The crystal structure of the bromo derivative of a cyclic diterpene has been determined. The crystals are of orthorhombic symmetry with a = 9.00Å, b = 31.46Å, c = 7.31Å. The structure was solved by the normal heavy atom procedure, and refined by least-squares to a residual of R = 0.064. The molecule is shown to be phyllocladan-15-yl bromoacetate, rather than to be a neo-atisirane derivative, as had been expected.

The ketoester, 3-ethoxycarbonyl-12-methoxy-5βH-18,19-bisnor-podocarpa-8,11,13-trien-4-one(80). The numbering used throughout this thesis is that proposed by J.W. Rowe (personal communication to Dr. R. C. Cambie) in 'The Common and Systematic Nomenclature of Cyclic Diterpenes', 3rd Revision, Oct. 1968, to be submitted to the IUPAC Commission on Organic Nomenclature has been prepared from 12-hydroxypodocarpa-8,11,13-trien-19-oic acid (1), via the C4 ketones (76) and (77). Both of these isomers were obtained from the methoxyalkene (11), which was isolated in a pure form during selective epoxidation of the methoxyalkene mixture (11 - 13) obtained from the decarboxylation of 12-methoxypodocarpa-8,11,13-trien-19-oic acid (2) with lead tetraacetate. This reaction has been investigated, and a mechanism accounting for the formation of nondecarboxylation products is proposed. The unsaturated ketoester (102) was prepared in low yield via the saturated C12 ketone (24), obtained from 12-hydroxypodocarpa-8,11,13-trien-19-oic acid (1) by Birch reduction of the aromatic ring. A number of ring A epoxide derivatives of the acid(1) were prepared and cleaved by a variety of reagents to give the 3-oxygenated derivatives (145), (146), and (152), together with compounds possessing a contracted ring A. Some ring B lactone derivatives of the acid (1) were prepared, and their stereochemistry and conformation were determined by n.m.r. spectroscopy. The conformation of some ring B-substituted derivatives was also determined by n.m.r. methods. The bromination of some 7-ketopodocarpa-8,11,13-triene derivatives was investigated, and reasons are advanced for the non-stereospecific reactions observed in some cases.

The first part of this thesis describes the use of several diterpenoids which are readily available in New Zealand for the synthesis of compounds possessing ambergris-type odours. Homologues and analogues of known odorants have been prepared with a view to correlating their odour properties with molecular structure. A new series of 1,3-dioxa-compounds has been found to exhibit odour properties characteristic of this class of odorants. In the second part a simple procedure for the conversion of totarol into a C-12 oxygenated derivative is reported. This derivative has been oxidatively degraded using ozone, and the degradation product has been converted into a possible intermediate for the synthesis of a ring-C system found in naturally occurring nagilactones. The rearrangement of several aryl phosphates under the action of n-butyllithium is reported.

This thesis describes the chemical, physico-chemical and structural studies of two types of copper compounds which interact with molecular oxygen in their formation. The first type is an intensely coloured species based on the ligand oxalyldihydrazide. The divalent metal and the ligand react together with simple carbonyl compounds and molecular oxygen in basic conditions to form blue species the nature of which has been the subject of conjecture for many years. This work shows that the metal is trivalent in the highly coloured states and that it acts as an oxidative catalyst with ascorbic acid. The copper (III)/copper (II) potential has been established as +0.244V by polarography. Compounds including acetaldehyde and acetone as the carbonyl component have been crystallized in monoclinic space groups. In both instances X-ray diffraction studies show that the metal is co-ordinated to a 6-5-6-5 macrocycle formed by a condensation reaction between two oxalyldihydrazide molecules and two carbonyl moieties. The co-ordination is via four deprotonated 'amide' nitrogen atoms and is of square-planar geometry. A structural study of oxalyldihydrazide has also been undertaken and comparisons are made with the co-ordinated species. The second type of compound studied is a Cu4OX6L4 cluster. It was made from a copper(I) precursor and studies with oxygen-18 gas show that formation requires oxidation to copper(II) followed by hydrolysis. Infrared evidence based on the Cu-O stretch (500-580 cm-1) is presented. Attempts to include both fluorine and iodine as the halogen component suggest that only chlorine and bromine may fill such a role.

