Content uploaded by Rajinikanth Rajagopal
Author content
All content in this area was uploaded by Rajinikanth Rajagopal on Apr 01, 2019
Content may be subject to copyright.
* Corresponding author:
Rajinikanth Rajagopal
email: rajinikanth.rajagopal@canada.ca
Detritus / Volume 05 - 2019 / pages 66-74
https://doi.org/10.31025/2611-4135/2019.13785
© 2018 Cisa Publisher. Open access article under CC BY-NC-ND license
HIGH-SOLID ANAEROBIC CO-DIGESTION OF FOOD WASTE AND
DAIRY MANURE: A PILOT SCALE STUDY AT LOW-TO-MODERATE
TEMPERATURE CONDITIONS
Rajinikanth Rajagopal 1,*, Bernard Goyette 1 and Jean-François Hince 2
1
Sherbrooke Research and Development Center, Agriculture and Agri-Food Canada, 2000 College Street, Sherbrooke,
Québec J1M 0C8, Canada
2
Bio-Terre Systems, 740 rue Galt Ouest, Sherbrooke, Québec J1H 1Z3, Canada
Article Info:
Received:
30 November 2018
Revised:
22 February 2019
Accepted:
22 February 2019
Available online:
31 March 2019
Keywords:
High-solid anaerobic digestion
Biomass adaptation
Fruit and vegetable waste
Low-to-moderate temperature
Methane yield
ABSTRACT
Treating organic solid wastes economically is a challenge, predominantly in cold
and high-altitude regions. Objective of this research was to determine the operating
strategies to reduce the start-up phase of high-solid anaerobic digestion (HSAD) pro-
cess and to improve the digestion of food waste (mainly fruits and vegetable wastes
[FVW]) with or without animal manure in a low-cost AD system at 20-25°C. In addition,
this study aimed to obtain the basic design criteria for starting up of scaled-up HSAD
system using adapted liquid inoculum. Inoculum to feedstock ratio was varied from
6:1 to 3:1. The organic loading rate (OLR) expressed as volatile solids (VS) and oper-
ational cycle length was varied from 0.44 -2.1 KgVS Kginoculum
-1 d-1 and 33 -14d, respec-
tively. Obtained results show that methane (CH4) production from FVW was feasible
at low-to-moderate temperature and specic methane yield of 0.4-0.6 L gVS
-1 was ob-
served even at high OLR. CH4 conversion rates and its quality were not affected, while
maintaining the operational stability (e.g. no acidication or VFA accumulations).
CH4 content reached over 60% and remained almost steady. Results also suggest
that HSAD process at 25°C is comparatively efcient in saving heat energy and at the
same time obtains the CH4 values close to mesophilic conditions. This means that
the smaller size digester (in the case of HSAD) is preferred as there is no waste dilu-
tion involved and also suitable for cold countries. Using this concept, livestock pro-
ducers can play a role in reducing GHG emissions while also earning C-offset credits.
1. INTRODUCTION
Mining bioenergy from biomasses is an effective alter-
native energy resource that can be used in an environmen-
tal friendly way and requires less energy production (Zheng
et al., 2012). Various biomasses derived from the carbona-
ceous wastes of human, livestock animals and natural re-
sources that could be utilized as renewable energy resourc-
es. According to Environment Canada (2013), there has
been growing interest in managing the organic-fraction of
the municipal waste stream in recent years. In Canada, bi-
odegradable material such as food waste (FW) represents
nearly 40% of the residential waste stream; therefore diver-
sion of organic materials is crucial to attain high diversion
targets. Municipal wastes (table, activated sludge, etc.) are
rich in protein, fat and ber materials, which can be effec-
tively treated using anaerobic digestion (AD) biotechnolo-
gy. However, the high levels of non-ber carbohydrate and
fat contents present in FW could lead to fast acidication.
Furthermore, accumulation of ammonia is attributed to its
high content of proteins (Braguglia et al., 2018). Co-diges-
tion of FW organics with other organic fractions in AD can
enhance a better nutritional supply and lessen the inhibit-
ing elements, such as ammonia and fat/lipid (Khairuddin
al., 2015).
Currently, municipalities must pay to transport and
dispose of these by-products in landlls/composting. An
interesting option for municipalities would be to pay local
farmers to receive and process these materials in AD bi-
oreactors. For farmers, the co-digestion of cow manure
(CM)+FW could increase the recovery of green energy,
production of litter for the herd and organic nitrogen fer-
tilizer for crops. Nevertheless, handling litter poses a sig-
nicant cost on dairy operations. Largely, animal waste
is considered an appropriate co-substrate due to its high
alkalinity, low C/N ratio and diverse macro- and micronu-
trients required by the anaerobic consortium. From previ-
ous studies, it is worth to note that, the mixtures of FW and
animal waste are typically composed by low percentages
67
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-74
of FW (Zhang et al. 2012; Agyeman and Tao, 2014) and
thus, further research is necessary to process high amount
of FW with minimal biological inhibition. Most of the AD
treatment of FW is carried out predominantly at mesophilic
temperatures (35-37°C) and few installations have been
reported on the thermophilic AD (50-55°C) operation. In
fact, thermophilic temperatures results in larger grade of
digester imbalance and higher risk for ammonia inhibition
than mesophilic conditions (Yang et al., 2015). Considering
the cold weather conditions, mesophilic/thermophilic pro-
cesses are constrained by the amount of energy needed to
heat the AD systems to maintain the desired temperature.
