Experimental results with a 50 kW solar reactor producing 10-15 g/h single-wall carbon nanotube's rich soot have shown that good quality product was obtained with helium and rather bad product with argon. This result is explained using a computational fluid dynamics model of the reactor accounting for fluid flow, heat transfer and mass transfer The bad results in argon were linked to the quenching of carbon vapor in the vaporization zone that resulted in the growth of carbon nanoparticles in spite of carbon-metal clusters (single-wall carbon nanotube precursor). By contrast, the vaporized materials (C, Ni, and Co species) at the target surface were well mixed in helium, an single-wall carbon nanotubes (SWNTs) might grow in the annealing and cooling zone at the backside of the reactor The same numerical approach may be used to design modifications of the reactor in order to favor the growth of SWNTs.