Sustainability 2019, 11, 3875; doi:10.3390/su11143875 www.mdpi.com/journal/sustainability
Anaerobic Digestion of Food Waste with
Unconventional Co-Substrates for Stable Biogas
Production at High Organic Loading Rates
and Thomas A. Trabold
The Water Center, University of Pennsylvania, Philadelphia, PA 19147, USA
Golisano Institute for Sustainability, Rochester Institute of Technology, Rochester, NY 14623, USA
* Correspondence: email@example.com; Tel.: +1-585-475-4696
Received: 23 June 2019; Accepted: 9 July 2019; Published: 16 July 2019
Abstract: Anaerobic digestion (AD) is widely considered a more sustainable food waste
management method than conventional technologies, such as landfilling and incineration. To
improve economic performance while maintaining AD system stability at commercial scale, food
waste is often co-digested with animal manure, but there is increasing interest in food waste-only
digestion. We investigated the stability of anaerobic digestion with mixed cafeteria food waste
(CFW) as the main substrate, combined in a semi-continuous mode with acid whey, waste bread,
waste energy drinks, and soiled paper napkins as co-substrates. During digestion of CFW without
any co-substrates, the maximum specific methane yield (SMY) was 363 mL gVS
loading rate (OLR) of 2.8 gVSL
, and reactor failure occurred at OLR of 3.5 gVSL
substrates of acid whey, waste energy drinks, and waste bread resulted in maximum SMY of 455,
453, and 479 mL gVS
, respectively, and it was possible to achieve stable digestion at OLR as
high as 4.4 gVSL
. These results offer a potential approach to high organic loading rate digestion
of food waste without using animal manure. Process optimization for the use of unconventional co-
substrates may help enable deployment of anaerobic digesters for food waste management in urban
and institutional applications and enable increased diversion of food waste from landfills in heavily
Keywords: food waste; co-digestion; biogas; specific methane yield; organic loading rate; process
Residential, institutional, and industrial sectors generate large volumes of food waste
throughout the year. Though some fraction is productively utilized, most of the generated food waste
is landfilled. In the United States, a few regions have moved towards legislation that bans sending
food waste to landfills, and it is expected that this practice will decline as it is the most inefficient use
of the material and does not adequately utilize its embodied energy and water content. Anaerobic
digestion is currently the most efficient commercial-scale use of food waste, as the end products are
nutrient-rich digestate and biofuel that can be converted to electrical or thermal energy. Food waste
not only provides an inexpensive substrate for anaerobic digestion but also significantly improves
biogas production relative to systems that convert manure or sewage sludge alone . However, a
relatively small fraction of anaerobic digester plants worldwide co-digest food waste with manure,
sewage sludge, or lignocellulosic biomass. Most digesters in the US are farm-based and digest dairy
manure and agricultural residues. While food waste is co-digested in small amounts with manure,
there is a need for major developments in achieving food waste-only digestion to accommodate
Sustainability 2019, 11, 3875 2 of 15
increasing rates of landfill diversion. There are certain challenges associated with food waste-only
digestion because, without any co-substrates, process instability can result from insufficient trace
elements that regulate enzyme reactions. Therefore, the conventional practice has been to digest food
waste, at a relatively low fraction, with primary substrates of animal manure or sewage sludge. Use
of co-substrates in anaerobic digestion of food waste helps to maintain process stability, either by
balancing the carbon-to-nitrogen ratio or by providing trace minerals and buffering action. Table 1
provides a sample of the literature exploring various co-substrates used in food waste digestion. The
majority of the studies have used animal manure and sludge as the co-substrates, except for a few
studies, which involved rice straw , fats, organic fraction of the municipal solid waste , and fats,
oils, and grease (FOG; ). It is essential to explore newer co-substrates as food waste generation
continues to rise, and existing digesters that co-digest food waste with manure cannot handle all the
additional waste generated.
There are two classes of methane-forming bacteria: acetoclastic and hydrogenotrophic
methanogens . Food waste is rich in nitrogen, generally attributed to the presence of proteins.
Therefore, food waste digestion often leads to ammonia formation, which is toxic to acetoclastic
methanogens . While hydrogenotrophic methanogens continue to produce methane, food waste
lacks certain trace elements that lead to the accumulation of volatile fatty acids. Therefore, ammonia
formation and volatile acid accumulation are the major process challenges in food waste digestion
that can be addressed by ammonia stripping or trace element supplementation. There are several
efforts at laboratory and pilot scale for in situ removal of ammonia using side-stream gas stripping
, gas mixing , ultrasonication , microwave irradiation , and other physical and chemical
methods . Trace element and mineral supplementation has also been proposed by various
researchers and organizations as an alternative approach to achieving a stable process ([6,12–14]). A
recent review study on the stability issues in food waste digestion recommended a need for intensive
process monitoring and microbial management to address instability issues associated with food
waste digestion . In another review focusing on anaerobic digestion of food waste, none of the
research studies reviewed used a secondary food waste stream as a co-substrate in food scrap
Table 1. Literature studies of food waste digestion with various co-substrates (NR: not reported; NA:
Food Waste Type Co-
[gVSL−1d−1] Operation Mode
Semi-continuous with only FW
fed on day 1 and 2, CM fed on
day 3; the sequence repeated
NA 508 50 
CFW CM 1 15
Semi-continuous with once a
day feeding and discharge 2 317 NR 
kitchen waste Biowaste NR 10.9
Semi-continuous with two
times a day feeding, five days a
NR 420 7 
sludge 5 1 Batch 0.5 494 21 
CFW Rice straw 1 5 Batch 5 392 NR 
Vegetable waste none 75 1.4 Semi-continuous with once a
day feeding and discharge NA 250 25 
sludge 0.5 2
Semi-continuous with once a
day feeding and discharge 1 633 20 
and grease 5 4
Semi-continuous with once a
day feeding and discharge 6.7 318 16 
OLR: organic loading rate; FW: food waste; SMY: specific methane yield; HRT: hydraulic retention time.
Anaerobic digestion (AD) is a viable method for the conversion of food waste and other organic
materials into methane-rich biogas. However, when applied at high organic loading rates, using only
food waste as feedstock can lead to an unstable process. The methods recommended in the previous
studies mentioned above, such as ammonia stripping and trace element supplementation, have been
Sustainability 2019, 11, 3875 3 of 15
proven effective in achieving process stability, but they add significantly to process cost by requiring
additional infrastructure or chemicals. The present study evaluated non-manure based co-digestion
of mixed cafeteria food waste (CFW) without the addition of any trace nutrients or recovery methods.