In addition, these processes are inhibited by free ammonia
toxicity while treating N-rich wastes. Lowering the tempera-
ture to 20°C could assure good methane yields and stabili-
ty by co-digesting FW+CM (Rajagopal et al., 2017); however
limited information have been reported on the successful
AD operation at temperatures below 35°C.
After the extensive research work done by Agriculture
and Agri-Food Canada (AAFC) and Bio-Terre Systems
(BTS) on the development of low-temperature AD system
for treating high solid content wastes like solid separated
animal manure and carcass (Massé et al., 2014, Rajagopal
et al., 2014; Saady and Massé, 2015), focus is now put on
the capacity of this new technological approach to play a
capital role in the organic waste management challenges
that several smaller municipalities are facing. Previous
studies were performed using laboratory-scale digesters
(30-120 L) to test different solid content manures with or
without liquid inoculum percolation-recirculation mode of
operation. It has been established that high solid AD can be
successfully operates with manure up to 35% TS content
(Saady and Massé, 2015). Thus, this paper aims to demon-
strate the operational feasibility of high-solid anaerobic
digestion (HSAD) system treating CM+FW at low-to-mod-
erate temperature conditions and to encourage small-scale
municipalities or farmers to adopt this technology at low
cost. In addition digestion of FW as a sole feeding source
with recirculation of liquid inoculum were performed. The
special emphasis was given to evaluate the biodegradation
of the organic waste and the optimal operation conditions
(such as OLR, cycle length) based on the organic matter
reduction and methane production. To obtain this, different
approaches were used such that lab-scale operations (30
L active volume) were performed in parallel and compared
with a scaled-up HSAD process (3 m3 active volume, that
is to say 100 times bigger than the lab-scale digesters) to
determine its feasibility of digesting high solid content fruit
and vegetable wastes (FVW) with or without solid dairy
manure. Liquid inoculum was used to start the pilot-scale
operation and the biomass adaptation procedures (liquid
to solid inoculum) were experimented.
2. METHODS AND MATERIALS
2.1 Experimental set-up and operating strategy
Scale-up testing was conducted in a container-type pi-
lot digester (2.43 m wide x 2.43 m height x 6.1 m length)
developed by Bio-Terre Systems and was installed outdoor.
This system is equipped with insulation, heating system,
gas collection and liquid inoculum percolation-recircula-
tion provisions. Provision was made to collect the liquid
percolate using a 50.8 mm liquid collection valve located at
the bottom of the container. The mixture of organic waste
and solid inoculum were lled in four numbers of plastic
bin-containers, each with a total capacity of 1 m3, primarily
to ease the waste handling procedures compared to bulk
loading of waste materials into the container itself. This pi-
lot-scale container type digester can accommodate a total
of 8 plastic bin-containers (i.e. total capacity of 8 m3). How-
ever, during this start-up phase of the study, four bin-con-
tainers were evenly lled with the mixture of solid inoculum
and organic waste before being put in the digester contain-
er, such that a total volume of about 3 m3 of waste mixture
were fed per cycle (33-14 d). The facility was also equipped
with a weighing scale to measure the mass of all materials
fed into the digester. Solid inoculum, organic waste, struc-
tural agent and the nal mixture were weighted for mass
balance purposes. The plastic bins handling were done
with a tractor, while the inoculum and organic wastes were
mixed using a S70 Bobcat. The mixed material was trans-
ferred to the respective bin-containers, weighted and load-
ed into the pilot-scale digester. Bioreactor was then sealed
during the entire treatment cycle length. Biogas production,
temperature and pressure were monitored through on-line.
Sampling of the material was done at the beginning and
the end of each cycle of operation.
At the start of each cycle, liquid inoculum was added
to enhance the microbial activity. The liquid leachate was
collected from the container with a 101.6 mm valve and
was transferred into a storage bin (1 m³) by gravity. Once
or twice a week, the leachate was recirculated back to the
feed mixture to maintain good humidity level and to im-
prove the waste-biomass contact. Throughout the treat-
ment cycle, temperature in the container bioreactor was
monitored daily using thermocouples installed in the con-
tainer and also in the organic waste mixture. Hot summer
uctuated the temperatures in the head space up to 28-
29°C and the heating system was adjusted to maintain the
temperatures between 20-25°C accordingly.
In parallel, a portion of the feed mixture was collected
separately and used to ll lab-scale digesters for closer
monitoring purposes. 50-L HSAD digesters (30-L active
volume) were operated in parallel at 25°C and gas produc-
tion was monitored using mass ow meters. The operating
protocol was maintained similar to that of pilot-scale ex-
periments.
2.2 Inoculum and feedstock sources
The liquid inoculum was obtained from an on-going
semi-industrial scale bioreactor treating diluted liquid CM
and FW mixture at our research facility. 500-L of the liquid
inoculum was taken to develop the solid inoculum, which
was then mixed with 200 kg of straw bedding and 44 kg
of raw solid dairy manure (without bedding, TS:19-20%).
