Bioresource Technology 224 (2017) 197–205

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Dry fermentation of manure with straw in continuous plug flow reactor:

Reactor development and process stability at different loading rates
Regina J. Patinvoh ⇑, Adib Kalantar Mehrjerdi, Ilona Sárvári Horváth, Mohammad J. Taherzadeh
Swedish Centre for Resource Recovery, University of Borås, Sweden

h i g h l i g h t s g r a p h i c a l a b s t r a c t

New plug flow reactor developed for

continuous dry digestion processes.

Reactor worked successfully for
230 days using untreated manure
with straw at 22% TS.

Methane yield of 56% of the
theoretical value was obtained at OLR
of 4.2 gVS/L/d.

OLR of 6 gVS/L/d caused process
instability with 41% VS removal
efficiency.

a r t i c l e i n f o a b s t r a c t

Article history: In this work, a plug flow reactor was developed for continuous dry digestion processes and its efficiency
Received 3 October 2016 was investigated using untreated manure bedded with straw at 22% total solids content. This newly
Received in revised form 1 November 2016 developed reactor worked successfully for 230 days at increasing organic loading rates of 2.8, 4.2 and
Accepted 3 November 2016
6 gVS/L/d and retention times of 60, 40 and 28 days, respectively. Organic loading rates up to 4.2 gVS/
Available online 5 November 2016
L/d gave a better process stability, with methane yields up to 0.163 LCH4/gVSadded/d which is 56% of
the theoretical yield. Further increase of organic loading rate to 6 gVS/L/d caused process instability with
Keywords:
lower volatile solid removal efficiency and cellulose degradation.
Dry fermentation
Plug flow reactor
Ó 2016 Elsevier Ltd. All rights reserved.
Continuous process
Process stability
Reactor development

1. Introduction Visvanathan, 2013; Kothari et al., 2014). Dry anaerobic digestion

technology is designed to process more organic wastes per reactor
Anaerobic digestion of organic wastes for biogas production has volume with total solids (TS) content greater than 20% (Demirer &
been successfully applied but there is a need for improved pro- Chen, 2008; Fernández et al., 2008) treating food wastes, manure
cesses and design of less expensive reactors. In this vein the drive bedded with straw, garden wastes and other high solid waste frac-
for dry anaerobic digestion processes is increasing both in research tions. In comparison with the wet anaerobic digestion process, this
and the industry because of technical simplicity in design, together process allows higher organic loading rates (OLR), less pretreat-
with low construction and operational costs (Karthikeyan & ment and gives better economic feasibility (Karthikeyan &
Visvanathan, 2013) since the reactor volume is minimized and it
is easier to handle the digestate residue.
⇑ Corresponding author. Animal wastes and crop residues are one of the most abundant
E-mail address: [email protected] (R.J. Patinvoh). waste fractions generated worldwide, the amount of manure

http://dx.doi.org/10.1016/j.biortech.2016.11.011
0960-8524/Ó 2016 Elsevier Ltd. All rights reserved.
198 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205

