In this work, Direct Numerical Simulations (DNS) have been performed on a n-heptane/air reacting mixing layer with detailed, finite-rate chemistry, to investigate primarily the interactions between aromatic chemistry and turbulent transport. Aromatic species are of critical importance for pollutant formation since their concentrations control directly the soot nucleation rates. The first objective of this DNS study is to provide an a posteriori validation of the previously-proposed reduced-order relaxation model for aromatic species [Xuan and Blanquart, Combust. Flame]. This model was designed for the chemistry tabulation of these species and their chemical source terms. The evolution of the chemical source terms and mass fractions of different aromatic species, from the DNS, will be compared to the model predictions. Results of this DNS will also provide a posteriori justification for different chemistry tabulation strategies for different categories of species. The Second objective of this DNS is to provide a database for the modeling of subgrid-scale scalar fluxes for Large-Eddy Simulations (LES). Various assumptions made by the commonly-used LES closure models in turbulent reacting flow simulations, for instance the dynamic smagorinsky model, will be examined.