Given the ecological problems associated to the CO2 emissions of fossil fuels, plasma technology has gained interest for conversion of CO2 into value-added products. Microwave plasmas operating at atmospheric pressure have proven to be especially interesting, due to the high gas temperatures inside the reactor (i.e. up to 6000 K) allowing for efficient thermal dissociation of CO2 into CO and O2. However, the performance of these high temperature plasmas is limited by recombination of CO back into CO2 once the gas cools down in the afterglow. In this work, we computationally investigated several quenching nozzles, developed and experimentally tested by Hecimovic et al., [1] for their ability to quickly cool the gas after the plasma, thereby quenching the CO recombination reactions. Using a 3D computational fluid dynamics model and a quasi-1D chemical kinetics model, we reveal that a reactor without nozzle lacks gas mixing between hot gas in the center and cold gas near the reactor walls. Especially at low flow rates, where there is an inherent lack of convective cooling due to the low gas flow velocity, the temperature in the afterglow remains high (between 2000 and 3000 K) for a relatively long time (in the 0.1 s range). As shown by our quasi-1D chemical kinetics model, this results in a important loss of CO due to recombination reactions. Attaching a nozzle in the effluent of the reactor induces fast gas quenching right after the plasma. Indeed, it introduces (i) more convective cooling by forcing cool gas near the walls to mix with hot gas in the center of the reactor, as well as (ii) more conductive cooling through the water-cooled walls of the nozzle. Our model shows that gas quenching and the suppression of recombination reactions have more impact at low flow rates, where recombination is the most limiting factor in the conversion process.