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

includes bibliographical references and index

PhD; 2023; Centre for Sustainable Technologies

The biological routes (such as anaerobic digestion) for tapping the energy stored in
lignocellulose are centered on the lignocellulose's herbaceous (leafy) parts. The resource
abundance of ~200BT further confers an advantage to lignocellulose from the standpoint of
resource availability. The resource availability of 3.14-3.9t/ha/yr from crop residues and
9.5t/ha/yr from forest leaf litter has the potential to translate to 461-930m3 CH4/ha/yr from
crop residues and 1098-1696m3 CH4/ha/yr from forest leaf litter. Therefore, the potential to
harvest energy from leafy lignocellulose via the anaerobic digestion route is significant.
Further, lignocellulose-based fuels emit up to 83% less carbon dioxide emissions than fossilbased fuels. Despite this, lignocellulose is rarely used as the dominant feedstock in anaerobic
digesters. The significant diversity in the lignocellulose with over 3,74,000 known species and
structural variation within the same variety and even the same plant (growing in a different
environment) makes behavior generalization and uniform reactor design difficult. In addition
to this, the low bulk density of the lignocellulose (18-50kg/m3
) confers significant operational
challenges- in conventional slurry-based reactors, the feedstock is rendered afloat and results
in eventual “scum” formation. These challenges need to be overcome to recover the energy
stored in lignocellulose, primarily through anaerobic digestion.
A decentralized application of anaerobic digesters fed with locally available lignocellulosic
residues is required to capitalize on the energy production potential of the lignocellulose. This
needs a more diverse pool of substrates that can be used for feeding in anaerobic digesters
and a deeper understanding of the anaerobic conversion of the lignocellulose. This would
enable faster feedstock selection and local waste stabilization and reduce feedstock
transportation costs, which account for up to 38% of total operational costs.
The dominant fraction of lignocellulose comprises of pectin, hemicellulose, cellulose, and
lignin. A significant variation during anaerobic digestion would, therefore, be regulated by the
compositions and interaction between these sub-components. Approaching anaerobic
degradation of the lignocellulose from this standpoint would enable the development of
general patterns that can help explain the variationsin the digestion of the lignocellulose. This
is the central approach of the thesis. First, the work was directed towards tracking the physical
(bulk density changes, total solids (TS) and volatile solids (VS) loss), chemical (composition),
and fermentation (biogas production potential, biogas production rates, methanogenic
growth, and colonization) parameters and then investigating the interaction between the
parameters to identify the general patterns to explain the variation in digestion behavior. The
objective was to find discernible patterns based on the lignocellulosic feedstock’s inherent
property, which would aid in estimating the potential of lignocellulose as a substrate and
predict the behavior during anaerobic digestion and gas production. Ten lignocellulosic
feedstocks represented the broad categories of agro residues (Saccharum officinarum,
Sorghum, Zea mays, Oryza sativa), dicot tree leaves (Tecoma stans, Broussonetia papyrifera,
Senna spectabilis) and dicot land weeds (Parthenium hysterophorus, Sphagneticola trilobata,
and Synedrella nodiflora). The intent was to have a test pool representative of the diversity
typical in the local lignocellulose.
Solid-state Stratified Bed Reactor (SSBR), a leach bed reactor, operated in fed-batch mode with
little or no feedstock preparation, was selected as the reactor system for the study. Leach bed
reactor bypasses the low-density induced operational challenges associated with the
lignocellulose. This was one of the main reasons for the selection of SSBR. Therefore, 10 SSB
reactors, each fed with a single lignocellulosic feedstock, were operated in a fed-batch mode
for 88-95d.