bedded with straw produced increases daily as the number of system to avoid the release of gas and air entering the system. This
housed dairy herd increases. Cattle manure bedded with straw part can be considered as an inactive zone in the inlet system. A
usually have a TS greater than 20%, farmers can use low cost anaer- shutoff valve was connected at the inlet with a piston rod (made
obic digesters to convert these enormous waste streams to biogas, gas tight by O-rings). The inlet storage volume is about 0.256 L
thereby improve water quality, reduce methane and nitrous oxide with 0.0053 L of feedstock per rotation.
emissions and improve soil fertility. Dry anaerobic digestion is Main Zone – the main zone of the reactor was made of a Poly-
therefore a better option for processing these wastes. Since biogas methylmethacrylate (PMMA) material with 110 mm outer diame-
production involves a complex biological process, monitoring of ter and 5 mm thickness for easy handling and to make the reactor
the process is essential to avoid process instability and failure of transparent. An impeller, installed on a hexagonal shaft that runs
the digester (Drosg, 2013). For enhanced performance of dry anaer- through the reactor connected the inlet to the outlet. The hexago-
obic digestion processes, a suitable reactor is required; considering nal shaft had two oil seals that prevented leakage of materials and
the substrate composition, amount of substrate to be treated, and several O-rings making the different parts gas tight. The impeller
process economy of the reactor. allowed mixing of the feedstock at the bottom part of the inlet,
Plug flow reactors have been reported to be efficient for dry and transported the materials very slowly towards the outlet tak-
anaerobic digestion processes. These reactors are inexpensive ing several rotations: 100 cm3 per rotation but it can also depend
and easy to build which make them a suitable technology to on the viscosity of feedstock and working volume of reactor.
improve the livelihoods of farmers (Lansing et al., 2010, 2008). Plug Outlet zone – this part was also made of polyethylene material
flow reactors have also been reported to have the highest success with an outlet pipe for the collection of the digestate residue. The
rate in the United States, where 42% out of the 242 anaerobic material was transferred from inlet towards outlet at the end of the
digesters operating at livestock farms in 2015 were plug flow reactor by rotating of the impeller shaft; as a result the digested
designs (USEPA, 2016). Nevertheless, some shortcomings, such as residue was discharged through the outlet pipe while new materi-
lower mass transfer due to lack of mixing, thermal stratification als were added. The impeller handle was connected to the outlet
and solid sedimentation problems have been reported (Lansing for manual rotation. In this zone, the gas outlet was also connected
et al., 2010). These problems can be minimized by the use of impel- where the daily volume of biogas produced was measured by the
lers in plug flow reactors. The impellers allow minimal mixing for tipping device in the automated methane potential test system
better performance in the reactors. In high solid digestion pro- (AMPTs); meanwhile with an inserted thermocouple the tempera-
cesses, however, continuous mixing have been reported to indicate ture of the reactor was controlled.
unstable performance at high OLR and it was observed that the Safety measures – a construction was made with a polyethylene
continuously mixed unstable reactor became stable when the mix- material at the inlet and outlet with specific rubber connectors for
ing level was reduced (Stroot et al., 2001). Researchers have also sealing. If pressure inside the reactor would increase beyond 10 kPa
investigated the effectiveness of plug flow reactors on manure due to blockage in the gas outlet or for any reason, this part of the
and other substrates with solid content in the range of 11–14% reactor could open up avoiding an explosion. All interior parts were
TS (Adl et al., 2012; Cantrell et al., 2008). There have been studies carefully selected in order to avoid any chemical reaction with the
on dry digestion of different substrates in batch reactors but little experimental materials as well as any corrosion during the test.
information is available on dry digestion in continuous plug flow Fig. 1 shows a schematic diagram of the reactor and the exper-
reactors. The innovation of this paper in meeting this gap was imental set up together with other accessories, such as heater and
the development of a novel type solid-state plug-flow laboratory water bath for maintaining the temperature, sampling point for
reactor to treat substrates at higher TS levels, i.e. greater than 20%. biogas composition analysis and the AMPTs system for measuring
The efficiency of this reactor was then investigated in a contin- the biogas produced.
uous dry anaerobic digestion process treating manure bedded with
straw at 22% TS content. The main objective was to identify the 2.2. Substrates and inoculum
critical OLR, above which instability can occur in the reactor. The
process was monitored by measuring VFA/Alkalinity ratio, pH, The substrate, cattle manure bedded with straw, was collected
volatile fatty acid (VFA) and total ammonia nitrogen concentra- from a cattle farm outside Borås (Sweden) and used as feedstock
tions regularly. This is vital because the cost of starting the entire for the continuous anaerobic digestion process. During the experi-
process over again far outweighs monitoring of the process. mental period two different batches, with similar content of total
solids (TS) and volatile solids (VS) were obtained from the same
farm. The manure was shredded manually to reduce the particle
2. Materials and methods size of straw; then it was characterized, weighted and stored in
plastic containers at 20 °C to prevent biodegradation until further
2.1. Reactor design use. During experiment, weighted frozen substrate was defrosted
at room temperature and thoroughly mixed to gain a homogenized
A horizontal plug flow reactor (Fig. 1) for continuous dry anaer- feed before use. Sludge used as inoculum was obtained from a
obic digestion process was designed and made-up at the University digester treating waste water sludge and operating at mesophilic
of Borås, Sweden. This reactor has 9.2 L total volume, 1565 mm conditions (Vatten and Miljö i Väst AB, Varberg, Sweden). The
length and maximum inside pressure of 10 kPa. It was mounted inoculum was filtered through a 2 mm porosity sieve to remove
on a base surface with which a clamp to the two edges of the reac- sand, plastic and other unwanted particles after which it was accli-
tor was associated for suspending the reactor. Mesophilic (37 °C) mated for five days in an incubator at 37 °C prior to use. The inocu-
conditions were maintained by circulating water from a heater, a lum was centrifuged at 10,000g for 10 min to obtain a TS content of
thermostatic water bath (GD 100, Grant instruments Ltd., Cam- 7.8 ± 0.24%. Table 1 shows the most important characteristics of
bridgeshire, UK), through a water jacket (150 mm outer diameter the substrate and the inoculum used during the investigations.
with 4 mm wall thickness and capacity of 5 L) surrounding the
reactor. The reactor was then shielded with a 10 mm thick- 2.3. Experimental procedure
Styrofoam to avoid heat loss.
The reactor was sub-divided into 3 zones as follows. Inlet zone – The substrate was inoculated with inoculum to start-up the
this part was made of polyethylene material with sealed buffer reactor, keeping a volatile solids (VS) ratio (VSsubstrate to VSinoculum)
 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205 199

Fig. 1. Schematic diagram of developed reactor and experimental set-up.