The co-digestion of CFW with other food sector waste materials, including acid whey (AW), waste
energy drinks (ED), wasted bread (WB), and paper napkins (PN), was studied to evaluate stability
issues in food waste digestion. In addition, the digestion of CFW without using any co-substrate and
cow manure as a co-substrate were studied as controls. This research focused on understanding the
process stability issues with food waste digestion through analysis of various parameters and
improving the process using non-manure-based co-substrates. Supplementing CFW regularly with
other food sector wastes with consistent composition was thought to improve process stability during
CFW digestion. The safe range of operating parameters is also recommended based on the results
obtained in this study. Use of unconventional co-substrates offers a potential advantage in situations
where it is not practical to haul digestate long distances for field spreading, for example,
decentralized digesters at universities, hospitals, or food processing plants. Facilitating “food waste
only” systems is also important in diverting a higher volume of food waste from landfills.
2. Materials and Methods
The term “reactor” is used to denote the experimental setup used in this study, whereas the term
“digester” is used in a real-world sense to emphasize the practical importance of this research.”
Cafeteria food waste” refers to mixed pre- and post-consumer waste from the university dining halls
and “food waste” is a generic term that relates to both CFW and food processing waste. The term
“food waste-only digestion” is used to emphasize the importance of unconventional (non-manure)
substrates originating from the food sector to produce energy using anaerobic digestion. The co-
digestion mixes used in each reactor are listed in Table 2.
Table 2. Co-substrates used in each reactor, with cafeteria food waste (CFW) as the primary
Reactor Co-Substrate Abbreviation
R1 None -
R2 Acid whey AW
R3 Energy drink ED
R4 Waste bread WB
R5 Paper napkins PN
R6 Cow manure CM
2.1. Food Waste and Co-Substrate Characterization
The substrates utilized in this study were analyzed for physical and chemical characteristics.
Physical parameters measured in our laboratory included pH, total solids (TS), and volatile solids
(VS). A third-party laboratory analyzed the samples for all the other parameters listed in Tables 3,4,5.
The pH of the substrates was recorded using a Seven Compact pH meter (Mettler Toledo, Columbus,
OH, USA). The TS and VS were analyzed according to standard Environmental Protection Agency
(EPA) method 1684 . The chemical oxygen demand (COD) was examined for each substrate using
a standard COD analyzer (Hach DR 3900, Loveland, CO, USA; Method 8000).
Sustainability 2019, 11, 3875 4 of 15
Table 3. Characteristics of cafeteria food waste (CFW) estimated as an average of several samples
from different batches. Standard deviations provided for n >2.
Characteristics of CFW
Physical and Chemical Properties, n = 6
pH 4.2 ± 0.3
TS % 23.8 ± 2.9
VS/TS % 90.9 ± 2.4
VS % 22.9 ± 1.2
COD g/L 197 ± 42
Ash % 1.8 ± 0.8
Calorific value, kJ/kg (n = 2) 23,098
Macronutrients, n = 4
Crude protein (CP) % 13.3 ± 10
Available protein % 13 ± 9.9
Soluble protein % of CP 53 ± 4.2
Lignin % 0.7 ± 0.9
Starch % 10.5 ± 8.5
Simple Sugars % 7.5 ± 4
Crude fat % 13.4 ± 11
Minerals, n = 4
Calcium, ppm 1225 ± 1014
Potassium, ppm 5950 ± 4088
Magnesium, ppm 425 ± 263
Phosphorous, ppm 1900 ± 1449
Sodium, ppm 4013 ± 2984
Iron, ppm 23 ± 17
Zinc, ppm 12.8 ± 8.4
Copper, ppm 2.7 ± 1.5
Manganese, ppm 6.5 ± 4.7
Molybdenum, ppm ND
Sulfur, ppm 1525 ± 1187
Chlorine ion, ppm 7125 ± 5227
Elemental composition, n = 2
Carbon % 52.4
Hydrogen % 7.4
Nitrogen % 3.3
Oxygen % 30.6
ND: not detected. TS: total solids; VS: volatile solids; COD: chemical oxygen demand.
Sustainability 2019, 11, 3875 5 of 15
Table 4. Physical and chemical properties and elemental composition of co-substrates. Standard
deviations provided for n >2.
Physical and Chemical
Properties, n = 3 AW ED WB PN CM
pH 4.2 ± 0.2 3.3 ± 0 NM NM 6.8 ± 0.5
TS % 2.9 ± 0.1 0.7 ± 0.1 95.6 ± 0.7 94.5 ± 3.5 10.3 ± 1.4
VS/TS % 73.3 ± 0.2 80.2 ± 5.0 89.4 ± 2.9 86 ± 6.9 83.5 ± 0.8
VS % 2.1 ± 0.1 0.6 ± 0.1 85.4 ± 2.3 81.4 ± 9.4 8.6 ± 1.2
COD g/L 43 ± 4.0 11 ± 0.1 1167 ± 97 1176 ± 142 97 ± 6.0
Ash % 4.2 ± 0.2 0.2 ± 0 10.2 ± 2.8 13.1 ± 5.9 1.7 ± 0.3
Elemental composition (single measurements)
Carbon % 1.5 0.5 45.2 44.8 54.8
Hydrogen % 11.2 10.8 6.6 6.1 NM
Nitrogen % 0.5 0.5 2.2 0.3 3.6
Oxygen % NM NM 42.6 47 NM
Sulfur, ppm 30 1000 1700 300 2600
Phosphorus, ppm 875 1 1854 29 3000
NM: not measured. AW: acid whey; ED: energy drinks; WB: wasted bread; PN: paper napkins; CM:
cow manure; TS: total solids; VS: volatile solids; COD: chemical oxygen demand.
Table 5. Composition of selected co-substrates.
Mineral Composition of Selected Co-Substrates
Minerals AW WB CM
Calcium, ppm 1000 ± 282 1100 2340
Potassium, ppm 125 ± 35 2300 2190
Magnesium, ppm 100 ± 0 400 670
Sodium, ppm 290 ± 85 7300 395
Iron, ppm ND 76 163
Zinc, ppm 3.5 10 188
Copper, ppm ND 2 72
Manganese, ppm ND 7 205
Molybdenum, ppm ND ND 2.6
Chlorine ion, ppm ND ND 620
ND: not detected. AW: acid whey; WB: wasted bread; CM: cow manure.