Formaldehyde is an important industrial chemical due to its unique properties, coupled with its low cost and the basic availability of the raw materials from which it is produced. It is generated principally by the catalytic oxidation of methanol using a heated stationary catalyst at approximately atmospheric pressure. Formaldehyde is used in the manufacture of a wide variety of commercial products, the most significant of which in New Zealand, are the urea-formaldehyde resins. Such resins are used in the production of reconstituted wood products such as: particleboard; medium density fibreboard; plywood; laminated and finger jointed wood products; and in the modification of textiles and papers. Insulating building foams have also been formulated using urea-formaldehyde resins. Formaldehyde release from products containing urea-formaldehyde adhesives has been well documented in the literature over recent years. It has been shown that in some instances the levels of airborne formaldehyde inside dwellings and commercial premises may exceed various industrial threshold limit values for an eight hour working day. In New Zealand the current industrial permissible time weighted average value for formaldehyde in air is set at a maximum level of 2 ppm (2.4 mg of formaldehyde per metre3 of air). Natural levels of formaldehyde in air do exist and have been measured as being in the vicinity of 0.12 to 0.39 parts per billion. In terms of being a health hazard, formaldehyde was found to be a primary irritant of the respiratory airways and a skin sensitizer in some individuals. Some suggestion has been made that formaldehyde can also be carcinogenic. This suggestion should be treated with caution as it derives from preliminary experiments on rats and mice exposed to very high formaldehyde levels. Over the years a variety of test methods have been applied to the measurement of formaldehyde which is released from formaldehyde-bearing materials. A dynamic testing method, in the form of a wind-tunnel was adopted in this study for measuring the emission rate of formaldehyde directly from the surfaces of urea-formaldehyde bonded particleboard. Emissions were quantified in terms of the weight of formaldehyde emitted in mg from one square metre of panel surface over one hour, under the prevailing climatic conditions. It was observed that formaldehyde emissions were as great as about 12 mg h-1 m-2 from panels tested soon after manufacture. After five years of storage of the particleboard panels in ventilated conditions, the formaldehyde emission rates were measured as being between 0.1 and 1.1 mg h-1 m-2. It was apparent that formaldehyde emissions from particleboard bonded with a urea-formaldehyde resin with a urea to formaldehyde molar ratio of 1 to 1.5, reached a basal level approximately 12 months from the time of manufacture. The formaldehyde emitted over the initial 12 month period was equated to the formaldehyde which was originally present dissolved in the resin solution and the formaldehyde which was cleaved from the urea-formaldehyde polymer network during high temperature pressing. Formaldehyde lost from the particleboard after the initial 12 month storage phase was probably due to hydrolysis of the resin under ambient conditions. The latter was termed "evaporable" formaldehyde as opposed to "free" formaldehyde which was emitted in the first 12 month period. The fluctuations in the basal emission rates of formaldehyde from particleboard were attributed to changes in the microclimate associated with the particleboard. Particleboard panels with a higher density surface layer of wood tended to have higher formaldehyde emission rates in the early period of storage compared with particleboard having a lower density of wood in the surface layer. Emissions of formaldehyde from the edges of freshly sawn particleboard tended to be as much as 30 to 60 times greater than emissions from the faces of the original panels. After extended storage, emissions from the edges of particleboard tended to approximate the levels of emissions from the faces of particleboard. It was demonstrated that the presence of an air boundary layer adjacent to the particleboard surface has a significant effect on the formaldehyde emission rate from the panel surface. Below air speeds of 5 cm s-1 over the panel surface, it appeared that the rate of diffusion of formaldehyde through the air boundary layer controlled the rate of diffusion of formaldehyde from the particleboard surface into the bulk air. At air velocities in excess of 5 am s-1 the effect of the air boundary layer diminished. It was estimated that within enclosed spaces, such as houses with a relatively low air-turnover rate, the diffusive resistance of air boundary layers to formaldehyde emissions from particleboard surfaces could be significant in lowering the formaldehyde levels inside the house. The use of paints, paint-coated wallpaper and polymer-coated wallpaper to cover particleboard surfaces tended to reduce the