Table of contents

The development of structural models for predicting the external morphology of crystalline materials is presented and discussed in terms of their applications to molecular crystals. The predicted crystal morphologies of a number of molecular materials including alpha -sulphur, naphthalene, benzoic acid and hexamine are presented using the Bravais-Friedel-Donnay-Harker, attachment energy and Ising models. The results of the various models are compared both against each other and against the experimentally observed morphologies.

Elastic constants for the study case of CO₂ have been obtained in the pair-potential rigid molecule approximation, using the crystal packing programme PCK83. The non-bonded interatomic interactions are described by (exp-6-1) potential energy functions. Strain energy curves and surfaces are fitted with polynomial expressions by a least squares technique. An analysis of the results is presented, yielding values for the constants c₁₁, c₁₂ and c₄₄ that compare well with existing data.

The relationships between the molecular packing and the morphology of rubredoxin crystals, a metalloprotein, have been studied in detail through the periodic bond chains (PBC) method of Hartman and Perdok (1955). To carry out the PBC analysis the authors considered that the crystallizing particle consisted of the protein molecule plus a tightly bound layer of solvent molecules. Molecular interactions such as hydrogen bonds, van der Waals contacts, ion pairs and hydrophobic free energy have been systematically analysed. In every case, they involved direct protein-protein and protein-solvent-protein contacts. The structure of the respective growth layers (slices) and of the PBCS has been determined with the help of molecular graphics. To get a quantitative insight into crystal packing forces they have estimated, for different faces, the interaction energy of the crystallizing particle with its nearest neighbours in the respective underlying growth layers. There is an inverse correlation between this energy and the morphological importance of the faces. The role of hydrogen bonding is also discussed.

The systematic-type studies carried out on small molecule crystallization are of limited use when dealing with proteins. The reasons for this are that proteins have a much higher molecular weight, typically ranging from 10000 to 250000 daltons and are of lower symmetry than small molecules. Protein crystals are 'soft' and sensitive to small changes in external conditions since they have a high solvent content. There are small binding energies between protein molecules in the crystal lattice. This means that there are a large number of potential attachment sites which are almost as favourable as the small number of specific sites through which the ordered array is formed. Despite these problems, some general rules have been established and procedures devised that can be followed when attempting to crystallize a new protein. These rules and procedures are illustrated by using specific examples of crystals grown from monomeric and multisubunit proteins isolated from a variety of bacterial and animal sources.

In Europe some 10% of men and some 20% of women have gallstones. In most cases these stones remain silent and do not present any problems. However, about one in five of the people with gallstones do develop symptoms which necessitate treatment. Gallstones may be classified according to their composition as being cholesterol rich, mixed or pigment. Cholesterol gallstones are the most common. About 70% of gallstones in English patients are cholesterol rich. This review of studies of the formation of gallstones focuses on the process of cholesterol crystallization from bile. In bile, cholesterol is present in mixed micelles and vesicles, and one of the major problems in this field has been the isolation of these biliary cholesterol complexes without disruption, or modification, of their content. Calcium is a major component of cholesterol gallstones and is considered to play an important role in vesicle aggregation which is believed to be an important prerequisite to cholesterol crystal deposition. Both pronucleation and antinucleation factors are found in gallbladder bile. High protein concentrations and high pH have also been implicated.

The morphology of caprolactam crystals depends strongly on the solvent from which they are grown. Qualitative explanations of the observed growth forms are either based on the interaction between the solvent molecules and the growth units or on an orientational ordering of solvent molecules at the solid-liquid interface which affects the step free energy.

Hydroxyapatite (HAP) was precipitated from solutions containing dissolved proteins at a range of concentrations, and the precipitates examined by scanning electron microscopy. In the absence of protein, HAP formed granular aggregates; however, as protein concentration was increased the size of these aggregates decreased, and the surface appeared smoother. This effect was measured by the spatial grey level method of image texture analysis in which texture parameters (energy, entropy, inertia and local homogeneity) are calculated from a co-occurrence matrix. A co-occurrence matrix was calculated from each digitized scanning electron micrograph. Energy, entropy and local homogeneity were linearly correlated with the protein concentration in the original solution from which the HAP was precipitated.