2.2. Inoculum and Substrates
The inoculum used in the experiments was obtained as effluent from a running digester that co-
digests industrial food waste with cow manure. As the effluent contains certain unused nutrients, it
is necessary to pre-incubate the inoculum under anaerobic conditions to minimize the amount of
biogas produced by these available nutrients during the experiment. The pre-incubation stage helps
to deplete the residual biodegradable organic material in the effluent. The effluent was incubated for
7 days at 370C to obtain a degassed inoculum. After the pre-incubation step, the inoculum was
analyzed for TS, VS, and pH. The primary substrate used in this study was mixed cafeteria food waste
(CFW) obtained from one of the university dining halls and contained approximately 50% by weight
of both pre- and post-consumer wastes. Table S1 shows the fraction of principal components in CFW.
The mixed CFW was weighed and ground using a blender (Vitamix) to a particle size sufficiently
small to pass through a 2 mm sieve. The co-substrates studied in this work were acid whey (AW),
waste energy drinks (ED), waste bread (WB), paper napkins (PN), and cow manure (CM). Acid whey
was obtained from a local cheese manufacturer and stored in several small vials at 40C to avoid
repeated thawing. A local digester provided the energy drink cartons; this digester co-digests large
Sustainability 2019, 11, 3875 6 of 15
volume of caffeinated drinks with cow manure and vegetable waste. The packaged food and drinks
contribute to a significant amount of the food sector waste processed by digesters in upstate New
York. Week-old store-bought white bread was used to simulate the waste bread. The bread was cut
into small pieces and dried at 750C for 6 h and powdered for long-term storage. Paper napkins used
in this study were unsoiled Tork H1 ® white paper towels. Paper towels were milled to 2–4 mm
particle size before using in experiments. Cow manure was obtained from a local farm-based digester
and stored in several small containers in the refrigerator. Vegetable waste (VW), in an independent
experiment (Supplementary Information), was prepared by grinding only pre-consumer vegetable
waste from the cafeteria.
2.3. Reactor Start-Up
A total of 6 reactors were used in this study, each with 2.2 L total volume and 1.8 L working
volume. Six reactors were used in a standard configuration as provided by Bioprocess Control (Lund,
Sweden), and connected to the Automated Methane Potential Test System (AMPTSII). The AMPTSII
system continuously measures biomethane production and is designed to work in a semi-continuous
mode, with manual feeding at discrete time intervals. During the start-up, the reactors were filled
with 1.8 L of pre-incubated inoculum. The reactors were then purged with nitrogen gas to create
anaerobic conditions before start-up. All the reactor outlets were connected to the AMPTSII detector
system. A gas sampling T-valve with a self-closing septum was connected between each reactor outlet
and detector to obtain biogas samples for daily compositional analysis. The volume of biogas
withdrawn for sample analysis was not included in the calculation of total daily biogas production.
A 30-day hydraulic retention time (HRT) was maintained for all experiments, and the influent and
effluent flow rates were adjusted manually at 60 mL d−1. All the reactors were incubated at 37 ± 2°C
in a water bath incubator. The digester contents were mixed using a built-in stirrer shaft rotating at
160 rpm with 10 s ‘ON’ and 50 s ‘OFF’ cycles. During the first 14 days of the start-up phase, all the
reactors were fed with CFW at an organic loading rate (OLR) of 0.5 gVSL−1d−1. On the 15th day, the
OLR was increased to 1.4 gVSL−1d−1 and maintained at this level until 45 days had elapsed. The
experiments conducted to test the effect of OLR started after the 45-day start-up phase. The HRT was
kept constant throughout the experiment period. As the feeding and the effluent withdrawal were
carried out manually, exact volumes of the feed and effluent were recorded. The reactor working
volumes were accordingly calculated on a daily basis, to account for any error in feeding and
withdrawal of reactor contents.
2.4. Semi-Continuous Anaerobic Digestion Experiments
The OLR ranged from 1.4 gVSL−1d−1 to 5.5 gVSL−1d−1 after the start-up phase described above. In
the first reactor (R1), the ground CFW was diluted with tap water to attain the required OLR. The pH
and mineral composition of tap water were not accounted for in calculating the feed composition. In
the second reactor (R2) containing CFW:AW mix, acid whey was used instead of water; however, VS
content of AW was adjusted for calculating OLR. The third reactor (R3) contained a mixture of CFW
and energy drink (ED). As the ED contains negligible VS, it was directly used to dilute the CFW to
the required OLR instead of water. The substrates and co-substrates used in each reactor are
summarized in Table 2. The CFW-to-co-substrate ratios for WB (R4), PN (R5), and CM (R6) were
chosen based on a pre-optimization study, conducted soon after the start-up phase. In this pre-
optimization work, all the reactors were maintained at 1.4 gVSL−1d−1, and different combinations of
test mixtures were studied for biogas production for 15 days. Based on these results, CFW co-
digestion with 10 ± 2% WB, 70% CM, and 5–8% PN by weight were selected to investigate the further
effect of increasing OLR; see Table S2 in Supplementary Materials. The reactors were fed every 24 h
according to the experimental design presented in Table S3. These experiments continued for
approximately 100 days, including the time required for recovery of the CFW-only reactor after
failure. Slight variations in daily HRTs due to sample preparation error are reported in Table S4 of
the Supplementary Materials.
Sustainability 2019, 11, 3875 7 of 15
Some of the parameters of interest for growth and product formation are pH, volatile fatty acid
(VFA) concentration, alkalinity, VFA-to-alkalinity ratio, dissolved ammonium concentration, daily
biogas production, daily methane, carbon dioxide and hydrogen sulfide composition of biogas, and
specific methane yield. These parameters were measured on a regular basis and discussed in this
3.1. Food Waste Characteristics
Although different CFW samples were obtained from the same source, considerable variation in
the properties was observed, as indicated in Table 3. The pH of the mixed CFW used in this study
only affected the reactor pH by 0.2 to 0.3 units immediately after feeding. However, the variation can
be significant when digesting in a large-scale digester with continuous feeding cycles. As
methanogens are very sensitive to changing environments, it is necessary to maintain a balance
between substrates and co-substrates. Therefore, if the main feedstock for biogas production does not
possess homogeneous composition, a co-substrate with consistent properties must always be used.
Various physical and chemical properties of CFW and other co-substrates are summarized in Table
4; Table 5. The digestate characteristics were determined only at OLR of 2.8 gVSL−1d−1 and provided
in Table S5 in the Supplementary Materials.
