The objective of this thesis is to describe the mixing in high temperature gas phase reactions.The Selective Non-Catalytic Reduction of NOx (referred as the SNR process) using NH3 as reductant was chosen as reaction system. This in-furnace denitrification process is made at around 1200 - 1300 K by injection of NH3 with carrier gas into the flue gas. NH3 can react with NO and form N2, but a competing reaction path is the oxidation of NH3 to NO.The SNR process is briefly described and it is shown by chemical kinetic modelling that OH radicals under the present conditions will initiate the reaction of NH3 by formation of NH2 and NH radicals.Mixing in reacting gas phase systems is described by an empirical mixing model (the droplet diffusion model). The mixing process is separated into macro- and micromixing. The macromixing is assumed to be ideal while the micromixing is modelled by molecular diffusion. The SNR process is simulated using the mixing model and an empirical kinetic model based on laboratory experiments.A bench scale reactor set-up has been built using a natural gas burner to provide the main reaction gas. The set-up has been used to perform an experimental investigation of the mixing in the SNR process using injection of NH3 with carrier gas into the flue gas in crossflow by a quartz nozzle.Experiments were made with variation in NH3 flow, carrier gas flow, carrier gas composition (O2 concentration) and reactor temperature. Natural gas has been used as an addition to the injected gas as well.The effects of the NH3 flow and natural gas addition were as expected from earlier studies in laboratory reactors and pilot plants.The experiments indicates that the SNR process was only dependent on the O2 concentration in the flue gas without any effect due to variation of the O2 concentrations in the injected gas between 0 - 20 vol%.Using a nozzle with a diameter of 1.9 mm the reduction of NO is dependent on the carrier gas flow for temperatures above 1200 K (1100 K when natural gas is added).It is shown that this effect can not be described by macromixing using a simple reactor model. The difference in the NO outlet concentration for varied carrier gas flow seems to have a maximum at 1350 K and is then decreasing for higher temperatures. This is in good agreement with an analysis of the micromixing effects.The mixing effect observed in the experiments can be described by the momentum ratio between the injected jet and the flue gas in crossflow both for the 1.0 mm and 1.9 mm nozzle, indicating that for momentum ratios above 30 there is no further improvement of the mixing. For decreasing momentum ratios below 30 the NO outlet concentration is increasing for temperatures above 1200 K. For temperatures below 1200 K the NO outlet concentration is unaffected because of lower reaction rates.The droplet diffusion model is used to model the experimental results and it can describe the influence of the carrier gas flow with a successful result.