Fresh dairy CM was collected at the experimental farm of
the Sherbrooke Research and Development Center. Feed
mixture was evenly distributed into 4 bin-containers (3-m3
active volume) and was placed in the container type biore-
actor for a period of 56 d primarily to allow a complete ad-
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-7468
aptation of the inoculum. A second cycle was started with
the addition of solid dairy manure and straw. Afterwards,
for another 106 d of reaction, co-digestion was started by
the addition of a small portion of FVW. Followed to this ad-
aptation phase, 100% of the organic loading was provided
by the FVW mixture.
Raw FVW was collected from local providers, which
was then weighted and grinded with a rototiller mounted on
a brush cutter. It mainly comprised potato and carrots peal-
ing, salad, potato, apple, banana, pineapple, orange, broc-
coli, onion, carrots and other rotten food materials. General
description and specic weight of the waste was taken
at every waste collection point. Overall, the proportion of
fruits was slightly higher than the vegetables (51%/49%).
The material was stored at 4°C before utilisation.
2.3 Sampling and analysis
For a feeding operation, a batch of about 250 kg of
waste was grinded together and evenly distributed into 150-
L barrels to have a more homogenous feedstock (for e.g.
6 barrels received 40 kg each of the same grinded batch).
Sampling was done accordingly by taking 10% of the dis-
tributed waste into the sampling box (for e.g. 40 kg distrib-
uted per barrel; 4 kg sample). Since majority of the mate-
rial was comprised of rotten food, the grinding was quite
easy and the particulate size was maintained smaller than
25.4 mm. For the inoculum and digested material, the sam-
pling was done in the mixing container (steel garbage bin).
Liquid leachate samples (100-150 mL) were taken on
a weekly basis from the liquid inoculum reservoir; and
the HSAD reactors’ samples were also collected at the
beginning and at the end of each treatment cycle. These
samples were analysed for total solids (TS), volatile solids
(VS), volatile fatty acids (VFA), chemical oxygen demand
(COD), pH and total Kjeldahl nitrogen (TKN). COD was de-
termined by the closed reux colorimetric method (APHA,
1992). The concentration of VFAs were measured using a
Perkin Elmer gas chromatograph model 8310 (Perkin Elm-
er, Waltham, MA), mounted with a DB-FFAP high resolution
column. TS and VS were determined using standard meth-
ods (APHA, 1992). pH value was measured using PH meter
(model, TIM840, France). TKN was analyzed using a Kjeltec
auto-analyzer (TECATOR 1030, Tecator AB, Hoganas, Swe-
den) using the macro-Kjeldahl method (APHA, 1992). Dai-
ly biogas production was measured by GFM mass ow
meters (Aalborg, USA). Biogas composition (methane,
carbon dioxide, H2S and nitrogen) was determined with a
HachCarle 400 AGCgas chromatograph (Hach, Loveland,
CO). The column and thermal conductivity detector were
operated at 80°C.
3. RESULTS AND DISCUSSIONS
3.1 Start-up of pilot-scale HSAD and adaptation of
solid biomass (Phase 1)
The initial two cycles were performed predominantly
to adapt the liquid inoculum to the high solid content op-
eration. Biogas quantity and compositions were followed
to measure the biological activity. Figure 1 presents the
specic methane yield (SMY) obtained for the initial two
cycles of operation. At the end of the rst cycle (37 d), the
dairy manure was not entirely converted into biogas. For in-
stance, the SMY reached only 58% of the expected conver-
sion in comparison to laboratory scale operation (0.170 LCH4
gVS
-1). This indicates that residual organic material was still
present in the digester and hence, second cycle was given a
longer reaction period (i.e. about 127 d). At the end of sec-
ond cycle, SMY obtained was about 0.244 LCH4 gVS-1. But
more importantly, the cumulative SMY for both the cycles
was 0.166 LCH4 gVS
-1, which was the expected for the dairy
manure digestion. For both cycles, the biogas composition
was stabilized with a CH4 content of around 40%.
3.2 Performance of pilot-scale HSAD treating FVW
waste (Phase 2: Low-loading conditions)
From day 128 onwards, pilot HSAD system was fed
with FVW and the performance was monitored in terms of
organic matter destruction, VFA accumulation, biogas con-
FIGURE 1: Start-up of HSAD and solid inoculum adaptation: Evolution of SMY.
69
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-74
centration and its quality, and SMY. For this phase of study,
the ratio of solid inoculum to FW was maintained at 6:1 to
limit the possibility of shock loading conditions to the bac-
teria. Organic loading rate (OLR) was maintained around
0.44-0.49 kgVS kginoculum
-1 d-1 and the results in terms of bio-
gas and cumulative methane production, and SMY are pre-
sented in Figure 2 (a-b). For both cycles, biogas production
preceded fairly quick start-up with no lag phase after each
feeding. It is to be noted that, about 77% of total biogas
production was attained within 18 d and 12 d for cycle 3
(cycle length: 34 d) and 4 (cycle length: 28 d), respectively.
High SMY values recorded for cycle 3 (i.e. 1.104 LCH4 gVS
-1) in
comparison to cycle 4 (0.625 LCH4 gVS
-1) were probably due
to the digestion of residual VS accumulated from previous
cycles of operation. Thus, longer reaction period was given
to cycle 3 to allow a complete digestion of the remaining
VS in the bioreactor. In addition, the biogas quality meas-
ured, especially after the feeding regime presented incon-
sistency, because of the operation procedure. As the feed-
ing was done in a batch mode, the bioreactor was opened
at the end of each cycle in order to be loaded with a new
material. As a result, the digester’s headspace was lled
with ambient air, which diluted the biogas for the initial few
days of a cycle. It took ve days for both cycles to ramp up
the biogas quality to 55% of CH4 in the measured biogas.