3.2. Process Monitoring
Figure 1a shows the average pH of each reactor at different OLRs. In R1, the daily pH dropped
to 6.2 on the seventh day from an initial pH of 7.3 at OLR of 4.4 gVSL−1d−1. The pH drop caused reactor
failure and led to an excessive CO2 fraction in the produced biogas. The reactors with acid whey
(AW), waste bread (WB), energy drink (ED), and cow manure (CM) as co-substrates maintained the
daily pH between 7.3 and 7.5 throughout the experimental duration. There were no major issues with
process stability with these co-substrates. With paper napkins as the co-substrate (R5), the reactor pH
varied between 6.8 and 7.3 at different OLRs. However, pH variation was minimal at each OLR,
indicating a steady process. If the pH drops below 6.8, it is advised to reduce the OLR as lower pH
values correspond to VFA accumulation and imply the reactor is undersized for a given OLR.
Analyzed at discrete time intervals, the pH of all the digesters reduced by 0.2 to 0.3 units immediately
after feeding, and recovered within 2 h.
At OLR of 4.4 gVSL−1d−1, R1 reached a total VFA concentration of 3375 mg (CH3COOH) L−1 on the
7th day, where the reactor produced less than half of the daily methane, leading to reactor failure.
Therefore, the OLR of the CFW reactor was reduced to 3.5 gVSL−1d−1 for further experiments after an
initial pH adjustment; however, a high average VFA concentration of 2288 mg (CH3COOH) L−1 at this
OLR indicated reactor overload. Therefore, it is recommended to keep the OLR between 1.4 and 2.8
gVSL−1d−1 to anaerobically digest the CFW without any co-substrates. The other reactors (R2—R6)
maintained an acceptable VFA concentration at 4.4 gVSL−1d−1, ranging between 508 and 818 mg
(CH3COOH) L−1. When the OLR was increased further to 5.5 gVSL−1d−1, only R2 had a VFA
concentration lower than 600 gVSL−1d−1, whereas all the other reactors had VFAs ranging between
1087 and 1307 mg (CH3COOH) L−1. These results (Figure 1b) indicate that acid whey may be a viable
non-manure substrate to co-digest CFW at high OLR where it is not convenient to haul manure, for
example, in institutional applications, such as hospitals and universities.
The average alkalinity levels of each reactor are shown in Figure 1c for different OLRs. When R1
failed to produce methane at higher OLR, it had an alkalinity of 2488 mg CaCO3 L−1. At lower OLRs,
the alkalinity ranged between 3700 and 4200 mg CaCO3 L−1, which is lower than that observed in
manure digesters; however, there were no observed instability issues. For other digesters with food
waste co-substrates, the alkalinity levels ranged between 4324 and 7307 mg CaCO3 L−1. At alkalinity
levels above 6500 mg CaCO3 L−1, the reactors did not perform well concerning methane production
even though there was no observed reactor failure. The average variation in alkalinity is shown in
Figure 1c. Methane production did not increase significantly in R2 and R3 and reduced in R4, R5, and
Sustainability 2019, 11, 3875 8 of 15
R6. It is important to maintain an acceptable VFA to alkalinity ratio (V/A) during digester operation,
with municipal digesters typically operating at a V/A ratio below 0.3 (Aquafix,
https://teamaquafix.com/anaerobic-digester-upset-troubleshooting/). The CFW reactor had a V/A of
one when it failed. In R4 and R5, the methane production rate was reduced when V/A reached a value
of 0.3. The other reactors R2, R3, and R6 always maintained an acceptable V/A. Average ammoniacal
nitrogen (NH3-N) concentration throughout the experimental duration varied from 368 to 1132 mg
L−1 in all the reactors, as shown in Figure 1d. Ammoniacal nitrogen levels did not change significantly
in any reactor, even at higher OLRs, and remained relatively stable throughout the experimental
period. In a review article , it was reported that a broad range of ammoniacal nitrogen
concentrations is inhibitory, ranging from 1700 mg L−1 to 15,000 mg L−1. Methanogens can acclimate
to increasing ammonia concentration with time. Therefore, it was difficult to recommend a safe
operating zone for ammonia during biogas production.
Figure 1. Average observed values of (a) pH; (b) total volatile acids [mgCH3COOH L−1]; (c) alkalinity
[mg CaCO3 L−1]; (d) ammoniacal nitrogen [mg NH3-N L−1]. Co-substrates: R1—none; R2—AW (acid
whey); R3—ED (energy drinks); R4—WB (wasted bread); R5—PN (paper napkins); R6—CM (cow
The average daily biogas production rate and methane, carbon dioxide, and hydrogen sulfide
composition of biogas are shown in Figure 2a through 2d. In R1, the average methane level reached
a high of 62% at 2.8 gVSL−1d-1 but was reduced to 44% at 4.4 gVSL−1d−1, indicating a need to stop
feeding and let the reactor stabilize for several days to attain a normal methane production level. In
R2, the methane level reached as high as 71% at 4.4 gVSL−1d−1 with a daily average of 66%, and this
reactor maintained greater than 62% average methane concentration throughout the measurement
period. With energy drink (ED) as co-substrate in R3, the biogas contained a high concentration of
H2S, even though it showed higher methane levels of 60 to 66%. With ED as co-substrate, the H2S
concentration was greater than 2000 ppm for the first two OLRs and reduced to 1194 ppm at 5.5
gVSL−1d−1. The average daily H2S concentrations ranged from 78 to 83 ppm for R2, 82 to 326 ppm for
R4, 91 to 391 for R5, and 139 to 810 ppm for R6 within the safe operating OLR region. The H2S levels
of R2, R4, and R5 were significantly lower than R6.
Sustainability 2019, 11, 3875 9 of 15
Figure 2. Biogas properties: (a) average daily biogas production rate; (b) methane fraction; (c) CO2
fraction; (d) H2S concentration [ppm]. Co-substrates: R1—none; R2—AW (acid whey); R3—ED
(energy drinks); R4—WB (wasted bread); R5—PN (paper napkins); R6—CM (cow manure). For R3,
the H2S concentration reached a detection limit of 2000 ppm of the instrument.