Table 1
Characteristics of Substrate and Inoculum used during experiment (standard devi-
corresponding retention times of 60, 40 and 28 days. For the exper-
ation based on at least duplicate measurements).
imental period (start-up and OLR1) the first batch of manure was
Parameters Manure with Straw Anaerobic sludge utilized, while another batch of the feedstock was used during
Total solids (%) 22.29 ± 2.78 7.80 ± 0.24 the experimental period of OLR 2 and OLR 3. Each OLR condition
Volatile solids (%)a 70.44 ± 2.68 40.46 ± 1.12 was kept until a period of at least one corresponding retention
Moisture (%) 77.72 ± 2.78 92.20 ± 0.30 time. The reactor was fed regularly in every second day and the
Ash (%)a 29.56 ± 2.68 59.54 ± 1.12
Fat content (%)a 1.37 ± 0.24 ND
digestate residue was withdrawn before feeding, and kept for anal-
Total carbon (%)a 39.14 ± 1.49 22.48 ± 0.62 ysis, while biogas composition was monitored daily. Process
Kjeldahl Nitrogen (%)a 2.33 ± 0.30 4.16 ± 0.15 parameters, such as pH, VFA/Alkalinity ratio, VFA and total ammo-
C/N 16.80 ± 0.64 5.4 ± 0.15 nia nitrogen concentration, TS, VS, as well as lignin, cellulose and
pH 8.81 ± 0.45 8.19 ± 0.30
hemicellulose contents in the digestate residue were also mea-
Bulk density (kg/m3) 967.20 ± 14.18 1001.33 ± 17.38
Protein Content (%TS) 14.56 ± 0.30 26.00 ± 0.15 sured to monitor the digestion process. The reactor was mixed
Total COD (gCOD/gVSadded) 0.73 ± 0.02 ND manually (once daily) with the impeller by moving its content to
BMPtheoretical (LCH4/gVSadded) 0.290 ± 0.01 ND the inlet and back to the outlet to minimize stratification in the
ND = Not determined; COD = chemical oxygen demand; C/N = carbon nitrogen reactor. Overall, the experiment lasted for 230 days.
ratio.
a
Dry basis. 2.4. Theoretical BMP of experimental feedstock

The feedstock being solid was prepared according to Zupančič

and Roš (2012) by blending 1 g of the sample with water (dilution
at 1:2 in the reactor. The pH of the mixture was adjusted to 7.47 by factor of 50) to reduce the particle size and allow homogenization
2 M HCl solution. Furthermore, a nutrient solution with composi- (Raposo et al., 2012). Thereafter, theoretical methane potential of
tion according to Angelidaki et al. (2009) was also added into the the substrate used during digestion process was calculated from
reactor. The experiment was started at batch mode until no further the amount and chemical oxygen demand (COD) concentration of
gas production was monitored (40 days of digestion period). After the actual feeding using Eq. (1) (Nielfa et al., 2015) assuming that
that the continuous feeding operation was started with an OLR of the equation is valid for any substances or products (Tarvin &
2.8 gVS/L/d. The experiments continued with gradually increasing Buswell, 1934).
OLRs, thereby decreasing retention times with the aim of main-
taining a fixed working volume. The OLR values investigated were nCH4 RT
BMPthCOD ¼ ð1Þ
2.8 gVS/L/d (OLR 1), 4.2 gVS/L/d (OLR 2) and 6 gVS/L/d (OLR 3) with pVSadded
200 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205

where: conductivity detector (Perkin-Elmer, USA), with an inject tempera-

BMPthCOD = theoretical yield under laboratory condition ture of 150 °C. The carrier gas was nitrogen operated with a flow
R = gas constant (R = 0.082 atm L/mol K) rate of 20 ml/min at 60 °C. A 250-ll pressure-lock gas syringe
T = working temperature (310 K) (VICI, precious sampling Inc., USA) was used for taking samples
P = atmospheric pressure (1 atm) for the gas composition analysis.
VSadded = volatile solids of the substrate added (g) VFAs in the digestate filtrates were measured using a high-
nCH4 = methane produced (mol) determined according to Eq. (2) performance liquid chromatograph (HPLC, Waters 2695, Waters
Corporation, Milford, MA, USA) equipped with an RI detector
COD (Waters 2414, Waters Corporation Milford, MA, USA) and a
nCH4 ¼ g