At the end of treatment cycles 3 and 4, the methane con-
centration was 64% and 67%, respectively. The biogas with
higher methane concentration indicates the good adapta-
tion of biomass to the high-solid content process.
Organic matter mass balance was performed using TS
and VS analysis. Inoculum, feedstock and digested mate-
rial samples were taken and analysed to perform a mass
balance approach. Table 1 illustrates the values from low
loading operating conditions. The operation of the HSAD
technology led to a great conversion of organic material
into biogas based on the mass balance calculations. From
the total 83 kg of VS fed, only 11 kg was accumulated in
the inoculum at the end of the last cycle. It represented a
reduction of 88%, which was the expected level of degra-
dation for that nature of waste (>85%). The mass balance
was not otherwise conclusive for the TS accumulations.
Based on the values presented in Table 1, an accumula-
FIGURE 2 (a,b): Performance of Phase-2 (low OLR) HSAD process: a) Biogas production and SMY evolution; b) Cumulative methane pro-
duction.
(a) Biogas production - Phase 2
Daily biogas production
(L/d)
(b) Cumulative methane production - Phase 2
Cumulative methane production
(L CH4)
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-7470
tion of 65 kg was present at the end of the treatment cycle.
However, only 10 kg was provided from the VS accumu-
lation and about 22 kg was resulted from the inorganic
fraction of the solids. It is not apparent that all the residual
32 kg was contributed by the liquid inoculum (at 1.5 to 2%
TS). Consequently, it was observed that rocks and sands
accumulated in the inoculum. Those inert materials were
probably introduced into the system during the bobcat op-
eration and were not from the feeding susbtrates. It can be
explained by some portion of the solid accumulation val-
ues measured in the mass balance.
The total organic mass increased to about 27% of the
total feed weight. The majority of the weight loss was not
due to the VS degradation but as a result of water content
released throughout the digestion process. With about 90%
water content, the feedstock ended-up with large quantity
of water, which was then released while digesting the or-
ganic material. From the mass balance, 765 kg of water
was included in the feedstock, in which, 548 kg was taken
out of the inoculum in 3 cycles (liquid inoculum production
and water vapor in the biogas). About 164 kg of water was
accumulated in the solid inoculum increase. According to
the data, about 53 kg of water (6.9% of all water) was not
accounted from the balance. The three digestion cycles
showed that high solid digestion of FW was effective in
converting organic material into good quantity and quality
of biogas. The mass balance indicates that for each kg of
FW fed to the solid inoculum, about 0.27 kg was accumu-
lated in the inoculum and about 0.65 kg of liquid inoculum
was released.
3.3 Performance of pilot-scale HSAD treating FVW
waste (Phase 3: High-loading conditions)
The purpose of this phase of study was to increase the
OLR by decreasing the ratio of inoculum to feedstock, such
that same size bioreactor can process more waste materi-
als with short retention times. From day 232 onwards, the
proportion of solid inoculum to dairy manure was retained
at 3:1. The similar operating strategy was followed for this
phase of study as that of Phase 2. The same testing ap-
proach was used to follow this series of tests by using con-
tainer bioreactor for the mass balance and the laboratory
scale bioreactor for the biogas measurements. OLR was
maintained around 1.6 to 2.1 kgVS kginoculum-1 d-1 and cycle
length was controlled at 14-16 d. Results in terms of biogas
and cumulative methane production, and SMY are present-
ed in Figure 3 (a-b). Although, 5th and 6th cycles were oper-
ated at high OLR conditions, AD preceded fairly quick start-
up with no lag phase after feeding with FVW. More than
75% of the total biogas production was attained within 7-d
for both cycles. SMY values in the range of 0.400-0.520 LCH4
gVS
-1 were recorded. The obtained values are comparable to
the AD of semi-dry mixed municipal FW (SMY: 0.401±0.01
LCH4 gVS-1) [Rajagopal et al., 2017]
The short retention times and high OLR conditions
seemed to have a little impact on the inoculum for the
5th cycle of operation (SMY: 0.400 L LCH4 gVS-1) but not
enough to imbalance the process. However, for the sub-
sequent cycles (Figure 3, a-b), the SMY increased by 30%,
which indicates the good activity of the biomass. This was
conrmed by the less VFA accumulations (total content be-
low 900 mg L-1) and high buffering capacity of the digester
(digester pH: 7.2-7.5). pH of the substrate was acidic (4.0-
4.5), but however, in the reactor it was in the neutral range.
There was no sign of inhibition or nutrient deciency at
these operating conditions.
The methane concentration during the rst 4 days of
operation was around 58% (for both cycles). At the end of
the treatment cycle, the same was increased to 66% and
63%, respectively. Better performance were obtained even
at higher loading conditions and short treatment cycles
due to the good adaptation of the biomass. These results
were comparable to that of laboratory scale study, particu-
larly in terms of SMY (i.e. 0.4-0.5 m3 CH4 kgVS
-1), and meth-
ane content of about 62-72%. The quality of biogas in the
pilot-scale digester increased with time and remained al-
most stable thereafter. This species that the smaller size
digester (in the case of HSAD) is preferred as there is no
waste dilution involved. This could reduce a major part in
the capital investment on the construction of digesters and
also suitable for cold countries. Further experiments were
continued to optimise the inoculum to feed ratio, high OLRs
and short treatment cycle lengths.