Figure 3a shows the average methane production rate, and Figure 3b depicts the specific
methane yields (SMYs) of different reactors at each given OLR. In CFW digestion without any co-
substrates, observed SMY was 352 ± 46 mL gVS−1d−1 at 1.4 gVSL−1d−1 and 363±28 mL gVS−1d−1 at 2.8
gVSL−1d−1 with all the other process parameters within the acceptable range. Therefore, if CFW has to
be digested alone, it is recommended to keep the OLR below 2.8 gVSL−1d−1, preferably between 1.5
and 2.0 gVSL−1d−1. However, it was possible to digest food waste at high OLRs using the previously
identified unconventional co-substrates. For example, with paper napkins as the co-substrate (R5), a
maximum SMY of 381 ± 30 mL gVS−1d−1 was observed at 2.8 gVSL−1d−1. The reactors R2, R3, R4, and
R6 showed maximum SMYs of 455 ± 31, 453 ± 20, 479 ± 29, and 372 ± 41 mL gVS−1d−1, respectively, at
OLR of 4.4 gVSL−1d−1 with all other process parameters within acceptable ranges. Acid whey, energy
drinks, and waste bread were the most efficient co-substrates with higher methane yield and lower
VFA levels. The SMYs of all the reactors reduced significantly at 5.5 gVSL−1d−1 compared to lower
OLRs, indicating reactor overload. A higher SMY from all the co-digestion mixtures compared to
CFW indicated a synergistic relationship between CFW and co-substrates. Reactors R2, R3, and R4
showed a higher observed SMY compared to R6, where cow manure was the co-substrate.
Sustainability 2019, 11, 3875 10 of 15
Figure 3. (a) Daily average methane production rate [mL d
]; (b) Specific methane yield (SMY) [mL
]. Co-substrates: R1—none; R2—AW (acid whey); R3—ED (energy drinks); R4—WB (wasted
bread); R5—PN (paper napkins); R6—CM (cow manure).
Biogas productivity is an indicator of process stability and must be monitored on a regular basis.
Reactors R2, R3, and R4 showed increasing productivities with increasing OLR up to 4.4 gVSL
(Figure 4a), and productivity either reduced or remained the same at a loading of 5.5 gVSL
all reactors except R2 and R6. No further increase in productivity indicates the onset of reactor
overload and implies that the OLR should be reduced to maintain stable operation. The productivity
in R1 did not increase further after an OLR of 2.8 gVSL
, and the productivity decreased in R5 at
the highest OLR tested. Therefore, it is suggested that OLRs be kept below 4.4 gVSL
whey, bread, and energy drinks are used as co-substrates with CFW, below 2.8 gVSL
digesting CFW with cow manure and paper napkins, and between 1.4 and 2.0 when digesting CFW
alone. The degradability or percent degradation signifies waste management efficiency because it is
directly proportional to the amount of food waste converted into biogas. All the reactors had a lower
biodegradability at the highest OLR, as indicated in Figure 4b. A significant improvement in
biodegradation from 1.4 to 2.8 gVSL
was observed in all the reactors. However, at 4.4 gVSL
only R2, R3, and R4 had higher COD removal. The COD removal reduced in all the reactors at 5.5
except for R5.
Figure 4. Parameters indicative of process efficiency: (a) biogas productivity expressed as the volume
of biogas produced per unit volume of the reactor per day; (b) average fractional COD (chemical
oxygen demand) degradation at each OLR (organic loading rate). Co-substrates: R1—none; R2—AW
(acid whey); R3—ED (energy drinks); R4—WB (wasted bread); R5—PN (paper napkins); R6—CM
Anaerobic digestion involves a synergistic metabolism between different classes of microbes:
hydrolyzing bacteria, acetogens, acidogens, and methanogens. These microbial communities differ
Sustainability 2019, 11, 3875 11 of 15
significantly in their morphology, optimum conditions for growth and product formation, and
sensitivity to changing microenvironments. Therefore, it is necessary to monitor different process
parameters to maintain a healthy balance between microbial populations and achieve a steady
The nutrient composition of the substrates directly affects microbial growth and biogas
production. The pH of the substrate supports faster acclimatization of the microbial population to
changing environments and solubilizes certain nutrients for easy uptake by microbes. Maximum
biogas production was observed when the pH of the food waste was 7.0, with a significant reduction
in biogas production at pH of 5.0 and 9.0 under batch conditions . The carbohydrates, lipids, and
proteins contribute to maintaining a healthy carbon-to-nitrogen ratio (C/N). Proteins act as a nitrogen
source upon degradation into ammonia. A high level of proteins in the substrate, for example, from
meat products, corresponds to lower C/N and leads to process instabilities by producing excessive
The daily pH of the anaerobic digester is known to impact digester performance by affecting the
mass transfer rate. In a substrate containing a high concentration of ammoniacal nitrogen, pH affects
the ratio of free ammonia (NH3) to the ionized form of ammonia (NH4+) . As the pH increases,
ammonia toxicity increases due to the increase in free ammonia. Methanogens are susceptible to
higher concentrations of ammonia and, therefore, they consume VFAs at a slower rate. Slower VFA
consumption leads to their accumulation and creates a low pH environment in the digester.
Therefore, if acetogens outnumber the methanogens, pH will drop, which can inhibit methanogens,
and ultimately lead to digester failure. A balanced metabolism of acetogens and methanogens helps
in maintaining the pH of a digester within the optimum range. A pH of 6.8–7.5 for a healthy
population of methanogens  has been suggested in earlier literature studies for manure and co-
digestion. Even though pH cannot be a single parameter that determines digester stability, for food
waste-only digestion, it is recommended to maintain a pH of 7.2–7.8 based on the observations in this
study. The pH drop was observed to be much faster below a level of 7.2 in food waste digestion.
Though pH is not an early indicator of process instability, it is important to maintain a constant
digester pH and, hence, VFA levels at all times. Maintaining lower concentrations of VFAs helps both
in attaining higher methane production and better waste conversion efficiency. Because of the
inhibitory effects of volatile fatty acids, it is required to monitor the VFA concentration in the digester
on a regular basis. Consistently elevated levels of VFAs indicate digester overload and can ultimately
lead to digester failure. Researchers have studied different strategies to decrease the negative effect
of VFAs, including co-digestion, the addition of certain metal ions like Ca2+ , and reagents that
increase alkalinity . Another approach suggests that a discontinuous feeding profile can avoid
VFA accumulation in the digester . While the existing literature is ambiguous on acceptable levels
of VFA, a total VFA concentration of below 800 mg (CH3COOH) L−1 at all times would be
recommended in this study to maintain optimum digester operation. In the event that VFA levels rise
above the recommended value, the temperature of the reactor must be reduced by 3–50C, the feeding
must be stopped (or OLR appropriately reduced) for a few days until the pH becomes normal, or the
pH should be adjusted to 7.2 with a base additive, such as sodium hydroxide.
