Anaerobic digestion characteristics of lignocellulosic feedstocks under solid-state stratified bed mode of fermentation

Abstract

Lignocellulosic feedstocks have become potential materials for conversion to biomethane but possess varying compositions and densities (18-50kg/m3), requiring alternative approaches for evolving sustainable anaerobic digesters. Solid-state anaerobic digesters such as solid-state stratified bed reactors (SSBR) can take such low-density biomass without extensive pre-processing. However, reactor optimization is needed due to the significant difference in physical and chemical steric properties of lignocellulose. This study attempts to identify appropriate combinations of these properties to identify ideal fermentation conditions beforehand. Temporal changes in physical, chemical, and fermentation properties were recorded as a function of SRT for ten lignocellulosic feedstocks- three dicots land weeds, three dicot tree leaves, and four agro residues in an SSBR, to represent a wide variety of feedstock in the test pool. The rates of degradation described by loss of TS, VS, and gain in bulk density indicate that for most feedstocks, there is an initial rapid weight loss ascribed to loss of extractives and some hemicellulose and an initial rapid gain in bulk density. After this period, decomposition slows down. This turning point, termed the inflection point, appears to be a good representation of economic or ideal SRT (18-25d in dicots and 46-60d in agro residues). The degradation behavior pattern of lignocellulose in SSBR- the extent of degradation, rate of degradation, and achievable average methane production rates could be predicted well by knowing the composition of feedstocks in terms of hot water extractable (HWE), oxalate extractable pectin (Ox), hemicellulose (HC), cellulose (C) and lignin (L) content expressed as ((HWE+Ox+Hc))/((C+L)), with R2=0.78, 0.81, and 0.85, respectively. It is the first time a single parameter has been able to predict multiple degradation behavior for diverse lignocellulose. At the same time, methanogenesis is addressed with a biofilm bed on the digesting feedstock. These digesting (spent) feedstocks had high specific methanogenic assay levels reaching 20L CH4/kg residual TS/d (hydrogenotrophic) and 35L CH4/kg residual TS/d (aceticlastic). Feedstocks with cellulose concentration >27% of TS recorded methane production potential between 7-10.4L/kg feed TS/d through the hydrogenotrophic route of methanogenesis. In addition, the TSMA, HSMA, and ASMA evaluation on SRT at the VS loss inflection point showed a negative correlation with “HWE+Ox+HC” (R2=0.77, 0.77, and 0.56, respectively), while a high concentration of “HWE+Ox+HC” conferred net high degradability to feedstocks. A new approach to seeding- a surrogate to substrate to inoculum ratio (S/I) is proposed for SSAD seeded with solid inoculum source after its SMA has been quantified – “(Kg TS fed/kg digestate TS)” and “SMA/S” (L TSMA required/kg TS fed). These developments show the potential to use new and untried feedstocks (and combinations) while avoiding various intermediate stages of feedstock trials and validation and end-of-life use of spent lignocellulose as high-activity biofilm support. On comparing SSBR with BMP, it was found that SSBR yields reached ≥69% of BMP yield but needed~ 53-73d to catch up. This, however, can be improved by introducing compacted feed and increasing the substrate contact with the methanogens. Methane yields estimated from BMP could be predicted with the formulated parameter- ((HWE+HC))/((C+L+Ox)). It is the first time such a close level of biomethane production potential has been predicted using the physicochemical properties of diverse feedstocks.