Influence of nitrogen-rich substrates on biogas production and on the methanogenic community under mesophilic and thermophilic conditions.

Grass silage was evaluated as a possible substrate in anaerobic digestion for generation of biogas in mesophilic and thermophilic long-term operation. Furthermore, the molecular biological parameter Metabolic Quotient (MQ) was evaluated as early warning system to predict process disturbance. Since this substrate is rich in nitrogen, high ammonia concentration of up to 2.2 g * kgFM -1 emerged. The high buffer capacity of the ammonium/ammonia system can disguise upcoming process acidification. At organic loading rates (OLR) below 1.0 kgVS * m-3 * d-1 (VS: volatile solids) for thermophilic and below 1.5 kgVS * m-3 * d-1 for mesophilic reactors, stable processes were established. With increasing OLR, the process was stressed until it broke down in the thermophilic reactors at an OLR of 3.5 kgVS * m-3 * d-1 or was stopped at an OLR of 4.5 kgVS * m-3 * d-1 in the mesophilic reactors. Mainly propionic acid accumulated in concentrations of up to 6.5 g * kgFM -1 . Due to the high buffer capacity of the reactor sludge, the chemical parameter TVA/TIC (ratio of total volatile acids to total inorganic carbon) did not clearly indicate process disturbance in advance. In contrast, the MQ indicated metabolic stress of the methanogens before process breakdown and thus showed its potential as early warning system for process breakdown. During the whole experiment, hydrogenotrophic methanogens dominated. In the thermophilic reactors, Methanoculleus IIA-2 sp. 2 and Methanothermobacter wolfeii were dominant during stable process conditions and were displaced by Methanobacterium III sp. 4, a possible new bioindicator for disturbances at these conditions. In the mesophilic reactors, mainly Methanobacterium III sp. 4 was dominant at stable, stressed and acidified processes. A hitherto uncultivated genospecies, Methanobacteriaceae genus IV(B) sp. 3 was determined as possible new bioindicator for mesophilic process disturbance.