Alkalinity levels lower than the optimum indicate VFA accumulation and can be maintained by
using an appropriate co-substrate that has the natural buffering ability or by using external agents
like calcium carbonate or sodium bicarbonate. The use of waste materials like egg shells and lime
mud from pulp and paper processing has been proposed for maintaining digester alkalinity . In
a manure-only digester maintained at pH = 7.4, normal alkalinity levels were observed to be 5500 mg
CaCO3L−1 to maintain stable operation . As manure-only digesters are known to run stably for a
long time, this value can also serve as a basis for average required alkalinity levels in food waste-only
digestion. Alkalinity also affects the digestate characteristics by changing the phosphorus (struvite)
removal efficiency .
Ammoniacal nitrogen refers to the nitrogen from free ammonia (NH3), and ammonium ions
(NH4+) in the digester are the end products of protein, amino acid, and urea degradation. Ammoniacal
nitrogen levels did not change significantly in any reactor, even at higher OLRs, and remained
Sustainability 2019, 11, 3875 12 of 15
relatively stable throughout the experimental period. In a review article by Chen et al. , it was
reported that ammoniacal nitrogen concentrations are inhibitory in the range of 1700 to 15,000 mg
L−1. Methanogens have a capability to acclimatize to increasing ammonia concentration with time.
Therefore, it was difficult to recommend a safe operating zone for ammonia during biogas production
in the current experimental campaign. Free ammonia is known to affect methanogenic activity by
inhibiting the methane-producing enzymes or by diffusing into the microbial cells, causing proton
imbalance or potassium deficiency . Liu et al.  observed that NH4+-N concentration of 1000
mg L−1 or higher was inhibitory in anaerobic digestion of municipal solid waste leachate using an
expanded granular sludge anaerobic reactor . These authors also achieved a higher COD removal
efficiency by maintaining the NH4+-N concentrations below 500 mg L−1. A lower concentration of
NH4+ or free ammonia is beneficial to anaerobic digestion, as these compounds serve as a nitrogen
source for microbes. Inhibition effects of ammoniacal nitrogen depend on the type of substrate, the
presence of other metal ions , and process conditions like temperature and pH.
Biogas production is by far the most important parameter to monitor in anaerobic digestion.
Measuring the biogas production daily helps to identify any stability issues arising during the
process. Biogas contains two main components, methane and carbon dioxide, with small amounts of
hydrogen sulfide (H2S), ammonia, nitrogen, and hydrogen. In a well-controlled digester, the methane
percentage of biogas varies between 55% and 65%. In a continuously fed digester at steady-state,
daily biogas composition should remain constant over time. Methane content below 55%, or CO2
content above 35–40%, indicates VFA accumulation and inhibition in the activity of methanogens.
Methane, which is a result of volatile solids destruction, is the final product of the anaerobic digestion
pathway, suggesting that higher methane production indicates better waste processing efficiency of
In addition to methane, it is necessary to monitor the H2S concentration of biogas. Though
sulfides help in maintaining the alkalinity similar to ammonia, higher sulfide concentrations are toxic
to methanogenic bacteria. Also, higher H2S concentration in biogas demands additional
infrastructure for purification before its use. The higher H2S levels with an energy drink as the co-
substrate may be attributed to the presence of taurine. Taurine, or 2-aminoethanesulfonic acid, acts
as a source of sulfur for anaerobic bacteria. These microbes dissimilate taurine to produce sulfite,
which is a nutrient source. The microbes then carry out sulfite respiration through sulfate reductase
enzyme, and sulfides are excreted out of the cells . Sulfide, excreted as hydrogen sulfide gas,
makes a major component of biogas. The energy drink also contains caffeine, which is a well-known
stimulant of biogas production . Therefore, it is important to characterize the feedstocks for the
presence of specific substrates that may cause unusual problems even after being stimulatory to
biogas production. In the current work, daily average biogas production was higher with acid whey
and waste bread compared to cow manure as a co-substrate. Therefore, acid whey and waste bread
can make better co-substrates than cow manure for CFW digestion.
Specific methane yield is the volume of the methane produced per gram of volatile solids added
per day. The SMY relates to the extent of biodegradability of each substrate. Babaee and Shayegan
 investigated the effect of OLR on vegetable waste digestion in a scale reactor operating at steady-
state. They suggested an OLR of 1.4 gVSL−1d−1 as the design criterion, with a SMY of 250 mL gVS−1
d−1. This paper also recommended an OLR of 1.4 gVSL−1d−1 for vegetable waste digested in semi-
continuous mode, as elevated VFA concentration was observed at higher OLR. Vegetable waste had
a SMY of 198 ± 48 mL gVS−1d−1 when mixed continuously, and 350 ± 90 mL gVS−1d−1 with intermittent
mixing in an independent experiment (not included in the graphs; see Supplementary Materials Table
S6). Specific methane yield is dependent on waste composition and process conditions. An OLR limit
of 1.5 gVSL−1d−1 was suggested in a previous study for mixed food waste digestion without any co-
substrates, yielding a SMY of 371 mL gVS−1d−1 . Food waste digestion at considerably high OLR
of up to 5.6 gVSL−1d−1 was achieved using special strategies like thermophilic digestion  and lipid
Sustainability 2019, 11, 3875 13 of 15
Use of unconventional co-substrates helped enhance anaerobic digestion of food waste at high
organic loading rates. These co-substrates generally resulted in increased daily methane production,
higher methane fraction in biogas, improved waste degradation, and process stability. Pure CFW
digestion is challenging at higher organic loading rates, and process instabilities are often observed;
however, it is possible to digest CFW if OLR is consistently kept lower. Our results show that, during
pure CFW digestion, a high level of volatile acid accumulation occurs, indicating poor degradation.
Digesting CFW at low OLRs will need a considerably larger volume of the digester than conventional
substrates for the same OLR, increasing the upfront capital cost.
In summary, it may not be practical to build “food waste only” digesters to enable higher OLR.