emission rate of formaldehyde to below the detection threshold of the method of measurement (0.01 mg h-1 m-2). Some building materials other than urea-formaldehyde bonded particleboard were shown to emit measureable quantities of formaldehyde. These materials were of cellulosic-origin and emitted as much as 0.3 mg of formaldehyde h-1 m-2. The latter were known to not contain urea-fonnal6ehyde bonding agents. The levels of formaldehyde inside a showhome containing urea-formaldehyde bonded particleboard as wall and floor cladding were measured as ranging from 1.2 to 7.2 mg m-3, but these air-borne concentrations of formaldehyde decreased to between 0.7 and 1.3 mg m-3. after an interval of three months. A mathematical model was derived for the prediction of the maximum expected steady state levels of formaldehyde inside dwellings containing formaldehyde-emitting materials. The model took into account such parameters as: the emission rate of formaldehyde from the emitting surface; the surface area of the exposed emitting material; and the volumetric air flow rate through the dwelling. The model in general tended to indicate higher formaldehyde levels when tested using the parameters derived from measurements taken in the showhome. This model could conceivably find use in the design of buildings in order to minimize the levels of formaldehyde in the indoor environment. The lack of sufficient information in the literature relating to the mass transfer of formaldehyde through polymers prompted the investigation into the permeability characteristics of polymeric materials to formaldehyde. The steady state transmission rates of formaldehyde through free polymer films, such as plastic films, free paint and free polyurethane varnish coatings, were measured using a diffusion cell operated at atmospheric pressure. The mass diffusion principle associated with this permeation testing device best paralleled the real-life situation where either polymer films (plastic films) or polymer coatings (paint or varnish coatings) may be in contact with formaldehyde and water vapour, at atmospheric pressure. As a source of penetrant in these studies a dilute aqueous solution of formaldehyde was used to yield molecular formaldehyde. It was calculated that an aqueous solution containing 69 mg L-1 of formaldehyde gave rise to a concentration in air of formaldehyde, above the solution, of 1 mg m-3 at 30°C. A solution with such a concentration of dissolved formaldehyde was chosen arbitrarily for permeation studies. Levels of air-borne formaldehyde inside enclosed airspaces such as in houses lined with urea-formaldehyde bonded particleboard, with a low air turn-over, could reasonably be expected to be in the region of 0.1 to 100 mg m-3. The use of carbon-14 formaldehyde was favoured in this research over the use of carbon-12 formaldehyde. This gave rise to the possibility of permeability isotope effects. Based on the information presented in this thesis and on the experimental variability in test results, it appeared that the effects of a permeability isotope effect were negligible in comparison. The transmission rates of formaldehyde through the following types of polymer films were measured: low density polyethylene; high density polyethylene; high density-low density polyethylene blended polymer; polypropylene; poly(ethy1ene terephthalate); poly(viny1idene dichloride)-coated Poly(ethy1ene terephthalate); "Paraform", a multi-layered food packaging polymer film based on cellophane; Poly(ethylene vinyl acetate), a multi-layered food packaging polymer film; plasticised poly(vinyl chloride); polystyrene; nylon-6; and cellophane. At 30°C transmission rates of formaldehyde ranged from 1.2 x 10-7 μg cm-1 s-1 for cellophane to 1.2 x 10-l2 μg cm-1 s-1 for poly (ethylene vinyl acetate) polymer films, respectively. There was some indication that the transmission rate of formaldehyde through low density polyethylene decreased with increasing film thickness. It also appeared that an increase in the water contents of nylon-6 and cellophane films lead to an increase in the transmission rate of formaldehyde. With the exception of the poly(ethylene terephthalate)-types of polymer films and the polystyrene film, all other polymer film types exhibited typical Fickian behaviour, in that the permeation of formaldehyde through each polymer film was steady state. The permeation of formaldehyde through poly (ethylene terephtha1ate)-type films and polystyrene film appeared to be anomalous in that the mass transfer process did not appear to show Fickian-type behaviour. The transmission rates of formaldehyde through free alkyd-type paint films, free vinyl-based paint films and through free polyurethane varnish films, were measured at 30°C. The permeation of formaldehyde through free alkyd-type paint films and free polyurethane varnish films typified Fickian behaviour, in that after a short time-lag the steady state permeation of formaldehyde occurred. The transmission rates of formaldehyde through