The influence of the ortho and para isomers on the crystal morphology of meta-chloronitrobenzene (m-CNB) was observed for both melt-grown crystals and crystals grown from solution using acetone as a solvent. The role of the O-CNB and p-CNB isomers in modifying the m-CNB morphology is the same, irrespective of whether growth takes place from the solution or the melt. The morphological changes are therefore much more influenced by the presence of other isomers than by the acetone solvent.

An account is given of the origins, in the molecular electronic properties and crystal structure, of the large quadratic optical nonlinearities found in some organic crystals. Particular cases, in which powerful second-harmonic generation or large Pockels effects have been found, chosen to illustrate the influence of different aspects of molecular or crystal structure, are introduced. Three of these, MBANP, NMBA and DAN are described in greater detail in the light of recent experimental and theoretical work.

As part of the authors' on-going studies aimed at reaching an understanding of biologically controlled mineralization the precipitation of strontium sulphate has been investigated in the presence of hydrophobic and hydrophilic behenic acid (CH₃(CH₂)₂₀COOH) Langmuir-Blodgett (LB) films by electron microscopy and electron diffraction. The films exhibited most marked effects at low supersaturations and both nucleation and crystal growth processes were affected. In the absence of a film, tabular rhombic crystals are deposited with (001) forming the major growth face. Crystallization in the presence of hydrophobic films at low supersaturations (10 K_sp) produces thin, disc-like crystals which are nucleated on the (010) crystal face of celestite. Hydrophilic substrates of 10 K_sp and 30 K_sp produce open-floret or needle shaped crystals, respectively, deposited on the films. The diffraction patterns obtained from these crystals cannot be assigned using the orthorhombic unit cell of celestite. At higher supersaturations, three-dimensional crystals of celestite based on symmetry related (011) and (210) faces are produced which show little contact with the film surface. Substrate-ion interactions are considered in order to understand the influence of the substrate on crystal nucleation. For precipitation in the presence of the hydrophobic film it is proposed that surface topography directly affects nucleation. For precipitates formed at low supersaturations in the presence of the hydrophilic film there is no direct surface: ion spatial correlation and deformed structures of celestite are observed.

The structural and stereochemical factors involved in the oriented nucleation of CaCO₃ under compressed Langmuir monolayers have been studied by optical and electron microscopy, and X-ray and electron diffraction. The effect of headgroup charge, stereochemistry and changes in supersaturation have been investigated. At total (Ca)=9 mM, monolayers with carboxylate headgroups induce the oriented nucleation of calcite with the (11.0) axis perpendicular to the organic surface. At (Ca)=4.5 mM, oriented vaterite with the (00.1) axis normal to the monolayer, is deposited. The common features of these interactions are: Ca accumulation at the stearate headgroups, geometric matching of lattice distances (calcite only) and stereochemical correspondence between carboxylate and carbonate groups at the monolayer/crystal interface. In contrast, nucleation under positively charged octadecylamine monolayers was independent of (Ca) and gave vaterite aligned either along the (00.1) or (11.0) crystallographic axes. Although Ca binding is absent on these films, there is the possibility of stereochemical matching between bidentate binding of bicarbonate and the corresponding anions in the (00.1) and (11.0) crystal surfaces. Crystals (calcite and vaterite) grown under liquid-phase stearate monolayers were of increased size and uniformity in particle size distribution compared with those nucleated under fully compressed films. CaCO₃ crystallization under octadecanol monolayers was inhibited and crystals grown in the presence of cholesterol films were the same as in the control experiments.

Experiments were carried out to study the precipitation of calcium carbonate resulting from the mixing of aqueous solutions of calcium nitrate and sodium carbonate in a continuous-flow crystallizer at 25 degrees C in the pH range 8.5 to 10.5, reagent concentration 0.025 to 0.2 mol/l and solution residence times of 5 to 20 min. The population balance model is used to estimate nucleation, growth and agglomeration rates. Under the conditions studied, agglomeration of individual, previously formed crystals largely determines particle size and at higher reagent concentrations results in a bimodal distribution. The crystal form depends on the concentration of reagents, pH and residence time.