This study sh owed tha t, when C FW is mixed w ith wide ly avail able was tes like acid whey, caffeinat ed
energy drinks, waste bread, paper napkins, and conventional co-substrates, such as cow manure, the
digestion process can be stabilized and lead to higher biogas yield and biodegradability. In addition,
co-digesting food waste with bread, acid whey, and paper napkins has been shown to reduce
hydrogen sulfide emissions and ammoniacal nitrogen in the digester. Because of the synergistic effect
offered by the co-substrates, the primary feedstock (i.e., CFW or mixed food waste) is utilized more
efficiently, leading to increased biodegradability. It is, therefore, recommended that CFW be co-
digested with more homogenous substrates that do not frequently change in their composition to
achieve a long-term steady-state process. Co-digestion has a beneficial impact in reducing the design
volumes of reactors, making provisions for treating large amounts of food waste. It is important to
note that commercial-scale digesters may process more than two food sector waste materials at the
same time, in which case, the interaction between these wastes may vary. Because the reported
experiments were carried out in semi-continuous mode with once daily feeding, at higher organic
loading rates, there is a possibility of nutrient shock soon after the feeding, which could have affected
some of the measured parameters like total volatile acids. Such an effect may be less likely in a
continuous digester with multiple feeding cycles per day. The substrates used in this work are
generated in large amounts in New York State, but the use of other potential co-substrates like waste
cooking oil, grease trap waste, and fruit and vegetable processing wastewater need further
evaluation. Co-digestion would be expected to have beneficial impacts on reducing digester volume
and water footprint of anaerobic digestion processes while increasing food waste management
throughput and renewable electricity production. The co-substrates suggested through this study
may help in the deployment of decentralized digesters in settings where it is not practical to haul
manure, for example, in institutional or commercial installations, such as hospitals, universities, or
large grocery stores.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: Sample
kinetics of cafeteria food waste (CFW) digestion, Table S1: Components of mixed cafeteria food waste (CFW),
Table S2: Optimization to choose the ratio of CFW to co-substrates, Table S3: Experimental design for feeding
CFW and co-substrates at different OLRs, Table S4: Observed hydraulic retention times at different OLRs, Table
S5: Digestate characteristics at OLR = 2.8 gVSL−1d−1, Table S6: Process monitoring during digestion of mixed
cafeteria food waste (CFW) and vegetable waste (VW) with continuous mixing.
Author Contributions: All the experiments were designed and performed by SH with guidance from TT, and
both authors contributed to preparing this manuscript.
Funding: Provided by the New York State Pollution Prevention Institute (NYSP2I) through a grant from the
NYS Department of Environmental Conservation, which also provided a Graduate Research Assistantship for
S. Hegde. Any opinions, findings, conclusions, or recommendations expressed are those of the authors and do
not necessarily reflect the views of the Department of Environmental Conservation.
Conflicts of Interest: The authors declare no conflict of interest. The sponsors had no role in the design,
execution, interpretation, or writing of the study.
1. Ebner, J.H.; Labatut, R.A.; Lodge, J.S.; Williamson, A.A.; Trabold, T.A. Anaerobic co-digestion of
commercial food waste and dairy manure: Characterizing biochemical parameters and synergistic effects.
Sustainability 2019, 11, 3875 14 of 15
Waste Manag. 2016, 52, 286–294, doi:10.1016/j.wasman.2016.03.046.
2. Pei, Z.; Liu, J.; Shi, F.; Wang, S.; Gao, Y.; Zhang, D. High-solid Anaerobic Co-digestion of Food Waste and
Rice Straw for Biogas Production. J. Northeast. Agric. Univ. 2014, 21, 61–66, doi:10.1016/S1006-8104(15)30021-
3. Stan, C.; Collaguazo, G.; Streche, C.; Apostol, T.; Cocarta, D.M. Pilot-scale anaerobic co-digestion of the
OFMSW: Improving biogas production and startup. Sustainability 2018, 10, 1939, doi:10.3390/su10061939.
4. Martín-González, L.; Colturato, L.F.; Font, X.; Vicent, T. Anaerobic co-digestion of the organic fraction of
municipal solid waste with FOG waste from a sewage treatment plant: Recovering a wasted methane
potential and enhancing the biogas yield. Waste Manag. 2010, 30, 1854–1859,
5. Kotsyurbenko, O.R.; Chin, K.-J.; Glagolev, M.V.; Stubner, S.; Simankova, M.V.; Nozhevnikova, A.N.;
Conrad, R. Acetoclastic and hydrogenotrophic methane production and methanogenic populations in an
acidic West-Siberian peat bog. Environ. Microbiol. 2004, 6, 1159–1173, doi:10.1111/j.1462-2920.2004.00634.x.
6. WRAP. Operators Briefing Note; Defra Research Project, U.K. 2012. Available online:
http://www.wrap.org.uk/sites/files/wrap/Operators%20Briefing%20Note.pdf (accessed on 12 November
7. Serna-Maza, A.; Heaven, S.; Banks, C.J. Ammonia removal in food waste anaerobic digestion using a side-
stream stripping process. Bioresour. Technol. 2014, 152, 307–315, doi:10.1016/j.biortech.2013.10.093.
8. Serna-Maza, A.; Heaven, S.; Banks, C.J. In situ biogas stripping of ammonia from a digester using a gas
mixing system. Environ. Technol. 2017, 38, 3216–3224, doi:10.1080/09593330.2017.1291761.
9. Cho, S.-K.; Lee, M.-K.; Kim, D.-H.; Yun, Y.-M.; Jung, K.-W.; Shin, H.-S.; Oh, S.E. Enhanced anaerobic
digestion of livestock waste by ultrasonication: A tool for ammonia removal and solubilization. Korean J.
Chem. Eng. 2014, 31, 619–623, doi:10.1007/s11814-013-0284-4.
10. Lin, L.; Yuan, S.; Chen, J.; Xu, Z.; Lu, X. Removal of ammonia nitrogen in wastewater by microwave
radiation. J. Hazard. Mater. 2009, 161, 1063–1068, doi:10.1016/J.JHAZMAT.2008.04.053.
11. Krakat, N.; Demirel, B.; Anjum, R.; Dietz, D. Methods of ammonia removal in anaerobic digestion: A
review. Water Sci. Technol. 2017, 76, 1925–1938, doi:10.2166/wst.2017.406.
12. Ariunbaatar, J.; Esposito, G.; Yeh, D.H.; Lens, P.N.L. Enhanced Anaerobic Digestion of Food Waste by
Supplementing Trace Elements: Role of Selenium (VI) and Iron (II). Front. Environ. Sci. 2016, 4, 8,
13. Facchin, V.; Cavinato, C.; Fatonea, F.; Pavan, P.; Bolzonell, D. Effect of trace element supplementation on
the mesophilic anaerobic digestion of foodwaste in batch trials: The influence of inoculum origin. Biochem.
Eng. J. 2013, 70, 71–77, doi:10.1016/J.BEJ.2012.10.004.
14. Banks, C.J.; Zhang, Y.; Jiang, Y.; Heaven, S. Trace element requirements for stable food waste digestion at
elevated ammonia concentrations. Bioresour. Technol. 2012, 104, 127–135, doi:10.1016/j.biortech.2011.10.068.