the free alkyd-types of paint film ranged from 1.8 x 10-9 μg cm-1s-1 to 3 2 x l0-9 μg cm-1 s-1 whereas transmission rates of formaldehyde through free polyurethane varnish films ranged from 4.1 x 10-10 μg cm-1 s-1 to 2.4 x 10-8 μg cm-1 s-1. It appeared that the "two-pack" or "catalytic-curing" type of polyurethane varnish film had a lower permeability to formaldehyde compared with either "one-pack" ("air-curing") or "one-pack" ("moisture-curing") varieties of polyurethane varnish. By contrast the transmission rates of formaldehyde through the vinyl-based types of paint films tended to range from 3.7 x 10-7 μg cm-1 s-1 to 6.2 x10-7 μg cm-1 s-1. A study of the temperature dependency of the permeation of formaldehyde through the following polymer film types was made: low density polyethylene; plasticised poly(vinyl chloride); polystyrene; polypropylene; high density polyethylene; cellophane; and nylon-6. The transmission rates of formaldehyde through each of these polymer film types increased with increasing temperature, over the temperature range of 30°C to 50°C, thus typifying Arrhenius behaviour. The energies of activation for the permeation of formaldehyde were calculated as ranging from 16 1 kcal mol-1 (67.4 kJ mol-1) for plasticised poly (vinyl chloride) to 33.1 kcal mol-1 (139 kJ mol-1 for cellophane, respectively. These activation energies were in general higher than the energies of activation for the permeation of other penetrants of similar molecular weight, through the same polymer types. The observed energies of activation for the permeation of formaldehyde through polymers were indicative of a strong interaction between formaldehyde and the various polymers. Desorption studies of formaldehyde from polymer films tested after permeability measurement, indicated that formaldehyde was irreversibly sorbed either onto or into the following polymer film types: low density polyethylene; poly(ethy1ene vinyl acetate); polypropylene; polystyrene; and plasticised poly(vinyl chloride). Formaldehyde was found to desorb from nylon-6 and cellophane polymer film types, indicating a different type of interaction between formaldehyde and these two polymer film types. The autoradiography of polymer films and free polymer coatings following permeability testing showed that the pattern of sorbed carbon-14 formaldehyde exhibited by each polymer film type or free polymer coating type, differed according to the type of polymer. Clustering of sorbed carbon-14 formaldehyde in polymer films containing polyethylene polymers may have been indicative of the formation of formaldehyde-formaldehyde polymers or perhaps could have been due to preferred pathways or sites for the diffusion of formaldehyde through the polymer film. The following types of polymer film tended to show a uniform sorption pattern of formaldehyde: nylon-6; cellophane; "Paraform"; free alkyd paint films; free vinyl-based paint films; and free polyurethane varnish films. The uniform appearance of the sorption pattern of carbon-14 formaldehyde may have been indicative of a higher solubility coefficient of permeation for formaldehyde in the latter polymer types. On the other hand, insufficient carbon-14 formaldehyde was sorbed by the poly(ethylene terephthalatel-type and polystyrene films in order to obtain autoradiographs. Both the poly(ethylene terephthalate)-types of films and the polystyrene film had a relatively low permeability to formaldehyde and appeared to also have a low solubilizing effect on formaldehyde. The data presented in this thesis are discussed in relation to the practical problem that initiated the study. Sufficient detailed information has been collected to enable informed decisions to be taken over the choice of formaldehyde-emitting materials used in the building and the food packaging industries.

The fully regio-controlled synthesis of C1a-C12a mono-seco colchicine 133, as well as a number of structurally interesting analogues, has been achieved. The key feature employed in our approach to troponoid compounds of this type has been the incorporation of synthetic equivalents for α-tropolone anions 84 and 87 via dehydrobrominative ring expansion of C7-substituted7-bromobicyclo[4.1.0]heptan-2,3- and -3,4-diones respectively. The hydroxyl functionality present in tropolone 143 has been found to be a suitable progenitor for the C7-acetamido group in mono-seco colchicine 133. Studies directed towards the enantio-controlled introduction of the C7-acetamide moiety via asymmetric reduction of the prochiral ketone 152 are discussed. Various strategies to effect B-ring closure in mono-seco colchicine 133 and its congeners are presented. The novel preparation and dehydrogenative dimerisation of desacetamido mono-seco colchicine 174 is reported. An efficient synthesis of the naturally occurring tropone nezukone (16) has been accomplished via the novel rearrangement of bicyclic methylene cyclopropane 209. Evidence for the intermediacy of heptafulvenes in this conversion is described.

