility points, imposing a limit on the range of salt concentration favorable for growth processes. In systems with no added salts at pH 7.0, the solution metastability was slightly higher at 10oC than at 40oC. This would suggest a future further examination of the salt system at a lower temperature, say of 10oC. To develop a batch crystallizer, the growth kinetic data of the material have to be known. Throughout the growth studies, the alpha-amylase crystals obtained were lozenge-shaped thin plates. Apparently, habit was not influenced by the crystallization conditions chosen. Similar to other proteins crystallized in bulk, the growth rate of alpha-amylase demonstrated a second-order dependence upon supersaturation. Importantly, the alpha-amylase demonstrated crystal growth rate dispersion (GRD) under all the conditions tested. To simplify the analysis of growth kinetic results, the seed crystals used were common history (CH) seeds whose growth rates are proportional to their sizes. The spread of growth rates (CV) was 0.54 for the unsieved CH seeds used. Due to GRD, growth rate coefficient data varied with crystal size. For instance, in solutions containing 5 wt% ammonium sulfate at pH 7.0 and 25oC, the growth rate coefficient for seed crystals initially at 20 m was 2.47 m/hr. This order of magnitude was equivalent to that of many other proteins. Although being small, industrial crystallization was feasible with these kinetics, as demonstrated by the sample design calculations included. To improve the design, it is recommended to further examine the solubility, metastability and growth kinetics of the above system at other temperatures to obtain a wider growth rate range. As the important phenomenon of growth rate dispersion has seldom been examined for protein and enzyme materials in the crystallization literature, this study is a significant contribution to this area.