In order to investigate the factors contributing to detonation failure, the velocity-composition relationships of the following gaseous mixtures near the limiting concentrations were investigated in a one inch diameter cylindrical tube: System SH = 2H2 + 02 separately diluted with H2, 02, He, A, CO2 and• NH3; System SD = 2D2 + 02 diluted with D2, 02, He and A; and System Sc = C2N2 + 02 diluted with C2N2, 02, He and A. In all cases, detonations were initiated by passing a detonation in a stoichiometric (2:1) hydrogen-oxygen mixture into the test mixture. Velocities were determined by measuring the time taken for the detonation front to pass between successive detection stations, a known distance apart. In additions several hitherto unobserved phenomena were noted: (i) Mixtures diluted with argon showed considerably wider detonation limits than when diluted with helium despite the fact that they should be hydrodynamically similar. In the same way, mixtures containing hydrogen showed wider detonation limits than corresponding mixtures containing deuterium. These mass effects could only be partly explained by existing theories for detonation failure. Other explanations are examined. (ii) Reproducible fluctuations in the detonation velocity-composition relationships were found near the limiting concentrations. These were found to be substantially due to a spin head following the same path down the detonation tube on successive occasions, thereby causing a displacement in the measured time intervals. (iii) In some mixtures containing hydrogen or deuterium, a second, lower velocity regime was found to co-exist with the "normal" detonation velocity regime which alone prevailed throughout the entire composition range. The measured velocities in the lower velocity regime were generally about 30% lower than in the "normal" regime. There were strong indications that this was due to the formation of hydrogen (deuterium) peroxide in the detonation front. Possible explanations for, and implications of, these novel phenomena are discussed.

An experimental investigation has been made of the effect of mass transfer on the heat transfer and drag force for evaporating and burning liquid drops. An experimental technique was developed which allowed freely falling drops of size less than 0.5 m.m. to be studied. A photographic record of a drop's flight down a furnace containing stationary air was obtained for the liquids: - water, heptane, pentane, benzene, ethyl alcohol and methyl alcohol. Experiments were carried out at varying furnace temperatures up to 1000°C; all the fuel drops ignited at about 700°C, and above this temperature burning drops with envelope flames were studied. From the drop size and distance-time records, the Nusselt number, drag coefficient and Reynolds number were calculated. For drops at low evaporation rates, the values for the Nusselt number were close to the values obtained by previous investigators who had used relatively large suspended drops evaporating into low temperatures. Ls the rate of evaporation increased, the values for the Nusselt number decreased. Similarly for burning drops it was found that as the mass transfer rate increased, the Nosselt values were lowered. The drag coefficient values for drops at low evaporation rates were close to the standard curve for solid spheres. As the intensity of mass transfer increased, the drag coefficient values were reduced. From the experimental results it has been shown how the reduction of the Nusselt number and drag coefficient can be correlated with the intensity of mass transfer.

It is found that in the low temperature gas phase oxidation of hydro-carbons, within specific regions of temperature and pressure, a periodic process may be set up which shows itself as a sequence of low intensity ignitions (cool flames). In an attempt to decide between alternative explanations of this oscillatory behaviour, the cool flame reaction of ether/air mixtures has been studied in a continuous stirred tank reactor employing variable heat transfer. Under such conditions it has been found possible to establish a continuous sequence of cool flames of constant amplitude and frequency. The effects of various experimental parameters have been investigated and it has been shown that heat transfer factors greatly affect the behaviour of the periodic process. A thermo-kinetic model has been postulated, and a theoretical analysis of the stability of the resulting equations has been carried out. Substitution of estimated parameter values in the stability equations has shown that a satisfactory agreement exists between theoretically predicted behaviour and experimental results. Accordingly the cool flame process, as observed under conditions of continuous flow, is postulated as being thereto-kinetic in nature involving the decomposition of a semi-stable active intermediate in a periodic manner.

