Chemical systems with standard reaction enthalpies (ΔH0r) and an activation energies (Ea) of a few tens of kcal/mol are generally well described by the unimolecular reaction rate theory. The goal of this work is to determine the limits of this model by studying the kinetics of various chemical reactions with a small ΔH0r and Ea respectively. The theoretical model for the evaluation of the experimental data has been developped by Troe to determine reaction rate constants in the different pressure regimes. Based on a relaxation method an experimental technique has been specially developped to investigate chemical reactions with small ΔH0r and for Ea. A fast temperature jump, induced by a pulsed CO2 laser, disturbes instantaneously the chemical system, and the subsequent relaxation of the system is observed spectrospically. First, this experimental technique has been proved with a chemical system of known chemical kinetics, i.e. the gas phase dimerization of nitrogen dioxide to dinitrogen tetroxid. The rate constants of this system have been determined by other methods, such as flash photolysis, laser photolysis or in shock wave tubes. T-jump experiments have been carried out in the presence of helium, from 0.3 to 200 bar total pressure, corresponding to the so called reaction fall-off. This wide range of pressures, extending over 4 decades, allows the extrapolation of the reaction rate constant in the low and high pressure limits with good confidence. Additionally, mesurements in the temperature range from 255 to 273 K lead to the following reaction rate constant : Experimental limits of the technique rise from non-thermalisation effects prior to the reaction, when direct V-V energy transfer occurs between the IR-sensitizer and dinitrogen tetroxid. This transfer is particularly observed with a small sensitizer dinitrogen tetroxid ratio. The next system studied is the N2O3 NO2 + NO reaction, which has a never smaller ΔH0r. In this case, non-thermalisation effects are not observed, leading to the conclusion of direct V-V energy transfer between the sensitizer and N2O4, and not NO2. Experiments between 225 and 260 K, and 0.5 to 200 bar of total argon pressure, allow the rate constant to be determined as: Like the N2O4/NO2 system, the N2O3 NO2 + NO reaction is also fairly well described by the kinetic theory of unimolecular reactions. Differences between the calculation according to Troe's model and experimental values are indeed below 30%. Finally, reactions with much smaller ΔH0r and Ea, like the cis/trans isomerisation of acrolein and 1,3-butadiene, have been studied. Experimental results are in good accordance with the theory for the low pressure limit reaction rate constant. However, there is a large discrepancy for the high pressure limit reaction rate constant, where the theory predicts rate constant about 100 times higher than the experimentals results. Further more, the reaction rate constant also depends on the concentration of the isomer present in the mixture. This surprising result can be explained by the formation of dimers or even polymers which decrease the reaction rates. These results are however in accordance with those obtained by Bauer and al. for the syn<->anti isomerisation of methylnitrite and aziridine, as well as those obtained by Quack in the HF dimer. These authors have also observed discrespancis between the theoretical prediction and experimental results in the high pressure regime.