Effect of benzene doping on soot precursors formation in non-premixed flames of producer gas (PG)

Abstract

The study of soot precursors formation in the combustion of model Producer Gas (PG) from wood gasifiers with benzene as tar surrogate is presented. Experimental temperature and concentration profiles of 4 non-premixed flames—1 neat PG and 3 benzene-doped flames (0.2, 1 and 2%vol)—collected in a counterflow burner were compared with CHEMKIN simulations using 3 kinetic mechanisms of PAHs formation: ABF coupled with GRI3, and two mechanisms developed by the Institute of Combustion Technology of the German Aerospace Center labeled here as DLR2012 and DLR2018, the latter being a partial upgrade of the prior version mainly including C1-C4 oxidation optimization. The neat-PG flame did not yield detectable aromatics while only toluene is produced in the flame with 0.2%vol of benzene, suggesting a minor contribution of pathways leading to the larger aromatics. DLR2018 reproduces the correct order of magnitude for most of the compounds, but notable quantitative and qualitative deviations were observed for naphthalene (A2) in the richer-benzene flames (1 and 2%vol). This discrepancy is a consequence of DLR2018 considering the reaction between i-C4H5 and benzene as the most influential A2 formation pathway in the examined flames. DLR2018 reaction-path and sensitivity analyses were used to gain insights in the mechanism performance and the benzene influence. It was found that addition of H to phenyl producing benzene is one of the reactions chiefly affecting the concentration profiles of the measured species. Lower rates taken from literature for this reaction yielded some improvements for most of the species in both flames, including the reduction of the initial naphthalene underestimation. Additional modifications of A2 reaction rates were tested, identifying changes leading to a better agreement between experiment and simulations. This analysis aims at recognizing the factors underlying the observed disagreements in predicting some species concentration in the studied flames. The obtained insights should be included in the next iteration of PAH reaction model development.