ð2Þ
64 mol biohydrogen-ion exchange column (Aminex HPX-87H, Bio-Rad,
Hercules, CA, USA) operating at 60 °C. A UV absorbance detector
(Walters 2487), operating at 210 nm wavelength was used in series
2.5. Analytical procedure with a refractive index (RI) detector (Walters 2414) operating at
60 °C.
Moisture content, pH, total nitrogen, TS and VS were deter-
mined according to biomass analytical procedures (APHA-
3. Results and discussion
AWWA-WEF, 2005). Total nitrogen contents were measured using
the Kjeldahl method, and the protein content was estimated by
3.1. Reactor development
multiplying the Kjeldahl nitrogen content with a factor of 6.25
according to Gunaseelan (2009). The total carbon was obtained
During preliminary studies a syphon mechanism was applied
by correcting the total dry weight carbon value for the ash content
initially at the inlet of the reactor to make it gas tight and avoid
(Haug, 1993; Zhou et al., 2015). The fat content was determined
entering of air during feeding. This mechanism worked well at
using the Soxhlet extraction procedure (Carpenter, 2010). Alkalin-
lower TS rates up to 11%. However, it was very troublesome when
ity measured as the total inorganic carbonate and was determined
working with higher TS concentrations; i.e. when feedstock with
by the Nordmann titration method according to Lossie and Pütz
13% TS was applied, this completely blocked the syphon system
(2008). Samples were centrifuged at 4000g for 15 min and then
for the feed. After that, the inlet part of the reactor was modified
5 ml of the supernatant was titrated with 0.1 N sulfuric acid to
to a sealed buffer system described in Section 2.1 and shown in
pH 5, the titration was then continued until pH 4.4 was reached
Fig. 1. The shutoff valve connected at the inlet was always closed
in order to determine the VFA concentration measured as acetic
at the start of feeding to avoid gas leakage. When materials were
acid equivalent. Digestate samples for total ammonia nitrogen con-
fed in, the piston rod was inserted and tightened, to avoid the
centration were centrifuged (15 min at 4000g); supernatant
release of any gas. After this, the valve was opened, and the mate-
diluted 50 times in deionized water to a final volume of 5 ml. Sam-
rial was pressed down into the reactor via the piston rod and then
ples were then analyzed using Ammonium 100 test kit (Nanocolor,
the valve was closed again. The impeller in this reactor allows min-
MACHEREY-NAGEL GmbH & Co. KG. Germany) and concentration
imal mixing by moving the reactor content manually to the inlet
measured using Nanocolor 500D Photometer (MACHEREY-NAGEL
and back to the outlet for better performance of the process. This
GmbH & Co. KG. Germany). The free ammonia concentration was
new modification worked well even with an inlet feed of 22% TS
calculated from the total ammonia nitrogen concentration and
without any blockage at the inlet. The developed plug flow reactor
pH values of samples according to Kayhanian (1999).
worked successfully treating cattle manure bedded with straw at
Total cellulose, hemicellulose and lignin content of the solid
22% TS for 230 days using organic loading rates of 2.8, 4.2, and
fraction of the digestates were determined according to NREL pro-
6 gVS/L/d at retention times of 60, 40 and 28 days respectively.
tocols (Sluiter et al., 2011). The digestate was centrifuged at 4000g
for 15 min and the solid fraction was washed with 100 ml distilled
water and then air dried until moisture content was less than 10%. 3.2. Substrate characterization
The samples were then hydrolyzed using 72% H2SO4 in a water
bath at 30 °C for 60 min, samples were stirred every 5 min to The most important characteristics of the substrate as collected
ensure uniform hydrolysis, and then a second hydrolysis was per- from the cattle farm are shown in Table 1. The substrate contains
formed using 4% H2SO4 in an autoclave at 121 °C for 60 min. Mono- 22.29% TS of which 70.44% are organic matter. The carbon nitrogen
meric sugars contained in the hydrolysis liquid were determined (C/N) ratio was 16.8:1 which is within the required range for stable
by HPLC (Waters 2695, Waters corporation, Milford, MA, USA). anaerobic digestion process. Previously, optimal C/N ratios
Mannose, glucose, galactose, xylose and arabinose were analyzed between 20:1 and 35:1 have been stated (Habiba et al., 2009;
using Aminex HPX-87P column (Bio-Rad) at 85 °C and 0.6 mL/ Kayhanian, 1999), also, other researchers, i.e. Friehe et al. (2010),
min ultrapure water as eluent. Acid soluble lignin (ASL) was deter- Pagés Díaz et al. (2011) reported a wider range of between 10:1
mined using a UV spectrophotometer (Libra S60, Biochrom, Eng- and 30:1 as optimal C/N ratios. However, the availability of carbon
land) at 320 nm. Acid insoluble lignin (AIL) was gravimetrically and nitrogen for the microorganisms is more significant than the
determined as residual solid after hydrolysis corrected with ash actual C:N ratio calculated on the basis of elementary composition
content. The ash content was determined as the remaining residue of the feedstock. Hence the C:N ratio of straw might be overesti-
after keeping the samples in the muffle furnace at 575 °C for 24 h. mated due to the limited availability of carbon fraction in the lig-
The percentage of cellulose degraded was calculated according to nocellulosic structure of straw (Herrmann et al., 2016). The
Zhou et al. (2015). theoretical methane potential of the feedstock used in this exper-
The daily volume of biogas produced was measured by the tip- iment was calculated to be 0.290 LCH4/gVSadded.
ping device in the Automatic Methane Potential Testing System
(AMPTS, Bioprocess control AB, Lund, Sweden), which is based on 3.3. Biogas production
the principle of water displacement buoyancy. The composition
(methane and carbon dioxide) of the produced gas was determined The startup batch digestion period ran for 40 days, the daily gas
using a GC (Perkin-Elmer, USA) equipped with a packed column production was high at the beginning; hence the digestion process
(60  1.800 OD, 80/100, Mesh, Perkin Elmer, USA), and a thermal started up properly and the feedstock were degraded until the end
 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205 201