A study of the reaction of oxygen with pyrolytic carbon filaments at 50)1 pressure showed a well defined maximum in the rate and the hysteresis effect, in the temperature range 900-2000°C. The hysteresis changed sense at high temperatures, about 1500°C, giving hysteresis above in the same sense as found earlier at 0.76u. This change over in the hysteresis was shown to occur at lower temperatures for lower pressures. It was also shown that differences in the quantities of carbon reacted and in filaments used were not responsible for different hysteresis effects observed previously. The Nagle Strickland-Constable theory was found to give a considerably better prediction of the maximum and minimum if account was taken of the variation of active sites as found by the hysteresis results. A simple mechanism is proposed based on annealing and creation of active sites which gives a good fit to the hysteresis data at 50 and 0.76u. Reaction in N20 was not found to affect subsequent reaction in 02 and vice versa, suggesting different active sites are involved. Various additives were added to the filament. Boron and Silicon were found to be without effect, while Fe, W, Zr, U, Mo and Pt all catalysed the reaction. The basic mechanism of the reaction did not appear to be changed fundamentally by catalysis, maxima in the rates still being evident. There was also a strong suggestion of hysteresis. The oxygen balances suggested in most cases that lower oxides still existed on the surface, but at the carbon-metal interface it was thought likely free metal or carbides were the active agents. The experimental evidence strongly suggests an electronic mode of catalysis.

Foaming and non-foaming systems have been distilled in a sieve plate column. It has been confirmed that the plate efficiency of the foaming (surface tension positive) system is higher than that of the non-foaming (surface tension negative) system under foam inducing conditions. At high vapour rates and low weir heights a liquid phase dispersed 'spray regime' is encountered. Under these conditions the surface tension positive system gives a lower plate performance than the negative system. It is thought that the difference is caused by the higher proportion of small drops present in the dispersion for the negative system. For both types of system the performance in the spray regime is found to increase with decrease in the absolute value of surface tension. This is in qualitative agreement with an analogy based on spray atomisation, and consistent with the effect of surface tension on drop size. A model is proposed for the spray regime which assumes a distribution of drop sizes projected upwards at initial velocities which are a function of drop size. The model has been used to deduce values for the drop size distribution parameters from the measured dispersion density profiles. It is found that the deduced mean drop size increases with surface tension. The dependence is similar to that predicted by the spray analogy. Values of plate performance in the spray regime estimated from the deduced drop size distribution with the help of the model are close to the measured values of plate performance.

Longitudinal dispersion of gas has been studied in two straight tubes of length 24m. and diameters 0.684 cm and 2.77 cm. respectively. Responses to pulse stimuli were recorded with a katharometer. Flowing Nitrogen was traced by Ethylene, Argon or Helium, to study the effect of Schmidt number over the transition from laminar to turbulent flow. Results are presented for 200< Re< 10⁴, and a theoretical study of dispersion has been made for the whole turbulent regime. For Re < 500 results confirm the Taylor—Aris solution for the dispersion coefficient, DL. For larger laminar Reynolds numbers results in the small tube showed better agreement with the theory than in the larger one. Dispersion in turbulent flow was in general described by a single dispersion coefficient. Values of the dispersion parameter (DL/ūd) were higher than predicted by theory (up to Re = 10⁴) and showed less variation with Schmidt number. Over a small region in the transition regime, response curves could not be described fully by a single dispersion coefficient because they were asymmetrical. This phenomenon has been analysed qualitatively in terms of a proposed model of transitional flow, involving intermittent slugs of turbulence, and the resulting secondary flow patterns.

Certain components of fuel oil ash have been tested against the refractory oxides silica and alumina. The solubility of alumina in vanadium pentoxide ranges from 1.4% at 750°C to 3.3% at 1050°C. The addition of sodium sulphate to this melt causes the solubility to rise to a slight maximum at 14 mole % Na2SO4 and from there decrease steadily to zero in the pure salt. The solubility of silica in vanadium pentoxide ranges from 1.8% at 750°C to 2.9% at 1050°C and the addition of sodium sulphate to this melt causes solubility to decrease until 18 mole % Na2SO4, from there it increases to a maximum at 58 mole % Na2SO4 and decreases to almost zero in the pure salt. The corrosion rates of alumina and silica in Na2SO4/V205 melts have been studied. The addition of sodium sulphate to vanadium pentoxide increases the corrosivity of the latter. The addition of vanadium pentoxide to sodium sulphate increases the corrosivity of the latter. At compositions between 30 and 70 mole % Na2SO4 in V205 corrosion rates are similar for both refractories. The diffusion coefficients for silica and alumina in these melts have been evaluated and for both systems a maximum occurs at 30 mole % Na2SO4. The kinetics of the reaction between sodium sulphate and vanadium pentoxide to produce vanadates have been investigated. Activation energies vary between 20 and 32 kcal/mole. Attempts were made to hot press firebrick to theoretical density. Results show that this should be possible at 1350°C under 1.15 kg/mm2. The mechanism is viscous flow, above 1100°C with an energy of activation of 55 kcal/mole.