Since the container type bioreactor was not providing
accurate biogas production, operation follow up was done
using mass balance approach. All the inoculum, feedstock
and digested materials were weighted in the container be-
fore and after the digestion process. Samples were taken
and analysed to perform a mass balance analysis. Values
from all high loading cycles are presented in Table 2. Mass
balance based on organic matter values was performed
using TS and VS analysis. Inoculum, feedstock and digest-
ed material samples were taken and analysed to perform a
mass balance approach. Table 2 presents the values from
high-loading operating conditions. The increase of the
loading rate modied the mass balance obtained from the
last two cycles. While organic fractions were converted up
to 87% in the rst three cycles, the high loading operation
TABLE 1: Mass balance values for the low loading conditions (Inoculum to feed ratio of 6:1).
Cycle # Inoculum
start weight
Feedstock
addition TS fed VS fed Inoculum
end weight
Inoculum
TS in
Inoculum
TS out
Inoculum
VS in
Inoculum
VS out
32096 kg 389 kg 60 kg 45 kg 2190 kg 306 kg 362 kg 236 kg 257 kg
41673 kg 275 kg 26 kg 22 kg 1673 kg 276 kg 285 kg 196 kg 188 kg
51171 kg* 186 kg 18 kg 16 kg 1225 kg 183 kg 184 kg 123 kg 120 kg
Total 4940 kg 500 kg 104 kg 83 kg 5088 kg 765 kg 831 kg 555 kg 565 kg
* Cycle 5 was done with 2/3 of the available inoculum. 1/3 remaining was used for the high loading testing
71
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-74
cycles obtained only 44% reduction. This low conversion
obtained was contradictory with the SMY measured using
the biogas production, which indicates a proper organic
fraction digestion. In that case, SMY would be considered
more precisely since the biogas measurement was done
on a continuous basis with instruments, while the mass
balance was performed based on one-time sample taken
on a large pile of heterogeneous material. Even though all
precautions were taken to assure the proper sampling, it
may be possible that the sample contains a piece or non-di-
gested FW that contaminated the sample. Once again, the
mass balance was not conclusive for the TS accumulation.
Based on the values presented, the treatment cycle led to
an accumulation of 44 kg. This value was not realistic as
only 24 kg was provided from the VS accumulation and 5
kg was from the inorganic fraction of solids. It is not evi-
dent that that all the remaining 15 kg was contributed by
the liquid inoculum (at 1.5 to 2% TS). Similar to the previ-
ous cycle of operation, the presence of inert material could
explain the portion of the solid accumulation measured in
the mass balance.
The total organic mass increase was about 88% of
the total feed weight. The majority of the weight loss was
not due to the VS degradation but from the water content
released throughout the digestion process. Similar to the
previous cycles, with a ~90% water content, the feedstock
brings great quantity of water, which was released during
the digestion process. From the mass balance, 503 kg of
water was included in the feedstock. From that, only ~49
kg was taken out of the inoculum in 2 cycles (liquid inoc-
ulum production and water vapor in the biogas). About
447 kg of water was accumulated in the solid inoculum.
According to the data, another 15 kg of water (3.0% of all
water) was missing from the balance. The water accumula-
tion in the solid inoculum was challenging and this showed
an indicative of a lack of structure in the solid.
FIGURE 3: Performance of Phase-3 (high OLR) HSAD process: a) Biogas production and SMY evolution; b) Cumulative methane produc-
tion.
(a) Biogas production - Phase 3
(b) Cumulative methane production - Phase 3
Daily biogas production
(L/d)
Cumulative methane production
(L CH4)
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-7472
Due to this reason, the further experiments were sup-
plemented with more structural material and the optimisa-
tion study was carried out with this modications. Struc-
tural agent (100 kg of dry straw) was used to increase the
solid content and the draining capacity of the solid inocu-
lum. The addition of straw increased considerably the vol-
ume of the solid inoculum and hence a small portion of
initial solid inoculum was used. Wasted food residues was
then fed to the newly mixed inoculum and it is to be noted
that the straw weight was not considered in the feeding
ratio calculation. The biogas production obtained was very
different from the previous cycles. Starting at day 266, this
cycle produced a small initial biogas production peak and
maintained its production level for a longer period of time.
It took about additional 2 days to reach the same conver-
sion factor as the previous cycles. Reaction period was in-
creased to 21 days to allow better adaptation to the straw
addition.
On the contrary, cycle 8 (day 287) did not perform well
as expected. However, the liquid inoculum percolation
was good, henceforth no additionnal structural agent was
added thereafter to the solid inoculum. It took 25 days to
reach the expected conversion factor of 0.5 LCH4/g VS fed.
This delay in the biological conversion of organic solids to
methane indicates the process imbalance. The hypothesis
posed was that the C/N balance was too high with the ad-
dition of straw and nitrogen was lacking in the process. In
order to help the biological process to resume its activity
level, dairy manure was added to the feed material on cy-
cle 9 (day 323). Dairy manure was used to rebalance the
C/N ratio and the other required minerals for the digestion
process. The total feeding ratio was kept around 3:1 (see
Table 2) but 28.5% of the feeding weight was resulted from
dairy manure (rest was FW). The addition of dairy manure
and FW was treated in about 17 days, which was an im-
provement in comparison to 25 days of operation obtained
previously.