of this batch digestion period. The methane content in the pro- gradually to around 5 L/d after 18 days, and thereafter it remained
duced gas reached 63% ± 3.86 at the end of the startup (batch) per- stable. The process performed differently when OLR 3, i.e. 6 gVS/L/
iod as shown in Fig. 2b. This start-up period was allowed until gas d was introduced. There were larger fluctuations observed in the
production nearly stopped (40 days) after which the continuous daily gas production and only a slightly stable period could be
digestion mode with feeding and withdrawing in every second achieved after 20 days maintaining a daily biogas production of
day was started. around 7 L/d. The reactor was then fed keeping the OLR 3 (6 gVS/
Fig. 2a illustrates the daily biogas production at different OLR. L/d) for an additional corresponding retention time of 28 days to
Experiments were carried out using progressive organic loading check the stability of the process under this condition. However,
rates; the loading rate of 2.8 gVS/L/d (OLR 1) produced only about the daily gas production could not be kept stable with decreasing
1 L/d in the initial days. Then it gradually increased to around 3 L/d values towards the end of this period. In parallel, the VFA concen-
after 36 days and remained at the same level until the end of the tration continued to increase (Fig. 4a) leading to increased VFA/
OLR1 (2.8 gVS/L/d) period. Similar pattern was observed for the alkalinity ratios with a maximum of around 0.9 at the end of the
period when OLR 2, (4.2 gVS/L/d) was applied, it started from period (Fig. 3a) showing instability in the process. To examine if
3.5 L/d daily methane production initially which increased VFA could be reduced and restore stability in the process, the

Fig. 2. Biogas production (a) and biogas composition (b) at different organic loading rate (OLR) during experiments. The symbols represent daily biogas production (}),
cumulative biogas production (h), methane composition (r) and carbon dioxide composition (j).
202 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205

Fig. 3. pH and VFA/Alkalinity ratio variation (a) and composition of cellulose, hemicellulose and lignin in digestate (b) at different organic loading rate (OLR) during
experiments. The symbols represent pH (}), VFA/Alkalinity ratio (h), total lignin (r), cellulose (j) and hemicellulose (N). Presented values are mean values of duplicate
measurements with error bars as standard deviation between the two values.

OLR was then reduced to 4.2 gVS/L/d (OLR 2) again. As a conse- straw in the cattle manure. Hydrolysis of the cellulose in untreated
quence of the reduced load the biogas production reduced gradu- straw has been reported to be the rate limiting step (Noike et al.,
ally and remained stable after 16 days maintaining a daily biogas 1985) especially when it is mixed with cattle manure (Myint &
production around 4.2 L/d which is slightly lower than that Nirmalakhandan, 2006). However, our result is similar to the
obtained when the same loading rate (OLR 2) was used previously. methane yield of 0.170 LCH4/gVSadded obtained by Kusch et al.
The average methane content in the biogas was 64.9%, 65.1% (2008) when the digestion of horse dung with straw was investi-
and 63.3% for OLR of 2.8, 4.2 and 6 gVS/L/d respectively. Overall, gated in batch-operated solid phase digestion.
OLR up to 4.2 gVS/L/d gave a better process stability with methane
yield of 0.163 LCH4/gVSadded accounting around 56% of the theoret- 3.4. Process performance
ical methane yield of 0.290 LCH4/gVSadded as shown in Table 2. The
methane yield at OLR 2.8 gVS/L/d was slightly lower than the yield Fig. 3a shows the variation of pH and VFA/Alkalinity ratio at dif-
at OLR 4.2 gVS/L/d probably due to slight differences in the feed- ferent OLR while Fig. 4a illustrates the variation of total and indi-
stock composition, since these two experimental periods were per- vidual VFA for all the experiments carried out. At 2.8 gVS/L/d
formed with two different batches of the substrate. This lower (OLR 1), the pH was increased slightly to around 8 initially but
methane yield could be associated with the presence of untreated dropped to around 7.5 after 4 days of digestion and it remains
 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205 203

Fig. 4. Total and individual volatile fatty acids (VFA) variation (a) and Total ammonia nitrogen with free ammonia concentration (b) at different organic loading rate (OLR)
during experiments. The symbols represent Total VFA (}), Acetic (h), Butyric (4), isobutyric (), isovaleric (⁄), propionic (s), valeric (|), Total ammonia nitrogen (r) and free
ammonia (j).

