The burning of metallic materials in oxygen-enriched environments involves extreme temperatures and complex chemical reactions and thermodynamic processes that are not well characterised. This burning process is of significant scientific interest among materials scientists and engineers. From a fire safety standpoint, the burning of bulk metallic materials is of particular interest to many industries including medical, SCUBA, gas liquefaction, welding, foundry operations, aviation and aerospace where high pressure gaseous or liquid oxygen systems are used. Accidental fires have occurred in these systems with the consequences ranging from equipment loss to catastrophic loss of plant or system and human life. Understanding the burning mechanisms of metallic materials, knowing which materials are more susceptible to ignition and sustained burning, and gaining an appreciation of the conditions and configurations necessary to support the burning of metallic materials is of importance to the safe design, operation, and maintenance of oxygen systems.
Standardised test procedures have been developed by organisations such as NASA and ASTM to determine a metallic material's propensity to sustain burning when subjected to an overwhelming ignition source. These tests involve igniting a vertically mounted 3.2 mm diameter metallic rod at its base, and documenting its upward burning characteristics. The tests are often studied in more detail by video recording, temperature and pressure recording, interpretation and use of melting rates (i.e. the Regression Rate of the Melting Interface (RRMI)), post-test analysis of the reaction products, and mathematical and chemical modelling of the burning mechanisms. Presented in this dissertation is further information, test data and discussion providing the results from a range of newly developed experimental and analytical techniques, to augment the current knowledge base in the field.
A review of past and present literature highlighted a number of general observations concerned with the burning of pure iron and aluminium. The analysis of existing test data, temperature and pressure considerations to predict likely burning modes, and energy balance calculations raised a number of additional issues regarding the burning mechanisms of metallic rods. In particular, the question was raised as to whether or not the molten drops which detach from burning iron and aluminium rods are fully reacted, as past work has assumed. This is an especially important consideration as previous modelling work often assumed the melting rate was equal to the reaction rate, that is, as the fuel (rod) is melted, it reacts. To answer this important question, an experimental procedure was developed to catch the detaching drops in, and to drop burning rods into, a quenching media to rapidly extinguish the burning mass and remove it from the oxidising atmosphere. The tests were videotaped and the post-test samples were subjected to various microanalysis techniques to determine the extent of reaction and to infer aspects relating to the burning mechanism. A number of standard tests were also conducted with iron, aluminium and aluminium-based alloys and composites to study their burning behaviour and to generate post-test samples for microanalysis. Samples were sectioned and studied using Scanning Electron Microscopy (SEM), Electron Probe Micro-Analysis (EPMA), micro X-Ray Diffraction (XRD) and various optical microscopy techniques. To compare and contrast the iron and aluminium results, other metallic materials such as nickel, cobalt, silicon, and titanium were also studied. Samples burned in reduced gravity conditions were also investigated.
It was found, for the conditions and configurations investigated, that only a small amount of the detaching mass was reacted in the aluminium samples and a moderately larger amount was reacted in the post-test iron samples. Accordingly, through microanalysis, both systems showed large amounts of unreacted material in the detaching drops, highlighting the need to differentiate between the observed melting rate (RRMI) and the oxidation reaction rate. These results suggest that at the test conditions, the majority of the burning reactions take place heterogeneously on the surface and within the molten liquid drop. The materials studied showed a range of different burning characteristics and displayed a wide range of features when subjected to microanalysis. Qualitative models of the burning mechanisms of iron and aluminium are presented and the suitability of these models to explain features observed in the burning of other metallic materials is discussed.
Although some specific materials have unique characteristics, many features can be applied to burning metallic materials in a more general sense. Most notable of these is the apparent tendency for the molten liquid drop to absorb/adsorb high levels of dissolved or chemically reacted oxygen. The level of oxygen incorporation often exceeds that which might be predicted by stoichiometry. This may be occurring in a small layer on the molten drop surface. After detachment, any unreacted material within the drop may continue to burn, absorbing/adsorbing further oxygen. During cooling, any "excess oxygen" is evolved from the molten drop, leaving behind the stable oxides for the system. The thermodynamic and chemical properties of the specific metal/oxide system will determine how this oxygen is present (dissolved, bridge species, ionic melts etc.).
The burning system generally consists of a solid rod with a molten ball attached. The molten ball has (from the inside outwards) a solid melting core, liquid metal with oxygen, a reaction surface and oxide product often with excess oxygen. In some of the systems (e.g. Al), some vapour phase burning can/will occur, however, this is predominantly due to the rod configuration being studied. That is, the properties of the product oxide are such that it can encapsulate the fuel and allow vapourisation before exposure to oxidiser.
The attached molten drop can encircle (surround) the solid/melting rod which assists heat transfer into the solid rod. This effect is dramatically amplified in reduced gravity leading to enhanced heat transfer and faster RRMI. Boiling temperatures at elevated pressures, and other thermophysical properties of metallic materials in particular the melting temperature, the thermal conductivity of the solid and the heat of formation of the oxide, can provide insight into the burning mechanisms of cylindrical metal rods affecting the measured RRMI and threshold pressures.
All of the metallic systems investigated in this work would benefit from further experimentation and analysis. The iron system needs quenching experiments performed at other pressures. It is likely that burning at higher pressure will result in less unburned metal detaching from the burning rods and lower pressures resulting in more. A certain pressure may be reached, where a true reaction rate will be represented by the measured RRMI. Additionally, different diameter rods should be burned and quenched with similar trends expected, that is, a certain diameter may be achieved whereby the melting rate results in all of the liquid metal reacting before drop detachment occurs. This could provide a critical diameter (dcrit) at a certain pressure and a critical pressure (Pcrit) for a given diameter. These parameters would be valuable to the modelling of burning iron rods, perhaps leading to a better understanding of the test sample's geometry (configuration) and it's effect on the observed results.