The conversion factor reached was lower due to the
lower biogas conversion potential of the dairy manure.
With the fed proportion, the expected conversion factor to
be reached was 0.33 L CH4/g VS. The 10th cycle (day 337)
was done using FVW alone and the operation went relative-
ly well by obtaining a biogas conversion rate of 0.487 LCH4/
gVS within 14 days. A signicant biogas production peak
was obtained for the rst 4 days and then stabilised. Once
again, the addition of low solid FVW dragged the solid in-
oculum to a low solid content (~15% TS). Structural agent
may be needed for the next treatment cycle (Table 3).
Similar to the lower loading conditions, the biogas qual-
ity measured throughout the digestion process was highly
variable mainly because of the operational procedures. De-
pending on a specic cycle, it took between 4 to 10 days to
reach 55% of CH4 in the measured biogas. If we consider
that around 15-L of headspace was present in the labora-
tory scale bioreactor, we can use the dilution formula to
estimate the biogas quality needed to bring the measured
biogas composition at 55%:
C
1
V
1
+ C
2
V
2
= C
3
V
3
(1)
Where:
C1: Methane concentration of the headspace
V1: Volume of the headspace
C2: Methane concentration of the produced biogas
V2: Volume of biogas produced
C3: Methane concentration measured
V3: Volume of biogas and headspace
The calculated methane concentration is presented in
Table 4. At the end of the treatment cycles, the methane
concentration measured was between 54% and 63%. Sol-
uble COD reduction of about 80-90% was obtained during
this operation and TKN concentration of about 5-8 g N/L
did not hinder the AD process. Except for the cycle 7, re-
sults indicate that adaptation of the biomass to the high
loading conditions can be achieved but it may require
longer reaction period. The impact of structural agent addi-
tion is yet to be fully understood.
The values obtained from the mass balance indicate
that nearly 45% of TS and VS were degraded during the
treatment cycles. A large uctuation in the solids reduc-
tions were measured throughout the cycles. This high var-
iation was probably due to the high heterogeneity level of
the material making it hard to mix and sample represent-
atively. Larger pieces of FW (e.g.: whole orange, potato,
TABLE 2: Mass balance values for the low loading conditions (Inoculum to feed ratio of 3:1).
Cycle # Inoculum
start weight
Feedstock
addition TS fed VS fed Inoculum
end weight
Inoculum
TS in
Inoculum TS
out
Inoculum
VS in
Inoculum VS
out
5a816 kg 279 kg 26 kg 24 kg 1 017 kg 138 kg 154 kg 89 kg 99 kg
6b1 205 kg 280 kg 21 kg 19 kg 1495 kg 181 kg 209 kg 118 kg 132 kg
7c559 kg 167 kg 15 kg 14 kg 588 kg 123 kg 114 kg 92 kg 84 kg
8 568 kg 232 kg 20 kg 18 kg 559 kg 111 kg 109 kg 82 kg 75 kg
9d1039 kg 368 kg 37 kg 32 kg 1196 kg 188 kg 208 kg 136 kg 158 kg
10 1071 kg 336 kg 21 kg 18 kg 1381 kg 187 kg 220 kg 141 kg 163 kg
Total 5 258 kg 1 662 kg 140 kg 125 kg 6 236 kg 876 kg 1 014 kg 610 kg 711 kg
Average 876 kg 277 kg 23 kg 21 kg 1039 kg 146 kg 169 kg 102 kg 119 kg
a: Cycle 5 was done with 1/3 of the available inoculum. 2/3 remaining was used for the low loading testing.
b: Cycle 6 was done with 2/3 of the available inoculum. 1/3 remaining was used for the 1:1 loading testing.
c: Cycle 7 had 62 kg of straw added (not included in the solid inoculum) adding 52 kg of TS, 48 kg of VS to the initial inoculum.
d: Cycle 9 feeding was done with 28.5% w/w of dairy manure and the rest was food waste.
73
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-74
avocado, corn, etc.) were harder to digest and some resid-
ual FW pieces were identied after the digestion process.
TS and VS characterisation of the inoculum was then be
greatly affected by the presence of those undigested ma-
terials.
A 45% VS reduction was considered a low value com-
pared to the SMY measured with the biogas production.
In the literature, FW degradation can be as high as 85%
during anaerobic process leading to a SMY of 0.5 to 0.6
LCH4 gVS
-1. In this experience, SMY was considered more
precisely since the biogas measurement was done on a
continuous basis with precision instruments. Whereas
the mass balance was done based on one sample taken
in a large pile of heterogeneous material. Even though all
precautions were taken to assure the proper sampling, the
samples still contained small portion or pieces of non-di-
gested FW.
During the high loading cycles, the total mass of the
inoculum increased by 18.6% (TS accumulation of 138
kg) partially due to the structural agent addition and water
retention. The average TS content of the solid inoculum
was maintained above 15% by the addition of straw. FW
material is generally a low solid content material that may
not be a good source of feedstock for high solid content
digestion if used as a sole source. Structural agent is nec-
essary to maintain the high solid level in the bioreactor.