15. Li, L.; Peng, X.; Wang, X.; Wu, D. Anaerobic digestion of food waste: A review focusing on process stability.
Bioresour. Technol. 2018, 248, 20–28, doi:10.1016/j.biortech.2017.07.012.
16. Xu, F.; Li, Y.; Ge, X.; Yang, L.; Li, Y. Anaerobic digestion of food waste—Challenges and opportunities.
Bioresour. Technol. 2018, 247, 1047–1058, doi:10.1016/j.biortech.2017.09.020.
17. Wang, X.; Lu, X.; Li, F.; Yang, G.; Banks, C.; Humphreys, P. Effects of Temperature and Carbon-Nitrogen
(C/N) Ratio on the Performance of Anaerobic Co-Digestion of Dairy Manure, Chicken Manure and Rice
Straw: Focusing on Ammonia Inhibition. PLoS ONE 2014, 9, e97265, doi:10.1371/journal.pone.0097265.
18. Zhang, Y.; Banks, C.J.; Heaven, S. Co-digestion of source segregated domestic food waste to improve
process stability. Bioresour. Technol. 2012, 114, 168–178, doi:10.1016/j.biortech.2012.03.040.
19. Nayono, S.E.; Gallert, C.; Winter, J. Co-digestion of press water and food waste in a biowaste digester for
improvement of biogas production. Bioresour. Technol. 2010, 101, 6987–6993,
20. Prabhu, S.M.; Mutnuri, S. Anaerobic co-digestion of sewage sludge and food waste. Waste Manag. Res. 2016,
21. Babaee, A.; Shayegan, J. Effect of Organic Loading Rates (OLR) on Production of Methane from Anaerobic
Digestion of Vegetables Waste. In World Renewable Energy Congress-Sweden; Linköping University
Electronic Press: Linköping, Sweden. 2011; pp. 411–417, doi:10.3384/ecp11057411.
22. Zahan, Z.; Othman, M.Z.; Rajendram, W. Anaerobic Codigestion of Municipal Wastewater Treatment Plant
Sludge with Food Waste: A Case Study. BioMed Res. Int. 2016, 2016, 1–13, doi:10.1155/2016/8462928.
Sustainability 2019, 11, 3875 15 of 15
23. EPA. METHOD 1684: Total, Fixed, and Volatile Solids in Water, Solids, and Biosolids, Environmental Protection
Agency: Columbia, WA, USA, 2001.
24. Chen, Y.; Cheng, J.J.; Creamer, K.S. Inhibition of anaerobic digestion process: A review. Bioresour. Technol.
2008, 99, 4044–4064, doi:10.1016/j.biortech.2007.01.057.
25. Jayaraj, S.; Deepanraj, B.V.S. Study on the effect of ph on biogas production from food waste by anaerobic
digestion. In Proceedings of the 9
Annual Green Energy Conference, Tianjin, China, 25–28 May 2014.
26. Shah, A.A.; Nawaz, A.; Kanwal, L.; Hasan, F.; Khan, S.; Badshah, M. Degradation of poly(ε-caprolactone)
by a thermophilic bacterium Ralstonia sp. Strain MRL-TL isolated from hot spring. Int. Biodeterior.
Biodegrad. 2015, 98, 35–42, doi:10.1016/j.ibiod.2014.11.017.
27. Kumar, D.; Singh, V. Dry-grind processing using amylase corn and superior yeast to reduce the exogenous
enzyme requirements in bioethanol production. Biotechnol. Biofuels 2016, 9, 228, doi:10.1186/s13068-016-
28. Zhang, J.; Wang, Q.; Zheng, P.; Wang, Y. Anaerobic digestion of food waste stabilized by lime mud from
papermaking process. Bioresour. Technol. 2014, 170, 270–277, doi:10.1016/j.biortech.2014.08.003.
29. Cavaleiro, A.J.; Pereira, M.A.; Alves, M. Enhancement of methane production from long chain fatty acid
based effluents. Bioresour. Technol. 2008, 99, 4086–4095, doi:10.1016/j.biortech.2007.09.005.
30. Labatut, R.A.; Gooch CA. Monitoring of Anaerobic Digestion Process to Optimize Performance and
Prevent System Failure. Proc Got Manure? Enhancing Environ. Econ. Sustain. 2012, 209–225. Available
online: https://ecommons.cornell.edu/handle/1813/36531 (accessed on 30 January 2017)
31. Liu, X.; Xiang, L.; Song, Y.; Qian, F.; Meng, X. The effects and mechanism of alkalinity on the phosphate
recovery from anaerobic digester effluent using dolomite lime. Environ. Earth Sci. 2015, 73, 5067–5073,
32. Gallert, C.; Bauer, S.; Winter, J. Effect of ammonia on the anaerobic degradation of protein by a mesophilic
and thermophilic biowaste population. Appl. Microbiol. Biotechnol. 1998, 50, 495–501,
33. Liu, A.; Ren, F.; Lin, W.Y.; Wang, J.Y. A review of municipal solid waste environmental standards with a
focus on incinerator residues. Int. J. Sustain. Built Environ. 2015, 4, 165–188, doi:10.1016/j.ijsbe.2015.11.002.
34. Zhang, A.Y.; Sun, Z.; Leung, C.C.J.; Han, W.; Lau, K.Y.; Li, M.; Lin, C.S.K. Valorisation of bakery waste for
succinic acid production. Green Chem. 2013, 15, 690, doi:10.1039/c2gc36518a.
35. Cook, A.M.; Denger, K. Metabolism of taurine in microorganisms: A primer in molecular biodiversity? Adv.
Exp. Med. Biol. 2006, 583, 3–13, doi:10.1007/978-0-387-33504-9-1.
36. Prabhudessai, V.; Ganguly, A.; Mutnuri, S. Effect of caffeine and saponin on anaerobic digestion of food
waste. Ann. Microbiol. 2009, 59, 643–648, doi:10.1007/BF03179203.
37. Liu, C.; Wang, W.; Anwar, N.; Ma, Z.; Liu, G.; Zhang, R. Effect of organic loading rate on anaerobic
digestion of food waste under mesophilic and thermophilic conditions. Energy Fuels 2017, 31, 2976–2984,
38. Li, D.; Sun, Y.; Guo, Y.; Yuan, Z.; Wang, Y.; Zhen, F. Continuous anaerobic digestion of food waste and
design of digester with lipid removal. Environ. Technol. 2013, 34, 2135–2143,
© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).