A series of structures of light atom molecules is presented. The molecules, in the main, are physiologically active or potentially active. The molecules have been divided into five broad categories. Their structure and properties are discussed. Section 1 presents some organic natural products largely isolated during the chemotaxonomic investigation of coastal marine organisms from New Zealand and Australian waters including the first porphyrin isolated from a sponge. Also presented are the structures of an isostegane and an intermediate in the synthesis of potentially odoriferous compounds. Section 2 presents compounds related to the drug amsacrine m-AMSA, of known chemotherapeutic use and other compounds produced in the study of chemotherapeutic agents, some as templates for elaboration into bis- and tris- intercalators. Section 3 presents anthraquinone derivatives related to the synthesis of compounds of known chemotherapeutic activity. One of these is a stable ozonide. Section 4 presents the study of 4, 4'-bipyridinium cations, compounds related to paraquat and exhibiting interesting electrical properties. Section 5 presents two compounds related to the study of colchicine, the most commonly used drug in the treatment of gout and a compound of other biological interest. One is an active analog the other is an intermediate in the regiocontrolled synthesis of colchicine itself.

Molecules containing transition metal-carbon multiple bonds have been recognised as active species in catalytic reactions and key intermediates in stoichiometric reactions. It is therefore important to gain an understanding of the reactivity of these functional groups. The metal-carbon double bond is introduced in Chapter 1 with a literature review and the following two chapters are concerned with the synthesis and properties of a new class of Group VIII metal carbene complexes containing the prototype carbene ligand, =CH2. The synthetic route to the first of these complexes involves a new low-valent osmium complex OsCl(NO) (PPh3)3. The synthesis of this complex from a reformulated complex containing the novel peroxycarbonyl ligand is discussed in Chapter 2, followed by simple reactions including the synthesis of a range of acetylene, allene and heteroallene π-adducts. The reactions establish that OsCl(NO)(PPh3)3 reacts as the osmium analogue of IrCl(CO)(PPh3)2. The methylene-ligand synthesis involves the addition of diazomethane to OsCl(NO)(PPh3)3 to give the structurally characterised Os(CH2)Cl(NO)(PPh3)2. The reaction has been extended to give the first isolable ruthenium and iridium methylene complexes. The reactions of these complexes establish the nucleophilicity of the double bond and two Lewis acid adducts, Os(CH2AuI)Cl(NO)(PPh3)2 and [Os(CH2S(O)O)Cl(NO)(PPh3)2]ClO4, have been structurally characterised. The reaction with CO and isocyanides to form ketene and ketenimine ligands illustrates a different mode of reaction that is interpreted in terms of the non-innocent participation of the nitrosyl ligand. The change from nucleophilicity to electrophilicity of the double bond with a reduction of metal electron density is illustrated by reactions of Os(CH2Cl)Cl2(NO)(pph3)2 which implicate [OS(CH2)Cl2(NO)(PPh3)2]+ as an intermediate. Os(CH2Cl)Cl2(NO)(PPh3)2 is formed by addition of Cl2 to Os(CH2)Cl(NO)(PPh3)2. In Chapter 4 a different type of metal-carbon double bond is discussed. The complexes contain the relatively uncommon dichlorocarbene ligand and the synthesis of the first such iridium complex, Ir(CCl2)Cl3(PPh3)2 is described. The utility of the dichlorocarbene ligand in the synthesis of a range of carbene ligands including monochlorocarbene complexes is illustrated. The most useful single reaction of the dichlorocarbene complexes is that forming a metal-carbon triple bond. A review of the literature concerning carbyne complexes is presented in Chapter 5 and the synthesis and reactions of a number of carbyne complexes is discussed in Chapter 6. Like the methylene complexes, the carbyne complexes such as Os(CPh)Cl(CO)(PPh3)2 contain nucleophilic metal-carbon bonds and Os(CHPh)Cl2(CO)(PPh3)2 resulting from HCl addition has been structurally characterised. However, just as the reactivity can be changed with the methylene complexes, the reduction of metal electron density in complexes such as [Os(CR)Cl2 (CNR')(PPh3)2]ClO4 (R = C6H4-4-NMe2, R' = C6H4-4-Me) makes the triple bond electrophilic. Two methods for synthesising these octahedral carbyne complexes are described. One, involving a reaction of complexes such as Os(CPh)Cl(CO)(PPh3)2 with oxygen and the formation of another peroxycarbonyl ligand, the other a reaction with iodine to give complexes such as Ru(CPh)ClI2(PPh3)2. Reactions with chlorine by contrast lead to a group of reactive chlorophenyl carbene complexes. Where appropriate, the reactivity of the complexes described is interpreted in terms of frontier-orbital and charge controlled reactions using the current models for the carbene and carbyne complexes.