Frossling's15 series solution of the boundary layer equations for forced convective transfer around a general axi-symmetric body cannot be applied to spheroids, because the required series expansions cannot be obtained in terms of the cartesian type boundary layer co-ordinates commonly used. To overcome these difficulties, the boundary layer equations were derived in orthogonal curvilinear co-ordinates. A new series solution was developed in which these equations were reduced to two infinite sets of ordinary differential equations by the use of series expansions for the velocity and temperature or concentration functions in powers of the curvilinear co-ordinate along the surface. A programme to generate and solve the general ordinary differential equation of each set was obtained in a form which can be used for all axi-symmetric bodies. Solutions of the first four or five equations of each set were obtained for spheres, discs, and oblate spheroids with width to height ratios of 0.8125, 0.625, 0.4375 and 0.25 at Prandtl or Schmidt numbers between 0.7 and 100. The exact solution predicts accurately, local rates of transfer up to θ a 70° for spheres and up to θ - 50° for spheroids, where θ is the parametric angle from the front stagnation point. To measure local rates of mass transfer from Beg's double exposure photographs of the sublimation of naphthalene oblate spheroids in air,a method was developed in which the spheroidal angle of points on the surface can be determined without projection of the whole image. Higher magnification and, therefore, more accurate measurements were obtained than with the original method by Beg14. Comparison of the local rates predicted by the exact solution, and the approximate solutions by Beg14 and Rojey91 revealed that Rojey's solution predicted local rates equally as well as tha exact solution for all oblate spheroidal shapes. 14. Beg.S.A. Ph.D. thesis, Imperial College, London 1966. 15. Frossling N. N.A.C.A. T.M. 1432 91. Rojey A. M.Sc. thesis, Imperial College, London 1967.

As a first step in the study of dumping or weeping, a theoretical model for the mechanism of bubble formation has been set up. Equations have been derived for pressures involved in a single orifice bubbling system and for radial expansion of a spherical bubble. The theory has known limitations; nevertheless a fair agreement with the experiments has been found over a certain range in the 'dumping region'. A simplified model for behaviour of a sieve tray, based on the single orifice theory, has been suggested. A simplified criterion for dumping in a single orifice or a sieve tray is described. It is shown that the mean pressure in the gas chamber under a plate and the liquid pressure behind a bubble rising above a plate has considerable effects in the dumping studies. A fair qualitative analysis of the behaviour of a system for predicting different degrees in dumping rate has been made possible using the present theory. Experiments have been carried out for measuring bubble frequency, pressure fluctuations in gas chamber and dumping rates. High speed cine photographs have been taken both from a three dimensional apparatus and from a two dimensional apparatus, the latter mainly for distinct observations of the dumping phenomenon.

The Maximum Common Substructure (MCS), along with numerous graph theory techniques, has been used widely in chemoinformatics. A topic which has been studied at Sheffield is the hyperstructure concept - a chemical definition of a superstructure, which represents the graph theoretic union of several molecules. This technique however, has been poorly studied in the context of similarity-based virtual screening. Most hyperstructure literature to date has focused on either construction methodology, or property prediction on small datasets of compounds. The work in this thesis is divided into two parts. The first part describes a method for constructing hyperstructures, and then describes the application of a hyperstructure in similarity searching in large compound datasets, comparing it with extended connectivity fingerprint and MCS similarity. Since hyperstructures performed significantly worse than fingerprints, additional work is described concerning various weighting schemes of hyperstructures. Due to the poor performance of hyperstructure and MCS screening compared to fingerprints, it was questioned whether the type of maximum common substructure algorithm and type had an influence. A series of MCS algorithms and types were compared for both speed, MCS size, and virtual screening ability. A topologically-constrained variant of the MCS was found to be competitive with fingerprints, and fusion of the two techniques overall improved active compound recall.