During the test cycles, a total amount of 1 522 kg of wa-
ter was fed with the FW to the digester. The majority of
the feedstock water (840 kg) stayed in the solid inoculum
(likely absorbed by the straw) while another 620 kg was
released from the digester (liquid inoculum and water va-
por in the biogas). The liquid inoculum can be used for
further treatment cycle but excess will have to be used.
Further analyses are needed to establish the nutrient con-
tent of the liquid inoculum, which is required to establish
its fertilizer potential.
3.4 HSAD technology: Final discussion and conclud-
ing remarks
The biogas production, especially in terms of SMY and
CH4 concentration in biogas, yielded from the low-to-mod-
erate temperature HSAD in this study were found to be
similar or marginally higher than that of mesophilic/ ther-
mophilic AD. For instance, Zhang et al. (2012) obtained a
greater digester stability while codigesting FW with cat-
tle slurry, in 1:4 ratio, at OLR of 2 kgVS m-3 d-1 in meso-
philic conditions (36°C). At similar OLR and temperature,
Agyeman and Tao (2014) determined the effects of FW
particle size on co-digestion with dairy manure, in 1:1 ra-
tio, in which they obtained a SMY up to 0.47 LCH4 g-1VS-
fed (for coarse-grinded FW). Alternatively at thermophil-
ic conditions (55°C), Castrillón et al. (2013) used lower
proportions of FW (i.e. 10% FW and 90% CM) to obtain
the stability, which corresponded to a SMY of 0.3 LCH4
g-1VSfed . In the present study, lowering the temperature
to 20-25°C assured a good SMY and operational stabili-
ty even for the mono-digestion of FVW. The reason could
probably be due to that fact that at lower temperature, re-
duced hydrolysis of complex organics have declined the
acidogenesis and thus reduced the proportion of CO2 in
biogas and additional production of acetate from CO2 and
H2 by homoacetogens and the decrease of the resulting
acetate would upsurge the amount of CH4 in biogas (Wei
et al., 2014). It is to be noted that low temperature AD is
particularly well adapted to the treatment of several or-
ganic wastes because of lower free ammonia nitrogen
Cycle # Inoculum: food ratio OLR (g VS/Kginoculum/day) Specic methane yield (SMY)* (L CH4/g VS)
5 2.9 : 1 2.07 0.396
6 2.7 : 1 1.57 0.520
7 3.3 : 1 1.06 0.584
8 2.45 : 1 0.91 0.612
9** 2.82 : 1 1.87 0.321**
10 3.2 : 1 1.21 0.487
* The gas production cannot be measured accurately with the container type bioreactor. SMY was obtained from lab-scale bioreactor lled with the same
material as the container.
** For this cycle, dairy manure was added with the FVW
TABLE 3: Values obtained for the high loading treatment cycle (3:1).
TABLE 4: Calculated biogas composition during the rst portion treatment cycle (Inoculum to feed ratio 3:1).
Cycle # Days to reach 55% CH4
in biogas
Volume of biogas
measured before 55%
Calculated % CH4
in rst biogas
% CH4 at the end of
treatment cycle
5 4 292.1 L 57.8% 60%
6 4 291.6 L 57.8% 63%
7 10 236.8 L 58.5% 54%
8 6 135.5 L 61.1% 63%
9 6 183.2 L 59.5% 62%
10 7 241.1 L 58.4% 62%
Average 6.2 230.05 L 58.9% 60.7%
R. Rajagopal et al. / DETRITUS / Volume 05 - 2019 / pages 66-7474
concentrations than in mesophilic/thermophilic AD pro-
cesses. In addition to this, the present study also validat-
ed that solid CM can be used as a co-substrate provided
a sufcient buffering capacity to FVW digestion by syner-
gizing the effect of microorganisms and handling the high
OLR. Nevertheless, further optimization will be essential
to validate and improve the performance of HSAD at rela-
tively short cycle length and high OLR.
In order to implement this HSAD technology for larger
scale operations, economic cost analysis is essential for
the nancial success. The amount of methane produced
in this process can directly inuence the economic ben-
ets, as this can substitute other fuels used for cook-
ing, heat, light, or electricity. For example, it is critical to
have methane concentration greater than 50% and H2S
concentration less than 1% for running a generator fue-
led with biogas (Lansing et al., 2008). In this study, HSAD
technology met the minimum conditions to power a
generator and coupled with the quantity of methane pro-
duced. Methane content of 65% in biogas has an overall
energy potential of approximately 23 MJ m-3. According
to U.S EPA (2016), over 38.4 million tons of US food waste
generation was reported in 2014. Since food makes up
over 20% of municipal solid waste combusted with ener-
gy recovery or landlled in US, 12 commercial food dis-
posal bans are often seen as a precarious step towards
the long-term waste reduction goals (U.S. EPA, 2016).
Keeping in mind that the caloric value of biogas is in the
range of 6 kWh m-3 (which is equivalent to 0.5 L of die-
sel oil) (Kashyap et al., 2003), the proposed low-to-mod-
erate temperature HSAD process could lead to saving of
an enormous amount of fuel per year. Furthermore, by
trapping methane from municipal or agricultural wastes
for heat or electricity also diminishes direct atmospheric
methane emissions and with it, greenhouse gas impact.
