In this study the pathways involved in N2O formation over a commercial LNT catalyst are addressed, when using H2 as a reductant. For this purpose, flow microreactor experiments coupled with FT-IR under operando conditions are used. The results indicate that N2O formation occurs both at the lean-to-rich (primary N2O) and rich-to-lean (secondary N2O) transitions. Primary N2O originates at the reduction front due to the presence of partially reduced PGM sites that do not readily dissociates NO released from the stored NOx. Undissociated NO couples with N-adatoms leading to the formation of N2O. At variance, secondary N2O originates upon oxidation with NO/O2 of reducing species left adsorbed on the catalyst surface (adsorbed CO, isocyanates and possibly NH3) during the rich phase, as pointed out by FT-IR spectroscopy. The concentration of such adsorbed species is however limited and hence the formation of secondary N2O is much smaller than that of primary N2O, when using H2 as reducing agent. The emissions of N2O reduce upon increasing the temperature, and above 250 °C N2O formation is negligible. Finally the reactivity of N2O with adsorbed NOx species (nitrites) and with the actual reductants (H2 and NH3) is also investigated, to provide further indications concerning the pathways leading to N2O emissions. It is found that N2O does not react with NOx species stored downstream the reduction front; at variance both H2 and NH3 may reduce N2O to N2 and water at rather low temperatures. The role of this reaction on N2O emission is herein discussed.