The reactions and transformations of atmospheric primary particles have been studied using the surface analysis techniques of X-ray Photoelectron Spectroscopy (XPS), Auger Electron Spectroscopy (AES), Scanning Electron Microscopy (SEM) and Secondary Ion Mass Spectrometry (SIMS). Particles emitted from a steel mill were studied at their point of emission, during transport through the atmosphere and after deposition. The mill was located in a coastal region removed from other major industrial particle sources which made the identification of the mill particulate emissions possible in ambient samples. The mill utilizes indigenous titanomagnetite ironsand and coal as the raw materials. There is evidence of the surface enrichment of volatile trace elements such as Zn, S, Na, K and P from the raw materials in the particulate emissions. Particle samples were collected from sites within the mill which represented different stages in the iron-making process, including the rotary kilns, electric melting furnaces and the vanadium recovery plant. This established an inventory of particulate emissions. Mill sourced particles were then identified and studied in air samples collected downwind of the mill on silver-coated nucleopore filters. AES and SIMS depth profiling studies indicated the mill particles became surface enriched in sea salt components such as S, predominately as sulfate, Cl- and Na+ after atmospheric aging. This was attributed to their coagulation with the marine derived ambient aerosol. SIMS ion imaging and mass spectral analysis suggested a characteristic "fingerprint" of the mill emissions could be distinguished in the ambient aerosol up to 8 km downwind of the mill. An experimental rig was constructed to model the interaction between the mill particles and the natural marine aerosol. A bubble nebuliser was developed to produce an artificial sea salt aerosol which was reacted with a metal powder in the fluidised bed of the rig. The metal powder was then aged under controlled relative humidity (RH) conditions. It was shown that an iron powder, after reaction and aging at 75% RH, developed an aqueous surface layer which quickly led to electrochemical corrosion, dissolution and oxidation of the particle surface. Evidence of this corrosion of metal particles occurring in the environment was found in a SEM-EDX study of the mill particles deposited on pine needles downwind of the mill. It was argued that these reactions would increase the bio-availability of the particle components. A model which incorporated these observations was developed to describe the morphogenesis of atmospheric primary particles during aging in the New Zealand environment. XPS was used to study ambient aerosols deposited on both botanical and artificial passive sampling surfaces. Plant leaves were shown to be excellent collectors of particulate material and were suited to analysis by XPS. Site differences in the atmospheric aerosol load and composition were detected on vine leaves located on a hill side (high NaCl) and at a roadside (high sulfate and silicates) in a rural area. Zn and Fe species were detected by XPS on pine needles up to 2 km downwind of the steel mill.