Measurements have been made of the growth rate of the (100) faces of ADP crystals grown under a constant supersaturation of 1% from static solution containing mixtures of MnCl₂4H₂O and KMnO₄ at pH=2.5. Under these conditions, manganese is incorporated into the crystal and the growth rate shows a strong oscillatory character in the early stages of growth. These oscillations dampen as growth progresses. X-ray transmission topography of larger single crystals that are grown under similar conditions and which also incorporate manganese ions reveals the development of strain bands in the (100) growth sector. A possible mechanism of lattice strain generation by incorporation of manganese ions, and its influence on growth rate variation in the (100) growth sector is discussed.

There have been many previous studies of the IV-III phase transition in ammonium nitrate and water is known to play a mechanistic role. This work examines the effect of other solvents and additives on the transition kinetics of single crystals and combines these data with structural information to examine the overall mechanism of the transition. All the data are consistent with a novel form of solvent mediation in which the interface between phases is a microscopic layer of ammonium nitrate solution.

A survey is given of new types of crystallographic faces and forms occurring on modulated crystals which have to be indexed with four integers (hklm). These faces have so far been found on modulated crystals with an average beta -K₂SO₄ structure which can be grown from an aqueous solution and the mineral crystal calaverite (AuTe₂) which can be grown from the melt. Interesting data on crystal growth obtained from in situ observations are presented. Among others, spiral-like patterns can be recognized. The occurrence on quasicrystals of crystal forms to be indexed with six integers (hklmno) is discussed and it is shown that these faces can in principle be explained by a generalization of the Hartman-Perdok theory. Crystal forms (hklm) and (hklmno) occurring on modulated and quasicrystals respectively can be explained by the higher than three-dimensional crystallography of de Wolff, Janner and Janssen (1977) where these crystals are considered as three-dimensional cuts out of a higher than three-dimensional truly translational invariant structure.

The structures of the serpentine minerals (Mg₃Si₂O₅(OH)₄) consist of composite layers made up of a sheet of SiO₄ tetrahedra linked to a sheet of MgO₂(OH)₄ octahedra. The dimensions of the tetrahedral and octahedral sheets do not match exactly, and in some serpentines the layers bend to accommodate this misfit. In the tiny fibrils of chrysotile asbestos, the layers curve round forming concentric cylinders, single or multiple spirals. A rare form, polygonal serpentine, has fibrils with a polygonal cross section, made up of flat layers in sectors. The number of sectors n is almost always either 15 or 30. An ideal geometrical model for the growth of sectors is described. At each new layer added around the polygonal fibre, five more unit cells are required around the perimeter than in the previous layer; the same number of extra unit cells is required around each successive cylindrical or spiral layer in chrysotile. Single Mg octahedra are added to new layers at the edges of sectors and this determines the Miller indices of the sector boundaries. The angle between the bounding faces will not in general be exactly 360/n. Minimal misfit occurs only for polygonal microstructures with 15 or 30 sectors, the number depending on the exact cell dimensions of the flat layers. For the curved layers of chrysotile, the cell dimensions predict a better and better fit as the number of sectors increases, as would be expected for cylindrical layers.

Some preliminary results of artificial growth of the incommensurately modulated mineral calaverite are presented. Using different growth techniques such as Bridgman and Kyropoulos the authors obtained relatively large single crystals of AuTe₂ (up to 50 mm in length) with facets. The crystals were characterized by use of SEM imaging and microprobe, Laue backscattering and optical goniometry. Using a four-dimensional description for the morphology, indexing of both main and satellite faces was possible.

The authors report structural results of mechanical alloying binary mixtures of pure elemental powders as a function of milling time. X-ray line broadening techniques are used to follow the lattice destabilization caused by interdiffusion of elements, which is favoured by a negative heat of mixing. Experimental evidence is outlined on the formation of highly distorted solid solutions and intermetallic compounds, which, due to the milling action, collapse to an amorphous condition when a critical density of defects is reached.

The fundamental steps involved in an electrochemical reaction are reviewed with respect to their role in crystallization at the electrode surface. The reaction system is seen to be sensitive to the flow environment and the nature of the electrode surface. The electrode potential can influence many features including reaction rate, chemical and phase composition, the extent of adsorption, orientation and texture. Aspects of electrocrystallization are illustrated by considering cases of metal deposition. Particular emphasis is placed upon the development of surface roughness and the formation of metal powders during copper deposition from aqueous acid sulphate solutions.