Studies on the biomethane production potential yield ranged from 136±12L/kg VS for
Sphagneticola to 402±8L/kg VS for Broussonetia. A general BMP prediction parameter was
formulated during the study, with potential application with both monocots and dicots. The
formulated parameter concluded- feedstocks with (LPM)
(LPB)
ratio <2 and (HWE+HC)
(C+L+Ox)
>1.16, yield
>338L CH4/kg VS. These feedstocks reflected a good potential as substrates for anaerobic
digestion. [LPM = Lag Phase of Methane, LPB= Lag Phase of Biogas].
The understanding developed from the lignocellulose degradation in SSBR indicated that the
feedstocks experience simultaneous loss of the sub-components. However, the rates and
extent of the weight loss experienced by the sub-components differed. For example, hot
water extractives experienced maximum weight loss(79-95%), while lignin achieved minimum
weight loss(16-82%). During digestion, lignocellulose tendsto undergo densification naturally
(the final density was up to 26 times the feeding density). Therefore, insights from the bulk
density changes of the digesting feedstock were used to recommend the initial bulk density
of the lignocellulosic feedstock between 350-400kg/m3
. Further, in situations where the
introduction of the compressed feed is not feasible, a feeding frequency of once in four days
was found to be adequate.
Pectin and hemicellulose were identified as the dominant components conferring structural
integrity to the lignocellulosic feedstocks studied. High concentrations of these components
in the feedstock also corroborated with high weight loss and high methane gas production.
The study recorded weight loss of up to 92% VS and biogas production rates of up to 1.17L/L/d.
Further, it was observed that lignocellulose has varying rates of digestion. For most
feedstocks, VS loss, bulk density, and biogas production data exhibited a two-component fit.
Based on the VS loss inflection point SRT, in a well seeded reactor, a retention time of ~18d
for dicots and between 24-46d for agro residueshould be practiced to improve reactor space
efficiency.
Further, a dimensionless parameter based on the lignocellulosic composition (HWE+Ox+HC)
(L+C)
was evolved to assess fermentation behavior. This factor was used as a general parameter
(for use with both monocots and dicots) to explain the variation in the extent of VS loss
achieved (R
2=0.78), the rate of the VS loss (R
2=0.81), and the methane production rates
obtained (R2=0.85). Feedstocks with the (HWE+Ox+HC)
(L+C)
between 1.96-2.3 translated to VS loss
between 78-92%, the daily rate of VS flux (till inflection point) between 35-59g VS loss/kg VS
fed/d, and average methane production rate between 0.47-0.72L/L/d, and thus show
potential as primary substrates in SSBR. [HWE= hot water extractives, Ox= oxalate extractives,
HC= hemicellulose, L=lignin, C=cellulose].
Digesting feedstocks showed a high level of colonization of the methanogens. Methane
production potential of the (methanogen colonized) spent lignocellulose recorded high
activity of up to 53L CH4/kg residual TS/d. High colonization reflected the potential of the
spent lignocellulose as a natural support for methanogen-rich biofilm. In this light, agro
residues (Zea, Oryza, and Sorghum) with 36-46L CH4/kg residual TS/d and high shelf life due
to lower degradation rates were recommended.
However, the methanogenic activity of such digesting biomass significantly varied among the
lignocellulosic feedstocks. Rates varied from 8L CH4/kg residual TS/d for Senna to 53L CH4/kg
residual TS/d for Broussonetia. Results indicated a clear substrate preference for colonization.
Therefore, Specific Methanogenic Assay (SMA) was compared based on per unit of the
introduced feedstock TS to obtain a common basis to quantify the activity. This approach
factors in the implications of the differential rates of weight loss for the comparison and is
hence a more inclusive index. Comparison of the SMA per unit of the introduced feedstock
TS to lignocellulosic composition showed a preference for cellulose for colonization of
hydrogenotrophic methanogens. Feedstocks with cellulose concentrations between 27-34%
recorded HSMA values between 7-10.4L CH4/kg feed TS/d. Further, the TSMA, HSMA, and
ASMA evaluation on SRT at the VS loss inflection point showed a negative correlation with
“HWE+Ox+HC” (R
2=0.77, 0.77, and 0.56, respectively). While a high concentration of
“HWE+Ox+HC” conferred net high degradability to feedstocks. The possible cause was the
low availability of the substrate for colonization on these feedstocks. Alternatively, feedstocks
with a high concentration of these components were considered to confer an environment
more conducive to the growth of suspended bacteria, explaining the high degradation and gas
production observed for these feedstocks despite low recorded SMA activity. Studies on
determining the dominant route of methanogenesis concluded that the dominant route of
methanogenesis is dynamic during digestion stages and is also substrate specific. As a general
pattern, feedstocks with VFA flux (>45% VS loss under 3-4d) showed general dominance of the
aceticlastic route of methanogenesis.