Table 2 Anaerobic digestion process has been reported to be working

Experimental Results at different loading rate (values are averages with standard favourably without acidification possibility when this ratio is less
deviation).
than 0.3 (Drosg, 2013). When the OLR was increased to 4.2 gVS/
Parameter OLR 1 OLR 2 OLR 3 L/d, the pH remained still within the favourable range around 7.6
Loading rate (gVS/d) 13.80 20.70 29.90 throughout this period and the VFA/Alkalinity ratio was below
Organic Loading rate (OLR) 2.80 4.20 6.00 0.3. So, an appropriate buffering capacity and high stability of the
(gVS/L/d) process was observed for OLR of 2.8 and 4.2 gVS/L/d and at corre-
Retention time 60 40 28
sponding retention times of 60 and 40 days.
Biogas produced (L/gVSadded) 0.23 ± 0.01 0.25 ± 0.01 0.23 ± 0.01
Average methane content (%) 64.90 ± 1.75 65.10 ± 2.81 63.30 ± 2.06 No VFA accumulation was observed during the process applying
Methane yield (LCH4/gVSadded) 0.15 ± 0.01 0.163 ± 0.01 0.146 ± 0.01 2.8 gVS/L/d (OLR 1) and 4.2 gVS/L/d (OLR 2) as shown in Fig. 4a.
Initial VS concentration (%) 17.17 17.17 17.17 This signifies the complete degradation of the intermediary pro-
Final VS concentration (%) 4.31 ± 0.64 7.31 ± 1.19 10.10 ± 0.82 duced volatile fatty acids. Fast consumption of VFAs was also
VS removal efficiency (%) 74.87 ± 3.75 57.42 ± 6.96 41.17 ± 4.80
reported by Liu et al. (2007) as a proof of stable and properly func-
tioning process without danger of failure. However, when the OLR
steady afterwards which is within the favourable pH range of 7–8 was increased to 6 gVS/L/d with corresponding decreased reten-
for anaerobic digestion (Drosg, 2013). The VFA/Alkalinity ratio was tion time of 28 days the stability of the system deteriorated. At this
lower (Fig. 3a) than the reported failure limit value of 0.3 which 6 gVS/L/d loading (OLR 3), the process started to show signs for
shows that the process was stable during these conditions. instability already during the first retention time of 28 days,
204 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205