For instance, diverting one ton of FW through HSAD re-
duces greenhouse gas emissions by nearly one ton of CO2
equivalents, as compared to landlling (Environment Can-
ada, 2013). Similarly, Canadian livestock produces about
180 MT of manure every year. Treating 20% of manure
would lead to a prot of $55M in terms of carbon credit
and bioenergy generated can replace electricity/natural
gas to a value close to $100M per year. Thus, the asset
of this low-to-moderate temperature HSAD technology is
that it provides an integrated solution to municipal and
agricultural waste streams without any dilutions.
4. CONCLUSIONS
This study validated the robustness of low-to-moderate
temperature HSAD technology, which can be employed to
treat high-solid content wastes such as dairy manure and
FW. SMY of 0.4-0.6 LCH4 gVS-1 was obtained even at high OLR
(1.6 to 2.1 kgVS kginoculum
-1 d-1) and short cycle length (14-
16 d), which is comparable to the laboratory scale study.
Lowering the temperature to 20-25°C even favoured the
mono-digestion of FVW. The mode of operation (process,
temperature) along with the acclimation of liquid biomass
to solid inoculum at step-wise increase in OLR ensured a
high stabilisation of the digestion process without inhibi-
tion. Compared to higher-temperature digestion process,
more energy is available for farm uses and thus farmers
could adopt this technology at affordable cost. Further re-
search is being performed to determine the optimal oper-
ating conditions.
ACKNOWLEDGEMENT
Authors thank Agri Innovation Program (Project No.
103) for providing nancial support.
REFERENCES
Agyeman, F.O., Tao, W., 2014. Anaerobic co-digestion of food waste
and dairy manure: effects of food waste particle size and organic
loading rate. J. Environ. Manage. 133, 268–274.
APHA., 1992. Standard methods for the examination of water and
waste water. 18th ed. Washington DC, USA: American Public
Health Association.
Braguglia, C.M., Gallipoli, A., Gianico, A., Pagliaccia, P., 2018. Anaerobic
bioconversion of food waste into energy: a critical review. Biore-
sour. Technol. 248, 37–56.
Castrillón, L., Marañón, E., Fernández-Nava, Y., Ormaechea, P., Quiro-
ga, G., 2013. Thermophilic co-digestion of cattle manure and food
waste supplemented with crude glycerin in induced bed reactor
(IBR). Bioresour. Technol. 136, 73–77.
Environment Canada., 2013. Technical Document on Municipal Solid
Waste Organics Processing. ISBN:978-1-100-21707-9. (Viewed on-
line on September 2017).
http://www.compost.org/English/PDF/Technical_Document_MSW_
Organics_Processing_2013.pdf
Kashyap, D.R., Dadhich, K.S., Sharma, S.K., 2003. Biomethanation under
psychrophilic conditions: a review. Bioresour. Technol. 87, 147–53.
Khairuddin, N., Manaf, L.A., Halimoon, N., Ghani, W.A.W.A.K.,Hassan,
M.A., 2015. High solid anaerobic co-digestion ofhousehold organic
waste with cow manure. Procedia Environ.Sci. 30, 174–179.
Lansing, S., Botero, R.B., Martin, J.F., 2008. Waste treatment and biogas
quality in small-scale agricultural digesters. Bioresour. Technol. 99
(13), 5881-5890.
Massé, D.I., Saady, N.M.C., Rajagopal, R., 2014. Psychrophilic dry an-
aerobic digestion of high solids content dairy manure: long-term
operation. Biological Engineering Transactions 7(3): 99-112.
Rajagopal, R., Massé, D.I., Saady, N.M.C., 2014. Low-temperature an-
aerobic co-digestion of swine carcass and swine manure: Impact
of high swine carcass loading rate. Transactions of the ASABE 57
(6), 1811-1816.
Rajagopal R, Bellavance D, Rahaman S., 2017. Psychrophilic anaero-
bic digestion of semi-dry mixed municipal food waste: For North
American Context. Process Safety and Environmental Protection
105, 101-108.
Saady, N.M.C., Massé, D.I., 2015. A start-up of psychrophilic anaerobic
sequence batch reactor digesting a 35% total solids feed of dairy
manure and wheat straw. AMB Expr 5:55.
U.S. EPA (Environmental Protection Agency)., 2016. Food Waste
Management in the United States, 2014.https://www.epa.gov/
sites/production/les/2016-12/documents/food_waste_manage-
ment_2014_12082016_508.pdf (accessed 10/11/2017)
Wei, S., Zhang, H., Cai, X., Xu, J., Fang, J., Liu, H., 2014. Psychrophilic
anaerobic co-digestion of highland barley straw with two animal
manures at high altitude for enhancing biogas production. Energ.
Convers. Manage. 88, 40-48.
Yang, L., Huang, Y., Zhao, M., 2015. Enhancing biogas generation per-
formance from food wastes by high-solids thermophilic anaero-
bic digestion: effect of pH adjustment. Int. Biodeterior. Biodegrad.
105, 153–159.
Zhang, Y., Banks, C.J., Heaven, S., 2012. Anaerobic digestion of two
biodegradable municipal waste streams. J. Environ. Manage. 104,
166-174.
Zheng, Y.H., Wei, J.G., Li, J., Feng, S.F., Li, Z.F., Jiang, G.M., Lucas, M.,
Wu, G.L., Ning, T.Y., 2012. Anaerobic fermentation technology in-
creases biomass energy use efciency in crop residue utilization
and biogas production. Renew. Sust Energy Rev 16: 4588–4596.
