The characterisation of thin films and the film/substrate interface is technologically very important. It is through characterisation that the properties of thin films are understood and this understanding leads to methods for optimising coatings for a wide range of applications. Fragmentation testing is one such characterisation technique applicable to the quantitative comparison of the mechanical properties of brittle film/compliant substrate systems. Several aspects of fragmentation testing are discussed in this thesis; the macro - and microstructural characterisation of film/substrate systems with transmission electron microscopy (TEM), the theoretical framework and experimental application of a model to account for variation in the coating thickness and fragmentation testing at elevated temperature. A novel route for preparing TEM specimens from brittle film/compliant substrate systems is demonstrated. This technique is purely mechanical and gives a means of assessing both the film microstructure and an upper limit to the film thickness. The application of the method allowed for the TEM assessment of Cr films on polyethylene teraphthalate (PET) and amorphous-AlxOy films on Cu. Conventional sample preparation of TEM samples is not possible for the Cr/PET system and it is shown that conventional preparation leads to artefacts in the AlxOy/Cu system. The most practical application of fragmentation testing is in the comparison of films produced with different compositions, deposition parameters or substrate surface conditions. To ensure a fair and truly quantitative comparison it is important to ensure only the experimental variable changes between samples. An analytical analysis of film thickness variation in a fragmentation testing experiment shows that variation should be separated into two length scales, variation over a horizontal range shorter, roughness, or longer, unevenness, than the maximum crack spacing. Roughness is important as it leads to a systematic error in the interfacial shear stress if not assessed and accounted for. The application of neighbour ratios to compensate for the effects of unevenness is shown to bring AlxOy/Cu and TiN/Cu systems into compliance with the prediction that the largest spacing between cracks is twice as large as the smallest. This prediction has never previously been observed. The failure of the neighbour ratio approach for Cr films on PET and polyimide (PI) is an indicator that quantitative calculation of interfacial shear stress is not valid for brittle film/polymer systems. The presence of out-of-plane film bending during straining of the Cr/PET and Cr/PI systems confirms this. As many coated systems are processed or operated at elevated temperatures, the application of fragmentation testing characterisation to elevated temperatures is of technological interest. Testing a model AlxOy/Cu system at room temperature and at 650C at high, 0.035 per second, and low, 0.000017 per second, strain rates showed that diffusional interfacial sliding as seen by Jobin et al. (Acta Mater., 1992) did not take place. It was additionally shown that testing at higher strain rates than found in the literature led to a broadening of the crack spacing distribution. This effect is attributed to the effective fracture toughness of the film/substrate interface being dependent on the strain rate. Elevated temperature testing of a Ti/PI adhesion layer system for flexible electronics showed a stark change in cracking behaviour and film adhesion energy with temperature. The behaviour of adhesion layers with temperature is important to understand as changes in properties like those observed could lead to unexpected failure of flexible circuitry. The reasons for the change in behaviour of the Ti coating are discussed.