though the pH was still within the favourable range, i.e. an average during the process at all loading rates examined. However, the
of 7.7. However, the VFA/Alkalinity ratio increased to 0.4 and the highest cellulose degradation was obtained at organic loading rate
acetic acid concentration increased to 2100 mg/l which signifies of 2.8 gVS/L/d (OLR 1); an average of 60.9% ± 6.37, while at 4.2 gVS/
a slightly unstable process (Drosg, 2013). There was fluctuations L/d (OLR 2) an average of 48.8% ± 6.73 cellulose degradation was
in the daily biogas production (Fig. 2a) as well at this loading rate observed, and finally 6 gVS/L/d (OLR 3) having the lowest cellulose
(OLR3). Continuing the loading at the same conditions, i.e. 6 gVS/L/ degradation of an average of 30.1% ± 7.05. For the same reason as
d, into the second retention time period caused considerably dis- explained above the amount of cellulose degraded decreases as
turbances in the system. As shown in Fig. 3a, the VFA/Alkalinity the loading rate increases and with reducing retention time.
ratio increased up to 0.9 and the total VFA content also increased
gradually up to around 7000 mg/L (Fig. 4a) which is a typical sign
4. Conclusions
for overloading. The acetic acids increased up to around 3800 mg/L
and the propionic acid concentration increased slightly to around
The new plug flow reactor developed can operate successfully
1000 mg/L during the process. Propionic acid accumulation within
for continuous dry digestion of manure bedded with straw at
the range of 250–1000 mg/L has been reported as an indication of
22% TS when operated at OLR of 2.8 and 4.2 gVS/L/d with retention
instability in the process (Drosg, 2013) and concentrations greater
time of 60 and 40 days respectively. OLR of 6 gVS/L/d and retention
than 1000 mg/L has been reported as a first sign for overloading in
of 28 days favoured process instability decreasing the VS removal
the system (Björnsson et al., 1997). Running the process at a higher
efficiency and cellulose degradation. Digestion of manure bedded
OLR (6 gVS/L/d) led to a shorter retention time of 28 days, since the
with straw without pretreatment at 22% TS was successful,
reactor was kept at a fixed working volume during the whole
methane yield 0.163 LCH4/gVSadded counting around 56% of the
experimental period. This shorter retention time could cause the
theoretical yield with 57% VS removal efficiency was obtained at
removal of the slow-growing methanogenic community, which
organic loading rate of 4.2 gVS/L/d.
led to process instability as well as overloading of the reactor.
Despite the increase in VFA the pH still remained within the
range of 7.6 and 8 during the second retention period of 6 gVS/L/ Acknowledgements
d (OLR3). Weiland (2010) has also reported that manure can have
a surplus of alkalinity which stabilizes the pH value even at higher This work was financially supported by Västra Götalandsregio-
VFA accumulation and as such will not always result in pH drop. To nen (Sweden).
examine if the VFA could be reduced and restore the process back
to stability, the loading rate was reduced again to 4.2 gVS/L/d (OLR References
2). It was observed that the VFA/Alkalinity ratio reduced gradually
to 0.2 (Fig. 3a) and the VFA accumulation reduced drastically Adl, M., Sheng, K.C., Xia, Y.H., Gharibi, A., Chen, X., 2012. Examining a hybrid plug-
(Fig. 4a) and so the process was restored back to stability. flow pilot reactor for anaerobic digestion of farm-based biodegradable solids.
Int. J. Environ. Res. 6 (1), 335–344.
During digestion, at OLR of 2.8 and 4.2 gVS/L/d the total ammo- Angelidaki, I., Alves, M., Bolzonella, D., Borzacconi, L., Campos, J.L., Guwy, A.J.,
nia nitrogen concentration was between 1350 and 2000 mg/l Kalyuzhnyi, S., Jenicek, P., Van Lier, J.B., 2009. Defining the biomethane potential
(Fig. 4b) and the VFA/Alkalinity ratio was below 0.3 which shows (BMP) of solid organic wastes and energy crops: a proposed protocol for batch
assays. Water Sci. Technol. 59 (5), 927–934.
good buffering capacity. However, at OLR of 6 gVS/L/d, the total APHA-AWWA-WEF, 2005. Standard Methods for the Examination of Water and
ammonia nitrogen concentration was slightly reduced to around Wastewater. American Public Health Association (APHA), American Water
1000 mg/l probably due to reduced time for feedstock degradation Works Association (AWWA), Water Environment Federation (WEF),
Washington DC.
and as such the buffering capacity was slightly lower. This together Björnsson, L., Mattiasson, B., Henrysson, T., 1997. Effects of support material on the
with the VFA accumulation resulted to higher VFA/Alkalinity ratio pattern of volatile fatty acid accumulation at overload in anaerobic digestion of
showing instability in the process. The free ammonia concentration semi-solid waste. Appl. Microbiol. Biotechnol. 47 (6), 640–644.
Cantrell, K.B., Ducey, T., Ro, K.S., Hunt, P.G., 2008. Livestock waste-to-bioenergy
throughout the process at different OLR was between 47 and generation opportunities. Bioresour. Technol. 99 (17), 7941–7953.
145 mg/l which shows that the reported tolerable free ammonia Carpenter, C., 2010. Determination of fat content. In: Nielsen, S.S. (Ed.), Food
concentration of 150 mg/l (Kayhanian, 1999) was not exceeded, Analysis Laboratory Manual. Springer, US, pp. 29–37.
Demirer, G.N., Chen, S., 2008. Anaerobic biogasification of undiluted dairy manure
hence the process was not inhibited by free ammonia.
in leaching bed reactors. Waste Manage. 28 (1), 112–119.
Drosg, B., 2013. Process Monitoring in Biogas Plants. IEA Bioenergy 978-1-910154-
3.5. Volatile solids removal efficiency 03-8.
Fernández, J., Pérez, M., Romero, L.I., 2008. Effect of substrate concentration on dry
mesophilic anaerobic digestion of organic fraction of municipal solid waste
An important factor for assessing the efficiency of an anaerobic (OFMSW). Bioresour. Technol. 99 (14), 6075–6080.
digestion process is following up the reduction in VS. As shown in Friehe, J., Weiland, P., Schattauer, A., 2010. Fundamentals of anaerobic digestion. In:
Table 2, 2.8 gVS/L/d (OLR 1) has the highest VS removal efficiency, Guide to Biogas From Production to Use, 5th revised ed., Fachagentur
Nachwachsende Rohstoffe e. V. (FNR).
an average of 74.87 ± 3.75%, while 6 gVS/L/d (OLR 3) has the lowest Gunaseelan, N.V., 2009. Biomass estimates, characteristics, biochemical methane
with an average of 41.17 ± 4.80%. It was observed that as VFAs get- potential, kinetics and energy flow from Jatropha curcus on dry lands. Biomass
ting accumulated the removal efficiency reduces. The removal effi- Bioenergy 33 (4), 589–596.
Habiba, L., Hassib, B., Moktar, H., 2009. Improvement of activated sludge
ciency decreases as the OLR increases, since with increasing OLR stabilisation and filterability during anaerobic digestion by fruit and
the retention time in the digester decreases, so that the organic vegetable waste addition. Bioresour. Technol. 100 (4), 1555–1560.
matter loaded into the digester has not enough time for the degra- Haug, R.T., 1993. Feed conditioning – physical and chemical. In: The Practical
Handbook of Compost Engineering. Taylor & Francis, pp. 247–257.
dation, especially not if difficult to digest fractions, as untreated Herrmann, C., Kalita, N., Wall, D., Xia, A., Murphy, J.D., 2016. Optimised biogas
straw, is present in the feedstock. production from microalgae through co-digestion with carbon-rich co-
substrates. Bioresour. Technol. 214, 328–337.
Karthikeyan, O.P., Visvanathan, C., 2013. Bio-energy recovery from high-solid
3.6. Degradation of lignocelluloses
organic substrates by dry anaerobic bio-conversion processes: a review. Rev.
Environ. Sci. Bio/Technol. 12 (3), 257–284.
As shown in Fig. 3b, the amount of cellulose and hemicellulose Kayhanian, M., 1999. Ammonia inhibition in high-solids biogasification: an
decreases gradually compared to those of the initial composition in overview and practical solutions. Environ. Technol. 20 (4), 355–365.
Kothari, R., Pandey, A.K., Kumar, S., Tyagi, V.V., Tyagi, S.K., 2014. Different aspects of
the feed during the digestion process, while the lignin content dry anaerobic digestion for bio-energy: an overview. Renewable Sustainable
increases. This shows that the carbohydrates were consumed Energy Rev. 39, 174–195.
 R.J. Patinvoh et al. / Bioresource Technology 224 (2017) 197–205 205