This thesis deals with the nucleation, growth and defect structure of anodic oxide films formed on bismuth in aqueous electrolytes. The early stages in the formation of continuous anodic layers of bismuth oxide were studied using the technique of cyclic voltammetry. During this period, the oxide, Bi2O3, covered the surface by a process of simultaneous thickening and spreading of patches. The study of oxide nucleation classified the metal surface into two different areas with different kinetics of oxide nucleation. The ratio of the two areas varied according to the history and original preparation of the surface. Film nucleation was also studied by the galvanostatic step and potentiostatic step methods. The rate of spreading of the oxide patches over the surface was shown to be controlled by the magnitude of the surface area still uncovered. Film thickening was studied using both galvanostatic and cyclic voltammetric techniques. The high field growth law i = Ā exp (BĒ) where i is the current density and Ē the field in the oxide layer, was found with parameters Ā = (1 ± 0.5) x 10-5 A cm-2 B = (2.0 ± 0.5) x 10-6 cm V-1 This value of B gives an activation distance for high-field ion transport, a* = 0.2 nm, comparable to the radius of a lattice site. These parameters were determined for very thin films in strongly alkaline electrolyte (pH 13; thickness <30 nm). At greater film thickness, cracking of the film gave a porous layer on top of a continuous barrier layer, and the apparent value of the parameter B increased to 1.7 x 10-5 cm v-1. All previous work on the bismuth anodic film thickening process has been affected to some extent by cracking of the film. In this work, cracking of the film was confirmed by microscopic observation, including scanning electron microscopy. Dissolution of the film, giving breakdown of the oxide layer and pitting of the metal, was an important phenomenon. It became particularly significant for pH <8. The thickness attained by the anodic bismuth oxide layer was limited by this process to only 4 nm at pH 5, increasing to over 200 nm at pH 13. The cathodic reduction of the oxide proceeded from the outer surface inwards, and a rough, porous metal surface resulted. A model involving electron injection from the metal into the oxide, diffusion of electrons through the film and their reaction at the outer surface has been proposed. Microscopic observation confirmed the porous nature of the electro-reduced surface. The transient conductivity of the bismuth anodic film has been investigated, and the effects of standing at open-circuit in the electrolyte or of heating in air studied. The galvanostatic method was used. The observed effects could be adequately explained as being due to the presence ofanon-stoichiometric, n-type (excess metal) layer at the film-solution interface. Cyclic voltammetry was used to investigate the possibility of nucleation of anodic bismuth oxo-halides from solutions containing halide ions. Solutions containing fluoride, chloride, bromide or iodide at pH 4-9, phosphate and phosphate with added methanol at pH 8.5 were used. Consideration of thermodynamic stability would indicate the formation of new phases, but results showed that, in the comparatively short time scale of the experiments, kinetic factors resulted in the formation of anodic Bi2O3, except when the solution contained iodide at low pH, when the oxo-iodide, BiOI, was probably formed. Room lighting had no effect on the cyclic voltammetric or galvanostatic measurements. The transient response of anodic films on bismuth to an intense flash of white light was studied. In the case of a thick anodic oxide film (240 nm), the results were interpreted in terms of photo-excitation of the film, producing electron-hole pairs which were separated by the applied field. The recombination process was best described as rate-limiting thermal excitation of trapped electrons. The lifetime of an electron in the conduction band of anodic Bi2O3 was determined: τc = 23 ms. The photo-response of a very thin anodic oxide film (4 nm) involved internal photo-emission of electrons from the metal into the oxide. Formation of a positive space charge by the injection of extra cations from the metal into the film followed. The transient photo-potential (galvanostatic experiments) or photo-current (potentiostatic experiments) showed the decay of both excess electronic and ionic space charges. The density of electron traps in the anodic oxide film was determined as NT > 3 x 1018 cm-3. When the anodic film was formed in a solution containing bromide or iodide at pH 5, an inversion of the sign of the primary photo-effect from that observed in the absence of additives (4 nm thick film) was seen, and interpreted in terms of the photo-excitation of halide ions incorporated in the film. The lifetime of a conduction electron in anodic BiOI was determined as 1.5 ± 0.2 ms, much less than in anodic Bi2O3 (23 ms). In all the studies of the photo-effect, Gauss' theorem was used to relate the observed photo-potential (galvanostatic experiments) to the photo-induced charge separation. An approximate value of the conduction electron mobility in anodic Bi2O3 was thus obtained: μ ≃ 5 x 10-8 cm2 V-1 s-1. Parameters for the evolution of hydrogen on the bare bismuth metal surface, were obtained: b = 0.11 V, log10(i0/Acm-2) = -10.3 and (∂(log10i0)/∂(PH))η = 0 where η = E – 60 pH mV.