Further, considering the high level of colonization of the methanogens on the spent
lignocellulose, spent lignocellulose was proposed as a potential inoculation source in the solidstate reactor. Furthermore, SMA was proposed to estimate the inoculation potential of the
available source of inoculum. Using the spent lignocellulose as an inoculation source also
showed the potential to further reduce the reactor size by decreasing the required volume of
the reactor to be dedicated for methanogenic inoculum.
Solid-state anaerobic digesters (SSAD) have emerged as the preferred reactor system for the
anaerobic degradation of the particulate substrates. However, the optimization of SSAD still
is inadequate. The basis for assessing the inoculation requirements, loading rates in SSAD,
recycling rate requirements, etc., in a leach bed reactor is poorly defined. Thus to move
forward in the optimization of the SSAD, “S/I analog"– kg feed TS/kg inoculum TS and
“SMA/S”– (TSMA required)/(kg feed TS) were proposed as potential parameters for making
inoculation and OLR decisions for the reactor.
As the digestion progressed, the colonization of the methanogens increased, and
concomitantly the potential of the reactor to handle the VS flux also increased. This increase in methanogenic colonization also rendered higher methane yield obtained from SSBR. A
comparison of the methane yield obtained from SSBR with BMP studies showed that under a
period of 53-73d, SSBR yield was ≥69% of BMP yield. It is important to remember that SSBR
was a scale-up from BMP (batch, liquid anaerobic digester) by a factor of 100-159. In some
feedstocks like Sphagneticola and Parthenium that experienced non-VFA induced inhibition in
BMP, SSBR seemed to be a more robust reactor system for these feedstocks and hence a better
reactor choice. No active inhibition in gas production was observed with these feedstocks in
SSBR, and bio-methane production rates of 0.23 and 0.55L/L/d, respectively, were observed.
To investigate the implications of the colonization further, the increasing methanogenic
potential of colonized methanogens on the digesting feedstocks in SSBR was compared with
the VS flux production from the digesting feedstocks. It was observed that the potential of
the colonized methanogens to handle the VS flux eventually surpassed the generated VS flux
from the digesting feedstock. This was referred to as the equalization point. This implied the
theoretical existence of the sub-starvation conditions for methanogens beyond this SRT was
interpreted as the potential to decrease the inoculum-occupied volume of the reactor or
increase the OLR. The proposed concept of equalization point opens a new direction for solidstate reactor design and operation and can significantly reduce the reactor volume required
for waste stabilization and concomitant gas production. This concept should be built upon by
extensive experimentation to quantify the implications of the colonizing methanogens and
the existence of the equalization point to estimate the limits to which the OLR can be pushed.
The proposed concept of equalization point, S/I analog to estimate the reactor seeding
requirements in SSAD, use of spent lignocellulose as alternative inoculum source, estimating
the methanation potential of spent lignocellulose using SMA to estimate its potential as
inoculum, and formulated composition based (general) prediction parameter for estimating
the achievable methane production rates, rate and extent of VS loss, across diverse
lignocellulosic feedstock in SSBR, opens a new direction for SSAD optimization, faster
feedstock selection, and for the decentralized deployment of SSAD (SSBR) for conversion of
locally available lignocellulosic feedstocks to bioenergy.