Kusch, S., Oechsner, H., Jungbluth, T., 2008. Biogas production with horse dung in Raposo, F., De la Rubia, M.A., Fernández-Cegrí, V., Borja, R., 2012. Anaerobic
solid-phase digestion systems. Bioresour. Technol. 99 (5), 1280–1292. digestion of solid organic substrates in batch mode: an overview relating to
Lansing, S., Martin, J.F., Botero, R.B., da Silva, T.N., da Silva, E.D., 2010. Methane methane yields and experimental procedures. Renewable Sustainable Energy
production in low-cost, unheated, plug-flow digesters treating swine manure Rev. 16 (1), 861–877.
and used cooking grease. Bioresour. Technol. 101 (12), 4362–4370. Sluiter, A., Hames, B., Ruiz, R., Scarlata, C., Sluiter, J., Templeton, D., Crocker, D., 2011.
Lansing, S., Víquez, J., Martínez, H., Botero, R., Martin, J., 2008. Quantifying Determination of Structural Carbohydrates and Lignin in Biomass. Laboratory
electricity generation and waste transformations in a low-cost, plug-flow Analytical Procedure. NREL/TP-510-42618.
anaerobic digestion system. Ecol. Eng. 34 (4), 332–348. Stroot, P.G., McMahon, K.D., Mackie, R.I., Raskin, L., 2001. Anaerobic codigestion of
Liu, G., Zhang, R., Sun, Z., Li, X., Dong, R., 2007. Research progress in anaerobic municipal solid waste and biosolids under various mixing conditions—I.
digestion of high moisture. Agric. Eng. Int. CIGR J. 9 (13), 1–10. Digester performance. Water Res. 35 (7), 1804–1816.
Lossie, U., Pütz, P., 2008. Targeted control of biogas plants with the help of FOS/TAC. Tarvin, D., Buswell, A.M., 1934. The methane fermentation of organic acids and
Practice Report, Laboratory Analysis Titration FOS/TAC. HACH LANGE. carbohydrates 1, 2. J. Am. Chem. Soc. 56 (8), 1751–1755.
Myint, M., Nirmalakhandan, N., 2006. Evaluation of first-order, second-order, and USEPA, 2016. US Environmental Protention Agency. AgSTAR Data and Trends –
surface-limiting reactions in anaerobic hydrolysis of cattle manure. Environ. Anaerobic digester facts and trends <https://www.epa.gov/agstar/agstar-data-
Eng. Sci. 23 (6), 970–980. and-trends#adfacts> Accessed 16/10/18.
Nielfa, A., Cano, R., Fdz-Polanco, M., 2015. Theoretical methane production Weiland, P., 2010. Biogas production: current state and perspectives. Appl.
generated by the co-digestion of organic fraction municipal solid waste and Microbiol. Biotechnol. 85 (4), 849–860.
biological sludge. Biotechnol. Rep. 5, 14–21. Zhou, C., Liu, Z., Huang, Z.-L., Dong, M., Yu, X.-L., Ning, P., 2015. A new strategy for
Noike, T., Endo, G., Chang, J.E., 1985. Characteristics of carbohydrate degradation co-composting dairy manure with rice straw: addition of different inocula at
and the rate-limiting step in anaerobic digestion. Biotechnol. Bioeng. 27 (10), three stages of composting. Waste Manage. 40, 38–43.
1482–1489. Zupančič, G.D., Roš, M., 2012. Determination of chemical oxygen demand in
Pagés Díaz, J., Pereda Reyes, I., Lundin, M., Sárvári Horváth, I., 2011. Co-digestion of substrates from anaerobic treatment of solid organic waste. Waste Biomass
different waste mixtures from agro-industrial activities: kinetic evaluation and Valor. 3, 89–98.
synergetic effects. Bioresour. Technol. 102 (23), 10834–10840.